The Drosophila wing is covered by an array of distally pointing hairs. This tissue planar polarity is regulated by the frizzled pathway. The function of the grainy head transcription factor is essential for the function of the frizzled pathway. grainy head mutant cells fail to localize planar polarity proteins at either the proximal or distal sides of wing cells and produce multiple hairs of abnormal polarity. Levels of the Starry night protein are strongly reduced in grainy head mutants in both larval wing discs and pupal wings, which is sufficient to account for much of the polarity phenotype. In addition, grh has frizzled pathway independent functions during the development of the adult cuticle (Lee, 2004).
grh function is required for several different processes during the differentiation of the adult Drosophila epidermis. These include the function of the fz dependent tissue polarity pathway, pigmentation, the timing of differentiation, epidermal hair morphogenesis and wing vein/blade specification. The Grh protein was originally isolated by virtue of its ability to bind to DNA in a sequence specific manner and to regulate the expression of target genes. These and later experiments led to the conclusion that grh functions as a transcription factor for development specific gene regulation. Experiments on vertebrate homologs of grh also suggest a similar cellular function. It is likely that it serves a similar function in the development of the adult epidermis (Lee, 2004).
The analysis of grh function in regulating gene expression appears complex. The first studies on grh argued that it acted as a positive regulator of Ddc and Ubx expression. Curiously, although Grh was isolated by virtue of its ability to bind to a sequence essential for the neuronal activation of Ddc, grh mutations alter the epidermal and not neuronal expression of Ddc. More recently it was found that grh positively regulates tll expression and negatively regulates ventral dpp expression (Lee, 2004).
The function of the grh transcription factor is shown in this study to be required for the function of the fz pathway in the wing. In the absence of grh function the Fz, Dsh and Vang proteins fail to accumulate apically and the levels of the Stan protein are dramatically decreased. Furthermore, Stan levels are increased in cells with two versus one copy of grh. Thus, stan expression is directly related to grh dose suggesting that stan might be a direct target of Grh. The direct relationship between stan expression and grh dose is seen in both pupal wing cells where Stan is localized assymetrically and in third instar wing disc cells where it is evenly distributed. Thus, it is concluded that the decreased levels of Stan protein in grh cells is not due to a failure of assymetric localization. Grh does not affect Stan stability; stan expression from the endogenous stan gene is altered, consistent with Grh having an important role in promoting stan transcription. It is suspected that this could be due to a direct interaction of Grh protein with stan genomic DNA. stan does not appear to be highly enriched in putative Grh binding sites but this may be a reflection of the variability in identified Grh binding sites not providing an ideal consensus site. It is also concluded that the decreased level of Stan protein is neither the cause or effect of the the delay in hair morphogenesis in grh cells. Thus far, all of the proteins that localize assymetrically are co-required for the asymmetric localization of the others, however only Stan is required for the apical accumulation of all of the other proteins. The alterations in tissue polarity protein localization seen in grh mutant cells could be explained entirely by the effect of grh on Stan expression. It remains possible however, that grh could be important for the expression of several or all members of the tissue polarity group. These experiments did not allow the assessment of possible changes in Fz or Vang levels due to altered expression of these genes, since the localization was examined of proteins produced from transgenes that did not utilize the normal promoters. Decreased levels of Dsh were not seen by antibody staining, however the staining background was relatively high in these experiments which could have hidden a modest effect on Dsh levels. The finding that Arm cortical localization is not altered in grh clone cells indicates that apical-basal polarity is not altered and suggests that gross cellular physiology is not altered in grh clones (Lee, 2004).
While it is possible that the grh mutant planar polarity phenotype could be due solely to a lack of stan expression in grh mutant cells, this may not be the case since there are a number of differences between the phenotypes of grh and stan clones. For example, the multiple hair phenotype of grh is much stronger than stan. There is also a difference in the non-autonomy of grh and stan clones. For mutations in both of these genes the domineering nonautonomy of clones is much weaker than that of fz or Vang. However, the weak domineering nonautonomy is seen much more frequently with grh than stan clones, suggesting that grh mutations alter the expression of additional tissue polarity genes or other cellular genes that interact with the planar polarity system. Genetic screens for enhancers or suppressors of the dominant negative grhFK2131 allele could be useful in identifying such genes (Lee, 2004).
grh has both fz pathway dependent and independent functions during wing development. Epistasis experiments showed that the ectopic wing vein, cuticle pigmentation, disturbed marginal bristle row and extreme multiple hair cell phenotypes of grh mutations are not altered in a null fz, in or mwh mutants. Thus, it is quite likely that some of the target genes whose transcription is altered by grh mutations are not part of the fz pathway (Lee, 2004).
grh cells are often dramatically delayed in hair morphogenesis. This is not seen in cells mutant for fz or stan and hence is unlikely to be an indirect consequence of a failure of stan expression or in the inactivation of the frizzled pathway. The time course of pupal development is controlled by ecdysone and it is possible that grh functions as part of the ecdysone cascade. The delay in hair morphogenesis could be due to a failure to induce the expression of genes such as kojak, where a loss of function results in a similar delay (Lee, 2004).
The grh multiple hair cell phenotype differs from that of downstream members of the fz pathway such as inturned, in not showing the typical fz/in abnormal polarity pattern and in the hairs being much more erect. The identity of the targets responsible for this phenotype are unkown. The grh hair phenotype is somewhat reminiscent of that seen with mutations in genes such as Rho kinase or crinkled (myosin VII) suggesting these or related genes as possible targets (Lee, 2004 and references therein).
The transcription of the Ddc gene has previously been shown to be regulated by grh and Ddc activity is required for melanization. Is ddc likely to be the target gene whose altered expression leads to the lowered pigmentation of grh clone cells? This is certainly possible but it seems unlikely to be the entire story. Ddcts2 flies raised at the restrictive condition have more profound pigmentation defects than grh clones. However, clones of ddc null alleles typically have a less severe pigmentation phenotype than grh clones due to partial rescue of the pigmentation phenotype by neighboring cells (i.e. ddc displays submissive cell non-autonomy). Based on these observations it is argued that grh must have other targets that contribute to the decreased pigmentation (Lee, 2004).
The data reported in this paper argue that grh has multiple functions during the development of the adult epidermis. In this context it is not clear to what extent grh functions in a permissive fashion to promote the expression of developmentally important genes and/or to promote changes in gene expression that are associated with the differentiation of the adult cuticle. The data are consistent with grh functioning in both ways. The requirement for grh for the expression of stan was seen at multiple stages consistent with grh having a permissive role. The effects on the timing of hair morphogenesis are consistent with, but do not demand an instructive role (Lee, 2004).
The class III phosphatidylinositol-3 kinase [PI3K (III)] regulates intracellular vesicular transport at multiple steps through the production of phosphatidylinositol-3-phosphate [PI(3)P]. While the localization of proteins at distinct membrane domains are likely regulated in different ways, the roles of PI3K (III) and its effectors have not been extensively investigated in a polarized cell during tissue development. This study, in vivo functions of PI3K (III) and its effector candidate Rabenosyn-5 (Rbsn-5) were examined in Drosophila wing primordial cells, which are polarized along the apical-basal axis. Knockdown of the PI3K (III) subunit Vps15 resulted in an accumulation of the apical junctional proteins DE-cadherin and Flamingo and also the basal membrane protein beta-integrin in intracellular vesicles. By contrast, knockdown of PI3K (III) increased lateral membrane-localized Fasciclin III (Fas III). Importantly, loss-of-function mutation of Rbsn-5 recapitulated the aberrant localization phenotypes of beta-integrin and Fas III, but not those of DE-cadherin and Flamingo. These results suggest that PI3K (III) differentially regulates localization of proteins at distinct membrane domains and that Rbsn-5 mediates only a part of the PI3K (III)-dependent processes (Abe, 2009).
Cell polarity along the apical-basal axis is essential for the function of epithelial cells. This polarity is formed and maintained by distinct localization of membrane spanning and associated proteins, to apical, lateral or basal membrane domains. Membrane proteins localized to the apical or basolateral plasma membrane are endocytosed into early and apical or basolateral endosomes. For example, horseradish peroxidase (HRP) administered to the apical cell surface is incorporated into the apical early endosome. By contrast, HRP or dimeric IgA administered to the basolateral cell surface or transferring receptor (TfR) in the basolateral domain are internalized into the basolateral early endosome, which remain distinct. Sorting of proteins for transcytosis, recycling and degradation takes place in these early endosomes. The proteins, incorporated into apical and basolateral early endosomes, meet in common endosomes, a process that can be observed within 15 min after the onset of internalization in MDCK cells. The significance of keeping the apical and basolateral early endosomes distinct is thought to ensure that proteins from the apical and basolateral plasma membrane remain apart before the sorting processes proceeds. Although it is plausible that the trafficking of proteins in distinct membrane domains is regulated differently, the factors involved in such a differential regulation remain elusive (Abe, 2009).
