Capulet and Slingshot share overlapping functions during Drosophila eye morphogenesis
© Lin et al.; licensee BioMed Central Ltd. 2012
Received: 2 March 2012
Accepted: 30 April 2012
Published: 30 April 2012
Skip to main content
© Lin et al.; licensee BioMed Central Ltd. 2012
Received: 2 March 2012
Accepted: 30 April 2012
Published: 30 April 2012
CAP/Capulet (Capt), Slingshot (Ssh) and Cofilin/Twinstar (Tsr) are actin-binding proteins that restrict actin polymerization. Previously, it was shown that low resolution analyses of loss-of-function mutations in capt, ssh and tsr all show ectopic F-actin accumulation in various Drosophila tissues. In contrast, RNAi depletion of capt, tsr and ssh in Drosophila S2 cells all affect actin-based lamella formation differently. Whether loss of these three related genes might cause the same effect in the same tissue remains unclear.
Loss-of-function mutant clones were generated using the MARCM or EGUF system whereas overexpression clones were generated using the Flip-out system. Immunostaining were then performed in eye imaginal discs with clones. FRAP was performed in cultured eye discs.
Here, we compared their loss-of-function phenotype at single-cell resolution, using a sheet of epithelial cells in the Drosophila eye imaginal disc as a model system. Surprisingly, we found that capt and ssh, but not tsr, mutant cells within and posterior to the morphogenetic furrow (MF) shared similar phenotypes. The capt/ssh mutant cells possessed: (1) hexagonal cell packing with discontinuous adherens junctions; and (2) largely complementary accumulation of excessive phosphorylated myosin light chain (p-MLC) and F-actin rings at the apical cortex. We further showed that the capt/ssh mutant phenotypes depended on the inactivation of protein kinase A (PKA) and activation of Rho.
Although Capt, Ssh and Tsr were reported to negatively regulate actin polymerization, we found that Capt and Ssh, but not Tsr, share overlapping functions during eye morphogenesis.
Remodeling of the actin cytoskeleton is controlled by various groups of actin-binding proteins that function at different steps to promote dynamic F-actin assembly and disassembly. Cyclase-associated protein (CAP) homologs have been suggested to act as actin monomer sequestering proteins through their C-terminal actin-binding domains to suppress the spontaneous polymerization of actin [1, 2]. Cofilin, an actin depolymerization factor, severs and depolymerizes older F-actin from the pointed end of the filament . Cofilin is inactivated by phosphorylation of an N-terminal serine by LIM kinase  and activated by removal of the phosphate by Ssh, a cofilin phosphatase .
capt and tsr encode the Drosophila CAP and cofilin orthologues, Capulet (Capt) and Twinstar (Tsr), respectively. Loss-of-function mutations in capt, ssh and tsr all cause increased ectopic accumulation of F-actin in various Drosophila tissues that leads to defects in epithelial morphogenesis [5–9]. For example, it has been shown  that Capt, by affecting actin polymerization, alters cell shape and therefore affects the distribution of Hh. However, RNAi depletion of capt, tsr, and ssh in Drosophila S2 cells affects actin-based lamella formation differently .
Drosophila retinal differentiation starts in third instar larvae when a moving MF sweeps across the developing eye disc in a posterior-to-anterior direction. Cells anterior to the MF proliferate and have a large apical surface whereas cells within the MF undergo apical constriction. Moreover, cells posterior to the MF differentiate and form ommatidial clusters that secrete Hh proteins [11, 12], whereas the remaining interommatidial cells (ICs) relax and regain a large apical surface. The anteriorly-diffused Hh activates the expression of target genes such as dpp and atonal in a strip of cells immediately ahead of the MF . The transient activation of the Hh pathway in these cells converts Ci from a repressor (Ci75) to an activator (Ci155). Hh-mediated apical constriction in the MF requires RhoA/Rho kinase and other unidentified kinases that act through both myosin II and Diaphanous to respectively cause the enrichment of activated myosin II and F-actin in the apical cortex of these cells [9, 14]. Rho kinase activates myosin II via phosphorylation of Ser19 of the MLC.
In this report, we compared the loss-of-function phenotypes of capt, tsr, and ssh using the sheet of epithelial cells in the Drosophila eye imaginal disc as a model system. We demonstrated that capt and ssh, but not tsr, mutants within and posterior to the MF similarly showed: (1) largely complementary accumulation of excessive F-actin and p-MLC, and (2) hexagonal cell packing with discontinuous AJs in the eye epithelial cells. We also found that these phenotypes depended on the inactivation of PKA and activation of Rho. Thus, Capt and Ssh have common functions in regulating actin depolymerization during eye morphogenesis.
