|Home | About | Journals | Submit | Contact Us | Français|
Most tissues display several features of cellular polarization. Besides the ubiquitous epithelial polarization in the apical-basal (A/B) axis, many epithelia (and associated organs) display a planar cell polarization (PCP). Recently, a crosstalk between the PCP and A/B polarity determinants has been suggested, i.e. the activity or stability of the PCP factor Frizzled is regulated by the A/B determinants aPKC and Bazooka in the Drosophila eye. We have systematically investigated genetic and physical interactions between the Drosophila A/B factors and the core PCP component Strabismus(Stbm)/Van Gogh(Vang). The A/B determinant Scribble was found to interact both genetically and physically with Stbm/Vang. We demonstrate that Scribble binds Stbm/Vang through its PDZ domain 3 and that it cooperates with Stbm/Vang in PCP establishment. Our data indicate that Scribble, in addition to its role in A/B polarity, has a distinct requirement in PCP establishment in the Drosophila eye and wing. We define a scribble allele that is largely PCP specific. Our data show that Scribble is part of the Stbm/Vang PCP complex and further suggest that it might act as an effector of Stbm/Vang during PCP establishment.
In metazoans, epithelial tissues exhibit two types of polarity: (1) the Apical/Basal (A/B) polarity perpendicular to the epithelial plane, and (2) Planar Cell Polarity (PCP; often also referred to as tissue polarity) within the plane of the epithelium. Both types of polarity are required for the development, morphogenesis, and function of most (if not all) tissues and organs. Defects in both types of polarity are also associated with several genetic diseases (reviewed in Simons and Mlodzik, 2008).
Genetic screens in Drosophila and C. elegans have identified 3 protein complexes playing a major role in A/B polarity establishment and maintenance (see also Supplemental Fig. S1): (1) The Crumbs (Crb) complex, constituted of Crb, Stardust (Sdt) and Patj; (2) the Par3/Bazooka (Baz) complex, comprised of Baz, Par6 and atypical protein kinase C (aPKC); and (3) the Scribble (Scrib) complex, comprising Scrib, Lethal giant larvae (Lgl), and Discs large (Dlg) (for reviews see (Bilder, 2004; Macara, 2004; Nelson, 2003)). In Drosophila, the Crumbs complex is the most apical, followed by the Baz/Par3 complex, which is associated with the apical junctional areas, and the Scrib complex is localized to the septate junctions just basal to the adherens junction (for reviews see Bilder, 2004; Macara, 2004; Nelson, 2003). Many of the proteins belonging to these 3 complexes contain PDZ protein-protein interaction domains. In particular Sdt, dPatj, Baz, Par6, Scrib and Dlg all contain multiple PDZ domains.
Planar cell polarity (PCP) was first described in Drosophila and is manifest in almost all external structures derived from imaginal disc epithelia (Adler, 2002; Klein and Mlodzik, 2005; Strutt, 2003). This is most evident in the distal orientation of wing hairs, the posterior orientation of cellular hairs and sensory bristles on the thorax and abdomen, and the very ordered arrangement of ommatidia in the eye. A number of genes have been defined as core PCP genes, as their mutations affect PCP features in most if not all tissues. These include genes encoding for the membrane associated factors such as the 7-pass transmembrane (TM) protein Frizzled (Fz), the atypical cadherin Flamingo/Starry Night (Fmi/Stan), and the 4-TM protein Strabismus/Van Gogh (Stbm/Vang), as well as for the cytoplasmic factors Disheveled (Dsh) and Diego (Dgo), which associate with Fz, and Prickle (Pk), which binds Stbm/Vang (reviewed in (Adler, 2002; Klein and Mlodzik, 2005; Lawrence et al., 2007; Strutt, 2003). This first set of core PCP factors is referred to as the Frizzled group (or also Fz/Stan-group) (Lawrence et al., 2007). A second core group of PCP factors centered around the proto-cadherins Fat and Dachsous acts in parallel to the Fz/Stangroup (Lawrence et al., 2007).