One of the key molecules regulating membrane trafficking is PI3K (III), a heterodimer of Vps34p and Vps15p/p150, which produces phosphatidylinositol-3-phosphate (PI(3)P). PI(3)P is found to localize with early endosome and internal vesicles of multivesicular bodies (MVBs) in mammalian cells in culture. Genetic and pharmacological analysis, using yeast and mammalian cells in culture, suggests that PI3K (III) is required for five distinct processes. These are: (1) the fusion of clathrin-coated vesicles and early endosomes as well as the fusion between early endosomes; (2) the recycling from early endosomes back to the Golgi complex or other destinations; (3) the entry of proteins into the lysosomal degradation pathway; (4) the formation of internal vesicles of MVBs and (5) autophagy. Moreover, inactivation of PI3K (III) by Vps34 mutation leads to an expansion of the outer nuclear membrane and an abnormal reduction of the LDL receptor at the apical membrane in C. elegans. In Drosophila, dVps34 mutation results in defective endocytosis of the apical membrane protein Notch and a defective onset of autophagy. It has been suggested that PI3K (III) utilizes different effectors at apical and basolateral endosomes. However, the role of PI3K (III) in the regulation of protein localization at different membrane domains has remained unclear (Abe, 2009 and references therein).
To understand the various functions of PI3K (III), it is crucial to clarify which downstream effectors are involved in each of the processes it regulates. PI3K (III) is thought to exert its function through the recruitment of proteins that contain PI(3)P-binding motifs such as FYVE or PX domains. Among such proteins, Rabenosyn-5 (Rbsn-5) has been shown to contribute to endosome fusion and recycling processes in mammalian cells. Genetic studies on C. elegans and Drosophila also show that Rbsn-5 is essential for receptor-mediated endocytosis and endosome fusion, although it is not clear whether or not Rbsn-5 is involved in other PI3K (III)-related phenomena (Abe, 2009).
To determine how the proteins in distinct membrane domains are regulated by PI3K (III) and its effector Rbsn-5 this study analyzed Drosophila wing development. This provides a good model since wing primordial cells have a clear polarity along the apical-basal axis. In addition a number of membrane proteins are known to be transported in an organized manner along the apical-basal axis. For example DE-cadherin, a cell adhesion protein and Fmi, a planar cell polarity (PCP) core protein, are localized in the apical junctions or zonula adherens (ZA), whereas the cell adhesion molecules FasIII and β-integrin are localized in lateral and basal membranes, respectively. This study found that inactivation of PI3K (III) in the wing primordial cells by knockdown of dVps15 affects the localization of these membrane proteins differently. In particular, it was found that dVps15 knockdown results in the accumulation of FasIII at the lateral membrane, whereas it results in intracellular accumulation of DE-cadherin, Fmi and β-integrin. Importantly, inactivation of Rbsn-5 shows accumulation of FasIII and β-integrin at the lateral membrane and intracellular vesicles, respectively, but no effects of DE-cadherin and Fmi localization (see in contrast Mottola, 2010). These results provide evidence for a differential regulation of protein localization by PI3K (III) and Rbsn-5 at distinct membrane domains (Abe, 2009).
This study demonstrated that PI3K (III) differentially regulates the localization of proteins at distinct membrane domains. The intracellular accumulation of Fmi, DE-cadherin and β-integrin induced by the dVps15 knockdown might be due to defects in the degradation pathway, since the maturation of MVBs and the lysosomal trafficking were defective in these cells. However, unlike these proteins, Fas III did not accumulate in the intracellular compartments, but rather accumulated at the surface of the lateral plasma membrane. It is possible that PI3K (III) regulates proteins at the lateral membrane differently from those localized at other membrane domains. It is also possible that PI3K (III) regulates Fas III in a different way, irrespective of the membrane domain to which it is localized. Whichever is the case it will be important to elucidate the mechanism underlying this difference in a future study (Abe, 2009).
Rbsn-5, a FYVE domain-containing protein, shares a part of the functions of PI3K (III), in that it is necessary for the regulation of Fas III and β-integrin localization, but not that of DE-cadherin and Fmi localization. Although the Rbsn-5C241 null mutant clones may not completely lack Rbsn-5 activity, the requirement of Rbsn-5, or at least the requirement of an appropriate amount, differs between these proteins with respect to normal trafficking. It appears that Rbsn-5 preferentially controls the events at the basolateral regions, given that Rbsn-5 is necessary for the formation of large endosomes at the basal region, whereas it is indispensable for the formation of actin bundles at the apical surface (Abe, 2009).
PI3K (III) has been implicated in the differential regulation of vesicle trafficking at apical and basolateral regions. For instance, a reduction of PI(3)P dissociates EEA1, a FYVE-domain containing protein essential for early endosome fusion, selectively from basolateral endosomes. However, which proteins, including EEA1, regulate the different trafficking pathways downstream of PI3K (III) has remained unknown. Rbsn-5 has been proposed to be a PI3K (III) effector, since Rbsn-5 harbors a FYVE domain. The current results provide further evidence supporting a possible functional interaction between these two molecules, based on their genetic interaction on the wing morphogenesis and the PI3K (III)-dependent Rbsn-5 immunostaining. Importantly, the different requirement of Rbsn-5 for trafficking at apical junction and basolateral membrane domains suggests that Rbsn-5 may a selective regulator under the control of PI3K (III) (Abe, 2009).
To address whether Stan has an intercellular adhesion activity, stan was expressed in Drosophila S2 cells that exhibit a very weak self-aggregating property. Cell aggregates are formed by transfection of the S2 cells with a cDNA construct that encodes the full-length Stan protein (Usui, 1999).
During planar polarization of the Drosophila wing epithelium, the homophilic adhesion molecule Flamingo localizes to proximal/distal cell boundaries in response to Frizzled signaling; perturbing Frizzled signaling alters Flamingo/Starry night distribution, many cell diameters distant, by a mechanism that is not well understood. diego, a tissue polarity gene, encodes a protein comprised of six ankyrin repeats that colocalizes with Flamingo at proximal/distal boundaries. Diego is specifically required for polarized accumulation of Flamingo and drives ectopic clustering of Flamingo when overexpressed. It is suggested that Frizzled acts through Diego to promote local clustering of Flamingo, and that the clustering of Diego and Flamingo in one cell nonautonomously propagates to others. Localized Frizzled signaling would modify the properties of the Diego protein at proximal-distal boundaries, increasing Diego's ability to promote clustering of itself and Flamingo. A bias in Diego localization might then be enhanced via homophilic interactions between Flamingo molecules on adjacent cells. This model is consistent with the reciprocal effects that Flamingo and Diego have on each other's localization (Feiguin, 2001).
Does cortical localization of Diego depend on Flamingo? The distribution of Diego was examined in clones of Flamingo mutant cells. Diego is undetectable at the cortex of flamingo mutant cells and is also absent from both proximal and distal interfaces between wild-type and flamingo mutant tissue. These data suggest that Diego is localized to the cortex via Flamingo, either directly or indirectly. Like Diego, Flamingo also fails to accumulate at either the proximal or distal interface between wild-type and Flamingo mutant clones. This has been interpreted to mean that Flamingo is normally enriched both proximally and distally and that homophilic interactions are required for its polarized accumulation. Similarly, it can be argued that Diego is normally enriched on both proximal and distal sides of the cell and that stable cortical Diego localization occurs only where homophilic Flamingo interactions are possible (Feiguin, 2001).
Clearly, Flamingo is required for cortical accumulation of Diego. To ask whether it was also sufficient, that is, whether Diego simply accumulates wherever Flamingo is present, the distribution of Diego was examined in cells that overexpressed Flamingo. Although some Flamingo can be detected at cell boundaries when the protein is overexpressed, Flamingo accumulates most obviously in bright spots inside the cell. Since Flamingo is a transmembrane protein, these spots might either represent protein that is trapped in the secretory pathway or that has accumulated in the endocytic pathway. In Flamingo-overexpressing cells, Diego protein appears less abundant on the cortex overall, and what Diego remains is no longer restricted to proximal-distal cell boundaries. Diego does not colocalize with Flamingo inside the cell where Flamingo accumulates to the highest level. This indicates that other proteins besides Flamingo must promote the cortical localization of Diego. Colocalization of Diego and Flamingo upon transfection of S2 cells could not be seen, indicating that other proteins may regulate whether Diego and Flamingo are present in the same complexes. Nevertheless, at least at the cortex, dominant mislocalization of Flamingo by overexpression can relocalize Diego (Feiguin, 2001).
