1. Developmental Biology

High hedgehog signaling is transduced by a multikinase-dependent switch controlling the apico-basal distribution of the GPCR smoothened

  1. Marina Gonçalves Antunes
  2. Matthieu Sanial
  3. Vincent Contremoulins
  4. Sandra Carvalho
  5. Anne Plessis  Is a corresponding author
  6. Isabelle Becam  Is a corresponding author
  1. Université Paris Cité, CNRS, Institut Jacques Monod, France
https://doi.org/10.7554/eLife.79843
Share this article
Cite this article
  1. Marina Gonçalves Antunes
  2. Matthieu Sanial
  3. Vincent Contremoulins
  4. Sandra Carvalho
  5. Anne Plessis
  6. Isabelle Becam
(2022)
High hedgehog signaling is transduced by a multikinase-dependent switch controlling the apico-basal distribution of the GPCR smoothened
eLife 11:e79843.
https://doi.org/10.7554/eLife.79843
  2. Download BibTeX
  3. Download .RIS

Abstract

The oncogenic G-protein-coupled receptor (GPCR) Smoothened (SMO) is a key transducer of the hedgehog (HH) morphogen, which plays an essential role in the patterning of epithelial structures. Here, we examine how HH controls SMO subcellular localization and activity in a polarized epithelium using the Drosophila wing imaginal disc as a model. We provide evidence that HH promotes the stabilization of SMO by switching its fate after endocytosis toward recycling. This effect involves the sequential and additive action of protein kinase A, casein kinase I, and the Fused (FU) kinase. Moreover, in the presence of very high levels of HH, the second effect of FU leads to the local enrichment of SMO in the most basal domain of the cell membrane. Together, these results link the morphogenetic effects of HH to the apico-basal distribution of SMO and provide a novel mechanism for the regulation of a GPCR.

Editor's evaluation

In this paper, Gonçalves-Antunes and colleagues uncovered that the morphogen Hedgehog regulates the activity and subcellular localisation of Smoothened through vesicular trafficking. In particular, they demonstrated that Smoothened trafficking favours recycling and basal enrichment depends on its phosphorylation signature in a Hh-concentration-dependent manner. This work will interest a wide readership as it links Hh's functions as a morphogen with Smoothened's subcellular localisation.

https://doi.org/10.7554/eLife.79843.sa0

Introduction

During development, signaling pathways control epithelial morphogenesis by acting on cell proliferation, differentiation, survival, and migration. Epithelial cells uniquely display an apico-basal (Ap-Ba) polarity with differential distribution of phospholipids and protein complexes between the different membrane domains (for review see Ikenouchi, 2018). This leads to functionally separated subregions with distinct properties and physiological functions, such as the microvilli in the apical domain, cell-cell adhesion junctions in the lateral domain, and cell-matrix adhesion in the basal domain. Finally, the apical and basal regions are in contact with extracellular environments, which can differ in the nature and dose of signaling molecules. For all these reasons, the control of signaling receptor distribution among these specific subdomains of the plasma membrane is expected to be critical for correct signal transduction.

The conserved hedgehog (HH) signals play major roles in the development of metazoans. Initially identified in the fly model, HH signaling is involved in the promotion, development, and/or metastasis of numerous types of tumors and drugging this pathway is a major goal for cancer therapies (for review see Ingham, 2022). In flies, HH controls the patterning of many structures including the wing imaginal disc (WID), which has been instrumental in the study of HH signaling (for review see Hartl and Scott, 2014). In this epithelial structure, HH emanating from the posterior (P) cells signals to the anterior (A) cells near the A/P boundary, thus controlling the expression of target genes in a dose-dependent manner. Of note, HH molecules form two gradients in the WID: an apical gradient, required for long-range, ‘low HH’ responses (as in the expression of decapentaplegic [dpp]) that depend on glypicans, and a basal one, for short-distance, ‘high HH’ responses (as in the anterior expression of engrailed [en]), that is based on the transport of HH by exosomes associated to filopodia-like structures oriented in the A/P axis (D’Angelo et al., 2015 and González-Méndez et al., 2017).

HH transduction requires Smoothened (SMO), a G-protein-coupled receptor (GPCR) like protein. In the absence of HH, the HH co-receptor Patched (PTC) inhibits SMO, probably by depleting accessible cholesterol from the outer leaflet of the plasma membrane (Kinnebrew et al., 2021 and see for review Radhakrishnan et al., 2020), which promotes the formation of a repressor form of the transcription factor cubitus interruptus (CI). The binding of HH to PTC inhibits its negative effect, which blocks the cleavage of CI, leading to the transcription of pathway target genes by full-length CI. SMO acts as a scaffold to transduce HH signaling to CI via an intracellular complex (called HTC for HH Transduction Complex) bound to its cytoplasmic C-terminal domain, which includes CI and a protein kinase called Fused (FU) (Malpel et al., 2007; Robbins et al., 1997; Sisson et al., 1997). SMO activation is associated with conformational switches, both in its cytoplasmic C-terminal domain and in its extracellular domains; these events correlate with its clustering, which seems critical for the downstream activation of the pathway (Fan et al., 2012; Shi et al., 2011; Su et al., 2011; Zhao et al., 2007).

Several labs—including ours—have highlighted the role of endocytic trafficking and the importance of post-translational modifications in the regulation of SMO’s levels, localization, and activation. SMO activation in the presence of HH is associated with changes in its localization: from internal vesicles to the plasma membrane in Drosophila and from the cell body to the primary cilium in mammals (Denef et al., 2000 and Huangfu et al., 2003). Both events are positively controlled by extensive phosphorylation of SMO’s intracellular tail by multiple kinases (for review see Chen and Jiang, 2013). Despite significant differences, the processes involved are remarkably conserved as illustrated by the fact that human SMO can be relocalized in response to HH to the surface of fly cells (De Rivoyre et al., 2006). In Drosophila, many kinases (protein kinase A [PKA], casein kinase I [CKI], GPCR kinase 2, casein kinase 2, Gilgamesh, atypical protein kinase C [aPKC]) are implicated, which regulate SMO activation and accumulation at the membrane (Apionishev et al., 2005; Chen et al., 2010; Jia et al., 2010; Jia et al., 2004; Li et al., 2016; Maier et al., 2014 and Zhang et al., 2004). We have also identified a phosphorylation-based positive feedback loop between SMO and the FU kinase, which is required for the response to the highest doses of HH (Alves et al., 1998). In this process, the initial activation of SMO promotes the recruitment at the cell membrane of FU and its activation, which then further phosphorylates SMO, leading to an enhanced accumulation of the SMO/FU complex at the cell surface and high signaling activation (Claret et al., 2007; Sanial et al., 2017).

Are these events polarized along the Ap-Ba axis and what is their link with the gradients of HH? Previous studies indicated that SMO is unevenly distributed along the Ap-Ba axis of the WID epithelial cells (Denef et al., 2000; Jiang et al., 2014; Sanial et al., 2017). Here, by specifically labeling the population of SMO at the plasma membrane, we show that it is unevenly distributed along the Ap-Ba axis and that HH acts in a dose-dependent manner to increase its accumulation in the most basal region. Blocking the endocytosis of SMO or following its fate after endocytosis, reveal that SMO is initially targeted to the apical membrane and that HH does not dramatically affect its apical endocytosis but affects its post-endocytic fate, favoring recycling over degradation. Moreover, we provide evidence that the HH-dependent basolateral enrichment of SMO relies on a two-step action of the FU kinase, first apically to enhance SMO localization at the cell surface before stabilizing it in the basolateral region. Altogether, these data support a model which connects the HH-dose-dependent activation of SMO to its vesicular trafficking and Ap-Ba localization.

Results

High levels of HH promote a basolateral enrichment of cell surface SMO

As HH promotes the accumulation of SMO at the plasma membrane, we set out to determine (i) whether the population of SMO that is present at the cell surface is differentially distributed along the Ap-Ba axis and (ii) whether its Ap-Ba distribution is affected by HH. For that purpose, a fusion between the extracellular N-terminus of SMO and the enzymatic self-labeling SNAP-tag (Tirat et al., 2006), called SNAP-SMO, (Sanial et al., 2017) was overexpressed in the dorsal compartment of the WID (see Figure 1A–A’ for the organization of this disc). This fusion was fully functional as its expression under smo’s endogenous promotor (from a BAC construct) rescued a loss of smo function (Figure 1—figure supplement 1A, B-B”). Moreover, its overexpression has no effect on HH signaling (Figure 1—figure supplement 1C-C”, see also Sanial et al., 2017). After dissection, the fraction of SNAP-SMO present at the cell surface (thereby called Surf SNAP-SMO) was specifically labeled using a non-liposoluble fluorescent SNAP ligand (Figure 1—figure supplement 2A, A’) before being imaged.

Figure 1 with 3 supplements see all
Download asset Open asset
Cell surface Smoothened (SMO) is unevenly distributed along the apico-basal (Ap-Ba) axis and high levels of hedgehog (HH) promote its basolateral enrichment.

(A–A’) (A) Left: scheme of a wing imaginal disc (WID) with the posterior (P) compartment, where HH (in green) is produced, the anterior (A)/P boundary is represented by the dotted dark blue line and the dorsal/ventral (D/V) boundary by the dotted yellow line. The fading green color represents the anterior gradient formed by HH (across about 12 rows of cells). The apGal4 driver used to express SNAP-smo, is expressed in the dorsal compartment. The circle represents the wing pouch. The rectangle indicates the region that is shown in the XY images. Right: the x (D/V), y (A/P), and z (Ap-Ba) axis used for imaging are in yellow, blue, and red, respectively. Here and in all the XY images, the dorsal compartment is to the top. (A’) Scheme of an antero-posterior XZ section of the wing pouch with the columnar epithelium at the bottom (c.e., which is represented in (A) and studied here) and the peripodial epithelium at the top (p.e.), with their respective apical sides facing a lumen (lu.). (B–B”) Confocal XY sections along the Ap-Ba axis (as indicated) of a third instar larva WID expressing UAS SNAP-smo (driven by apGal4) in the dorsal compartment and labeled with a non-liposoluble fluorescent SNAP substrate. The scale bar represents 20 µm. Here and in the other XY images: the dorsal part of the wing pouch is shown, the discs are oriented anterior to the left, dorsal up, and the A/P boundary is represented by a vertical dotted white line drawn based on the absence of cubitus interruptus (CI) labeling in the P compartment (see Figure 1—figure supplement 2B). Here, in C–C” and F–F” the images were acquired using the confocal LSM980 spectral Airyscan 2, 63×. (C–C”) Confocal XZ antero-posterior image in the dorsal part of an apGal4; UAS SNAP-smo third instar larva wing pouch labeled for Surf SNAP-SMO with a non-liposoluble fluorescent SNAP substrate (C, green in C’’) and immunolabeled for CI-F (C’, red in C’’). Merged image in (C’’). Based on CI staining, four regions (represented between the C and C’ images and separated by dotted white lines) are identified along the antero-posterior axis: the P region (green), and three A regions: CI-A (low levels of full-length activated CI, in purple), CI-F (higher levels of CI full-length, in pink) and CI-R (CI repressor not detected by the anti-CI antibody used here, in red). This image was reconstituted from the XY stack that includes the three sections shown in B–B”. Here and in the other XZ images: the discs are oriented anterior to the left, apical up, the A/P boundary is represented by a vertical dotted white line, and the scale bars are 10 µm. (D–E) Graphs showing the mean intensity (D) and the relative intensity (calculated for each region along the Ap-Ba axis as the ratio of its integrated density over the integrated density of the height of the disc, which is designated the column) (E) of Surf SNAP-SMO in the apical, lateral, and basal domains (in light, medium, and dark blue, respectively) of the CI-R (R), CI-F (F), CI-A (A), and P regions. N=33. Here and in the other figures, all the quantifications were done using projections of eight XZ sections that were directly acquired using an SP5 AOBS confocal, 40×. The error bars represent the SD and the statistical analysis was performed using paired t-test for the mean intensities and Wilcoxon matched-pairs signed rank tests for the relative intensities. (F–F”) Confocal XZ images in the dorsal part of WIDs from apGal4; UAS SNAP-smo, hhts2 flies in which HH was inactivated at the restrictive temperature prior to dissection and labeling of Surf SNAP-SMO (F, green in F’’) and CI-F (indicated here as CI, in F’, red in F’’). Merged image in (F’’).

Figure 1—source data 1

Mean intensity of Surf SNAP-SMO along the apico-basal axis.

https://cdn.elifesciences.org/articles/79843/elife-79843-fig1-data1-v1.xlsx
Figure 1—source data 2

Relative intensity of Surf SNAP-SMO along the apico-basal axis.

https://cdn.elifesciences.org/articles/79843/elife-79843-fig1-data2-v1.xlsx

We imaged XY sections of wing discs labeled for SNAP-SMO at different positions along the Ap-Ba axis (Figure 1A, B, B’ and B”). As expected, Surf SNAP-SMO levels are lower in the more A cells (called thereby far anterior, FA) away from the HH source than in the A cells abutting the A/P boundary (and that respond to HH) or in the P cells (where HH signaling is activated due to the lack of ptc expression) (Méthot and Basler, 1999). Strikingly, the increased accumulation of SNAP-SMO in the cells in which HH signaling is activated is particularly visible in the lateral section (Figure 1B’) and even more in the basal section (Figure 1B”) than in the apical region of the cells (Figure 1B). This basolateral enrichment is also highlighted in the reconstituted antero-posterior XZ sections (Figure 1A’, C and C’’).

