The Drosophila Frizzled (Fz) and Frizzled2 (DFz2) proteins function as receptors for Wingless (Wg) in tissue culture cells. While previous results indicate that loss of function for fz results in tissue polarity defects, the loss-of-function effects of Dfz2 are not known. The requirements of fz and Dfz2 during neurogenesis have now been examined. Both Fz and DFz2 function in Wg signaling, and loss of either of the two affects the same subset of neuroblasts as those affected by loss of wg. While these defects are partially penetrant in embryos lacking either fz or Dfz2, the penetrance is significantly enhanced in embryos lacking both. Since the penetrance of the CNS phenotypes is not complete in double mutants, additional components that allow some degree of Wg signaling must exist in vivo (Bhat, 1998).

In the ventral nerve cord of the Drosophila embryo, wg is expressed in row 5 cells within a segment. It is nonautonomously required for the formation and specification of row 4 neuroblasts as well as for the formation of a few neuroblasts in row 5 and most neuroblasts in row 6. Among those neuroblasts that are affected in wg mutants, NB4-2, a row 4 neuroblast that gives rise to the RP2/sib lineage, has been one of the most studied neuroblasts in the CNS. The RP2 and its sibling cell are formed from the first GMC of NB4-2 (this GMC is known as GMC-1 or GMC4-2a). In this lineage, wg is required for both the formation and specification of this neuroblast. The elimination of maternal and zygotic fz gene products causes loss of NB4-2->GMC-1->RP2/sib lineage and failure in the formation of row 6 neuroblasts in the ventral nerve cord of the Drosophila embryo (Bhat, 1998).

It has been argued that Fz might not function in the Wg signaling during wing and eye development, and that in these tissues, while DFz2 functions to receive the Wg signal, Fz receives the signal from some other Wnt. If DFz2 is solely responsible for receiving the Wg signal in the CNS, elimination of Dfz2 should have eliminated the NB4-2 lineage in a manner similar to loss of function for wg. Given that the wg-like CNS defects in Dfz2 embryos are only partially penetrant, as is the case with fz mutants, the simplest explanation is that there is a genetic redundancy between fz and Dfz2 and both function in the transduction of the Wg signal. The observation that the penetrance of the RP2/sib lineage phenotype is significantly enhanced in embryos lacking both fz and Dfz2 activities certainly reinforces this view. These results are also consistent with the observation that during epidermal patterning, while the intracellular localization of Arm is not significantly affected in embryos missing Dfz2, it is nearly lost in embryos missing both the activities (Bhat, 1998 and references).

Many epithelial cells are polarized along the plane of the epithelium, a property termed planar cell polarity. The Drosophila wing and eye imaginal discs are the premier models of this process. Many proteins required for polarity establishment and its translation into cytoskeletal polarity were identified from studies of those tissues. More recently, several vertebrate tissues have been shown to exhibit planar cell polarity. Striking similarities and differences have been observed when different tissues exhibiting planar cell polarity are compared. This study describe a new tissue exhibiting planar cell polarity -- the denticles, hair-like projections of the Drosophila embryonic epidermis. the changes in the actin cytoskeleton that underlie denticle development are described in real time, and this is compared with the localization of microtubules, revealing new aspects of cytoskeletal dynamics that may have more general applicability. An initial characterization is presented of the localization of several actin regulators during denticle development. Several core planar cell polarity proteins are asymmetrically localized during the process. Finally, roles for the canonical Wingless and Hedgehog pathways and for core planar cell polarity proteins in denticle polarity are described (Price, 2006).

Among the hallmarks of PCP in structures as diverse as Drosophila wing hairs to stereocilia in the mammalian ear is polarization of the actin cytoskeleton. The polarized actin cytoskeleton underlying wing hair polarity has been described and defects in polarization in fz and dsh mutants have been documented. Microtubules (MTs) are also polarized in developing wing hairs, and disruption of either actin or MTs disrupts wing hair formation. The data suggest that basic features of cytoskeletal polarity in pupal wing hairs are also seen in denticles. Denticles, like wing hairs, arise from polarized actin accumulations – in denticles this occurs along the posterior cell margin. Further, like wing hairs, denticles all elongate in the same direction. The less detailed analysis of dorsal hairs suggests that they also arise from polarized actin accumulations, but these are more complex; different cell rows accumulate actin either along the anterior or posterior cell margin (Price, 2006).

The effect of Wg and Hh on denticle development is mediated in part by their regional activation of the Shaven-baby transcription factor (Ovo), which is necessary and sufficient for cells to generate actin-based denticles. Therefore genes that are targets of Shaven-baby are likely to be triggers for actin accumulation and cytoskeletal rearrangements. Wg and Hh signaling may also trigger polarization of cellular machinery that is not typically thought to be involved in PCP – e.g. the polarity of Arm that was observed. It will be useful in the future to examine whether proteins polarized during germband extension, such as Bazooka, are also polarized during denticle formation. Mutations in both hh and wg also affected the normal changes in cell shape accompanying denticle formation – rather than elongating along the dorsal-ventral axis, cells remain columnar. A similar failure of cells to polarize during dorsal closure is observed in wg mutants. These effects may reflect alterations in cell polarization or cytoskeletal regulation. It will be of interest to determine whether changes in cell shape are coupled to the establishment of cytoskeletal polarity (Price, 2006).

Thus far the analysis of actin in wild-type and mutant pupal wings has been restricted to snapshots in fixed tissue. This was extended by examining F-actin in developing denticles in real time, revealing features of polarization that have not been noted previously; these features may be shared with wing hairs or other polarized structures. The initial cytoskeletal change observed was actin accumulation all across the apical surface of the cell. This actin gradually 'condenses', becoming more restricted to the posterior cell margin and forming distinct condensations, which then brighten and sometimes merge. They then elongate, all in the posterior direction. It will be interesting to learn whether the dynamic aspects of condensation involve de novo actin polymerization and/or collection of preexisting actin filaments (Price, 2006).

