Neurotransmitters that are made, packaged, and stored in presynaptic compartments are functionally inert, their precisely controlled release and delivery for juxtacrine activation a signature property of synaptic signaling. In order to investigate whether the release of Hh might be regulated at cytoneme synapses, we analyzed Hh signaling in genotypes that express different amounts of Hh. We tested whether amounts of Hh and Hh signaling in recipient cells are proportional to Hh production, as might be expected of constitutive, unregulated release by producing cells, or if they are independent of production as might be expected of regulated release and delivery.

We first monitored Hh signaling in wing discs with genotypes that vary the number of wildtype (WT) hh genes and hh transgenes (Figure 1B). The hh transgenes were (1) BAC plasmids containing the WT hh transcription unit in a genomic fragment of 40 kb (HS) or 101 kb (HL), or HS-GFP, the 40 kb genomic fragment into which GFP has been recombined in frame (Hh:GFP) (Chen et al., 2017); and (2) a Hh:GFP construct expressed under GAL4 control. Flies without a functional hh gene die as embryos, but haploid flies with one BAC transgene (encoding Hh [HS, HL, or HS-GFP]) had normal appearance and wing discs had normal morphology (Chen et al., 2017). Hh:GFP encoded by this transgene is therefore presumed to be a functional surrogate for the normal, WT protein.

To investigate how the production of hh RNA and Hh protein in the wing disc scale with gene dosage, we measured amounts of hh RNA by qPCR and amounts of Hh protein by monitoring α-Hh antibody staining (Figure 1C). Genotypes with 1 (1× WT), 2 (2× WT), 3 (2× WT, 1× BAC), or 4 (2× WT, 2× BAC) hh genes had amounts of hh RNA in wing discs which increased with gene dosage and scaled linearly. Hh protein in the P compartment (Hh-producing) cells of the wing blade primordium increased 2.9× in genotypes with 3 hh genes versus 1. These results show that in this range of gene copy number, both hh RNA and Hh protein are produced in direct proportion to gene dosage. We also measured hh RNA in wing discs with 2 WT hh genes and transgenes containing hh-Gal4 and UAS-hh:GFP, and observed that the amount increased approximately 15× over WT.

To determine if different amounts of Hh expression change signaling, growth, and patterning, we examined several parameters which are sensitive to Hh signaling: expression of a gene targets of Hh signal transduction, and size and shape of the wing and wing disc. Expression of the ptc gene in the wing disc is upregulated by Hh signaling in a band of cells at the A/P compartment border. The width of this band decreases under conditions of reduced Hh signal transduction (Molnar et al., 2011), and increases under conditions in which Hh signaling is elevated (Cheng et al., 2012; Wang and Holmgren, 1999). As shown in Figure 1D,E, the size of the Ptc band did not change in discs with 1, 2, 3, or 4 hh gene copies, despite the differences in hh RNA and Hh protein amounts (Figure 1D,E, Figure 1—figure supplement 1 and 2). In contrast, Hh overexpression driven by hh-Gal4 changed both the size and shape of the Ptc band. Expression of Knot (Kn), which like Ptc expression is regulated by Hh signaling (Vervoort et al., 1999), did not change in genotypes with 1–4 hh gene copies (Figure 1—figure supplement 3).

We monitored wing size in the different genotypes by measuring the length (A/P border length from wing tip to wing base) and width (length of a line extending the posterior crossvein from anterior to posterior margin). These wing dimensions were not dependent on hh gene copy or expression (Figure 1F,G). We monitored wing shape by comparing the length/width ratios of the different genotypes, and determined that only the hh-Gal4-driven (~15×) overexpression changed it significantly (Figure 1G). L3 wing discs were also similar in appearance except for the hhGal4 genotype, which had an outgrowth anterior to the wing blade primordium (Figure 1H). The only apparent change we noted to the adult morphology was the decrease in length/width ratio.