In wild-type cells adjacent to Flamingo-overexpressing cells, Diego disappears from the proximal-distal boundaries and appears to accumulate to a higher level at the boundary with the Flamingo-overexpression domain, perpendicular to its normal proximal-distal pattern. Furthermore, overaccumulation of Diego is also seen on the cell boundary opposite that which is in contact with the Flamingo-overexpressing cells. In some wings, reorientation of both Diego and Flamingo occurs up to three cell diameters away from the overexpression stripe. Thus, despite the fact that loss of Flamingo activity in clones has an essentially autonomous effect, depleting Diego and Flamingo only from the directly adjacent cell boundary, flamingo overexpression causes nonautonomous relocalization of both itself and Diego over several cell diameters (Feiguin, 2001).
Is Diego required for Flamingo localization? The distribution of Flamingo in wild-type versus homozygous diego mutant wings was compared. In all of the wild-type wings examined, the Flamingo protein is enriched on proximal-distal boundaries by 27 hr after puparium formation. In contrast, Flamingo in dgo380 wings localizes uniformly around the junctional region at this time, similar to its distribution in frizzled mutant wings. The ratio of Flamingo staining intensity on proximal-distal versus anterior-posterior cell boundaries was quantified in images from four wild-type wings and four dgo380 mutant wings. In wild-type wings, Flamingo is enriched between 2.5- and 3.2-fold on proximal-distal boundaries at 27 hr after puparium formation. In contrast, the Flamingo in dgo380 mutant wings is not significantly enriched (ratios ranged between 0.8 and 1.3). These data suggest that Diego promotes Flamingo polarization (Feiguin, 2001).
Flamingo sometimes polarizes its distribution in parts of dgo380 mutant wings at later times (29-30 hr after puparium formation), but the axis of polarity is often abnormal, presaging the whorls of misoriented hairs observed in dgo380 adult wings. In all cases, the effect of diego is specific to Flamingo and does not affect the distribution of cortical actin. Clonal analysis shows that diego loss of function does not affect Flamingo localization outside of diego clones. These data suggest that Diego increases the efficiency of Flamingo polarization and promotes its correspondence with the proximal-distal axis determined by Frizzled. The fact that Flamingo is localized similarly in frizzled and diego mutant cells may indicate that Frizzled acts through Diego to mediate Flamingo polarization (Feiguin, 2001).
To assess the effects of excess Diego protein on polarity, the effect of Diego overexpression on Flamingo localization was examined. When Diego protein is overexpressed with the ptc:GAL4 driver, it still localizes exclusively to the junctional region, but is present in large aggregates with no proximal-distal polarity. Flamingo protein accumulates in these cortical clusters, and its distribution is no longer polarized along the proximal-distal axis. These data suggest that Diego specifically promotes clustering of Flamingo. It is suggested that under wild-type conditions, Diego-mediated Flamingo clustering is regulated in response to Frizzled signaling and occurs only at proximal-distal boundaries (Feiguin, 2001).
Interestingly, Diego overexpression has nonautonomous effects on Flamingo localization in adjacent cells. In the first row of wild-type cells adjacent to Diego overexpressers, Flamingo is missing from proximal-distal boundaries and presumably relocalizes to maximize homophilic interactions with the clustered Flamingo on the Diego-overexpressing cells. Interestingly, Flamingo protein levels are also modestly elevated on the cell boundary opposite that of the Diego-overexpressing cells. As a result, the Flamingo in the second row of cells seems less unambiguously polarized to proximal-distal boundaries than in the cells at a greater distance from the overexpression stripe (Feiguin, 2001).
The ability of adjacent pupal wing cells to compare the activity of different tissue polarity genes and orient hairs accordingly is intrinsic to the mechanism of planar polarization. When artificial expression gradients are produced, Frizzled and Flamingo cause opposite effects on hair polarity; hairs point away from cells with higher Frizzled expression but toward cells with higher Flamingo expression. To ask how cells with differing levels of Diego behaved, the adult wings from UAS:dgo * ptc:GAL4 crosses were examined. Wing hairs were oriented from high- to low-expressing cells in ptc:GAL4;UAS:dgo wings -- similar to the effects caused by Frizzled and the opposite of those caused by Flamingo. These data suggest that Diego responds positively to Frizzled signaling and that the subsequent Diego-mediated clustering of Flamingo has negative effects on Frizzled signaling activity (Feiguin, 2001).
The mechanism by which Frizzled proteins mediates cell polarization events is not well understood. In wing epithelial cells, one consequence of Frizzled signaling is the polarization of Flamingo. Diego protein has been shown in this study to be required for normal polarization of Flamingo. Diego is itself localized at proximal-distal boundaries in response to Frizzled signaling. In the wing, proximal-distal localization of Diego and Flamingo proteins is interdependent. Perturbing the distribution of either protein disrupts that of the other. In the absence of Flamingo, Diego fails to localize to the cortex at all. However, in diego mutant cells, Flamingo is cortically localized but fails to polarize efficiently or accurately along the proximal-distal axis. The localization of each of these proteins can also be perturbed by overexpression of the other. When Diego is overexpressed, it accumulates in large nonpolarized junctional clusters that recruit high levels of Flamingo protein as well. In contrast, overexpression of Flamingo has a dispersive rather than a clustering effect on Diego in the same cell, causing it to delocalize from proximal-distal boundaries and assume a more uniform cortical distribution. Furthermore, in cells adjacent to Flamingo overexpressers, Diego is dominantly relocalized with Flamingo to face the overexpression domain. Clearly, Flamingo and Diego each affect the localization of the other (Feiguin, 2001).
It is suggested that Diego associates with the cortical membrane and helps to polarize the cell in response to Frizzled signaling. Diego's cortical localization and ability to cluster make this model more likely than one in which Diego promotes polarized delivery of the Flamingo protein from the Golgi apparatus. How might Diego and Flamingo polarize in response to Frizzled? Although one possibility is that Diego mediates polarization by directly interacting with components of the activated Frizzled pathway, the interdependence of Flamingo and Diego localization speaks against such a linear model. A more reasonable model is that localized Frizzled signaling might modify the properties of the Diego protein at proximal-distal boundaries, increasing its ability to promote clustering of itself and Flamingo. This initial bias in localization caused by clustering might then be potentiated via homophilic interactions between Flamingo molecules on adjacent cells. This model is consistent with the reciprocal effects that Flamingo and Diego have on each other's localization (Feiguin, 2001).
The ability of polarization signals to propagate from cell to cell is one of the most intriguing aspects of tissue polarity and is one of the least understood. It has been known for some time that locally perturbing Frizzled signaling has long-range effects on the polarity of the surrounding tissue. This repolarization is reflected, at the molecular level, by the altered distribution of Flamingo and Diego. Interestingly, in the absence of any direct disruption of Frizzled, Flamingo overexpression causes nonautonomous repolarization of Diego and Flamingo that propagates over several cell diameters. Both Flamingo and Diego accumulate ectopically to a high level at the interface between frizzled mutant and wild-type tissue. The data suggest that the accumulation of Diego and Flamingo at frizzled clonal boundaries might play an important role in propagating repolarization of these proteins into the surrounding tissue (Feiguin, 2001).
How might this propagation occur? Although it is easy to understand how Flamingo-mediated homophilic interactions could recruit Flamingo and Diego from a directly adjacent cell to the boundary with the overexpression domain, it is less clear how this would result in a corresponding accumulation on the opposite side of the cell. One possibility is that when Diego and Flamingo are polarized via Flamingo-mediated homophilic interactions, a locally acting intracellular signal is generated that prevents similar interactions nearby. Flamingo and Diego might then tend to accumulate on the opposite side of the cell, where such interactions were not discouraged. Increased levels of Diego and Flamingo on the opposite side of the cell might, in turn, recruit Diego and Flamingo from the next cell in line. Such a mechanism might be sufficient to propagate polarization of these proteins over many cell diameters (Feiguin, 2001).
Acquisition of planar cell polarity (PCP) in epithelia involves intercellular communication, during which cells align their polarity with that of their neighbors. The transmembrane proteins Frizzled (Fz) and Van Gogh (Vang) are essential components of the intercellular communication mechanism, as loss of either strongly perturbs the polarity of neighboring cells. How Fz and Vang communicate polarity information between neighboring cells is poorly understood. The atypical cadherin, Flamingo (Fmi), is implicated in this process, yet whether Fmi acts permissively as a scaffold or instructively as a signal is unclear. This study provides evidence that Fmi functions instructively to mediate Fz-Vang intercellular signal relay, recruiting Fz and Vang to opposite sides of cell boundaries. It is proposed that two functional forms of Fmi, one of which is induced by and physically interacts with Fz, bind each other to create cadherin homodimers that signal bidirectionally and asymmetrically, instructing unequal responses in adjacent cell membranes to establish molecular asymmetry (Chen, 2008).