In the Drosophila eye, the PCP genes control not only the correct cell fate choice within the R3/R4 equivalence group, but also the orientation of the entire ommatidial units with respect to their neighbors and the entire eye field. In the adult eye, this is translated into the chiral arrangement of ommatidia, forming a mirror image across the dorso-ventral (D/V) midline, the equator (for reviews see Klein and Mlodzik, 2005; Strutt, 2003; see also Fig. 1A). Cell specification of the photoreceptor R3 vs. R4 subtype is the first and most critical step during PCP establishment in the third instar eye imaginal disc. Each ommatidium initially emerges as a symmetrical precluster posterior to the morphogenetic furrow (MF), where it then becomes “asymmetric” following the action of the PCP genes, with the precursor cell closer to the equator becoming specified as R3 and its polar neighbor as R4 (Klein and Mlodzik, 2005; Strutt, 2003) (also Fig. 1A). The specification of the R3/R4 photoreceptor pair is subsequently followed by a 90° rotation of the ommatidial preclusters towards the equator, eventually causing the dorsal and ventral halves of the eye to form a mirror image across the equator (for reviews see Klein and Mlodzik, 2005; Strutt, 2003; also Fig. 1B). PCP establishment in the fly wing results in the asymmetric formation of a single actin-rich prehair on the distal vertex of each wing cell, reflecting an orientation in the proximal-distal (P/D) axis. This is preceded by the asymmetric localization of the core PCP factors within the P/D axis (Mlodzik, 2002; Strutt, 2003).
As a prerequisite to PCP establishment and the regulatory interactions among the PCP core factors, all PCP factors are apically localized in epithelial cells (Adler, 2002; Klein and Mlodzik, 2005; Strutt, 2003). Their ensuing interactions lead to the first detectable asymmetric localization, either within the P/D/ axis in the wing, the D/V axis in the eye, or the A/P axis in the body wall. In the eye, this is most evident across the R3/R4 membrane border: Fz, Dsh and Dgo localize to the R3 side, whereas Stbm and Pk localize to the R4 side of that border. This is also mirrored in the genetic requirements of the core PCP factors, with Fz, Dsh and Dgo being required in R3 for its cell fate induction, and Stbm and Pk acting in the R4 precursor (Fanto and Mlodzik, 1999; Tomlinson and Struhl, 1999; Wolff and Rubin, 1998; Zheng et al., 1995).
Recent work has provided evidence for a crosstalk between the PCP and the A/B polarity determinants. For example, Fz activity or stability is regulated by phosphorylation mediated by aPKC and antagonized by Bazooka (Djiane et al., 2005) and Dsh has been shown to regulate Lgl to participate in A/B polarity (Dollar et al., 2005). Finally, Dlg and Scrib interact with Stbm/Vang respectively during Drosophila sensory organ asymmetric cell divisions, and during the alignment of mouse inner ear choclear cilial cells (Bellaiche et al., 2004; Lee et al., 2003; Montcouquiol et al., 2003).
We systematically investigated the Drosophila A/B factors whether they can interact with Stbm/Vang and found that Scrib interacts both genetically and physically with Stbm/Vang. Here, we demonstrate that Scrib, in parallel to its major role in A/B polarity, is required specifically for PCP establishment in the Drosophila eye and wing. We define a scrib allele that is largely PCP specific, which interestingly has very similar molecular defects to the PCP specific mouse Scrib1 allele Circletail. Our data suggest that Scrib is part of a Stbm/Vang protein complex and that it might act as an effector of this complex during PCP establishment.
Based on recent evidence for a crosstalk between A/B and PCP factors (Bellaiche et al., 2004; Djiane et al., 2005; Dollar et al., 2005; Montcouquiol et al., 2003), we systematically tested whether stbm can interact genetically with any of the A/B polarity factors. We used the hypomorphic stbm153 allele, which displays mild PCP phenotypes covering a wide range of defects, including chirality defects, misrotations, and rare symmetrical ommatidia in the eye (this allele is wrongly described as a null in Flybase; it is a phenotypically defined hypomorphic allele; Rawls and Wolff, 2003). With a total of 38% +/- 6.9 affected ommatidia (Fig. 1C,F; compare to wild-type, Fig. 1B) this allele appeared well suited for genetic modification studies (Rawls and Wolff, 2003). The homozygous stbm153 background was combined with heterozygous conditions for alleles of the A/B determinants containing PDZ domains (see Supplemental Fig. S1 for schematic representation of A/B polarity factors).
