Tài liệu Báo cáo khoa học: Diego and friends play again Old planar cell polarity players in new positions doc

Thus, in this review we focus only on a selected set of unexpected recent findings including a the discovery of a new role for dgo in the apical recruitment and⁄ or maintenance of PCP com

Diego and friends play again

Old planar cell polarity players in new positions

Jo´zsef Miha´ly, Tama´s Matusek and Csilla Pataki

Institute of Genetics, Biological Research Center, Hungarian Academy of Sciences, Szeged, Hungary

Functional tissues are comprised of polarized cell

types Cellular polarization can be manifested in

many different ways, depending on the orientation

and axis of polarity Well known examples include

the Drosophila ovary and embryo, where all major

body axes are determined in a single cell; neuronal

cells that typically exhibit axonal-dendritic polarity

and epithelial cells that are characterized by

apical-basal polarity In many instances, however, tissue

dif-ferentiation also requires the coordination of cell

polarity within the plane of a tissue – a feature

referred to as planar cell polarization (PCP) or tissue

polarity for short Although PCP can be observed

throughout the animal kingdom (vertebrate examples

include fish scales, bird feathers and hairs in

mam-mals, or the neurosensory epithelium in the inner

ear), the regulation of such coordinated cell

polariza-tion events has been best studied in the fruitfly,

Dro-sophila melanogaster

PCP in flies is most evident in the wing, which is covered by uniformly polarized, distally pointing hairs,

in the epidermis, where sensory bristles and trichomes point to the posterior, and in the eye, where PCP results in a mirror symmetry arrangement of the ommatidia or unit eyes Polarization in these tissues is controlled by the gene products of the PCP genes, mutants of which impair planar organization Some of the PCP genes, which have been placed into the core group, appear to affect polarity in all of the tissues, whereas others function in a tissue-specific way The core group includes the seven-pass transmembrane receptor frizzled (fz), the cytoplasmic signal transducer dishevelled (dsh), the cytoplasmic LIM domain protein prickle (pk), the atypical cadherin flamingo (fmi), the four-pass transmembrane protein strabismus (stbm) and the ankyrin repeat protein diego (dgo) [1–9] Gen-etic analysis of the PCP genes indicates that polarity establishment can be subdivided into three major steps

Keywords

Diego; Drosophila; Four-jointed; inturned;

tissue polarity

Correspondence

J Miha´ly, Institute of Genetics, Biological

Research Center, Hungarian Academy of

Sciences, H-6726 Szeged, Temesvari krt.

62, Hungary

Fax: +36 62 433503

Tel: +36 62 599687

E-mail: mihaly@brc.hu

(Received 21 February 2005, accepted

27 April 2005)

doi:10.1111/j.1742-4658.2005.04758.x

The formation of properly differentiated organs often requires the planar coordination of cell polarization within the tissues Such planar cell polar-ization (PCP) events are best studied in Drosophila, where many of the key players, known as PCP genes, have already been identified Genetic analy-sis of the PCP genes suggests that the establishment of polarity conanaly-sists of three major steps The first step involves the generation of a global polarity cue; this in turn promotes the second step, the redistribution of the core PCP proteins, leading to the formation of asymmetrically localized signa-ling centers During the third step, these complexes control tissue-specific cellular responses through the activation of cell type specific effector genes Here we discuss some of the most recent advances that have provided valuable new insight into each of the three major steps of planar cell polarization

Abbreviations

MF, morphogenetic furrow; PCP, planar cell polarization.

First, a long-range polarity signal is set up At present,

the molecular nature of this signal is unclear, but it is

believed that the atypical cadherins fat (ft) and

dach-sous (ds), the type II transmembrane protein

four-join-ted(fj) and the transcriptional repressor atrophin (atro)

are all involved in the generation or the modulation of

this long-range positional cue [10–20] In the second

major step, the core PCP proteins redistribute and

build up asymmetrically localized multiprotein

com-plexes Finally, these apical membrane-associated

signaling centers control the tissue-specific cellular

responses through the activation of cell-type specific

effectors

While this remains a very general model that

ignores important details, many such details, together

with exciting new findings suggesting that the core

mechanisms of PCP regulation are conserved from

flies to human, have recently been summarized in

excellent reviews [21–25] Thus, in this review we

focus only on a selected set of unexpected recent

findings including (a) the discovery of a new role for

dgo in the apical recruitment and⁄ or maintenance of

PCP complexes [26]; (b) the demonstration that

secretion is not required for Fj function, but instead

it acts intracellularly in the Golgi apparatus [27];

