The Clathrin adaptor AP-1 and the Rab-stabilizing chaperone Stratum act in two parallel pathways to control the activation of the Notch pathway in Drosophila

Drosophila sensory organ precursors divide asymmetrically to generate pIIa/pIIb cells whose identity relies on the differential activation of Notch during cytokinesis. While Notch is present apically and basally relative to the midbody at the pIIa-pIIb interface, only the basal pool of Notch is reported to contribute to Notch activation in the pIIa cell. Such proper intra-lineage signalling therefore requires appropriate apico-basal targeting of Notch, its ligand Delta and its trafficking partner Sanpodo. We previously reported that AP-1 and Stratum regulate the intracellular trafficking of Notch and Sanpodo from the trans-Golgi network to basolateral membrane. Loss of AP-1 or of Stratum caused mild Notch phenotype. Here, we report that the concomitant loss of AP-1 and Stratum result in the stabilization of the apical pool of Notch, Delta and Spdo, the loss of the basal pool of Notch at the pIIa-pIIb interface, and is associated with activation of Notch in the two SOP daughters. We propose that AP-1 and Stratum control two parallel pathways towards plasma membrane and that Notch intra-lineage signalling could also occur at the apical pIIa-pIIb interface.

activated by transmembrane ligands present at the plasma membrane of adjacent cells.
Following binding to Notch, endocytosis of the ligand induces pulling forces driving a change in the conformation of the Notch extracellular domain thereby unmasking the S2 cleavage site of Notch (Gordon et al., 2015;Langridge and Struhl, 2017;Meloty-Kapella et al., 2012;Seo et al., 2016;Shergill et al., 2012;Wang and Ha, 2013). This regulated cleavage is followed by a constitutive proteolytic cleavage of Notch by the gamma secretase complex (Mumm et al., 2000;Struhl and Adachi, 2000) giving rise to the Notch intracellular domain, a polypeptide that is translocated into the nucleus and that acts as a transcriptional coactivator (Artavanis-Tsakonas et al., 1999;Bray, 1998;Kopan and Ilagan, 2009). Because proteolytic activation of the Notch receptor is by definition irreversible, Notch activation needs to be tightly controlled in time and in space. The model system of asymmetric cell division of the sensory organ precursors (SOPs) in the pupal notum of Drosophila has been instrumental to identify the site of Notch activation at the cell surface. SOPs are polarized epithelial cells that divide asymmetrically within the plane of the epithelium to generate two daughter cells whose fate depends on the differential activation of Notch signalling (Schweisguth, 2015). It relies on the unequal partitioning of the two cell fate determinants Neuralized (Neur) and Numb in the anterior SOP daughter cell (Le Borgne and Schweisguth, 2003;Rhyu et al., 1994). Neur promotes the endocytosis of Delta (Le Borgne and Schweisguth, 2003) while Numb inhibits the recycling of Notch and its cofactor Sanpodo (Spdo) towards the plasma membrane to instead promote their targeting towards late endosomal compartments (Cotton et al., 2013;Couturier et al., 2013;Johnson et al., 2016;Upadhyay et al., 2013).
Consequently, the anterior cell adopts the pIIb identity while Notch is selectively activated in the posterior cell that adopts the pIIa fate. Combination of live-imaging and FRAP experiments using GFP-tagged Notch revealed that proteolytic activation of Notch occurs during SOP cytokinesis and that a specific pool of Notch receptors located basal to the midbody is the main contributor to the signalling in the pIIa cell (Trylinski et al., 2017). These data imply a polarized trafficking of Notch, Delta and Spdo towards this specific subcellular location during cytokinesis. We previously reported that the clathrin adaptor complex AP-1 regulates the polarized sorting of Notch and Spdo from the trans-Golgi network (TGN) and the recycling endosomes towards the plasma membrane (Benhra et al., 2011). Loss of AP-1 causes stabilization of Notch and Spdo at the adherens junctions following SOP division, a phenotype associated with a mild Notch gain-of-function phenotype (GOF). More recently, we reported that Stratum (Strat), a chaperone regulating Rab8 recruitment, controls the exit from the Golgi apparatus and the basolateral targeting of Notch, Delta and Spdo (Bellec et al., 2018). As for AP-1, loss of Strat causes a mild Notch GOF phenotype. These data can be interpreted as AP-1 and Strat acting in the same transport pathway and therefore being simple fine-tune regulators of Notch-Delta trafficking and activation. However, as AP-1 and Strat/Rab8 both regulate basolateral trafficking, an alternative explanation could be that AP-1 and Strat function in two parallel pathways . In this scenario, loss of one of the two components could be compensated by the other. Accordingly, the concomitant loss of AP-1 and Strat would exhibit a stronger phenotype.