Strikingly, homozygous stbm153 in combination with heterozygous mutant alleles scrib1 or scrib673 (both strong loss-of-function/LOF alleles) showed an enhancement of the stbm153 associated PCP defects (Fig. 1D-F). In contrast, none of the other A/B determinants tested for dominant modification of stbm153 did show an interaction. In particular strong LOF alleles for dlg (dlgm52, dlgIP20, dlgm35, dlgw55) and lgl (lgl4w3), the partners of scrib during A/B polarity establishment, did not modify the PCP defects of stbm153 (Fig. 1F), suggesting that the interaction between scrib and stbm/Vang is specific. Similarly, none of the other A/B polarity genes tested showed a modification of the stbm153 PCP phenotype (Fig. 1F and not shown).
In parallel to the genetic approach, we tested in a yeast two-hybrid assay whether the intracellular carboxy-tail (C-tail) of Stbm would display a protein-protein interaction with these A/B polarity determinants, using their PDZ domains as prey (see Supplemental Fig. S1 for schematics of the A/B factors). Interestingly, the Stbm C-tail interacted with the PDZ domains of Scrib and Dlg, whereas no interaction was observed between the Stbm C-tail and the Baz, Par6, Sdt or dPatj PDZ domains (Fig. 2A). A Stbm-Dlg interaction has been reported in another context and thus this interaction could be used as a positive control (Bellaiche et al., 2004; Lee et al., 2003). Similarly, mouse Stbm (Vangl2/Looptail) has been shown to bind mouse Scrib1 (Montcouquiol et al., 2003; Montcouquiol et al., 2006), supporting our results.
Taken together, these experiments suggested a specific interaction between Stbm/Vang and Scrib in the PCP context. We thus focused our efforts on dissecting the potential role of scrib in PCP establishment.
In order to confirm the two-hybrid results and also define which PDZ domain(s) of Scrib interacts with Stbm, we performed GST-pull down assays using a GST-Stbm C-tail, and radiolabeled in vitro translated PDZ domains of Scribble. As the Stbm C-tail is known to interact with itself (Jenny et al., 2003), it served as a positive control. The Stbm C-tail pulled down PDZ domains 3-4 as well as itself, but it did not interact with PDZ domains 1-2 (Fig. 2B). When individual PDZ domains of Scrib were tested, Stbm C-tail pulled down PDZ 3 efficiently and only a weak interaction was detected with PDZ 4 (Fig. 2C; again no interaction was detected with PDZ 1 or 2). These results are consistent with what has been observed with the mouse proteins (Montcouquiol et al., 2006). We conclude that PDZ 3 domain is the domain of Drosophila Scrib that specifically binds to the Stbm C-tail.
A conserved PDZ binding motif (PBM) has been identified for the last 4 amino acids in the C-tail of Stbm/Vang (Wolff and Rubin, 1998). We thus next tested whether the PBM was necessary for the interaction between Stbm and the Scrib PDZ 3 domain and performed the equivalent pull down assays using a GST-Stbm C-tail with its last 4 amino acids deleted (StbmΔPBM). Unexpectedly, StbmΔPBM still effectively bound the Scrib PDZ 3 domain (albeit slightly less efficiently as compared to wild-type Stbm (Fig. 2B, right panel). This was confirmed using GST-PDZ3 of Scrib and testing its binding to either radiolabeled Stbm-C-tail or Stbm-C-tailΔPBM (Fig. 2C). Again the binding was not significantly affected by the presence or absence of the PBM and thus we conclude that the PBM is not necessary for the interaction.
In order to analyze the functional requirements of the Stbm PBM in vivo, we generated transgenic flies expressing stbmΔPBM. To avoid the caveats of exogenous (or ectopic) expression of stbm, we expressed stbmΔPBM under the control of its endogenous promoter and control sequences of stbm (stbm-stbmΔPBM; see Methods). Several independent insertions on the third chromosome, rescued the stbm mutant phenotype to differing degrees with up to 90% (Fig. 2D and not shown; the differences in rescue efficiency resulted most likely from position effects, as similar variations were observed with the equivalent stbm-stbmwt transgenics). These data indicated that the PBM is not essential for stbm function in vivo and are also consistent with the observation that stbmΔPBM expressed under the control of an actin promoter was also able to rescue stbm- PCP phenotypes (Bastock et al., 2003) (although a different group has suggested that the PBM is functionally important; T. Wolff, personal communication). Taken together with the molecular interaction data, these results indicated that the Stbm PBM is neither necessary for the physical Stbm/Vang-Scrib association in vitro nor the stbm/Vang function in vivo.
