The conserved Hippo tumor suppressor pathway is a key kinase cascade that controls tissue growth by regulating the nuclear import and activity of the transcription co-activator Yorkie. Here, we report that the actin-Capping Protein αβ heterodimer, which regulates actin polymerization, also functions to suppress inappropriate tissue growth by inhibiting Yorkie activity. Loss of Capping Protein activity results in abnormal accumulation of apical F-actin, reduced Hippo pathway activity and the ectopic expression of several Yorkie target genes that promote cell survival and proliferation. Reduction of two other actin-regulatory proteins, Cofilin and the cyclase-associated protein Capulet, cause abnormal F-actin accumulation, but only the loss of Capulet, like that of Capping Protein, induces ectopic Yorkie activity. Interestingly, F-actin also accumulates abnormally when Hippo pathway activity is reduced or abolished, independently of Yorkie activity, whereas overexpression of the Hippo pathway component expanded can partially reverse the abnormal accumulation of F-actin in cells depleted for Capping Protein. Taken together, these findings indicate a novel interplay between Hippo pathway activity and actin filament dynamics that is essential for normal growth control.
INTRODUCTION
The conserved Hippo (Hpo) signal transduction pathway has emerged as a crucial regulator of tissue size in both Drosophila and mammals (Edgar, 2006). In Drosophila, the Hpo pathway comprises a kinase cascade in which the Hpo kinase binds the WW domain adaptor protein Salvador (Sav) to phosphorylate and activate the Warts (Wts) kinase (Pantalacci et al., 2003; Tapon et al., 2002; Udan et al., 2003; Wu et al., 2003). Wts, in turn, facilitated by Mats (Lai et al., 2005; Wei et al., 2007), phosphorylates and prevents nuclear translocation of the transcriptional co-activator Yorkie (Yki) (Huang et al., 2005; Oh and Irvine, 2008). This leads to transcriptional downregulation of target genes that positively regulate cell growth, survival and proliferation, including the Drosophila inhibitor of apoptosis protein 1 (Diap1; thread – FlyBase), expanded (ex), Merlin (Mer) and wingless (wg) in the inner distal ring, within the proximal wing imaginal disc (Reddy and Irvine, 2008). The upstream components Ex, Hpo and Wts are also thought to regulate Yki through a phosphorylation-independent mechanism, by directly binding to Yki, sequestering it in the cytosol and thereby abrogating its nuclear activity (Badouel et al., 2009; Oh et al., 2009).
Multiple upstream inputs are known to regulate the core Hpo kinase cassette at various levels (Grusche et al., 2010; Halder and Johnson, 2010). Thus, the atypical cadherin Fat was identified as an upstream component of the Hpo pathway and was proposed to transduce signals from the atypical cadherin Dachsous (Ds) and Four-jointed (Fj), a Golgi-resident kinase that phosphorylates Fat and Ds (Bennett and Harvey, 2006; Cho et al., 2006; Hamaratoglu et al., 2006; Silva et al., 2006; Willecke et al., 2006). Moreover, the two Ezrin-Radixin-Moesin (ERM) family members, Ex and Mer have been reported to lie upstream of the Hpo pathway (Hamaratoglu et al., 2006). Mer and Ex can heterodimerize (Badouel et al., 2009; Boedigheimer et al., 1993; Bretscher et al., 2002; McCartney et al., 2000) and are believed to exert their growth suppression activity by activating the Hpo kinase. However, how the different inputs that feed into the core kinase cassette are coordinated to regulate Yki activity is unknown.
ERM proteins form a structural linkage between transmembrane components and actin filaments (F-actin) (McClatchey and Giovannini, 2005). For instance, mammalian Mer binds numerous cytoskeletal factors, including actin, and appears to act as an inhibitor of actin polymerization (McClatchey and Giovannini, 2005; Scoles, 2008). Interestingly, the Merlin-actin cytoskeleton association is required for growth suppression and inhibition of epidermal growth factor (EGFR) signaling (Curto and McClatchey, 2008). Moreover, F-actin depolymerization promotes activation of the Hpo orthologs MST1 and MST2 in mouse fibroblasts (Densham et al., 2009). These observations suggest a role for F-actin dynamics in modulating Hpo pathway activity.
Actin filament growth, stability and disassembly are controlled by a plethora of actin-binding proteins. Among these, the Capping Protein (CP) heterodimer, composed of α (Cpa) and β (Cpb) subunits, acts as a functional heterodimer to restrict the accessibility of the filament barbed end, inhibiting addition or loss of actin monomers (Cooper and Sept, 2008). In Drosophila, mutations in either cpa or cpb, lead to accumulation of F-actin within the cell and give rise to identical developmental phenotypes that are tissue specific. In the wing blade (BL), the most distal domain of the imaginal disc, cpa and cpb prevent cell extrusion and death, but they are not required for this function in the most proximal domain, the prospective body wall and the hinge wing imaginal disc (Janody and Treisman, 2006). The Cofilin homolog Twinstar (Tsr) and the Cyclase-associated protein Capulet (Capt) also restrict actin polymerization: Tsr severs filaments and enhances dissociation of actin monomers from the pointed end (Bamburg, 1999), whereas Capt sequesters actin monomers, preventing their incorporation into filaments (Gottwald et al., 1996).
