Apical Basal Cell Polarity Determinant Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Proper establishment and maintenance of apical-basal polarity is critical for normal development. Alterations of apical-basal polarity are often associated with cancer in vertebrates. In Drosophila, abnormal expression of apical-basal determinants, such as overexpression of Crumbs (Crb) or loss of Scrib, can lead to loss of cell polarity and proliferation control, which are two hallmark of cancer. Several models have been proposed to explain the overgrowth phenotypes. For example, expansion of apical domain may cause the accumulation of receptors that deregulate many growth controlling pathways and thus lead to the overgrowth phenotype. Alternatively, the polarity complex proteins may specifically modulate one or more growth control pathways (Vaccari and Bilder 2005; Hariharan and Bilder 2006). However, the pathways through which apical-basal polarity determinants affect growth remain unclear.

I specifically investigated how apical determinant Crb regulates growth. Crumbs is a transmembrane domain protein that localizes apically with Patj and stardust (sdt) to establish and maintain cell polarity. I found that Crb acts through the Hpo pathway to regulate growth. The genetic data presented below indicate Crb regulates growth and cell polarity through different motifs in its interacellular domain and identify a pathway through which Crb affects growth.


4. 1. Crumbs gain of function causes overgrowth and induces Hippo target genes expression

Overexpression of full length Crumbs (CrbFL), or a truncated version of Crb that lacks the extracellular domain (Crbintra) during wing development by using C765-gal4 results in overgrown adult wings (Figure 4.1A-C, and data not shown). Similarly, overexpression of CrbFL or Crbintra along the anterior-posterior compartment boundary by using decapentaplegic-Gal4 (dpp-Gal4) causes significant enlargement of the overexpression domain in wing discs (Figure 4.1D-G). The enlargement of the overexpression domain is associated with extra cell proliferation that is revealed by higher levels of bromodeoxyuridine (BrdU) incorporation. BrdU incorporation marks cells in S-phase of the cell cycle (Figure 4.1D,E). In contrast, cell size remains unaffected in the overexpression region. Therefore, we conclude that overexpression of Crb promotes cell proliferation in wing discs.

To gain insight into the pathway through which Crb induces overgrowth, we tested for effects on the Hippo pathway, a conserved growth control pathway that specifically regulates cell number but not cell size (Harvey and Tapon 2007; Pan 2007; Reddy and Irvine 2008). We assayed the expression of the Hippo pathway component ex using a lacZ enhancer trap insertion into the ex locus (ex-lacZ) (Hamaratoglu, Willecke et al. 2006). ex is regulated by the Hippo pathway in a negative feedback loop in multiple imaginal discs and is a widely used lacZ reporter to reveal the activity of the Hippo pathway (Hamaratoglu, Willecke et al. 2006). We found that overexpression of full length Crb or Crbintra caused strong upregulation of ex-lacZ (Figures 4.1F,G and 4.4A), similar to the effects seen with defects in Hippo signaling and Yki overexpression (Hamaratoglu, Willecke et al. 2006; Willecke, Hamaratoglu et al. 2006). We thus conclude that Crb overexpression upregulates Hippo target gene expression.

4. 2. Mutation in crumbs causes overgrowth and inhibition of Hippo pathway activity

To determine whether loss of crb also regulates growth, I characterized the phenotypes of crb mutant cells in imaginal discs and in adult tissues. In order to generate tissues nearly wholly mutant for crb, we flipped chromosomes carrying crb11A22 (the null allele) against chromosomes carrying a Minute mutation with GFP or white+ pigmented marker by using either ey-FLP or ubx-FLP. We found that crb mutant tissues, such as heads and wings, are enlarged (Figure 4.2A-D) with venation defects in the wing as was previously observed (Richardson and Pichaud 2010). To assay the effect of loss of Crb function in the regulation of cell proliferation, we analyzed the pattern of BrdU incorporation in the posterior of the eye discs. In wild-type discs, cells posterior of the morphogenetic furrow undergo an additional round of cell division, known as the second mitotic wave. After the second mitotic wave, cells cease proliferation and start to differentiate into photoreceptors. In contrast to wild-type eye discs, crb mutant cells showed ectopic incorporation of BrdU. (Figure 4.2E, arrowhead). This result suggests that Crb is required to arrest cell cycle progression in the region posterior to the morphogenic furrow. We thus conclude that Crb is required to restrict cell proliferation and maintain appropriate organ size.