Based on the interactions described above, we next tested whether scrib was required for PCP establishment. Scrib is a critical factor of Apical/Basal polarity establishment and maintenance, and thus scrib- mutant epithelial cells loose A/B polarity leading to multi-layered cell sheets with rounded cells (Bilder et al., 2000; Bilder and Perrimon, 2000) (see also Supplemental Fig. S2). It is therefore not possible to study PCP in the null allele condition. We therefore analyzed a hypomorphic scrib allele, scrib5. This allele was useful for our study as A/B polarity appears to be largely unaffected (Zeitler et al., 2004) and, importantly, it encodes for a protein lacking the last two PDZs domains, PDZ 3 and 4 (Fig. 3A), which bind Stbm physically in vitro (see above; Fig. 2).
Using this allele, we performed clonal analyses in developing eyes and wings using the FRT/FLP system (Golic and Lindquist, 1989). Clones of the scrib5 allele show normal A/B polarity in imaginal discs, with apical and basal factors being correctly localized (Supplemental Fig. S2C and data not shown). In adult eye sections we could however only recover small scrib5 clones. In such clones, we noticed occasional R3/R3 symmetrical ommatidia as well as rotation and chirality defects (Fig. 3B; highlighted by yellow arrowheads, and data not shown). In very rare larger clones, the overall eye morphology showed significant abnormalities in ommatidial architecture with fused photoreceptors, malformed rhabdomeres, and loss of photoreceptors (Fig. 3B’), making it impossible to quantify and analyze the eye PCP phenotype in detail (the morphology defects are likely due to a potential requirement of full length Scrib during the formation of the rhabdomeres and morphogenesis of photoreceptors in pupal eye development, which requires changes in the apical/basal polarization of these specialized neurons; Pellikka et al., 2002; Tepass and Harris, 2007).
To bypass the late photoreceptor morphogenesis requirement of scrib in the eye, we performed a clonal analysis in eye imaginal discs using the molecular markers for the R3/R4 precursors: mδ0.5-lacZ, which is expressed predominantly in R4 (Cooper and Bray, 1999), and psq-Gal4, UAS-GFP (psq>GFP, Weber et al., 2008), which is expressed differentially in both cells of the pair with high levels in R3 and low levels in R4. Analysis of scrib5 clones using the mδ0.5-lacZ marker, showed typical PCP defects with misrotations, chirality flips, and a fairly high frequency of loss of the R4 marker (~15%), suggesting that these clusters were R3/R3 symmetrical clusters (Fig. 4A-A”). Interestingly, we never observed a symmetrical expression of mδ0.5-lacZ in both R3/R4 precursor cells (reflecting R4/R4 type clusters), which can be observed in for example fz mutant clones (not shown, Cooper and Bray, 1999; Zheng et al., 1995). As expected, in clones of stronger alleles, like the null allele scrib1, severe defects in A/B polarity were observed (e.g. the apical markers dPatj and DE-Cad were mislocalized; Supplemental Fig. S2B; Bilder and Perrimon, 2000) and none of the neuronal markers (ELAV and mδ0.5-lacZ) were expressed (Supplemental Fig. S2A). It was thus not possible to assess PCP in the scrib null allele.
Analysis of scrib5 mutant clones with the psq>GFP marker confirmed and refined the phenotypic analysis with respect to PCP defects. Strikingly, ~25% of R3/R4 precursor pairs displayed identically high levels of GFP expression, reflecting R3/R3-type clusters (Fig. 4B-B”). Taken together with the mδ0.5-lacZ R4 marker results (where no R4/R4-type clusters were detected), we conclude that these clusters represent indeed R3/R3-type symmetrical clusters. Similar to the mδ0.5-lacZ marker, we observed ~15% of chirality flips (Fig. 4B-B”; example highlighted by red arrowhead) and frequent misrotations. This latter defect was best visualized with DE-Cad as marker for the overall orientation of the preclusters (Fig. 4C-C”). In summary, this scrib5 clonal analysis in the eye with the two independent and complementary markers mδ0.5-lacZ and psq-GFP revealed a range of defects very similar to and characteristic of core PCP determinants (~40-45% of all clusters; similar to strong hypomorphic alleles of fz or stbm/Vang; e.g. Rawls and Wolff, 2003).
