- Email: [email protected]
Cortactin modulates cell migration and ring canal morphogenesis during Drosophila oogenesis
Mechanisms of Development 121 (2004) 57–64 www.elsevier.com/locate/modo
Cortactin modulates cell migration and ring canal morphogenesis during Drosophila oogenesis Ka´lma´n Somogyi, Pernille Rørth* European Molecular Biology Laboratory, Developmental Biology Programme, Meyerhofstrasse 1, 69117 Heidelberg, Germany Received 13 August 2003; received in revised form 20 October 2003; accepted 20 October 2003
Abstract Cortactin is a Src substrate that interacts with F-actin and can stimulate actin polymerization by direct interaction with the Arp2/3 complex. We have isolated complete loss-of-function mutants of the single Drosophila cortactin gene. Mutants are viable and fertile, showing that cortactin is not an essential gene. However, cortactin mutants show distinct defects during oogenesis. During oogenesis, Cortactin protein is enriched at the F-actin rich ring canals in the germ line, and in migrating border cells. In cortactin mutants, the ring canals are smaller than normal. A similar phenotype has been observed in Src 64 mutants and in mutants for genes encoding Arp2/3 complex components, supporting that these protein products act together to control specific processes in vivo. Cortactin mutants also show impaired border cell migration. This invasive cell migration is guided by Drosophila EGFR and PDGF/VEGF receptor (PVR). We find that accumulation of Cortactin protein is positively regulated by PVR. Also, overexpression of Cortactin can by itself induce F-actin accumulation and ectopic filopodia formation in epithelial cells. We present evidence that Cortactin is one of the factors acting downstream of PVR and Src to stimulate F-actin accumulation. Cortactin is a minor contributor in this regulation, consistent with the cortactin gene not being essential for development. q 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: EMS1; Cytoskeleton; Motility
1. Introduction Cortactin was originally identified as a major substrate of the Src tyrosine kinase (Wu et al., 1991). The name Cortactin reflects that the protein binds to F-actin and that it localizes to the cell cortex, including membrane ruffles and lamellipodia (Wu and Parsons, 1993). These features, plus the presence of an SH3 domain and proline-rich regions in the Cortactin protein, suggested that Cortactin might link signaling events to the actin cytoskeleton. Phosphorylation of Cortactin stimulated by Src modulates its activity in vivo (Huang et al., 1997). Cortactin phosphorylation and subcellular localization are also affected by receptor tyrosine kinases (RTKs) (Maa et al., 1992; Zhan et al., 1993). The cortactin gene was also identified as EMS1, a putative oncogene encoding one of the transcripts amplified in certain human carcinomas (Schuuring et al., 1993). Directed overexpression of EMS1/ * Corresponding author. Tel.: þ 49-6221387109; fax: þ 49-6221387166. E-mail address: [email protected] (P. Rørth). 0925-4773/$ - see front matter q 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.mod.2003.10.003
cortactin was subsequently shown to increase the motility of and invasion of fibroblasts (Patel et al., 1998) and the metastatic potential of breast cancer cells (Li et al., 2001). Cortactin also enriched in ‘invadopodia’ from invasive tumor cells, cellular protrusions associated with degradation of extracellular matrix (Bowden et al., 1999). Together, these studies suggest that Cortactin may play a role in promoting cell motility of cancer cells, as well as of normal cells in response to growth factor stimulation. Recent studies have shown that Cortactin can directly influence actin polymerization. The Arp2/3 complex is an important nucleator of actin polymerization. It stimulates actin polymerization while binding to actin filaments and can thereby induce formation of a branched actin network (Higgs and Pollard, 2001). The Arp2/3 complex consists of multiple proteins and is very conserved through evolution, from yeast to man (Higgs and Pollard, 2001). Cortactin can bind to and activate the Arp2/3 complex as well as stabilize the resulting F-actin network (Uruno et al., 2001; Weaver et al., 2001). Cortactin may act synergistically with proteins of the WASP family, which are potent regulators of Arp2/3
58
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
activity (Weaver et al., 2002). In addition to WASP and N-WASP, the WASP family includes the SCAR/WAVE proteins. The WASP family of proteins is regulated by GTPases of the Rho subfamily, namely CDC42 and Rac: WASP by direct binding (Higgs and Pollard, 2001), SCAR/WAVE by dissociation of an inhibitory complex (Eden et al., 2002). Cortactin can also associate with the adaptor protein WIP (WASP interacting protein), which binds monomeric actin and thus may aid Cortactin in actin nucleation (Kinley et al., 2003). Arp2/3 and its regulators are required for extensions of lamellipodia and other actinrich structures, but also for intracellular trafficking events such as endocytosis (Moreau et al., 1997). Cortactin has been suggested to link endocytosis to Arp2/3 mediated actin polymerization in mammalian cells. Cortactin is associated with dynamin2 and with vesicles (McNiven et al., 2000). Consistent with this type of function, inhibition of Cortactin inhibits receptor mediated endocytosis (Cao et al., 2003). A Cortactin protein has been identified in Drosophila (Katsube et al., 1998), but not in yeast. It is possible that other proteins substitute for Cortactin in yeast. It is also possible that the function of Cortactin is specific to higher eukaryotes that, contrary to yeast, use tyrosine kinases for cell –cell communication and cell regulation. There are multiple RTKs in Drosophila, which control cell growth, differentiation and morphogenesis including EGFR, insulin receptor and PDGF/VEGF receptor. The Drosophila genome also encodes a number of non-receptor tyrosine kinases such as Src. There are two Src genes, Src42 and Src64. Src42 may act downstream of RTKs to modify their signaling (Takahashi et al., 1996). Src64 is specifically required for proper morphogenesis of actin-rich structures in the female germ line, called ring canals (Dodson et al., 1998). Drosophila Cortactin protein is localized to the cell cortex and shows protein– protein interactions similar to that of mammalian Cortactin (Katsube et al., 1998). To understand the function of Cortactin in vivo, we have generated complete loss-of-function mutants of Drosophila cortactin. Cortactin mutants are fully viable and are fertile, but have subtle defects during oogenesis. Our analysis suggests that Cortactin acts downstream of RTKs and Src in vivo. The analysis also indicates that Cortactin is only a minor mediator of the effects of tyrosine kinases on the actin cytoskeleton.
2. Results Mutations in the cortactin locus were generated by imprecise excision of a P-element located upstream of the transcription start site (Fig. 1A, see also Section 4). One of the four deletion alleles precisely removes the cortactin transcription unit (cort M7) and two others most of the coding region (cort D4 and cort A4). All four alleles showed the same phenotypes, but the M7 and D4 alleles were used for further analysis. The cortactin mutant flies are
Fig. 1. Cortactin mutants and protein distribution in egg chambers. (A) The cortactin locus and the four cortactin mutants. The genomic region removed in each of the deletion mutants is indicated below (see Section 4 for details). (B,C) Anti-Cortactin staining of stage 9 egg chambers from wild type (B) and cort M7/cort M7 mutant females (C). The mutant encodes no Cortactin protein, so the staining in C represents background staining. The cortical staining as well as staining of border cells (arrow) and ring canals (arrowheads) is specific. In these and all other micrographs of egg chambers, anterior is to the left. Border cells migrate from the anterior of the egg chamber to the oocyte. (D) Higher magnification view of follicle cells stained with anti-Cortactin showing staining along the cell cortex. Scale bars indicate 20 mm in B and C, 10 mm in D.