These findings indicate that the processes of production, maturation, secretion, transport, and uptake that disperses Hh in the wing disc generates a distribution in targets cells that is insensitive to changes in Hh amounts that 1–4 hh gene copies generate. The sensitivity to an excess of ~15-fold normal Hh generated by hh-Gal4 reveals that the capacity of the process which buffers against changes in amounts of Hh production is limited. We do not know which step or steps in the process might be overwhelmed by the ~15× excess—if, for instance, Hh released in this overexpression condition is processed and modified normally, or if overexpressed Hh exits by the normal route. The important point is that there is a system that can compensate for different amounts of production. Previous studies characterized the robustness of Hh signaling to variations in either production or response, assuming that release from producing cells is constitutive and that robustness is solely an attribute of the signal transduction process (Li et al., 2018; Zhang et al., 2020). We consider two possible alternatives.

If the amount of Hh taken up by the recipient, target cells are proportional to production, each recipient cell might scale the outputs of Hh signal transduction relative to its neighbors. This mechanism might adjust relative responses independently of absolute amounts, determining growth and pattern by the slope of the concentration gradient across a field of cells. This type of mechanism was proposed for the morphogen gradient of Dpp in order to model the effects of mosaic ectopic activation induced by the expression of a constitutively active Dpp receptor (Rogulja and Irvine, 2005). Alternatively, the amount of Hh transferred from producing to responding cells might be regulated independently of the amount produced. To distinguish between these mechanisms, we quantified Hh in recipient cells of the wing blade primordium.

Quantification of Hh detected by antibody staining in the entire A and P compartments of the wing primordium of normal discs (2× WT) showed that approximately 5.2% was in the A compartment (n=7; standard deviation [SD]=1.2%). We next monitored Hh with α-Hh antibody in the small rectangular region composed of equal portions of the A and P compartments (Figure 1A) in genotypes with 1, 2, 3, or 4 hh genes. As expected, Hh amounts were proportional to gene dosage in the P compartment portion of the rectangular region. The cells in this region produce Hh and this result is consistent with the analysis of the entire P compartment (described above) and the finding that Hh RNA and protein scale with gene dosage. Analysis of the anterior portion of the rectangular region revealed that Hh amounts were not detectably different in genotypes with 1, 2, 3, or 4 hh genes (Figure 2A,B, Figure 2—figure supplement 1). In sum, these results are consistent with the idea that most Hh produced in the P compartment is not taken up or retained by A compartment cells, and that Hh uptake is not linked directly to production. We also examined the intracellular localization of Hh to determine if it colocalized with Rab7, which concentrates in late endosomes. Association with Rab7 would suggest that a fraction of Hh produced in P compartment cells is destined for turnover, not release (Figure 1—figure supplement 4).

To characterize the relationship between Hh production and delivery further, we used an α-GFP antibody to analyze genotypes with one Hh:GFP-encoding BAC transgene (BHS-GFP) together with either zero, one, or two (untagged) WT hh genes (Total genes: 1: BH-GFP; −/−; 2: BHS-GFP; +/−; and 3: BHS-GFP; +/+) (Figure 1B). As depicted in Figure 2C, α-GFP antibody staining of GFP-tagged Hh that is titrated with different amounts of untagged Hh distinguishes between constitutive and regulated delivery in these genotypes. If delivery of Hh to the A compartment is proportional to gene dosage and not regulated, Hh:GFP amounts in the A compartment are expected to be unaffected by co-production of untagged Hh, and Hh:GFP remains constant as gene dosage and production increases. However, if delivery is regulated, the fraction of Hh:GFP in the A compartment is expected to decrease as the fraction of untagged Hh increases in proportion to total gene copy and production.

Analysis of wing discs with one Hh:GFP BAC (BHS-GFP) and zero, one, or two WT hh genes shows that Hh:GFP in the producing cells of the P compartment was not diminished by the presence of hh genes that encode untagged Hh (Figure 2D,E). This is consistent with the idea that the production of both Hh RNA and protein are proportional to gene copy. In contrast, Hh:GFP amounts in the Hh-receiving cells of the A compartment decreased in proportion to number of hh genes that encode untagged Hh. This shows that Hh:GFP was diluted by the presence of untagged Hh, a result which is consistent with the amounts of Hh we detected in the A compartment with α-Hh antibody in genotypes with 1, 2, 3, or 4 hh genes (Figure 2A,B). We conclude that the amount of Hh in the A compartment was constant and did not scale with production.