The observation that Crb overexpression induces Hippo target gene expression raises the question of whether loss of crb also affects Hippo signaling. To answer this question, we monitored Hippo pathway activities by using the ex-lacZ reporter. We found that expression of the ex-lacZ reporter is autonomously upregulated in crb mutant clones. This effect was especially prominent in the hinge region of wing discs (Figure 4.2F, arrowhead). My results indicate that Crb is required for proper regulation of Hippo target genes.

Notably, the phenotypes of Crb overexpression on growth and Hippo signaling are similar but not stronger than those of crb loss of function. The similarity between the loss and gain of function phenotypes of Crb indicates that wild-type levels of Crb are essential for normal functioning of the Hippo pathway.

4. 3. Crumbs genetically interacts with Hippo pathway components

As described previously, the overgrowth phenotypes of crb mutants resemble those seen in loss of Hippo signaling. However, the crb mutant phenotypes are not as strong as those of hpo mutant clones. The difference is most evident in the pupal retina. hpo mutant retinae show a large excess of interommatidial cells (Udan, Kango-Singh et al. 2003) whereas crb mutant retinae showed no extra interommatidial cells (Figure 4.3A). The weak phenotype of crb in pupal retina is very similar to that of ft, ex, and mer (REF), components of two upstream branches of the Hippo signaling pathway. Abolishing both branches causes a stronger phenotype than depleting either single one alone. mer;fat and mer;ex double mutants show synergistic phenotypes, such as many extra interommatidial cells (Hamaratoglu, Willecke et al. 2006; Silva, Tsatskis et al. 2006; Willecke, Hamaratoglu et al. 2006) which was not observed in the single mutants. To test whether Crb acts upstream in the Hippo pathway in parallel to Mer or Fat, we examined the crb mutant pupal retinae in either Mer or Fat knocked down background. Similarly, we found that the crb mutant pupal retinae in a Mer knock down background showed extra interommatidial cells while knocking down Mer by GMR-Gal4 driven UAS-merRNAi in retinae did not result in extra interommatidial cells (Figure 4.3B). In contrast, the crb mutant pupal retinae in a Fat knocked down background do not have synergistic effects. Thus, we conclude that Crb can synergize with Mer to regulate cell number in the pupal retina.

In addition, loss of crb interacts genetically with D, an unconventional myosin that functions downstream of Fat (Mao, Rauskolb et al. 2006). Knocking down crb in the wing by nubbin-Gal4 (nub-Gal4) driven UAS-crbRNAi resulted in slightly larger wings compared to wild-type wings (Figure 4.3C,D). Overexpression of D in the developing wing caused weak overgrowth phenotypes (Figure 4.3E). Interestingly, overexpression of D in addition to knock down of crb resulted in synergistic effects and significantly overgrown wings (Figure 4.3F). We conclude that Crb genetically interacts with components in the Hippo pathway.

4. 4. Yorkie is required for Crumbs induced phenotypes

In order to more directly test the hypothesis that Crb functions through the Hippo pathway, we further investigated whether the Hippo pathway is necessary for the growth control function of Crb. We tested whether Yki is required for the overgrowth phenotype caused by Crb overexpression. Overexpressing Crbintra in the wing by nub-Gal4 causes lethality when the crosses are incubated at 25°C. We found that heterozygosity for yki rescued the lethality induced by overexpressing Crb. Similarly, when the crosses are incubated at 18°C, heterozygosity for yki suppressed the overgrowth phenotype induced by overexpressing Crb in the wing (Figure 4.3G,H). Additionally, the overgrowth phenotype and induction of ex-lacZ caused by hedgehog-Gal4 (hh-Gal4) driven Crbintra overexpression in the wing discs can be reversed by knocking down Yki via RNAi (Figure 4.3I-K). Therefore, we conclude that Yki is required for the overgrowth and Hippo pathway target gene induction caused by Crb overexpression. Thus, Crb acts upstream of Yki in the Hippo pathway to regulate growth.

