Par3 cooperates with Sanpodo for the assembly of Notch clusters following asymmetric division of Drosophila sensory organ precursor cells

In multiple cell lineages, Delta-Notch signalling regulates cell fate decisions owing to unidirectional signalling between daughter cells. In Drosophila pupal sensory organ lineage, Notch regulates the intra-lineage pIIa/pIIb fate decision at cytokinesis. Notch and Delta that localise apically and basally at the pIIa-pIIb interface are expressed at low levels and their residence time at the plasma membrane is in the order of minutes. How Delta can effectively interact with Notch to trigger signalling from a large plasma membrane area remains poorly understood. Here, we report that the signalling interface possesses a unique apico-basal polarity with Par3/Bazooka localising in the form of nano-clusters at the apical and basal level. Notch is preferentially targeted to the pIIa-pIIb interface, where it co-clusters with Bazooka and its cofactor Sanpodo. Clusters whose assembly relies on Bazooka and Sanpodo activities are also positive for Neuralized, the E3 ligase required for Delta activity. We propose that the nano-clusters act as snap buttons at the new pIIa-pIIb interface to allow efficient intra-lineage signalling.

Notch is the receptor of an evolutionarily conserved cell-cell signaling pathway 40 controlling fate acquisition in numerous developmental processes throughout 41 metazoan development (Kopan and Ilagan, 2009). Within many cell lineages, 42 following the division of a precursor cell, Notch activation regulates binary fate 43 choice between daughter cells (Bertet et al., 2014;Bivik et al., 2016;Dong et al., 44 2012;Ohlstein and Spradling, 2007;Pardo-Saganta et al., 2015;San-Juan and 45 Baonza, 2011). In the majority of cases, Notch receptor is activated by 46 transmembrane ligands present in adjacent cell. Following binding to Notch, 47 endocytosis of the ligand generates pulling forces driving a change in the 48 conformation of the Notch extracellular domain leading to the exposure of the S2 49 cleavage site of Notch (Gordon et al., 2015;Langridge and Struhl, 2017;Meloty-50

136
Remodeling of apico-basal polarity during SOP division leads to an atypical 137 pIIa-pIIb interface 138 139 As SOP undergoes a specific remodeling of polarity modules during division 140 (Bellaiche et al., 2001a), we started by investigating the resulting apicobasal 141 polarity and remodeling of junctional complexes on the forming pIIa-pIIb cell 142 interface during cytokinesis from which Notch is activated. We previously 143 reported that formation of the novel adhesive pIIa-pIIb interface, visualized with 144 DE-Cadherin-GFP (E-Cad), is assembled with similar kinetics to that of 145 epidermal daughters (Founounou et al., 2013). Here, we first live-monitored and 146 compare the remodeling of septate junctions (SJs) and markers in SOP versus 147 epidermal cells as they undergo cytokinesis. All fluorescent markers are inserted 148 at the locus, giving rise to functional reporters expressed at physiological level. In 149 every case, the transition from metaphase to anaphase was considered to be t0 150 (time is indicated in min:sec). Dlg-GFP (Woods and Bryant, 1991) and Baz. Crb-GFP is detected faster at the new apical pIIa-pIIb interface than 165 between epidermal daughters (Fig. 1A, S1C; t7±2 min at pIIa-pIIb interface 166 intracellular domain of Notch tagged with GFP into the nucleus of the pIIa cell 198 (Fig. 1D,D'', (Bellec et al., 2018;Bellec et al., 2021;Couturier et al., 2012;199 Trylinski et al., 2017). As Baz, Notch is not detected in clusters at the lateral 200 interface of epidermal daughter cells (Fig. S1F). As epithelial cells are tightly 201 packed, we first determined the origin of the Notch signal presents between the 202 SOP daughter cells. Indeed, during epithelial cells cytokinesis, the dividing cell 203 maintains membrane contacts with the neighboring cells forming a ménage à 204 quatre that is progressively resolved as the cell progresses towards abscission 205 (Daniel et al., 2018;Founounou et al., 2013;Guillot and Lecuit, 2013;Herszterg 206 et al., 2013;Morais-de-Sa and Sunkel, 2013;Wang et al., 2018). Because of its 207 duration, the contact is particularly noticeable within the plane of SJs where 208 epidermal cells maintain contact in the form of fingers like protrusions connected 209 to the SOP midbody (t5, Fig. S1B) until the entire belt of SJ is reformed (about 90 210 min) (Daniel et al., 2018)