To quantify the effects of HH on Surf SNAP-SMO levels and distribution, we measured (using XZ projections of eight sections) both its mean intensity (sum of pixel values over the number of pixels) and the integrated density (sum of pixel values) in three regions along the Ap-Ba axis (Figure 1—figure supplement 2C): (i) the apical region, estimated here as the 15% most apical region based on Discs large (DLG) staining of the septate junctions, (ii) the basal region, arbitrarily defined as being the 10% most basal part, and (iii) the lateral or intermediate region in between the two others (Figure 1—figure supplement 2C). Since HH acts as a morphogen in the wing disc, we performed these quantifications in four regions across the wing disc epithelium, based on the co-immunodetection of the transcription factor CI (Figure 1C’ and C”, Figure 1—figure supplement 2C): the P compartment (green, where CI is not expressed) and three anterior regions: (i) the CI-R region (red, corresponding to the FA region), where CI is processed into its shorter repressor form that is not detectable with the antibody used here, (ii) the CI-F region (pink), corresponding to cells that receive medium to low levels of HH that lead to the stabilization of full-length CI, and (iii) the CI-A region (purple), which corresponds to the cells nearer to the A/P border, where CI-F is very active (and called CI-A) but present at low levels due to the repression of ci by the anterior expression of en promoted by high HH levels (Roberto et al., 2022). The comparison of the mean intensities of Surf SNAP-SMO in these four regions shows that HH progressively increases the levels of SNAP-SMO, with Surf SNAP-SMO being 1.7-fold more abundant in the P region than in the CI-R region (p-value) (Figure 1D, gray columns). These effects are seen in all the regions along the Ap-Ba axis (Figure 1D, the light, medium, and dark blue colors correspond to the apical, lateral, and basal regions, respectively). However, the calculation of the relative intensities of SNAP-SMO (calculated as the ratio of the integrated density of each region along the Ap-Ba axis over the integrated density of the three regions together, called column) shows that this increase in SNAP-SMO levels is unequal along the Ap-Ba axis, with a relative decrease in the apical fraction associated with a relative enrichment of the lateral and basal fractions (Figure 1E).

Importantly, a similar increase in the basolateral localization of SMO in the anterior region abutting the A/P and the P regions is also seen for immunolabeled endogenous SMO (Figure 1—figure supplement 3A, A”, B), or when we labeled Surf SNAP-SMO expressed at an endogenous level from a BAC construct (Figure 1—figure supplement 3C, C”). Note that when we also specifically labeled the intracellular fraction of SNAP-SMO, neither its levels nor its Ap-Ba distribution was affected by HH (Figure 1—figure supplement 2A, A”’, D, D’, E).

To ensure that these changes in the Ap-Ba distribution of Surf SNAP-SMO were indeed due to HH, we performed the same experiment when the HH function was inactivated. For that purpose, we looked at Surf SNAP-SMO in a genetic context homozygous for a thermosensitive allele of hh (hhts2) (Ma et al., 1993). At restrictive temperature, the function of HH is reduced, and both the anterior expression of en and the reduction in CI-F levels in the cells abutting the A/P boundary are suppressed (Figure 1F’ and F”, Figure 1—figure supplement 2F,F”). Under that condition, Surf SNAP-SMO is no longer accumulated in the P and the anterior cells near the A/P and is no longer enriched in the basolateral region of these cells (Figure 1F, F’ and F”).

In summary, together these data provide evidence that SMO is asymmetrically distributed along the Ap-Ba axis, and that HH acts in a dose-dependent manner on the Ap-Ba distribution of Surf SNAP-SMO, leading to an increase in the lateral and especially basal population in presence of the highest levels of HH.

HH controls the fate of SMO post-endocytosis

To understand how the distribution of Surf SNAP-SMO along the Ap-Ba axis is established, we looked at the consequences of blocking its endocytosis.

For that purpose, we first used a thermosensitive mutation of the shibire (shits), a Drosophila Dynamin ortholog, which is central for the scission of coated vesicles (van der Bliek and Meyerowitz, 1991). Blocking SHI activity—for less than an hour—leads to an accumulation of Surf SNAP-SMO in both compartments of the WID, with a stronger accumulation in the apical region of the cells (Figure 2A, B, Figure 2—figure supplement 1A, B).

Figure 2 with 1 supplement see all
Download asset Open asset
Blocking Smoothened (SMO) endocytosis favors its accumulation in the apical region independently of hedgehog (HH).

(A–B) XZ confocal images of wing imaginal discs (WID) from shi+; apGal4; UAS SNAP-smo (A, called ctr for control in C and C’) or shits; apGal4; UAS SNAP-smo (B, called shits in C and C’) male flies put at the restrictive temperature for 30 min before dissection, labeling of Surf SNAP-SMO and imaging under the same conditions. Note that here and in all the following figures the images were taken using an SP5 AOBS confocal, 40× oil. (C–C’) Quantification of the mean intensities (C) and relative intensities (C’) of Surf SNAP-SMO in the different regions of the disc (far anterior [FA] and posterior [P]) and along the apico-basal (Ap-Ba) axis, as indicated. The mean intensity is also shown for the whole epithelial column (gray). N=17 and 10 for ctr and shits flies, respectively. (D–E”) XZ confocal images of WIDs from apGal4, Gal80ts; UAS SNAP-smo (D) (ctr) or apGal4, Gal80ts/+; UAS SNAP-smo/UAS YFP-rab5CA (E–E”) flies put at the restrictive temperature for 24 hr (to allow YFP-rab5CA expression) before dissection and labeling for Surf SNAP-SMO. In the latter case, E” is a merged image with Surf SNAP-SMO in green and YFP-RAB5CA in red. Note that for better visualization of SMO, imaging of YFP-rab5CA and ctr discs were not acquired under the same conditions. (F) The relative intensity of Surf SNAP-SMO along the Ap-Ba axis in apGal4, Gal80ts; UAS SNAP-smo ctr discs (called ctr, n=11) and apGal4, Gal80ts/+; UAS SNAP-smo/UAS YFP-rab5CA (called rab5CA, n=8) discs. (G) The relative intensity of Surf SNAP-SMO along the Ap-Ba axis in apGal4; UAS SNAP-smo (called ctr, n=20) and apGal4/UAS smurf RNAi; UAS SNAP-smo (called smurf RNAi, n=6) discs. Note that YFP-rab5CA or smurf RNAi overexpression has a much stronger effect than shits inactivation, reflecting that endocytosis is blocked for a much shorter time in the latter case. Here and in all the following figures, the XZ images correspond to a unique single XZ section.

Figure 2—source data 1

Mean intensity of Surf SNAP-SMO for control and shits flies.

https://cdn.elifesciences.org/articles/79843/elife-79843-fig2-data1-v1.xlsx
Figure 2—source data 2

Relative intensity of Surf SNAP-SMO for control and shits flies.

https://cdn.elifesciences.org/articles/79843/elife-79843-fig2-data2-v1.xlsx
Figure 2—source data 3

Relative intensity of Surf SNAP-SMO for control and RAB5CA flies.

https://cdn.elifesciences.org/articles/79843/elife-79843-fig2-data3-v1.xlsx
Figure 2—source data 4

Relative intensity of Surf SNAP-SMO for control and smurf RNAi flies.

https://cdn.elifesciences.org/articles/79843/elife-79843-fig2-data4-v1.xlsx

Quantification of the mean intensities in the FA region (corresponding to CI-R, no HH) and P region are shown in Figure 2C. It reveals that the increase in Surf SNAP-SMO levels is comparable in both regions and that it occurs all along the Ap-Ba axis of the cells, with an especially strong increase in the apical region. Calculation of the relative intensities (Figure 2C’) confirms the relative enrichment of Surf SNAP-SMO in the apical region of the cells and shows that it is associated with its relative decrease in the lateral region. Note that the effects seen here are specific to the inactivation of shits by the restrictive temperature as (i) the distribution of Surf SNAP-SMO is similarly affected when comparing shits flies at permissive and restrictive temperature (Figure 2—figure supplement 1E), whereas (ii) shits and shi+ control (ctr) flies have indistinguishable distributions when kept at the permissive temperature (+).

To confirm these results, we also blocked SMO trafficking just after endocytosis (in early endosomes), using a constitutively active RAB5 tagged with a YFP (YFP-RAB5CA) that is locked in the GTP bound state (D’Angelo et al., 2015). When YFP-RAB5CA is expressed for 24 hr, Surf SNAP-SMO strongly accumulates in YFP-RAB5CA positive endocytic vesicles, which are almost exclusively located in the apical region (Figure 2D and E–E”, Figure 2—figure supplement 1C, D, D”). This leads to a decrease in the relative abundance of its lateral fraction and to a lesser extent, of its basal fraction (Figure 2F). These effects are seen both in the presence (P region) and absence (FA region) of HH but are slightly weaker in its presence. Of note, YFP-RAB5CA overexpression also led to the accumulation of endogenous SMO or SNAP-SMO expressed at endogenous levels in apical YFP-RAB5CA positive endocytic vesicles (Figure 2—figure supplement 1G-G”,"H-H"").

Finally, we ruled out an indirect effect of a general block of endocytosis, as we obtained similar results when we specifically blocked the endocytosis of SNAP-SMO by downregulating (by RNA interference, RNAi) the expression of smurf, which encodes an E3 ubiquitin ligase known to promote SMO endocytosis by mediating its ubiquitylation (Li et al., 2018; Figure 2G).

In conclusion, blocking SMO endocytosis/in early endosomes by three different means reveals that (i) SMO endocytosis is not dramatically affected by HH and (ii) SMO endocytosis occurs all along the Ap-Ba axis but more apically than basolaterally. They also show that newly synthesized SMO is initially addressed and subsequently endocytosed at the apical membrane.

Endocytosed SMO is targeted from the apical to the basolateral region

The above data indicate that the strong stabilization of SMO in response to HH is likely due to a reduction of its degradation after endocytosis. To study the fate of Surf SNAP-SMO after its endocytosis, we performed an endocytosis assay in which Surf SNAP-SMO labeling was followed by a chase. After the chase, the subcellular localization of Surf labeled SNAP-SMO shifted: its presence at the cell surface decreases, and this is associated with increased localization in dot-like structures that likely correspond to endocytic vesicles as some of them colocalize with RAB7, an endosomal marker required for the trafficking between late endosomes and the lysosome (Vanlandingham and Ceresa, 2009; Figure 3A and B and Figure 3—figure supplement 1A, A’, A”, B, B’, B”). The global levels of Surf labeled SNAP-SMO also decrease during the chase in the whole disc with a reduction of 44% in the far anterior cells (no HH) but of only 22% in the posterior cells (with HH) (Figure 3C, Figure 3—figure supplement 1C). In the absence of HH, the apical and lateral regions are slightly more affected than the basal regions (with a 47% and 44% decrease for the former compared to 28% for the latter) (Figure 3C, Figure 3—figure supplement 1C’). This leads to an increase in the relative abundance of Surf labeled SNAP-SMO in the basal region (associated with a slight but non-significant decrease of the apical fraction) (Figure 3D). By contrast, in the presence of HH (P region), the reduction is much more pronounced in the apical region (38%) than in the intermediate (28%) and basal region (which is barely affected, 4%) (Figure 3C, Figure 3—figure supplement 1C’). This leads to a relative redistribution in the P region (compared to the FA region) of endocytosed Surf labeled SNAP-SMO from the apical to the basal and intermediate regions (Figure 3D).

Figure 3 with 1 supplement see all
Download asset Open asset
Hedgehod (HH) controls the fate of endocytosed Smoothened (SMO), favoring its recycling and leading to its basolateral accumulation.

XZ confocal images of dissected wing imaginal discs expressing UAS SNAP-smo fixed immediately (0 min) (A) or 15 min (B) after labeling of Surf SNAP-SMO. Quantification of the mean intensity (C) and of the relative intensity (D) in the far anterior and posterior regions of the disc and along the apico-basal axis. N=20 for 0 min and 23 for 15 min. Int: intermediate. The discs were treated in the same conditions, all the images were acquired under the same conditions and the dynamic range was normalized to allow a better comparison of Surf SNAP-SMO.

Figure 3—source data 1

Mean intensity of Surf SNAP-SMO at T0 and at T15'.

https://cdn.elifesciences.org/articles/79843/elife-79843-fig3-data1-v1.xlsx
Figure 3—source data 2

Relative intensity of Surf SNAP-SMO at T0 and T15'.

https://cdn.elifesciences.org/articles/79843/elife-79843-fig3-data2-v1.xlsx

In summary, our data show that, both with and without HH, SMO endocytosis leads to a change in its distribution along the Ap-Ba axis, in favor of the basal and intermediate regions, suggesting that SMO could undergo transcytosis. Moreover, HH reduces the degradation of endocytosed SMO and favors its presence in the intermediate and basal regions. As HH does not dramatically affect SMO endocytosis, this indicates that HH controls the fate of endocytosed SMO by reducing its degradation and increasing its recycling.

Phosphorylation by the PKA/CKI and FU kinases regulates the Ap-Ba localization of SMO at the cell surface

Given that the PKA and FU kinases positively regulate SMO activation and accumulation at the membrane, we analyzed their effect on the Ap-Ba localization of SMO by looking at forms of SMO mimicking (S to D replacements) or blocking (S to A substitution) these phosphorylations.