Nonclassical cadherins generally exhibit weak homophilic binding in vitro, raising the possibility that they regulate signaling rather than adhesion. Moreover, after cell-cell recognition, cadherins are thought to function either homophilically and symmetrically or heterophilically and asymmetrically between cells. This study shows that the atypical cadherin Fmi acts homophilically to communicate PCP signals between neighboring cells, yet its action is asymmetric, serving to link the accumulation of Fz on one cell boundary with Vang on the adjacent cell boundary, and vice versa. These data lead to the proposal of a model in which Fmi exists in two functional forms on opposite sides of intercellular borders, one of which selectively and cell-autonomously interacts with Fz (F-Fmi), and the other with Vang (V-Fmi). The native form of Fmi is V-Fmi, but upon interaction with Fz, V-Fmi is converted to F-Fmi. It is inferred that Fmi homodimers consist preferentially of opposite forms, thereby producing asymmetric function of the complex. By virtue of this mechanism, Fmi-mediated intercellular signaling communicates information about PCP protein asymmetry between neighboring cells (Chen, 2008).
How might Fmi achieve its homophilic yet asymmetric function? Although two splice forms exist, a single form can fulfill both V- and F-Fmi functions. A second possibility is that different stoichiometries of Fmi interact on opposite sides of the boundary -- for example, cis-dimers of Fmi might behave as V-Fmi whereas monomers function as F-Fmi. However, a reproducible proximal-distal difference in levels of tagged Fmi have not been detected when expressed in a mosaic pattern. A third possibility is that posttranslational regulation results in two distinct forms of Fmi that are selectively recruited or retained, directly or indirectly, by Fz or Vang. A fourth model is that V-Fmi and F-Fmi are alternate conformers or modified forms of Fmi where conversion of V-Fmi to F-Fmi depends on interaction with Fz. Although it is not possible to distinguish between the latter three possibilities, the finding that Fz and Fmi directly interact favors models in which Fz physically alters the properties of V-Fmi, thereby inducing the F-Fmi form. Extensive evidence shows that interacting proteins can modify the activity of cadherins. Of note, Xenopus Fz7, which mediates convergent extension during gastrulation, has been reported to directly bind a protocadherin through its extracellular domain. Detailed molecular and structural studies will be required to determine the precise nature of V-Fmi and F-Fmi and how they interact with Fz and Vang (Chen, 2008).
During PCP signaling, cells each receive a signal that orients polarization. Cells then consolidate this information by amplifying the asymmetry in a process that involves communicating and aligning polarity with surrounding cells. By signaling to a neighbor that a given cell boundary is enriched for either Fz or Vang, asymmetric Fmi homodimers transmit this information bidirectionally between cells. In the wild-type, amplification through feedback control is required to produce sharp differences between Fz and Vang levels on adjacent cell surfaces. Pk, Dgo, and Dsh are required for this amplification, though they are not required for intercellular signaling per se. The mechanism by which this amplification occurs is unknown, but the result is a mutual exclusion of Fz and Vang from a given region of the cell surface. Fmi therefore serves to link the action of feedback loops in neighboring cells, assuring a coordinated polarization (Chen, 2008).
It has been proposed that asymmetric placement of core PCP proteins is itself the signal that controls morphological polarization. However, an alternative model has been proposed in which the absolute, scalar value of Fz activity within each cell varies in a gradient across the tissue in response to an unidentified ligand. Scalar Fz levels are proposed to be refined by an averaging process between neighboring cells. According to this view, asymmetric PCP protein localization is only an epiphenomenon and is not required for function. It is suggested that several observations are inconsistent with this model. (1) The model predicts that the extent of domineering nonautonomy near fz mutant clones should vary according to position in the Fz activity gradient. However, its extent was reported to be equal throughout the abdomen, and no proximal-distal difference is evident in the fly wing. (2) FzΔCRD rescues polarity in fz null mutant flies despite deletion of the CRD, leaving little protein to which an extracellular ligand might bind (Chen, 2008).
In essence, the scalar Fz model argues for a Fz gradient across the tissue, whereas the asymmetric protein localization model invokes gradients of core PCP protein localization or activity within each cell but not across the tissue. As an additional test to distinguish between these two models, small wild-type islands of ~20 cells in size surrounded by fmi mutant cells were exsmained. These wild-type cells are prevented from communicating with and receive no repolarizing signal from the surrounding fmi mutant cells. The scalar Fz model predicts that these small wild-type islands should still be directly responsive to the proposed morphogen gradient and should therefore generate a normal Fz activity slope, resulting in normal polarity. In contrast, because the asymmetric PCP protein model posits only subcellular gradients of PCP protein activity but no tissue level gradient of Fz or other core PCP protein activity, each cell's tendency to align polarity with its neighbors could lead to other patterns of local alignment. Consistent with this latter possibility, prehairs in many of these islands exhibited PCP defects and formed swirling patterns. Because there is no discontinuity as one follows the polarity of cells in these islands, the scalar Fz activity model cannot accommodate this result without invoking an Escher's staircase of infinitely rising Fz levels. In contrast, the asymmetric protein model easily explains this result by organizing proximal and distal PCP protein domains in a spoke-like pattern between the cells of the swirl, as is indeed observed. In light of the evidence presented in this study that Fmi homodimers can instructively generate asymmetric Fz and Vang localization and locally align the polarity of neighboring cells, a model is favored in which instructive protein localization mediated by Fmi homodimers is itself the signal that transmits PCP information between cells. It is thought that this is the only known example of a cadherin homodimer providing dissimilar signals across intercellular boundaries (Chen, 2008).
The Frizzled (Fz) receptor is required cell autonomously in Wnt/β-catenin and planar cell polarity (PCP) signaling. In addition to these requirements, Fz acts nonautonomously during PCP establishment: wild-type cells surrounding fz- patches reorient toward the fz- cells. The molecular mechanism(s) of nonautonomous Fz signaling are unknown. Un vivo studies identify the extracellular domain (ECD) of Fz, in particular its CRD (cysteine rich domain), as critical for nonautonomous Fz-PCP activity. Importantly, biochemical and physical interactions have been demonstrated between the FzECD and the transmembrane protein Van Gogh/Strabismus (Vang/Stbm). This function precedes cell-autonomous interactions and visible asymmetric PCP factor localization. The data suggest that Vang/Stbm can act as a FzECD receptor, allowing cells to sense Fz activity/levels of their neighbors. Thus, direct Fz-Vang/Stbm interactions represent an intriguing mechanism that may account for the global orientation of cells within the plane of their epithelial field (Wu, 2008).
The data suggest that the Fz ECD, including the CRD, acts as a Vang/Stbm ligand in nonautonomous signaling. How do these data and interpretations fit with other existing results and models? The Fz CRD is clearly dispensable for canonical Wg signaling in vivo. Previous studies have shown that it is essential for PCP signaling; however, the specifics of when and where have been controversial. A recent paper suggests that the CRD is not strictly required for PCP establishment, as FzΔCRD can partially rescue fz mutant phenotypes in the wing. However, some PCP defects remain. It is also worth noting that in experiments where the multiple wing hair phenotype of fz wings were assayed, this serves as a marker for late stage cell-autonomous Fz functions and does not address whether FzΔCRD is functional in intercellular nonautonomous communication. The experiments indicate that FzΔCRD does not fully rescue fz PCP phenotypes and does not affect domineering nonautonomy of fz mutant clones. Thus, it is concluded that the CRD of Fz is necessary for cells to send polarizing signals to neighboring cells (Wu, 2008).
Genetic and physical interaction data suggest that at the early PCP signaling stage (14-24 hr APF), Vang/Stbm functions as a receptor for FzCRD. As such, it would appear that Vang/Stbm senses how much Fz (activity) is present on adjacent cells and relays this information, causing a cell to orient toward the neighboring cell with lower Fz level/activity. The conclusion that Fz 'signals' and Vang 'receives the signal' is consistent with previous models, in that a fz− cell at the clone boundary will orient toward the center of the mutant area as it compares levels of its two neighboring cells (one of which is the wild-type cell adjacent to the clone). Similarly, it has been shown that Vang is not needed in the 'sending' cell, which is consistent with the result that fz− Vang− double mutant clones behave like fz− clones. How do these observations fit with the nonautonomous behavior of Vang− clones? In Vang− mutant cells, all Fz protein accumulates at the membranes abutting wild-type cells; wild-type cells at the clonal border would therefore presumably detect more Fz in Vang− cells. The model would predict that this relocalization of Fz causes these cells to orient away from the Vang− neighbors. This interpretation is also consistent with the Vang− phenotype being suppressed in fz− Vang− double mutant clones, suggesting that the nonautonomous effect is mediated largely through Fz (Wu, 2008).