It is only in late condensations that enrichment was seen of any of the actin regulators that were examined. Arp3 and Dia are weakly enriched in late condensations, with enrichment increasing as denticles elongate, and Ena is enriched even later. Of course, the localization of these actin regulators to developing denticles does not by itself demonstrate that they play an important role there, but it is consistent with the possibility that they have a role in actin remodeling associated with denticle elongation. To test this hypothesis, genetic analyses will be necessary. This presents significant obstacles, since Arp2/3 and Dia are required for much earlier events (syncytial stages and cellularization), while maternal Ena plays a role in oogenesis, complicating analysis of loss-of-function mutants. Surprisingly, none of these actin regulators localizes in an informative fashion during the initial formation of actin condensations (though APC2 localizes there during this time). Thus additional regulators functioning during early denticle development need to be identified. Studies of cytoskeletal regulation in the larger adult sensory bristles may guide this. EM studies, the use of cytoskeletal inhibitors, and FRAP, which has proved informative in studies of wing hairs and bristles, may reveal how actin in denticles is assembled. Finally, it will be important to study in denticles additional actin regulators that regulate bristle development (Price, 2006).

What signals regulate denticle polarity? As examples of PCP have proliferated, understanding of the signals that instruct cells about their orientation in epithelial sheets has evolved. Certain features are shared in many, if not all, tissues. Fz receptors play a key role. Other core polarity proteins including Dsh, Fmi, Van Gogh/Strabismus and Prickle act in many if not all places. The current data extend this analysis to the denticles. Intriguing differences were found between the phenotypes of loss of Wg or Hh signaling, in which polarity was severely altered or abolished and loss of proteins that play dedicated roles in PCP, such as embryos null for either fz or stbm, which exhibit more subtle defects. A strong polarity bias was retained in these latter mutants, with cells in the posterior denticle rows correctly polarized and only cells in the anterior two rows making frequent mistakes. Interestingly, occasional mistakes are also observed in wild-type embryos (albeit at much lower frequency) and these are also restricted to the anterior most rows. This is in strong contrast to the effects of these mutants in the wing disc, where they globally disrupt polarity (Price, 2006).

One possible reason for this difference is the different scales of the tissues. The embryonic segment is only 12 cells across, while the wing disc encompasses hundreds of cells. Many core polarity proteins help mediate a feedback loop that amplifies an initially small difference in signal strength between the two sides of a wing cell. Perhaps the small scale of the embryonic segment makes this reinforcement less essential. It is also intriguing that the polarity is most sensitive to disruption in the anterior two denticle rows. If signal emanated from the posterior, signal strength might be lower in the anteriormost cells, rendering the reinforcement process more important. The lower frequency of defects in pk¹ mutants may also reflect the reduced role of the feedback loop, but this is subject to the caveat that pk is a complex locus with different mutations having different consequences. Future work will be needed to test these possibilities (Price, 2006).

Significant questions also remain about the signal(s) activating Fz receptors during PCP. Wnts were initial candidates, since Fz proteins are Wnt receptors. In vertebrates, this may be the case – Wnt11 regulates convergent extension and Wnt proteins can regulate PCP in the inner ear. By contrast, Drosophila Wnt proteins may not play a direct role. The Wg expression pattern in the eye and wing discs is not consistent with a role as the PCP ligand. Detailed studies of PCP in the eye and abdomen are most consistent with the idea that neither Wg nor other Wnt proteins are polarizing signals, but suggest that Wg regulates production of a secondary signal [dubbed `X'). Recent work suggests that Fj, Ds and Fat may be this elusive signal, with Drosophila Wg acting as an indirect cue of polarity. In fact, one cannot rule out the possibility Wnt11's role in vertebrate convergent extension is also indirect (Price, 2006).

Roles were found for Wg, Dsh and Arm in establishing denticle polarity. At face value, Arm's role is surprising, since the current view is that the Wg pathway diverges at Dsh, with a non-canonical branch mediating PCP and the canonical pathway playing no role in this. However, the data do not imply that Arm is required in denticle PCP per se. Wg acts in a paracrine feedback loop to maintain its own expression. In embryos maternally and zygotically mutant for arm alleles that cannot transduce Wg, Wg expression is lost by late stage 9. Thus, even though Arm is not in the non-canonical pathway, loss of Arm could still disrupt PCP indirectly due to the loss of Wg expression (Price, 2006).

While the data demonstrate that Wg is required for denticle PCP, two things suggest its role is indirect. wg mutants retain segmental periodicity in denticle orientation, suggesting that polarity is not totally disrupted, while in hh mutants there is no segmental periodicity. Second, when Wg signaling was reduced but did not eliminated, many cells retained normal polarity and there was segmental periodicity to which cells lost polarity or exhibited polarity reversals. This is consistent with the idea that Wg regulates production of another ligand. In fact, Wg's role may be even more indirect – given the more dramatic effect of hh, Wg's primary role in polarity may be to maintain Hh expression (this is also consistent with a requirement for canonical pathway components like Arm). Global activation of Hh signaling in the ptc mutant also disrupts polarity. Hh thus remains a possible directional cue. In the abdomen, Hh also plays an important role in polarity, but it does not seem to be the directional cue either but rather regulates its production; this may also be the case in the embryo. Thus the precise roles for canonical Wg and Hh signaling in denticle polarization must be addressed by future experiments. If neither Wnts nor Hh are directional signals, what is? Data from the eye, wing and abdomen suggest roles for Ds, Fj, Fat and Fmi but details differ in different tissues. It thus will also be useful to examine Ds, Fj and Fat's roles in embryonic PCP (Price, 2006).

Immunostaining of pupal wings shows that the FZ protein is evenly distributed throughout the wing (Park, 1994). This suggests that FZ protein is not a graded morphogen. FZ proteins localize to the apical region of pupal cells (Park, 1994).

The Drosophila eye is composed of several hundred ommatidia that can exist in either of two chiral forms, depending on position: ommatidia in the dorsal half of the eye adopt one chiral form, whereas ommatidia in the ventral half adopt the other. Chirality appears to be specified by a polarizing signal with a high activity at the interface between the two halves (the 'equator'), which declines in opposite directions towards the dorsal and ventral poles. Here, using genetic mosaics, it is shown that this polarizing signal is decoded by the sequential use of two receptor systems. The first depends on the seven-transmembrane receptor Frizzled (Fz) and distinguishes between the two members of the R3/R4 pair of presumptive photoreceptor cells, predisposing the cell that is located closer to the equator and having higher Fz activity towards the R3 photoreceptor fate and the cell further away towards the R4 fate. This bias is then amplified by subsequent interactions between the two cells mediated by the receptor Notch (N) and its ligand Delta (Dl), ensuring that the equatorial cell becomes the R3 photoreceptor while the polar cell becomes the R4 photoreceptor. As a consequence of this reciprocal cell fate decision, the R4 cell moves asymmetrically relative to the R3 cell, initiating the appropriate chiral pattern of the remaining cells of the ommatidium (Tomlinson, 1999).