First, we looked at the constitutively active SNAP-SMOPKA-SD, which mimics SMO fully phosphorylated by PKA and CKI kinases (Jia et al., 2004). As expected from previous data with SMOPKA-SD in cultured cells (Jia et al., 2004; Sanial et al., 2017), Surf SNAP-SMOPKA-SD accumulates at the cell membrane both in the anterior and posterior compartments of the wing disc (Figure 4A and B, Figure 4—figure supplement 1A-A’, B-B’). Moreover, even in the absence of HH (FA cells), its accumulation in the apical region decreases in favor of the lateral and even more of the basal region, similar to what is seen with Surf SNAP-SMOWT in presence of HH (Figure 4E–E”). These effects in the FA region are partially suppressed by mutations that prevent the phosphorylation by the FU kinase, with especially a strong reduction of the basal localization of Surf SNAP-SMOPKA-SD FU-SA, compared with Surf SNAP-SMOPKA-SD (Figure 4C and E–E”, Figure 4—figure supplement 1C-C’). In contrast, Surf SNAP-SMOPKA-SD FU-SD, which accumulates at high levels at the cell surface in both compartments of the discs, shows a further decrease in its accumulation at the apical and lateral cell surface in A cells compared to SMO PKA-SD (Figure 4D and E–E”, Figure 4—figure supplement 1D-D’). This effect is associated with a strong basal enrichment.

Figure 4 with 1 supplement see all
Download asset Open asset
The apico-basal (Ap-Ba) distribution of Smoothened (SMO) is controlled by phosphorylation by the protein kinase A (PKA)/ casein kinase I (CKI) and Fused (FU) kinases.

(A–D) Confocal images of wing imaginal discs labeled for Surf SNAP-SMOWT (A), SNAP-SMOPKA-SD (B), SNAP-SMOPKA-SD FU-SA (C), and SNAP-SMOPKA-SD FU-SD (D). Note that for better visualization of SMO, imaging of the different forms of SNAP-SMO was not done under the same conditions. See also the corresponding false-color images in Figure 4—figure supplement 1A. (E–E”) Quantification of the Ap-Ba distribution of the different forms of SNAP-SMO, as indicated. N=21 WT, 10 PKA-SD, 17 PKA-SD FU-SA, and 18 PKA-SD FU-SD discs, respectively. (See Figure 4—figure supplement 1E) showing the reproducibility of Surf SNAP-SMOWT distribution. Note that the effects of the PKA-SD mutations are even stronger than the effect of hedgehog (HH) on SNAP-SMOWT, probably because in presence of HH, only a fraction of the SMO population is phosphorylated (Sanial et al., 2017).

Figure 4—source data 1

Relative intensity of Surf SNAP-SMO for different SMO mutants.

https://cdn.elifesciences.org/articles/79843/elife-79843-fig4-data1-v1.xlsx

The phosphorylation of SMO by the PKA is known to promote its activation and accordingly, the expression of SNAP-SMOPKA-SD leads to both an anterior expansion of anterior en and the CI-A regions, two outcomes of high HH signaling (compare the dorsal region in which SNAP-SMOPKA-SD is expressed to the ctr ventral region in Figure 4—figure supplement 1F-F”). In contrast, expression of SMOPKA-SD FU-SA under the same conditions led to a reduction in both the anterior en and CI-A domains (Figure 4—figure supplement 1G-G”).

Overall, these results provide evidence that the phosphorylation by the PKA/CKI recapitulates the effects of HH on Surf SMO Ap-Ba distribution, leading to a relative apical depletion and basolateral enrichment. They also show that this effect is dependent upon further phosphorylation of SMO by FU, which both especially enhances its distribution in the basal region and is required for high HH signaling.

FU is required first apically and then basally to promote high SMO activity

Our data strongly link the basal localization of SMO and its activation by the FU kinase. FU is present everywhere along the Ap-Ba axis, with some enrichment in the apical region. It is both diffused in the cytoplasm and present in vesicular puncta, some of which colocalize with SMO (Claret et al., 2007; Figure 5—figure supplement 1A, A’, B-B”). To understand where FU acts on SMO, we trapped it in the apical or in the basolateral region of the cells using the GRAB bipartite system (Harmansa et al., 2017). In this method, GFP tagged FU (GFP-FU, known to be fully functional Malpel et al., 2007) and whose Ap-Ba distribution is similar to that of endogenous FU (Figure 5—figure supplement 1C), is trapped in either the apical or basolateral domain via its binding to an intracellular GFP nanobody fused to an apical (T48) or basolateral (NVR1) transmembrane domain. For that purpose, we expressed both GFP-FU and its trap in the dorsal part of the disc and we immunolabeled endogenous SMO (Figure 5, note that in B–B”’ and D–D”’, the Z sections correspond to anterior YZ sections across the dorso-ventral axis, with the ventral compartment serving as an internal control).

Figure 5 with 2 supplements see all
Download asset Open asset
Trapping the Fused (FU) kinase to the apical region promotes the stabilization of Smoothened (SMO) at the cell surface.

Confocal images of wing imaginal discs coexpressing (using the apGal4 driver) GFP-fu and T48 or NVR1 fused to the mcherry (mche). XY sections are shown in (A–A”’ and C–C”’), anterior YZ sections in (B–B”’ and D–D”’). The mche is shown in (A, B, C, and D), GFP-FU in A’, B’ C’, and D’, green in the merged images A”’, B”’, C”’, and D”’ and immunolabeled endogenous SMO in A”, B”, C” and D”, red in the merged images A”’, B”’, C”’, and D”’. Note that contrary to SNAP-SMO, endogenous SMO is also present in the peripodial membrane (black arrow above the SMO images). D: dorsal, V: ventral, A: anterior, P: posterior, ap: apical, and ba: basal. Here and in Figure 6, the scale bar for XY section represents 50 µm.

As expected, expression of GFP-fu with mCherry (mche)-T48 leads to strong apical relocalization of GFP-FU (compare Figure 5B’ to Figure 5—figure supplement 1C), while its coexpression with mche-NVR1 leads to its basolateral enrichment (compare Figure 5D’ to Figure 5—figure supplement 1C). Apical tethering of GFP-FU (with mche-T48) promotes the accumulation of SMO both in the anterior (Figure 5A” and B”, Figure 5—figure supplement 2D) and posterior (Figure 5A”, Figure 5—figure supplement 2A”’, D’) compartments. This effect is homogeneous along the Ap-Ba axis, with only a slight increase in the basal relative distribution of SMO (Figure 5—figure supplement 2D”). Importantly, a similar effect is seen with Surf SNAP-SMO (expressed at endogenous levels from the BAC construct), indicating that T48-tethered GFP-FU stabilizes SMO at the cell surface (Figure 5—figure supplement 2C-C”’). By contrast, basolateral trapping of GFP-FU (with mChe-NVR1) has only a slight effect on SMO accumulation (Figure 5C”–D” and Figure 5—figure supplement 2B”’) and seems to induce its vesicular localization.

Next, we tested the effects of trapping GFP-FU with mche-T48 or NVR1 on HH signaling. We monitored the accumulation of CI and the expression of the low-HH target dpp (using a dpp-LacZ reporter, dpp-Z), the medium/high-HH target ptc, and the high-HH target, en. While expression of GFP-FU alone has no effect (Claret et al., 2007), coexpression of GFP-FU with mche-T48 leads to the ectopic activation of medium levels of HH signaling as it promotes the ectopic expression of dpp-Z (Figure 6—figure supplement 1A) and ptc throughout the whole anterior compartment (although its expression at the A/P border is slightly reduced) (Figure 6A). On the other hand, high-level HH signaling was reduced as en expression was decreased near the A/P border but expanded at low levels throughout the anterior region (Figure 6B). By contrast, trapping GFP-FU with mche-NVR1 has no effect on dppZ and ptc expression and only weakly reduces the anterior en expression (Figure 6C and D, Figure 6—figure supplement 1B).

Figure 6 with 1 supplement see all
Download asset Open asset
Trapping the Fused (FU) kinase to the apical region is sufficient to promote medium but not high hedgehog (HH)/ Smoothened (SMO) signaling.

Confocal XY images of discs expressing GFP-fused with mche-T48 in fuWT (A–B) or knockout (fuKO) context (E–F), with mche-NVR1 in fuWT (C–D) or alone in fuKO context (G–H) and immunolabeled for Patched (PTC) (A, C, E, and G) or EN (B, D, F, and H). In (A–D), the normal pattern of expression of these genes is visible in the ventral region of the discs.

To ensure that these effects were not due to an indirect effect of mche-T48-trapped GFP-FU on endogenous FU, we repeated this experiment in a fu mutant background. As all currently used fu mutants display complex genetic interactions suggesting that they may not be null alleles, we knocked out the fu gene by CRISPR, leading to a deletion (called fuKO) that removed the entire fu transcribed region, except for the 3’UTR region (Figure 6—figure supplement 1C). This mutation leads to the total suppression of anterior en expression, and a very strong reduction of ptc expression (see the ventral region of the discs shown in Figure 6E, F, G and H). These effects are similar but stronger than what was shown for strong fu mutants (Alves et al., 1998). In this context, expression of GFP-FU trapped with mche-T48 has a similar effect to that in presence of the endogenous functional FU protein: ectopic anterior ptc expression, reduced en expression near the A/P, (see the dorsal region of the discs shown in Figure 6E and F). Strikingly, while the expression of GFP-FU alone suppresses the effect of the fuKO mutation on HH signaling with the restoration of ptc and anterior en expression (see the dorsal region of the discs shown in Figure 6G and H), it was unable to promote the effect of T48-tethered GFP-FU on ptc and anterior en. These data indicate that the effects seen when GFP-fu is coexpressed with mche-T48 are indeed due to its apical tethering and do not require the presence of endogenous FU.

In summary, anchoring FU to the apical membrane, but not to the basolateral one, increases the levels of SMO at the cell surface. It also leads to its constitutive activity, promoting low-medium HH signaling but blocking very-high HH signaling. This suggests (i) that only apical FU can activate SMO and stabilize it at the cell surface and (ii) that both the high HH-induced basal enrichment of SMO and the expression of high HH targets may require a second input of FU on SMO in the basolateral region.

Discussion

Understanding how HH controls the fate of SMO in a polarized epithelium is central to understanding how this GPCR can be activated both in physiological and pathological conditions. Here we provide evidence that supports a model (Figure 7) whereby (i) SMO is initially addressed to the apical membrane before being transcytosed to the basolateral region, (ii) HH controls the post-endocytic fate of SMO likely by enhancing its recycling, especially in the basolateral region, (iii) very high levels of HH favors local trapping of SMO in the most basal region, and finally that (iv) these effects rely on an SMO phosphorylation-barcode determined by the sequential action of the PKA/CKI and FU kinases, with FU acting in a two-step manner.

Figure 7
Download asset Open asset
Model: hedgehog (HH) controls the fate of endocytosed Smoothened (SMO) via a multikinase phosphorylation barcode.

SMO is initially targeted (curved arrow) to the apical plasma membrane (presented in pink), where it is endocytosed (vesicle on the top, with a pink membrane). Endocytosed SMO can then be sent to the lysosome for degradation (vesicle on the left with gray membrane), recycled to the apical membrane, or transcytosed to the basolateral membrane (represented in blue), where it can, in turn, be re-endocytosed (followed by recycling or degradation). HH controls several of these steps via the graded action of the proteing kinase A (PKA)/casein kinase I (CKI) and Fused (FU) kinases. The double arrow represents the diffusion of SMO in the basolateral plasma membrane. In absence of HH (left panel), a large fraction of endocytosed SMO is targeted for degradation at the expense of its recycling (reduced recycling is indicated by gray arrows). In that situation, the C-terminal cytoplasmic domain of SMO is not phosphorylated and adopts a closed, inactive conformation. Conversely, in presence of HH (see middle and right panels) the degradation of endocytosed SMO is reduced (gray arrows), in favor of its recycling (green arrows), especially in the basolateral region. These effects of HH lead to an increased accumulation of SMO at the cell surface and require the phosphorylation of the C-terminal cytoplasmic domain of SMO by the PKA/CKI. This triggers a change in the conformation of this region, which is associated with the activation of SMO. In the presence of intermediate levels of HH (middle panel), SMO accumulation and/or activation reaches a threshold that in turn activates FU, which then retroacts on SMO, further enhancing the stabilizing and activating effects of the PKA/CKI kinases. The fact that FU is required for that effect in the apical region, suggests that the PKA/CKI also acts in this domain. In the presence of high levels of HH (right panel), FU is further activated and promotes the enrichment of SMO in the most basal region by phosphorylating its C-terminal cytoplasmic domain, which likely increases its clustering. This favors SMO’s trapping in the most basal region, which leads to high levels of HH signaling. The green arrow on the bottom right corner indicates a putative diffusion trapping mechanism, the gray bottom arrow indicates that recycling may be subsequently reduced.

The stabilization of SMO induced by HH could result from a reduction of its internalization or of its degradation after internalization. Our results indicate SMO endocytosis is little affected by HH and that HH acts on endocytosed SMO, shifting its fate toward recycling rather than degradation. This involves the phosphorylation of SMO by the PKA/CKI, whose effects are further enhanced by a secondary action of FU. Notably, phosphorylation of the β-adrenergic receptor by the PKA has also been shown to increase its recycling to promote its resensitization (Gardner et al., 2004).

We have previously provided evidence for a double positive SMO-FU feedback loop behind ‘high HH’ signaling: FU recruitment at the plasma membrane by SMO leads to the first level of FU activation, which in turn further activates SMO, which further increases FU activation. Here, tethering of FU to the apical membrane—but not to the basolateral one—is sufficient to ectopically promote both the stabilization of SMO at the cell membrane and the activation of low/medium HH targets. However, in contrast to what is seen in presence of very high levels of HH or when SMO is fully hyperphosphorylated (SMOPKA-SD FU-SD), apical FU does not promote high levels of HH signaling and does not lead to a basal enrichment of SMO. Together, these data support the existence of the second effect of FU on SMO, which would occur in the basolateral region and lead to a basal accumulation of SMO, promoting very high HH signaling. Note that the aPKC was also reported to positively modulate SMO activity and to favor (directly or indirectly) its basolateral localization (Jiang et al., 2014). However, contrarily to the phosphorylation of SMO by FU, the phosphorylation by the aPKC does not seem to affect the ‘high HH’-dependent basal localization of SMO, nor ‘high HH’ signaling.