Similar to the eye, marked scrib5 clones in wings also displayed characteristic PCP defects. These were manifest in cellular misorientations as apparent in “waves” and “whorls” (Fig. 3C,D), typical for mutations in PCP factors. Multiple wing hair defects were also observed (Fig. 3E). These defects are already apparent during development in pupal wings as reflected in the misorientation of actin hairs (Fig. 5A-A’; note that some non-autonomous PCP defects in wild-type cells neighboring mutant clones can be observed, examples marked by yellow arrowheads; see Discussion). In addition, PCP defects are also apparent in pupal wings in scrib5 clones by their failure to resolve the PCP factor localization to the membranes in the proximal-distal (P/D) axis (Fig. 5B-C’ and quantified in Fig. 5D). The Fmi staining pattern in scrib5 clones resembles in part the defect seen in stbm6 clones, in that in both genotypes Fmi accumulates at the border between wild-type and mutant cells (such cells are highlighted by yellow dots in Fig. 5B’ and C-C’, compare to stbm6 clones in Fig.5 E-E’). The main difference between scrib5 and stbm6 clones is that mutant scrib5 cells still display Fmi around the apical cortex, albeit without enrichment in the P/D axis (quantified in Fig. 5D), whereas Fmi is reduced apically in stbm6 clones (Fig. 5E-E; as was also shown earlier).
Taken together, these data indicate that in the wing, like in the eye, scrib5 mutant tissue displays characteristic PCP factor like phenotypes, suggesting that scrib, besides its critical role in A/B polarity establishment, shares a distinct PCP specific function with similarity to stbm/Vang (see below and Discussion).
As Scrib can bind to Stbm we also tested whether Scrib localization is affected in PCP mutants. Surprisingly, we did not see detectable Scrib localization defects in either stbm6 (Fig. 5F-F’) or fzP21 clones (not shown). This can be explained with the vast majority of Scrib being associated with the A/B determinant complexes and thus a small change in localization due to a PCP association of Scrib appears undetectable.
Next we wished to determine the cellular requirement of scrib in the context of the R3/R4 pair, in analogy to the data established for example for the core PCP factors fz and stbm/Vang (Weber et al., 2008; Wolff and Rubin, 1998; Zheng et al., 1995). In order to do this, we performed a mosaic analysis using the psq>GFP marker, a reliable marker for differential R3/R4 specification used in previous PCP studies (Weber et al., 2008). We have analyzed 121 clusters mosaic within the R3/R4 pair. Strikingly, when precursor for R3 was mutant and the R4 precursor wild-type, the large majority of such clusters showed wild-type R3/R4 marker expression (Fig. 6A-C’). In contrast, when the precursor for R3 was wild-type and R4 mutant, the majority of such mosaic clusters (73%) displayed chirality defects (31.7% showed inverted chirality, and 41.3% were symmetrical; Fig. 6B-D). Our results thus indicate that scrib is required in the R4 precursors for R4 cell specification. Taken together with the genetic and physical interactions, our data suggest that scrib cooperates with stbm/Vang in R4 precursors and together they contribute to R4 cell fate specification.
To further address the potential interaction of scrib and stbm/Vang, we used a gain-of-function non-autonomous signaling assay in the wing (Wu and Mlodzik, 2008). Unlike Fz, which readily causes non-autonomous reorientation of hairs outside the expression domain, Stbm/Vang does so only poorly. For example, Stbm driven with dppGal4 (at 25°C) in a stripe along the A/P compartment boundary does only very rarely lead to cell orientation defects (Fig. 3G). Similarly, dppGal4 driven Scrib causes only very subtle defects (Fig. 3I). When Stbm and Scrib are co-expressed, however, robust cell reorientation of cells outside the expression domain is observed in all wings (Fig. 3H). The direction of re-orientation is towards the expression domain, opposite to the effect of Frizzled (Wu and Mlodzik, 2008), which is what one would predict for a strong Stbm effect. These data further support the notion that Scrib cooperates with Stbm/Vang in PCP signaling (see Discussion).
All core PCP factors are apically localized in Drosophila epithelia, which is a prerequisite for their function in PCP (e.g. Wu et al., 2004) and subsequently as a consequence of their molecular interactions they segregate into asymmetrically localized patterns in the R3/R4 precursors and in pupal wing cells (reviewed in Strutt, 2003; Klein and Mlodzik, 2005; Seifert and Mlodzik, 2007). Stbm for instance is found on the R4 side at the R3/R4 border and on the proximal side pupal wing cells.