Here, we investigate the relationship between the control of the actin cytoskeleton and Hpo pathway activity. We show that the actin-binding proteins CP and Capt, but not Tsr, enhance Hpo signaling activity. Moreover, we uncover a new relationship between the Hpo pathway and the machinery that regulates F-actin, and reveal that Hpo signaling activity, like CP, limits F-actin accumulation at apical sites independently of Yki. Finally, we propose that regulation of an apical F-actin network by Hpo signaling activity and CP sustains Hpo pathway activity, thereby limiting Yki nuclear import and the activation of proliferation and survival genes.
MATERIALS AND METHODS
Fly strains and genetics
The cpa, tsr and cpb alleles, in addition to the UAS-HA-cpa transgenic flies and the generation of clones have been previously described (Janody and Treisman, 2006). Other fly stocks used were ex697 (lacZ reporter) (Hamaratoglu et al., 2006); sav3 (Tapon et al., 2002); Mer4 (LaJeunesse et al., 1998); matse235 (Lai et al., 2005); hpo42–47, a gift from D. Pan (Wu et al., 2003); exe1 (Boedigheimer et al., 1993); wtslatsX1 (Xu et al., 1995); UAS-hpoM11.1 (Pantalacci et al., 2003); UAS-ykiS168A (Dong et al., 2007); UAS-yki, UAS-Mer1-600 (LaJeunesse et al., 1998); UAS-yki-GFP (Oh and Irvine, 2008); UAS-ex (Udan et al., 2003); UAS-Fat (Sopko et al., 2009); Diap1::lacZ (Hay et al., 1995); sd-Gal4 (Klein and Arias, 1998); hh-Gal4, a gift from T. Tabata (University of Tokyo, Japan); ap-Gal4; tub-Gal80ts (Bloomington Stock Center, IN, USA); UAS-hpo-IR (Pantalacci et al., 2003); UAS-yki-IR4005R-2; UAS-capt-IR5061R-2 and UAS-wts-IR12072R-1 (National Institute of Genetics, NIG, Japan); UAS-cpa-IR7009 (Vienna Drosophila Research Center, VDRC, Austria).
Immunohistochemistry
Third instar larval imaginal discs were stained with antibodies as described (Lee and Treisman, 2001). Antibodies used were mouse anti-Arm [N2 7A1, 1:10; Developmental Studies Hybridoma Bank (DSHB)], rabbit anti-Hth (1:500; A. Salzberg, Technion-Israel Institute of Technology, Haifa, Israel), guinea pig anti-Hth (1:3000; R. S. Mann, Columbia University Medical Center, New York, NY, USA), mouse anti-Nub (1:10; S. M. Cohen, Institute of Molecular and Cell Biology, Singapore), rabbit anti-Ex (1:200; A. Laughon, University of Wisconsin, Madison, WI, USA), guinea pig anti-Mer (1:4000; R. Fehon, University of Chicago, IL, USA); mouse anti-Wg (4D4, 1:10; DSHB), rabbit anti-Yki (1:100) (Oh and Irvine, 2008), rabbit anti-Caspase 3 (1:500; BD Bioscience), pMLC (1:10; Cell Signaling), mouse anti-β-galactosidase (1:200; Promega), mouse anti-HA (1:5000; Covance), mouse anti-BrdU (1:200; Roche). Rhodamine-conjugated Phalloidin (Sigma) was used at a concentration of 0.3 μM. BrdU labeling was assayed as described (Johnston and Schubiger, 1996). Secondary antibodies were from Jackson ImmunoResearch, used at 1:200 or Molecular Probes, used at 1:500. Fluorescence images were obtained on a Leica TCS NT or a LSM 510 Zeiss confocal microscope. The NIH Image J program was used to perform measurements. To measure roundness, clones recovered in the proximal wing epithelium that were either FRT40A, wild type or FRT42D, wild type or FRT40A, cpbM143 or FRT42D, cpa107E, were outlined and the circularity was calculated as 4πιA/L2, where A is the clone area and L is the perimeter. To measure the surface area of cpb mutant clones with no GFP and of sibling twin spots with two copies of GFP, clones recovered in the proximal wing disc epithelia were outlined separately. The surface area for each clone was calculated in pixels and converted to μm2 (where 1024 pixels corresponds to 387.5 μm). To measure the length of the Nub-Hth expression domain, the apical surface of the Nub-Hth positive domains were outlined on optical cross-sections (blue lines in Fig. 1E-G) and the length was measured on each disc for each genotype. To quantify the intensity of BrdU signal, the posterior and anterior compartment of hh>GFP and hh>cpa-IRC10 wing discs were outlined separately for each disc and the intensity levels were calculated as the sum of the gray values of all the pixels in the selection divided by the number of pixels for each compartment. Values were then normalized with those of hh>GFP control discs. To quantify the intensity of Phalloidin signals, a region of interest (ROI) of 100 × 100 pixels (~40 cells) was selected. The sum of the gray values was measured for each ROI, applied twice to the posterior and to the anterior compartments (on each side of the dorsal/ventral and anterior/posterior boundaries) on standard confocal sections comprising the apical surface. The ratio of signal between the posterior and anterior compartments was calculated for each disc. Statistical significance was calculated using a two-tailed t-test.