Although the entire basolateral membrane is overall considered as a unique membrane domain in which proteins and lipids freely diffuse, many examples of membrane subregionalization exist (for review see Trimble and Grinstein, 2015). Here, we provide evidence that hyperactivated SMO can be enriched in the most basal region. It could in part be due to a partial reduction of its basal endocytosis (as suggested by our experiments), but it likely also involves other mechanisms that were shown to restrain the localization of transmembrane proteins (for review see Trimble and Grinstein, 2015). For instance, it could involve an active oriented displacement of endocytic vesicles carrying SMO and FU directly to the basal domain by the kinesin COS2, which is known to be required for high HH signaling and to transport SMO and FU along microtubules in cultured fly cells (Farzan et al., 2008). Alternatively, diffusion trapping or partitioning phenomena (for review see Trimble and Grinstein, 2015) are also known to lead to local protein enrichment, for instance in axons (Ashby et al., 2006). Here, the changes in SMO electrostatic charges, conformation, and/or clustering, which result from its hyperphosphorylation, could favor such processes (Shi et al., 2013; Zhao et al., 2007).

Regardless of the mechanism leading to this basal subpopulation of SMO, our data show that its presence is correlated to high levels of SMO activation and the basal gradient of HH. Although we cannot exclude that this basal localization is a consequence rather than a cause of SMO activation, for instance, a desensitization mechanism, our data along with published results strongly suggest that it is crucial to promote high HH signaling. Indeed, our results strongly connect SMO basal localization to its ‘high activation’ as: (i) high HH leads to SMO basal localization, (ii) the phosphorylation of SMO by FU is required for both SMO basal accumulation and its highest level of signaling activity, and (iii) on the contrary blocking FU in the apical region reduces both events. We propose that in presence of high levels of HH, SMO could be trapped in basal specialized lipid microdomains that enhance signaling, similarly to what has been shown for the regulation of several GPCR by lipids rafts (for review see Villar et al., 2016). This possibility is supported by many reports showing that both in flies and in mammals SMO responds to changes in its lipid environment (for review see Radhakrishnan et al., 2020 and Zhang et al., 2021) with an emerging key role of accessible cholesterol (Kinnebrew et al., 2021). Notably, in Drosophila WID cells, SMO also was shown to relocalize in response to HH to cholesterol-rich raft lipid microdomains in the plasma membrane, where it forms higher-order structures (e.g. oligomers) that are required for high HH signaling, but the Ap-Ba localization of these rafts was not addressed (Shi et al., 2013). Such microdomains could constitute signaling platforms acting on SMO structure (Zhao et al., 2021) enhancing the oligomerization of SMO and the HTC (Shi et al., 2011) and /or the interaction of SMO with specific HH signaling protein(s) (as seen with the EvC proteins in the cilia, see below [Dorn et al., 2012; Yang et al., 2012]).

Several lines of evidence point toward the conservation of a diffusion-trapping mechanism that would lead to the activation of SMO via its subcompartmentalization in specific membrane domains of the plasma membrane. Indeed, in response to Sonic HH, the entry of SMO in the primary cilium of mammalian cells, which also depends on its phosphorylation (Chen and Jiang, 2013), was shown to involve the lateral diffusion of SMO from the cell body into the cilium membrane (Milenkovic et al., 2009) and is followed by its spatial restriction in a ciliary distinct compartment named the EvC zone (Dorn et al., 2012; Yang et al., 2012). This involves the phosphorylation-dependent interaction of SMO with two components of this zone, EvC and EvC2, both acting downstream of SMO to alleviate the negative effects exerted by the Suppressor of FU (SUFU). In that respect, it is worth noting, that in the Drosophila WID, the negative effects of SUFU need to be suppressed by FU for high ‘HH signaling’ (Alves et al., 1998).

Materials and methods

Drosophila strains and genetics

Request a detailed protocol

All smo transgenes, except for the BAC (CH322-98K24) SNAP-smo, which was introduced at the landing site 9725 on 3 R (at 75A10), were introduced into the landing site (9738) on 3 R (at 99F8), using the PhiC31 integration system to ensure that they are expressed at similar levels (Bateman et al., 2006) by BestGene Inc the fuKO mutant was generated by inDroso functional genomics using CRISPR/Cas9 mediated genome editing (Jinek et al., 2012).

The flies were kept at 25°C with three exceptions: for shibirets experiments, the flies were kept at 18°C before being switched, at the third instar larval stage, to 30°C for 30 min; for rab5CA and hhts2 experiments, flies were kept at 18°C for 7 to 8 days before being switched to 29°C for 24 hr.

For the genotypes of the Drosophila strains used here, see Appendix 1.

WID SNAP labeling

Request a detailed protocol

For surface labeling: third instar larvae were dissected in Shields and Sang M3 Insect cell complete medium and incubated for 10 min at 25°C (or at 29°C for BAC (CH322-98K24) SNAP-smo experiments and 30°C for shits experiments) with the SNAP-Surface Alexa Fluor (NEB) in complete medium, then fixed 20 min at room temperature in 4% paraformaldehyde, and washed three times 10 min in PBS + 0.3% Triton (PBST). This was followed by immunolabeling.

For intracellular labeling: after surface labeling and fixation, discs were incubated with SNAP-Cell TMR-Star (NEB) and SNAP-Surface Block (NEB) for 20 min at 25°C (for intracellular labeling), followed by immunolabeling.

For immunolabeling, discs were incubated with the primary antibody overnight at 4°C. They were then washed three times with PBST and incubated for 2 hr at RT with the secondary antibody in PBST before another three washes of 10 min in PBST. Discs were then mounted in Citifluor (Biovalley).

For the different concentrations and details on the chemicals used see Appendix 1.

Fly wings

Request a detailed protocol

For wing harvesting, flies were collected in ethanol 70%, the wings were then dissected in water, and mounted in Hoyer’s medium. Pictures were taken with a Zeiss Lumar stereomicroscope and the AxioVision software.

Analysis of SMO trafficking

Request a detailed protocol

For the pulse-chase experiments, dissected WIDs were incubated for 10 min at 25°C with SNAP-Surface Alexa Fluor, rinsed, and incubated for 15 min with SNAP-Surface Block before fixation and immunolabeling.

Image acquisition, processing, and quantification

Request a detailed protocol

Images in Figure 1 and Figure 5—figure supplement 1B-B” were acquired using the confocal Zeiss LSM980 spectral Airyscan 2, 63× oil. All the other images were acquired using the confocal Leica confocal SP5 AOBS, 40× oil. XZ and YZ stacks of discs were acquired with sections every 1 µm. Microscope settings were chosen to allow the highest fluorescence levels to be imaged under non-saturating conditions. Image data were processed and quantified using ImageJ software (National Institute of Health).

An ImageJ macro was designed to quantify SNAP-SMO fluorescence from the Z projections (average intensity) of the stack (8 sections). Firstly, the macro determines the shape and limits of the disc. It then asks the user to select a point in the disc (selected at the A/P boundary) from which rectangular regions from apical to basal, with the same width, will be drawn across the disc. Following this, each region is divided into three smaller ones: the apical/subapical, lateral, and basal subregions. The macro considers apical/subapical and basal subregions to be 15% and 10%, respectively, of the disc’s thickness. The 15% value was fixed based on the immunostaining of the septate junctions by DLG. The 10% value is arbitrary and based on image analysis. The macro proceeds to measure the raw integrated density and the mean density of the different regions. Finally, with the help of CI immunostaining, four regions are selected across the disc (as shown in Figure 1—figure supplement 2C) to do the quantification. For macro description, see Appendix 2.

Statistics and data representation

Request a detailed protocol

Statistical analysis was carried out using GraphPad Prism 9. The sample size was chosen large enough (n≥8) to allow assessment of the statistical significance. Sample numbers are indicated in the figure and source data for each experiment. N-numbers indicate biological replicates, meaning the number of biological specimens evaluated (e.g. the number of wing discs). When comparing the A and P compartments within the same disc a paired t-test (for the mean intensities) or a Wilcoxon matched-pairs signed rank test (for the relative intensities) was used (Figure 1). When comparing different discs (in all other figures), a Mann-Whitney test was used. The p-values are shown in the corresponding source data.

Plasmids

All expression vectors, except for the BAC (CH322-98K24) SNAP-smo and the fuKO mutant, were constructed by the Gateway recombination method (Invitrogen). The BAC (CH322-98K24) SNAP-smo was generated from an attB-P[acman]-Ap BAC using recombineering mediated gap-repair (Venken et al., 2009). All mutated regions were verified by sequencing.

We used the Gateway Technology (Invitrogen following the manufacturer’s instructions) to introduce the SNAP-smoPKA-SD, SNAP-smoPKA-SD FU-SA, or SNAP-smoPKA PKA-SD FU-SD transgenes in the vector pUASt-GW-attB (constructed by A Brigui by insertion of the GW recombination cassette C3 at the EcoRI site of the pUASt-attB plasmid [GI EF362409]) for PhiC31 germline transformation, respectively. Prior to that, the PCR products obtained from the coding sequence (without the termination codon) of a smo wild type cDNA were inserted into pENTR/D-TOPO by directional TOPO Cloning. Mutations leading to the S to A and S to D changes of the PKA/CKI sites were inserted into pENTR/D-TOPO-snap-smo by replacement of a region with a similar region coming from smoPKA-SD/SA from Jia et al., 2004, leading to pENTR/D-TOPO-snap-smoPKA-SA/PKA-SD. The mutations leading to the S to A and S to D replacements of the FU phosphosites were introduced into pENTR/D-TOPO-snap smoPKA-SD by replacement of a region by a similar region coming from pENTR/D-TOPO-smo smoFU-SD/SA (Sanial et al., 2017) leading to pENTR/D-TOPO-snap-smoPKA-SD FU-SA and pENTR/D-TOPO-snap-smoPKA-SD FU-SD. All constructs were checked by sequencing the fragments produced by PCR and their junctions.

The BAC transgene was generated from an attB-P[acman]-Ap BAC. Briefly, a functional smo BAC (CH322-98K24) was modified using recombineering mediated gap-repair (Venken et al., 2009) to introduce snap cDNA at the N-terminus of smo after the codon encoding the Serine 33 at the end of the signal peptide sequence (Alcedo et al., 1996). pENTR/D-TOPO-snap-smo was used as a template to amplify the snap cDNA. For snap cDNA insertion into smo mRNA (present in CH322-98K24), first, the primers rpsL-neo/smo mRNA 360/F and rpsL-neo/smo mRNA 462/R (see Appendix File 1 key resource table) were used to amplify the rpsL-neo cassette. Second, the cassette was replaced by snap cDNA amplified using the primers pEnSnapSmo/smo mRNA 360/F and pEnSnapSmo/smo mRNA 462/R (see M&M table). Insertions were confirmed using different sequencing primers: smo mRNA 289/Seq/F, smo mRNA 512/Seq/R, Rpsl/neo/273/Seq/F, and snap/203/Seq/F (see Appendix File 1 key resource table).

For fuKO mutant generation, two guide RNAs (gRNA) and a double-strand DNA plasmid donor containing the fluorescent marker DsRed, were used to lead homology-directed repair in the fu locus. The gRNA1 anneals 545 bp upstream the ATG, while the gRNA2 anneals 26 bp upstream the TAG. Approximately, 2.6 Kb were removed, including the 5’UTR but not the 3’UTR, and replaced with the coding sequence of DsRed fluorescent marker, which provides a marker to distinguish fuKO larvae.