homozygous viable and show no visible abnormalities. The stock becomes completely homozygous without selection, showing that there is no requirement for maternal contribution of cortactin, nor is there a large growth disadvantage due to loss of cortactin. However, we noticed that the females had somewhat reduced fertility and therefore examined oogenesis in more detail. During oogenesis, development proceeds in egg chambers consisting of centrally located germ line cells (15 ‘nurse cells’ and one oocyte) surrounded by a simple monolayer epithelium, the follicular epithelium. Fig. 1B shows a wild type egg chamber at stage 9 stained with a Cortactin antibody. The control for specificity of the antibody is shown in Fig. 1C, an egg chamber of the same stage from a cortactin deletion mutant. At this stage, a small cluster of follicle cells called border cells, delaminate from the follicular epithelium, invade the germ line cluster
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
and migrate to the oocyte. Cortactin was specifically enriched in migrating border cells (indicated by an arrow in Fig. 1B,C). Cortactin was also enriched at the actin rich structures called ring canals (arrowheads in Fig. 1B). Ring canals are formed early in oogenesis in place of the cleavage furrow after incomplete cytokinesis of germ line cells. Proper function of ring canals is necessary for transfer of material from nurse cells to the oocyte. Finally, Cortactin protein was detected along the cortex of the follicle cells (Fig. 1D). A number of specific defects were observed in cortactin mutants. These phenotypes were confirmed to be due to loss of cortactin, as they could be rescued by a transgene driving ubiquitous expression of cortactin cDNA under control of a tubulin promoter. One cortactin phenotype was a mild defect in ‘dumping’, transfer of bulk cytoplasmic material from nurse cells to the oocyte (Fig. 2A,B). This phenotype is similar to that observed in Src64 mutants (Dodson et al., 1998), but is also seen in other mutants. In the case of Src64, the defect was correlated to the presence of Src64 protein at the ring canals and to reduced size of the ring canals. As shown in Fig. 1B, Cortactin is also present at ring canals. To determine whether cortactin affected ring canal morphogenesis, we quantified their size at stage 10. Ring canal size was significantly reduced in the cortactin mutant
Fig. 2. Phenotypes of cortactin mutants in the ovary. (A) A stage 11 wild type egg chamber. (B) The mild ‘dumping defect’: a ‘stage 11’ mutant egg chamber, staged by overall size and F-actin network accumulation in the nurse cells. The relative size of the oocyte is too small indicating a defect in transfer of material to the oocyte. (C) Ring canals of stage 10 wild type and mutant egg chambers and size quantification with standard error indicated. The outer diameter of ring canals closest to the oocyte was measured in multiple egg chambers ðn . 87Þ: (D) Quantification of egg size from wild type females as well as from cort M7/cort M7 and from Src64 D17/Src64 D17 mutant females (n . 107). (E) Discontinuities in the follicular epithelium observed in some cortactin mutant egg chambers. Normally, the epithelium covers all of the germ cells. Between the arrowheads, follicle cells are missing. Scale bars indicate 20 mm in A, B and E, 10 mm in C.
59
relative to wild type (Fig. 2C). Dumping defects usually result in smaller eggs. Eggs from cortactin mutant mothers and from Src64 mutant mothers were on average smaller than wild type (Fig. 2D). The hatching rate of the eggs from mutant mothers was also decreased. Thus, mutations in cortactin affect the actin rich ring canals, and processes dependent on function of ring canals, in a manner similar to mutations in Src64. Mutations in components of the Arp2/3 complex (Arpc1 or Arp3 mutants) also specifically affect ring canal morphogenesis (Hudson and Cooley, 2002). With low penetrance (around 6% of stages 5 –9 egg chambers), cortactin mutants also display discontinuities in the follicular epithelium (Fig. 2E), which is not observed in the wild type situation. A similar phenotype has been observed in other mutants such as crumbs (Tanentzapf et al., 2000), where the affected protein is thought to have a basic function in establishing the epithelium. The phenotype suggests that Cortactin has a role in epithelial integrity. The low penetrance indicates that this function mostly can be compensated by other proteins. Border cell migration was also abnormal in cortactin mutants (Fig. 3). As cortactin has been proposed to be important for migration of normal and invasive mammalian cells, this defect seemed particularly interesting. To quantify border cell migration, the progression of the cluster (bracket in Fig. 3A) was compared to the stretching of the epithelium on the surface of the egg chamber (arrowheads in Fig. 3A) as well as to the growth of the
Fig. 3. Border cell migration defects in cortactin mutants. (A) A wild type stage 9 egg chamber. Migration of the border cell cluster (bracket) has proceeded as far as the progression of cells sliding on the outside of the egg chamber (arrowheads). (B) A cort M7/cort M7 mutant stage 9 egg chamber showing delays in border cell migration. (C) Quantification of migration in cort M7/cort M7 mutant females and control (n . 367 for stage 10 and n . 245 for stage 9). Delay means border cells more than one nurse cell diameter more anterior than the expected position (as in B). (D) Quantification of migration in cort M7/cort M7 mutant clones at stage 9 (n ¼ 26 for germ line clones, n ¼ 52 for border cells clones). The control is heterozygous (cort/þ ) egg chambers from the same sample. Scale bars indicate 20 mm.
60
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
oocyte. In cortactin homozygous mutant females, border cell migration was delayed (Fig. 3B and quantification in Fig. 3C). However, migration did not arrest completely, as the clusters managed to reach their target, the oocyte, at stage 10 (Fig. 3C). To determine whether the delay was due to a defect in the border cells themselves or in the migration substrate, the germ line cells, we analyzed mosaic egg chambers (Fig. 3D). When follicle cells including border cells were mutant for cortactin, the delay was apparent. In contrast, if only germ line cells were mutant for cortactin, the migration was normal. Thus, cortactin is required cell autonomously for border cell migration to occur efficiently. To gain some insight into how Cortactin might affect migration of border cells, we analyzed how follicle cells respond to overexpression of Cortactin (Fig. 4). The follicle cells forming the simple monolayer epithelium that covers the oocyte are related to border cells, but normally express lower levels of Cortactin. Clones of follicle cells expressing high levels of Cortactin showed increased accumulation of F-actin (Fig. 4A). Interestingly, these cells also form short, Cortactin- and F-actin-rich protrusions or filopodia (Fig. 4B, compare to the control in Fig. 4C). Thus, Cortactin overexpression is sufficient to promote actin filament accumulation. It is also sufficient to induce formation of short cellular extensions in epithelial cells. These cellular
effects may be mechanistically related to the ability of mammalian Cortactin to promote formation of membrane protrusions and membrane ruffles or waves in cultured cells. There were several reasons to suspect that the effect of Cortactin on the actin cytoskeleton could be related to effects of RTKs. We have previously shown that border cell migration is guided by two RTKs, namely PDGF/VEGF receptor (PVR) and EGFR (Duchek and Rørth, 2001; Duchek et al., 2001). Also, an activated form of PVR induced robust formation of actin-rich extensions in follicle cells in a Rac dependent manner (Duchek et al., 2001). Finally, mammalian Cortactin has been suggested to act as a link between RTKs such as PDGF receptor and the actin cytoskeleton. We therefore decided to examine the effect of PVR signaling on Cortactin protein in the follicular epithelium. Overexpression of wild type PVR was sufficient to increase signaling in follicle cells slightly, resulting in a small increase in F-actin accumulation in the cell. This is most visible at the basal F-actin network (Fig. 5A0 ). PVR overexpression also resulted in clear recruitment and/or stabilization of Cortactin at the cell cortex (Fig. 5A00 ). Cortactin protein was not simply recruited by the increased amount of F-actin, as the subcellular localization of Cortactin was distinct from that of F-actin. In addition, the level and the localization of other actin-associated proteins
Fig. 4. Cortactin overexpression in follicle cells induces F-actin accumulation and filopodia. (A,B) A large clone of follicle cells overexpressing Cortactin adjacent to wild type cells, stained with anti-Cortactin (A,B) and phalloidin to label F-actin (A0 ,B0 ). (C) A control clone overexpressing GFP. All images are from stage 10 egg chambers, optically sectioned in the plane of the follicular epithelium to show the most basal aspect of the cells. Note the prominent filopodia visualized by anti-Cortactin (and phalloidin) in B. Scale bar indicates 10 mm.
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
61
Fig. 5. Overexpression of PVR or activated PVR (l-PVR) leads to Cortactin protein accumulation. (A) Overexpression of wild type PVR leads to a small increase in F-actin in the cell (A0 ) and to Cortactin protein accumulation at the cell cortex (A00 ). (B,C) Expression of constitutive active form of PVR (l-PVR) or of Src (Src42CA) leads to increased accumulation of Cortactin protein in the cells as well as disruption of the normal cell shape. (D) Overexpression of PVR does not affect accumulation of another actin-binding protein, moesin. Egg chambers from females of the genotype slboGal4/UAS-PVR (A) or slboGal4/UASl-PVR (B) or slboGal4/UAS-Src42 CA (C), stained with anti-PVR (A,B), phalloidin (A0 ,B0 ,C0 ) and anti-Cortactin (A00 ,B00 ,C00 ). (D) Egg chambers from females of the genotype slboGal4/UAS-PVR, stained with anti-PVR (D), phalloidin (D0 ) and anti-moesin (D00 ). Overlay of all three is shown to the right. Images are from stage 10 egg chambers, optically sectioned in the plane of the follicular epithelium. At this stage, slboGal4 drives expression in centripetal cells and more sporadically in the adjacent main body follicle cells. The pattern of PVR and l-PVR overexpression is visualized directly by the anti-PVR antibody (A and B); the pattern of Src42 overexpression is expected to be the same. The arrowheads in A and D point to one of the main body follicle cells with high PVR expression. Scale bar indicates 10 mm.