4. 5. Crumbs regulates growth and cell polarity through different domains

Crb is a single-pass transmembrane protein with a relatively short intracellular domain of only 37 a.a. The extracellular domain of Crb contains 29 epidermal growth factor like repeats and 4 laminin-A globular domain-like repeats. The intracellular domain of Crb is conserved and contains two conserved protein binding motifs (Figure 4.4). The juxtamembrane motif (JM) is a FERM domain binding motif, that has been reported to bind to the FERM-domain of Yurt (Laprise, Beronja et al. 2006) and forms complexes with β-spectrin and Moesin. The C-terminal PDZ domain binding motif (PBM) has been shown to bind to Sdt and thus form a complex with Patj to regulate apical-basal polarity in various tissues, including embryonic epithelial cells, follicle cells, and pupal retina (Klebes and Knust 2000; Bachmann, Schneider et al. 2001; Hong, Stronach et al. 2001; Izaddoost, Nam et al. 2002). Crb overexpressed in the pupal retina was mislocalized throughout the cell and was sufficient to recruit Patj to the basolateral membrane. Similarly, overexpression of Crb in the embryo also caused redistribution of Sdt throughout the cell. The effects on Patj and Sdt specifically require the PBM but not the JM (Klebes and Knust 2000). To test whether Crb utilizes the same motif and the same mechanism to regulate growth and cell polarity, we quantified the overgrowth phenotypes by using ImageJ and monitored Hippo signaling activity when overexpressing Crb with the different motifs by using dpp-Gal4 (Figure 4.5A-D). The relative size of different genotypes is calculated by comparing the ratio of the expression domains area marked by GFP expression to the overall size of the discs. By statistical analysis, we found that the overexpression regions of full length Crb and Crbintra are around three fold larger than that of the corresponding area in wild-type discs (Figure 4.5F). Interestingly, mutation of the JM or removal of both motifs abrogated the growth effects, while deletion of the PBM still allowed for growth effects similar to those of intact Crbintra (Figure 4.5F).

As mentioned previously, overexpression of Crbintra caused overgrowth phenotypes and the induction of the Hippo reporter ex-lacZ (Figure 4.1G, 4.5A). Consistent with the quantification results, mutation of the JM or removal of both motifs completely abolished these effects (Figure 4.5B,D). In contrast, overexpressing Crb without the PBM still resulted in the induction of ex-lacZ and the overgrowth phenotype (Figure 4.5C).

In addition to ex-lacZ, we examined the transcriptional expression of a Hippo target gene, Diap1, by using a reporter transgene, Diap1-GFP while we overexpressed different Crb deletion constructs by engrail-Gal4 (en-Gal4) in the posterior compartment of the wing discs. We also assayed the effects of en-Gal4 driven overexpression of different Crb deletion constructs on the expression of Wingless (Wg), which is regulated by Hippo signaling in the hinge region of the wing discs. The overexpression of full length Crb and Crbintra similarly induced Diap1-GFP and Wg expression (Figure 4.6A-C,4.7A-C). However, while mutation of the JM or removal of both JM and PBM motifs abrogated the growth effects, deletion of the PBM only still retained the ability to affect growth (Figure 4.6D-F,4.7D-F).

In summary, our data show that the effects of Crb on the Hippo pathway required the JM but not the PBM. Therefore, Crb regulates growth and cell polarity through different domains and thus through different mechanisms.