Appendix 1

Appendix 1—key resources table
Reagent type (species) or resourceDesignationSource or referenceIdentifiersAdditional information
Genetic reagent (D. melanogaster)apGal4Weihe et al., 2001FLYB: FBtp0084979FlyBase symbol: P{ap-GAL4.U}
Genetic reagent (D. melanogaster)UAS SNAP-smoWTSanial et al., 2017
Genetic reagent (D. melanogaster)UAS SNAP-smoPKA-SDThis paperGenerated by BestGene Inc
using the PhiC31
integration system
Genetic reagent (D. melanogaster)UAS SNAP-smoPKA-SD FU-SAThis paperGenerated by BestGene Inc
using the PhiC31
integration system
Genetic reagent (D. melanogaster)UAS SNAP-smoPKA-SD FU-SDThis paperGenerated by BestGene
Inc using the PhiC31
integration system
Genetic reagent (D. melanogaster)hhtsHeemskerk and DiNardo, 1994FLYB: FBal0031490
Genetic reagent (D. melanogaster)shibiretsvan der Bliek and Meyerowitz, 1991FLYB: FBal0015610
Genetic reagent (D. melanogaster)apGal4 Gal80tsGift from F Schweisguth
Genetic reagent (D. melanogaster)UAS-YFP-rab5CAHugo J. Bellen,
Baylor College of Medicine		RRID:BDSC_9773
Genetic reagent (D. melanogaster)UAS smurf RNAiPerkins et al., 2015RRID:BDSC_40905
Genetic reagent (D. melanogaster)UAS mCherry-NVR1-GFP nanobodyHarmansa et al., 2017RRID:BDSC_68175
Genetic reagent (D. melanogaster)UAS mCherry-T48-GFP nanobodyHarmansa et al., 2017RRID:BDSC_68178
Genetic reagent (D. melanogaster)UAS GFP-fusedRuel et al., 2003
Genetic reagent (D. melanogaster)dpp-LacZ, apGal4Gift from R Holmgrendpp-LacZ RRID:BDSC_12379
Genetic reagent (D. melanogaster)w1118Hazelrigg et al., 1984FLYB: FBal0018186FlyBase symbol: Dmel\w1118
Genetic reagent (D. melanogaster)BAC SNAP- smoThis papersnap insertion via
recombineering mediated
gap-repair. Fly generated
by BestGene Inc using
the PhiC31 integration system.
Genetic reagent (D. melanogaster)fuKO/FM0BarThis paperfu knockout (KO) mutant
fly generated via CRISPR/
Cas9 by inDROSO
Antibodyanti-PTC (mouse monoclonal)DSHB, 
Martín et al., 2001		DSHB Cat# Drosophila Ptc (Apa 1), RRID:AB_528441IF(1:50)
Antibodyanti-EN (mouse monoclonal)DSHB, University of Iowa, USA, 
Patel et al., 1989		DSHB Cat# 4D9 anti-engrailed/invected, RRID:AB_528224IF(1:50)
Antibodyanti-DLG (mouse monoclonal)DSHB, Parnas et al., 2001DSHB Cat# 4F3 anti-discs large, RRID:AB_528203IF(1:50)
Antibodyanti-SMO (mouse monoclonal)DSHB, Lum et al., 2003DSHB Cat# Smoothened (20C6), RRID:AB_528472IF(1:100)
Antibodyanti-Rab7 (mouse monoclonal)DSHB, University of Iowa, USA,DSHB Cat# Rab7, RRID:AB_2722471IF(1:25)
Antibodyanti-FU (rabbit polyclonal)Gift from Ruel et al., 2003IF(1:100)
Antibodyanti-β-Galactosidase (rabbit polyclonal)MP Biomedicals085597-CFIF(1:100)
Antibodyanti-CI (rat monoclonal 2A1)Gift from Motzny and Holmgren, 1995IF(1:5)
Antibodyanti-mouse IgG Alexa Fluor Plus 555 (goat polyclonal)Thermo Fisher ScientificThermo Fisher Scientific Cat# A32727, RRID:AB_2633276IF(1:200)
Antibodyanti-rat IgG Alexa Fluor Plus 647 (goat polyclonal)Thermo Fisher ScientificThermo Fisher Scientific, Cat# A-21247, RRID: AB_141778IF(1:200)
Antibodyanti-rabbit IgG Alexa Fluor Plus 555 (goat polyclonal)Thermo Fisher ScientificThermo Fisher Scientific, Cat# A32732, RRID: AB_2633281IF(1:200)
Antibodyanti-rabbit IgG Alexa Fluor Plus 488 (Goat polyclonal)Thermo Fisher ScientificThermo Fisher Scientific, Cat# A32731, RRID: AB_2633280IF(1:200)
Antibodyanti-rabbit IgG Alexa Fluor Plus 647 (goat polyclonal)Thermo Fisher ScientificThermo Fisher Scientific, Cat# A-21245, RRID: AB_253581IF(1:200)
Antibodyanti-mouse IgG cross adsorbed 488 (chicken polyclonal)Thermo Fisher ScientificThermo Fisher Scientific, Cat# A-21200, RRID: AB_2535786IF(1:200)
Antibodyanti-mouse IgG Dylight 649 (donkey polyclonal)Jackson Immuno715-495-151IF(1:200)
Recombinant DNA reagent (D. melanogaster)BAC smoBAC Resources PACCH322-98K24https://bacpacresources.org/home.htm
Recombinant DNA reagentBAC snap-smoThis papersnap cDNA introduced in
CH322-98K24 after
Serine 33 codon of smo
Recombinant DNA reagentpENTR/D-TOPO-snap smoSanial et al., 2017Template for snap amplification
Recombinant DNA reagentpENTR/D-TOPO-snap smoPKA-SDSanial et al., 2017
Recombinant DNA reagentpENTR/D-TOPO-snap smoPKA-SD FU-SASanial et al., 2017
Recombinant DNA reagentpENTR/D-TOPO-snap smoPKA-SD FU-SDSanial et al., 2017
Recombinant DNA reagentpUASt-GW-attBBrigui et al., 2015
Sequence-based reagentrpsL-neo/smo mRNA 360 /FEurofins GenomicsPCR primersAGGTTGCGATCTTATGCCTGTGGGTGGTCGCAGACGCATCGGCCAGTTCGggcctggtgatgatggcgggatcg
Sequence-based reagentrpsL-neo/smo mRNA 462 /REurofins GenomicsPCR primersAGTTCCACATCCGACTGCTGCGCACTTGCGGGCGTTGTGCTGCCGAACTTGGCtcagaagaactcgtcaagaaggcg
Sequence-based reagentpEnSnapSmo/smo mRNA 360 /FEurofins GenomicsPCR primersAGGTTGCGATCTTATGCCTGTGGGTGGTCGCAGACGCATCGGCCAGTTCG
Sequence-based reagentpEnSnapSmo/smo mRNA 462 /REurofins GenomicsPCR primersAGTTCCACATCCGACTGCTGCGCACTTGCGGGCGTTGTGCTGCCGAACTTGGC
Sequence-based reagentsmo mRNA 289/Seq/FEurofins GenomicsPCR primersGACTCGCCTCTGGCAAATGG
Sequence-based reagentsmo mRNA 512/Seq/REurofins GenomicsPCR primersTGCCCTTCTTGGCGTACAGTCGG
Sequence-based reagentRpsl/neo/273/Seq/FEurofins GenomicsPCR primersGAACTCCGCGCTGCGTAAAGTATG
Sequence-based reagentsnap/203/Seq/FEurofins GenomicsPCR primersCCTACTTTCACCAGCCTGAG
Chemical compound, drugSNAP-Surface Alexa Fluor 488New England BiolabsCat#S9124S3.3 µM
Chemical compound, drugSNAP-Surface Alexa Fluor 546New England BiolabsCat#S9132S3.3 µM
Chemical compound, drugSNAP-Surface BlockNew England BiolabsCat#S9143S2 µM
Chemical compound, drugSNAP-Surface Alexa Fluor 647New England BiolabsCat#S9136S3.3 µM
Software, algorithmFijiSchindelin et al., 2012Fiji, RRID:SCR_002285http://fiji.sc/
Software, algorithmAdobe PhotoshopAdobeRRID: SCR_014199https://www.adobe.com/products/photoshop.html
Software, algorithmAdobe IllustratorAdobeRRID:SCR_010279http://www.adobe.com/products/illustrator.html
Software, algorithmAdobe InDesignAdobeRRID:SCR_021799https://www.adobe.com/fr/products/indesign.html
Software, algorithmPrism – GraphPad 9Graph Pad SoftwareRRID:SCR_002798https://www.graphpad.com/scientific-software/prism/

Data availability

All data generated or analysed during this study are included in the source data. The script is provided in Source Code 1.

References

  24. Book
    1. Ingham PW
    (2022) Chapter one - Hedgehog signaling
    In: Soriano PM, editors. Current Topics in Developmental Biology. Academic Press. pp. 1–58.
    https://doi.org/10.1016/bs.ctdb.2022.04.003

Decision letter

  1. Vilaiwan M Fernandes
    Reviewing Editor; University College London, United Kingdom
  2. Claude Desplan
    Senior Editor; New York University, United States

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

[Editors' note: this paper was reviewed by Review Commons.]

https://doi.org/10.7554/eLife.79843.sa1

Author response

General Statements

First of all, we would like to thank the reviewers for their insightful comments;

In summary, taking into account all the reviewers’ comments and advice we have now included a substantial amount of novel data which rely in a large part, on the construction of two novel genetic tools: a BAC SNAP-smo construct that allows expression of SNAP-SMO at endogenous levels and the first fu KO mutant (generated by CRISPR).

We now propose the following main changes in the Figures (the different sections of the text have been changed accordingly):

The Figure 1 has been completely reorganized.

(i) For simplification, we have removed the former panels A-A”, B-B”, E, E’ (data on Total SNAP-SMO) and E”. We have moved the data on Intra SNAP-SMO to the figure supplement 2.

(ii) We have added diagrams of the wing imaginal disc.

(iii) We now provide better images of SMO subcellular localization and apico-basal distribution thanks to an enhanced resolution confocal system: panels B-B” and C-C” that replace panels F-F” and G-G” respectively clearly show the basal localization of SMO in the “+HH cells “ (in the P and the CI-A region near the A/P).

(iii) We have kept the graph (former panel H) on the relative intensity in the four regions and added the graph on the relative intensity from the former sup Figure 1E.

(iv) We now show that the basal localization of SMO is suppressed when HH is inactivated (hhts allele, panels F-F”).

It is now associated to three figure supplements.

  • Figure 1—figure supplement 1 shows that (i) SNAP-SMO carried by a BAC is fully functional and (ii) that overexpression of SNAP-SMO has no effect on HH signaling.

  • Figure 1—figure supplement 2 includes (i) a schematic description of the labeling procedure (former Sup 1 F-F”) and of the quantification method (former Sup1 A), (ii) the data on Intra SNAP-SMO, and (iii) the validation of the inactivation the hhts mutant at restrictive temperature.

  • Figure 1—figure supplement 3 demonstrates that the HH-induced basal localization of Surf SNAP-SMO is not an artefact due to its fusion to the SNAP tag nor its overexpression. It shows the distribution of (i) immunolabeled endogenous SMO (former Figure S1 B-B” and C’ panels, the panel B has been removed for simplification as it did not provide novel informative data), (ii) Surf SNAP labelled SNAP-SMO expressed at endogenous levels from.

The Figure 2 is now associated to the Figure 2—figure supplement 1 that includes the former panels of Figure S2 and to which we added a control at 18°C (F) and experiments that show that blocking early endosomes (by the expression of RAB5CA) has the same effect on endogenous SMO or SNAP-SMO expressed at endogenous levels (G-G”’, H-H”’) that it has on overexpressed SNAP-SMO.

The Figure 3 is now associated to a sup Figure, Figure 3—figure supplement 1, in which we show that after the chase, Surf SNAP-SMO localizes with the vesicular marker RAB7 (A-A”’, B-B”’) and provide graphs in which we quantify the reduction of Surf SNAP-SMO during this chase(C-C’).

In Figure 4—figure supplement 1 (former S3), we have added (i) a graph showing the reproducibility of the distribution of Surf SNAP-SMOWT in the Figures 1 and 4 and (ii) images showing the effects of mimicking or blocking phosphorylation of SMO on HH signaling.

The Figure 5 is now associated to two figure supplements:

  • Figure 5—figure supplement 1 includes the former S4 panels A-A’, B-B”, to which we have added a control XZ section showing GFP-FU without the GRAB tether.

  • Figure 5—figure supplement 2 includes the former S4 panels C-C”’, D-D”’ and E-E’ to which we added data that shows that apical grabbing of GFP-FU leads to an increase in Surf SNAP-SMO (expressed at endogenous levels from a BAC).

In Figure 6, for simplification we have removed the data on CI-F and on dpp. We have also added data showing that tethering FU to the apical region is not suppressed in the absence of endogenous FU (CRISPR KO mutant of fu).

It is now associated to a figure supplement (Figure 6—figure supplement 1) showing the effects of the Nvr1 and T48 tethering of GFP-FU on dpp and a scheme on the construction of the fu KO mutant.

In Figure 7, the model has been simplified.

Responses to reviewer #1:

Specific points:

1. The phenotypes under study are subtle – we're talking about a shift of a few percent of Smo from the most apical region to basal. The work appears to have been carefully done, though, with sufficient replicates to allow these small differences to be detected using statistics. However, many of the conclusions are based on a subsequent comparison of the differences of differences (e.g. "a reduction of 46% in the far anterior cells but of only 25% in the posterior cells" p. 9). Given the magnitude of the uncertainty (standard deviation) in the individual measurements, it's not always clear that these differences between the differences are really significant, and this is not tested statistically. In light of the subtlety of some of the effects, applying a statistical analysis to these comparisons would strengthen the conclusions.

We have now statistically compared the % of decrease in Surf SNAP-SMO between t0 and t15 (calculated as (t0-t15)/t0) in the FA (-HH) and P regions (+HH). This confirms the significance of these comparison. This comparison is shown in the Figure 3—figure supplement 1C, C’.

Concerning the shits experiments, we removed the part on comparing the % (lines 147 to 153) as it was unnecessary long, and the effects presented in the graph C were sufficient and statistically significant.

2. The images in Figure 1 show a pronounced enrichment of total SNAP-Smo in the apical domain of both far A and P cells, which is not seen when staining for endogenous Smo (shown in Supplement). This could give a bit of a misleading impression about the "normal" distribution of Smo, which is more uniform. Can the authors comment on this?

We have now built a transgenic line carrying a smo BAC construct in which we inserted a SNAP tag at the exact same position than in the UAS SNAP-smo construct. We show that this construction is fully functional (now shown in Figure 1—figure supplement 1 A, B-B”, see also text lines 114-116). Labelling of Surf SNAP-SMO expressed in this condition is shown in Figure 1—figure supplement 3C-C”, see also text lines 158-159. No strong apical enrichment is seen. As SNAP-SMO is initially targeted to the apical membrane before being endocytosed, a possibility is that apical endocytosis might be a limiting step when high amounts of SNAP-SMO are synthetized, leading to its apical accumulation. However, SNAP-SMO expressed for the BAC at endogenous levels is -as the case for endogenous SMO and overexpressed Surf-SNAP-SMO-, enriched in the basolateral region of the cells in the Ci-A and P regions, which is the main subject of this work. Note that we could not correctly quantify these images due to a low signal to noise ratio and strong photobleaching, which prevented us from doing Z stacks.

We now also show that endogenous SMO and SNAP-SMO expressed from the BAC at endogenous levels are also initially targeted to the apical region where they are endocytosed in a RAB5 dependent manner, indicating that (SNAP-SMO overexpression does not affect its apical targeting nor its apical endocytosis). This is now shown in Figure 2—figure supplement 1G-G”, H-H”. See also the text lines 201-204.

Altogether these results strongly validate the use of this UAS SNAP-smo construct.