We applied an extracellular staining protocol that detects antibodies bound to preparations of non-permeabilized and unfixed cells (Strigini and Cohen, 2000). We stained wing discs with a high titer α-GFP antibody that detected BAC-encoded Hh:GFP, and analyzed genotypes with 1 or 3 genes that encode untagged Hh. For reference, total GFP fluorescence was determined in discs processed with our standard immunohistochemistry protocol that includes detergent permeabilization and formaldehyde fixation. GFP fluorescence increased in P compartment cells and decreased in A compartment cells in proportion to hh gene copy (Figure 2E). The extracellular staining protocol detected basolateral GFP fluorescence (Figure 2—figure supplement 2), consistent with the observations of Callejo et al., 2011 who reported that Hh moves to the basolateral compartment prior to localization to basolateral cytonemes and export (Callejo et al., 2011). Hh:GFP detected by the extracellular staining protocol was less than the total, as the fluorescence was visible only in the most basal optical sections and required higher laser power and gain settings (see Materials and methods), but direct quantitative comparisons are not possible because of the different protocols. Extracellular staining decreased in both P and A compartments in the presence of untagged Hh (Figure 2F), indicating that Hh exposed on the exterior of P compartment cells is gated. Although the results suggest that Hh taken up by A compartment cells represents the population of externalized Hh present on the surface of P compartment cells, they do not reveal whether the release of externalized Hh is regulated.

To investigate whether regulated delivery is also a feature of Dpp signaling, we monitored Dpp signaling and dispersion in genotypes with different numbers of dpp genes. We created a Dpp-encoding BAC transgene (BD) that rescues dpp haploinsufficiency: animals with one WT dpp and one BD (+/dpp^H46; +/BD) allele are viable and their wing size is comparable to WT flies (Figure 3A,B), indicating that the Dpp BAC is a functional substitute for a WT dpp gene. To monitor different amounts of dpp expression, we compared wing discs with two or four dpp genes (two genes: +/+; four genes: +/+; BD/BD). First, to examine proportionality between dpp gene copy and Dpp protein, we stained wing discs with antibody that recognizes the prodomain of unprocessed Dpp (Akiyama and Gibson, 2015; Panganiban et al., 1990). Staining in cells that produce Dpp was approximately double in the four copy genotype compared to the two copy genotype (Figure 3C,D). Second, we examined wing size, which is sensitive to different amounts of Dpp signaling. Mutant conditions that decrease Dpp signal transduction reduce wing disc growth and mutant conditions that elevate signal transduction cause overgrowth (Capdevila and Guerrero, 1994; Spencer et al., 1982). We found that wing size did not differ between genotypes with two or four dpp genes (Figure 3A,B). Third, we asked if signal transduction increases with gene dosage and Dpp production. α-pMAD staining, a readout of Dpp signaling, forms a band that coincides with and straddles dpp expressing cells in WT discs. The width of the pMAD-staining band was not different in wing discs with two or four gene copies (Figure 3E,F), indicating that increased Dpp production does not increase signal transduction. Similarly, expression of Spalt (Salm), which is regulated by Dpp signaling in the wing disc (de Celis et al., 1996), did not change in genotypes with two or four dpp gene copies (Figure 3—figure supplement 1). In sum, these results show that Dpp signaling in the wing disc is insensitive to increased levels of Dpp production.

To monitor Dpp distributions, we examined discs stained with α-Dpp antibody to compare genotypes with one Dpp:Cherry knock-in allele (generated by CRISPR-mediated recombination) and either one or three dpp genes that encode untagged protein (Figure 3G). The experimental setup and rationale are similar to the analysis of Hh:GFP depicted in Figure 2C. Evaluation of the two genotypes showed that levels of Dpp:Cherry fluorescence in producing cells did not change with the production of untagged Dpp (Figure 3H,I), and that levels of Dpp:Cherry fluorescence in non-producing, receiving cells decreased in proportion to the number of genes that encode untagged Dpp (Figure 3I’). This finding, that the amount of Dpp:Cherry in target cells decreased as the ratio of tagged:untagged Dpp declined, is consistent with the idea that transmission of Dpp to targets is regulated.