In A/B polarity establishment, the main role of scrib is to control apical localization of apical factors like Crb (Bilder and Perrimon, 2000). We therefore asked whether the role of scrib in the PCP context could be to properly localize Stbm to the apical junctional regions. Comparing Stbm/Vang localization between wild-type and scrib5 mutant cells in eye imaginal discs, we observed no apparent differences: Stbm/Vang was apically localized in mutant cells indistinguishably from neighboring wild-type cells (Supplemental Fig. S3). These data indicate that the role of Scrib in PCP is not primarily to localize Stbm/Vang to apical junctional complexes and suggest a distinct and specific function of scrib in PCP establishment (see Discussion).
We have analyzed the requirements of the apical-basal (A/B) determinant Scribble (Scrib) during PCP establishment in Drosophila. Our data indicate that scrib acts generally in PCP (as shown by classical PCP defects in the wing and eye) with phenotypes comparable to stbm/Vang mutants (or also other members of the Fz-group of core PCP factors). Genetic and molecular data suggest that scrib cooperates with stbm/Vang and might possibly function as an effector of the core PCP factor Stbm/Vang.
Due to the subapical cortical localization requirement of the core PCP factors (Wu et al., 2004), we were interested to test for interactions between the PCP proteins and the A/B determinants. Earlier work has identified dPatj as a specific binding partner of Fz, required for restriction of Fz activity (Djiane et al., 2005). Among the major A/B polarity factors tested for interaction with Stbm, only Scrib interacted both physically and genetically. As shown earlier for the other members of the Scrib complex (Scrib, Dlg, Lgl; Lee et al., 2003), we also found Dlg to physically bind to the Stbm C-term. However, we did not detect any genetic interactions between stbm and dlg in PCP contexts. These data suggest that the interaction between Stbm and Scrib is specific and restricted to PCP establishment, which is also supported by genetic interactions of the respective genes in the mouse (Montcouquiol et al., 2003; see below).
A structure/function analysis of Scrib in the context A/B polarity in C. elegans and Drosophila has established that the LRR domains are necessary for targeting Scrib to the basolateral membrane, while PDZ domains 1 and 2 play an important role by concentrating Scrib proteins at the septate junctions (Albertson et al., 2004; Legouis et al., 2003; Zeitler et al., 2004). A specific role for the PDZ 3 and 4 domains was not observed in this context. Our data indicate that the Scrib PDZ 3 domain is necessary and sufficient for its physical interaction with Stbm. The in vivo data with the scrib5 allele, in which the PDZ domains 3 and 4 are deleted, demonstrate that these domains are critical for PCP function (our work, see below).
Taken together our data suggest that scrib plays not only a major role in A/B polarity establishment through its LRR and PDZ 1 and 2 domains, but that it also has important and distinct requirements in PCP establishment through the PDZ 3 and 4 domains and their interaction with Stbm. How does this relate to the reported Scrib-Stbm interaction in vertebrates? Analyses with mammalian Scrib proteins have suggested that PDZ domains 2 and 3 of mScrib1 can bind Vangl2, although, like in our studies, the PDZ domain 3 had the strongest affinity for the Vangl2 (Stbm) C-tail (Kallay et al., 2006; Montcouquiol et al., 2006).
The physical binding studies indicate that the PBM of Stbm/Vang is not necessary for Scrib PDZ 3 binding to Stbm (although the binding between Stbm C-tail and PDZ 3 is somewhat stronger in the presence of the PBM; Fig. 2C). However, analyses of the mammalian orthologs of scrib and stbm (mouse scrib1 and vangl2, respectively) suggested that the Vangl2 PBM is required for its interaction with mScrib1 (Kallay et al., 2006; Montcouquiol et al., 2006). In functional assays in vivo, the PBM of Stbm/Vang appears dispensable and StbmΔPBM rescues the stbm/Vang mutant as well as StbmWT does. This is observed in both our rescue experiments in the eye using the endogenous promoter of stbm/Vang, as well as in experiments using heterologous expression cassettes (the actin promoter) assaying rescue in the wing (Bastock et al., 2003). We conclude that although the Stbm/Vang PBM can make its interaction with Scrib more efficient in Drosophila it is largely dispensable here.