The focus for the chirality choice maps to the presumptive R3/R4 photoreceptor pair; it maps specifically to the R3 cell. Although these two cells lie adjacent in the ommatidial precluster, fate-mapping experiments, together with histological analysis, indicate that they are initially distant from one another, separated along the equatorial axis by at least the remaining three cells that will enter the precluster. During normal development, the member of the pair that is closest to the equator invariably chooses to develop as an R3 cell, while the remaining member develops as an R4 cell. This asymmetric cell fate choice determines the chirality of the ommatidium. The R3/R4 cell fate decision is governed by the relative difference in activity of Frizzled (Fz) protein in the two cells comprising the presumptive R3/R4 pair. The cell with higher Fz activity becomes R3, while the remaining cell becomes R4. Since fz encodes a serpentine receptor-like protein, Fz is a candidate receptor for a factor X, consistent with the notion that the presumptive R3/R4 pair decodes the factor X gradient into a relative difference in Fz activity between the two cells. The difference in Fz activity between the two members of the presumptive R3/R4 pair biases a process of lateral specification between the two cells mediated by the receptor Notch (N) and its ligand Delta (Dl). In essence, the cell with higher Fz activity appears to be better at sending the Dl signal while the cell with lower Fz activity is better at receiving the signal. This difference is probably then amplified by a feedback mechanism in which receptor activation blocks ligand production in the same cell, while the loss of receptor activity leads to enhanced ligand production in the other cell. As a consequence, the N transduction pathway is fully induced in one cell but silenced in the other. The resulting disparity in N signal transduction is both necessary and sufficient to specify the reciprocal R3 and R4 cell fates and to determine the chirality of the ommatidia (Tomlinson, 1999).

Two models are presented for how Fz activity within the R3/R4 pair might bias the N-Dl interaction. In the first model, a scalar model, factor X activity positively regulates Fz activity, whereas in the second it negatively regulates Fz activity. A convention of factor X activity has been adopted; factor X is considered as being high at the equator and low at the poles. However, the results can be equally well explained if factor X has the opposite distribution, high at the poles and low towards the equator. In this scenario the positive and negative influences of factor X on Fz activity in the two models will be reversed. A difference in the levels of Fz activity between the two cells of the R3/R4 pair determines which cell will become R3 and which will become R4. Because the presumptive R3 cell lies closer to the equator than the presumptive R4, it will detect higher levels of factor X activity and consequentially will have a higher level of Fz activity. The higher levels of Fz activation then bias the subsequent N-Dl interaction so that the cell with greater Fz activity becomes a dedicated Dl signaling cell, while its partner, with less Fz activity, becomes a dedicated Dl receiving cell. For example, the level of Fz activity in each cell could govern the activity or level of expression of a component of the N-Dl signaling mechanism, such as N itself or Dl. Even a small difference in signaling capacity between the two cells would then bias the N-Dl system of feed-back regulation causing the cell with initially higher N transducing capacity to become a dedicated Dl receiving cell (and hence R4), while the remaining cell becomes a dedicated Dl sending cell (R3). Experimentally induced changes in Fz activity, which reverse the relative difference in Fz activity between the presumptive R3 and R4 cells, might cause a corresponding reversal in direction of N-Dl signaling and the R3/R4 cell fate decision. An issue raised by this model is whether the presumptive R3 and R4 cells have the capacity to meter accurately what are likely to be small differences in their absolute levels of factor X. Consider that up to 15 ommatidia can form along the equatorial-polar axis of each half of the eye, with the presumptive R3 and R4 cells located next to each other in each cluster and separated by several cell diameters from their counterparts in neighboring clusters. More than 75 cells are arrayed in each half of the eye along the Eq/Pl axis and any two neighboring cells at any position within that array would need to faithfully decode the factor X gradient. However, the results of fate-mapping analysis and histological studies suggest that the presumptive R3 and R4 cells are initially located at a distance from one another, separated along the Eq/Pl axis by the remaining three cells of the precluster (the presumptive R2, R8 and R5 cells). This raises the possibility that the presumptive R3 and R4 may meter factor X abundance when they are located at opposite ends of these cell lines before they come to lie next to each other. The physical separation between the two cells at this time would allow them to sample a broader segment of the factor X gradient, much as the separation between the two tips of the forked tongue of a snake facilitates the detection of odor gradients (Tomlinson, 1999).

In a second model, the vectoral model, each cell in the retina, and hence both the presumptive R3 and R4 cells, detects the gradient of factor X activity across its diameter and decodes it into a steeper gradient of Fz activity within the cell. Specifically, it is proposed that the factor X activity might negatively regulate Fz such that Fz activity is highest within each cell on its polar side and lowest on its equatorial side. Within the precluster, the presumptive R3 cell presents its polar face to the equatorial face of the presumptive R4 cell. The differential Fz activities across the two faces then bias the subsequent N-Dl interaction between the two cells. In this case, it would be argued that the bias is likely to be mediated through the direct and local modulation of a component of the N-Dl signaling apparatus, e.g. a post-translational modification of N or Dl activity along the surface of one of the two cells, where they abut. When Fz is overexpressed in the presumptive R4 cell, Fz activity is enhanced throughout the cell, including at the surface comprising the equatorial face where it abuts the polar face of the R3 cell. As a consequence, any influence of the polarizing factor X gradient would be subsumed, reversing the N-Dl signaling predisposition that would otherwise occur. Conversely, when Fz activity in the presumptive R3 cell is abolished, the polar face of the R3 cell would now have lower Fz activity than the adjacent, equatorial face of the R4 cell. As in the overexpression case, this change would reverse the direction of N-Dl signaling and hence reverse the reciprocal cell-fate choices made by the two cells (Tomlinson, 1999).

A key difference between the two models is that the scalar model requires cells to meter the absolute concentration of factor X via the level of activation of Fz protein, whereas the vectorial model requires them to detect and then amplify a relative difference of factor X concentration across the cell's diameter. Another difference is that factor X is predicted to have opposite effects on Fz activity, activating Fz in the scalar model and inhibiting Fz in the vectorial model. Hence, it is difficult to envisage how both mechanisms could work in concert, and as a consequence, they are viewed as mutually exclusive (Tomlinson, 1999).