Western blotting
Proteins were extracted from wing imaginal discs expressing UAS-fat, UAS-cpa-IRD4 or UAS-wts-IR12072R-1 under the control of sd-GAL4, at 25°C, using a lysis buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1 mM EDTA pH 7.4, 1% NP-40) in the presence of protease inhibitors (Roche) and phosphatase inhibitors (Sigma). After adding the lysis buffer, samples were frozen in liquid nitrogen and subsequently spun at 13,000 g for 15 minutes at 4°C. Protein amounts in the supernatant were quantified by the Bradford method to normalize quantities of extract used for loading between sd>w1118 discs and sd>fat; sd>cpa-IRD4; sd>wts-IR overexpressing wing discs. Total extracts were prepared according to NuPAGE Bis-Tris Mini Gel Protocol (Invitrogen) and run on a 10% NuPAGE Bis-Tris Gel (Invitrogen). Wet electrotransfers of gels to nitrocellulose membranes (Amersham Hybond-P, GE Healthcare) were performed at 30 mA overnight at 4°C. Blots were blocked for 2 hours at room temperature in PBST (0.2% TritonX-100 in PBS) containing 5% low-fat milk and incubated overnight at 4°C in PBST containing 5% low-fat milk supplemented with rabbit anti-pYkiS168 (1:1000) or rat anti-Yki (Yki69 1:1000; N. Tapon, Cancer Research UK, London, UK) or mouse anti-βTub (E7; 1/2000; DSHB). Blots were washed four times in PBST and incubated for 1 hour in PBST containing 5% low-fat milk supplemented with ECL anti-rabbit IgG-conjugated HRP or anti-rat IgG-conjugated HRP ECL anti-mouse IgG-conjugated HRP (1:5000 dilution; GE Healthcare). After four washes in PBST, blots were developed using Amersham ECL Plus Western Blotting Detection System (GE Healthcare). The same membrane was first probed with anti-pYkiS168, followed by Yki and anti-tubulin antibodies after stripping.
Molecular biology
To make the UAS-cpa-IR construct, a 447 bp DNA fragment was amplified by PCR from genomic DNA of Oregon flies, using the oligonucleotides dscpa5′ (ATTATCTAGAGGTACCAGCTCTGTTTTTGGAAAGGC containing XbaI and KpnI sites) and dscpa3′ (ATTATCTAGATTATTGCGTCTTCAGTTCCT containing a XbaI site) and subcloned into the XbaI site of pWIZ (Lee and Carthew, 2003) to give rise to the pWIZcpa3-447 construct. The same PCR product digested with XbaI was then subcloned into the SpeI site of the pWIZcpa3-447 construct. Off-targets were analyzed using dsCheck (http://dscheck.rnai.jp/) (Naito et al., 2005). No off-targets from 429 mers were predicted. Three independent transgenic lines (UAS-cpa-IRC10, UAS-cpa-IRB4 and UAS-cpa-IRD4) were generated by standard methods (Spradling and Rubin, 1982).
RESULTS
Capping Protein suppresses proliferation and overgrowth of proximal wing tissue
To understand the role of CP in epithelia, we analyzed in more detail groups of cells mutant for cpa and cpb in the proximal wing imaginal disc, where cell polarity and organization of the epithelium monolayer is not affected (Janody and Treisman, 2006). We noticed that the boundaries of mutant clones were abnormally straight (Fig. 1A), and the clones were also rounder than their wild-type twin spots (Fig. 1C), suggesting that mutant cells minimize contacts with wild-type neighbors. When induced in first instar larvae, 40% of cpb mutant clones displayed increased staining of the apical marker Armadillo (Arm) around or within the clones (arrows in Fig. 1A). Optical cross-sections through the wing disc showed that the mutant tissue was folded, but maintained the columnar shape of a polarized epithelium (Fig. 1B). Fold formation is therefore unlikely to be caused by a change in cell shape but might result from excess proliferation, giving rise to cyst-like clones that appear rounder than wild-type clones. To evaluate the growth rate of cells mutant for cpb, we induced mutant clones at first instar larvae and measured in the proximal wing domain, the surface area of cpb mutant clones with no GFP, compared with their wild-type twin spots that contain two doses of GFP. cpb mutant clones were on average 25% larger than wild-type twin spots (see Fig. S1A in the supplementary material). However, a mixture of very small (blue arrow in Fig. S1B in the supplementary material) and big clones (blue asterisk in Fig. S1B in the supplementary material) were recovered, suggesting that only some cpb mutant clones could overgrow.