Loss-of-function mosaic clones were generated using the MARCM system . Entirely capt E636 or ssh 1-63 mutant eyes were generated using the EGUF (eyeless-Gal4 UAS-FLP) system of recombination . Flip-out clones were generated by P[act5C > y+ > GAL4] P[UAS-GFP.S65T]/CyO.
For immunostaining, third instar larval eye imaginal discs were dissected and fixed in 4% paraformaldehyde. Antibodies used were: rat anti-DE-cad (DCAD2; 1:50; Hybridoma Bank), mouse anti-Arm (N2-7A1; 1:40; Hybridoma Bank), rabbit anti-pSer19-MLC (1:10; Cell Signaling Technology), rat anti-Ci155 (2A1; 1:1; Hybridoma Bank), rabbit anti-Egfr (1:50; ), Alexa 594-phalloidin (1:200; Invitrogen), and Cy3- and Cy5-conjugated secondary IgGs (Jackson Immuno Research Laboratories).
Images were acquired using a 63× NA1.4 Oil Plan-Apochromat objective lens on a confocal microscope (LSM510, Carl Zeiss). The fluorescence intensity and apical surface area were quantified using Zeiss LSM software. The percentage of cells with x-sided polygons was quantified manually by measuring the number of neighboring cells. For fluorescence recovery after photobleaching (FRAP), third instar larval eye discs with ubi-DE-cad-GFP were dissected and placed in a drop of serum-free M3 medium at room temperature. A small area (1500 nm2) was photobleached. Fluorescence intensity in the bleached region or tricellular junction was measured at each time point using Zeiss LSM software.
Next, we determined whether loss of Capt in the ICs behind the MF affected the distribution of the actomyosin network. Unlike wild-type ICs where both p-MLC and F-actin were barely detectable, we observed strong apical accumulation of F-actin into a ring-like structure in sub-cortical regions of the capt E636 mutant ICs (Figure 1D, E). Of note, the fragmented AJs formed in regions where F-actin rings from two adjacent capt E636 mutant ICs came close together (arrowheads in upper inset of Figure 1E′) whereas AJs failed to form at the tricellular junctions where F-actin was absent (arrowhead in lower inset of Figure 1 E′). Surprisingly, we also observed dramatic accumulation of p-MLC at the apical cortex of capt E636 mutant ICs (arrowhead in Figure 1F), with the levels of p-MLC accumulation much higher than wild-type ICs (315 ± 12%, n = 21 cells). Importantly, this strong p-MLC staining outlined the cell morphology and partially overlapped with the F-actin ring (Figure 1F). Thus, fragmented AJs assemble at regions where cadherin/catenin, F-actin and myosin II overlap but do not assemble at the tricellular junction where F-actin is missing.
Next, we determined the dynamics of DE-Cad within the fragmented AJs by analyzing the distribution of ubi-DE-cad-GFP in cultured entirely capt E636 mutant eye discs. As expected, DE-Cad-GFP localized to the middle of each side of the hexagon but was missing at the tricellular junction (Figure 1G). In tests of fluorescence recovery after photobleaching (FRAP) in an area (1500-nm2) of the fragmented AJs in capt E636 mutant cells, we found no significant recovery of fluorescence signal for over 55 s (Figure 1H). This is in contrast to the control, which showed faster recovery of DE-Cad-GFP fluorescence when we performed FRAP in the wild-type AJs (1500-nm2) of ubi-DE-cad-GFP discs (Figure 1H). Moreover, the fluorescence intensity in the tricellular junction of capt E636 mutant cells was very low, and we found that the fluorescence signal in a small area (350-nm2) of a tricellular junction did not significantly change for up to 3 minutes (Figure 1I, J). Together, our results suggest that the fragmented AJ was relatively stable and did not diffuse laterally towards the tricellular junction.
Ssh positively regulates cofilin/Tsr activity to promote F-actin severing and depolymerization . We thus looked next at clones mutant for Tsr, using the tsr N96A mutation. Compared to the endogenous MF cells, tsr N96A mutant cells in the MF showed enlarged apical surface as reported . However, unlike cap E636 and ssh 1-63 MF cells, tsr N96A MF cells contained largely normal levels and distribution of F-actin and p-MLC that overlapped with AJs (Figure 3D,E). Moreover, these cells did not adopt a hexagonal cell shape (Figure 3D,E). tsr N96A mutant cells posterior to the MF did show strong accumulation of cortical F-actin (Figure 3F) and, possibly via apical constriction, often led to groove formation when the clones were large (data not shown). Thus, although capt, ssh and tsr all affected the apical surface, capt and ssh differed from tsr in their effects on F-actin polymerization, MLC phosphorylation and cell-shape change within the MF.