One way in which a cell might amplify a difference in factor X abundance across its diameter would be to polarize the distribution of Fz itself in response to the ligand. Although the mechanisms are not yet clear, there are several precedents for the ability of cells to respond to shallow gradients of extracellular signals by polarizing the distribution of receptors and cytoskeletal components (e.g. during shmooing of yeast cells in response to mating pheromone, and the extension of pseudopods in Dictyostelium in response to cAMP). It is notable that Fz activity gradients have been implicated in the control of cell polarity in other epithelial tissues in the fly, particularly the wing. However, there is, at present, no indication that N signaling is involved in the establishment of cell polarity in this context. It is therefore suggested that Fz signaling generally mediates the establishment of cell polarity without requiring an N-Dl feedback amplification step. Hence, the involvement of N-Dl signaling in establishing ommatidial chirality may reflect a special attribute of this system, perhaps to allow the polarity of just two cells, the presumptive R3/R4 pair, to be used as a cue to control the pattern of a much larger ensemble of cells, the ommatidium (Tomlinson, 1999).

When Fz activity is absent throughout eye development, the Fz-dependent bias should be eliminated and each cell of the presumptive R3/R4 pair should have an equal chance of becoming either the signaling cell (R3) or the receiving cell (R4). Under these conditions, the choice of which cell becomes R3 and which becomes R4 would be determined by a stochastic variation, which gives one of the two cells a slight advantage that is then amplified by the N-Dl feedback mechanism. This explains why ommatidia in fz mutants 'choose' their chirality randomly when both cells have the same N gene dosage, but non-randomly when there is a 3:2 differential in N gene dosage. However, in fz minus eyes, approximately one third of the ommatidia are symmetrical, indicating that the R3/R4 distinction has not been resolved. One possible explanation is that the interaction between the R3 and R4 cells may be limited to only a few hours and this may be too short to ensure that a stochastic variation will arise and be amplified by the N pathway in all fz minus ommatidia. Timing may not be as critical in wild-type ommatidia, because the bias from Fz signaling is sufficient to ensure an appropriate resolution during this relatively brief interval (Tomlinson, 1999).

Asymmetric partitioning of cell-fate determinants during development requires coordinating the positioning of these determinants with orientation of the mitotic spindle. In the Drosophila peripheral nervous system, sensory organ progenitor cells (SOPs) undergo several rounds of division to produce five cells that give rise to a complete sensory organ. The asymmetric divisions that give rise to these cells have been visualized in developing pupae using green fluorescent protein fusion proteins. Spindle orientation and determinant localization are tightly coordinated at each division. Furthermore, two types of asymmetric divisions exist within the sensory organ precursor cell lineage: the anterior-posterior pI cell-type division, where the spindle remains symmetric throughout mitosis, and the strikingly neuroblast-like apical-basal division of the pIIb cell, where the spindle exhibits a strong asymmetry at anaphase. In both these divisions, the spindle reorients to position itself perpendicular to the region of the cortex containing the determinant. On the basis of these observations, it is proposed that two distinct mechanisms for controlling asymmetric cell divisions occur within the same lineage in the developing peripheral nervous system in Drosophila (Roegiers, 2001).

The dynamics of Pon-GFP localization and the coordination of spindle positioning and Pon-GFP crescent localization were examined in frizzled (fz) mutant pupae. The Frizzled protein is required for maintaining planar polarity during SOP divisions and thus the proper orientation of sensory bristles on the thorax. During the pI division in wild-type pupae, Pon-GFP forms an anterior crescent and is segregated to the anterior pIIb cell after mitosis. The divisions of the pI cells on the pupal notum are asynchronous, but the A-P orientation of the divisions is invariant. In fz^r54 mutant pupae, the pI division occurs with an orientation that is random with respect to the A-P axis. Moreover, in time-lapse confocal images of pI divisions in fz mutants, the Pon-GFP crescent had drifted between 45° and 90° from its initial position by late anaphase in four out of ten SOPs. This result suggests that Fz may have a role in stabilizing the positioning as well as orienting the Pon crescent during mitosis (Roegiers, 2001).

The coordination of Pon crescent with spindle orientation was examined in the fz mutant background. Despite the mispositioning of the crescent relative to the A-P axis, the dynamics of spindle orientation are similar to those of the wild type. In most cases (90%), the mitotic spindle positions itself perpendicular to the Pon-GFP crescent, regardless of the relationship of the crescent to the A-P axis. The spindle has been observed to seesaw during the accumulation of Pon-GFP to one side of the cell; however, in about one-tenth of cases, the mitotic spindle is positioned parallel to the Pon-GFP crescent, but appeared to have both spindle poles anchored to the cortex. The spindles also have an abnormal curved morphology throughout mitosis. Notably, at telophase these cells appear to segregate the bulk of the Pon-GFP to one of the two daughter cells, despite the abnormal position of the spindle during metaphase and anaphase (Roegiers, 2001).

Owing to this rescue of the asymmetric localization of Pon at anaphase, cell-fate transformations have never been reported in fz mutants, and no loss of external sensory structures have been observed that would indicate a mis-segregation of Numb protein in the pI division. In SOPs co-expressing Pon-GFP and Tau-GFP in a wild-type background, spindles are always oriented perpendicular to the Pon-GFP crescent at anaphase. The orientation of subsequent divisions within the SOP lineage does not depend on Fz. This analysis reveals that in 14 out of 18 SOPs, the division pattern of the SOP lineage occur in the same orientation relative to pI as in the wild type, despite the randomization of the pI division with respect to the A-P axis. In the 4 out of the 18 remaining SOPs, the divisions were asymmetric, but slight deviations from the wild-type orientation were observed, such as the pIIa division occurring at a 45° angle from the pI division. Together, these results indicate that Fz participates in establishing the correct orientation and stabilization of the position of the Pon-GFP crescent in the pI cell, and that it may participate in the coordination of spindle orientation and crescent formation (Roegiers, 2001).