To further assess the cellular requirement for CP, we generated stable lines of transgenic flies carrying an inverted repeat construct capable of expressing intron-spliced hairpin dsRNA for cpa under Gal4-UAS control (UAS-cpa-IR). Flies expressing these constructs displayed similar phenotypes to flies mutant for the cpa or cpb allele (Janody and Treisman, 2006), including cell extrusion and death within the distal wing epithelium (see Fig. S1C-D in the supplementary material) and apical F-actin accumulation (see Fig. S1E-F in the supplementary material). To investigate the consequences of knocking down cpa on tissue growth in the proximal wing domain, we drove expression of cpa-IR in the developing wing using wing disc-specific drivers. Expressing cpa-IR with apterous-Gal4 (ap-Gal4) induced abnormal growth of the dorsal wing disc compartment, characterized by the formation of extra folding (Fig. 1D; see Table S1 in the supplementary material). To quantify the amount of overgrowth, we expressed cpa-IR with a driver that would induce overgrowth of the proximal wing tissue without triggering tissue misfolding. scalloped-Gal4 (sd-Gal4) drives GFP expression in the whole wing disc primordium at early larval stages (Klein and Arias, 1998). By mid-third instar larvae, GFP was then restricted to the central BL and distal hinge (DH) domains (see Fig. S2A in the supplementary material) and progressively faded in this later region that co-expresses the two markers Nubbin (Nub) and Homothorax (Hth) (see Fig. S2B-C in the supplementary material). As expected, expressing dsRNA directed against Hpo with sd-Gal4 triggered an expansion of 55% (P<0.001) of the DH domain, as measured on cross-sections by an increase of the apical length of the Nub/Hth-positive domain (Fig. 1H and compare Fig. 1G with 1E). Upon expression of two independent cpa dsRNA, the DH domain also showed a 21.5% and 26.2% increase (P<0.03) in length, respectively (Fig. 1H and compare Fig. 1F with 1E). We conclude that CP limits the growth of the distal hinge tissue.
To test the effects of loss of CP on cell proliferation, we analyzed the pattern of bromodeoxyuridine (BrdU) incorporation, which identifies cells undergoing DNA synthesis. We observed reproducible increases in BrdU labeling in cells expressing cpa-IR in the whole posterior compartment of the wing imaginal disc under hedgehog-Gal4 (hh-Gal4) control (Fig. 1I). We quantified this effect by measuring the intensity of BrdU signal in posterior versus anterior wing compartments in hh>cpa-IR and control hh>GFP wing discs. After normalization with hh>GFP control discs, we found that signal intensity was 34% higher (P<0.004) in cells expressing cpa-IR (Fig. 1J). Taken together, these data indicate that CP limits cell proliferation in the whole wing disc epithelium but restricts tissue overgrowth only in the proximal wing disc epithelium.
Capping Protein inhibits ectopic expression of Yorkie target genes
The Hpo signaling pathway is a crucial regulator of tissue size (Edgar, 2006). We therefore tested whether loss of CP affects Hpo pathway target genes. Indeed, cpa or cpb mutant clones accumulated Ex (Fig. 2A) and Mer (see Fig. S4C in the supplementary material) at the apical cell surface (see Fig. S4B,D in the supplementary material) and Wg in the inner distal ring, within the proximal wing (Fig. 2C). Increase expression of Ex (Fig. 2B), Wg (Fig. 2D, arrows in D′) and Mer (see Fig. S2D in the supplementary material) could also be observed when we drove cpa-IR with hh or ap-Gal4. We then tested whether Ex upregulation occurs at the level of transcription. Expression of a lacZ enhancer trap insertion into the ex gene (ex-lacZ) was upregulated in cpa mutant clones (Fig. 2E) and in ap>cpa-IR-expressing discs (Fig. 2F). Moreover, cpb mutant clones (Fig. 2G) or hh>cpa-IR-expressing cells (Fig. 2H) also upregulated expression of a lacZ enhancer trap insertion in the Diap1 gene. Interestingly, we also noticed that loss of CP affected Hpo target genes non-cell autonomously. Whereas ex-lacZ and diap1-lacZ were upregulated in wild-type cells adjacent to the CP mutant tissue (blue arrows in Fig. 2E′,G′), Ex protein levels were reduced (blue arrow in Fig. 2A). This suggests that CP affects Yki transcriptional activity cell autonomously but also non-autonomously in wild-type neighboring cells. In agreement, Yki colocalized with the nuclear transcription factor Hth in cpb mutant clones (Fig. 2I) and in wild-type cells adjacent to the clone border (blue arrow in Fig. 2I′), whereas it showed a diffuse staining in wild-type cells located a few rows far from the mutant clone (white asterisk in Fig. 2I). We also analyzed by western blot phosphorylation of Yki at Ser168, a crucial phosphorylation site induced upon activation of the Hpo pathway (Dong et al., 2007). Like Wts-depleted tissues (sd>wts-IR), extracts from sd>cpa-IR contained decreased phospho-Yki levels, but did not display changes in the total level of Yki (Fig. 2J). Thus, consistent with the observed changes in Yki localization, loss of CP results in a decrease in phosphorylated Yki and upregulation of Yki target genes. Moreover, we observed a strong enhancement of growth and Wg expression when we combined expression of cpa dsRNA with hpo or wts dsRNA (see Fig. S3A-F and Table S1 in the supplementary material). By contrast, expressing a yki dsRNA or overexpressing hpo could suppress the growth of CP-depleted tissue (see Fig. S3G-K and Table S1 in the supplementary material). Together, we conclude that CP limits Yki nuclear import cell autonomously and non-cell autonomously, thereby inhibiting its transcriptional activity.
