key: cord-0257125-v7g4v7or
authors: Chanet, Soline; Huynh, Jean-René
title: Collective cell sorting requires contractile cortical waves in germline cells
date: 2020-01-06
journal: bioRxiv
DOI: 10.1101/2020.01.06.895847
sha: e7c64ec37dc7d348c6fc3570890e2640b6210e0d
doc_id: 257125
cord_uid: v7g4v7or

Encapsulation of germline cells by layers of somatic cells forms the basic unit of female reproduction called primordial follicles in mammals and egg chambers in Drosophila. How germline and somatic tissues are coordinated for the morphogenesis of each separated unit remains poorly understood. Here, using improved live-imaging of Drosophila ovaries, we uncovered periodic actomyosin waves at the cortex of germ cells. These contractile waves are associated with pressure release blebs, which project from germ cells into somatic cells. We demonstrate that these cortical activities, together with cadherin-based adhesion, are required to sort each germline cyst as one collective unit. Genetic perturbations of cortical contractility, blebs protrusion or adhesion between germline and somatic cells induced failures to encapsulate any germ cells or the inclusion of too many germ cells or even the mechanical split of germline cysts. Our results reveal that germ cells play an active role in the physical coupling with somatic cells to produce the female gamete.

In this study, we used hydrogel-based live imaging, genetics and image analyses to 86 investigate the role of germ cells during the very first steps of encapsulation. We show that 87 germline cysts actively generate forces both to maintain their position in the germarium while 88 being surrounded by FCs, and to preserve their integrity as groups of 16 cells. To investigate a potential role of germ cells during encapsulation, we first looked at 95 actomyosin dynamics in the germline. To monitor the dynamics of the actomyosin 96 cytoskeleton, we used a GFP-tagged version of the regulatory light chain of the non-muscle 97 the germarium without interfering with stem cells and the formation of germline cysts. 165

Depleting chic or zip in germ cells effectively reduced cortical myosin intensity and waves 166 frequency compared to a control knocked-down (ctl-RNAi) (Figure 2a , b, Movie 7). We 167 obtained similar reduction of cortical contractility in germline cysts mutant for ROCK (rok 2 ) 168 ( Figure S2a ). We also tested the consequences of increasing cortical contractility by inhibiting 169 mbs, the myosin binding subunit of the myosin phosphatase, which dephosphorylates and 170 inhibits myosin. Depleting mbs in the germline (mbs-RNAi), increased waves frequency 171 ( Figure 2a, b, Movie 7) . 172

The most striking phenotype induced by both reducing or increasing contraction waves in 173 the germline was an abnormal number of germ cells per egg chamber, detected during mid-174 oogenesis (12 to 36h after leaving the germarium) (Figure 2c To test if abnormal egg chambers originated from encapsulation defects, we imaged live 199 the early steps of encapsulation in hydrogel. We followed the displacement of individual cysts 200 along the anterior-posterior (a-p) axis, at the time they started to be separated by FCs (in 201 region 2b and 3 of the germarium). We used cap cells at the anterior tip of the germarium as a 202 fixed reference point and measured cyst displacement by tracking their movements over 1 to 203 2h (Figure 3a In chic-RNAi and zip-RNAi, we also noticed that cysts in region 3 instead of being round 220 adopted an elongated shape along the a-p axis compared to controls (Figure 3c, increased 221 aspect ratio). Live-imaging showed cysts being squeezed and sometimes cut into several parts 222 by ingressing FCs (Figure 3a , zip-RNAi lower panel, Movie 8 and 9). Cysts splitting events 223 would give rise to egg chambers with less than 16 germ cells. In addition, different parts of 224 the split cysts could also be packaged with adjacent cysts generating egg chambers with more 225 than 16 cells. These results indicated that cortical contractility also conferred stiffness to 226 germline cysts, preventing them from being squeezed and cut by surrounding FCs. 227

In mbs-RNAi, cysts movements and speed were also increased compared to control 228 conditions. We measured a mean displacement speed of 0.0203 +/-0.0029 µm.min -1 . Whereas 229 in zip-RNAi and chic-RNAi cysts tend to be more frequently pushed toward the anterior, mbs-230