We investigated Wg dispersion by analyzing three genotypes with different numbers of functional wg genes: 1 (wg⁺/wg⁻), 2 (wg⁺/wg⁺), and 2^{+overexpression} (wg-Gal4 UAS-Wg:GFP; wg⁺/wg⁺), and monitoring expression and distribution of Wg, as well as the Wg gene targets senseless (sens) and Distal-less (Dll). α-Wg antibody staining showed that Wg production is proportional to gene copy number in discs with one and two wg genes, and that the wg-Gal4 driver generated approximately seven times more Wg than a single endogenous gene (Figure 4A,A’,B). Despite the differences in expression between these genotypes, the amount of Wg in the neighboring cells that received Wg was unchanged (Figure 4A,A’,B). The fraction of Wg present in the neighboring cells relative to the total produced in the wing blade decreased with increasing functional gene dosage, from approximately 41% (one copy) to 24% (two copies), and 3.5% (seven functional equivalents). α-Sens antibody detects two narrow, parallel stripes of expression that are immediately adjacent to but do not overlap the Wg-expressing cells (Nolo et al., 2000), and the patterns of Sens expression were not detectably different in these genotypes (Figure 4A”,C,C’). α-Dll antibody detects a region of expression that overlaps Wg-expressing cells (Zecca et al., 1996), and the pattern of Dll expression also was not different in these genotypes (Figure 4—figure supplement 1). These findings indicate that Wg transmission to target cells is regulated.

We investigated the distribution of Wg using the extracellular staining protocol and detected low levels of basolateral Wg in Wg-producing (Figure 4—figure supplement 1). Similar to Hh (Callejo et al., 2011), Wg moves to the apical surface before relocating to the basolateral surface (Yamazaki et al., 2016), and the small fraction of basolateral, extracellular Wg suggests that insertion in the basolateral membrane is gated. To determine if this portion of the Wg population is sensitive to Wg production amounts, we examined discs that expressed Wg from one or two wg genes, or from two wg genes combined with wgGal4 driven overexpression. We found that extracellular Wg of both producing and receiving cells increased with gene dosage and expression (Figure 4D). This protocol does not distinguish between protein on either the cell surface or associated with cytonemes, but these results suggest that delivery and release to receiving cells from the basolateral membrane is rate-limiting.

Previous studies in several different contexts showed that the number of cytonemes correlates positively with signal transduction activity. Whereas cells with low signaling activity have few cytonemes, cells with higher levels of signaling have more (Bischoff et al., 2013; Chen et al., 2017; Du et al., 2018; González-Méndez et al., 2017; Huang et al., 2019; Huang and Kornberg, 2016; Mattes et al., 2018; Roy et al., 2011; Roy et al., 2014). This correlation also holds for mutant conditions that change cytoneme numbers or signaling: an example is overexpression of Ihog which increases cytoneme stability and Hh signaling in the wing disc (González-Méndez et al., 2017).

We first analyzed Hh signaling and cytoneme densities in the air sac primordium (ASP), a tracheal branch that is physically attached to the wing disc and which extends cytonemes to the disc that mediate the uptake of Hh, Dpp, and Bnl (Chen et al., 2017; Du et al., 2018; Hatori and Kornberg, 2020; Roy et al., 2014). To investigate the relationship between cytonemes and Hh production amounts, we monitored ASP cytonemes in genotypes with different numbers of hh genes.