Simultaneous loss of AP-1 and Strat causes a penetrant Notch GOF phenotype
To test this prediction, we induced clones of cells homozygote mutant for a null mutation of Strat (Strat KO CRISPR, (Bellec et al., 2018) in which the µ subunit of the AP-1 complex (also known as AP-47) was silenced using the previously described dsRNA (Benhra et al., 2011). Silencing of AP-47 prevents the assembly of functional AP-1 complexes and gives similar Notch GOF phenotypes as two independent AP-47 null mutant alleles  and AP-47 deletion by CRISPR ((Benhra et al., 2011) and data not shown)). Silencing of AP-47 is hereafter referred to as loss of AP-1. In wild-type sensory organ (SO), the pIIb cell divides two times to generate the internal cells among which is one neuron, labelled using an anti-Elav. The pIIa cell divides once to generate the external cells among which is one socket cell, labelled with Suppressor of Hairless (Su(H)). In agreement with previous studies, Notch GOF phenotypes corresponding to an excess of socket cells were observed in 6,7% and 7,5% of SO mutant for strat or depleted of AP-1, respectively ((Bellec et al., 2018;Benhra et al., 2011), Fig.1C). However simultaneous loss of AP-1 and Strat led to 49% of transformed SOs (Fig.1B-C). Among them, 34% of SOs were transformed at the SOP division, i.e. SOP divided to generate two pIIa cells (n= 87/254, Fig.1C and data not shown).
These lineage data indicate that AP-1 and Strat function in two distinct and complementary pathways to regulate Notch activation.

Numb is correctly partitioned upon loss of AP-1 and Strat
Because AP-1 interacts with Numb and loss of Numb exhibits a Notch GOF, we wondered if the phenotype observed upon the loss of AP-1 and Strat would result from the loss of Numb function. We previously reported that the unequal partitioning of Numb is unaffected by the loss of AP-1 (Benhra et al., 2011) or by the loss of Strat (Bellec et al., 2018). Because unequal segregation of Numb is under the control of polarity genes, we first monitored the overall cell polarity (Gho and Schweisguth, 1998;Kraut et al., 1996;Schober et al., 1999).
We noticed that, as in AP-1 mutant, the cuticle looks thinner and less pigmented than in the control situation ( Fig.S1A-B'). These phenotypes could be caused by a defective AP-1 dependent transport of the Menkes Copper transporter ATP7a that regulates cuticle pigmentation (Holloway et al., 2013;Norgate et al., 2006) and by a defective apical secretion of cuticle components that may rely on AP-1 function as does the apical glue granule secretion in salivary glands (Burgess et al., 2011). Despite these cuticle defects, the localization of the junctional markers DE-Cadherin and Coracle, as well as that of the polarity determinant Par3, were unaffected by the loss of AP-1 and Strat indicating that the overall cell polarity is unaffected ( Fig.S1C-D). We next monitored the localization of Numb during the division of strat mutant SOPs depleted of AP-1. Anti-Numb staining showed that Numb localizes asymmetrically during prometaphase in control and in absence of Strat and AP-1 ( Fig.2A-B) and live-imaging using a Numb::GFP crispr (Bellec et al., 2018) revealed that Numb is unequally partitioned in the anterior SOP daughter cell, as seen in the control situation ( Fig.2C-D). Thus, the Notch GOF phenotype cannot be explained by a defective Numb unequal partitioning during SOP division and this rather suggests that it could be due to a defect in the polarized trafficking of Notch signalling actors.