4. 6. Crumbs is required for Expanded membrane localization

Crb is localized to the apical membrane where Fat, Ex, and Mer localize. Given that Crb is a transmembrane protein, it may function as a receptor of Hippo signaling. These facts raise the question of whether Crb interacts with upstream components of Hpo signaling and thus acts upstream in the Hippo pathway.

We first asked whether Fat, Ex, Mer, and Crb affect each other's localization. We found that the correct localization of Crb is not affected in fat, ex, and mer mutant cells (Figure 4.8A-C). Rather, it has been reported that ex or fat mutant cells had higher levels of Crb at the membrane (Genevet, Polesello et al. 2009; Hamaratoglu, Gajewski et al. 2009) while mer mutant cells, similar to wild-type cells, had normal amounts of Crb. (Figure 4.8C, arrowhead). However, loss of Crb leads to Ex mislocalization. In crb mutant cells, Ex was largely absent from the apical membrane and diffused into the cytoplasm at the basal lateral region (Figure 4.9A,B). When crb mutant cells were produced in a Minute background, which grows slower during development, they often did not have cytoplasmic Ex (Figure 4.9C). This observation suggests that Ex may have been degraded. Our data indicated that Crb may regulate the localization and/or stability of Ex. Interestingly, we observed that Ex is lost from the membranes of wild-type cells that are adjacent to crb mutant cells. The localization of Ex thus results in fork-like localization patterns at crb mutant clone borders (Figure 4.9C, arrowhead). Similarly, the localization of Crb as well as that of Patj, at crb mutant clone borders also formed fork-like localization patterns (Figure 4.9D,E, arrowheads). It indicates that Crb homophilically interacts with Crb molecules on neighboring cells through its extracellular domain and this interaction is required for its localization(Tanentzapf, Smith et al. 2000; Izaddoost, Nam et al. 2002). Crb may thus be required non-autonomously for Ex localization, as well as that of Patj, in neighboring cells.

In contrast, Fat and Mer are not lost from the membrane of crb mutant cells. Therefore, Crb is not required for the localization of Mer or Fat (Figures 4.10A,B, arrowheads). Crb is thus specifically required for the localization of Ex to the subapical membrane, but not other Hippo pathway components.

The requirement of Crb for Ex localization prompted the question of whether Crb overexpression is sufficient to cause the redistribution of Ex. To answer this question, we further investigated the effect of Crb overexpression on Ex localization. Full length Crb and Crbintra that are ectopically expressed in various tissues localize throughout the cell (Klebes and Knust 2000; Izaddoost, Nam et al. 2002). Because overexpressed Crb is very potent and often causes strong overgrowth phenotypes and morphological defects, it is difficult to assay protein localization of the genetically manipulated cells. To bypass this problem, we utilized temperature-sensitive Gal80 in combination with hh-Gal4 to further fine-tune its expression temporally. Crosses were kept at 18°C and shifted to 30°C for either 5 hours or 1day before being assayed. We found that after 5 hour induction of Crb overexpression, the total amount of Ex in cells is reduced (Figure 4.9G), while the amount of basolaterally localized Ex is increased (Figure 4.9I). Crb overexpression in embryonic epithelial cells also results in similar effects on Sdt (REF). We found that overexpression of CrbintraΔJM does not cause Ex relocalization (Figure 4.9H,J). Consistent with the requirement for growth, the JM domain is necessary for the effect on Ex localization. We conclude that overexpressed Crb is sufficient to relocalize Ex. This supports our model that Crb is essential for apical localization of Ex. Our data indicate that Crb, in particular the JM domain, regulates Ex localization and/or stability.

It has been reported that the level of Ex is decreased in fat mutant clones in a D dependent manner (Irvine). However, the deregulation of Ex is not observed in ft, d double mutant clones (Feng and Irvine 2007). To test whether Crb regulates Ex independently of Fat, we examined the effect of loss of Crb on Ex in a d mutant background. We found that removing D does not rescue the loss of Ex in crb mutant clones. Ex was still mislocalized in crb mutant cells in a d mutant background (Figure 4.10C, arrowhead). Our data indicate that Crb regulates Ex membrane localization in a D independent manner. Moreover, D localization, as well as Fat localization, remained intact in crb mutant clones (Figure 4.10D, arrowhead). Thus, our data support the idea that Crb regulates Ex membrane localization through a Fat and D independent mechanism.