. Analysis of the borders of clones of cells expressing 211
NiGFP (Bellec et al., 2018) enables to determine the origin of the NiGFP signal 212 detected along the pIIa-pIIb interface ( Fig. 2A-B

'). When a SOP expressing 213
NiGFP is dividing next to epidermal cells expressing untagged Notch ( Fig. 2A), 214 NiGFP signal is detected at the apical pIIa-pIIb interface, and basally between 215 the pIIa-pIIb nuclei ( Fig. 2A', t21). In the converse situation, when a SOP 216 expressing untagged Notch is dividing next to epidermal cells expressing NiGFP, 217 no GFP signal is detected at the apical and basal pIIa-pIIb interface (Fig. 2B,B'). 218 Analyses of clone boundaries also revealed that low NiGFP signal was detected 219 at the boundary between SOP daughters and the neighboring epidermal cells. 220 This is not observed in epidermal cell where Notch equally partitions along the 221 plasma membrane ( Fig. 2A', B', and Fig 3A). These data show that, following 222 SOP division, Notch is preferentially transported towards or stabilized at the pIIa-223 pIIb interface where signaling takes place. 224 As Notch resembles Baz localisation at the pIIa-pIIb interface, we investigated 225 their localisation simultaneously by co-imaging NiGFP together with Ubi-Baz-226 mCherry (Bosveld et al., 2012). We first observed that Ubi-Baz-mCherry 227 colocalises with NiGFP in punctae along the pIIa-pIIb apical interface as well as 228 in the basally located, lateral clusters (Fig. 2C). The NiGFP/Baz lateral clusters 229 do not correspond to spot AJs (McGill et al., 2009) as E-Cad and Baz do not 230 colocalise at the lateral pIIa-pIIb interface (Fig. 2D). We next monitored the 231 dynamics of Ubi-Baz-mCherry and NiGFP clusters using high spatiotemporal 232 resolution acquisitions. Kymographs of these acquisitions (Fig. 2 E,E') show that, 233 at the apical pIIa-pIIb interface and even more markedly at the lateral interface, 234 the Ubi-Baz-mCherry and NiGFP tracks colocalise to a greater extent compared 235 to epidermal-epidermal interface , raising the possibility that Baz and 236 Notch act together in space and time to contribute to pIIa/pIIb identities that we 237 next investigated. 238 239