and (c) the finding that the inturned (in) gene

prod-uct is localized proximally in the wing cells [28],

although it was previously considered to be a PCP

effector directly regulating wing hair outgrowth at

the distal vertex

PCP in the Drosophila wing and eye

In Drosophila, tissue polarity has been studied in a

number of different body regions but it is been best

understood in the wing and eye In the wing, each

cell produces a single distally oriented hair at its

distal vertex (Fig 1A) These structures are apical

membrane outgrowths that are stiffened by actin and

microtubule elements Mutations in PCP genes

dis-rupt wing hair polarity in several different ways

(Fig 1B) Some of them, such as fz, alter hair

orien-tation and also in this case hairs often form in the

center of the cell instead of the most distal part [29]

Certain other mutations, however, such as in, multiple

wing hair (mwh) and fuzzy (fy) do not affect hair

ori-entation, but result in the formation of multiple hairs

from a single cell [29,30] Thus, PCP genes appear to

regulate three major aspects of wing hair

develop-ment: they restrict wing hair outgrowth to the

distal-most part (distal vertex) of the cell, they control hair

orientation and they determine the number of hairs

produced

In contrast to the wing, where each individual cell normally becomes polarized, PCP in the eye is reflected

in the arrangement of a group of cells corresponding

to the unit eyes called ommatidia (Fig 1C) Each ommatidium consists of 20 cells including eight photo-receptor cells and 12 supporting cells Sectioning of the adult eye reveals that the ommatidia are chiral struc-tures as photoreceptor cells aquire an asymmetrical trapezoidal shape within each of these multicellular units [31] Interestingly, the ommatidia in the dorsal half of the eye all adopt the same chirality, but this is opposite to that adopted in the ventral half resulting

in a mirror symmetry arrangement (Fig 1D) The line where the dorsal and ventral ommatidia meet corres-ponds to the dorsal–ventral midline, often called the equator This spectacular planar organization is settled during imaginal disc development, a few hours after the photoreceptor preclusters emerge from the mor-phogenetic furrow (MF) of the eye-antennal disc The preclusters first become asymmetric and adopt dorsal

or ventral chirality; subsequently they rotate accord-ingly, i.e 90
clockwise in the dorsal clusters and 90
counterclockwise in the ventral clusters (Fig 1C) It has recently become clear that the key to PCP genera-tion in the eye is the step deciding the fate of the R3⁄ R4 photoreceptor cells It has been demonstrated that an fz PCP pathway dependent cell fate specifica-tion in the R3⁄ R4 pair is required for correct chirality choice and rotation of the whole ommatidial cluster [32,33] In agreement with this, PCP mutations can alter the chirality choice, resulting in chirality flips or symmetric ommatidia, and can also lead to various rotation defects (Fig 1D,E)

Asymmetric localization of the core PCP proteins

A major breakthrough towards an understanding of the molecular mechanisms controlling PCP came with the discovery that the core PCP proteins accumulate asymmetrically in cells [7,9,16,34–42] The first set of key observations established that, although the core PCP proteins in the wing cells are initially found in uniformly distributed complexes, a few hours before prehair formation they undergo relocalization and become differentially enriched along the proximal– distal axis, displaying a peculiar zigzag pattern Fz and Dsh become localized to distal cell membranes, whereas Stbm and Pk localize to the proximal side, while Fmi is found on both sides of the apical mem-brane (Fig 2A) Asymmetric PCP protein distribution can also be observed in the developing eye disc [16,37,38,40], although it is only in the precursor cells

of the R3⁄ R4 photoreceptors where protein

relocaliza-tion takes place leading to a transiently asymmetric

localization Interestingly, Fz and Dsh become

locali-zed on the R3 side, Stbm and Pk on the R4 side, and

Fmi on both sides of the R3⁄ R4 boundary (Fig 2B)

Thus, the PCP protein distribution at the R3⁄ R4

boundary in the eye displays striking similarities to

that of the distal⁄ proximal cell border in the wing

and hence the R3⁄ R4 cell boundary appears to be

functionally equivalent to the distal⁄ proximal cell

boundary in the wing Consistent with the protein

distribution in the eye, for fz and dsh there are genetic

requirements in R3 [32], for stbm and pk in R4 [5,41],

and for fmi in both R3 and R4 [37]