In the distal wing disc epithelium, loss of CP does not lead to tissue overgrowth but triggers cell extrusion and death (Janody and Treisman, 2006). We therefore wondered whether Yki target genes were also upregulated in dying CP mutant cells. To analyze the expression of Yki target genes in this region, we examined wing imaginal discs 36 hours after induction of cpa or cpb mutant clones, allowing us to recover mutant cells still present within the epithelium. In these cells, Ex (see Fig. S4A in the supplementary material) and Mer (see Fig. S4C in the supplementary material) were also upregulated. Moreover, in the eye imaginal disc, cpa or cpb mutant cells also upregulated diap1-lacZ (see Fig. S4E in the supplementary material), Ex (see Fig. S4F,G in the supplementary material) and Mer (see Fig. S4H in the supplementary material). We therefore conclude that the effect of CP on Yki activity is not specific to the proximal wing disc epithelium. Although in the proximal wing disc epithelium, loss of CP triggers upregulation of Yki target genes and tissue growth, in other tissues, such as in the distal wing epithelium, despite inhibition of Hpo signaling activity, other factors such as the polarity status might prevent CP mutant cells from surviving and growing.
F-actin accumulation per se does not trigger activation of Yorkie target genes
CP inhibits the turnover of actin monomers at the filament barbed end (Schafer et al., 1995). We therefore wondered whether accumulation of F-actin per se could disrupt Hpo signaling activity. We first examined mutations in the Cofilin homolog twinstar (tsr), which accumulates F-actin around the entire cell cortex (Janody and Treisman, 2006) (see Fig. S5A in the supplementary material). tsr mutant clones neither accumulated Wg (Fig. 3A), nor upregulated ex-lacZ (Fig. 3C). However, when we expressed a dsRNA directed against the cyclase-associated protein capulet (capt), which restricts actin filament accumulation near the apical surface (Janody and Treisman, 2006) (see Fig. S5B in the supplementary material), we saw that posterior hh>capt-IR-expressing cells accumulated Wg (arrows in Fig. 3B) and upregulated ex-lacZ (Fig. 3D). These data indicate that Yki target genes are transcriptionally upregulated when Capt levels are reduced. Together, these observations argue that accumulation of F-actin is not sufficient to trigger Yki activity, but support a role for an apical F-actin network in limiting Yki activity.
Hippo signaling activity regulates F-actin independently of Yorkie
Mammalian Mer has been shown to inhibit actin polymerization (McClatchey and Giovannini, 2005; Scoles, 2008). Moreover, the association of Merlin with the actin cytoskeleton is required to suppress growth (Curto and McClatchey, 2008). Because CP inhibits actin polymerization and limits Yki activity, one way by which Ex and Mer could regulate Hpo signaling activity is through the control of F-actin. We therefore wondered whether Drosophila Ex and Mer also have an effect on F-actin. When we induced ex mutant clones, we observed that, like cells in which Cpa levels were reduced (see Fig. S1E-F in the supplementary material), mutant cells accumulated F-actin near the apical surface (Fig. 4A; see Fig. S6A in the supplementary material). However, loss of Mer had only a weak effect on F-actin accumulation (see Fig. S7A-B in the supplementary material). Interestingly, clones mutant for hpo (Fig. 4B; see Fig. S6B in the supplementary material), sav (Fig. 4C; see Fig. S6C in the supplementary material), mats (Fig. 4D; see Fig. S6D in the supplementary material) and wts (Fig. 4E; see Fig. S6E in the supplementary material) also accumulated apical F-actin, indicating that Ex and the whole kinase cassette inhibit F-actin formation. As Ex is an ERM family member, which can form a structural linkage between transmembrane components and the actin cytoskeleton (McClatchey and Giovannini, 2005), we then tested whether Ex prevents F-actin accumulation independently of its effect on the kinase Hpo. Overexpressing hpo suppressed growth and seemed to prevent F-actin accumulation in ex mutant clones (blue arrow in Fig. 4F). Some clones constricted apically and appeared to contain increased apical F-actin (blue arrow in Fig. 4G). However, this effect is likely to result from a reduction of the apical cell surface in constricted tissues. Wild-type clones overexpressing hpo were also very small in size and seemed to slightly reduce apical F-actin (blue arrow in Fig. 4H). Therefore, the effect of Ex on F-actin might be through activation of the Hpo kinase. We conclude that Hpo pathway activity reduces F-actin at apical sites.