Fz-Vang/Stbm interactions are dependent on the presence of the Fmi (also known as Stan) protein but are independent of the core PCP factor Dsh. Similarly, they are independent of Pk, which mediates the cell-autonomous requirement for Vang/Stbm. It is important to note that Vang/stbm mutants affect fz− nonautonomy differently from pk mutants: Vang/stbm− backgrounds suppress the domineering nonautonomy of fz− clones (consistent with the model), whereas pk− mutants enhance the nonautonomous effects. These data suggest that during early nonautonomous PCP signaling, Fz-Vang/Stbm effects are independent of Pk. It is thus likely that there are two distinguishable phases of Fz-Vang/Stbm interactions: the nonautonomous phase addressed here (14-24 hr APF and a later autonomous phase involving Dsh and Pk (Wu, 2008).
The simplest interpretation of the data suggests signaling from Fz to Vang/Stbm during nonautonomous signaling. It cannot excluded, however, that the interaction is bidirectional and that Fz activity is also influenced by binding to Vang/Stbm (in a Dsh-independent manner). Nevertheless, comparing the gain-of-function data of Fz and Vang/Stbm, the effects of Fz in repolarizing neighboring cells are always robust, whereas those with Vang/Stbm are milder and more cell autonomous. Despite this observation, bidirectional signaling is possible, either through the Fz-Vang/Stbm interaction or through their links to the atypical cadherin Fmi as suggested in several models. The data do not exclude an instructive role for a Fmi-Fmi interaction as proposed earlier. Indeed, this latter idea is supported by the observation that nonautonomous fz− clonal phenotypes are not completely suppressed in a Vang− mutant background, as some nonautonomy is still observed (~25% of fz− clones still display weak nonautonomy in a Vang− background) (Wu, 2008).
Fmi has recently been shown to associate with Fz, and thus the homophilic cell adhesion behavior of Fmi could also contribute to an instructive directional signal. The observations that Fz can associate extracellularly with Vang/Stbm (this work) and within the membrane with Fmi suggest a complex scenario. A cross-cell interaction mediated by the homophilic Fmi interaction could display asymmetric properties, as Fmi-Fz and/or Fmi-Vang complexes could have different qualities and signal in either direction. However, fmi null clones show little nonautonomous behavior (a 1 cell wide effect), while the fz− and fz− Vang− clones with widespread nonautonomy are nevertheless striking. Fmi causes significant nonautonomy when overexpressed, and this effect seems not to depend on the presence of Fz or Vang in the overexpressing clone. Multiple parallel mechanisms are thus likely to exist that contribute to cell-cell communication in transmitting the polarity signal (Wu, 2008).
In conclusion, this study provides molecular evidence for a mechanism of nonautonomous Fz signaling through direct interactions with Vang/Stbm on neighboring cells. It remains unclear how the levels of the initial Fz-Vang/Stbm interaction are established in wild-type. Both Fz and Vang/Stbm are expressed evenly and their initial subcellular localization is not polarized. Thus, in wild-type, the generation of a polarized Fz-Vang/Stbm interaction (across cells) must be mediated by other factors that modify Fz, Vang/Stbm, or their interaction in a graded manner (Wu, 2008).
The planar polarization of developing tissues is controlled by a conserved set of core planar polarity proteins. In the Drosophila pupal wing, these proteins adopt distinct proximal and distal localizations in apicolateral junctions that act as subcellular polarity cues to control morphological events. The core polarity protein Flamingo (Fmi) localizes to both proximal and distal cell boundaries and is known to have asymmetric activity, but the molecular basis of this asymmetric activity is unknown. This study examined the role of Fmi in controlling asymmetric localization of polarity proteins in pupal wing cells. Fmi was found to interact preferentially with distal-complex components, rather than with proximal components, and evidence is presented that there are different domain requirements for Fmi to associate with distal and proximal components. Distally and proximally localized proteins cooperate to allow stable accumulation of Fmi at apicolateral junctions, and evidence is presented that the rates of endocytic trafficking of Fmi are increased when Fmi is not in a stable asymmetric complex. Finally, evidence is provided that Fmi is trafficked from junctions via both Dishevelled-dependent and Dishevelled-independent mechanisms. A model is presented in which the primary function of Fmi is to participate in the formation of inherently stable asymmetric junctional complexes: Removal from junctions of Fmi that is not in stable complexes, combined with directional trafficking of Frizzled and Fmi to the distal cell edge, drives the establishment of cellular asymmetry (Strutt, 2009).
The differing ability of overexpressed Fmi to modulate Fz:Dsh and Stbm:Pk levels at junctions could be explained by a number of mechanisms. One likely hypothesis is that Fmi may require a cofactor for a robust interaction with Stbm, and that this cofactor is limiting when Fmi is overexpressed. Alternatively, Fmi may require posttranslational modification or a conformational change to interact with Stbm, and a factor needed for this modification is limiting. The cytoplasmic C-terminal tail of Fmi is a likely region to mediate an interaction with Fz:Dsh or Stbm:Pk; therefore, a truncated form of Fmi was constructed, in which this region is either absent or replaced with GFP (Strutt, 2009).
When overexpressed in pupal wing cells, FmideltaIntra is much more efficient at recruiting Fz and Dsh to junctions than full-length Fmi, an effect similar to that caused by removal of stbm or pk. Stbm is still reduced at junctions, although less than when full-length Fmi is overexpressed. This suggests that the C-terminal intracellular domain of Fmi is dispensible for the interaction of Fmi with Fz:Dsh and, importantly, that Fz:Dsh no longer have to compete with Stbm:Pk for access to Fmi (Strutt, 2009).
Interestingly, two isoforms of Fmi have been identified, one of which contains a PDZ binding motif (PBM) at its C terminus. It is possible that loss of the PBM alone could account for the failure of overexpressed Fmi or FmideltaIntra to associate with Stbm:Pk. However, this is unlikely, because Fmi that lacks the PBM can rescue the planar polarity phenotype of fmi mutants (Strutt, 2009).
Endogenous Fmi is thought to be localized on both proximal and distal cell boundaries. This was confirmed by expressing CFP-tagged Fmi at physiological levels in clones in pupal wings, and it was observed that levels of staining appear similar at each end of the cell, consistent with the homophilic-interaction model. Notably, expression of a GFP-tagged form of FmideltaIntra results in its preferential localization to distal cell edges, where Fz and Dsh also localize (Strutt, 2009).
Interestingly, junctional localization of FmideltaIntra-EGFP is not dependent on endogenous, full-length Fmi, suggesting that this molecule is still able to participate in homophilic interactions. Hence, the ability of FmideltaIntra-EGFP to functionally rescue the polarity phenotype of fmi null mutant clones was investigated. If FmideltaIntra-EGFP interacts preferentially with the distal Fz:Dsh complex, then Stbm recruitment to junctions inside clones would be compromised. Consequently, FmideltaIntra-EGFP:Fz complexes inside the clone would preferentially interact with Fmi:Stbm outside the clone, leading to a reversal in polarity on proximal clone edges. Importantly, this prediction is upheld, and fmi clones rescued with FmideltaIntra-EGFP exhibit weak proximal polarity inversions, such that trichomes point away from the clone, and polarity proteins are recruited to the clone boundary (Strutt, 2009).
Nevertheless, Stbm localizes asymmetrically inside the clone, although not always at the correct site, whereas in a fmi null mutant it lacks any asymmetric localization. Thus, FmideltaIntra-EGFP must retain some ability to interact with Stbm. To confirm this, the ability of full-length Fmi or FmideltaIntra-EGFP to interact with Fz and Stbm in Drosophila S2 cells was analyzed. In this assay, Fmi and FmideltaIntra-EGFP are recruited to sites of cell contact, as a result of homophilic interactions between their extracellular domains. Cotransfection of Fz or Stbm with either full-length Fmi or FmideltaIntra-EGFP in Drosophila S2 cells results in the recruitment of both to sites of cell contact (Strutt, 2009).
Interestingly, if S2 cells were transfected with either Fz or Stbm and then mixed, weak recruitment is also observed to sites of cell contact, arguing that their extracellular domains can interact independently of Fmi. Nevertheless, recruitment was weaker and less frequent than when Fmi was cotransfected, suggesting that Fmi:Fmi interactions are more important than Fz:Stbm interactions in stabilizing complexes between adjacent cells (Strutt, 2009).