See also the response to reviewer 2, point 1 p8.

3. Although everything is quantified, in a couple of cases the quantification doesn't seem to match the eye test based on what is shown in the images, or the staining is inconsistent from figure to figure. Can the authors comment on the following points?

3A. The graph in S1D' shows no difference in intracellular SNAP-Smo mean intensity distribution between FA and P cells. In the matching image in Figure 1D', it looks a lot like intracellular Smo levels are much lower in P and Hh-responding A cells than in far A cells – both in the number of Smo+ punctae and in their staining intensity.

We thank the reviewer for this comment. We apologize for what is an error in the images frame and we now provide the images with a larger section of the P region and with the Ci staining that allows identification of the different regions. See Figure 1—figure supplement 2D,D’.

3B. The differences in surface SNAP-Smo distribution in far A (apical dot like structures, presumably endocytic vesicles) versus P cells (on membranes, fewer vesicles) shown in Fig, 1C and Dare not consistent in other staining’s of the same genotype (e.g. Figure 2A, Figure 4A).

Some differences reflect the biological variation from one disc to another, which justifies the importance of systematic quantification. In this case, the estimated average intensity of SNAP-SMO labeling was determined after quantification of Z projections, each composed of eight sections, from between 10 and 23 discs and provides more robust information than a single image.

Moreover, we have now also compared the apico-basal distribution of SNAP-SMO in the experiments quantified for the Figures 1 and 4 and this confirms that the apico-basal distribution of SNAP-SMO is in fact very reproducible. We have added this data as Figure 4—figure supplement 1E.

Note also that some comparisons are not equivalent, such as the comparison between (i) Z projection and Z sections, (ii) experiments done at different temperature (Figure 2 compared to Figures 1, 3 and 4), and now (iii) experiment done with different confocals as the confocal Zeiss LSM980 spectral Airyscan (63X, reconstructed Z sections) in the novel (Figure 1) and the confocal SP5 AOBS images (40X, direct Z imaging) in almost all the other Figures.

4. The experiments comparing apical versus basolateral trapping of FU are a bit complicated to interpret. What is the evidence that NVR1 leads to basolateral enrichment of FU? Put another way, how does FU localization when co-expressed with NVR1 compare to its distribution when just expressing FU alone?

We have now added in an image of GFP-FU in Figure 5—figure supplement 1C. Comparison with the Figure 5B’, D’ and Figure 5—figure supplement 1C’, D’ confirms that GFP-FU is indeed relocalized by the Nvr1 and T-48 traps, respectively. See text lines 285-286.

5. One caveat to this experiment is that apical targeting of FU will presumably lead to a much higher FU density since the apical domain is fairly small. Spreading the same amount of FU over the entire basolateral domain will result in much less concentration of the protein. Since Fu clustering plays an important role in its activation, can the authors rule out that the basolateral trapping doesn't have much effect because Fu activity is much lower in this condition?

We cannot formally exclude the possibility raised here to explain the ectopic ptc expression (but not the reduced en expression). However, to our knowledge, there is no data indicating that increased levels of FU could promote its dimerization nor its activation (on the contrary, overexpression of FU has no effect on HH signaling for instance see Claret et al.). On the contrary, FU levels are decreased in response to HH (Ruel et al., 2013). Moreover, an effect via FU clustering would not explain the effect of T48-anchored FU on SMO localization as (Shi et al., 2011) also showed that dimerized FU activates CI independently of SMO.

Note also that many labs (Kalderon, Jian, our …) have overexpressed FU, with or without a tag, even at very high levels, and activation of FU or the pathway has not been observed.

Even if this apical tethering would act indirectly by increasing FU clustering, this would not change the fact that it allows FU-GFP to activate the medium to low HH-targets, but not high levels of the anterior en. This also indicates that a secondary event is necessary for the highest level of activation.

We did not see the points 6 and 7.

8. The conclusion that a basolateral function of FU is missing in the apical targeting experiment would be clear if the experiment was done in a fu mutant background. However, in wild-type cells, shouldn't there be the usual abundance of endogenous FU to carry out the function of FU in the basolateral domain?

We have now repeated this experiment in a fu null mutant background (thanks to a fuKO mutant that we built by the CRISPR method). In this context, apical trapping of GFP-FU has similar effect in the absence as in presence of the endogenous FU protein. They indicate that apically trapped GFP-FU does not act via endogenous FU and reinforce our conclusion that FU is required basally for the full activation of High HH targets. Note that we also checked that GFP-FU rescues the effects of the fuKO allele. These data are presented in Figure 6E-H. See also the text lines 316-332.

Minor comments

p 9 – It's not clear to me why the authors conclude that endocytosis of Smo occurs all along the apical-basal axis in Figure The data in E suggest that endocytosis is primarily occurring apically.

Our data show that a strong block of endocytosis (Rab5 CA) leads to an accumulation of SMO in the apical region almost exclusively, revealing that SMO is initially targeted to the apical region where it undergoes a first endocytosis. However, partial blocking of SMO endocytosis (under shi inactivation conditions) leads to an increased accumulation of SMO in both the apical and the basolateral region, which reveals that SMO redistributed to the basolateral region can undergoes a secondary endocytosis.

Figure S4E – What condition is being analyzed? I couldn't find it in the legend.

We apologized for this lack of information. This point has now been added in the legend of this Figure now called Figure 5—figure supplement 2.

Significance

There have long been bits and pieces of data about Smo cycling to the plasma membrane and being internalized, and Ptc affecting this process to keep Smo levels low in the absence of Hh. However, there is no clear picture of the route that Smo takes in response to Hh, at least in flies. (In mammals there is good evidence that Smo goes to cilia upon its activation.) The authors have used an innovative approach to try to address this, by fluorescently labeling surface Smo and trying to follow its fate. While subtle, the conclusions that Smo is delivered to the apical domain, endocytosed, and in response to Hh moves basally are generally convincing, with some exceptions noted above. The data showing that this is controlled by PKA/CKI and FU phosphorylation of Smo are also clear. Aside from these mechanistic experiments, the rest is more descriptive and, as the authors note, the results don't clearly distinguish whether the basal localization of Smo is a consequence rather than a cause of Smo activity. Nonetheless, it's a quality manuscript that will be of interest to many in the Hh field, and likely to people who are more generally interested in GPCR signaling.

Concerning the distinction between a consequence or a cause, we formally agree. We have tried to test the “cause hypothesis” by using the GRAB system to target mutant and wild-type forms of SMO-GFP. However, we did not succeed in relocalizing SMO as on the contrary, we observed that SMO-GFP had the ability to relocalize the anchors T48 (Author response image 1) and NVR1. We therefore see no other possibility to discriminate between these two hypotheses.

Author response image 1
Download asset Open asset

We would however like to mention that while numerous published data show a strong correlation between SMO cell surface localization and SMO activation (i. e. (i) HH leads to both SMO activation and its relocalisation at the cell surface , (ii) SMO relocalisation to the cell surface is sufficient to activate SMO and (iii) blocking SMO phosphorylation by the PKA prevents both its localization at the cell surface and its activation) none of them- to our knowledge- show that this relocalisation is in fact required for its activation. Similarly, all our results strongly connect SMO basal localization to its “high activation” without formally proving that the second requires the first as: (i) high HH leads to SMO basal localization, (ii) the phosphorylation of SMO by FU is required for its basal accumulation, (iii) blocking FU in the apical region reduces both basal localization of SMO and its high activation. We have now added data (see Figure 4—figure supplement 1 F-F”, G-G”, lines 266-271) confirming that the phosphorylation of SMO by FU is required for expression of the high HH target gene en and some text in the discussion lines 390-394. This further links the connection between SMO basal localization and activation. See also our response to reviewer 3 p19.We want to stress that the significance of our data also has to be evaluated in light of what is known on the HH reception: First, HH is known to form a basal gradient supported by cytonemes and this basal gradient is necessary for high HH signaling (Bischoff et al., 2013). Second, the HH co-receptor PTC, is present both in the apical and basolateral/basal regions and was recently shown to inhibit SMO by removing accessible cholesterol from the outer leaflet of the plasma membrane, with HH binding to PTC blocking the passage of these cholesterol molecules. As the apical and the basolateral membranes behave as distinct membrane regions, with no exchanges of lipids and proteins; the basal molecules of PTC that receive HH from the basal gradient should need basal SMO to fully activate the pathway. Given all this and our data, we propose that FU would be required to send SMO to the basal region; due to the high levels of HH in that region, SMO would then be fully activated, leading to an increased activation of FU, which would then inhibit its partners and targets COS2 and SUFU to fully activate CI.

Responses to Reviewer #2:

Major comments:

1. While SNAP labeling technique is a powerful method for probing protein localization, the main question of this manuscript is whether overexpressed SNAP-Smo reports the true localization pattern of endogenous Smo. As shown in Figures 5B" and S1B as well as in the published literature, endogenous Smo does not show significant apical accumulation but is more basolateral, which is very different from the SNAP-Smo shown in Figure 1B. The authors claimed that SNAP-Smo is functional in vivo by citing Sanial et al. (2017) that overexpressed SNAP-Smo rescues adult wing defects caused by smo RNAi. However, Sanial et al. (2017) did not completely rule out whether an N-terminal SNAP tag would lead to Smo activation, or whether overexpressed SNAP-Smo represented an activated form of Smo. If this is the case, the studies described in this manuscript are less important for understanding the impact of the subcellular localization of Smo in Hh signaling activation. Furthermore, the dimerization requirement for Smo activation further complicates the situation. To vigorously establish the relevance of polarized SNAP-Smo distribution in Hh signaling, it is especially important to express SNAP-Smo at levels comparable to endogenous Smo or knock in the SNAP tag into the smo locus.

The effect of the tagging of SMO with the SNAP-tag and of SNAP-SMO overexpression on SMO apico-basal distribution is addressed in our responses to reviewer 1 point 2. In summary, using a recombinant BAC carrying a SNAP-SMO fusion (expressed under the endogenous promoter of smo) we show that, similarly to what is seen with overexpressed SNAP-SMO, SNAP-SMO expressed at endogenous levels (i) fully rescues a smo loss of function (viability, wild-type wings and targets, in Figure1—figure supplement 1A, B-B”, (see text, lines 113-116) and (ii) is initially addressed apically where it is endocytosed in RAB5 endosomes Figure2-supplement 1H-H”, text lines 202-204).

Concerning the possibility that SNAP-SMO overexpression would promote its activation, we provide new data, which shows that overexpression of SNAP-SMO has no effect on en expression nor the presence of the CI-A domain, two markers of high HH signaling. These results are now shown in Figure1—figure supplement 1C-C” (text lines 115-116) and confirm publications by numerous labs (Beachy, Kalderon, Thérond, Jia, Jiang, our lab etc…) showing that overexpression of SMO is not sufficient to activate it. These reports are based on the overexpression of various UAS smo constructs, untagged or with various tags inserted in various loci (N term or C term).

2. The authors claimed that newly synthesized SNAP-SMO proteins display "dot-like structures" when released from the endocytic block, likely corresponding to endocytic vesicles. This claim is untenable unless co-staining is performed with endocytic markers.

We now show colocalisation between some of these Surf SNAP-SMO “dot-like structures" and the endosomal marker RAB7. This is presented in Figure 3—figure supplement 1 A-A”, B-B”. See also the text lines 223-224.

Furthermore, there are no apparent differences between the mean intensities shown in Figures 3A and 3B. In some cases (lateral and basal), the mean intensity of the surf SNAP-SMO even increased after the 15-minunte chase, which is not consistent with the statistical analysis in Figure 3C.

As we now add in the Figure 3 legend, the discs were treated in the same conditions and all the images were acquired under the same conditions. However, the dynamic range was automatically adjusted for each image independently by Image J for a better visualization of the subcellular localization of SNAP-SMO. To allow a better eye comparison, we have now replaced these images by images in which the dynamic range is normalized.

3. The XZ images of the apical-basal distribution of SNAP-Smo and endogenous Smo shown in different figures are inconsistent. For example, surf SNAP-Smo accumulated apically in Figures 1D, 1G and Figure 2D, but not in Figures 2A and Figure 4A.

As explained in our response to reviewer 1 point 2, only the images 1D and 4A can be compared for Surf SNAP-SMO and we have now added in Figure 4—figure supplement 1E data showing that both sets of experiments display very similar distribution along the apico-basal axis both without and with HH, even if differences in the contrast of the images, may be misleading for the eye.

Furthermore, in Figure 5B', significant accumulation of endogenous Smo was observed when overexpressed Fu was captured in the apical region. However, the upregulation and basal accumulation of endogenous Smo in Figure S4C' were relatively weak under the same conditions. Although immunostaining results for SNAP-Smo or endogenous Smo may vary, it is necessary to provide more appropriate results consistent with statistical data.

We believe that there is a misunderstanding as in Figure 5B’, GFP-FU is shown, not SMO and it is enriched in the apical region due to its anchoring by T48. Endogenous SMO, which is shown in Figure 5B”, shows no apical enrichment. As mentioned, above (response to reviewer 1, point 4) we have added as a control the subcellular localization of GFP-FU in the absence of an anchor Figure 5-supplement 1C (text line 285-286). Note also that Figure 5 shows XZ sections in the anterior region (-HH situation, corresponding to the FA region); while in the previous Figure S4 (now Figure 5—figure supplement 1) the XZ sections are in the posterior region (+HH situation): this explains why the effect are weak in the P compartment where SMO is already stabilized and activated.