such as moesin (Fig. 5D) and a-spectrin (data not shown) were not visibly affected by PVR overexpression. Expression of a constitutive active form of PVR (l-PVR) resulted in more robust F-actin accumulation but also disruption of the normal cell shape (Fig. 5B0 ). The activated receptor was not restricted to the cell cortex but was present in vesicles throughout the cell (Fig. 5B). The constitutive active PVR also induced accumulation of Cortactin throughout the cell (Fig. 5B00 ). Thus, PVR activation in follicle cells affects the accumulation and subcellular localization of Cortactin protein, primarily resulting in more Cortactin at the cell cortex of normal epithelial cells. The ability of Cortactin to induce F-actin accumulation and filopodia formation in conjunction with the effect of PVR on Cortactin protein suggested that Cortactin might act
downstream of PVR with respect to control of the actin cytoskeleton. To determine if this might be the case, we performed an epistasis experiment. The effect of activated PVR (l-PVR) on F-actin accumulation and cell shape in follicle cells (see Fig. 5B) was scored using three categories of severity. Quantification was done by blindly scoring severity of the phenotype in many egg chambers. In each experiment we compared follicle cells which are mutant for cortactin to a control, wild type background. We saw a small, but statistically significant decrease in the severity of the l-PVR induced phenotypes in the cortactin mutant background (Fig. 6A). This result is consistent with Cortactin acting downstream of PVR, but also shows that the effect of PVR on the actin cytoskeleton does not strictly require Cortactin. Another factor that appears to act
62
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
Fig. 6. The F-actin accumulation induced by activated PVR (l-PVR) and activated Src is reduced in cortactin mutant background. Quantification of F-actin accumulation and cell shape changes according to strength of phenotype. (A) Comparison of slboGal4/UAS-l-PVR (control) and slboGal4/UAS-l-PVR; cort 2/cort 2 (cort 2/cort 2); n . 286 for each. (B) Comparison of slboGal4/UAS-Src42 CA (control) and slboGal4/UAS-Src42 CA; cort 2/cort 2 (cort 2/cort 2); n . 421 for each. The difference between control and cortactin mutants is statistically significant according to a x2 test ðP , 0:001Þ: All samples were scored blindly.
downstream of PVR is the Rac activator Mbc (related to DOCK180 and Ced-5). In the same assay, removal of mbc has a much more pronounced effect (Duchek et al., 2001). Activation of Rac can cause translocation of Cortactin to the cell periphery in mammalian cells (Weed et al., 1998). Thus, cortical Cortactin accumulation could be one of the downstream effects of Rac activation by PVR. In conclusion, Cortactin appears to contribute to the effects of PVR on actin, but is largely redundant with other factors. A contributing, but not essential, function is also consistent with cortactin not being an essential gene. Expression of an activated form of Src (Src42CA, Tateno et al., 2000) in border cells and other follicle cells showed a phenotype similar to that of activated PVR. It completely blocked border cell migration and disrupted cell shape and the actin cytoskeleton of follicle cells (Fig. 5C). Src64 has a specific function in the female germ line, which coincides with a function of Cortactin as discussed above. Src42 appears to function more generally in somatic cells and is required for viability. Unfortunately, it is technically not feasible to make Src42 mutant clones to determine whether Src42 function is required in border cells. However, we could determine whether Cortactin might act downstream of activated Src42 by performing an epistasis experiment equivalent to that with activated PVR. Removal of cortactin resulted in a small, but significant, decrease in the severity of the activated Src-induced phenotype (Fig. 6B). This is consistent with a role of Cortactin downstream of Src. It also shows that Src can affect the cytoskeleton independently of Cortactin.
3. Discussion In this report, we present the first genetic analysis of the conserved actin regulator cortactin. Drosophila cortactin contributes to the regulation of cell migration and other processes during oogenesis, but is not essential for these processes and not required for viability of the animal. Cortactin was originally isolated as a substrate of the Src kinase. Further biochemical characterization of mammalian
Cortactin indicated that it could directly affect actin polymerization by stimulating activity of the Arp2/3 complex in addition to stabilizing the F-actin network. Analysis of mammalian cells has also indicated that phosphorylation by Src is critical for Cortactin activity in affecting cell behavior such as cell migration. In fact, a nonphosphorylatable form can act as a dominant negative protein (Huang et al., 1998). Phosphorylation by Src can also affect the ability of Cortactin to crosslink F-actin (Huang et al., 1997), but phosphorylation may not directly affect Arp2/3 interaction (Weaver et al., 2001; Uruno et al., 2001). Genetic analyses in Drosophila indicate that Src, cortactin and Arp2/3 components control at least some of the same processes in vivo, rather than Src and Arp2/3 being associated with different aspects of Cortactin function. For example, in the germ line of the ovary, mutations in Src64, cortactin or components of the Arp2/3 complex all affect the actin-rich structures called ring canals in a similar way (Dodson et al., 1998; this study; Hudson and Cooley, 2002). This supports the notion that Src, Cortactin and Arp2/3 act together in vivo. However, the interactions may be quite indirect. Perhaps Src induced phosphorylation of Cortactin influence its localization or stability in vivo, which in turn affects the ability of Cortactin to stimulate Arp2/3. The effects of Cortactin in border cells and other follicle cells allowed us to investigate its relationship to a specific RTK, namely PVR. This was of some interest as Cortactin has been implicated both in stimulation of actin polymerization per se, and in endocytosis. The two functions could reflect the same actin-regulatory biochemical function of Cortactin. PVR signaling stimulates actin polymerization in follicle cells and down-regulation of PVR via endocytosis might limit this effect. Our genetic analysis indicates that cortactin plays a positive role downstream of PVR in stimulating actin polymerization, rather than a negative role by limiting receptor signaling. We cannot rule out that Cortactin also contributes in a minor way to stimulating receptor endocytosis, in these cells or in other cells. However, genes encoding other proteins that directly control endocytosis such as a-adaptin or Hrs are essential in Drosophila (Gonzalez-Gaitan and Jackle, 1997; Lloyd
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
et al., 2002). Some mutants have phenotypes that reflect upregulation of specific signaling pathways indicating that endocytosis is important for limiting activity of these signaling pathways in vivo (Berdnik et al., 2002; Lloyd et al., 2002; Seto et al., 2002). One of the useful features of Drosophila as a model system is that distinctive defects are produced upon mis-regulation of different signaling pathways. However, no such visible phenotypes were seen in the cortactin mutants. It is perhaps surprising that cortactin in not an essential gene. Many modulators of essential processes in the cell such as dynamics of the cytoskeleton, cell adhesion or cell signaling are very well conserved in higher eukaryotes. They may add to the robustness and fidelity of the regulation, but they may only be essential to the organism if their absence completely changes the behavior or fate of specific, important cells. In mammals, there are often multiple closely related genes and simple redundancy between these gene products may explain an absence of phenotypes in knockout mice. In Drosophila, this type of simple redundancy is less frequent. For example, there is no evidence for another cortactin gene in the sequenced Drosophila genome. However, more distantly related genes may have overlapping functions. Subtle phenotypes may also reflect that one process can be regulated in multiple ways. Combining multiple mutations can then be used to genetically help define which genes and pathways overlap in function.
63
ubiquitin-GFP to heat shock. Clones of follicle cells overexpressing Cortactin were obtained by ‘flipout’ Gal4, subjecting females of the genotype hs-FLP/Actin5C-FRTstop-FRT-Gal4; UAS-cortactin/þ (control: UAS-GFP) to heat shock and dissecting 3 days later. Quantification of phenotypes due to l-PVR and SrcCA overexpression (Fig. 6) was done by judging the increase in F-actin accumulation as well as the associated changes in cell shape and assigning each egg chamber to a category (as done previously, see Fig. 6 in Duchek et al., 2001). The samples were scored blindly. For antibody staining, ovaries were fixed with 4% paraformaldehyde and subsequent incubations were in PBS þ 0.1% triton X-100, with 5% normal goat serum as blocking reagent. Primary antibodies were anti-Cortactin (Katsube et al., 1998), anti-PVR (Duchek et al., 2001) and anti-moesin (Edwards et al., 1997). Fluorescent secondary antibodies (Jackson ImmunoResearch) were used together with rhodamin – phalloidin (Molecular Probes). All images were captured using confocal microscopy (Leica) and single optical sections are shown.
Acknowledgements We are grateful to Manabu Takahisa for the Cortactin antibody and to Daniel Kiehart for the antibody to moesin. We also thank Takashi Adachi-Yamada for UAS-Src42CA flies and to Ann Mari Voie for embryo injections.
