Reggie-1/flotillin-2 promotes secretion of the long-range signalling forms of Wingless and Hedgehog in Drosophila

Abstract

The lipid-modified morphogens Wnt and Hedgehog diffuse poorly in isolation yet can spread over long distances in vivo, predicting existence of two distinct forms of these morphogens. The first is poorly mobile and activates short-range target genes. The second is specifically packed for efficient spreading to induce long-range targets. Subcellular mechanisms involved in the discriminative secretion of these two forms remain elusive. Wnt and Hedgehog can associate with membrane microdomains, but the function of this association was unknown. Here we show that a major protein component of membrane microdomains, reggie-1/flotillin-2, plays important roles in secretion and spreading of Wnt and Hedgehog in Drosophila. Reggie-1 loss-of-function results in reduced spreading of the morphogens, while its overexpression stimulates secretion of Wnt and Hedgehog and expands their diffusion. The resulting changes in the morphogen gradients differently affect the short- and long-range targets. In its action reggie-1 appears specific for Wnt and Hedgehog. These data suggest that reggie-1 is an important component of the Wnt and Hedgehog secretion pathway dedicated to formation of the mobile pool of these morphogens.

Figure 1

Overexpression of reggie-1 leads to loss of a subset of Wg responses. (A) Wild-type adult wing. (B, C) Adult wings overexpressing reggie-1 by en-Gal4 (B) or MS1096-Gal4 (C) lose wing margin (arrows); inserts show where these lines drive expression. Panel B also shows an Hh phenotype: broadening between veins 3 and 4 (yellow bar). (D–H) Wing imaginal discs of late third instar larvae. Ventral is up, posterior is right. White arrows indicate the A/P border; en-Gal4 expresses to the right from the arrows, marked by GFP. en-Gal4-driven reggie-1 downregulates Wg short-range targets Cut (D′) and Sens (E′), but not the long-range target Dll (H′). Wild-type Sens is shown in panel F. The ZNC visualized by the gap in BrdU staining (F′) is autonomously eliminated by reggie-1 (G′).

Figure 2

Overexpression of reggie-1 leads to enhanced spreading of Wg. (A) Wild-type wing disc expressing GFP in the posterior compartment showing normal Wg and Dll staining. (B–D) Overexpression of reggie-1 by en-Gal4 (B, D) or dpp-Gal4 (C) results in broadening of the Wg diffusion domain (B′, C′), loss of Sens expression (C′′) and vast increase in the extracellular Wg staining (D′). White parentheses in panel C mark dpp-Gal4 expression zone. (B′′′) shows a high-power magnification of panel B′; the green line demarcates reggie-1 overexpression. Arrowheads mark Wg puncta appearing far from the production zone.

Figure 3

Overexpression of reggie-1 affects Wg-producing cells. (A–F) Wild-type (wg-Gal4) adult wing (A) and wing disc (B, D–F). (G–L) wg-Gal4; UAS-reggie-1 adult wing (G) and wing disc (H, J–L). Reggie-1 overexpressed by wg-Gal4 loses wing margin structures (red arrows in panel G), reduces Sens expression and wing pouch size (J, K) and broadens the Wg gradient (H) as compared with wild-type discs (B). Wg gradient is eroded at the apical level: panels C and I show representative pixel intensity scans of wild-type and UAS-reggie-1 discs; the dorso-ventral border is at position 25 μm. Panels F and L show extracellular anti-Wg stainings oriented at 90° in relation with other panels. (M–P) Multiple small tub-Gal4; UAS-Wg clones (marked by loss of GFP) induce a reduction in the endogenous Wg production (yellow arrow in panel N) and ectopic Sens induction (white arrows in panel O). (R–U) Multiple small clones coexpressing Wg and reggie-1 produce more diffusive Wg (compare the diffuse anti-Wg staining in panel S with localized staining in panel N), incapable to reduce endogenous Wg (S) or induce ectopic Sens (T).

Figure 4

Loss of reggie-1 results in shortening of the Wg gradient and target gene expression. reggie-1-mutant discs show more narrow total (B) and extracellular (D) Wg gradient formation than wild-type discs (A, C) stained in parallel, but wg transcription (E, F) and Sens expression (G, H) are unchanged. In contrast, reggie-1^−/− discs have narrowed Dll expression domains (marked by dotted lines in I, J) and an overall threefold reduction in Dll expression levels (K, data shown as mean±s.e.m., n=8 discs); Student's t-test was used to determine statistical significance. (L) reggie-1^−/− somatic clones (marked by loss of GFP) intersecting the Wg-producing stripe reduce Wg spreading (L′′) and Dll expression (L′); the number of Wg particles emanating from the reggie-1^−/− Wg-producing cells (demarcated by white lines) is reduced (L′′′). (M–O) Wg spreading and the number and size of Wg particles are reduced in the domain expressing reggie-1 RNAi (marked by GFP); (N, N′) high-magnification of the wild-type and RNAi-expressing halves of same disc. Wg particles outside the Wg-producing stripe were counted separately in the wild-type and RNAi-expressing halves of identical size of the wing pouch (O); data are shown as mean±s.e.m., n=10 discs; paired t-test was used to determine statistical significance.