Baz contributes to Notch localisation after SOP division 240
After having establish that Ubi-Baz-mCherry/NiGFP clusters are specific to the 241 pIIa-pIIb interface, we asked whether Notch and Baz are mutually required for 242 clusters formation. In baz EH747 clones, a genetic and protein null allele of Baz To test if conversely Notch is required for Baz localisation in clusters, we 257 depleted Notch using RNAi or degradFP system (Caussinus et al., 2011). Both 258 approaches resulted in the reduction in Notch signal and Notch loss-of-function 259 phenotypes, including an excess of SOP specification due to defective lateral 260 inhibition, and pIIa to pIIb cell fate transformation (( Fig. 4 A-C) and data not 261 shown). Under these conditions, Baz still localises in clusters along the apico-262 basal pIIa-pIIb interface as in the wild type, indicating that Notch is dispensable 263 for Baz cluster assembly (Fig. 4 D,E,yellow arrowheads). Nonetheless, lateral 264 Baz clusters were never detected between two adjacent SOPs nor between a 265 SOP prior division and a pIIb-pIIb-like cell (data not shown). These data suggest 266 the remodeling of cell polarity taking place during SOP cytokinesis is a 267 prerequisite for the assembly of Baz/Notch lateral clusters. 268 As Baz was deemed necessary but not sufficient per se for Notch cluster 269 assembly, we next asked what key regulators of Notch-dependent binary fate 270 acquisition control Baz/Notch cluster assembly and/or dynamics. C'). Upon loss of Neur, Baz localises at the apical interface and in clusters at the 287 lateral pIIa-pIIb interface (Fig. 5D). By analyzing their dynamics at high spatio-288 temporal resolution, we found that Baz and NiGFP remain closely associated in 289 the apical and lateral clusters at the pIIa-pIIb interface upon loss of Neur (Fig.  290  5E). Collectively, these results argue that Neur, Delta, Baz and Notch localise 291 together in clusters at the pIIa-pIIb interface and that their numbers and signal 292 intensities depend on Neur and Delta activities. These data also suggest that the 293 clusters are assembled but fail to dissociate in the absence of Neur-mediated 294 activation of Notch. These data are fully consistent with that published previously in (Couturier et al., 303 2012;Trylinski et al., 2017) and further show that the lateral Notch clusters 304 accumulation persists until at least 36 min after anaphase onset. Upon silencing 305 of Numb, Baz localises in lateral clusters at the pIIa-pIIa like interface (Fig. 6D, 306 D') where it colocalises with NiGFP as revealed by the dynamics of the Baz-307 Notch clusters at high spatio-temporal resolution (Fig. 6E). 308 309 Together, these data indicate that Numb represses Notch activation and 310 negatively regulates the number of Notch-Baz clusters. As Numb is present and 311 regulates Notch endosomal trafficking in the anterior pIIb cell (Cotton et al., 2013;312 Couturier et al., 2013), our data suggest that Notch-Baz clusters are assembled 313 in the anterior cell Plutot a l'interface non? Pas de data montrant que c'est dans 314 la cellule anterieure upon loss of Numb Mais aussi en absence de Neur ou Delta, 315 and contribute to Notch activation in this cell. This model further suggests that 316 Numb acts antagonistically to Baz to promote Notch clusters assembly and/or 317 stability Neur et Delta aussi non?. To test this prediction, we overexpressed 318 Numb in the SOP and daughter cells and observed that NiGFP is no longer 319 detected along the pIIb-pIIb-like interface neither apically nor laterally (Fig. 6F). 320 While Baz localises uniformly at the apical SOP daughter cell interface, lateral 321 clusters are barely detectable (t14, Fig. 6F, bottom panels). Thus, Numb appears 322 to inhibit Baz-Notch cluster laterally, raising the possibility that Numb and Baz 323 could act antagonistically as proposed in vertebrates (Nishimura and Kaibuchi, 324 2007;Sun et al., 2016). Du coup, je crois avoir deja pose la question, mais si on 325 surexprime Neur ou Delta, a t on le meme phenotype? Quelle est le modele que 326 vous avez en tete les concernant? As Numb interacts with the NPAF motif of 327 Spdo to control Notch/Spdo endosomal trafficking, the above data are calling the 328 question of the relationship between Baz and Spdo that we next studied. In this study, we have characterized the remodeling of apico-basal cell polarity 357 occurring during SOP division leading to a specific pIIa-pIIb Notch signaling 358 interface. We report that Baz, but not aPKC, co-partitions with Notch, Spdo and 359 Neur in apical and lateral clusters. The assembly of these clusters requires Baz 360 and Spdo activities, and their number and dynamics are regulated by Delta, Neur 361 and Numb activities. In the absence of Numb, the number of clusters increases 362 while overexpression of Numb results in their disappearance suggesting that 363 Numb and Baz acts antagonistically. We propose that Notch/Baz/Spdo/Neur 364 clusters represent the Notch signaling units at the pIIa-pIIb interface. 365