4. Experimental procedures Standard fly-rearing conditions were used and transgenic flies were generated by standard injection technique. The tub-cortactin rescue construct was made by cloning the full length cortactin cDNA from LD29964 into pCasper4 containing the a-tubulin promoter. EPg35301 (cortactin) was identified in a gain-offunction screen for slbo suppressors, as described in Rørth et al. (1998), using the EPg element described in Mata et al. (2000). EPg35301 is inserted 622 base pairs upstream of the cortactin transcription start site, oriented such that it will drive expression of cortactin. One hundred and seventy-four excisions were generated by exposing flies to D2 – 3 transposase source, generating stocks from individual offspring and analyzing genomic DNA by PCR. For deletions in the cortactin locus, the PCR products were sequenced for precise mapping. M7 removes from 51 base pairs upstream of the EP to 15 base pairs downstream of the cortactin gene (total 3474 base pairs). D4 and A4 remove 2070 base pairs and 1522 base pairs downstream of the EP, deleting the gene to the middle of the fourth and the third exon, respectively. K8 goes from 97 base pairs upstream of the EP to the end of the first exon (total 885 base pairs). Mitotic clones of cortactin were induced by exposing larvae of the genotype hs-FLP/þ ; FRT82, cort/FRT82,
References Berdnik, D., To¨ro¨k, T., Gonzalez-Gaitan, M., Knoblich, J.A., 2002. The endocytic protein alpha-Adaptin is required for numb-mediated asymmetric cell division in Drosophila. Dev. Cell 3, 221 –231. Bowden, E.T., Barth, M., Thomas, D., Glazer, R.I., Mueller, S.C., 1999. An invasion-related complex of cortactin, paxillin and PKCmu associates with invadopodia at sites of extracellular matrix degradation. Oncogene 18, 4440–4449. Cao, H., Orth, J.D., Chen, J., Weller, S.G., Heuser, J.E., McNiven, M.A., 2003. Cortactin is a component of clathrin-coated pits and participates in receptor-mediated endocytosis. Mol. Cell Biol. 23, 2162–2170. Dodson, G.S., Guarnieri, D.J., Simon, M.A., 1998. Src64 is required for ovarian ring canal morphogenesis during Drosophila oogenesis. Development 125, 2883–2892. Duchek, P., Rørth, P., 2001. Guidance of cell migration by EGF receptor signaling during Drosophila oogenesis. Science 291, 131– 133. Duchek, P., Somogyi, K., Je´kely, G., Beccari, S., Rørth, P., 2001. Guidance of cell migration by the Drosophila PDGF/VEGF receptor. Cell 107, 17– 26. Eden, S., Rohatgi, R., Podtelejnikov, A.V., Mann, M., Kirschner, M.W., 2002. Mechanism of regulation of WAVE1-induced actin nucleation by Rac1 and Nck. Nature 418, 790–793. Edwards, K.A., Demsky, M., Montague, R.A., Weymouth, N., Kiehart, D.P., 1997. GFP-moesin illuminates actin cytoskeleton dynamics in living tissue and demonstrates cell shape changes during morphogenesis in Drosophila. Dev. Biol. 191, 103–117.
64
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
Gonzalez-Gaitan, M., Jackle, H., 1997. Role of Drosophila alpha-adaptin in presynaptic vesicle recycling. Cell 88, 767 –776. Higgs, H.N., Pollard, T.D., 2001. Regulation of actin filament network formation through ARP2/3 complex: activation by a diverse array of proteins. Annu. Rev. Biochem. 70, 649–676. Huang, C., Ni, Y., Wang, T., Gao, Y., Haudenschild, C.C., Zhan, X., 1997. Down-regulation of the filamentous actin cross-linking activity of cortactin by Src-mediated tyrosine phosphorylation. J. Biol. Chem. 272, 13911–13915. Huang, C., Liu, J., Haudenschild, C.C., Zhan, X., 1998. The role of tyrosine phosphorylation of cortactin in the locomotion of endothelial cells. J. Biol. Chem. 273, 25770–25776. Hudson, A.M., Cooley, L., 2002. A subset of dynamic actin rearrangements in Drosophila requires the Arp2/3 complex. J. Cell Biol. 156, 677 –687. Katsube, T., Takahisa, M., Ueda, R., Hashimoto, N., Kobayashi, M., Togashi, S., 1998. Cortactin associates with the cell–cell junction protein ZO-1 in both Drosophila and mouse. J. Biol. Chem. 273, 29672–29677. Kinley, A.W., Weed, S.A., Weaver, A.M., Karginov, A.V., Bissonette, E., Cooper, J.A., Parsons, J.T., 2003. Cortactin interacts with WIP in regulating Arp2/3 activation and membrane protrusion. Curr. Biol. 13, 384– 393. Li, Y., Tondravi, M., Liu, J., Smith, E., Haudenschild, C.C., Kaczmarek, M., Zhan, X., 2001. Cortactin potentiates bone metastasis of breast cancer cells. Cancer Res. 61, 6906–6911. Lloyd, T.E., Atkinson, R., Wu, M.N., Zhou, Y., Pennetta, G., Bellen, H.J., 2002. Hrs regulates endosome membrane invagination and tyrosine kinase receptor signaling in Drosophila. Cell 108, 261 –269. Maa, M.C., Wilson, L.K., Moyers, J.S., Vines, R.R., Parsons, J.T., Parsons, S.J., 1992. Identification and characterization of a cytoskeletonassociated, epidermal growth factor sensitive pp60c-src substrate. Oncogene 7, 2429– 2438. Mata, J., Curado, S., Ephrussi, A., Rørth, P., 2000. Tribbles coordinates mitosis and morphogenesis in Drosophila by regulating string/CDC25 proteolysis. Cell 101, 511 –522. McNiven, M.A., Kim, L., Krueger, E.W., Orth, J.D., Cao, H., Wong, T.W., 2000. Regulated interactions between dynamin and the actin-binding protein cortactin modulate cell shape. J. Cell Biol. 151, 187 –198. Moreau, V., Galan, J.M., Devilliers, G., Haguenauer-Tsapis, R., Winsor, B., 1997. The yeast actin-related protein Arp2p is required for the internalization step of endocytosis. Mol. Biol. Cell, 1361–1375. Patel, A.S., Schechter, G.L., Wasilenko, W.J., Somers, K.D., 1998. Overexpression of EMS1/cortactin in NIH3T3 fibroblasts causes
increased cell motility and invasion in vitro. Oncogene 16, 3227–3232. Rørth, P., Szabo, K., Bailey, A., Laverty, T., Rehm, J., Rubin, G.M., et al., 1998. Systematic gain-of-function genetics in Drosophila. Development 125, 1049–1057. Schuuring, E., Verhoeven, E., Litvinov, S., Michalides, R.J., 1993. The product of the EMS1 gene, amplified and overexpressed in human carcinomas, is homologous to a v-src substrate and is located in cellsubstratum contact sites. Mol. Cell Biol. 13, 2891–2898. Seto, E.S., Bellen, H.J., Lloyd, T.E., 2002. When cell biology meets development: endocytic regulation of signaling pathways. Genes Dev. 16, 1314–1336. Takahashi, F., Endo, S., Kojima, T., Saigo, K., 1996. Regulation of cell – cell contacts in developing Drosophila eyes by Dsrc41, a new, close relative of vertebrate c-src. Genes Dev. 10, 1645– 1656. Tanentzapf, G., Smith, C., McGlade, J., Tepass, U., 2000. Apical, lateral, and basal polarization cues contribute to the development of the follicular epithelium during Drosophila oogenesis. J. Cell Biol. 151, 891 –904. Tateno, M., Nishida, Y., Adachi-Yamada, T., 2000. Regulation of JNK by Src during Drosophila development. Science 287, 324–327. Uruno, T., Liu, J., Zhang, P., Fan, Y., Egile, C., Li, R., et al., 2001. Activation of Arp2/3 complex-mediated actin polymerization by cortactin. Nat. Cell Biol. 3, 259–266. Weaver, A.M., Karginov, A.V., Kinley, A.W., Weed, S.A., Li, Y., Parsons, J.T., Cooper, J.A., 2001. Cortactin promotes and stabilizes Arp2/3induced actin filament network formation. Curr. Biol. 11, 370–374. Weaver, A.M., Heuser, J.E., Karginov, A.V., Lee, W.L., Parsons, J.T., Cooper, J.A., 2002. Interaction of cortactin and N-WASp with Arp2/3 complex. Curr. Biol., 1270–1278. Weed, S.A., Du, Y., Parsons, J.T., 1998. Translocation of cortactin to the cell periphery is mediated by the small GTPase Rac1. J. Cell Sci. 111, 2433–2443. Wu, H., Parsons, J.T., 1993. Cortactin, an 80/85-kilodalton pp60src substrate, is a filamentous actin-binding protein enriched in the cell cortex. J. Cell Biol. 120, 1417–1426. Wu, H., Reynolds, A.B., Kanner, S.B., Vines, R.R., Parsons, J.T., 1991. Identification and characterization of a novel cytoskeleton-associated pp60src substrate. Mol. Cell Biol. 11, 5113–5124. Zhan, X., Hu, X., Hampton, B., Burgess, W.H., Friesel, R., Maciag, T., 1993. Murine cortactin is phosphorylated in response to fibroblast growth factor-1 on tyrosine residues late in the G1 phase of the BALB/c 3T3 cell cycle. J. Biol. Chem. 268, 24427–24431.