It has been demonstrated that all the PCP proteins

are required for the correct localization of each of the

others, suggesting that these molecules might act

together in a multiprotein complex However, detailed phenotypic analysis in the wing and eye has revealed that the different proteins might play different roles in the process of PCP protein localization In the wing, this is suggested by the fact that, while some PCP mutations (e.g fmi) impair the apical localization of the other proteins, others (e.g dsh) merely affect the asymmetric enrichment of PCP proteins without dis-rupting their apical localization In the eye, different PCP mutations affect the localization of the other PCP proteins in markedly different ways: (a) the apical pro-tein localization is compromised; (b) the asymmetric pattern is lost, but the apical localization remains unaffected; (c) asymmetric enrichment occurs, but in random orientation with respect to the equator (result-ing in chirality flips) Together, these observations sug-gest that PCP protein localization can be divided into

MF

B

D

3

C A

Normal R3/R3 R4/R4 Dorsal-Ventral

inversion

Misrotation E

Chirality flip

Dorsal

Ventral

Equator

1 2 3 5

6

7/8

1

2 5

6

7/8

Fig 1 Planar polarity and PCP phenotypes in the wing and eye (A) The establishment of PCP in the wing begins with actin accumulation at the distal vertex (middle cartoon) that will subsequently lead to the formation of a distally pointing hair (shown in green) (B) The absence of PCP genes can affect hair formation in different ways Hairs are sometimes disoriented, and the site of hair outgrowth is often not restricted

to the distal most part of the cell, or multiple hairs form in a single cell (mutant forms are indicated in red) (C) Ommatidial preclusters emerge from the morphogenetic furrow (MF) of the eye disc and initially form symmetric structures As eye development proceeds preclus-ters rotate 90
towards the equator, i.e dorsal cluspreclus-ters rotate clockwise, while ventral ones rotate counterclockwise At the end of this pro-cess the R3 ⁄ R4 cell pair acquires an asymmetric position within the cluster, and thus chirality also becomes established (R3 cells are highlighted in green, R4 cells in red) (D) The mirror symmetric structure of an adult eye can be disrupted by PCP mutations that can cause rotation defects, dorsal-ventral inversions, and loss of chirality resulting in symmetrical ommatidia with either R3 ⁄ R3 or R4 ⁄ R4 cell pairs (see enlarged on E).

two main phases: proteins first become localized to

adherens junctions in the apicolateral membrane, and

in the second stage they become asymmetrically

distri-buted along the proximal–distal axis in the case of the

wing, or on the R3⁄ R4 cell boundary in the eye

Addi-tionally, the asymmetric distribution in the eye must

be coordinated with respect to the dorso–ventral

mid-line

Apical localization requires Diego

What is the molecular mechanism that ensures the api-cal loapi-calization of the PCP protein complex, and how

is membrane recruitment achieved for the predicted cytoplasmic proteins Dsh, Pk and Dgo? It was recently proposed by Bastock et al [42] and subsequently reviewed in detail by Strutt [24] that the PCP proteins might act in a hierarchy to generate asymmetrically localized apicolateral complexes (Fig 3A) This model postulates that Fmi acts at the top of the hierarchy and is responsible for recruiting the other transmem-brane proteins, Fz and Stbm This is supported by the finding that in the absence of fmi negligible amounts

of any other PCP, protein can be detected in the apicolateral region (Table 1) In the simplest case, Fmi would recruit Fz and Stbm by direct protein–protein interactions, although no direct binding partner has so far been found for Fmi Despite this discrepancy, it is now well established that Fmi, Fz and Stbm are cer-tainly required for the membrane recruitment of the putative cytoplasmic proteins, Dsh, Pk and Dgo (Table 1) In accord with this, Fz has been shown to bind Dsh [43] and is able to recruit Dsh to membranes

in heterologous assays [44] Furthermore, physical interactions have been reported between Stbm and Dsh, and between Stbm and Pk [41,42], suggesting a model in which at least Dsh and Pk become apicolater-ally localized due to direct binding to Fz and Stbm Because in the absence of Dsh, Pk or Dgo, apicolateral recruitment of the other PCP proteins is not affected, but their asymmetric redistribution does not take place [9,42] (Table 1), it seemed reasonable to assume that, although Dsh, Pk and Dgo play negligible roles in apicolateral recruitment, they are required to promote the assembly and⁄ or the maintenance of asymmetric PCP complexes