to epidermal cells 368
Previous pioneer work unraveled that Par3/Par6/aPKC and Pins/Dlg polarity 369 modules are specifically relocated from apical-basal into posterior-anterior cortex 370 respectively during SOP mitosis (Bellaiche et al., 2001a;Roegiers et al., 2001). 371 Assembly of the Baz/Par6/aPKC is initiated by the phosphorylation of Par6 by the 372 mitotic kinase AurA (Wirtz-Peitz et al., 2008). Here, we report that during 373 cytokinesis coinciding with the presumptive proteolytic degradation of AurA, the 374 Baz /Par6/aPKC complex disassembles with aPKC redistributing like Crumbs in 375 apical intracellular compartments at the expense of its regular plasma membrane 376 localisation observed in epidermal cells. In contrast to aPKC, Baz redistributes 377 apically at the posterior pole of the pIIa cell and in the form of clusters at the pIIa-378 pIIb interface. Such lateral clusters of Baz are only found at the pIIa-pIIb interface 379 indicating that SOP specific remodeling polarity that takes place at SOP mitosis 380 is instrumental in clusters formation. Baz has been reported to be excluded from 381 the lateral plasma membrane following to Par1-mediated phosphorylation 382 (Benton and St Johnston, 2003). In addition, phosphorylation of Baz by Par1 383 activity is antagonized by type 2A protein phosphatase (PP2A) activity (Krahn et 384 al., 2009), and silencing of tws the regulatory B subunit of PP2A results in Notch 385 gain-of-function phenotype (Shiomi et al., 1994). However, we observed that Baz 386 lateral clusters are decorated with anti-phospho Baz S151, S980, or S1085 387 suggesting that while Par1-dependent phosphorylation is taking place efficiently, 388 it does not promote exclusion of Baz from the lateral plasma membrane (data not 389 shown). It is yet unclear how SOP polarity remodeling leads to Baz clusters 390 assembly and localisation laterally. The fact that Spdo and Notch, two The nanoscopic clusters of Baz are reminiscent to the clusters serving as AJ 396 assembly landmark, by repositioning Cadherin-Catenin clusters at apico-lateral 397 sites for assembly of spot AJ (SAJ, (McGill et al., 2009)). The Baz and Cad-398 Catenin clusters were shown to assemble independently and the number and 399 size of Cadherin-Catenin clusters are decreased in baz mutant as reported here 400 for Notch clusters. This is pointing towards a common function of Baz in the 401 control of cluster assembly, positioning and stability. 402 In addition to organize membrane nanoscopic clusters of Cadherin and Catenin,403 in vertebrates, Par3 also functions as a receptor for exocyst, a protein complex of 404 the secretory pathway required for the delivery of basolateral proteins to the 405 plasma membrane (Ahmed and Macara, 2017). It is interesting to note that Baz 406 clusters are exclusively located at the pIIa-pIIb interface. Analyses of NiGFP 407 clones borders revealed a preferential localisation of Notch at the pIIa-pIIb 408 interface instead of being equally partitioned at the plasma membrane. The clusters present at the pIIa-pIIb interface are positive for Notch, Spdo, Baz 568 and Neur. While Delta is also detected along the pIIa-pIIb interface on fixed 569 specimen (Bellec et al., 2021), DlGFP was reported to be barely detectable in 570 living pupae unless Neur-mediated Delta endocytosis was blocked (Trylinski et 571 al., 2017). This led to the proposal that newly synthesized Delta reaches the 572 plasma membrane and signals from there thus exhibiting a rapid 573 turnover/endocytosis. A direct implication of these findings is that the clusters are 574 present on both sides of the pIIa-pIIb interface as a kind of snap button with 575 Delta/Neur in the pIIb cell interacting in trans with Notch/Spdo in the pIIa cell. 576 Based on the role of Numb in Notch/Spdo trafficking in the pIIb cell, the fact that 577 Baz is enriched in the posterior pIIa cell at cytokinesis and the proposed 578 antagonism between Numb and Baz, we anticipate that Baz is located primarily 579 in the clusters on the pIIa cell side. As the time residence of Delta, Notch and 580 Baz in the cluster are very short (minute time scale), it implies that Delta can 581 interact with Notch in trans, be internalized in a Neur-dependent manner to 582 promote the S2 cleavage of Notch in the minute time scale. We propose that Baz 583 mediated clustering might be a mean to locally concentrate Notch/Spdo and 584 increase its ability to interact with Delta. 585 586 Sensory organ specific relocation of Crumbs and site of NICD production 587 The SO-specific localisation of Crb could be explained by the down-regulation of 588 Notch also contributes to NICD production, it is yet unclear whether NICD is 611 directly produced from the apical pIIa-pIIb interface or if basolateral relocation is 612 a prerequisite (Bellec et al., 2021;Couturier et al., 2012;Trylinski et al., 2017). Neur, we anticipate that Delta is bound to Notch in trans. In absence of Neur-627 mediated endocytosis of Delta that exerts pulling force on Notch, the clusters are 628 stabilized/not consumed. Numb interacts physically with Spdo to control the 629 subcellular localisation of the Notch/Spdo complex. In the control situation, Numb 630 is not detected in the Notch/Spdo clusters at the pIIa-pIIb interface, suggesting 631 that Notch and Spdo clusters at the interface are predominantly on the pIIa side. 632 Loss of Numb that leads to recycling of Notch/Spdo towards the plasma 633 membrane of the pIIb cell results in the increase in the number and intensity of 634 Notch/Spdo/Baz clusters at the pIIa-pIIb interface. Overexpression of Numb 635 causes the disappearance of Notch/Spdo clusters at the pIIa-pIIb interface. 636 By analogy to vertebrates, we anticipate that Numb by its ability to bind to Baz 637 (Nishimura and Kaibuchi, 2007), is somehow competing with Baz for the access 638 to Notch/Sdpo, and therefore formation of Notch/Spdo/Baz signaling clusters. 639 Based on the fact that loss of Spdo leads to a stronger reduction in Baz/Notch 640 clusters assembly, a simple prediction is that Baz interacts with Spdo/Notch. 641 Future studies will aim to investigate this possibility to determine which domain of 642 Baz, its ability to oligomerize in order to promote Notch/Baz/cluster assembly. 643