The data suggest that Fz:Dsh and Stbm:Pk complexes differ in their ability to associate with Fmi. Whereas endogenous levels of Fmi result in the formation of asymmetric complexes with Fz:Dsh on one side of the boundary and Stbm:Pk on the other, overexpressing Fmi favors Fz:Dsh recruitment. Furthermore, a C-terminally deleted form of Fmi preferentially localizes distally with Fz, and overexpression of this form has an even greater preference for Fz:Dsh recruitment. Thus, the C terminus of Fmi is important in promoting the interaction with Stbm:Pk. The Fmi truncation data could be explained simply by the possibility that the C terminus of Fmi contains a direct binding site for Stbm; however, this fails to explain why overexpressed full-length Fmi prefers to recruit Fz:Dsh. It is therefore proposed that the association of Fmi with Stbm:Pk requires a limiting factor that is saturated by Fmi overexpression. The most plausible hypothesis is a requirement for a cofactor for Stbm:Pk binding, but other possibilities include saturation of the machinery for a posttranslational modification or a conformational change in Fmi (Strutt, 2009).
The data also suggest that Fmi itself needs to associate with both proximal and distal components in order to be stably localized to apicolateral junctions. Although it can form homophilic dimers between adjacent cell membranes in tissue culture, in pupal wings Fmi does not localize strongly to apical junctions and presumably fails to form stable homodimers in trans. Fz on one side of the junction and Stbm:Pk on the opposite side stabilize Fmi at junctions, most likely by promoting homophilic interactions or preventing internalization. However, Fmi appears to be capable of forming complexes with either distal or proximal components alone, but these complexes (particularly the proximal complex) are apparently less stable at junctions. Taken together with overexpression experiments, this would suggest that the most stable configuration is Fz:Fmi on one side of the boundary and Fmi*:Stbm:Pk on the other (where Fmi* denotes the modified form able to preferentially associate with Stbm:Pk) (Strutt, 2009).
In order for an asymmetric complex to be stabilized across junctions, the extracellular domains must somehow 'look' different. One possibility is that the Fz and Stbm extracellular loops interact - a view supported by S2 cell data. Alternatively, the Fmi extracellular domain, when associated with either Fz or Stbm:Pk, could undergo a conformational change that promotes homophilic Fmi interactions (Strutt, 2009).
An intriguing question is why clones of cells that overexpress Fmi behave like fz loss-of-function clones. It is suggested that within the clones, excess Fmi associates with the entire available pools of both Fz and Stbm. However, there is still a pool of uncomplexed Fmi that can associate with Fmi:Fz in adjacent wild-type cells, forming the relatively stable Fmi-Fmi:Fz configuration, thus causing polarity to be reversed on distal clone boundaries. In support of this model, an identical nonautonomous effect is seen when FmideltaIntra is overexpressed, which itself interacts only poorly with Stbm but presumably can interact with Fmi:Fz in adjacent cells outside the clone (Strutt, 2009).
Interestingly, Fmi accumulates in excess at junctions in a dsh, stbm double mutant, whereas Fz does not. Thus, although Fz acts to stabilize Fmi at junctions, Fmi does not always need to associate with Fz in a stoichiometric fashion in order to be stabilized. Perhaps as long as there is some Fz associated with Fmi, this may permit local stabilization of other Fmi molecules in cis. Alternatively, this excess accumulation of Fmi might simply represent 'unstable' Fmi homodimers that are no longer being removed from junctions by the actions of Dsh and Stbm (Strutt, 2009).
The composition of the complex with which Fmi is associated appears to be critical for determining the frequency and manner by which Fmi is turned over from the plasma membrane. Most compellingly, Fmi accumulates more strongly in an enlarged endosomal compartment in Rab7TN mutant tissue when stbm and fz are absent than when they are present. Thus, it is suggested that more Fmi is resident in the endocytic pathway when it is unable to form stable asymmetric complexes. Fmi:Fz puncta have been observed that are selectively trafficked to distal cell edges. In the current experiments, these puncta colocalize with YFP-Rab4, suggesting that Fmi and Fz are recycled back to the plasma membrane by a Rab4-dependent mechanism. Furthermore, the increased intracellular and junctional levels of Fz and Fmi in dor mutant clones suggests that in addition to being recycled to the plasma membrane, a significant fraction of internalized Fmi and Fz is also sent for degradation. It is formally possible that the intracellular accumulation of Fmi and Fz seen when lysosomal trafficking is blocked by loss of Rab7 or in dor clones is due to their being sent for degradation immediately after synthesis (e.g., if damaged or misfolded); however this is unlikely because newly synthesized Fmi-ECFP appears first at junctions before been seen in puncta (Strutt, 2009).
Stbm has not been observed in large intracellular puncta, but it seems likely that it is also internalized and recycled, possibly together with Fmi, although it must do so by alternative pathways involving smaller or more rapidly recycling particles that are not visible by confocal microscopy. Indeed, the data suggest a potential role for Dsh and Stbm in regulating junctional levels of Fmi. A stbm mutant alone results in a loss of Fmi from junctions, consistent with a need for Stbm in stabilizing Fmi in asymmetric complexes. In contrast, loss of Dsh and Stbm together increases Fmi levels at junctions, suggesting a role for Stbm in internalization. It is suggested that the outcome of any interaction of Stbm with Fmi is dependent upon whether Fmi is able to form stable homodimers with Fz on the opposite cell membrane. In a wild-type situation, one could envisage that Fmi forms stable homodimers in a Fz:Fmi-Fmi*:Stbm configuration, and that both Dsh and Stbm promote internalization of any Fmi that is not in this configuration, the majority of which is subsequently recycled back to the plasma membrane. In dsh mutants, there is reduced internalization, but the effect on Fmi levels is subtle; Fz and Stbm are still present to promote Fmi homodimer formation, and Stbm still promotes internalization of any unstable Fmi. In contrast, in stbm mutants, the number of less stable Fmi complexes (associating only with Fz) is greatly increased, favoring internalization by Dsh. Finally in dsh, stbm double mutants, Fmi is again less stable (associating only with Fz), but there is no Dsh- or Stbm-mediated internalization, leading to an overall increase of Fmi at junctions (Strutt, 2009).
How do Dsh and Stbm regulate Fmi levels at junctions? Stbm contains potential interaction motifs for the endocytic adaptor AP2, but their role has not been functionally tested. In addition, in vertebrate Wnt signaling, there is evidence that Dsh interacts with the endocytic adaptor protein β-arrestin and mu2 subunit of AP2 to mediate Wnt/Fz endocytosis and downregulation of Wnt signaling. Interestingly, in planar polarity this is no evidence that Dsh directly mediates internalization of Fz, but the data rather point to Dsh promoting Fmi internalization when it is not associated with Fz. Instead, the trafficking of Fmi together with Fz into the lysosomal pathway is Dsh independent (Strutt, 2009).
In summary, it is proposed that a number of mechanisms exist by which Fmi contributes to the generation of asymmetry at the molecular level. First, the characterization of the previously inferred asymmetry in Fmi activity indicates that Fmi normally prefers to bind to Fz and requires a limiting factor for association with Stbm:Pk. Second, Fmi stability at junctions is dependent on both Fz and Stbm:Pk, with the most stable form being Fz:Fmi bound to Fmi*:Stbm. Finally, it is proposed that entry of Fmi into the endocytic trafficking pathway is decreased if it is in a stable complex, and this is regulated either by Dsh and Stbm or independently of Dsh and Stbm, depending on whether it is associated with Fz (Strutt, 2009).
An outstanding question is how these mechanisms translate into cellular asymmetry, such that in any particular cell, heterophilic polarity complexes preferentially form with Fz:Dsh at the distal junctions, rather than having heterophilic complexes in both orientations. It is thought that the acquisition of cellular asymmetry is likely to be driven by directional trafficking of Fmi:Fz, although other models, such as a mechanism for preferential stabilization of Fmi:Fz interactions at the distal cell edge, are also possible. In addition, it seems likely that an amplification mechanism would be required, although the molecular mechanisms remain to be elucidated (Strutt, 2009).
While this manuscript was in preparation, another manuscript was published, in which Fmi was proposed to mediate an asymmetric and instructive signal between proximal and distal complexes to generate asymmetry, and thus does not act merely as a scaffold for Fz:Stbm interactions across membranes. It is argued that the current data do not provide evidence for a specific signaling function of Fmi. Instead, the hypothesis is favored that the composition of the proximal and distal complexes is distinct, and that heterophilic complexes are inherently more stable than homophilic complexes. Together, removal of unstable nonasymmetric complexes through increased endocytic turnover, in concert with directional trafficking and an unknown amplification mechanism, may be sufficient to generate asymmetry without the need to invoke a specific signaling function for any components of the complexes (Strutt, 2009).