An interesting recent paper [26] has now questioned this simple interpretation and presented new insight into the mechanisms regulating the initial apical local-ization and subsequent maintenance of PCP com-plexes The research in this paper is focused on dgo, the least well characterized core PCP gene Previous work has shown that Dgo is colocalized with Fz and Fmi during polarity establishment in the wing, and apical Dgo localization depends on these proteins [9]

At that time, however, it was not possible to determine the precise subcellular localization of the dgo gene product Das et al have now reported that Dgo accu-mulates on the distal side of the wing cells Not sur-prisingly, in the eye Dgo becomes enriched on the R3 side of the R3⁄ R4 cell boundary, and it follows that, just like fz and dsh, dgo is genetically required in R3

2 3

8

2 3

4 5 8

Fz

Dsh

Fmi

Stbm

Pk

Dgo

A

B

32h APF 24h APF

Apical

Basal

Apical

Basal

Equatorial

Polar

2 3

8

2 3

4 5 8

Fig 2 Core PCP protein localization in the developing wing and

eye (A) During the initial phase of pupal wing development [up to

 24 h after prepupa formation (APF)] the protein products of the

core PCP genes are found in apically localized symmetric

com-plexes (shown on the left) However, at  24 h APF they

redistrib-ute into asymmetric complexes that are present transiently until

actin accumulation begins at  32 h APF Between 24 and 32 h

APF Fz, Dsh and Dgo are enriched on distal cell membranes, Stbm

and Pk accumulate on proximal membranes, while Fmi is found on

both sides (right panel) (B) Although core PCP protein localization

in the eye is somewhat more complicated than in the wing, it

appears that PCP protein distribution across the R3 ⁄ R4 cell

bound-ary is remarkably similar to that of the distal–proximal cell

boundar-ies in the wing Notably, after the initial phases of ommatidia

differentiation when PCP proteins do not show polarized

accumula-tion, five or six rows behind the morphogenetic furrow Fz, Dsh and

Dgo begin to preferentially accumulate on the R3 side, whereas

Stbm and Pk accumulate on the R4 side, and Fmi becomes

enriched on both sides of the R3 ⁄ R4 interface Developing

ommati-dia are shown in five-cell precluster stages before and after

asym-metric redistribution takes place, Row 4 and Row 7, respectively.

Color code of the PCP proteins is identical in both (A) and (B)

Num-bers on (B) indicate the identity of the photoreceptor precursor

cells.