Cortactin modulates cell migration and ring canal morphogenesis during Drosophila oogenesis Ka´lma´n Somogyi, Pernille Rørth* European Molecular Biology Laboratory, Developmental Biology Programme, Meyerhofstrasse 1, 69117 Heidelberg, Germany Received 13 August 2003; received in revised form 20 October 2003; accepted 20 October 2003
Abstract Cortactin is a Src substrate that interacts with F-actin and can stimulate actin polymerization by direct interaction with the Arp2/3 complex. We have isolated complete loss-of-function mutants of the single Drosophila cortactin gene. Mutants are viable and fertile, showing that cortactin is not an essential gene. However, cortactin mutants show distinct defects during oogenesis. During oogenesis, Cortactin protein is enriched at the F-actin rich ring canals in the germ line, and in migrating border cells. In cortactin mutants, the ring canals are smaller than normal. A similar phenotype has been observed in Src 64 mutants and in mutants for genes encoding Arp2/3 complex components, supporting that these protein products act together to control specific processes in vivo. Cortactin mutants also show impaired border cell migration. This invasive cell migration is guided by Drosophila EGFR and PDGF/VEGF receptor (PVR). We find that accumulation of Cortactin protein is positively regulated by PVR. Also, overexpression of Cortactin can by itself induce F-actin accumulation and ectopic filopodia formation in epithelial cells. We present evidence that Cortactin is one of the factors acting downstream of PVR and Src to stimulate F-actin accumulation. Cortactin is a minor contributor in this regulation, consistent with the cortactin gene not being essential for development. q 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: EMS1; Cytoskeleton; Motility
1. Introduction Cortactin was originally identified as a major substrate of the Src tyrosine kinase (Wu et al., 1991). The name Cortactin reflects that the protein binds to F-actin and that it localizes to the cell cortex, including membrane ruffles and lamellipodia (Wu and Parsons, 1993). These features, plus the presence of an SH3 domain and proline-rich regions in the Cortactin protein, suggested that Cortactin might link signaling events to the actin cytoskeleton. Phosphorylation of Cortactin stimulated by Src modulates its activity in vivo (Huang et al., 1997). Cortactin phosphorylation and subcellular localization are also affected by receptor tyrosine kinases (RTKs) (Maa et al., 1992; Zhan et al., 1993). The cortactin gene was also identified as EMS1, a putative oncogene encoding one of the transcripts amplified in certain human carcinomas (Schuuring et al., 1993). Directed overexpression of EMS1/ * Corresponding author. Tel.: þ 49-6221387109; fax: þ 49-6221387166. E-mail address: [email protected] (P. Rørth). 0925-4773/$ - see front matter q 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.mod.2003.10.003
cortactin was subsequently shown to increase the motility of and invasion of fibroblasts (Patel et al., 1998) and the metastatic potential of breast cancer cells (Li et al., 2001). Cortactin also enriched in ‘invadopodia’ from invasive tumor cells, cellular protrusions associated with degradation of extracellular matrix (Bowden et al., 1999). Together, these studies suggest that Cortactin may play a role in promoting cell motility of cancer cells, as well as of normal cells in response to growth factor stimulation. Recent studies have shown that Cortactin can directly influence actin polymerization. The Arp2/3 complex is an important nucleator of actin polymerization. It stimulates actin polymerization while binding to actin filaments and can thereby induce formation of a branched actin network (Higgs and Pollard, 2001). The Arp2/3 complex consists of multiple proteins and is very conserved through evolution, from yeast to man (Higgs and Pollard, 2001). Cortactin can bind to and activate the Arp2/3 complex as well as stabilize the resulting F-actin network (Uruno et al., 2001; Weaver et al., 2001). Cortactin may act synergistically with proteins of the WASP family, which are potent regulators of Arp2/3
58
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
activity (Weaver et al., 2002). In addition to WASP and N-WASP, the WASP family includes the SCAR/WAVE proteins. The WASP family of proteins is regulated by GTPases of the Rho subfamily, namely CDC42 and Rac: WASP by direct binding (Higgs and Pollard, 2001), SCAR/WAVE by dissociation of an inhibitory complex (Eden et al., 2002). Cortactin can also associate with the adaptor protein WIP (WASP interacting protein), which binds monomeric actin and thus may aid Cortactin in actin nucleation (Kinley et al., 2003). Arp2/3 and its regulators are required for extensions of lamellipodia and other actinrich structures, but also for intracellular trafficking events such as endocytosis (Moreau et al., 1997). Cortactin has been suggested to link endocytosis to Arp2/3 mediated actin polymerization in mammalian cells. Cortactin is associated with dynamin2 and with vesicles (McNiven et al., 2000). Consistent with this type of function, inhibition of Cortactin inhibits receptor mediated endocytosis (Cao et al., 2003). A Cortactin protein has been identified in Drosophila (Katsube et al., 1998), but not in yeast. It is possible that other proteins substitute for Cortactin in yeast. It is also possible that the function of Cortactin is specific to higher eukaryotes that, contrary to yeast, use tyrosine kinases for cell –cell communication and cell regulation. There are multiple RTKs in Drosophila, which control cell growth, differentiation and morphogenesis including EGFR, insulin receptor and PDGF/VEGF receptor. The Drosophila genome also encodes a number of non-receptor tyrosine kinases such as Src. There are two Src genes, Src42 and Src64. Src42 may act downstream of RTKs to modify their signaling (Takahashi et al., 1996). Src64 is specifically required for proper morphogenesis of actin-rich structures in the female germ line, called ring canals (Dodson et al., 1998). Drosophila Cortactin protein is localized to the cell cortex and shows protein– protein interactions similar to that of mammalian Cortactin (Katsube et al., 1998). To understand the function of Cortactin in vivo, we have generated complete loss-of-function mutants of Drosophila cortactin. Cortactin mutants are fully viable and are fertile, but have subtle defects during oogenesis. Our analysis suggests that Cortactin acts downstream of RTKs and Src in vivo. The analysis also indicates that Cortactin is only a minor mediator of the effects of tyrosine kinases on the actin cytoskeleton.
2. Results Mutations in the cortactin locus were generated by imprecise excision of a P-element located upstream of the transcription start site (Fig. 1A, see also Section 4). One of the four deletion alleles precisely removes the cortactin transcription unit (cort M7) and two others most of the coding region (cort D4 and cort A4). All four alleles showed the same phenotypes, but the M7 and D4 alleles were used for further analysis. The cortactin mutant flies are
Fig. 1. Cortactin mutants and protein distribution in egg chambers. (A) The cortactin locus and the four cortactin mutants. The genomic region removed in each of the deletion mutants is indicated below (see Section 4 for details). (B,C) Anti-Cortactin staining of stage 9 egg chambers from wild type (B) and cort M7/cort M7 mutant females (C). The mutant encodes no Cortactin protein, so the staining in C represents background staining. The cortical staining as well as staining of border cells (arrow) and ring canals (arrowheads) is specific. In these and all other micrographs of egg chambers, anterior is to the left. Border cells migrate from the anterior of the egg chamber to the oocyte. (D) Higher magnification view of follicle cells stained with anti-Cortactin showing staining along the cell cortex. Scale bars indicate 20 mm in B and C, 10 mm in D.
homozygous viable and show no visible abnormalities. The stock becomes completely homozygous without selection, showing that there is no requirement for maternal contribution of cortactin, nor is there a large growth disadvantage due to loss of cortactin. However, we noticed that the females had somewhat reduced fertility and therefore examined oogenesis in more detail. During oogenesis, development proceeds in egg chambers consisting of centrally located germ line cells (15 ‘nurse cells’ and one oocyte) surrounded by a simple monolayer epithelium, the follicular epithelium. Fig. 1B shows a wild type egg chamber at stage 9 stained with a Cortactin antibody. The control for specificity of the antibody is shown in Fig. 1C, an egg chamber of the same stage from a cortactin deletion mutant. At this stage, a small cluster of follicle cells called border cells, delaminate from the follicular epithelium, invade the germ line cluster
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
and migrate to the oocyte. Cortactin was specifically enriched in migrating border cells (indicated by an arrow in Fig. 1B,C). Cortactin was also enriched at the actin rich structures called ring canals (arrowheads in Fig. 1B). Ring canals are formed early in oogenesis in place of the cleavage furrow after incomplete cytokinesis of germ line cells. Proper function of ring canals is necessary for transfer of material from nurse cells to the oocyte. Finally, Cortactin protein was detected along the cortex of the follicle cells (Fig. 1D). A number of specific defects were observed in cortactin mutants. These phenotypes were confirmed to be due to loss of cortactin, as they could be rescued by a transgene driving ubiquitous expression of cortactin cDNA under control of a tubulin promoter. One cortactin phenotype was a mild defect in ‘dumping’, transfer of bulk cytoplasmic material from nurse cells to the oocyte (Fig. 2A,B). This phenotype is similar to that observed in Src64 mutants (Dodson et al., 1998), but is also seen in other mutants. In the case of Src64, the defect was correlated to the presence of Src64 protein at the ring canals and to reduced size of the ring canals. As shown in Fig. 1B, Cortactin is also present at ring canals. To determine whether cortactin affected ring canal morphogenesis, we quantified their size at stage 10. Ring canal size was significantly reduced in the cortactin mutant
Fig. 2. Phenotypes of cortactin mutants in the ovary. (A) A stage 11 wild type egg chamber. (B) The mild ‘dumping defect’: a ‘stage 11’ mutant egg chamber, staged by overall size and F-actin network accumulation in the nurse cells. The relative size of the oocyte is too small indicating a defect in transfer of material to the oocyte. (C) Ring canals of stage 10 wild type and mutant egg chambers and size quantification with standard error indicated. The outer diameter of ring canals closest to the oocyte was measured in multiple egg chambers ðn . 87Þ: (D) Quantification of egg size from wild type females as well as from cort M7/cort M7 and from Src64 D17/Src64 D17 mutant females (n . 107). (E) Discontinuities in the follicular epithelium observed in some cortactin mutant egg chambers. Normally, the epithelium covers all of the germ cells. Between the arrowheads, follicle cells are missing. Scale bars indicate 20 mm in A, B and E, 10 mm in C.