In vertebrates, the role of the PBM of Stbm/Vangl family members is controversial. In convergent extension (CE) in vertebrates, a process also regulated by the core Fz-group PCP factors (Seifert and Mlodzik, 2007; Wang and Nathans, 2007), it has been shown in Xenopus that increasing amounts of StbmΔPBM inhibit the cell movement indistinguishable from xStbm full length and that the Stbm/Vang PBM is therefore not required in CE (Darken et al., 2002). However, other studies suggest that the PBM is necessary for Stbm function in modulating CE in zebrafish and Xenopus, and that StbmΔPBM might antagonize StbmWT (Goto and Keller, 2002; Park and Moon, 2002). Without any functional rescue experiments in vertebrates this issue cannot be conclusively answered.
We have shown that scrib and stbm interact physically and genetically during PCP establishment in Drosophila. Similar observations were reported for PCP generation in the mouse cochlea (Montcouquiol et al., 2003; Montcouquiol et al., 2006). In the mouse, the scrib and stbm homologs, Circletail (Crc/mScrib1) and Vangl2 (originally named Looptail/Lp) also interact during neural tube closure (Kibar et al., 2001; Murdoch et al., 2001; Murdoch et al., 2003). Moreover, only the double homozygous Crc and Vangl2/Lp mutants have been so far shown to display craniorachischisis, the most severe form of neural tube defects (NTD) (Murdoch et al., 2003). NTDs are a consequence of morphogenetic movement defects during late stages of gastrulation and CE, highlighting the importance of the stbm and scrib cooperation in PCP associated processes. Most interestingly, the molecular defect in the mScrib1Crc mutation has been identified as a single base insertion in codon 947, causing a frame shift and resulting in a premature stop codon and truncation of the protein before the PDZ 3 and 4 domains (Murdoch et al., 2003). Thus the molecular lesion in the mScrib1Crc mice largely mimics the molecular mutation in the scrib5 allele in Drosophila, where also the protein is truncated before the PDZ 3 and 4 domains. This observation confirms the requirement of the PDZ domains 3 and 4 in PCP signaling and thus the scrib5 allele in Drosophila and the mScrib1Crc allele in mouse are likely to also share functional defects in the PCP context in the mouse and Drosophila.
It is worth noting that scrib is expressed in two isoforms during Drosophila embryogenesis, one giving rise the full length 1756 aa protein (called Scrib1) and a shorter form of 1247 residues (Scrib2), which is lacking the PDZ domains 3-4 (Li et al., 2001). Both isoforms are expressed in vivo during embryogenesis and thus one can speculate that it is functionally important to have the 2 different isoforms. Possibly, the Scrib2 isoform, missing PDZ domains 3 and 4 has a distinct function than the Scrib1 isoform (Bilder and Perrimon, 2000; Zeitler et al., 2004). In summary, we conclude that Scrib, through its interaction with Stbm/Vang, is an important PCP factor, and this function is evolutionarily conserved, requiring the PDZ domains 3-4.
The core PCP factors Stbm and Fz play opposing roles during PCP establishment, mutually antagonizing each other (Klein and Mlodzik, 2005; Seifert and Mlodzik, 2007; Strutt, 2003). In the R3/R4 PCP decision in the Drosophila eye fz is required in R3 (Zheng et al., 1995) and stbm/Vang is required in R4 (Wolff and Rubin, 1998). Mosaic analysis with scrib5 demonstrated that scrib is also required in R4 for its proper specification, very similar to stbm/Vang. In the wing, the LOF PCP phenotype of scrib5 clones looks also similar to the core PCP factors, in particular to stbm/Vang, with misoriented wing hairs and occasional “multiple wing hair” defects.