4. Capture of overexpressed GFP-Fu in the apical region resulted in a marked increase in endogenous Smo proteins, which accumulated slightly in the basal region, and downregulation of high Hh signaling targets, such as en. In contrast, capturing overexpressed GFP-Fu in the basolateral region had no apparent effects on Smo levels or apical-basal distribution. The authors concluded that apical-captured GFP-Fu activates and stabilizes Smo on the cell surface, while basal Fu is required for basal enrichment of Smo and activation of high-level Hh targets. However, previous studies have shown that Fu can dimerize. Therefore, the captured GFP-Fu may indirectly affect the localization of endogenous Fu, thereby affecting its function on Hh target expression. The amount and apical-basal distribution of endogenous Fu should be examined to eliminate this possibility.

As mentioned in our response to reviewer 1, point 8, we now show that the effects of the apically trapped GFP-FU on HH signaling do not depend on the presence of the endogenous FU protein.

Note also that we have no way to assess the localization of the endogenous FU due the lack of a tool that would allow us to distinguish endogenous FU from overexpressed GFP-FU.

5. Since this study highlights the effect of Hh and Fu on Smo protein levels and their apical-basal distribution, it is necessary to examine whether reduced expression of hh or fu affects the amount and polarized distribution of endogenous Smo.

To ensure that the changes in the apico-basal distribution of Surf SNAP-SMO were indeed due to HH, we performed the same experiment when HH function was inactivated, using a thermosensitive allele of hh. Under HH inactivation, Surf SNAP-SMO is no longer accumulated in the P and the anterior cells near the A/P, and is no longer enriched in the basolateral region of these cells. We have added this result in Figure 1F-F” and Figure 1—figure supplement 2F-F” (text lines 162-169).

Minor comments:

1. In the Discussion, the authors stated that aPKC, a kinase previously known to regulate polarized distribution of Smo, does not affect the "high Hh"-dependent basal localization of Smo. However, no relevant data are provided in the paper and no references are cited.

We apologize for this confusion: we meant that the aPKC was a kinase previously known to regulate polarized distribution of SMO but that it is not known whether the loss of aPKC affects the "high Hh"-dependent basal localization of Smo. We have changed the text accordingly to make it accurate (lines 368-370) The reference is Jiang et al., PNAS 2014 and has been added (line 368).

2. Overexpression of the captured GFP-Fu protein in the dorsal compartment should have no effects on Ci-FL levels in cells localized in the ventral compartment. However, in Figure 6D, Ci-FL levels and activation status (i.e. CiA) in the ventral compartment appear to be altered by overexpression of captured GFP-Fu in the dorsal compartment.

In Author response image 2, mostly the dorsal part of the wing pouch is shown. We apologize for that and you can see the full wing pouch in the image to the right. However, on the revised version of the MS to simplify this part we have decided to delete these data.

Author response image 2
Download asset Open asset

Nature and significance of the advance (e.g. conceptual, technical, clinical) for the field.The idea is new, but the evidence in the current manuscript is not convincing.

We hope that the additional data that we now provide with more control experiments and which include results with the BAC and the fuKO mutant make the novel manuscript convincing. See also our above response to the Reviewer 1 “Significance” section.

Context of the existing literature (provide references, where appropriate).

Li et al. (2016) PLoS Biol. 14, e1002481.

Li et al. (2018) Sci. Signal. 11, eaan8660.

Malpel et al. (2007) Dev. Biol. 303, 121.

Sanial et al. (2017). Development. dev.144782.

Shi et al. (2013). J. Biol. Chem. 288, 12605.

– Audience that might be interested in and influenced by the reported findings.

Developmental biologists and cell biologists who are interested in mechanisms of signal transduction

– Field of expertise with a few keywords to help the authors contextualize your point of view. Indicate if there are any parts of the paper that you do not have sufficient expertise to evaluate.

Development, signal transduction, endocytic transport, polarized protein distribution

Responses to reviewer #3:

Key findings:

HH signal causes more SMO protein to be central and basal in wing imaginal disc cells, with reduction apically.

Blocking endocytosis in two ways causes SMO accumulation in apical regions of cells. HH does not appear to affect SMO endocytosis.

Pulse-chase labeling showed movement of SMO from apical to more central and basal regions. Since this looks HH-independent, the authors conclude that HH regulates SMO abundance by altering degradation rather than endocytosis.

SMO altered to prevent cytoplasmic tail phosphorylation, or to mimic it, shows that apical-basal SMO localization is regulated by phosphorylation.

The most striking result in the paper is that apically localized over-produced FU kinase causes SMO to become active with respect to medium (not high)-level HH target gene activation and to accumulate at cell membranes.

Lines 351-2: "we cannot exclude that this basal [SMO] localization is a consequence rather than a cause of SMO activation".

See our response to reviewer 1 “significance” section and in the discussion lines 390-394.

The authors propose that basal localization driven by high HH brings SMO into a membrane environment where it gains activity and activates high-HH target genes.

Line 24: "plasma membrane subcompartmentalisation". While that may be correct, most of the data do not discriminate between plasma membrane and internal membranes. This makes it harder to make a model, such as the one in Figure 7.

We have deleted “plasma membrane subcompartmentalisation” from our summary.

Specific comments:

line 47 There are reviews more current about Hh and cancer therapies than the one mentioned, Briscoe & Therond 2013.

We have replaced this reference by Ingham PW. Curr Top Dev Biol. 2022 Line 47.

line 67 "cytotail" is jargon, not English, and should be replaced.

We have replaced it by “cytoplasmic C-terminal tail”everywhere.

line 99 "phosphomutants" is jargon, not English, and should be replaced.

We have replaced it by “phosphorylation mutant”.

Figure 1 would be helped by having a diagram showing an imaginal disc and the regions within it referred to in the text. This would support Figure 7 as well. Figure S1A is closer but still not a full-context view.

Such diagrams are now shown in Figure 1A and A’.

Also in Figure 1, DLG is used to stain. A search of the text reveals that the first mention of DLG is in line 405 in the methods, and what it stains is only mentioned at line 440. The use of it should be explained in text and legend with reference to Figure 1.

We thank the reviewer for noting that and this has been done. See line 134.

lines 118-120 Shouldn't there be a third category of cells near the A-P boundary that receive HH but do not produce it? This would be, I suppose, the region indicated by the purple A in Figure S1. This question is especially relevant since it is unclear whether any of the cells shown in Figure 1A-D represent the FA region-there is no labeling to tell us. And I see that in line 285 these cells are finally mentioned.

As mentioned at the beginning of this document, and following this very useful comment and the following suggestion to simplify Figure1, we have now totally reorganized this Figure, its three figure supplements and the corresponding text section. We now directly introduce the presentation of these different regions along the A/P axis.

line 122 Yes, stronger apical accumulation of SMO is visible in A, B, C, and D but…

line 123 I fail to see the increased basal accumulation; indeed, I can scarcely see any basal accumulation.

We now provide -in the novel Figure 1 XY and XZ- images taken with a Zeiss LSM980 confocal with a spectral Airyscan 2, 63X (rather than with a Leica confocal SP5 AOBS, 40X). These novel images clearly show the basal accumulation of Surf SNAP-SMO in presence of HH. However, given that the confocal Zeiss LSM980 spectral Airyscan acquisitions are lengthier and do not allow us to perform direct XZ sections, all the quantifications and most of the other images still correspond to the SP5 images (except for the Figure 5—figure supplement 1B-B”).

I do not know how to reconcile the A-D panels showing staining with the chart in E, which shows the 50% increase in P vs FA regions. Is that because the FA regions are not shown in A-D? E' and E' show scarcely any staining in basal regions, which is what I see in the stained discs. Perhaps the use of a basal marker, analogous to the use of DLG, would clarify the situation. Figure S1A at least has things labeled more clearly, and there too barely any staining is seen (A or P) in the basal region. In Figure S1 the anterior is divided into three regions depending on Ci staining (not shown) with the red R having an arrow to "FA". Yet the red R region is not very far anterior.

Overall, therefore, Figure 1 is quite confusing and does not seem to fully agree with the conclusions mentioned in the text. The references in the text to FA and P are not matched by any labels in Figure 1A-D. What I do see, in A, B, C, and D, is higher levels of surface SMO in lateral regions in the posterior vs anterior-and maybe the same for total SMO though A seems to contradict B in that regard. So at line 105, the accurate statement would seem to be "lateral" rather than "basolateral", and it is not clear that the protein being monitored is only cell surface protein. Part of the problem may be the lack of indication in the figure of what counts as "basal".

We thank the reviewer for this advice, which prompted us to reorganize and improve this Figure 1 and its associates figure supplements 1, 2 and 3, as we described at the beginning of this document.

The mentioned graph E does not address the apico-basal localization of SMO but shows the increase of SMO levels in response to HH. But, following the advice to simplify Figure 1 and given that it mainly confirmed a fact that has been well documented by many labs (including our), this part has been removed from our revised manuscript.

As mentioned in the text (former line 116, now in line 135) and in the former Figure S1A., we call “basal” the “10% most basal region of the Z sections”. To our knowledge, the basal region could be not defined by using a marker as we can do for the apical region.

The use of DLG and the definition of what we call basal are now clearly explained in the main text (line 134) as well as in the legends of Figure 1—figure supplement 2.

We hope that the confocal Zeiss LSM980 spectral Airyscan images convincingly display the basal accumulation of Surf SNAP-SMO in the presence of HH.

See for more details our response to reviewer 1 point 3, p 4.

Figure 1E' shows two sets of ***, which may be correct statistically but seem like small differences.

Despite the identification of many genes and proteins involved in HH signaling, little is really known on how the differences in HH dose and even less on how the basal and apical gradients are sensed and interpreted. Classically, such “all or none “responses may implicate regulatory loops that could contribute to “switch like “effects (Ashe et Briscoe Dev. 2006). In that context, quantitative imaging approaches as the one we developed, could allow -when associated to statistics to validate them- to identify effects that may be small but very relevant when considered in the context of the spatiotemporal regulation of a complex system. See also below.

line 119 Not really a model if they are HH-responsive cells.

We understand this point but we think that we should keep this term because strictly speaking the P cells do not really “respond to HH” in the way A cells near the A/P do: in the latter case, SMO is activated because the binding of HH to its receptor PTC inhibits the negative effect exerted by the later on SMO, while in the P cells, SMO is activated because PTC is absent (the ptc gene is not expressed in the P compartment).

line 123 "an increase"

This has been changed

line 129 "integraded" should be "integrated"

We thank the reviewer for noting these spelling errors and we have made these corrections.

lines 131-3 if it's not significant, why mention it?

We have now removed this in the result section and line 352 in the Discussion section.

line 135 in S1B, B' it is again unclear what is meant by basal, since the only staining seems to be lateral.

See above.

In S1C the key shows apical as light green, but nothing in the chart is light green. In contrast green bars in Figure 3D do not have a corresponding item in the key. The higher level of SMO in the posterior is visible. "Column" is defined in the text at line 130 but should be repeated in the S1 legend.

We apologize for these mistakes in the keys and we thank the reviewer for noticing it.

The colors of the graph have been changed on Figure3D. For simplification purpose, the panel S1C has been removed (see Figure1—figure supplement 3B), showing the images and the chart with the relative intensities.

We now mention that “Column corresponds to the entire height of the epithelium (i. e.100%)” in the legend of the current Figure1—figure supplement 2.

I can't see the basis in the staining for the right hand two (dark blue) bars in Figure S1C, since in B and B" the SMO staining seems weaker apically and laterally as well as basally in what may be FA cells vs P. That's why C' seems a better read of what's going on, i.e. no change in FA vs P in relative terms in any of the three regions (this is acknowledged in lines 150-2 but with different emphasis).

We have followed this advice and removed this panel.

lines 154-6 and Figure 1C' and D': Since the signal for intracellular Smo is so weak, I think the conclusion that HH does not affect it is also weak.

We show this data as the signal is detectable, was carefully quantified, and statistically analyzed. After organization of the Figure 1 and taking into account the present comment, this part has been moved to the Figure1—figure supplement 2.

lines 167-8 These too are small effects but supported by the pattern across four regions.

lines 168-9 Not sure what point is being made here. First of all both A and P compartments are being used here, and second the disc is a long-established tool for studying HH signaling.

See above our response to the comment.

lines 180-1 The proper control here (for Figure 2A) would be the shibire genotype without the temp shock, Figure 2E notwithstanding.

This control is now added to the Figure 2—figure supplement 1E and is now mentioned in the text line 190-191.

lines 204-210 If much of the normal signal is endocytosed SMO, then describing its location as "lateral" in the earlier paragraphs and figures is a bit misleading. Lateral suggests the sides of cells, not the inside. "Central" might be a better term.

We agree that after the chase we cannot distinguish between the labeled SNAP-SMO molecules which are still at the cell surface from the molecules that have been endocytosed. We have therefore changed “lateral” to “intermediate” in the Figure 3 (and the corresponding text lines 233, 236, 239 and 241).

lines 218-9 nice clear result

lines 228-9 Small differences but ok

line 232 "transcytosed" would mean traversing the cytoplasm to appear at the other (in this case basal) surface. Instead what's observed is an apparently internal central (and/or lateral) accumulation.

We agree that we cannot not conclude this in this section. This term was removed.

line 239 "cytotail" is not English

We have replaced it by “cytoplasmic C-terminal tail”.

lines 256-7 if the "basolateral" enrichment is a sign of the activation of SMO, why does a primarily negative regulator like PKA have an effect similar to the positive regulator HH? In contrast Fu is primarily a positive regulator.

Indeed, PKA was identified as a negative regulator of HH signaling in Drosophila as it negatively regulates CI. However, it has been shown by many labs (Jia, Jiang, Zhu, Beachy etc…) that it also acts as a positive regulator of SMO. To clarify this, we have changed “Given the importance of the phosphorylations of the C-terminal cytoplasmic domain of SMO” by “Given that the PKA and FU kinases positively regulate SMO activation and accumulation at the membrane” See lines 247-248.