We first asked if delivery of Hh to ASP cells is sensitive to amounts of Hh production by monitoring two conditions that are dependent on Hh signaling in the ASP: tissue morphology and expression of engrailed (en), which is a transcriptional target that is induced by Hh signaling in the wing blade and ASP (Guillen et al., 1995; Hatori and Kornberg, 2020). In WT, the ASP has a proximal narrow stalk and distal bulb (Figure 1A), and en expression is graded, with highest levels in the tip cells (Figure 5A,B). In mutant conditions with elevated Hh signaling (e.g., ectopic overexpression of Hh in the ASP), the stalk was absent and En expression extended to more proximal tracheal cells (i.e., the transverse connective; Figure 1A), whereas in mutant conditions with reduced Hh signaling (e.g., smoothened loss-of-function and Patched overexpression), the stalk was elongated and En expression was reduced. We found that in genotypes with 1–4 hh genes, neither ASP morphology nor extent of En expression changed (Figure 5A–C). In contrast, Hh overexpression driven by hh-Gal4 reduced the stalk and increased the extent of En expression. These results show that the ASP is insensitive to the different amounts of Hh produced by 1–4 gene copies, and are consistent with the idea that Hh delivery is regulated. The sensitivity to hh-Gal4 driven overexpression indicates that the capacity of the ASP system to buffer against different levels of expression is limited.

We next investigated the relationship between Hh production and ASP cytonemes. We analyzed the number of cytonemes in genotypes with 1, 2, 3, and 4 hh genes by marking ASP cytonemes with membrane-tethered Cherry (btl>CD8:Cherry). Cytonemes that extend from the distal tip of the ASP take up Hh and contain Ptc (Chen et al., 2017), and in the experimental genotypes with 1–4 hh genes, the number of distal tip cytonemes did not change (Figure 5D,E). We conclude that the number of cytonemes and amount of cytoneme-mediated Hh uptake are insensitive to conditions that reduce or increase Hh production by a factor of 2 relative to WT.

We also investigated the effects of Ihog over-expression, which stabilizes cytonemes. The extent of En expression in the ASP increased under conditions of Ihog over-expression (btlGal4 UAS-Ihog), consistent with the idea that stabilized cytonemes increased Hh uptake and signaling (Figure 5F). In the wing disc, Ihog over-expression in the ptc domain at the A/P compartment border (ptcGal4 UAS-Ihog) increased both the width of the Ptc band (compare Figure 1E, Figure 5G) and sensitivity to amount of Hh production. In genotypes with 1, 2 or 3 hh genes, the width of the Ptc band in discs with two copies increased approximately 35% relative to discs with one (ANOVA and Tukey; p<0.05), but no significant difference between discs with two or three copies (Figure 5G). These results show both that genetic conditions that increase cytoneme numbers also increase Hh signaling of wing disc cells, and that cells that overexpress Ihog have a limited capacity to respond to increases in Hh.

Morphogen signaling is a multi-step process that involves post-translational processes that prepare Hh, Dpp, and Wg in producing cells, feedback regulation in receiving cells, and extracellular proteins that influence activity. We investigated whether changes in the production of Hh, Dpp, or Wg affect the expression of genes that encode functions known to modulate signaling, because the expression of these genes might provide feedback regulation that compensates for changes in the amounts of proteins that are released or taken up. We might expect, for example, that the expression of a gene that provides negative feedback increases under conditions of increased signaling protein production. Shifted (Shf), for instance, encodes an extracellular factor that is required for the normal distribution of Hh, and Shf protein levels decrease in conditions of lowered signaling (Glise et al., 2005). We quantified shf expression in discs with one and four hh genes by quantifying shf transcripts with qPCR; no change in shf mRNA was detected in these genotypes (Figure 6A,B). This insensitivity to the tested changes in Hh amounts suggests that Shf does not control Hh release. brinker (brk), pentagone (Pent; aka magu), short gastrulation (sog), and crossveinless-2 (Cv-2) negatively affect Dpp signaling. Brk is a transcriptional repressor of Dpp signal transduction whose expression is suppressed by Dpp. Pent, Sog, and cv-2 encode extracellular proteins that bind Dpp and negatively affect spread and signaling (Raftery and Umulis, 2012; Figure 6A). Ectopic Dpp signaling suppresses brk, pent, and sog and upregulates cv-2 expression (Raftery and Umulis, 2012; Yu et al., 1996). qPCR analysis detected no changes to expression of brk, pent, sog, or cv-2 in genotypes with two or four dpp genes (Figure 6B). Notum expression is induced by Wg signaling and encodes an extracellular deacylase of Wg that inhibits Wg signaling (Minami et al., 1999), but its expression is not influenced by changes to Wg gene number (Figure 6B). In sum, these data do not support the idea that expression of known modulators of the Hh, Dpp, and Wg pathways compensate for changes in amounts of morphogen production, and are consistent with the idea that release is regulated.