The absence of dgo does not affect the apical

localiza-tion or the asymmetric enrichment of PCP proteins in

the eye, but the asymmetric accumulation is apparently

randomized with respect to the equator In contrast,

the apical localization of Dgo is completely abolished

in an fz mutant tissue, and strongly reduced in fmi

clones, while the absence of stbm or pk although

indu-ces a short delay in Dgo localization, the overall

pattern remains largely normal, albeit randomized

Strikingly, however, in dgo, pk or dgo, stbm double

mutant clones, the apical localization of both Fz and

Fmi is strongly reduced Additionally, in the dgo, pk

combination, Stbm and Dsh also fail to form apically

localized complexes, although this might not reflect a

direct requirement as Fmi localization is compromised

as well The situation with pk, stbm double mutants is

more complex because the apical localization of Fmi

and Fz in the eye is lost anterior to the MF, whereas

the apical localization is hardly affected posterior to

the furrow even if the asymmetric distribution is

per-turbed [26] In contrast to the eye, Fmi and Fz

local-ization is severely reduced in pk, stbm double mutant

wing cells [42] Finally, despite the fact that single

mutants of pk and stbm do not significantly affect

api-cal Dgo loapi-calization in the eye, in pk, stbm double

mutant clones Dgo is much reduced at the apical

cor-tex Significantly, it has also been revealed by yeast

two-hybrid and GST pull-down assays that Dgo

inter-acts physically with Pk and Stbm

These results suggest that, opposite to what might

be expected from single mutant analysis, Pk and Dgo are also required for membrane localization of the PCP factors, though this is a redundant requirement between Dgo, Pk and Stbm These data led Das et al [26] to outline a model to explain how PCP complexes might be formed and maintained during the early phases of PCP establishment (Fig 3B) They propose that the cytoplasmic PCP proteins (Dsh, Dgo and Pk) initially recruited to the membrane by Fz and Stbm form a protein complex that is required to maintain Fmi apically In turn, apical Fmi promotes the main-tenance of the PCP complex at adjacent cell mem-branes and can also facilitate their signaling specific interactions This model is consistent with the available data and also supported by the protein–protein inter-action results However, in so far as Fmi is concerned, Das et al came to just the opposite conclusion to that

of Bastock et al who suggested that Fmi lies at the top of the hierarchy of apical PCP protein recruitment [42] In that view, Fmi is required for the initial mem-brane recruitment of Fz and Stbm (Fig 3A) While these opposing views might simply reflect tissue-specific differences between the eye and the wing, an fz, stbm double mutant analysis could be informative in respect

of the order of initial membrane recruitment If Fmi is

at the top, in such fz, stbm double mutants Fmi local-ization should not be significantly affected, whereas if

Fz and Stbm were the initial recruiters (as suggested

Stbm

Pk

Fmi

Fz

Dgo Dsh

Stbm

Pk

A

Dsh Dgo

Stbm

Pk

Dgo

Dsh

Stbm

Pk

B

Fig 3 Two possible models of apical PCP protein recruitment (A) One model, mainly based on data in the wing, proposed that Fmi lies on the top of the hierarchy of apical recruitment, and it is responsible for recruiting Fz and Stbm (top panel) Subsequently, these membrane proteins would recruit the putative cytoplasmic proteins, Dsh, Dgo and Pk (bottom panel) (B) The second model, based on eye data, sug-gests that Fz and Stbm would be the initial membrane recruiters of Dsh, Dgo and Pk (top panel), and these proteins would then be required

to maintain Fmi apically (blue arrows, bottom panel) In turn, Fmi would promote the maintenance of the whole core PCP complex at adja-cent cell membranes Black arrows represent the genetic requirements for apical recruitment, grey ovals indicate the nuclei (A) and (B) are modified figures after Bastock et al [42], and Das et al [26], respectively.

by Das et al.), apical Fmi should be lost in the fz, stbm

double mutant clone In fact, in the wing, unlike the

situation in the eye, stbm moderately reduces the level

of apically localized Fmi and fz also has a weak effect,

which could be used as an argument in favor of Fmi

localization being dependent on Fz and Stbm Such

simple assumptions, however, must be treated with

caution because one of the limitations of the genetic

approaches used during these experiments is that they

do not clearly distinguish between initial apical

recruit-ment and maintenance At present therefore it is not

possible to distinguish between the two alternatives

that have been put forward to explain the apical

recruitment and maintenance of PCP complexes

Moreover, as we know very little about the molecular

composition of the PCP complexes formed in vivo and

the feedback mechanisms that might help to stabilize them, it is clear that other models are also possible Nevertheless, the employment of double mutant analy-sis has proved to be a very useful tool to discover new aspects of PCP establishment in the Drosophila wing and eye It seems likely that the examination of further double mutant combinations will promote a deeper understanding of this process It would be interesting

to examine double mutant combinations between the

fz, dsh and the pk, stbm, dgo groups as single mutants

of these in most cases have either no effect or only a weak effect on apical localization Finally, it would also be of interest to compare the results of double mutant analyses in the wing and eye as this could yield further hints concerning the tissue specific differences already revealed by single mutant analysis

Table 1 Core PCP protein localization in the wing and eye in PCP mutant backgrounds This table summarizes the relevant aspects of pro-tein localization in the wing and the eye [7,9,26,34–42] Api., apical localization; Asy., asymmetric redistribution; D, normal but delayed local-ization; ND, not determined The ratio between filled and empty circles indicates the amount of properly localized proteins compared with wild-type level: ddd, wild type level; sss, complete loss of localization.

a Reduced Pk level at the apical membrane, but increased level in the cytoplasm [40–42] b Data not shown in [38] c Normal apical localiza-tion behind the MF, but loss of apical localizalocaliza-tion anterior to the MF [26].