Twinfilin (Twf), similar to ADF/cofilin, sequesters actin monomers and negatively regulates F-actin formation . However, we did not detect excessive F-actin accumulation, hexagonal cell shape, or discontinuous AJs when twf 110 null allele mutant clones were generated using MARCM within and posterior to the MF (Figure 3G and data not shown). Together, our results suggest that the effects in eye epithelial cells are capt/ssh-specific.
It was shown that Hh signaling acts through Ci to regulate Rho/Rok/myosin II to cause apical constriction . We next determined whether the accumulation of p-MLC and F-actin in capt mutant cells is dependent on the activation of the downstream effector Rho. To test this, we overexpressed a dominant negative form of Rho (RhoN19) in capt E636 mutant clones using MARCM. These clones tended to be small, however, we found that inhibition of Rho activity abolished capt-mediated accumulation of F-actin and p-MLC within and posterior to the MF (Figure 4F,G), indicating that the accumulation of p-MLC and F-actin in capt mutant cells required the activation of Rho. Together, our results suggest that the capt mutant phenotypes detected were dependent on the inactivation of PKA and activation of Rho.
Cofilin is activated by Ssh, the phenotypic differences between capt/ssh and tsr mutant MF cells are unexpected. However, as mentioned in the Introduction, RNAi depletion of tsr and ssh in Drosophila S2 cells affects actin-based lamella formation differently . Recent studies also show that loss of capt, but not tsr, leads to activation and nuclear import of Yorkie in the wing disc [27, 28]. Thus, the effect generated by the lack of cofilin (as in tsr mutant cells) is not completely the same as disrupting cofilin and p-cofilin recycling (as in ssh mutant cells). The reason for this is unclear. One possibility is that cofilin activity is regulated by multiple additional mechanisms, including dephosphorylation of phosphorylated cofilin by chronophin and general phosphatase . Alternatively, Capt and Ssh have a common function independent of cofilin.
Spot AJs in early embryonic epithelia contain homophilic DE-cad clusters in a stable microdomain and are bona fide sites of adhesion . Small but stable actin patches were suggested to underlie the stability of spot AJs . We found that, in actin-turnover defective capt/ssh mutant cells, fragmented AJs were stable and they assembled at regions where cadherin and high levels of F-actin overlapped. Thus, fragmented AJs and spot AJs might have a similar function but they differ in underlying F-actin organization. We showed that the fragmented AJs can also non-autonomously assemble along the interface between wild-type cells when they were surrounded by the capt E636 mutant cells. One possibility is that when there is a fragmented AJ formed at one side of a wild-type cell (along the interface between wild-type and capt E636 mutant cells), this fragmented AJ cannot link to the adjacent AJ (along the interface between wild-type cells) at the tricellular junction, and therefore this adjacent AJ also becomes fragmented once the associated actomyosin contracts.
It has been proposed that capt is required to prevent premature photoreceptor differentiation ahead of the MF . Here, we showed that the capt/ssh mutant phenotypes, including the apical surface area, cannot be detected anterior to the MF. By using the enrichment of EGF receptor to mark MF cells , we observed roughly 8–10 rows of enlarged capt mutant cells with enrichment of EGF receptor in a capt mutant clone spanning the MF (Additional file 1: Figure S3A). This is similar to the presence of about 10 rows of endogenous MF cells in the wild-type disc . Thus, the numbers of MF cells are largely the same in both wild-type and capt mutant discs. However, as the capt mutant MF cells have enlarged apical area, they might occupy positions anterior to the endogenous MF and thus cause the distortion of the MF toward the anterior of the MF (Additional file 1: Figure S3A). However, this distortion of the MF, at lower resolution, was previously interpreted as premature photoreceptor differentiation caused by anteriorly diffused Hh .
It was previously shown at low resolution that capt, ssh and tsr mutant cells all similarly cause accumulation of F-actin and enlarged apical area in the eye epithelia ,. Here, we found that capt and ssh, but not tsr, mutant cells share various phenotypes that are dependent on the inactivation of PKA and activation of Rho. In addition to the accumulation of F-actin as previously reported, these phenotypes also exhibited excessive accumulation of p-MLC, fragmented AJs and hexagonal cell shape.
We are grateful to J. Treisman, T. Uemura, L. Luo, C.-T. Chien, the Developmental Studies Hybridoma Bank and Bloomington Stock Center for providing reagents and stocks. This research was supported by grants from the National Science Council, Taiwan to J.-C. Hsu.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.