59
relative to wild type (Fig. 2C). Dumping defects usually result in smaller eggs. Eggs from cortactin mutant mothers and from Src64 mutant mothers were on average smaller than wild type (Fig. 2D). The hatching rate of the eggs from mutant mothers was also decreased. Thus, mutations in cortactin affect the actin rich ring canals, and processes dependent on function of ring canals, in a manner similar to mutations in Src64. Mutations in components of the Arp2/3 complex (Arpc1 or Arp3 mutants) also specifically affect ring canal morphogenesis (Hudson and Cooley, 2002). With low penetrance (around 6% of stages 5 –9 egg chambers), cortactin mutants also display discontinuities in the follicular epithelium (Fig. 2E), which is not observed in the wild type situation. A similar phenotype has been observed in other mutants such as crumbs (Tanentzapf et al., 2000), where the affected protein is thought to have a basic function in establishing the epithelium. The phenotype suggests that Cortactin has a role in epithelial integrity. The low penetrance indicates that this function mostly can be compensated by other proteins. Border cell migration was also abnormal in cortactin mutants (Fig. 3). As cortactin has been proposed to be important for migration of normal and invasive mammalian cells, this defect seemed particularly interesting. To quantify border cell migration, the progression of the cluster (bracket in Fig. 3A) was compared to the stretching of the epithelium on the surface of the egg chamber (arrowheads in Fig. 3A) as well as to the growth of the
Fig. 3. Border cell migration defects in cortactin mutants. (A) A wild type stage 9 egg chamber. Migration of the border cell cluster (bracket) has proceeded as far as the progression of cells sliding on the outside of the egg chamber (arrowheads). (B) A cort M7/cort M7 mutant stage 9 egg chamber showing delays in border cell migration. (C) Quantification of migration in cort M7/cort M7 mutant females and control (n . 367 for stage 10 and n . 245 for stage 9). Delay means border cells more than one nurse cell diameter more anterior than the expected position (as in B). (D) Quantification of migration in cort M7/cort M7 mutant clones at stage 9 (n ¼ 26 for germ line clones, n ¼ 52 for border cells clones). The control is heterozygous (cort/þ ) egg chambers from the same sample. Scale bars indicate 20 mm.
60
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
oocyte. In cortactin homozygous mutant females, border cell migration was delayed (Fig. 3B and quantification in Fig. 3C). However, migration did not arrest completely, as the clusters managed to reach their target, the oocyte, at stage 10 (Fig. 3C). To determine whether the delay was due to a defect in the border cells themselves or in the migration substrate, the germ line cells, we analyzed mosaic egg chambers (Fig. 3D). When follicle cells including border cells were mutant for cortactin, the delay was apparent. In contrast, if only germ line cells were mutant for cortactin, the migration was normal. Thus, cortactin is required cell autonomously for border cell migration to occur efficiently. To gain some insight into how Cortactin might affect migration of border cells, we analyzed how follicle cells respond to overexpression of Cortactin (Fig. 4). The follicle cells forming the simple monolayer epithelium that covers the oocyte are related to border cells, but normally express lower levels of Cortactin. Clones of follicle cells expressing high levels of Cortactin showed increased accumulation of F-actin (Fig. 4A). Interestingly, these cells also form short, Cortactin- and F-actin-rich protrusions or filopodia (Fig. 4B, compare to the control in Fig. 4C). Thus, Cortactin overexpression is sufficient to promote actin filament accumulation. It is also sufficient to induce formation of short cellular extensions in epithelial cells. These cellular
effects may be mechanistically related to the ability of mammalian Cortactin to promote formation of membrane protrusions and membrane ruffles or waves in cultured cells. There were several reasons to suspect that the effect of Cortactin on the actin cytoskeleton could be related to effects of RTKs. We have previously shown that border cell migration is guided by two RTKs, namely PDGF/VEGF receptor (PVR) and EGFR (Duchek and Rørth, 2001; Duchek et al., 2001). Also, an activated form of PVR induced robust formation of actin-rich extensions in follicle cells in a Rac dependent manner (Duchek et al., 2001). Finally, mammalian Cortactin has been suggested to act as a link between RTKs such as PDGF receptor and the actin cytoskeleton. We therefore decided to examine the effect of PVR signaling on Cortactin protein in the follicular epithelium. Overexpression of wild type PVR was sufficient to increase signaling in follicle cells slightly, resulting in a small increase in F-actin accumulation in the cell. This is most visible at the basal F-actin network (Fig. 5A0 ). PVR overexpression also resulted in clear recruitment and/or stabilization of Cortactin at the cell cortex (Fig. 5A00 ). Cortactin protein was not simply recruited by the increased amount of F-actin, as the subcellular localization of Cortactin was distinct from that of F-actin. In addition, the level and the localization of other actin-associated proteins
Fig. 4. Cortactin overexpression in follicle cells induces F-actin accumulation and filopodia. (A,B) A large clone of follicle cells overexpressing Cortactin adjacent to wild type cells, stained with anti-Cortactin (A,B) and phalloidin to label F-actin (A0 ,B0 ). (C) A control clone overexpressing GFP. All images are from stage 10 egg chambers, optically sectioned in the plane of the follicular epithelium to show the most basal aspect of the cells. Note the prominent filopodia visualized by anti-Cortactin (and phalloidin) in B. Scale bar indicates 10 mm.
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
61
Fig. 5. Overexpression of PVR or activated PVR (l-PVR) leads to Cortactin protein accumulation. (A) Overexpression of wild type PVR leads to a small increase in F-actin in the cell (A0 ) and to Cortactin protein accumulation at the cell cortex (A00 ). (B,C) Expression of constitutive active form of PVR (l-PVR) or of Src (Src42CA) leads to increased accumulation of Cortactin protein in the cells as well as disruption of the normal cell shape. (D) Overexpression of PVR does not affect accumulation of another actin-binding protein, moesin. Egg chambers from females of the genotype slboGal4/UAS-PVR (A) or slboGal4/UASl-PVR (B) or slboGal4/UAS-Src42 CA (C), stained with anti-PVR (A,B), phalloidin (A0 ,B0 ,C0 ) and anti-Cortactin (A00 ,B00 ,C00 ). (D) Egg chambers from females of the genotype slboGal4/UAS-PVR, stained with anti-PVR (D), phalloidin (D0 ) and anti-moesin (D00 ). Overlay of all three is shown to the right. Images are from stage 10 egg chambers, optically sectioned in the plane of the follicular epithelium. At this stage, slboGal4 drives expression in centripetal cells and more sporadically in the adjacent main body follicle cells. The pattern of PVR and l-PVR overexpression is visualized directly by the anti-PVR antibody (A and B); the pattern of Src42 overexpression is expected to be the same. The arrowheads in A and D point to one of the main body follicle cells with high PVR expression. Scale bar indicates 10 mm.