We then tested whether Hpo signaling activity controls F-actin through Yki activity. In wild-type clones, overexpressing Yki (Fig. 5A; see Fig. S6F in the supplementary material) or a constitutively active form of Yki (ykiS168A) (Fig. 5B; see Fig. S6G in the supplementary material) induced morphological defects but did not promote F-actin accumulation. Moreover, ex mutant clones in which Hpo signaling output is blocked by expressing yki dsRNA still accumulated F-actin at the apical surface (Fig. 5D; see Fig. S6I in the supplementary material), whereas wild-type clones expressing yki dsRNA did not significantly show disruption of the F-actin network (Fig. 5C; see Fig. S6H in the supplementary material). This indicates that Hpo signaling activity reduces apical F-actin independently of Yki.
Hippo signaling activity and Capping Protein regulate apical F-actin
An HA-tagged form of Cpa, which can fully rescue cpa mutant animals to viability (Janody and Treisman, 2006), colocalized with Ex at apical sites (Fig. 6A). Moreover, both CP and Hpo signaling activities regulate apical F-actin. We therefore asked whether increasing Hpo signaling modulates the effect of CP loss on F-actin. Overexpression of a constitutively active form of Mer with ap-Gal4 did not affect F-actin organization nor induced morphological defects (see Fig. S7C-D in the supplementary material). As cells overexpressing ex (Fig. 6B) or hpo (see Fig. S3H in the supplementary material) grow very poorly, we induced the expression of these genes late in development using a thermosensitive form of Gal80 expressed under the control of the ubiquitous tubulin promoter (tub-Gal80ts) to inhibit GAL4 activity at earlier stages. tub-Gal80ts, hh-Gal4; UAS-ex or UAS-hpo larvae switched from 18°C to 25°C at early third instar and dissected 48 hours later maintained ex (Fig. 6C) or hpo (Fig. 6F)-overexpressing cells within the wing disc epithelium and did not show significant disruption of the apical F-actin network. By contrast, Gal80ts; hh>cpa-IR wing imaginal discs displayed increased F-actin in the posterior compartment (Fig. 6D). Interestingly, ex overexpression partially suppressed the F-actin accumulation caused by knocking down Cpa (Fig. 6E). We quantified this effect by measuring the ratio of Phalloidin signals between the posterior and anterior compartments for each genetic combination (Fig. 6H). This ratio was 0.99 for Gal80ts; hh>GFP control discs. By contrast, Gal80ts; hh>cpa-IR wing imaginal discs contained 1.82 times more F-actin at the apical surface of posterior cells. This actin accumulation was significantly reduced by ex overexpression (P/A ratio: 1.15, P<0.005), although ex overexpression in wild-type discs had no significant effect on F-actin. Overexpressing hpo also seemed to partially suppress F-actin accumulation of cpa-IR expressing cells (P<0.07). These results indicate that Hpo signaling activity acts downstream or in parallel to CP on F-actin.
DISCUSSION
In this report, we show an interdependency between Hpo signaling activity and F-actin dynamics in which CP and Hpo pathway activities inhibit F-actin accumulation, and the reduction in F-actin in turn sustains Hpo pathway activity, preventing Yki nuclear translocation and upregulation of proliferation and survival genes (Fig. 7).