These results suggest that there are two fundamental types of asymmetric divisions in the developing Drosophila nervous system. During Drosophila development these two types of divisions are reiterated in different tissues at different times to generate cell-fate diversity. The divisions of the sensory organ precursor cell provide a unique system for studying different types of asymmetric cell divisions within the same lineage and how they might be coordinated. The orientations of the divisions are tightly regulated: two divisions occur along the A-P axis, and two divisions occur in the apical-basal orientation. In the pI division, which occurs along the A-P axis, the spindle is symmetric and reorients to align perpendicular to the crescent of Pon-GFP, and fz is important for the proper orientation of the crescent and appears to contribute to the coordination of spindle orientation and crescent positioning. In contrast, the spindle in the pIIb cell orients along the apical-basal axis and exhibits a strong size asymmetry. insc, a gene of central importance in coordinating spindle orientation and crescent formation in embryonic neuroblast divisions, also has an important role in orienting the mitotic spindle in the pIIb cell. These findings provide strong evidence that the pIIb division is a neuroblast-like division. It will be interesting to know whether other genes known to be involved in controlling the asymmetric divisions of neuroblasts, such as bazooka or partner of inscuteable, are also required for the pIIb division. The results may reveal general mechanisms for generating cell-fate diversity in Drosophila as well as in other species (Roegiers, 2001).

The frizzled gene of Drosophila encodes a transmembrane receptor molecule required for cell polarity decisions in the adult cuticle. In the wing, a single trichome is produced by each cell, which normally points distally. In the absence of frizzled function, the trichomes no longer point uniformly distalward. During cell polarization, the Frizzled receptor (visualized using Frizzled-Green fluorescent protein) is localized to the distal cell edge, probably resulting in asymmetric Frizzled activity across the axis of the cell. Furthermore, Frizzled localization correlates with subsequent trichome polarity, suggesting that it may be an instructive cue in the determination of cell polarity. This differential receptor distribution may represent a novel mechanism for amplifying small differences in signaling activity across the axis of a cell (Strutt, 2001).

To understand the asymmetric distribution of Fz-GFP, the distribution was studied in flies mutant for other genes involved in trichome polarity establishment. In clones of cells lacking starry night (stan) function, both the apical and PD localization of Fz-GFP is completely abolished. However, in cells lacking dsh function, in which Fz signal transduction is compromised, Fz-GFP apical localization is preserved, but there is no proximodistal (PD) localization, with a splotchy irregular distribution being seen instead. The same phenotype is observed for mutations in the prickle-spiny-legs (pk^pk-sple) and Van Gogh (Vang) genes. This would be consistent with the trichome polarity phenotypes of these mutations being due to a failure of Fz localization (Strutt, 2001).

Genetic data indicate that the polarity genes in, fy, and mwh act downstream of Fz/Dsh, inhibiting trichome formation where Fz is not active. In agreement with this, mutations in these loci do not alter Fz-GFP distribution despite trichome polarity being disrupted (Strutt, 2001).

The localization of Stan has also been reported to be disrupted in cells lacking dsh function but not in those lacking mwh. Therefore, whether Fz-GFP and Stan remain colocalized in different mutant backgrounds was tested. In clones of cells lacking dsh function, it was found that both Fz-GFP and Stan remain predominantly apical, and although the distribution of both appears diffuse, they nevertheless show broad colocalization. Similarly, in an in background, both Fz-GFP and Stan remain apical and colocalized to the PD cell boundaries (Strutt, 2001).

It is concluded that one important function of Stan is to localize Fz apically in the cell during polarity establishment. Stan may also play a role in localizing and/or anchoring Fz at the distal cell edge. However, in the absence of fz autonomous polarity signaling activity, neither Fz-GFP nor Stan is localized to PD boundaries. This leads to the speculation that the Fz receptor is responsible for receiving an extracellular polarity signal, and that the interpretation of this signal drives the localization of Fz to the distal cell edge and Stan to the distal and proximal cell edges, Stan thus acting downstream of Fz. Nevertheless, the data would equally well support Fz acting downstream of Stan; the codependence of the localization of both proteins would support this possibility, as would the observation that Stan has homology to G protein-coupled receptors. Thus far, no ligand has been identified for the Fz receptor in polarity signaling, and it is conceivable that in fact Fz might be activated by association with another transmembrane receptor, a role for which Stan is clearly a candidate (Strutt, 2001).

Both Fz-GFP and Stan localization are also downstream of fz nonautonomous signaling activity. One of the functions of Stan is to measure differences in fz activity between adjacent cells, as Stan accumulates on the boundary between fz⁺ and fz^- cells. These findings have been extended to show that Stan in fact accumulates on the boundaries between cells with different levels of fz nonautonomous signaling activity. This observation is consistent with there being a long-range gradient of fz nonautonomous signaling activity across the pupal wing, with each cell having different levels of fz nonautonomous activity relative to its neighbors (Strutt, 2001).

An intriguing observation is that in the cells bordering a clone deficient in only fz autonomous signaling, Stan appears to show a preference for localizing on the cell boundaries lying perpendicular to the clone boundary. Thus, in cells bordering the proximal and distal clone edges, Stan is sometimes seen lying on the lateral cell boundaries rather than the PD cell boundaries. Similarly, the same is true for Fz-GFP localization on the boundary of stan clones. This phenomenon seems to suggest that if Stan or Fz-GFP cannot localize on one of the PD boundaries of a cell, they show a preference for not localizing on the opposite PD boundary. This, in turn, opens up the possibility that there is an intracellular communication mechanism that couples Stan/Fz-GFP accumulation on one cell boundary to that on the opposite cell boundary. Interestingly, such a mechanism would enable the propagation of a wave of Stan/Fz-GFP polarization across the wing, starting from a single row of polarized cells at one edge. Obviously, such a mechanism for propagation of polarity would obviate the need for an external ligand gradient and would also argue against the existence of a long-range gradient of fz nonautonomous activity. However, long-range gradient models and cell-cell communication models for the propagation of polarity need not be mutually exclusive, and both may operate side by side (Strutt, 2001).

Taking these observations together, the following model is put forward for Fz function in the polarization of single cells in the developing wing. Initially, unlocalized Fz is required for the long-range propagation of a polarity signal. Fz is then recruited apically in a Stan-dependent manner and becomes stably localized at the distal cell edge in a process requiring Fz signaling and the activities of Stan, Dsh, Pk^pk-sple, and Vang. This Fz localization then restricts the site of trichome initiation to the distal cell vertex. It is possible that Fz signaling activates Stan molecules to bind both to Fz (in the same cell) and to Stan molecules in the adjacent cell, and so anchors Fz at the distal edge of the cell. Localization of Fz may lead to further increased Fz signaling (possibly through the effects of receptor clustering), which could, in turn, recruit more Stan and Fz. Over time, increased activity of clustered Fz receptors at the distal cell edge would lead to the majority of the Fz in the cell being recruited to this location. In heterologous systems, Fz activity leads to recruitment of Dsh to the cell membrane, so it is likely that Dsh is also present at the distal cell boundary. A precedent for Fz-dependent localization of a cytoplasmic protein during planar polarity establishment is provided by the observation that the Numb protein requires Fz activity for correct asymmetric subcellular localization during sense organ precursor cell divisions (Strutt, 2001).