such as moesin (Fig. 5D) and a-spectrin (data not shown) were not visibly affected by PVR overexpression. Expression of a constitutive active form of PVR (l-PVR) resulted in more robust F-actin accumulation but also disruption of the normal cell shape (Fig. 5B0 ). The activated receptor was not restricted to the cell cortex but was present in vesicles throughout the cell (Fig. 5B). The constitutive active PVR also induced accumulation of Cortactin throughout the cell (Fig. 5B00 ). Thus, PVR activation in follicle cells affects the accumulation and subcellular localization of Cortactin protein, primarily resulting in more Cortactin at the cell cortex of normal epithelial cells. The ability of Cortactin to induce F-actin accumulation and filopodia formation in conjunction with the effect of PVR on Cortactin protein suggested that Cortactin might act
downstream of PVR with respect to control of the actin cytoskeleton. To determine if this might be the case, we performed an epistasis experiment. The effect of activated PVR (l-PVR) on F-actin accumulation and cell shape in follicle cells (see Fig. 5B) was scored using three categories of severity. Quantification was done by blindly scoring severity of the phenotype in many egg chambers. In each experiment we compared follicle cells which are mutant for cortactin to a control, wild type background. We saw a small, but statistically significant decrease in the severity of the l-PVR induced phenotypes in the cortactin mutant background (Fig. 6A). This result is consistent with Cortactin acting downstream of PVR, but also shows that the effect of PVR on the actin cytoskeleton does not strictly require Cortactin. Another factor that appears to act
62
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
Fig. 6. The F-actin accumulation induced by activated PVR (l-PVR) and activated Src is reduced in cortactin mutant background. Quantification of F-actin accumulation and cell shape changes according to strength of phenotype. (A) Comparison of slboGal4/UAS-l-PVR (control) and slboGal4/UAS-l-PVR; cort 2/cort 2 (cort 2/cort 2); n . 286 for each. (B) Comparison of slboGal4/UAS-Src42 CA (control) and slboGal4/UAS-Src42 CA; cort 2/cort 2 (cort 2/cort 2); n . 421 for each. The difference between control and cortactin mutants is statistically significant according to a x2 test ðP , 0:001Þ: All samples were scored blindly.
downstream of PVR is the Rac activator Mbc (related to DOCK180 and Ced-5). In the same assay, removal of mbc has a much more pronounced effect (Duchek et al., 2001). Activation of Rac can cause translocation of Cortactin to the cell periphery in mammalian cells (Weed et al., 1998). Thus, cortical Cortactin accumulation could be one of the downstream effects of Rac activation by PVR. In conclusion, Cortactin appears to contribute to the effects of PVR on actin, but is largely redundant with other factors. A contributing, but not essential, function is also consistent with cortactin not being an essential gene. Expression of an activated form of Src (Src42CA, Tateno et al., 2000) in border cells and other follicle cells showed a phenotype similar to that of activated PVR. It completely blocked border cell migration and disrupted cell shape and the actin cytoskeleton of follicle cells (Fig. 5C). Src64 has a specific function in the female germ line, which coincides with a function of Cortactin as discussed above. Src42 appears to function more generally in somatic cells and is required for viability. Unfortunately, it is technically not feasible to make Src42 mutant clones to determine whether Src42 function is required in border cells. However, we could determine whether Cortactin might act downstream of activated Src42 by performing an epistasis experiment equivalent to that with activated PVR. Removal of cortactin resulted in a small, but significant, decrease in the severity of the activated Src-induced phenotype (Fig. 6B). This is consistent with a role of Cortactin downstream of Src. It also shows that Src can affect the cytoskeleton independently of Cortactin.
3. Discussion In this report, we present the first genetic analysis of the conserved actin regulator cortactin. Drosophila cortactin contributes to the regulation of cell migration and other processes during oogenesis, but is not essential for these processes and not required for viability of the animal. Cortactin was originally isolated as a substrate of the Src kinase. Further biochemical characterization of mammalian
Cortactin indicated that it could directly affect actin polymerization by stimulating activity of the Arp2/3 complex in addition to stabilizing the F-actin network. Analysis of mammalian cells has also indicated that phosphorylation by Src is critical for Cortactin activity in affecting cell behavior such as cell migration. In fact, a nonphosphorylatable form can act as a dominant negative protein (Huang et al., 1998). Phosphorylation by Src can also affect the ability of Cortactin to crosslink F-actin (Huang et al., 1997), but phosphorylation may not directly affect Arp2/3 interaction (Weaver et al., 2001; Uruno et al., 2001). Genetic analyses in Drosophila indicate that Src, cortactin and Arp2/3 components control at least some of the same processes in vivo, rather than Src and Arp2/3 being associated with different aspects of Cortactin function. For example, in the germ line of the ovary, mutations in Src64, cortactin or components of the Arp2/3 complex all affect the actin-rich structures called ring canals in a similar way (Dodson et al., 1998; this study; Hudson and Cooley, 2002). This supports the notion that Src, Cortactin and Arp2/3 act together in vivo. However, the interactions may be quite indirect. Perhaps Src induced phosphorylation of Cortactin influence its localization or stability in vivo, which in turn affects the ability of Cortactin to stimulate Arp2/3. The effects of Cortactin in border cells and other follicle cells allowed us to investigate its relationship to a specific RTK, namely PVR. This was of some interest as Cortactin has been implicated both in stimulation of actin polymerization per se, and in endocytosis. The two functions could reflect the same actin-regulatory biochemical function of Cortactin. PVR signaling stimulates actin polymerization in follicle cells and down-regulation of PVR via endocytosis might limit this effect. Our genetic analysis indicates that cortactin plays a positive role downstream of PVR in stimulating actin polymerization, rather than a negative role by limiting receptor signaling. We cannot rule out that Cortactin also contributes in a minor way to stimulating receptor endocytosis, in these cells or in other cells. However, genes encoding other proteins that directly control endocytosis such as a-adaptin or Hrs are essential in Drosophila (Gonzalez-Gaitan and Jackle, 1997; Lloyd
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
et al., 2002). Some mutants have phenotypes that reflect upregulation of specific signaling pathways indicating that endocytosis is important for limiting activity of these signaling pathways in vivo (Berdnik et al., 2002; Lloyd et al., 2002; Seto et al., 2002). One of the useful features of Drosophila as a model system is that distinctive defects are produced upon mis-regulation of different signaling pathways. However, no such visible phenotypes were seen in the cortactin mutants. It is perhaps surprising that cortactin in not an essential gene. Many modulators of essential processes in the cell such as dynamics of the cytoskeleton, cell adhesion or cell signaling are very well conserved in higher eukaryotes. They may add to the robustness and fidelity of the regulation, but they may only be essential to the organism if their absence completely changes the behavior or fate of specific, important cells. In mammals, there are often multiple closely related genes and simple redundancy between these gene products may explain an absence of phenotypes in knockout mice. In Drosophila, this type of simple redundancy is less frequent. For example, there is no evidence for another cortactin gene in the sequenced Drosophila genome. However, more distantly related genes may have overlapping functions. Subtle phenotypes may also reflect that one process can be regulated in multiple ways. Combining multiple mutations can then be used to genetically help define which genes and pathways overlap in function.
63
ubiquitin-GFP to heat shock. Clones of follicle cells overexpressing Cortactin were obtained by ‘flipout’ Gal4, subjecting females of the genotype hs-FLP/Actin5C-FRTstop-FRT-Gal4; UAS-cortactin/þ (control: UAS-GFP) to heat shock and dissecting 3 days later. Quantification of phenotypes due to l-PVR and SrcCA overexpression (Fig. 6) was done by judging the increase in F-actin accumulation as well as the associated changes in cell shape and assigning each egg chamber to a category (as done previously, see Fig. 6 in Duchek et al., 2001). The samples were scored blindly. For antibody staining, ovaries were fixed with 4% paraformaldehyde and subsequent incubations were in PBS þ 0.1% triton X-100, with 5% normal goat serum as blocking reagent. Primary antibodies were anti-Cortactin (Katsube et al., 1998), anti-PVR (Duchek et al., 2001) and anti-moesin (Edwards et al., 1997). Fluorescent secondary antibodies (Jackson ImmunoResearch) were used together with rhodamin – phalloidin (Molecular Probes). All images were captured using confocal microscopy (Leica) and single optical sections are shown.
Acknowledgements We are grateful to Manabu Takahisa for the Cortactin antibody and to Daniel Kiehart for the antibody to moesin. We also thank Takashi Adachi-Yamada for UAS-Src42CA flies and to Ann Mari Voie for embryo injections.