Concluding remarks 645
Due to the conservation of intra-lineage communication, it will be interesting to 646 investigate whether cell-cell communication interface exhibit an atypical apico-647 basal polarity and if Par3-dependent clustering of Notch also applies to pass a 648

Declaration of interest 669
The authors declare no competing interests 670 Drosophila melanogaster strains were grown and crossed at 25°C. 687

Material and methods
Somatic clones were generated using the FLP-FRT system with a hs-FLP. 688 Crosses were passed in new tubes every 2 days and then, FLP expression was 689 induced by at least 2 heat shocks (1h at 37°C) from embryonic stage for baz EH747 690 clones and from first instar larval stage for all the other clones. 691 The pnr-GAL4 driver was used to drive the expression of: UAS-Notch dsRNA; 692

Live-imaging and image analyses 796
Pupae aged of around 16h30 APF were prepared for imaging as described 797 previously (Daniel et al., 2018). Briefly, the pupa is positioned between a glass 798 slide and a coverslip coated with a thin layer of Voltalef, the coverslip being 799 supported anteriorly and posteriorly by columns made of 5 and 6 little coverslips. 800 Images were acquired at 25°C on a LSM 880 AiryScan or LSM TCS SPE and 801 processed using FIJI. 802

Statistical tests 805
Statistical difference between two conditions was evaluated by a F test followed 806 by a student t test using Microsoft Excel. Statistical significances were 807 represented as follows: * : p value

Colocalisationlevel measurement 829
In order to evaluate the degree of similarity of Baz and Notch clusters dynamics, 830 we generated kymographs from high time resolution (Δt = 2 s) acquisitions at 831 pIIa/pIIb new interface compared to epidermal/epidermal interfaces apically and 832 laterally (only at pIIa/pIIb interface). We then applied the coloc 2 plugin from FIJI 833 on the kymographs using the following settings: threshold regression = Costes, 834 PSF = 4.0. We choose to use the Mander's coefficient (Manders et al., 1993) 835 above autothreshold values to evaluate the colocalisation between NiGFP and 836 Ubi-Baz-mCherry tracks observed. Mander's coefficients represent respectively 837 the percentage of total signal from NiGFPchannel which overlaps with Ubi-Baz-838 mCherry signal and reciprocally the Ubi-Baz-mCherrysignal which overlaps with 839 NiGFP signal. 840 841