Figure 5

Cell culture experiments analysing the effects of reggie-1 on Wg and Hh secretion. (A) Effects of overexpression of reggie-1 on Wg secretion measured by Western blots. Serum albumin is shown as the loading control for the medium samples. (B) Effects of overexpression of reggie-1 or a reggie dominant-negative construct on secretion of Wg, Hh or secreted luciferase (sluc) measured by enzymatic assay. Bars represent means±s.e.m. from 12 experiments. P-values from Student's t-test are shown; ‘ns' means non-significant (P>0.05). (C) GFP-Wg media from panel A were subjected to sucrose-density ultracentrifugation. Thirteen fractions were collected (numbers and sucrose density in alternating fractions are shown on top). (D, E) S2 cells transfected with Wg together with reggie-1 (D) or EGFP (E) were co-cultured with cells independently transfected with DsRed. Accumulation of the anti-Wg signal in DsRed-cells is much higher if Wg was produced by reggie-1-transfected cells. (F, G) Pulse–chase Texas-red dextran colocalization assays. S2 cells transfected with Wg together with reggie-1 (F) or empty vector (G) were co-cultured with cells independently transfected with EGFP. Accumulation of Wg in dextran-positive early endosomes is much higher if Wg was produced by reggie-1-transfected cells (arrowheads in panel F). Note that unlike in panels D and E where identical confocal settings were used to record anti-Wg staining, the settings in panels F and G were independently optimized for highest resolution. (H) Reggie-1 does not stimulate Wg endocytosis cell-autonomously: S2 cells were transfected with Wg plus reggie-1-EGFP and co-cultured with cells independently transfected with reggie-1-DsRed. Incorporation of Wg into the Wg-receiving cells was the same whether they overexpressed reggie-1-DsRed or not (white arrows).

Figure 6

Modulations in reggie-1 levels affect Hh secretion, gradient formation and target gene expression. (A, B) Wild-type (A) and ap-Gal4; UAS-reggie-1 (B) discs stained for Ci, Ptc and Hh; ap-Gal4 drives expression below the white dotted line. Overexpression of reggie-1 results in a dramatic enhancement of Hh spreading into the anterior domain (white arrow (B′′)) and broadening of the Ptc expression (white arrows (B′)). High magnification (A′′′′, B′′′′) shows four anti-Ci staining zones, from right to left: (1) no staining in the posterior compartment, (2) low staining representing Ci^* (between the two green lines), (3) high staining of Ci-155 (between the green and blue lines) and (4) lower staining in the rest of the anterior representing Ci-75 (left from the blue line). Overexpression of reggie-1 strongly broadens the Ci^* zone (B′′′′). (C–H) Wild-type (E, F), hh-Gal4; UAS-reggie-1 (C, D) and reggie-1^−/− (G, H) discs stained for Ci (C, E, G), Col (C′, E′, G′), and extracellular Hh (D, F, H). (D′, F′, H′) are higher magnifications of panels D, F and H; the red arrows show the range of Hh diffusion into the anterior compartment (beginning of the arrows demarcates the A/P border). (C, C′′, E, E′′, G, G′′) White bars mark the Ci^* staining; red bars in (C′, E′, G′) mark Col expression.

Figure 7

Modulations in reggie-1 levels differently affect short-range and long-range targets of Wg and Hh depending on the morphogen production quantities. (A–C) Modelling Wg gradients using Equation 1. The parameters were as follows: v=18.7 molecules/s·cell; w=6 μm; a=3 μm; D=0.05 μm²/s and k=0.001427 s⁻¹ (‘wild-type gradient' in panels A and B). D was 0.01 μm²/s in ‘gradient in reggie-1^−/−' (A) and 0.25 μm²/s in ‘gradient in UAS-reggie-1' (B, C); w was 18 μm in ‘broad production zone' gradients (C). Yellow and blue arrows at the y-axis indicate the arbitrarily set [Wg] threshold levels for the expression of the short- and long-range targets respectively. The zone of expression of these targets is indicated by yellow and blue rectangles. Changes in the broadness of the target expression domains in panels A and B are indicated by grey arrows. (D, E) Ectopic Wg expressed by dpp-Gal4 (marked by UAS-GFP) together with UAS-reggie-1 spreads further away from the production zone (E), shows a more punctate staining (E′′) and induces ectopic Sens, and is a broader domain (E′) than ectopic Wg expressed without reggie-1 (D–D′′). GFP staining in (D′, E′) is in blue pseudocolour for a better visualization. (F) Expression of reggie-1 by hh-Gal4 at 25°C strongly reduces the size of posterior compartment and changes Hh responses: Hh now induces narrow Col (arrows in F′′) but broad Ci-155 (brackets in (F′)) zones. (G, H) High magnification shows enhanced diffusion of Hh into the anterior compartment upon reggie-1 overexpression (G) as compared with wild-type discs (H). The A/P border is shown with a thick yellow line and the range of Hh infiltration is marked with a thin yellow line.