Long-range patterning and the

Golgi-associated protein, Four-jointed

A few hours after PCP proteins have been recruited to

the apicolateral regions, they become asymmetrically

distributed How does this happen? Although the

answer to this question is largely unclear, it is believed

that redistribution occurs in response to a directional

signal that coordinates polarity with the axis of the

tis-sue It is also generally thought that the polarity signal

induces a bias in Fz activity along the proximal⁄ distal

axis of the wing and the equatorial⁄ polar axis of the

eye [16,18,36] Subsequently, the initially created subtle

difference in Fz activity on the opposite sides of the

cells would be amplified by intercellular feedback

mechanisms, leading to high Fz signaling on the distal

(i.e in the wing) and equatorial (i.e in the eye) sides,

and to low level signaling on the opposite sides [39]

While this is an attractive model, there are several

important points that need to be verified It is not yet

clear what the link is between differential Fz activation

and asymmetric redistribution What is the cause and

what is the consequence here, if there is a direct casual

relationship at all? Another important problem is that

the source and nature of the polarity signal remain

elu-sive, including the important question of whether it is

a long-range or a short-acting signal Models based on

the former possibility propose that polarity is

estab-lished as a result of interpreting the concentration of a

long-range signal (most probably a secreted factor)

present in a concentration gradient across the tissue

Alternatively, a locally acting short-range signal could

be used to polarize one cell, which would in turn

gen-erate a signal to polarize its neighbors via a signal

relay mechanism Finally, we note that there are

pro-found tissue-specific differences between the wing, eye

and abdomen, and thus the in vivo mechanism could

change from tissue to tissue, including a combination

of the long- and short-range models

Despite the fact that the molecular nature of the

mysterious polarity cue (often called factor X) is not

known, several genes have recently been implicated in

long-range signaling acting upstream of asymmetric

PCP protein redistribution A great body of work on

the developing wing, eye and abdomen has led to a

model in which the activity gradients of the atypical

cadherins ft and ds and the type II transmembrane

protein Fj generate or modulate the activity of a

long-range polarity signal [11–20] A feature almost

cer-tainly relevant to this issue is that the same Ft⁄ Ds ⁄ Fj

module is involved in the proximal-distal patterning of

the wing and leg [12,45–48], and hence it is tempting

to speculate that planar polarity establishment (at least

in the wing) is directly coupled to growth and pattern-ing of the tissue (see also [23]) Direct evidence in support of this view is missing, however, indicating that further work will be required to clarify the link between PCP and tissue patterning Whether the activity of the Ft⁄ Ds ⁄ Fj module will ultimately lead to the secretion of an Fz ligand or coordinate PCP in a different way is also an open question It appears that

to resolve these problems we need a better understand-ing of the signalunderstand-ing events between these proteins and, potentially, other pathways and proteins; and there is

a need to learn more about the molecular biology and biochemical properties of these proteins Some pioneer-ing experiments have already provided evidence that ft acts through the transcriptional corepressor Atro [17], support for the in vivo existence of a Ft–Ds hetero-philic interaction [19], and last but not least, revealed that Fj is a Golgi-associated protein [27]

Former studies demonstrated that fj is expressed in

a gradient in the developing wing and eye [11,12], and that clones of cells which either lack or ectopically express Fj cause both autonomous and nonautono-mous PCP defects in these tissues [11,12] Additionally,

an in vitro analysis has indicated that the extracellular C-term of the Fj protein can be cleaved, resulting in a secreted form [47] Together, these results strongly sug-gested that Fj functions as a secreted signaling mole-cule present in a gradient on the proximo–distal axis

of the wing and the dorso-ventral axis of the eye Strutt et al have now tested this idea directly by com-paring the signaling abilities of modified Fj forms that are either poorly cleaved, constitutively secreted or anchored to the Golgi [27] During this elegant set of experiments, they used overexpression and rescue assays to test the in vivo activities of the different Fj forms and concluded that secreted Fj is not the active form, but, unexpectedly, Fj acts intracellularly This is consistent with their antibody staining result that most