Line 29-5 FU-GFP (T48) is enriched apically, but is also present throughout the cell. FU-GFP (NVR1) is not enriched apically but is seemingly present throughout the cells. Therefore interpretation of any negative result (eg Figure 5C' and lines 290-2) depends on whether the amount of FU kinase is limiting or not.

Several publications reported that changes in fu dosage (either loss of one dose of the fu gene (see for instance Preat et al. 1990)) or its overexpression (see for instance Claret et al., 2007) have no effect. Moreover, we now show that the effects of apical trapping of GFP-FU on hh targets still occur in the absence of the endogenous FU protein. All this indicate that is not likely due to an effect on the amount of FU kinase. See Figure 6 E-H and lines 316-332. See also our response to reviewer1, point 5, p5.

line 286-8 Nice result.

line 288 explain what the normal "anterior" expression of en looks like-it's not everywhere.

Normal anterior en expression (also called anterior en or late en expression) occurs in the first rows (around three) of cells abutting the A/P boundary. Note that en is also expressed in the whole posterior compartment but this is independent of HH. We have added this information in the legend of the Figure 1—figure supplement1.

line 297 "enrichment"

This has been corrected.

line 294 the reference to "cell surface" came as a surprise, since the surface means apical, lateral, and basal. Figure 6 doesn't resolve surfaces and Figures5A', B', C', and D' do not clearly show what is happening to SMO localization with respect to cell surfaces.

We have been able to visualize Surf SNAP-SMO expressed at endogenous levels from the BAC construct. It shows expression of GFP-FU with T48 leads to accumulation of Surf SNAP-SMO. This result is shown in Figure 5—figure supplement 2C-C”’. Lines 299-301.

line 298 Logic not clearly explained here. For example, why can't the endogenous FU provide this proposed basal function?

We think that apically tethered GFP-FU recruits endogenous FU. This is in agreement with the fact that the effect of apically trapped GFP-FU does not require endogenous FU. See also our response above and to reviewer 1 point 8, p5.

line 632 "merged" not "merge"

This has been corrected

Figure 7 This diagram seems unnecessarily complex to me and should be simplified as much as possible.

One problem is that the design of the diagram makes it hard to see quantitive changes in SMO abundance in different regions of the cells, yet those changes are central to the paper.

How much the abundance changes occur inside cells vs on the plasma membrane is not clear, and the diagram should leave open both possibilities.Why are PKA/CKI excluded from the right panel? I don't think that including adherens junctions, integrins, and extracellular matrix does anything but make the model harder to understand.Golgi appears in the model but nowhere else in the paper, so it's not clear why it's brought in here.

We thank the reviewer for these suggestions. The diagram is now simplified and PKA/CKI added.

Significance

Hedgehog (Hh) signaling is employed in the shaping of many organs and tissues during the development of many species of animals. The mechanisms of Hedgehog signaling are important for understanding development, birth defects, evolution, and cancer. In this paper the authors examine mechanisms involving the transmembrane protein Smoothened (SMO), a positive regulator in the pathway from received HH signal to the activation of target gene transcription. Subcellular localization of SMO has been shown to be important to its activity state. In vertebrate cells, arrival of a HH signal causes SMO to accumulate on the surface membrane in primary cilia, one of which is present on the surface of every cell. No analogous structure has been discovered in the Drosophila cells used in the present study, but SMO subcellular localization has been found to vary in response to HH, including some of it moving to the cell surface. For example, the present authors found that a positive feedback loop between the FU kinase and SMO increase the accumulation of both proteins at the cell surface.

Hh signaling was discovered in Drosophila and found to be important both during embryonic development and in the imaginal discs, precursors of adult appendages and body wall. Here the authors analyze effects on SMO during Hh signaling in the wing imaginal disc. They find that HH causes SMO to become stabilized during steps of endocytosis and recycling, and that at the highest levels of HH signaling the SMO protein is enriched in basal domains of epithelial cells.

The authors have done a lot of precise and difficult immunostaining work to look at the localization of SMO. The results show often small changes (10% effects) that are statistically significant but leave open questions of how essential the subcellular localization of SMO is to its functions.

We believe that the answer to these points relies on the complexity of the system that we study and on the fact that in “real life” populations of signaling molecules are both heterogeneous and dynamic.

First, one has to have in mind that in response to HH, only a part of the SMO population is phosphorylated by the PKA/CKI at a given time and among those, a fraction is also phosphorylated by FU as well (see for instance Sanial et al. 2017). Thus, only a fraction of it is expected to be localized in the basal region. In that respect, the use of mutants that mimic or block its phosphorylation reduce this heterogeneity, making the effect of the phosphorylation more visible and allowing the dissection of the role of the phosphorylation (see Figure 4). For instance, SMOPKA-SD FU-SD, which mimics highly phosphorylated SMO, is more enriched basally than unphosphorylated SMO (SMOWT in absence of HH) and this effect is reduced when the Fu sites are mutated to prevent their phosphorylation. Of note, we now have included data showing how these mutations of SMO also promote its activity. It shows that SMOPKA-SD FU-SD can promote high HH signaling, while SMOPKA-SD FU-SA on the contrary, blocks it. All these results further strengthen the link between the basal localization and its signaling activity. See the Figure 4—figure supplement 1 F-F”, G-G” and our response to reviewer 1, p 6-7.

Second, these effects on SMO trafficking are likely very dynamic and we may have “caught” a transient basal localization of SMO. For instance, the “very active” basal fraction of SMO may undergo subsequent degradation leading to its desensitization.

Finally, the response to HH levels, may also require the additive, and even possibly synergetic effects of other regulatory loops as those that have been shown to regulate HH signaling (see for instance Strigini and Cohen 1997, Kent et al., 2006, Holmgren 2022).

See also our response to reviewer 1’s comment, “significance“ section, p6.

This paper is part of the ongoing effort in many labs to understand the molecular biology of HH signal transduction. It will therefore be of interest to others working in this arena, and has implications for developmental biology and cancer. The most striking advance in this paper is the importance of FU kinase subcellular localization and its impact on SMO function and HH target gene activation. The paper could be simplified and made more accessible by focusing it more on these results. Localization of SMO and other transducers has proven to be very important for understanding HH signaling in flies and mammals, so this paper provides some useful new ideas.

https://doi.org/10.7554/eLife.79843.sa2

Article and author information

Author details

  1. Marina Gonçalves Antunes

    Université Paris Cité, CNRS, Institut Jacques Monod, Paris, France
    Contribution
    Conceptualization, Formal analysis, Validation, Investigation, Methodology, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3818-4434

  2. Matthieu Sanial

    Université Paris Cité, CNRS, Institut Jacques Monod, Paris, France
    Contribution
    Formal analysis, Investigation, Visualization, Methodology, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0301-4710

  3. Vincent Contremoulins

    Université Paris Cité, CNRS, Institut Jacques Monod, Paris, France
    Contribution
    Investigation
    Competing interests
    No competing interests declared

  4. Sandra Carvalho

    Université Paris Cité, CNRS, Institut Jacques Monod, Paris, France
    Contribution
    Investigation
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8654-0539

  5. Anne Plessis

    Université Paris Cité, CNRS, Institut Jacques Monod, Paris, France
    Contribution
    Conceptualization, Formal analysis, Supervision, Writing - original draft, Writing - review and editing
    For correspondence
    anne.plessis@ijm.fr
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-9193-7031

  6. Isabelle Becam

    Université Paris Cité, CNRS, Institut Jacques Monod, Paris, France
    Contribution
    Conceptualization, Formal analysis, Supervision, Writing - original draft, Writing - review and editing
    For correspondence
    isabelle.becam@ijm.fr
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-4464-7880

Funding

Institut des sciences biologiques

  • Marina Gonçalves Antunes
  • Matthieu Sanial
  • Vincent Contremoulins
  • Sandra Carvalho
  • Anne Plessis
  • Isabelle Becam

Fondation ARC pour la Recherche sur le Cancer (JA20191209287)

  • Anne Plessis

EurGene

  • Marina Gonçalves Antunes

Universite de Paris Cité

  • Marina Gonçalves Antunes

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We are grateful to Drs M Crozatier, A Guichet, D Hipfner, J Jian, P Therond, F Schweisguth for generously sharing their fly lines and reagents; to A Benhmerah, A Guichet and S Léon and our colleagues from the Institut Jacques Monod for insightful discussions, Lisa Barbuglio, Andréa Mialet and Severine Nozownik for their technical help. We are very grateful to Drs. G D’Angelo and R Holmgren for sharing their expertise and for their insightful advice. The Apa 1, 4D9, 4F3, and 20C6 monoclonal antibodies were obtained from the Developmental Studies Hybridoma Bank, created by the NICHD of the NIH and maintained at The University of Iowa, Department of Biology, Iowa City, IA 52242. Drosophila embryo injections were carried out by BestGene Inc and by inDroso. We acknowledge the ImagoSeine core facility of Institut Jacques Monod, member of France-BioImaging (ANR-10-INBS-04) and certified IBiSA. This work was supported by the Centre National de la Recherche Scientifique CNRS, the Université de Paris, and the Fondation ARC pour la recherche sur le Cancer (JA20191209287). MGA was supported by the Université de Paris (CNRS and the Ecole Universitaire Génétique et Epigénétique Nouvelle Ecole (EUR G.E.N.E)).

Senior Editor

  1. Claude Desplan, New York University, United States

Reviewing Editor

  1. Vilaiwan M Fernandes, University College London, United Kingdom

Version history

  1. Preprint posted: January 20, 2022 (view preprint)
  2. Received: April 29, 2022
  3. Accepted: August 22, 2022
  4. Version of Record published: September 9, 2022 (version 1)

Copyright

© 2022, Gonçalves Antunes et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 564
    Page views
  • 117
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Marina Gonçalves Antunes
  2. Matthieu Sanial
  3. Vincent Contremoulins
  4. Sandra Carvalho
  5. Anne Plessis
  6. Isabelle Becam
(2022)
High hedgehog signaling is transduced by a multikinase-dependent switch controlling the apico-basal distribution of the GPCR smoothened
eLife 11:e79843.
https://doi.org/10.7554/eLife.79843

Categories and tags

Research organism

Further reading

    1. Developmental Biology

    Dynamic compartmentalization of the pro-invasive transcription factor NHR-67 reveals a role for Groucho in regulating a proliferative-invasive cellular switch in C. elegans

    Taylor N Medwig-Kinney, Brian A Kinney ... David Q Matus
    Research Article

    A growing body of evidence suggests that cell division and basement membrane invasion are mutually exclusive cellular behaviors. How cells switch between proliferative and invasive states is not well understood. Here, we investigated this dichotomy in vivo by examining two cell types in the developing Caenorhabditis elegans somatic gonad that derive from equipotent progenitors, but exhibit distinct cell behaviors: the post-mitotic, invasive anchor cell and the neighboring proliferative, non-invasive ventral uterine (VU) cells. We show that the fates of these cells post-specification are more plastic than previously appreciated and that levels of NHR-67 are important for discriminating between invasive and proliferative behavior. Transcription of NHR-67 is downregulated following post-translational degradation of its direct upstream regulator, HLH-2 (E/Daughterless) in VU cells. In the nuclei of VU cells, residual NHR-67 protein is compartmentalized into discrete punctae that are dynamic over the cell cycle and exhibit liquid-like properties. By screening for proteins that colocalize with NHR-67 punctae, we identified new regulators of uterine cell fate maintenance: homologs of the transcriptional co-repressor Groucho (UNC-37 and LSY-22), as well as the TCF/LEF homolog POP-1. We propose a model in which the association of NHR-67 with the Groucho/TCF complex suppresses the default invasive state in non-invasive cells, which complements transcriptional regulation to add robustness to the proliferative-invasive cellular switch in vivo.

    1. Chromosomes and Gene Expression
    2. Developmental Biology

    Transcription: A hub of activity

    Virginia L Pimmett, Mounia Lagha
    Insight

    Imaging experiments reveal the complex and dynamic nature of the transcriptional hubs associated with Notch signaling.

    1. Cell Biology
    2. Developmental Biology

    Cylicins are a structural component of the sperm calyx being indispensable for male fertility in mice and human

    Simon Schneider, Andjela Kovacevic ... Hubert Schorle
    Research Article

    Cylicins are testis-specific proteins, which are exclusively expressed during spermiogenesis. In mice and humans, two Cylicins, the gonosomal X-linked Cylicin 1 (Cylc1/CYLC1) and the autosomal Cylicin 2 (Cylc2/CYLC2) genes, have been identified. Cylicins are cytoskeletal proteins with an overall positive charge due to lysine-rich repeats. While Cylicins have been localized in the acrosomal region of round spermatids, they resemble a major component of the calyx within the perinuclear theca at the posterior part of mature sperm nuclei. However, the role of Cylicins during spermiogenesis has not yet been investigated. Here, we applied CRISPR/Cas9-mediated gene editing in zygotes to establish Cylc1- and Cylc2-deficient mouse lines as a model to study the function of these proteins. Cylc1 deficiency resulted in male subfertility, whereas Cylc2-/-, Cylc1-/yCylc2+/-, and Cylc1-/yCylc2-/- males were infertile. Phenotypical characterization revealed that loss of Cylicins prevents proper calyx assembly during spermiogenesis. This results in decreased epididymal sperm counts, impaired shedding of excess cytoplasm, and severe structural malformations, ultimately resulting in impaired sperm motility. Furthermore, exome sequencing identified an infertile man with a hemizygous variant in CYLC1 and a heterozygous variant in CYLC2, displaying morphological abnormalities of the sperm including the absence of the acrosome. Thus, our study highlights the relevance and importance of Cylicins for spermiogenic remodeling and male fertility in human and mouse, and provides the basis for further studies on unraveling the complex molecular interactions between perinuclear theca proteins required during spermiogenesis.