Planar cell polarity (PCP) refers to a second polarity axis orthogonal to the apicobasal axis in the plane of the epithelium. The molecular link between apicobasal polarity and PCP is largely unknown. During Drosophila eye development, differentiated photoreceptors form clusters that rotate independently of the surrounding interommatidial cells (ICs). This study demonstrates that both Echinoid (Ed), an adherens junction-associated cell adhesion molecule, and Flamingo (Fmi), a PCP determinant, are endocytosed via a clathrin-mediated pathway in ICs. Interestingly, it was found that Ed binds AP-2, an adaptor that acts between cargo receptors and clathrin-coated vesicles during endocytosis, and is required for the internalization of Fmi into ICs. Loss of ed led to increased amounts of Fmi on the cell membrane of non-rotating ICs and also to the misrotation of photoreceptor clusters. Importantly, overexpression of fmi in ICs alone was sufficient to cause misrotation of the adjacent photoreceptor clusters. Together, it is proposed that Ed, when internalized by AP-2, undergoes co-endocytosis with, and thereby decreases, Fmi levels on non-rotating ICs to permit correct rotation of ommatidial clusters. Thus, co-endocytosis of Ed and Fmi provides a link between apicobasal polarity and PCP (Ho, 2010).
This study found that Ed interacts with AP-2 and promotes Fmi internalization via a clathrin-dependent pathway in ICs. Moreover, loss of ed leads to accumulation of Fmi (and of several receptors/CAMs of AJs) on the membrane of these non-rotating ICs. This, together with the observation that overexpression of fmi in the non-rotating ICs is sufficient to cause photoreceptor misrotation, led to a proposal of an Ed-mediated co-endocytosis model to explain the rotation defects associated with the ed mutant tissue. Thus, the homophilic CAMs, Ed and Fmi, play crucial roles in ICs to allow coordinated rotation of ommatidial clusters (Ho, 2010).
Ed specifically regulates Fmi endocytosis in ICs. Although Fmi was previously shown to be endocytosed in the photoreceptor cluster, it is argued that Fmi levels in the rotating photoreceptor clusters are regulated by an Ed-independent mechanism. First, Fmi levels in the photoreceptor clusters are not affected in the ed mutant clones. Second, the Fmi distribution pattern in R3/R4 is largely unchanged even when Ed-GFP is overexpressed in photoreceptors by Elav-Gal4 to mimic the ed-expressing ICs. It was also demonstrated that Ed levels in ICs, but not in photoreceptor clusters, are regulated via an AP-2-dependent endocytic pathway. It remains unknown how Ed is downregulated in the photoreceptor cluster. Egfr signaling has been proposed to regulate the morphological and adhesive changes of cells within the photoreceptor cluster, and it is possible that Egfr signaling directly or indirectly downregulates the Ed levels in R8/R2/R5 and later in R3/R4 (Ho, 2010).
Interestingly, ed affects the levels of Fmi, DE-cad and Egfr, but not of Dlg. Therefore, Ed seems to only affect receptors/CAMs at AJs. One intriguing possibility is that Ed, via its interaction with AP-2, triggers the co-endocytosis of most, if not all, of the receptor/CAM at AJs. Although Ed has been shown to associate with Egfr, there is currently no evidence that Ed interacts directly with Fmi. Thus, Ed might undergo co-endocytosis either directly or indirectly. Although Ed contains five putative protein-sorting motifs, it is not the only molecule with a protein-sorting motif: Egfr, for example, also contains the YXXphi signal (phi is a bulky hydrophobic amino acid which could potentially interact with the AP-50 subunit of the AP-2 complex). It is possible that different receptors/CAMs might cooperate, via their interaction with AP-2 or other adaptors, to promote the co-endocytosis of other receptors/CAMs (Ho, 2010).
Because Ed facilitates the endocytosis of many receptors/CAMs of AJs, although to different extents, multiple functions of ed in the eye disc would be expected. Indeed, ed plays crucial roles in PCP (this study) and in Egfr signaling during eye development. It was previously shown that loss of ed leads to sustained MAPK (Rolled - FlyBase) activation only in cells of the proneural clusters and over several rows. This is consistent with the observation that Egfr is upregulated in an enlarged group of arc cells that contains two R8 photoreceptors as well as in cells of developing ommatidia up to two rows posterior to the MF. Thus, it is plausible that Ed was co-endocytosed with Egfr in the proneural clusters to downregulate Egfr activity within these cells and thus ensure that only one R8 is selected from the two to three R8 cell-equivalent group. When ed is absent, Egfr cannot be internalized efficiently and therefore persists on the membrane to cause sustained MAPK activation and multiple R8 selection. Although the level of Egfr is also upregulated in cells more posterior to the MF, these levels might not be high enough to cause sustained MAPK activation. In the wing disc, Ed also facilitates Notch signaling to promote mesothorax bristle patterning. In fact, Ed colocalizes with Notch/Delta in Hrs-containing early endosomes. However, it remains unclear whether the endocytosis of Ed plays any role in facilitating Notch signaling in the wing discs (Ho, 2010).
It has been shown that each photoreceptor cluster, as a group, moves independently of the adjacent ICs. Most rotation-specific genes identified thus far have been proposed to function mainly in the rotating clusters to modulate rotation. This study provides evidence that Ed plays crucial roles in the ICs to modulate ommatidial rotation. It is proposed that Ed, via co-endocytosis, reduces the level of Fmi on the non-rotating ICs to prevent homotypic interactions with the enriched Fmi on the rotating cluster. This allows free and coordinated rotation of photoreceptor clusters, a process regulated by effectors such as Zipper and Nemo. It is reasoned that in the absence of ed, as seen in ed mutant clones, the upregulated Fmi on the non-rotating ICs might affect the free rotation of ommatidial clusters not only within the ed clone, but also in the adjacent wild-type clusters abutting the ed clones. This might contribute, at least in part, to the non-autonomous effect of ed on ommatidial rotation. The dynamic and differential expression of Ed (and of its paralog Friend of Echinoid) in the rotating clusters and non-rotating ICs has also been proposed to modulate ommatidial rotation (Fetting, 2009). Thus, differential expression of Ed, Fmi, DE-cad and Friend of Echinoid in the rotating clusters and non-rotating ICs prevents the homotypic interaction of these four CAMs to allow free rotation of photoreceptor clusters. The largely complementary expression pattern between Ed and DE-cad/Arm in a photoreceptor cluster is similar to that observed during the generation of ed mutant clones in the wing discs, where Ed-non-expressing cells accumulate high levels of DE-cad/Arm and sort out from the surrounding Ed-expressing cells. Thus, cell sorting-like behavior of a photoreceptor cluster, mediated by differential expression of Ed and DE-cad/Arm, might help photoreceptors in the cluster to rotate as a group (Ho, 2010).
Fetting (2009) recently showed that ed genetically interacts with Egfr pathway members, and proposed that ed, via inhibiting Egfr signaling in the photoreceptors, regulates ommatidial rotation. However, this study demonstrated that, after row 2, Ed facilitates the endocytosis of Egfr only in the non-rotating ICs, but not in the photoreceptor clusters. Thus, if Ed indeed inhibits Egfr signaling in the photoreceptors as suggested, it probably employs mechanisms other than to reduce the levels of Egfr on the photoreceptors. It is currently unknown whether the effect of Ed on Egfr levels in ICs plays any role in the modulation of ommatidial rotation. Moreover, this study found that in the absence of ed, not only Fmi, Fz and Dsh (the R3-specific PCP proteins), but also Strabismus (Van Gogh) and its associated Prickle (the R4-specific PCP proteins) were all upregulated in ICs, but their enrichment at R3/R4 borders in the photoreceptor cluster was largely unaffected. As ed affects the levels of all the core PCP proteins tested in ICs, it remains unclear how ed only interacts genetically with the R3-specific PCP genes to modulate the degree of ommatidial rotation (Fetting, 2009). Finally, ed also weakly affects the initial R3/R4 specification, as a small proportion of ed mutant cells also show randomized chirality and symmetrical ommatidia. It is possible that ed might exert this effect through mechanisms other than the promotion of Fmi endocytosis. Alternatively, as overexpression of Fmi in the non-rotating ICs (generated by fmi-overexpressing clones) can affect both ommatidial rotation and reversal of R3/R4 cell fate of the adjacent clusters, it is possible that the Fmi upregulation in the ICs (generated by ed mutant clones) might also affect, to some extent, the asymmetric distribution of Fmi in R3/R4 and, thereby, the R3/R4 specification of the adjacent clusters (Ho, 2010).