Notch is enriched at the apical pIIa-pIIb interface upon loss of AP-1 and Strat
Elegant work from F. Schweisguth laboratory identified two pools of Notch at the pIIa-pIIb interface, apical and basal to the midbody, and provides the compelling evidence that only the subset of receptors located basal to the midbody contributes to signalling in the pIIa cell.
To investigate whether AP-1 and Strat affect the apico-basal distribution of Notch receptors, we monitored the dynamics of Notch::GFP crispr (Bellec et al., 2018) throughout the asymmetric division of the SOP. First of all, in the control situation, we confirmed the presence of two pools of Notch along the apical-basal pIIa-pIIb interface, in agreement with previous studies ((Trylinski et al., 2017), Fig.3A). Notch is transiently detected at the apical pIIa-pIIb interface at ~ 6 -9 min after the anaphase onset with a signal intensity peaking at 15-20 min prior to progressively disappear at ~ 3 0 min ( Fig.3A-B and Movie S1). Basal to the midbody, and in agreement with previous studies (Couturier et al., 2012;Trylinski et al., 2017), we also detected punctate structures positive for Notch, hereafter called lateral clusters, that appear ~ 6 -9 min following the anaphase onset and that persist longer than the apical pool of Notch (up to 45 min after the anaphase onset; Fig.3A). In striking contrast to the control situation, upon loss of Strat and AP-1, the Notch lateral clusters were not detected (Fig.3C). Instead, we found that Notch appears in higher amounts at the apical pIIa-pIIb interface cells To investigate this possibility, we first tried to directly test the ability of the apical pool of Notch to be cleaved and to be translocated into the nucleus. To this aim, we inserted the green-to-red photoconvertible Dendra2 in the cytoplasmic domain of Notch (Notch::Dendra crispr ) with the purpose of photoconverting the apical pool of Notch::Dendra crispr and testing its ability to translocate into the nucleus following proteolytic activation. While Dendra2 is detected at the apical plasma membrane upon loss of Strat and AP-1 using an anti-Dendra on fixed specimen (Fig.S2A), Notch::Dendra crispr signal is not detectable at the plasma membrane in live specimens (Fig.S2B). In time-lapse imaging, Notch::Dendra crispr signal is restricted to intracellular dotted structures, likely endosomal compartments ( Fig.S2B). This apparent discrepancy between fixed and living specimen could be due to a longer time of folding or maturation of the newly synthesized Dendra2 compared to the GFP probe, as already reported for other fluorescent probes (Couturier et al., 2014). This result is compatible with the idea that newly synthesized, e.g. not yet folded or matured, Notch::Dendra crispr receptors are targeted to the apical SOP daughter cells interface and are processed from this location upon loss of Strat and AP-1. Because the lack of detection of Notch::Dendra crispr at the plasma membrane prevented us to directly test our hypothesis of apical activation of Notch, we next investigated the localization of Spdo and Delta in SO mutant for Strat and depleted of AP-1.
Spdo and Delta are distributed with Notch at the apical pIIa-pIIb interface upon loss of

Strat and AP-1
The localization of Spdo and Delta was analysed on fixed specimens. As previously described, in the control situation, Spdo is faintly detected at the apical pole of SOP daughter cells, and localizes predominantly in endosomes in the pIIb cell, while in the pIIa cell, Spdo distributes not only in endosomes but also at the basolateral plasma membrane illustrates how the apico-basal targeting of Notch, Spdo and Delta along the pIIa-pIIb interface limits signalling between SOP daughter cells at the exit of mitosis. The question of whether this mechanism is conserved awaits further investigation, but given the key role played by Notch in the acquisition of cell fate in vertebrates, it is conceivable that the regulation described in this study may be general.

Drosophila stocks and genetics
Drosophila melanogaster stocks were maintained and crossed at 25°C. Mitotic clones were induced using the FLP-FRT technique using the hs-FLP and by heat shocking (2×60 min at 37°C) at second and early third instar larvae. pnr-GAL4 was used to drive the expression of the AP-47 dsRNA . The following stocks were used in this study: y w hs-FLP; Ubi-GFP nls, FRT40A; pnr-GAL4/TM6 Tb (Fig.1B

Immunofluorescence and antibodies
Pupae were aged for 16.5 to 18.5 h after puparium formation (APF) for SOPs and SOPs daughter cells analysis and were aged for 24 h to 28 h APF for lineage analysis. Pupae were dissected in 1x phosphate-buffered saline (1x PBS) and then fixed for 15 min in 4% paraformaldehyde at room temperature. Dissection and staining conditions were essentially as previously described (Le Borgne and Schweisguth, 2003). Primary antibodies used were

Imaging
Images of fixed nota were acquired with a Leica SPE confocal microscope and a Zeiss Airyscan microscope. The live imaging of Numb::GFP crispr and Notch::Dendra crispr was acquired with a Leica SPE confocal microscope. The live imaging of Notch::GFP crispr in wildtype SOP and in strat SOP expressing pnr-GAL4>AP-47 dsRNA was acquired with a Zeiss Airyscan microscope. All images were processed and assembled using ImageJ 1.48 and Adobe Illustrator.