The studies described in this Chapter functionally link the growth regulatory activity of Crb with Hippo signaling and thus identify Crb as a novel component of the Hippo pathway. We showed that Crb gain and loss of function cause overgrowth, ectopic proliferation, and the upregulation of Hippo pathway target genes. The overgrowth phenotypes of crb and the induction of Hippo target genes require Yki, indicating that Yki is epistatic to Crb. Moreover, loss of Crb genetically interacts and synergizes with mutations of Hippo pathway components. Furthermore, the proper level of Crb is required for the correct localization of Ex. Taken together, our data indicate that Crb acts through the Hippo pathway to regulate tissue size.

4. 7. Crb functions upstream in the Hippo pathway

To date, multiple inputs to the Hippo pathway have been identified, including the atypical cadherin Fat, and Mer. Nevertheless, the mutant phenotypes of upstream components, such as ft and mer, are generally weaker than those of downstream components, such as hpo and wts. Interestingly, ft;mer double mutants which abolish signals from those two different upstream branches display a stronger phenotype that resembles mutant phenotypes of downstream components. Similarly, loss of crb synergized with knocking down of mer in the pupal retina indicating that Crb and Mer may function in different upstream branches and cooperate to modulate Hippo pathway activity. This also supports the idea that Crb functions in the Ex branch and specifically regulates the localization of Ex but not that of Mer.

Proper level and localization of Crb appear to be essential for the correct localization of Ex to the sub-apical region of the plasma membrane. Crb loss and gain of function had reduced protein levels of Ex at the sub-apical plasma membrane even though the transcription level of ex was increased. Crb is likely to regulate Ex post-transcriptionally and affects the localization of Ex to the apical membrane. This further supports the model that Crb functions upstream of Ex in the Hippo pathway.

Furthermore, Crb itself is controlled by a negative feedback loop through the Hippo pathway (Genevet, Polesello et al. 2009; Hamaratoglu, Gajewski et al. 2009). Similar feedback mechanisms have been observed for several other Hippo pathway components (Genevet, Wehr et al. ; Hamaratoglu, Willecke et al. 2006; Hamaratoglu, Gajewski et al. 2009). Epithelial cells mutant for hpo and wts display elevated levels of Crb as well as Ex, Mer, Kibra, and Fat (Genevet, Wehr et al. ; Hamaratoglu, Willecke et al. 2006). At least for ex and kibra, the feedback depends on transcriptional regulation and is thus not simply a secondary consequence of the enlargement of the apical domain observed in Hippo pathway mutants. Rather, it constitutes a direct feedback loop in the Hippo pathway. Those feedback regulations may provide a homeostatic effect on the regulation of the Hippo pathway. Notably, the feedback regulation of Crb is not dependent upon transcriptional regulation (Tapon) indicating that more than one mechanism may contribute to achieve homeostasis of Hippo pathway activity.

4. 8. Expanded stability and membrane localization

Ex was largely absent from the apical membrane and diffused into the cytoplasm at the basal lateral region in crb mutant cells, while the Ex level at the apical domain is decreased and localized more basolaterally with ectopically expressed Crb. Interestingly, when we extended the duration of those genetic manipulations, such as by prolonged induction of Crb overexpression, the basolaterally localized Ex is no longer observed. This implies that Ex may be degraded when not localized properly. Because Ex is recruited by overexpressed Crb to a more basolateral region and thus degraded, it is unlikely that the interaction between Crb and Ex would stabilize Ex. Crb may act as a scaffold that is required to recruit Ex to the sub-apical membrane making it available for another unknown regulator to stabilize. Alternatively, Crb may make Ex unavailable for proteins that degrade Ex and normally localize basolaterally. In either case, it appears that the localization of Ex is important for its stability, and the degradation of Ex may be a potential regulatory mechanism of the Hippo pathway. Further understanding of how the presence of Crb affects Ex stability will offer insights into how Hippo signaling is regulated.