Members of the Flamingo cadherin family are required in a number of different in vivo contexts of neural development. Even so, molecular identities downstream from the family have been poorly understood. This study shows that a LIM domain protein, Espinas (Esn), binds to an intracellular juxtamembrane domain of Flamingo (Fmi), and that this Fmi-Esn interplay elicits repulsion between dendritic branches of Drosophila sensory neurons. In wild-type larvae, branches of the same class IV dendritic arborization neuron achieve efficient coverage of its two-dimensional receptive field with minimum overlap with each other. However, this self-avoidance was disrupted in a fmi hypomorphic mutant, in an esn knockout homozygote, and in the fmi/esn trans-heterozygote. A functional fusion protein, Fmi:3eGFP, was localized at most of the branch tips, and in a heterologous system, assembly of Esn at cell contact sites required its LIM domain and Fmi. It was further shown that genes controlling epithelial planar cell polarity (PCP), such as Van Gogh (Vang) and RhoA, are also necessary for the self-avoidance, and that fmi genetically interacts with these loci. On the basis of these and other results, it is proposed that the Fmi-Esn complex, together with the PCP regulators and the Tricornered (Trc) signaling pathway, executes the repulsive interaction between isoneuronal dendritic branches (Matsubara, 2011).
A yeast two-hybrid screen highlighted members of the Drosophila PET-LIM domain family: Prickle (Pk), Espinas (Esn), and Testin. Pk has been well studied as one of the core group members of PCP, whereas any physiological roles of Esn were unknown. Because of prominent expression of esn in the nervous system, Esn was considered to be a strong candidate for a Fmi binder in neurons, and this protein was analyzed further. The JM domain of Fmi interacted with the LIM domain of Esn or Pk in a targeted yeast two-hybrid assay and also in a coimmunoprecipitation (co-IP) experiment using Drosophila S2 cells. JM was dissected and the N-terminal subdomain (JM-A) was shown to bound LIM much more strongly than the C-terminal JM-B. Although the detection of the JM-LIM interaction was straightforward, no Fmi-Esn interaction was shown by co-IP when both full-length proteins were expressed in S2 cells. A technical difficulty with the demonstration of the Fmi-Esn interaction by co-IP is anticipated, considering that Fmi is a multipass transmembrane protein of 3579 residues. Nonetheless, attempts were made to demonstrate the physical interaction between the endogenous proteins by using a larval CNS lysate as a source material. In two out of two co-IP experiments, endogenous Esn was detected in the anti-Fmi immunoprecipitates, but not in the immunoprecipitates with negative control antibodies, although the efficiency of Esn co-IP with Fmi was low. These results imply that the association of Esn to Fmi might be transient and/or dependent on the cellular milieu (Matsubara, 2011).
The Fmi-Esn interaction was further investigated in a cell adhesion assay. Fmi-expressing S2 cells adhere to each other due to Fmi-Fmi homophilic interactions, and Fmi often accumulates at sites of cell-cell contacts. This assay was applied to test whether Esn was colocalized with Fmi at the cell contact sites; however, no clear conclusions could be drawn. As an alternative approach to assess the possibility of Fmi-Esn colocalization in cells, pupal wing epithelia was used where endogenous Fmi is enriched at cell boundaries. When Esn was ectopically expressed in the wing, it was colocalized with Fmi at cell boundaries. In contrast, an Esn mutant form that lacked the LIM domain (Esn?LIM) was diffusely distributed in the cells and did not exhibit obvious concentration at the boundaries where Fmi was accumulated. Furthermore, the full-length Esn was no longer localized at the apical cell contact sites when fmi was knocked down. These results show that Esn can be localized at cell contact sites in a LIM domain- and Fmi-dependent manner (Matsubara, 2011).
This study shows both physical and genetic interactions between Fmi and Esn, and has proposed a role of the Fmi-Esn interplay in contact-dependent repulsive signaling between isoneuronal dendritic branches of class IV da neurons. The LIM domain is shared by a number of proteins and is recognized as a modular protein-binding interface. Esn belongs to the Drosophila PET-LIM domain family, where two other members are Pk and Testin. The the co-IP results raise the possibility that the binding of Esn to Fmi in da neurons might be under the control of a conformational change and/or post-translational modifications of Esn. In fact, it has been shown for human Testin (hTes) that its LIM domains mediate the intramolecular interaction (Zhong, 2009). More specifically, the C-terminal LIM interacts directly with N-terminal hTes sequences, and this intramolecular association seems to preclude the docking of partners such as α-actinin. Furthermore, there might be a regulation at the level of tethering Esn to the plasma membrane. The CAAX motif at the C terminus of Esn could be the target of such a regulation. Potential candidates that may be involved in such a hypothetical regulation of Esn would be PCP regulators such as Vang (Matsubara, 2011).
A group of key molecules play pivotal roles in dendrite self-avoidance. That said, the underlying mechanisms involving these molecules (the Hpo-Trc kinase cascade, Dscam1, Tutl, and the Fmi-Esn interplay) seem to be complex, as suggested by the fact that phenotypic analyses of these genes show similarities as well as differences in terms of class specificity and self-avoidance versus tiling. Moreover, understanding of the molecular mechanisms is currently limited. For example, unknown are the identities of interbranch signals and their sensors that are connected to the Trc pathway. Genetic interactions have been sought between trc and Dscam1 or tutl, but positive results have not been obtained. This study found an interaction between fmi and trc; between fmi and fry, which encodes an activator of Trc; and between fmi and hpo, an upstream kinase. These genetic interactions by themselves do not necessarily support the idea that Fmi acts in close proximity to Trc, Fry, or Hpo; analysis of the Fmi-Esn and Trc-Fry complexes, and verification of their cross-talk, are required at the molecular level (Matsubara, 2011).
The search for genetic interactions was targeted toward PCP genes, and several positive results were found. It should be emphasized that PCP regulators play important roles not only in epithelial sheets, but also in rearranging cell populations. In connection with the avoidance behavior of the dendrites, intriguing analogies are found in contact inhibition of locomotion of neural crest cells in Xenopus and in stereotyping a trajectory of migrating facial motor neurons in zebrafish. Vertebrate homologs of Pk, Vang, and Dsh are required for neural crest cells to initiate sustained directional migration away from the contact with another neural crest cell, although roles of Fmi homologs were not examined in this system. Importantly, the PCP regulators are localized at the site of cell contact, leading to activation of RhoA. In the context of migration of the motor neurons, homologs of Fmi and Fz act in the surrounding neuroepithelium to prevent the integration of the neurons into the neuroepithelium, thus restricting them to the correct path (see Wada, 2009). Although the mechanisms are unknown, homologs of Fmi and Fz are also proposed to act together in axonal navigation and tract formation in mouse forebrain and C. elegans nerve cord on the basis of similar mutant phenotypes (Matsubara, 2011).
It is considered likely that Fmi, Esn, Vang, and Fz are localized closely within branches and activate RhoA locally upon contact with other isoneuronal branches, redirecting those branches away from one another. However, in contrast to prominent puncta of Fmi:eGFP at most of the tips, mCherry:Esn was not necessarily enriched there and distributed more diffusely inside branches. These differential distributions do not necessarily exclude the hypothesis. As for the Trc kinase and its activator, Fry, both physical and genetic interactions are substantiated, and yet, those two molecules show distinct features of subcellular localization in da neurons (Fang, 2010). In mammals, localization of Pk1, a mammalian member of the PET-LIM domain family, is regulated by degradation that depends on the interaction of E3 ubiquitin ligases with one of the other PCP regulators, Dvl2 (Narimatsu, 2009). It could be that the contact between isoneuronal branches may trigger the Fmi-Esn interaction locally, with subsequent stabilization of the bound Esn, which is otherwise degraded. Another, mutually nonexclusive possibility would be that the association/dissociation of Esn and Fmi may be coupled to a switch of Esn signaling activity. There could be a third intriguing possibility: Esn may play a role in the trafficking of Fmi, such as promotion of endocytic recycling of Fmi. It has been proposed that Fmi is subject to high rates of endocytic turnover in the pupal wing (Strutt. 2008; Strutt, 2011), and a cytoplasmic motif of a mouse homolog of Fmi, Celsr1, governs its internalization (Devenport, 2011; Matsubara, 2011 and references therein).
If the PCP proteins are localized at contact sites at least transiently, then how do they assemble on both sides of the interface? The assembly of the PCP proteins at cell boundaries has been best studied in the Drosophila wing epithelia, where the arrangement of the proteins is asymmetrical across proximodistal (P/D) cell boundaries: Vang and Pk at the proximal cell cortex, Fz and Dsh at the distal cortex, and Fmi at both. In the dendritic arbor of a single da neuron, would it be possible that some terminal branches are Vang-rich 'proximal' types, whereas others are Fz-rich 'distal' ones? Alternatively, a single terminal branch or even a single tip could be a mosaic of Vang-rich and Fz-rich membrane domains. Coexpression of Fz and Vang with distinct tags in the same neuron and tracking pairs of terminal branches that are approaching each other at high-enough resolution may answer this question (Matsubara, 2011).
Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.