The stable PD localization of Fz also requires Pk-Sple and Vang activity, with their loss having a similar effect on Fz-GFP localization as loss of Dsh activity. It is possible that, like Dsh, they are required for the transduction of the Fz signal, or they may be involved in the function of Dsh itself. Interestingly, Vang activity on only one side of the PD boundary is sufficient for Fz-GFP localization to occur. Further investigations of the biochemical activities of these proteins will be required to fully elucidate their roles in planar polarity establishment (Strutt, 2001).

During patterning of the Drosophila eye, the Notch-mediated cell fate decision is a critical step that determines the identities of the R3/R4 photoreceptor pair in each ommatidium. Depending on the decision taken, the ommatidium adopts either the dorsal or ventral chiral form. This decision is directed by the activity of the planar polarity genes, and, in particular, higher activity of the receptor Frizzled confers R3 fate. Evidence is presented that Frizzled does not modulate Notch activity via Rho GTPases and a JNK cascade as previously proposed. The planar polarity proteins Frizzled, Dishevelled, Flamingo, and Strabismus adopt asymmetric protein localizations in the developing photoreceptors. These protein localizations correlate with the bias of Notch activity between R3/R4, suggesting that they are necessary to modulate Notch activity between these cells. Additional data support a mechanism for regulation of Notch activity that could involve direct interactions between Dishevelled and Notch at the cell cortex. In the light of these findings, it is concluded that Rho GTPases/JNK cascades are not major effectors of planar polarity in the Drosophila eye. A new model is proposed for the control of R3/R4 photoreceptor fate by Frizzled, whereby asymmetric protein localization is likely to be a critical step in modulation of Notch activity. This modulation may occur via direct interactions between Notch and Dishevelled (D. Strutt, 2002).

Frizzled tagged with green fluorescent protein (Fz-GFP) exhibits a dynamic subcellular distribution from early stages of ommatidial differentiation. Ommatidia are born behind the furrow in rows polarized in the anteroposterior axis. In row 4, Fz-GFP is enriched on the apical membranes of the newly recruited R3/R4 pair but excluded from the region where they contact R2/R5. No Fz-GFP enrichment is apparent around R2/R5, but it does accumulate on the posterior side of R8. By row 6, Fz-GFP is no longer enriched in R3, except at the boundary with R4 and sometimes at the boundary with the anterior cone cell. Conversely, R4 still has strong accumulation around its perimeter, except where it contacts R5. This accumulation around R4 persists through row 8, but accumulation fades elsewhere. Thus, Fz-GFP is initially in a symmetric pattern in R3/R4 but rapidly resolves into an asymmetric pattern that is visible by the time ommatidial rotation occurs in row 6. Using antibodies against Dsh and Fmi, these proteins were found to colocalize with Fz and show the same dynamic distribution (D. Strutt, 2002).

N is also at highest levels in apical membranes of cells posterior to the furrow and in rows 4 through 6 it overlaps with Fz-GFP at the R3/R4 boundary (but shows no asymmetry). The localization of Fz-GFP (and Fmi/Dsh) to the R3/R4 boundary is therefore consistent with Fz/Dsh being able to directly modulate N activity in this location. However, if Fz/Dsh are differentially regulating N activity, a crucial requirement is that these complexes should be preferentially localized on one side of the R3/R4 boundary. Since this cannot be distinguished by light microscopy, genetic mosaics were created in which both R3/R4 had sufficient fz activity for normal signaling and fate determination, but only one of the pair carried the Fz-GFP transgene. Using this approach, it was found that Fz-GFP is more highly enriched on the R3 side of the R3/R4 boundary in row 4 and more posteriorly is found exclusively on the R3 side of the boundary. Thus, about two rows prior to ommatidial rotation, Fz-GFP is asymmetrically distributed across the R3/R4 boundary. Since studies in the wing demonstrate that Dsh adopts the identical asymmetric localization to Fz (and indeed their asymmetric localization is mutually dependent), it is inferred that Dsh is also differentially localized on the R3 side of the R3/R4 boundary (D. Strutt, 2002).

The polarity gene stbm is required for R4 fate: whether Stbm protein also shows an asymmetrical localization in R3/R4 was investigated using a Stbm-YFP transgene. Stbm-YFP is apically localized in cells posterior to the furrow, and, subsequently, its distribution is similar but distinct from that exhibited by Fz-GFP. In row 4, a symmetric pattern is observed, with Stbm-YFP around R3/R4, except where these cells contact R2/R5, and enriched on the posterior face of R8. This symmetric pattern is maintained until the ommatidia are already rotated in row 6 and more posteriorly. Staining then fades around R3, except where R3 contacts R4. Mosaic analysis revealed that, in contrast to Fz-GFP, Stbm-YFP is enriched on the R4 side of the R3/R4 boundary from row 4 onward, i.e., Stbm is on the opposite side of the boundary with Fz (D. Strutt, 2002).

Therefore Fz, Dsh, Fmi, and Stbm localize to the apical region of the R3/R4 cell boundary, where they become asymmetrically distributed prior to or concomitant with R3/R4 fate determination. Normally, Fz/Dsh are enriched on one particular side of the cell boundary, in the presumptive R3 cell. However, in mosaic ommatidia where one or other cell is mutant for polarity genes, the assembly of the asymmetrical complexes can be reversed. In all conditions examined, the polarity of Notch signaling between R3/R4 is consistent with the polarity of the asymmetric complexes, with Notch activity being lowest in the cell where Fz/Dsh accumulate. Finally, evidence is provided that the domain of N, which is known to interact directly with Dsh, is required for efficient R3/R4 fate decisions (D. Strutt, 2002).