4. Experimental procedures Standard fly-rearing conditions were used and transgenic flies were generated by standard injection technique. The tub-cortactin rescue construct was made by cloning the full length cortactin cDNA from LD29964 into pCasper4 containing the a-tubulin promoter. EPg35301 (cortactin) was identified in a gain-offunction screen for slbo suppressors, as described in Rørth et al. (1998), using the EPg element described in Mata et al. (2000). EPg35301 is inserted 622 base pairs upstream of the cortactin transcription start site, oriented such that it will drive expression of cortactin. One hundred and seventy-four excisions were generated by exposing flies to D2 – 3 transposase source, generating stocks from individual offspring and analyzing genomic DNA by PCR. For deletions in the cortactin locus, the PCR products were sequenced for precise mapping. M7 removes from 51 base pairs upstream of the EP to 15 base pairs downstream of the cortactin gene (total 3474 base pairs). D4 and A4 remove 2070 base pairs and 1522 base pairs downstream of the EP, deleting the gene to the middle of the fourth and the third exon, respectively. K8 goes from 97 base pairs upstream of the EP to the end of the first exon (total 885 base pairs). Mitotic clones of cortactin were induced by exposing larvae of the genotype hs-FLP/þ ; FRT82, cort/FRT82,
References Berdnik, D., To¨ro¨k, T., Gonzalez-Gaitan, M., Knoblich, J.A., 2002. The endocytic protein alpha-Adaptin is required for numb-mediated asymmetric cell division in Drosophila. Dev. Cell 3, 221 –231. Bowden, E.T., Barth, M., Thomas, D., Glazer, R.I., Mueller, S.C., 1999. An invasion-related complex of cortactin, paxillin and PKCmu associates with invadopodia at sites of extracellular matrix degradation. Oncogene 18, 4440–4449. Cao, H., Orth, J.D., Chen, J., Weller, S.G., Heuser, J.E., McNiven, M.A., 2003. Cortactin is a component of clathrin-coated pits and participates in receptor-mediated endocytosis. Mol. Cell Biol. 23, 2162–2170. Dodson, G.S., Guarnieri, D.J., Simon, M.A., 1998. Src64 is required for ovarian ring canal morphogenesis during Drosophila oogenesis. Development 125, 2883–2892. Duchek, P., Rørth, P., 2001. Guidance of cell migration by EGF receptor signaling during Drosophila oogenesis. Science 291, 131– 133. Duchek, P., Somogyi, K., Je´kely, G., Beccari, S., Rørth, P., 2001. Guidance of cell migration by the Drosophila PDGF/VEGF receptor. Cell 107, 17– 26. Eden, S., Rohatgi, R., Podtelejnikov, A.V., Mann, M., Kirschner, M.W., 2002. Mechanism of regulation of WAVE1-induced actin nucleation by Rac1 and Nck. Nature 418, 790–793. Edwards, K.A., Demsky, M., Montague, R.A., Weymouth, N., Kiehart, D.P., 1997. GFP-moesin illuminates actin cytoskeleton dynamics in living tissue and demonstrates cell shape changes during morphogenesis in Drosophila. Dev. Biol. 191, 103–117.
64
K. Somogyi, P. Rørth / Mechanisms of Development 121 (2004) 57–64
Gonzalez-Gaitan, M., Jackle, H., 1997. Role of Drosophila alpha-adaptin in presynaptic vesicle recycling. Cell 88, 767 –776. Higgs, H.N., Pollard, T.D., 2001. Regulation of actin filament network formation through ARP2/3 complex: activation by a diverse array of proteins. Annu. Rev. Biochem. 70, 649–676. Huang, C., Ni, Y., Wang, T., Gao, Y., Haudenschild, C.C., Zhan, X., 1997. Down-regulation of the filamentous actin cross-linking activity of cortactin by Src-mediated tyrosine phosphorylation. J. Biol. Chem. 272, 13911–13915. Huang, C., Liu, J., Haudenschild, C.C., Zhan, X., 1998. The role of tyrosine phosphorylation of cortactin in the locomotion of endothelial cells. J. Biol. Chem. 273, 25770–25776. Hudson, A.M., Cooley, L., 2002. A subset of dynamic actin rearrangements in Drosophila requires the Arp2/3 complex. J. Cell Biol. 156, 677 –687. Katsube, T., Takahisa, M., Ueda, R., Hashimoto, N., Kobayashi, M., Togashi, S., 1998. Cortactin associates with the cell–cell junction protein ZO-1 in both Drosophila and mouse. J. Biol. Chem. 273, 29672–29677. Kinley, A.W., Weed, S.A., Weaver, A.M., Karginov, A.V., Bissonette, E., Cooper, J.A., Parsons, J.T., 2003. Cortactin interacts with WIP in regulating Arp2/3 activation and membrane protrusion. Curr. Biol. 13, 384– 393. Li, Y., Tondravi, M., Liu, J., Smith, E., Haudenschild, C.C., Kaczmarek, M., Zhan, X., 2001. Cortactin potentiates bone metastasis of breast cancer cells. Cancer Res. 61, 6906–6911. Lloyd, T.E., Atkinson, R., Wu, M.N., Zhou, Y., Pennetta, G., Bellen, H.J., 2002. Hrs regulates endosome membrane invagination and tyrosine kinase receptor signaling in Drosophila. Cell 108, 261 –269. Maa, M.C., Wilson, L.K., Moyers, J.S., Vines, R.R., Parsons, J.T., Parsons, S.J., 1992. Identification and characterization of a cytoskeletonassociated, epidermal growth factor sensitive pp60c-src substrate. Oncogene 7, 2429– 2438. Mata, J., Curado, S., Ephrussi, A., Rørth, P., 2000. Tribbles coordinates mitosis and morphogenesis in Drosophila by regulating string/CDC25 proteolysis. Cell 101, 511 –522. McNiven, M.A., Kim, L., Krueger, E.W., Orth, J.D., Cao, H., Wong, T.W., 2000. Regulated interactions between dynamin and the actin-binding protein cortactin modulate cell shape. J. Cell Biol. 151, 187 –198. Moreau, V., Galan, J.M., Devilliers, G., Haguenauer-Tsapis, R., Winsor, B., 1997. The yeast actin-related protein Arp2p is required for the internalization step of endocytosis. Mol. Biol. Cell, 1361–1375. Patel, A.S., Schechter, G.L., Wasilenko, W.J., Somers, K.D., 1998. Overexpression of EMS1/cortactin in NIH3T3 fibroblasts causes
increased cell motility and invasion in vitro. Oncogene 16, 3227–3232. Rørth, P., Szabo, K., Bailey, A., Laverty, T., Rehm, J., Rubin, G.M., et al., 1998. Systematic gain-of-function genetics in Drosophila. Development 125, 1049–1057. Schuuring, E., Verhoeven, E., Litvinov, S., Michalides, R.J., 1993. The product of the EMS1 gene, amplified and overexpressed in human carcinomas, is homologous to a v-src substrate and is located in cellsubstratum contact sites. Mol. Cell Biol. 13, 2891–2898. Seto, E.S., Bellen, H.J., Lloyd, T.E., 2002. When cell biology meets development: endocytic regulation of signaling pathways. Genes Dev. 16, 1314–1336. Takahashi, F., Endo, S., Kojima, T., Saigo, K., 1996. Regulation of cell – cell contacts in developing Drosophila eyes by Dsrc41, a new, close relative of vertebrate c-src. Genes Dev. 10, 1645– 1656. Tanentzapf, G., Smith, C., McGlade, J., Tepass, U., 2000. Apical, lateral, and basal polarization cues contribute to the development of the follicular epithelium during Drosophila oogenesis. J. Cell Biol. 151, 891 –904. Tateno, M., Nishida, Y., Adachi-Yamada, T., 2000. Regulation of JNK by Src during Drosophila development. Science 287, 324–327. Uruno, T., Liu, J., Zhang, P., Fan, Y., Egile, C., Li, R., et al., 2001. Activation of Arp2/3 complex-mediated actin polymerization by cortactin. Nat. Cell Biol. 3, 259–266. Weaver, A.M., Karginov, A.V., Kinley, A.W., Weed, S.A., Li, Y., Parsons, J.T., Cooper, J.A., 2001. Cortactin promotes and stabilizes Arp2/3induced actin filament network formation. Curr. Biol. 11, 370–374. Weaver, A.M., Heuser, J.E., Karginov, A.V., Lee, W.L., Parsons, J.T., Cooper, J.A., 2002. Interaction of cortactin and N-WASp with Arp2/3 complex. Curr. Biol., 1270–1278. Weed, S.A., Du, Y., Parsons, J.T., 1998. Translocation of cortactin to the cell periphery is mediated by the small GTPase Rac1. J. Cell Sci. 111, 2433–2443. Wu, H., Parsons, J.T., 1993. Cortactin, an 80/85-kilodalton pp60src substrate, is a filamentous actin-binding protein enriched in the cell cortex. J. Cell Biol. 120, 1417–1426. Wu, H., Reynolds, A.B., Kanner, S.B., Vines, R.R., Parsons, J.T., 1991. Identification and characterization of a novel cytoskeleton-associated pp60src substrate. Mol. Cell Biol. 11, 5113–5124. Zhan, X., Hu, X., Hampton, B., Burgess, W.H., Friesel, R., Maciag, T., 1993. Murine cortactin is phosphorylated in response to fibroblast growth factor-1 on tyrosine residues late in the G1 phase of the BALB/c 3T3 cell cycle. J. Biol. Chem. 268, 24427–24431.