What is the role of Scrib in PCP? Localization studies of Stbm/Vang in scrib5 clones do not reveal appreciable defects in the A/B axis in the tissues analyzed, indicating that Scrib is not required for proper apical Stbm/Vang localization. Scrib protein localization is consistent with its role on A/B polarity throughout development. In addition, it appears to become transiently enriched at the subapical junctions within the proximo-distal axis in pupal wings (not shown), although it largely detected all around the subapical cortex. Thus, this effect is transient and very difficult to visualize due to the PCP independent “ubiquitous” Scrib localization at the junctional levels. Taken together with the physical binding data, the genetic interactions, and the cooperation (synergy) between Scrin and Stbm/Vang in the gain-of-function assay, we hypothesize that Scrib could function as a downstream effector of Stbm/Vang in PCP generation. In such a model, we would propose that the “PCP-activity” of Scrib is locally controlled by Stbm/Vang, either through local modification or recruitment to other factors or unknown means. Scrib has been implicated in the control of localized exocytosis through a complex with PIX (a Guanine nucleotide Exchange Factor/GEF for Arf; Audebert et al., 2004). One could imagine a model in which Stbm by recruiting Scrib would control exocytosis locally which could in turn affect the outcome of the PCP core factor interactions and or stability thus accounting for the effects on PCP in a general manner.
As discussed above the phenotypic features of scrib5 are similar to stbm/Vang mutant alleles. However, the defects in scrib5 are milder. In this regard scrib is similar to pk mutants, with Pk also being part of the Stbm/Vang complex displaying phenotypic defects that are milder than stbm/Vang mutant alleles. For example, Pk has no apparent function in the non-autonomous signaling role of Stbm/Vang (Adler et al., 2000). Interestingly, srcib5 mutant clones show some non-autonomous defects in wing tissue and Scrib cooperates with Stbm/Vang in non-autonomous signaling assay, which could suggest that Scrib mediates (in part) the non-autonomous function of Stbm/Vang (whereas Pk is strictly autonomous). A detailed dissection of the relationship of Scrib and Pk will be needed to address this further. Due to the general requirements and localization of Scrib in A/B polarity this is technically a very challenging task. Possibly, dissecting the functional requirement(s) of Stbm that are shared with Scrib will help to clarify this issue.
Flies were grown on standard fly medium at 25°C unless otherwise indicated. The following mutant stocks were used:
The following genotypes were used to induce clones and analyze them:
The mosaic analyses were performed as described (Weber et al., 2008), using psqF112Gal4, UAS-GFP, an excellent marker for R3/R4 differentiation with high expression of GFP in the R3 precurso and low expression of GFP in R4.
Imaginal discs were dissected and stained as described (Fanto and Mlodzik, 1999). The following primary antibodies were used:
All secondary antibodies were from Jackson Laboratories. Imaginal discs were mounted in Mowiol and viewed with a Zeiss LSM510 Meta Confocal microscope. Single optical sections or stacks of several sections processed using Image J and Photoshop are shown.
Adult eyes were embedded in Durcupan resin and tangential sections were taken at the equatorial region. 3-6 independent eyes with over 150 ommatidia each were analyzed per genotype. All images are oriented dorsal up, anterior left.
Stbm C-term (aa 294-584) was used as bait against the PDZ domain-containing Apical/Basal (A/B) determinants as preys. The A/B determinants were cloned as described (Djiane et al., 2005). Yeast cells were transfected with bait and prey plasmids using Match Maker kit (Clontech, according to the manufacturer’s protocol). Transfected cells were plated on selective media lacking Leu and Trp for 3 days to select for double transformants, and were then streaked on media lacking Leu, Trp and Ade in order to test for potential interactions.
GSTStbm_Cterm and GSTStbm_CtermΔPBM were cloned and expressed as described (Jenny et al., 2003). GSTScrib_PDZ3-4 (aa 1151-1333) was cloned into pGEX4-T1.
S35-radiolabeled PDZ domains of Scribble (aa 738-818; aa 947-1019; aa 738-1019; aa 1151-1234; aa 1264-1333; aa 1151-1333) and Stbm C-term and Stbm C-termΔPBM (Jenny et al., 2003), were generated by in vitro translation using the TNT Coupled Reticulocyte Lysate System (Promega, according to manufacturer’s protocols). Truncated isoforms of Scrib were cloned by PCR into a modified version of pβ-Globin vector and GST pull-downs were performed as described (Jenny et al., 2003).
We are grateful to Drs. David Bilder, Tanya Wolff, Chris Doe, and Manzoor Bhat, and the Bloomington stock center for Drosophila strains and antibodies. We thank S. Okello and A. Blitzer for technical help, W. Gault for careful reading and comments on the manuscript, and all members of the Mlodzik lab for helpful discussions and input. This work has been supported by NIH grants EY13256 and GM62917 to MM. Confocal microscopy was performed at the MSSM-Microscopy Shared Resource Facility.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.