Quantification of the enrichment of Notch at the apical interface
To quantify the signal of Notch::GFP at the apical interface, we measured the signal in a manually drawn area on sum slices of the two apical planes where the Notch signal at the apical interface is the strongest. We also measured the signal between two epithelial cells within an equivalent drawn area and on the same sum slices. Then, we calculated the following ratio: Average fluorescence intensity at the apical interface between SOP daughter cells/Average fluorescence intensity at the apical interface between epithelial cells. This ratio was calculated for each time point. In absence of AP-1 and Strat, the manually drawn area was minimized to avoid taking into account the apical punctate structures positive for Notch.

Quantification of the apical enrichment of Spdo
The fluorescence intensity was calculated with ImageJ, on two-cell stages, as previously described (Bellec et al., 2018). Briefly, the average fluorescence intensity was measured in a manually drawn area on sum slices of the two most apical planes, where Spdo is enriched.
The background noise was measured in the same way and subtracted from the apical intensity value.

Statistical analysis
Statistical analyses were carried out using the GraphPad Prism 6.05 software. A two-way ANOVA was performed for the quantification of the enrichment of Notch at the apical pIIa-pIIb interface with a multiple comparison Bonferroni test. For the quantification of the apical enrichment of Spdo, because data do not follow a normal distribution, we performed a Wilcoxon test. Statistical significances are represented as follows: not significant (ns)≥0.05; *P<0.05; **P<0.01; ***P<0.001 and ****P<0.0001.       u  m  b  l  o  c  a  l  i  z  e  s  a  t  e  n  d  o  s  o  m  e  s  a  n  d  c  o  n  t  r  o  l  s  t  h  e  e  n  d  o  s  o  m  a  l  s  o  r  t  i  n  g  o  f  n  o  t  c  h  a  f  t  e  r  a  s  y  m  m  e  t  r  i  c  d  i  v  i  s  i  o  n  i  n  D  r  o  s  o  p  h  i  l  a  .  C  u  r  r  e  n  t  b  i  o  l  o  g  y  :  C  B   2  3   ,  5  8  8  -5  9  3  .  C  o  u  t  u  r  i  e  r  ,  L  .  ,  T  r  y  l  i  n  s  k  i  ,  M  .  ,  M  a  z  o  u  n  i  ,  K  .  ,  D  a  r  n  e  t  ,  L  .  ,  a  n  d  S  c  h  w  e  i  s  g  u  t  h  ,  F  .  (  2  0  1  4  )  .  A  f  l  u  o  r  e  s  c  e  n  t  t  a  g  g  i  n  g  a  p  p  r  o  a  c  h  i  n  D  r  o  s  o  p  h  i  l  a  r  e  v  e  a  l  s  l  a  t  e  e  n  d  o  s  o  m  a  l  t  r  a  f  f  i  c  k  i  n  g  o  f  N  o  t  c  h  a  n  d  S  a  n  p  o  d  o  .  T  h  e  J  o  u  r  n  a  l  o  f  c  e  l  l  b  i  o  l  o  g  y   2  0  7 ,  i  n  h  i  b  i  t  s  m  e  m  b  r  a  n  e  l  o  c  a  l  i  z  a  t  i  o  n  o  f  S  a  n  p  o  d  o  ,  a  f  o  u  r  -p  a  s  s  t  r  a  n  s  m  e  m  b  r  a  n  e  p  r  o  t  e  i  n  ,  t  o  p  r  o  m  o  t  e  a  s  y  m  m  e  t  r  i  c  d  i  v  i  s  i  o  n  s  i  n  D  r  o  s  o  p  h  i  l  a  .  D  e  v  e  l  o  p  m  e  n  t  a  l  c  e  l  l   5   ,  2  3  1  -2  4  3  .  R  a  j  a  n  ,  A  .  ,  T  i  e  n  ,  A  .  C  .  ,  H  a  u  e  t  e  r  ,  C  .  M  .  ,  S  c  h  u  l  z  e  ,  K  .  L  .  ,  a  n  d  B  e  l  l  e  n  ,  H  .  J  .  (  2  0  0  9  )  .  T  h  e  A  r  p  2  /  3  c  o  m  p  l  e  x  a  n  d  W  A  S  p  a  r  e  r  e  q  u  i  r  e  d  f  o  r  a  p  i  c  a  l  t  r  a  f  f  i  c  k  i  n  g  o  f  D  e  l  t  a  i  n  t  o  m  i  c  r  o  v  i  l  l  i  d  u  r  i  n  g  c  e  l  l  f  a  t  e  s  p  e  c  i  f  i  c  a  t  i  o  n  o