Clusters counting 855
To count the number of NiGFP clusters between pIIa and pIIb nuclei, we 856 developed a macro working with FIJI (script available upon request). Briefly, first, 857 a threshold is applied to both pIIa and pIIb nuclei allowing for the delimitation of 858 the nuclei inside ROIs. Then an ovoid mask including both nuclei ROIs is 859 generated. From this mask, the initial nuclei ROIs are substracted to keep only a 860 "RenyiEntropy" is applied and finally the clusters are detected using an "Analyse 862 particles". At the end, the macro refers to the size and NiGFP fluorescent 863 intensity of each cluster detected. Note that two erroneous situations which and we fixed the minimal cluster area to 0,03µm 2 and the maximal cluster area to 874 0,2 µm 2 . As for the minimal intensity threshold, we found a linear correlation with 875 the apical mean fluorescence: intensity threshold = 0,2654 X apical mean 876 fluorescence + 227,6. We applied these 2 thresholds successively to the images 877   F  i  g  u  r  e  1  :  D  i  s  t  r  i  b  u  t  i  o  n  o  f  p  o  l  a  r  i  t  y  m  a  r  k  e  r  s  a  n  d  N  o  t  c  h  d  u  r  i  n  g  S  O  P  c  e  l  l  d  i  v  i  s  i  o  n   1114   T  i  m  e  -l  a  p  s  e  i  m  a  g  i  n  g  o  f  C  r  b  -G  F  P  (   A  ,  A S  3  :  D  e  l  t  a  r  e  g  u  l  a  t  e  s  t  h  e  d  y  n  a  m  i  c  s  o  f  N  i  G  F  P  a  t  t  h  e  p  I  I  a  -p  I  I  b  i  n  t  e  r  f  a  c  e   1240   (   A   )  T  i  m  e  l  a  p  s  e  i  m  a  g  i  n  g  o  f  N  i  G  F  P  (  g  r  e  e  n  )  a  t  t  h  e  n  o  v  e  l  i  n  t  e  r  f  a  c  e  b  e  t  w  e  e  n  p  I  I  a  -p  I  I  b  c  e  l  l  s   1241   i  d  e  n  t  i  f  i  e  d  u  s  i  n  g   n  e  u  r   H  2  B  -R  F  P  u  p  o  n  s  i  l  e  n  c  i  n  g  o  f  D  e  l  t  a  .  Y  e  l  l  o  w  a  r  r  o  w  h  e  a  d  s  s  h  o  w  l  a  t  e  r  a  l   1242   c  l  u  s  t  e  r  s  p  o  s  i  t  i  v  e  f  o  r  N  i  G  F  P  .  T  i  m  e  i  s  i  n  m  i  n  ,  s  c  a  l  e  b  a  r  s  a  r  e  5 μm. 1243  F  i  g  u  r  e  5  :  N  e  u  r  l  o  c  a  l  i  s  e  s  i  n  a  n  d  r  e  g  u  l  a  t  e  s  t  h  e  n  u  m  b  e  r  o  f  N  i  G  F  P  /  B  a  z  p  o  s  i  t  i  v  e  c  l  u  s  t  e  r  s  .   1249   (   A   )  L  o  c  a  l  i  z  a  t  i  o  n  o  f  N  e  u  r  -G  F  P  (  g  r  e  e  n  )  a  n  d  B  a  z  -S  c  a  r  l  e  t  (  r  e  d  )  i  n  c  o  n  t  r  o  l  .  N  e  u  r  a  n  d  B  a  z   1250   c  o  l  o  c  a  l  i  s  e  i  n  c  l  u  s  t  e  r  s  b  o  t  h  a  p  i  c  a  l  l  y  a  n  d  l  a  t  e  r  a  l  l  y  a  t  t  h  e  p  I  I  a  -p  I  I  b  i  n  t  e  r  f  a  c  e  (  n  =  8  S  O  P  s  )  .   1251   Y  e  l  l  o  w  d  a  s  h  e  d  l  i  n  e  d  e  l  i  n  e  a  t  e  s  t  h  e  e  n  l  a  r  g  e  d  a  r  e  a  s  p  r  e  s  e  n  t  e  d  i  n  t  h  e  i  n  s  e  t  s  (  r  i  g  h  t  p  a  n  e  l  s