Fgfr-Ras-MAPK signaling is required for apical constriction via apical positioning of Rho-associated kinase during mechanosensory organ formation

Abstract

Many morphogenetic movements during development require the formation of transient intermediates called rosettes. Within rosettes, cells are polarized with apical ends constricted towards the rosette center and nuclei basally displaced. Whereas the polarity and cytoskeletal machinery establishing these structures has been extensively studied, the extracellular cues and intracellular signaling cascades that promote their formation are not well understood. We examined how extracellular Fibroblast growth factor (Fgf) signals regulate rosette formation in the zebrafish posterior lateral line primordium (pLLp), a group of ∼100 cells that migrates along the trunk during embryonic development to form the lateral line mechanosensory system. During migration, the pLLp deposits rosettes from the trailing edge, while cells are polarized and incorporated into nascent rosettes in the leading region. Fgf signaling was previously shown to be crucial for rosette formation in the pLLp. We demonstrate that activation of Fgf receptor (Fgfr) induces intracellular Ras-MAPK, which is required for apical constriction and rosette formation in the pLLp. Inhibiting Fgfr-Ras-MAPK leads to loss of apically localized Rho-associated kinase (Rock) 2a, which results in failed actomyosin cytoskeleton activation. Using mosaic analyses, we show that a cell-autonomous Ras-MAPK signal is required for apical constriction and Rock2a localization. We propose a model whereby activated Fgfr signals through Ras-MAPK to induce apical localization of Rock2a in a cell-autonomous manner, activating the actomyosin network to promote apical constriction and rosette formation in the pLLp. This mechanism presents a novel cellular strategy for driving cell shape changes.

Fig. 1.

Fgfr-Ras-MAPK signaling in the pLLp is required for rosette formation. (A-D) Confocal projections showing rosette formation in the DMSO control (A) and with Fgfr (B), Ras (C) and MAPKK (D) inhibition in claudinB:EGFP zebrafish embryos at 30 hpf. Arrowheads indicate centers of the trailing rosettes. Pink arrow indicates nascent rosette. Brackets indicate rosette-free region. (A′-D′) pMAPK immunolabeling in single planes from projections in A-D. Fgfr and MAPKK inhibition with 100 μM SU5402 and 7 μM PD0325901, respectively, was from 28-30 hpf. Ras inhibition via hsp70:dn-Ras induction was at 28 hpf; embryos were fixed 2 hours following heat shock. (E-G) Stills from time-lapse movie of pLLp in a wild-type claudinB:EGFP embryo (supplementary material Movie 1). (F) Cells align apical ends along the midline (pink arrows). (G) Center of the nascent rosette (arrowhead). (E′-G′) Three-dimensional reconstructions of the highlighted cell in E-G. (H) ACIs for embryos treated with DMSO, SU5402, PD0325901 or following induction of hsp70:dn-Ras. n=180 cells from six embryos per condition. ^**P<0.0001, Wilcoxon test. Error bars indicate s.e.m. Scale bars: 20 μm in A-G; 4 μm in E′-G′.

Fig. 2.

Rho kinase activity is required for rosette formation, apical constriction and MRLC activation. claudinB:EGFP zebrafish embryos were treated from 28 to 30 hpf with the Rho kinase inhibitor Rockout. (A-D) ACIs were then measured for the leading-most 30 cells (C,D) and compared with DMSO-treated controls (A,B). Colors of surfaces in A,C correspond to cell positions in B,D. (E) ACI measurements from embryos treated with Rockout or DMSO. n=180 cells from six embryos per condition. ^**P<0.0001, Wilcoxon test. Error bars indicate s.e.m. (F-G′) Immunolabeling showing loss of pMRLC from the leading region following Rockout treatment as compared with control. (H) Quantification of the leading region, where pMRLC is not detected. ^*P<0.0001, Student’s t-test. Percentages were derived by counting the number of cells caudal to the distalmost pMRLC signal (i.e. the number of leading cells that lacked the signal) and normalizing to the total number of cells in the pLLp. Scale bars: 20 μm.

Fig. 3.

Fgf and MAPK signaling are required for localization of Rock2a and MRLC activation. (A-F′) Fgfr or MAPKK inhibition with SU5402 or PD0325901, respectively, from 28-30 hpf in claudinB:EGFP zebrafish embryos. Lower panels show sagittal view of leading pLLp region (left, EGFP; right, Rock2a). Post-treatment, Rock2a (A-C) and pMRLC (D-F) were assayed by immunolabeling. Yellow arrowheads indicate the caudal-most apical accumulation of Rock2a. Note that Rock2a is not localized to apical ends of cells in the leading region following treatments. pMRLC staining shows failure of leading-region MRLC activation following treatments. (G) Quantification of the leading region (performed as in Fig. 2H). Rock2a (n=6 embryos; P<0.03, ANOVA) and pMRLC (n=6 embryos; P<0.003, ANOVA) are not apically localized. Note there are fewer leading cells with apically localized Rock2a in Fgf-inhibited (39.9±1.1%) and MAPKK-inhibited (42.5±1.4%) embryos compared with the control (26.4±1.5%). For pMRLC staining: DMSO, 28.0±0.7%; SU5402, 34.4±0.6%; and PD0325901, 39.8±1.0%. Scale bar: 20 μm.

Fig. 4.

Ras-MAPK signaling mediates cell-autonomous apical constriction and Rock2a localization. (A,B) claudinB:EGFP-positive mosaic zebrafish embryos containing hsp70:dn-Ras donor cells (red) at 30 hpf. Note the lack of pMAPK labeling in dn-Ras cells. (C-D′) Before heat shock, both dn-Ras and wild-type donor cells are columnar; 2 hours after heat shock, the wild-type cell is constricted, whereas the dn-Ras cell remains columnar. (E) Quantification of ACIs of transplanted wild-type and hsp70:dn-Ras cells before and after heat shock. n=50 cells from ten embryos. ^**P<0.009, Wilcoxon test. n.s., not significant. (F-G′) Transplantation of hsp70:dn-Ras or wild-type cells causes no obvious differences in global Rock2a distribution. (H,H′) Rock2a distribution in a single transplanted cell shows failure of Rock2a apical localization when Ras is inhibited. (I) Ratio of apical to basal fluorescence intensity in transplanted wild-type cells and dn-Ras cells. n=15 cells from four embryos per condition. ^*P<0.03, ANOVA. Error bars indicate s.e.m. Scale bars: 20 μm in A,F; 2 μm in C.