Fj is localized to discrete spots inside the cells, the majority of which correspond to the Golgi apparatus Significantly, they also demonstrated that, although the Golgi-tethered Fj is not secreted, when over-expressed it is still able to produce nonautonomous polarity phenotypes To explain this finding, they pro-pose that Fj most likely acts by modulating the activity

of other proteins involved in intercellular signaling The best candidates as targets of Fj are the atypical cadherins Ds and Ft, which have been shown to act downstream of fj [16] The analysis of fj mutant clones revealed that Fj is clearly involved in the control of cell adhesion, and it regulates the intracellular distribu-tion of Ds and Ft, which are likely to bind to each other in vivo [15,18,19] This set of results has led to

the proposal that Fj may regulate cell adhesion by

modulating Ds⁄ Ft heterophilic interactions [15,18,19]

Although the molecular mechanism of this modulation

remains to be discovered, it seems conceivable that Fj

may have an enzymatic activity that is involved in the

post-translational modification of Ft and⁄ or Ds (or an

unidentified protein), in much the same way as Fringe

(Fng) regulates Notch (N) activity [49,50]

The idea that Fj mediates the post-translational

modification of Ft and⁄ or Ds has interesting

implica-tions as regards the potential redundancy at the level of

Fj function As ft and ds exhibit strong planar polarity

phenotypes when homozygous [10,14,16], while fj

dis-plays polarity phenotypes almost only on the

boundar-ies of mutant clones [12], the weak fj phenotype was

explained by redundancy However, the origin and

nature of this redundancy remain uncertain, largely

because of the lack of information on the molecular

function of fj In the light of recent data, it may be

spe-culated that, while Fj modulates Ft⁄ Ds activity, for

example, by adding certain type of post-translational

modifications on it, full Ft⁄ Ds activity would require

additional upstream inputs (e.g additional types of

post-translational modifications) Thus, Fj activity

would contribute to the activation of Ft⁄ Ds, but alone

it would not be sufficient to create a fully functional

form As a consequence, the absence of Fj would be

compensated by additional, an as yet unidentified

fac-tors also involved in Ft⁄ Ds modulation

Although the precise mechanisms by which these

ele-ments contribute to the formation of global polarity

cues have not been fully clarified, experiments of this

type offer promising new developments in the PCP

field, and underline the importance of biochemical

approaches in elucidating further details of PCP

estab-lishment

Inturned: a new turn in the game

While much attention has recently been paid to the

asymmetric localization of the PCP proteins and to

the mechanisms of action of the potential upstream

elements, novel studies on inturned (in), a planar

polarity effector gene, have led to the discovery of a

new level of complexity among the downstream PCP

elements as well Earlier experiments revealed that, in

the absence of in or fy, wing cells fail to restrict hair

outgrowth to the distal vertex; instead, they form

multiple hairs at ectopic locations [29] It has also

been shown that the in and fy loss-of-function

muta-tions are epistatic to fz, dsh and pk, suggesting that

they function downstream of the core PCP genes [29]

A simple interpretation of these results is that in and

fy act as inhibitors of hair formation, while one poss-ible function of fz would be to inhibit in and fy loc-ally to allow prehair initiation at the distal vertex [29] Although this hypothesis is consistent with the observations that the asymmetric accumulation of Fz and Dsh is not altered in fy or in mutants [34–36], the molecular details of the potential inhibitory mech-anisms were entirely missing It has now been repor-ted that the In protein becomes preferentially accumulated at the proximal edge of the pupal wing cells under the instruction of the core PCP genes [28] This pattern which is apparent several hours before prehair initiation, closely resembles the zigzag pattern typical for core PCP proteins; indeed, the In staining largely overlaps with that of Fz At present, it is not clear how In is recruited to the proximal side of the wing cells One tempting possibility is that In is recruited by Pk or Stbm as these proteins also accu-mulate there [39,42]; however, it was not possible to detect a direct interaction between In and Pk or Stbm with the yeast two hybrid system [28] Although such

an interaction is still possible, another alternative would be that Fuzzy (Fy) or Fritz (Frtz) or both are involved in In recruitment and localization This would be consistent with several different findings First, asymmetric In accumulation and In protein sta-bility have both been shown to depend on fy and frtz [28] Second, although it has not been studied in detail, fy appears to encode a transmembrane protein that could recruit In directly [51] Third, single or double mutants of fy, frtz and in display almost iden-tical wing hair phenotypes [29,52], suggesting that these genes function together in PCP and might inter-act with each other directly As we know very little

at the molecular level about Fy and Frtz, and their subcellular localization has not been described yet, In recruitment remains an open issue We note, however, that the combination of the above alternatives is also