Considering these results together, it is proposed that an extracellular polarity signal leads to the asymmetric assembly of a complex of planar polarity proteins at the boundary between the R3/R4 cell pair. This asymmetric complex then leads to asymmetric N activity between the cell pair and thus determines cell fate. Since no evidence is found that this regulation occurs via the proposed signaling cascade downstream of Fz/Dsh (i.e., Rho GTPases/JNK) and since manipulation of Dl transcription does not perturb polarity of Notch signaling, it is concluded that there must be an alternative pathway by which asymmetrical Fz/Dsh affects Notch activity (D. Strutt, 2002).

One favored mechanism for the modulation of N/Dl activity is via local interactions between N and asymmetrically localized proteins and, in particular, between the intracellular domain of N and Dsh. Four lines of evidence support the proposal that the regulation occurs at the cell cortex: (1) Fz/Dsh are in the same subcellular domain as N at the apical R3/R4 boundary during the critical stages of development when the cell fate decision is made; (2) the appearance of the asymmetric Fz/Dsh complexes is shortly prior to or concomitant with the appearance of a bias in N/Dl activity and ommatidial rotation; (3) direct interactions between N and Dsh have been previously demonstrated and proposed to be important for patterning in other tissues, and these interactions have been found to be repressive, consistent with Fz/Dsh being required in R3, where N activity is lowest and (4) deletion of the domain of N required for interactions with Dsh leads to less-efficient R3/R4 fate decisions (D. Strutt, 2002).

The model whereby asymmetric Fz/Dsh localization leads to downregulation of N activity on the R3 side of the R3/R4 boundary is further supported by studies in the Drosophila leg, where loss of planar polarity gene activity leads to ectopic activity of Notch. However, there are still unexplained observations: if the only role of the polarity genes is to inhibit N in R3, mutations in fmi, fz, or dsh (which result in no apical Dsh localization) should have high N activity in both R3/R4, not the reduced activity that is detected (D. Strutt, 2002).

This discrepancy might be explained if there are two phases to polarity gene regulation of N activity. One would be an activation/derepression of N, which would require symmetric protein localization of Fz/Dsh in R3/R4. The second would be linked to asymmetric protein localization, when Fz/Dsh would in turn become repressors of N activity in R3 (D. Strutt, 2002).

The asymmetric localization of Fz, Dsh, and Fmi in the eye develops in a similar manner to that seen in the pupal wing. Thus, the R3/R4 cell boundary appears analogous to the proximodistal wing cell boundaries, with the R3 side of the boundary, where Fz and Dsh are localized, being equivalent to the wing cell distal edge. Another of the polarity gene products, Stbm, is localized on the R4 side of the boundary, which is consistent with the requirement for stbm function in R4. By analogy to the wing, it is likely that Fmi is present on both sides of the R3/R4 boundary and Pk-Sple/Sple is on the R4 side (D. Strutt, 2002).

The adoption of the asymmetric pattern occurs in two phases. The first involves symmetric apicolateral localization of Fz, Dsh, Fmi, and Stbm in R3/R4 (and in all other cells except R2/R5); this is evident in ommatidial row 4. As in the wing, the initial apical recruitment of Fz is dependent on Fmi, and the recruitment of Dsh is in turn dependent on Fz. Subsequently, the distribution evolves rapidly into an asymmetric pattern. Adoption of asymmetry requires the function of dsh, stbm, and the LIM domain protein Prickle-Spiny-legs (Pk-Sple), and if any of these are missing, Fz distribution remains symmetric in ommatidial rows 5/6, and ommatidial rotation is delayed. It is likely that the asymmetry evolves through the same mechanisms as in the wing, where it has been proposed that an extrinsic signal leads to a small bias in Fz/Dsh signaling on either side of the cell boundary, which subsequently becomes amplified by feedback loops that lead to Fz/Dsh becoming concentrated on one side of the interface and Pk-Sple/Stbm on the other (D. Strutt, 2002).

One notable difference between the eye and the wing is that asymmetric Fz/Dsh distribution is eventually observed in stbm and pk-sple eye discs, but in both cases it occurs with a random bias and is delayed by about one to two ommatidial rows. This correlates well with the fact that the adult phenotypes of stbm and pk-sple exhibit a low incidence of achiral ommatidia. Conversely, in fmi, fz and dsh, negligible asymmetric protein localization occurs, and there is a relatively high proportion of 'achiral' ommatidia in the adult eye, suggesting that achirality is a result of poor asymmetric complex formation. In general, the aquisition of asymmetry also correlates with mDelta0.5 activity, particularly in pk-sple and sple mutations where its expression usually resolves into a single cell by row 10 (D. Strutt, 2002).

In the pupal wing, asymmetric localization of Fz/Dsh/Pk-Sple is proposed to involve a signaling feedback loop that amplifies an initially small bias in Fz/Dsh activity across the axis of each cell. In the eye, the N/Dl feedback loop was proposed to perform a similar function, amplifying an initially small difference in Fz/Dsh activity between R3/R4. With the observation that Fz/Dsh are also distributed in asymmetric complexes in the eye, it appears that both mechanisms are operating in R3/R4, although it is not clear why both would be required, since either alone should be sufficient to amplify small biases in signaling activity (D. Strutt, 2002).

One possible explanation is that use of both mechanisms increases the speed and robustness of the R3/R4 fate decision. A fast fate decision may be necessary because of the dynamic nature of eye patterning, in which the R3/R4 decision is only part of a complex series of events involving cell recruitment and movement to generate the final polarized ommatidium. It is also possible that a rapid decision is required because the extrinsic polarity cue is transient in nature. It is noted that the rapidity of the decision would be further enhanced if N/Dl signaling also influenced Fz/Dsh localization. While there is no direct evidence for this, it could explain the eventual, randomly polarized, asymmetric protein localization seen in stbm and pk-sple backgrounds in the eye. In this case, the inability of Fz/Dsh to efficiently localize asymmetrically in the absence of Stbm/Pk-Sple might lead to N/Dl making a stochastic decision that then leads to Fz/Dsh asymmetry. Conversely, in the pupal wing, where N/Dl are not active in planar polarity decisions, Stbm/Pk-sple activity would be absolutely required, since their absence would not be compensated for by the N/Dl feedback loop (D. Strutt, 2002).