To characterize how signaling proteins are apportioned among the cells they target, we analyzed Hh signaling in a uniquely positioned group of Hh-responding cells near the Hh-producing cells of the wing disc notum primordium. Cells in this region include cells of the wing disc A compartment, the ASP and myoblasts, mesenchymal cells that cover most of the disc A compartment in this region and extend over a portion of the P compartment as well. These myoblasts will develop into the flight muscles in the adult. Because of the close proximity of these cells to Hh-producing cells in the disc, and because no other Hh-producing cells are as close, we presume that the Hh they receive originates from this one source (Hatori and Kornberg, 2020). We designated an area 150 μm×150 μm that includes all the Hh-responding cells in the notum, ASP, and myoblasts, as a Hh ‘microenvironment’ (Figure 7A–C).

We investigated the behavior of Hh in the three cell populations of this microenvironment, testing if signaling in one tissue is influenced by the amount of Hh the others take up. The experiment distinguishes whether regulated Hh export creates a common pool of signaling protein that is shared among target cells, or if export is independently directed to target cells. If uptake is from a common pool, reducing the number of target cells is predicted to increase uptake and signaling in the target cells that remain. Experiments that increased amounts of extracellular and diffusible morphogens have shown that target cells are capable of responding to increases that are available for uptake. Examples include overexpression of a nonlipidated form of Hh (HhN) in the wing disc and addition of FGF-soaked beads in the chick limb bud; in both experiments, increased signaling in target cells was observed (Callejo et al., 2006; Cohn et al., 1995).

We genetically ablated the ASP and myoblasts (Figure 7D,E), and monitored Ptc expression as a readout of Hh signaling in the remaining tissues (Figure 7F–I). The ASP does not develop when tracheal cells overexpress Btl^DN (Du et al., 2018; Sato and Kornberg, 2002), a dominant-negative mutant FGFR protein (Reichman-Fried and Shilo, 1995) that inhibits signaling by disc-produced FGF ligand Branchless (Bnl) (Sato and Kornberg, 2002). Tracheal overexpression of Cut, a transcription factor that negatively regulates FGF signaling (Du et al., 2018; Pitsouli and Perrimon, 2013), also ablates the ASP (Figure 7—figure supplement 1A). Neither the presence of Btl^DN or overexpressed Cut in the tracheal cells had an apparent effect on the growth and morphogenesis of the disc or myoblasts (Figure 7—figure supplement 1A,B). In the absence of an ASP (and of the Hh target cells in the ASP), the amounts of Ptc in the disc and myoblasts were indistinguishable from controls (Figure 7F,G, Figure 7—figure supplement 1C). To ablate myoblasts, we ectopically expressed moleskin (msk) RNAi. Msk is a nuclear importer of the FGF transcriptional activator ERK (Vishal et al., 2017). mskRNAi expression in the myoblasts had no apparent effect on the wing disc (Figure 7—figure supplement 1E), but reduced the myoblast population and decreased the size of the ASP (Figure 7—figure supplement 1F). This ASP phenotype is consistent with our previous findings that ASP growth and morphogenesis are dependent on Notch signaling from the myoblasts (Huang and Kornberg, 2015). Whereas the total amount of Ptc in the reduced population of myoblasts decreased under conditions of msk expression, Ptc expression in the disc was indistinguishable from controls (Figure 7H,I, Figure 7—figure supplement 1G).

These experiments and the results are summarized in Figure 8A–H. The genotypes we tested eliminate either the ASP or myoblasts (Figure 8A–F), and showed that Hh signaling in the disc and myoblasts was not dependent on Hh uptake by the ASP (Figure 8G). In addition, despite the fact that the Hh-responding myoblasts in the microenvironment represent an area 20% greater than the area of Hh-responding disc cells at the stage these experiments were conducted, Hh signaling in the disc was not dependent on Hh uptake by myoblasts (Figure 8H). We conclude that the Hh target cells in the ASP and myoblasts do not change the delivery of Hh to wing disc cells, and therefore that the spatial patterns of Hh signaling and Hh transport in the disc form independently of the presence or absence of other target cells.

