Sara phosphorylation state controls the dispatch of endosomes from the central spindle during asymmetric division

During asymmetric division, fate assignation in daughter cells is mediated by the partition of determinants from the mother. In the fly sensory organ precursor cell, Notch signalling partitions into the pIIa daughter. Notch and its ligand Delta are endocytosed into Sara endosomes in the mother cell and they are first targeted to the central spindle, where they get distributed asymmetrically to finally be dispatched to pIIa. While the processes of endosomal targeting and asymmetry are starting to be understood, the machineries implicated in the final dispatch to pIIa are unknown. We show that Sara binds the PP1c phosphatase and its regulator Sds22. Sara phosphorylation on three specific sites functions as a switch for the dispatch: if not phosphorylated, endosomes are targeted to the spindle and upon phosphorylation of Sara, endosomes detach from the spindle during pIIa targeting.

5A showing that Sara co-immunoprecipitates with Sds22-GFP, but not with GFP.
Red boxes indicate part of the gel showed in Figure

Fly lines and fly handling
The UAS-Gal4 system was used for over-expression and RNAi experiments.
When using the Gal80 ts protein to modulate the levels of expression, animals were kept at 16°C until puparium formation, and then shifted to the temperature specified below for each experiment. Stocks used in this study were: Neur-Gal4 27 ; Pnr-Gal4 (Bloomington #3039); UAS-mRFP-Pon 7 ; tub- Animals of the following genotypes were studied, and pupae were shifted overnight to the indicated temperatures:

Scanning Electron Microscopy (SEM)
Quantification of the Neur phenotypes: To study the relevance of Sara in Notch signaling during SOP cell-fate assignation, we followed the rationale previously established 15 . We used a partial depletion of Neuralized in the center of the notum (using Pnr>Neur RNAi ), which allows many sensory organs to still undergo asymmetric cell fate assignation and to develop, as in wild-type, into structures containing at least the two external cells. However, these remaining structures are absent in Pnr>Neur RNAi , Sara 12 / Df(2R)48 transheterozygotes mutants, leading to a naked cuticle (Fig. 4C,F,I and Fig. 4K). Importantly, in these transheterozygotes mutants, the number of SOP is 5 times that of control ("Pnr>") animals (Fig. 4C, and see below), ruling out a potential early SOP specification defect that could have led to a loss of SOP mother cells. These Clonal analysis has been previously used to address whether Neuralized acts in pIIa or pIIb exclusively 6 . Upon initiation of a neuralized mutant clone in the SOP, only one of the two daughter cells is homozygous mutant, either pIIa or pIIb 6 . Under these conditions, no phenotype was observed whether pIIa or pIIb is mutant 6 . When pIIb is mutant, it can still inherit from the SOP the Neuralized anterior cortical localization domain (crescent) and this could rescue the loss of the neuralized functional gene; because there is also no mutant phenotype when pIIa is mutant, this might indicate that Neuralized act in pIIb, where the crescent is seen 6 . However, Neuralized is a cytosolic factor that is enriched in the anterior cortex, while the largest pool of the Neuralized protein can remain cytosolic and present in the mutant pIIa and pIIb ("perdurance"). Clonal analysis is therefore too slow to address the relevance of the biased localization of Neuralized at the pIIb cortex or Sara endosomes to the pIIa during asymmetric Notch signaling: it only informs about the perdurance of the protein. The same reasoning applies to Sara, which is a cytosolic protein that is enriched on endosomes.
Quantification of the Sara 3A and Sara F678A phenotypes: To study the relevance of Sara phosphorylation in Notch signaling during SOP cell-fate assignation, we studied the GFP-Sara 3A and GFP-Sara

Image processing
Image processing was performed using Fiji 32 and Matlab (MathWorks).

Detection and tracking of endosomes
Endosomes were detected using two different algorithms depending on the size and shape of the endosome. See also Supplementary Fig. 1. i) Initially all endosomes were detected by fitting 2D Gaussian functions to the pixel intensity profiles of the images as previously ii) When an endosome radius was found to be larger than the diffraction limit, the result of the Gaussian approach was discarded for this specific endosome. Instead, an intensity threshold was applied to segment the endosome and the endosome position was then determined by calculating the position of the center of mass.
Endosome tracking was then performed using a modified Vogel-algorithm 26 .
This algorithm requires an estimation of the motility of the endosomes (the 'diffusion' constant). Because endosome velocities varied significantly during the movie, this caused trajectories to be terminated prematurely when the endosome displacement was too long. These sub-trajectories were manually linked afterwards.

GFP-Sara levels in endosomes
GFP-Sara levels were calculated by dividing the fluorescent intensity in an endosome by the surface area of that endosome. For this purpose, endosomes were considered to be spherical with a diameter equal to the FWHM (Full Width Half Maximum) obtained from the 2D Gaussian fit, or from an approximation to the real shape when no 2D Gaussian fit was available. If an endosome was detected in multiple planes, fluorescence intensities of the endosome in these planes was summed. To segment the population of endosomes according to high and low Sara levels, a threshold (444 au) was chosen by classifying the Sara endosome collection into two groups of approximately the same number of endosomes.

Detection of the Pon-crescent and the division plane
Before cytokinesis, the Pon-crescent was used to determine the division plane that will separate the two daughter cells. We developed an algorithm to automatically detect the Pon-crescent in each image plane and calculate the division plane using the edges of the crescent, see Supplementary Fig. 1.  Figure 1F.

Endosomal asymmetry
The quantification of the proportion of endosomes targeted to the pIIa and the pIIb daughter cells (Fig. 2D, Fig. 4C and Fig. 5E) was performed as described in 1 . Briefly, the total endosomal intensity in the pIIa and the pIIb cells (IpIIa and IpIIb, respectively) was measured by integrating total intensity values in each slice of the first z-stack after abscission, after having subtracted the background and thresholded the endosomes; the percentage of endosomes in the pIIa cell was finally computed as IpIIa / (IpIIa + IpIIb).

Enrichment at the central spindle
Enrichments at the central spindle ( Fig. 1C-D, Fig. 2C, Fig. 2E and Fig. 5C-D) were calculated as in 11  The decay of the time of departure ( Fig. 1D and 5D) was measured the following way. First, the fold enrichment at the central spindle was measured as explained above. Then, the departure phase was defined: in the time period between -260s and +100s (relative to abscission), the departure phase start was defined as the time point in which the fold enrichment reached its maximal value; the departure phase end was defined as 100s after abscission.
During the departure phase, the fold enrichment was fitted to an exponential, where IntDen is the Integrated Density function of ImageJ/Fiji 32 (product of Area and Mean Gray Value). Finally, the calculated enrichments were averaged between cells (n=12 cells for the Control and n=6 cells for dsRNA Sds22) and presented as mean ± standard error of the mean.
Expression levels in GFP-Sara, GFP-Sara3A and GFP-SaraF678A Identical acquisition setup for live imaging was performed for each condition (GFP-Sara, GFP-Sara 3A and GFP-Sara P-values indicated in Supplementary Fig. 2E correspond to a Mann-Whitney Rank Sum Test run after a non-parametric Kruskal-Wallis test, the variables being not normal. P-values indicated in Supplementary Fig. 4B and S5C correspond to a Mann-Whitney Rank Sum Test.
All representations on graphs and values in the text are given as mean ± standard error of the mean.