Capping Protein and Hippo signaling activities inhibit F-actin
ERM proteins can form a structural linkage between transmembrane components and the actin cytoskeleton (McClatchey and Giovannini, 2005). Mammalian Mer appears to act as an inhibitor of actin polymerization (McClatchey and Giovannini, 2005; Scoles, 2008). Moreover, the Mer-actin cytoskeleton association has a crucial role for growth suppression and inhibition of EGFR signaling (Curto and McClatchey, 2008). In Drosophila, Mer and Ex are structurally related and appear to have partially redundant functions but vary in their requirement depending of the tissue or developmental stage (Boedigheimer et al., 1993; Boedigheimer et al., 1997; Hamaratoglu et al., 2006; McCartney et al., 2000). In imaginal discs, loss of ex shows stronger phenotypes when compared with those of Mer (Boedigheimer et al., 1993; Boedigheimer et al., 1997; Hamaratoglu et al., 2006; McCartney et al., 2000). Ex might also have a stronger requirement on F-actin dynamics, as loss of ex, but not that of Mer, triggered F-actin accumulation (Fig. 4; see Figs S6, S7 in the supplementary material). Surprisingly, loss of hpo, sav, mats or wts also triggered apical F-actin accumulation (Fig. 4; see Fig. S6 in the supplementary material), independently of Yki activity (Fig. 5; see Fig. S6 in the supplementary material). Ex is likely to affect F-actin through activation of the Hpo kinase cassette because in most ex mutant clones, overexpressing hpo suppressed F-actin accumulation. Some clones seemed to contain increased F-actin (Fig. 4). However, these clones also constricted apically, suggesting that the effect on F-actin levels results from a reduction of the apical cell surface and that in the absence of ex, differential activity of overexpressed hpo triggers cell shape changes. Together, these observations argue that Ex prevents apical F-actin accumulation through Hpo signaling activity but independently of Yki.
Loss of Hpo pathway activity or CP triggered apical F-actin accumulation (Figs 4, 6; see Figs S1, S3, S6 in the supplementary material). Ex localizes to the sub-apical region of epithelial cells (Oh et al., 2009; Reddy and Irvine, 2008), and colocalized with an HA-tagged form of Cpa (Fig. 6), Ex, Hpo, Sav and Wts all interact with each other through WW and PPXY motifs (Oh et al., 2009; Reddy and Irvine, 2008). Therefore, a pool of Hpo, Sav and Wts, localized at apical sites, could directly regulate an actin-regulatory protein. Hpo pathway activity might act downstream of CP on F-actin. In agreement with this, ex overexpression significantly suppressed F-actin accumulation in cells with reduced CP levels (Fig. 6). The role of Hpo signaling activity might be to inhibit an actin-nucleating factor, which adds new actin monomers to filament barbed ends free of the capping activity of CP. However, we cannot exclude that ex overexpression enhances the activity of residual Cpa in cells knocked down using RNAi, nor that Hpo pathway activity acts in parallel to CP on F-actin. Interestingly, although endogenous Ex is upregulated in cells lacking CP (Fig. 2; see Fig. S4 in the supplementary material), mutant cells still accumulated F-actin (Janody and Treisman, 2006). wts mutant clones also upregulated Ex, which, when overexpressed, suppresses growth of wts mutant clones (Badouel et al., 2009). Therefore, the increased levels of endogenous Ex in cells lacking either CP or wts appears to be insufficient to fully suppress the effects of loss of CP or wts on F-actin and growth, respectively.
Actin cytoskeleton genes control Hippo signaling activity
Our data indicates that CP inhibits Yki nuclear accumulation, activation of Yki target genes, and consequently overgrowth of the proximal wing epithelium (Figs 1, 2; see Fig. S4 in the supplementary material). Interestingly, Yki was also found to accumulate in nuclei of wild-type cells adjacent to the clone border. Consistent with a non-autonomous effect of CP loss on Hpo pathway activity, ex-lacZ and diap1-lacZ were upregulated in wild-type cells adjacent to CP mutant clones. However, Ex levels were reduced in wild-type neighboring cells (Fig. 2). Cells expressing different amounts of ds and fj also upregulate ex-lacZ, but show reduced levels of Ex (Willecke et al., 2008). Therefore, loss of CP might affect Fj or Ds levels, creating a sharp boundary of their expression. However, in contrast to clones overexpressing ds or mutant for fj, cell lacking CP also upregulated Ex and Mer inside the mutant clones, indicating that CP also acts cell-autonomously to promote Hpo signaling activity. CP might facilitate Yki phosphorylation by the Hpo kinase cassette as cpa-depleted tissues contained decreased phospho-Yki levels (Fig. 2). But, we cannot exclude the possibility that CP also favors the direct binding of non-phosphorylated Yki to Ex, Hpo or Wts (Oh et al., 2009). Further analysis will be required to elucidate the mechanisms by which CP restricts Yki activity cell autonomously and in wild-type neighboring cells.
Our results argue for a constitutive role of CP in Hpo pathway activity as Yki target genes were upregulated in the whole wing and eye imaginal discs (Fig. 2; see Fig. S4 in the supplementary material). However, loss of CP did not fully recapitulate the phenotype for core components of the hpo pathway (Edgar, 2006). Despite that, on average, cpb mutant clones located in the proximal wing disc domain were 25% larger than wild-type twin spots; 60% of mutant clones failed to grow (Fig. 1; see Fig. S1 in the supplementary material). Moreover, in the distal wing epithelium, reducing CP levels induces mislocalization of the adherens junction components Armadillo and DE-Cadherin, extrude and death (Janody and Treisman, 2006) (see Fig. S1 in the supplementary material). Furthermore, in Drosophila, CP also prevents retinal degeneration (Delalle et al., 2005; Johnson et al., 2008). This indicates that although loss of CP can, under certain conditions, result in tissue overgrowth due to inhibition of Hpo pathway activity, other factors such as the polarity status, also determine the survival and growth of the mutant tissue. Therefore, in addition to promoting Hpo pathway activity, CP has additional developmental functions in epithelia. However, we cannot exclude the possibility that, like most upstream inputs that feed into the Hpo pathway (Halder and Johnson, 2010), CP has a tissue-specific requirement in Hpo pathway activity.