Flies were cultured in standard cornmeal and agar medium at 25°C; all crosses were at 25°C, except the expression of Cut. To express Cut in the ASP, btl-Gal4/UAS-Cut; Gal80^ts/+ was incubated at 18°C until early L3 and transferred to 29°C until late L3.

Wing discs were dissected in phosphate-buffered saline (PBS), RNA was extracted using RNeasy Micro Kit (Qiagen), and cDNA was synthesized using the High Capacity RNA-to-cDNA Kit (Applied Biosystems). qPCR was performed with SensiFast Sybr green (Bioline). For each genotype, 3–4 replicates of 5 wing discs were analyzed. Actin was the internal control and fold differences in relative mRNA levels between genotypes were calculated as 2^−ΔΔCt.

Wing discs together with Tr2 trachea were dissected in PBS, fixed in 4% formaldehyde in PBS (25 min), and washed on a rocking rotator in PBS+0.3% Triton X-100 (PBST) (3×10 min), and incubated in Roche blocking solution (1 hr). Primary antibodies were diluted in Roche Blocking solution and incubated with discs overnight at 4°C (12 hr/overnight). Discs were washed in PBST (3×10 min) on a rocking rotator, incubated with secondary antibody diluted in Roche blocking solution (2 hr), washed in PBST (3×10 min), and after removal of PBST, mounted in Vectashield (Vector Labs). All procedures were at room temperature except for primary antibody incubation. The domains of α-Ptc, α-Sense, α-En, and α-pMAD staining were measured in ImageJ from single optical sections at the basolateral part of the wing disc.

Average intensity quantifications for projections of α-Hh, α-GFP, α-Cherry, and α-Wg staining were calculated for segments of optical sections spanning 20 μm from the most apical side of the wing pouch cells. This segment was chosen because the basal sides of the wing discs are folded and therefore problematic to quantify. Background measurements were taken in equivalent areas distant from staining regions and were subtracted.

The comparison between total and extracellular Hh is a relative estimate because fluors, antibodies, laser intensities, and gain settings were not the same for the two protocols. Wg detected by the extracellular staining protocol in producing and adjacent receiving cells was measured in the indicated 30 μm×90 μm area, with the producing area defined by a 30 μm×10 μm rectangle and the receiving area defined by two 30 μm×40 μm rectangles. Distance of fluorescence was measured manually. Size differences between discs were small and were not taken into account. All measurements of intensity were with ImageJ.

For Figure 8G, H, K, the dotted lines for the predicted extracellular pool were calculated assuming that the total area of Ptc expression is constant and an increase in remaining tissues compensates for the absence of ablated tissue. The ratio for directed release was set at 1.00 based on the assumption that Ptc-expressing area for the remaining tissues would not change under conditions of tissue ablation.

To observe cytonemes, unfixed preparations were observed using the hanging drop method (Huang and Kornberg, 2016). Images were acquired using the FV3000 Olympus Confocal microscope with GaAsP PMT detectors. Images were analyzed and processed with ImageJ and Photoshop.

Maximum intensity projection image of the whole volume of the ASP was used to count the number of cytonemes in the bulb. To calculate the density of cytoneme per µm, the number of cytonemes was divided by the perimeter of the bulb of the ASP.

As adopted from Torroja et al., 2004: third Instar larval wing discs were dissected and incubated in 3.7 mM Red dextran (lysine fixable, MW 3000, Molecular Probe) in M3 Media at 25°C (5 min), washed 5× in ice-cold M3 media (2 min each), fixed in 4% paraformaldehyde/PBS (40 min at 4°C), fixed in 4% PFA/PBS at room temperature (20 min), and washed 2× in PBT (10 min each). α-GFP staining was as above.

Error bars indicate SD and statistical significance was calculated by unpaired Student’s t-tests. P value of <0.05 is designated as statistically significant.