Cholesterol-modification of Hh is necessary for Hh-dependent graded cell fate specification in the dorsal epidermis. All panels show dorsal cuticle views of first instar larvae. (A) Wild type. (B) hh^AC. (C) hh^AC, engal4 UAS-hh-Np. (D) hh^AC, engal4 UAS-hh-N-CD2. (E) hh^AC, engal4 UAS-hh-N. (F) ptcgal4 UAS-Shi^DN. (G) disp glc. (H) disp glc, enGal4 UAS-hh-Np and (I) disp glc, enGal4 UAS-hh-N. (A) In the wild-type dorsal epidermis, Hh is secreted from type 1 cells and forms a morphogen gradient that allows the differentiation of cell types 1 to 3 in a concentration-dependent manner (scheme and cuticle). (B) In hh loss of function, only type 4 cells are present. (C) Expression of Hh-Np in En cells allows the rescue of the hh mutant phenotype. (D,E) Hh-N or Hh-N-CD2 expression induces some type 1 cells and a domain of type 2 cuticle, but in the case of Hh-N-CD2 naked cuticle was never formed at the midline (broken line in D). (F) In ptcgal4 UAS-Shi^DN larvae, the domain of type 2 cells is reduced, whereas the domain of type 3 cells is enlarged. (G) In disp glc loss of function, only type 4 cells are present. (H,I) In disp glc embryos, Hh-Np only induced some patches of type 1 cells (arrow in H), whereas Hh-N induced type 1 and some type 2 cells. Bars delimit the width of type 2 cuticle. (J) Histogram showing the width of the cuticle covered by type 2 and 3 cells in the genotypes shown in A,C-F. (K) A theoretical model based on the results shown in J, representing cell fate behaviours as a function of Hh range and activity. Numbers correspond to the type of fate adopted by the dorsal cells related to the distance from Hh-producing cells.

Cholesterol modification is required for membrane association and full activity of Hh. (A,B) Cl8 cells were treated with Cd to induce expression of either Hh-Np (top) or Hh-N (bottom). Permeabilised cells (A) or unpermeabilised cells (B) were subjected to immunostaining for Hh (green) and for the ER marker BiP (red). Hh-N and Hh-Np are present in large cytoplasmic vesicles marked with BiP (A). Hh-Np is also present at the plasma membrane (A). In non-permeabilised cells (B), Hh-N was undetectable, while Hh-Np was enriched at the plasma membrane as aggregates (arrows). Even without Cd induction, Hh accretions were faintly detectable because of the leaky promoter. (C, top) Media of cultured Cl8 cells expressing either Hh-N or Hh-Np were subjected to gel filtration on a Sephacryl S200 column. Hh-N is recovered only in the low molecular weight fractions corresponding to the Hh monomeric form (peak B). Hh-Np was greatly enriched in the high molecular weight fractions (peak A around 160 kDa). (C, bottom) When subjected to a Superose 6 column, a stable peak of Hh-Np was recovered around 200 kDa. Immunoreactivity was also observed in the highest molecular weight fractions. (D) Peak A- or B-containing fractions were applied to naive Cl8 cells that were subsequently fixed and stained for Hh. Hh-N was not detectable, but accretions of Hh-Np were present at the plasma membrane of peak A-treated cells (arrow). (E) Both peaks were tested for their ability to induce a Hh reporter gene. At similar concentrations, Hh-Np (peak A) is threefold more potent than Hh-N (peak B). (F) When peak A is added to naive cells, it stimulates Fu phosphorylation two to three times faster than peak B (the upper arrow on each blot indicates the phosphorylated Fu isoform). (G) Embryonic extracts were subjected to gel filtration on a Superose 6 column: a stable peak of Hh is visible around 200 kDa, but Hh is also present in the very high molecular weight fractions. The last fraction falls outside of the resolution of the column. (C-G) Representative images and blots from three independent experiments.

Cholesterol-modification is required for Hh homodimerisation. Hh-Np- or Hh-N S2-expressing cells were transfected with an HA-tagged form of Hh-Np (Hh-HA-Np). The presence of the Hh proteins in the culture media or in cell lysates was analysed by western blotting (upper panel). The three different Hh forms are easily distinguishable because of their different electromobility properties. Immunoprecipitation using an anti-HA antibody (lower panel) revealed that Hh-HA-Np co-precipitates with Hh-Np both in cell lysates and in culture media, but does not co-precipitate with Hh-N. The HA adduct imposes a low electromobility onto Hh, whereas the cholesterol adduct in Hh-Np is responsible for its higher electromobility when compared with Hh-N.

Disp is required for the apical Hh-Np LPS formation in the wing imaginal disc. (A-C) Wild type, (D-F) disp, (G-I) yw hs-flp; enGAL4/UAS-Hh-Np; FRT82 hh^AC/FRT82 tubGAL80 and (J-L) yw hs-flp; enGAL4/UAS-Hh-N; FRT82 hh^AC/FRT82 tubGAL80 imaginal wing discs stained for Ptc (red in A-C,G-L) or Arm (red in F) and Hh (green). (C,F,I,L) Confocal z-sections corresponding to the level of the broken line in A,B,H,K. (H,K) Enlargements of the discs shown in G,J, respectively. (A-C) In wild-type discs, Hh-Np is secreted from the P compartment and forms apically located accretions (thin arrows in A and C) that can also be observed in anterior cells. In the first two or three rows of A-responding cells, larger intracellular accretions containing Hh and Ptc can be observed (thick arrows in B and C). (D-F) None of these accretions is observable in Disp mutant discs. Posterior hh-null clones expressing Hh-Np and abutting the AP border stimulate Ptc overexpression across a six- or seven-cell wide area (G-I), while clones of similar size expressing Hh-N activate Ptc in only three or four rows of cells (J-L). In both cases, small clones also appeared in the A compartment close to the AP boundary (arrowheads in H and K), most probably owing to engal4 expression in this domain at late larval stages. Although in hh^null UAS-hh-Np expressing clones Hh is enriched at the apical side (I) [as is endogenous Hh-Np (C)], in hh^null UAS-hh-N expressing clones, Hh is enriched basally (L). In the latter case, we observed very basal Ptc-Hh accretions (thick arrows in L), whereas in the former case Ptc-Hh vesicles are concentrated more apically in the receiving cells (thick arrows in C,I). In all cases, Hh-Np LPSs are observable, probably owing to the surrounding wild-type cells (thin arrows in I and L).

Short-range activity of Hh-N in columnar cells.dpp-lacZ (red) expression in wild-type wing imaginal disc (A) or around clones of cells (marked by GFP in green) expressing either Hh-Np (B-D), Hh-N (E-G) or Hh-N-CD2 (H-J) produced in columnar cells using the flip-out technique. (D,G,J) Confocal z-sections (apical side upwards). Clones of cells expressing Hh-Np in the anterior compartment of the disc induced dpp expression in a six- or seven-cell wide region around the clone (B-D), while Hh-N was able only to activate dpp expression in a three- or four-cell region around the clone, even when the clones were bigger than Hh-Np expressing clones (E-G). Hh-N-CD2 expressing clones activated dpp in 1 row of cells around the clone (H-J).