4. 9. Crb regulates cell polarity and Hippo signaling through different mechanisms.

It is interesting that Crb coordinately interacts with cell polarity determinants and components of the Hippo tumor suppressor pathway. However, our data indicate that Crb regulates these two pathways by different domains and thus through different mechanisms. Crbintra, which is lacking the extracellular domain, is sufficient to mediate the functions of full-length Crb in modulating Hippo signaling and cell polarity (Wodarz, Hinz et al. 1995; Izaddoost, Nam et al. 2002). Specifically, the effects on Hippo signaling require the JM whereas the effects on cell polarity require the PBM, which binds to Sdt. These results indicate that the function of Crb in apical-basal polarity and growth control can be uncoupled.

The JM is a FERM-domain binding motif that can directly interact with the FERM domain protein Yurt during development (Laprise, Beronja et al. 2006). It has been shown that Yurt can negatively regulate Crb to control cell polarity (REF). However, yurt mutants do not have growth defects, unlike ex mutants. Several lines of evidence imply that Crb may bind to Ex directly. Consequently, the effects on the Hippo pathway by Crb loss and gain of function may be caused by loss of Ex which leads to decreased Hippo activity. First, the JM is a FERM-domain binding motif that is potentially capable of interacting with the FERM domain of Ex. Second, the JM is required for the Crb-induced growth phenotypes that are similar to those with loss of Hippo activity. Third, Crb is required for correct Ex localization. Indeed, it has been recently reported by Ling et al. that Crb can directly interact with the FERM domain of Ex through the JM. Nevertheless, the phenotypes of Crb overexpression are stronger than ex mutants and thus cannot simply be explained by loss of Ex. Therefore, Crb is likely to interact with another FERM domain protein that cooperates with Ex to regulate Hippo signaling. The identification of novel interaction partners for Crb will certainly shed light on the molecular mechanism of Crb's action.

As discussed previously, Crb regulates apical-basal polarity and growth by using different domains and thus through different mechanisms. Crb potentially mediates the crosstalk between the apical-basal polarity pathway and growth control signaling through the Hippo pathway. Notably, Crb is required for proper Crb localization on neighboring cells and is thus non-autonomously required for the localization of Ex and Patj at the apical membrane. Crb may simply function as a scaffold protein that is required for proper membrane localization of Ex and Patj. Alternatively, Crb may act as a receptor and transduce the extracellular cue to both the cell polarity pathway and the Hippo pathway. In this scenario, homophilic binding of Crb may coordinate growth and polarity information signal between cells. These results thus identify a cell-cell interaction dependent mechanism that is mediated by Crb and regulates Hippo pathway activity.

Do Crb homologs act through Hippo signaling in mammals? Three Crb homologs, Crb1-3, has been identified in mammals. However, it is not clear whether any of the vertebrate Crb homologs regulate growth. The intracellular domains of Crb1-3 are conserved and important for proper apical-basal polarity (Bazellieres, Assemat et al. 2009). Notably, Crb3 has been implicated as a tumor suppressor in immortalized mouse kidney epithelial cells (Karp, Tan et al. 2008). During the selection of tumorigenic cell lines, the expression of Crb3 is lost. Interestingly, overexpression of Crb3 can restore contact inhibition and cell polarity, and suppress tumor progression. In addition, the highly conserved Hippo signaling pathway has been implicated in tumor suppression in vertebrates (Harvey and Tapon 2007; Pan 2007; Reddy and Irvine 2008; Zhao, Lei et al. 2008). Therefore, our study placing Crb within the Hippo signaling pathway may have important implications for the study of cancer development and treatment.