A number of lines of evidence have previously suggested that Rho/Rac GTPases and the JNK cascade are required for ommatidial polarity decisions and, in particular, the R3/R4 fate decision. These include the following: overexpression of Fz or Dsh in the eye gives a polarity phenotype that is dominantly suppressed by RhoA, bsk, hep, and Djun; RhoA clones or expression of dominant-active/negative RhoA or Rac1 gives ommatidial polarity phenotypes; overexpression of dominant-active/negative JNK pathway components and human Jun elicits ommatidial polarity defects, and expression of a Dl enhancer trap is altered by overexpression of either fz or dsh or by activated human Jun, Hep, RhoA, or Rac1. These observations led to the hypothesis that higher levels of Fz/Dsh signaling in R3 result in higher activation of Dl transcription in R3 via a Rho GTPase/JNK cascade, biasing the N/Dl feedback loop to produce high N in R4 (D. Strutt, 2002).

Taken together, the phenotypic evidence from loss-of-function studies does not support a primary role for Rho GTPases/JNK cascades in the R3/R4 fate decision. But the weight of genetic evidence does support a secondary role for some of the proposed pathway components, possibly in the augmentation of polarity decisions driven largely by asymmetric localization of polarity proteins and direct repression of N activity. In addition, the observation that RhoA mutations result largely in defects in ommatidial rotation supports the hypothesis that RhoA acts downstream of the planar polarity genes in regulating this aspect of ommatidial polarity (D. Strutt, 2002).

Drosophila epithelia acquire a planar cell polarity (PCP) orthogonal to their apical-basal axes. Frizzled (Fz) is the receptor for the PCP signal, and Dishevelled (Dsh) transduces the signal. Unipolar relocalization of Dsh to the membrane is required to mediate PCP, but not Wingless (Wg) signaling. Dsh membrane localization reflects the activation of Fz/PCP signaling, revealing that the initially symmetric signal evolves to one that displays unipolar asymmetry, specifying the cells' ultimate polarity. This transition from symmetric to asymmetric Dsh localization requires Dsh function, and reflects an amplification process that generates a steep intracellular activity gradient necessary to determine PCP (Axelrod, 2001).

To investigate a possible role for Dsh membrane association during Fz/PCP signaling in vivo, Dsh subcellular localization during PCP signaling was examined in the developing wing. Transgenes were produced that express a Dsh::green fluorescent protein (GFP) C-terminal fusion, driven by native dsh regulatory sequences. One or two copies of these transgenes rescue dsh^v26 null mutants to viability and produce wild-type PCP, indicating that they fully replace the function of endogenous Dsh in both Wg and PCP signaling (Axelrod, 2001).

Fz and Fmi colocalize at proximal-distal boundaries at 30 h apf. Furthermore, the asymmetric pattern of Fz localization depends on Flamingo (Fmi), whereas the asymmetric pattern of Fmi localization depends on Fz. These data are consistent with the possibility that Fz, Dsh, and Fmi function together, perhaps in a complex, during PCP signaling, with both Fz and Dsh localizing to the distal edge, and Fmi apparently localizing to both the proximal and distal edges of the cell. A mutual dependence for asymmetric localization exists between these three proteins (Axelrod, 2001).

It was asked whether Dsh localization depends on upstream signaling through the Fz/PCP pathway by examining Dsh localization in a fz mutant background. In a fz^R52 null mutant, Dsh fails to accumulate at the membrane at 30 h apf. At 2 h apf, only the weak, perimembranous, methanol-sensitive enrichment of Dsh, reminiscent of that seen in wild-type third-instar discs, remains. Absence of membrane-associated Dsh from around 2 h apf through 30 h apf indicates that both the earlier, symmetric phase of Dsh-membrane association, as well as the late, asymmetric phase, are Fz dependent. Dsh-membrane association depends not simply on the presence of Fz protein, but also on its ability to signal. Disrupting the ability to localize Dsh to the membrane, either by mutating Dsh (dsh¹) or by blocking Fz function, produces a mutant PCP phenotype. Dsh-membrane association is therefore necessary to transduce the polarity signal (Axelrod, 2001).

To determine whether Fz signaling is sufficient to produce the asymmetric localization of Dsh, its localization was examined in a Fz expression gradient that alters the polarity pattern on the wing. Consistent with previous demonstrations, graded expression of ectopic Fz in the dpp or dll expression domains reorients hairs from high to low levels of Fz expression. In these wings, asymmetric Dsh localization realigns according to the Fz gradient. Therefore, both the membrane localization of Dsh and its asymmetry are dependent on signaling through Fz. In contrast, Dsh localization is normal in a mwh mutant, consistent with previous arguments placing Dsh upstream of Mwh in the polarity signaling pathway (Axelrod, 2001).

Dsh may translocate to the membrane from an existing pool, or may be stabilized at the membrane, increasing the total cellular Dsh content. Furthermore, Dsh is a phosphoprotein, and its phosphorylation state is potentially regulated during PCP signaling. Western blot analysis was therefore used to examine Dsh protein levels and phosphorylation state in pupal discs during PCP signaling. No significant difference in total Dsh levels was observed in wild type, fz^R52, or dsh¹ wings, indicating that membrane association represents a shift in Dsh localization from the cytoplasmic to the membrane compartment. However, more than half of the Dsh protein in wild type is in a hyperphosphorylated form, whereas very little of this form exists in fz^R52 or dsh¹ mutants. The PCP signal therefore results in phosphorylation of Dsh, and phosphorylation correlates with membrane localization, suggesting it is either required for, or is a response to, localization. This result is consistent with studies in Xenopus showing that XDsh phosphorylation and membrane association correlate with activity in convergent extension, a process homologous to PCP signaling, but not axis duplication, a ß-catenin mediated process (Axelrod, 2001).

Although both Fz and DFz2 transduce the Wg signal, only Fz can serve as a receptor for PCP signaling. Analysis of chimeras points to structural differences distal to the ligand binding domains as responsible for this difference. However, the question of how Fz specifically transduces two distinct signals, both of which require Dsh function, still remains. During late third instar, Wg signals through both Fz and DFz2 to establish the proneural clusters that give rise to bristles near the D/V boundary of the wing. However, no accumulation of Dsh is observed at membranes near the D/V boundary of third-instar wing discs. Furthermore, Dsh is not observed at membranes in embryos, nor in wing discs throughout third instar. During early pupal stages, when Dsh shows the earliest Fz-dependent membrane localization, no difference is observed between cells close to Wg expressing cells and those at greater distances. Recruitment of Dsh to the membrane is therefore a specific response to the Fz/PCP signal, and does not result from the Wg signaling activity of either Fz or DFz2 (Axelrod, 2001).