CP, Capt and Tsr all restrict F-actin assembly directly (Bamburg, 1999; Cooper and Sept, 2008; Gottwald et al., 1996). CP and Capt control F-actin formation near the apical surface and inhibit ectopic expression of Yki target genes, whereas Tsr acts around the entire cell cortex and has no effect on Yki target genes (Janody and Treisman, 2006) (Fig. 3; see Fig. S5 in the supplementary material). This argues that Hpo signaling activity is not affected by the excess of F-actin per se but provides significant support to the view that stabilization of an apical F-actin network by CP, Capt and Hpo signaling activity feeds back on the Hpo pathway to sustain its activity.
Our findings do not allow us to conclude where F-actin accumulation intersects Hpo signaling activity because both Hpo signaling activity and F-actin dynamics feedback to each other. For instance, hpo or ex overexpression suppressed growth of CP-depleted cells (see Fig. S3 in the supplementary material; data not shown). But, overexpressed ex and possibly hpo also suppressed F-actin accumulation of Cpa-depleted cells (Fig. 6). The control of F-actin by Hpo signaling activity and CP might constitute a parallel input, which sustains Hpo pathway activity. Alternatively, F-actin could act upstream of the core kinase cascade, which in turn feeds back to F-actin, to maintain its activity. The identification of additional actin cytoskeletal components that either promote Hpo pathway activity or act downstream of Hpo pathway activity on F-actin would help to discriminate between these possibilities.
Mechanisms by which F-actin regulates Hippo signaling activity
How F-actin influences Hpo signaling activity remains to be determined. The apical F-actin network, which regulates the formation and movement of endocytic vesicles from the plasma membrane (Smythe and Ayscough, 2006), might promote the recycling or degradation of Hpo pathway components. Increased F-actin at apical sites would, therefore, affect protein turnover. Alternatively, apical F-actin might act as a scaffold to tether Hpo pathway components apically. In support of this, Ex, Hpo, Sav, Wts and Yki could all interact between each other through WW and PPXY motifs at apical sites (Oh et al., 2009; Reddy and Irvine, 2008). Moreover, expression of a membrane-targeted form of Mats enhances Hpo signaling (Ho et al., 2010). Although Ex and Mer were properly localized in CP mutant cells (see Fig. S4 in the supplementary material), other members of the pathway might be mislocalized in the presence of excess F-actin. Interestingly, in mouse fibroblasts, the Hpo orthologs MST1 and MST2 colocalize with F-actin structures and are activated upon F-actin depolymerization (Densham et al., 2009), suggesting that by tethering Hpo pathway components, F-actin dynamics modulates their activity. Finally, the F-actin network might act as a mechanical transducer. Most of the mechanosensitive responses require tethering to force-bearing actin filaments (Vogel and Sheetz, 2006). Tissue surface tension has been proposed to be a stimulus for a feedback mechanism that could regulate tissue growth (Aegerter-Wilmsen et al., 2007; Hufnagel et al., 2007). The tension exerted by neighboring cells might be sensed at the cell membrane by the actin cytoskeleton and translated to the regulation of cell proliferation through the Hpo signaling pathway.
Acknowledgements
We thank S. Cohen, R. Mann, A. Laughon, A. Salzberg, B. Hay, N. Tapon, K. D. Irvine, D. J. Pan, R. G. Fehon, the Bloomington Drosophila Stock Center, the Drosophila Genomics Resource Center, the Vienna Drosophila Research Center (VDRC), the National Institute of Genetics (NIG) and the Developmental Studies Hybridoma Bank (DSHB) for fly stocks and reagents. The manuscript was improved by the critical comments of Nicolas Tapon, Jessica Treisman, Francesca Pignoni, Monica Bettencourt-Dias and Moises Mallo. This work was supported by grants (PTDC/SAU-OBD/73191/2006 and PTDC/BIA-BCM/71674/2006) from Fundação para a Ciência e Tecnologia (FCT). B.G.F., P.G., C.B.-P. and B.J. were the recipient of fellowships from FCT (SFRH/BPD/35915/2007, SFRH/BD/47261/2008, SFRH/BPD/46983/2008 and SFRH/BD/33215/2007, respectively).
References
Competing interests statement
The authors declare no competing financial interests.