RNAi cysts tend to move most frequently toward the posterior (Figure 3b ). This behavior was 231 more pronounced for cysts in region 3, that significantly moved faster toward the posterior 232 than ctl-RNAi in region 3 ( Figure S3 ). In the strongest instances, we observed collisions in the 233 posterior regions of the germarium between a fast moving cyst and an older cyst resulting in 234 the encapsulation of the two cysts together (Movie 9). In contrast to chic-RNAi and zip-RNAi, 235 we never observed cysts being split in mbs-RNAi. mbs-RNAi cysts were not squeezed but 236 remained round in region 3 with an aspect ratio close to 1 (Figure 3c To decrease homophilic interactions between germ and somatic cells, we depleted E-257

Cad (shotgun, shg in Drosophila) either in the germline using the bam-Gal 4 driver (ECad-258

RNAi germ, Figure S3a ) or in the FCs using traffic jam-Gal4, which is strongly expressed in 259 FCs (ECad-RNAi soma, Figure S3a The white-shRNA was used as a control (ctl-RNAi) because white is not expressed during 653 oogenesis. Depending on the strength of the shRNA, different drivers were used for 654 knockdowns in the germline. To knockdown zip and mbs we used nos-GAL4, either the 655 original stock or w; sqh::GFP; nos-GAL4 (generated using stocks previously described). To 656 knockdown chic and shg we used a bam-GAL4(x2) driver (containing two copies of the bam-657 GAL4 driver, (Clemot et al., 2018)): w; bam-GAL4(x2) or w; bam-GAL4(x2); Utr::GFP 658 (generated using stocks previously described). For gene knockdowns in follicle cells, we 659 used: w; Traffic jam-GAL4; Utr::GFP (generated using stocks previously described and from 660 Bloomington). Knock downs were performed at 29°C to increase the efficiency of the GAL4 661 driver. 662

To generate Flp out clones we used the stock: w hs-Flp; UASp-GFP; act-FRT-y+-FRT-663 GAL4 (generated using stocks from Bloomington). Heat-shocks were performed on early 664 pupae, 30 min at 37°C. Images were processed using Fiji and Imaris (Bitplane) and graphs were generated in Prism 708 (GraphPad). A bleach correction was applied to time-laps images. Images of the movies 709 represent a maximum intensity Z projection (15m). Waves and blebs frequency (Figure 1, 2  710 and 3) were measured in Fiji, every occurrence of a wave or bleb was counted, the resulting 711 number was then divided by the time of the recording. Wave occurrences were either counted 712 per cell (Figure 1f ) or per cysts (Figures 1i, 2b, 3b) . To measure the average number of blebs 713 per cyst ( Figure 5) , we counted the number of blebs per cyst for each time frame (every 30 714 sec) and divided by the number of frames. This method was used to take into account bleb 715 persistence. 716

Region 3 cyst aspect ratios (Figure 3) were measured in Fiji using the build-in toolbox, an 717 ellipse was fitted to the shape of the cyst (as indicated in Figure 3f ). 718

Cyst displacement were tracked in Imaris. Cap cells at the anterior tip of the germarium were 719 used as a fixed reference point. To estimate cyst displacement, we tracked the oldest ring 720 canal of each cyst, which is the widest and brightest, making it an easy object to track. We 721 projected displacement along the a-p axis using Imaris build-in toolbox and calculated the 722 mean speed over 1 to 2h. By convention, we conferred a negative value for displacement 723 speed toward the anterior and a positive value for displacement speed toward the posterior. 

2010. α-Catenin as a tension 608 transducer that induces adherens junction development

Epithelial polarity and morphogenesis

Regulation of zygotic gene 613 expression in Drosophila primordial germ cells

An Actin-615 Based Wave Generator Organizes Cell Motility

Signaling and adhesion activities of mammalian 617 beta-catenin and plakoglobin in Drosophila

Drosophila Rho-Associated Kinase (Drok) Links Frizzled-Mediated Planar Cell 620 Polarity Signaling to the Actin Cytoskeleton

Time-lapse images from a mosaic germarium containing wild-type (RFP+) germline cysts and 1075

rok 2 mutant cysts (marked by the absence of RFP), and expressing the myosin marker 1076 sqh::GFP. A wild-type cyst (asterisk) migrate and invade the position of a more posterior 1077 mutant cyst

conditions. Only cysts in region 3 that have already transition from disc-shape to a rounder 878 shape were considered. Violin plots with median and 25%-75% quartiles are shown. n = 21 879 cysts in region 3 for moe-TA::GFP; n = 12 cysts in region 3 for moe-TD::GFP. ns, non-880 significant. ** P < 0.01, Mann-Whitney U-test (performed on absolute speed values). 881 (f) Still images from a movie of a germarium overexpressing moe-TA::GFP in the germline.