a valid possibility In that scenario, Stbm and⁄ or Pk would be the key to the initial proximal recruitment

of In or Fy or Frtz, which in turn would promote the assembly of a functional In⁄ Fy ⁄ Frtz complex Previous observations have suggested that the spatial restriction of cytoskeleton activation and prehair for-mation to the distal vertex of the wing cells largely or entirely depends on the local (distal) accumulation and activation of Fz and Dsh, whereas the proximal accu-mulation of other PCP proteins has been only sugges-ted to play a role in the establishment of proximal and distal cortical domains It has now been found that local Fz⁄ Dsh signaling alone is not sufficient to restrict prehair formation to the distal vertex: surprisingly, this signaling has to be coordinated with downstream

regu-latory events that depend on the proximally localized

factor, Inturned [28] How can this proximally

recrui-ted protein ensure that hairs form at the distal edge?

Adler et al considered three alternatives: proximal In

might stimulate hair formation at the distal edge of

the neighboring cells; alternatively, proximal In might

organize the polarized intracellular transport of

cellu-lar components that play a role in hair morphogenesis;

and finally, the old idea that proximal In might

func-tion as an inhibitor of hair initiafunc-tion The available

data do not permit a clear distinction between these

alternatives, though the first appears very unlikely as it

predicts a strong nonautonomous effect, which has not

been seen in in mutants It is difficult to raise formal

arguments against the second alternative, but the third

seems to be the most appealing to us Following the

finding that In is proximally localized in wing cells, the

idea that In functions as an inhibitor, offers a refined

view of the proximal and distal cortical domains

According to this hypothesis, once the core PCP

pro-teins have redistributed asymmetrically, the proximal

and distal cortical domains become established

Subse-quently, Fz⁄ Dsh promotes hair formation at the

distal-most part of the cell, but only there, while at the same

time Stbm⁄ Pk inhibits hair formation in the proximal

part, acting through an In complex Thus, this model

predicts that the restriction of hair outgrowth

exclu-sively to the distal vertex of the cell requires that the

positively acting distal Fz⁄ Dsh PCP signal is paralleled

by an In inhibitory signal on the opposite side of the

cell that might form an intracellular gradient with its

high end at the proximal pole Interestingly, this

hypo-thesis may help to explain a previously described set of

results that were difficult to reconcile with a simple

lin-ear regulatory relationship between the core elements

and the in-like genes Notably, Lee and Adler reported

that a weak multiple wing hair phenotype induced by

hypomorphic in or fy alleles is strongly enhanced both

by the removal and by the overexpression of fz, fmi or

dsh[52] If we consider that, in the absence of fz, the

site of actin accumulation is already somewhat

‘delo-calized’, leading to a weak multiple hair phenotype,

the inhibitory model predicts that the introduction of

a hypomorphic in allele into this background will

enhance the multiple wing hair phenotype, because the

proximally acting inhibitor of cytoskeleton activation

is also impaired This prediction fits perfectly with the

published data If we now consider the case of Fz

overexpression, we find that the excess of Fz induces a

failure in asymmetric PCP protein redistribution,

lead-ing to an imperfect restriction of actin accumulation

even in the presence of In (which itself fails to undergo

proper localization in this situation) If Fz

overexpres-sion is accompanied by a parallel impairment of the inhibitor of hair formation (In), the original multiple wing hair phenotype is again expected to be enhanced,

as is indeed the case At present this model remains very speculative, but it is noteworthy that, if such opposite, but complementary In and Fz⁄ Dsh effects exist, this would represent a very powerful mechanism whereby hair initiation is restricted exclusively to the distal vertex, and it could contribute substantially to the remarkable precision with which the site of actin accumulation and, ultimately, wing hair number and orientation are determined Nevertheless, further experiments will clearly be required to resolve the intriguing questions of how the In⁄ Fy ⁄ Frtz module is linked to the core PCP complexes and what the in vivo function of In is

Perspectives

The earlier discovery that PCP proteins build up asym-metrically localized complexes was considered a major breakthrough in the field Much effort has subse-quently been devoted to elucidating how such com-plexes are formed and to clarifying the protein–protein interactions between the components However, despite the great progress regarding certain details, the overall picture remains unclear The mechanistic details on protein localization are largely missing, the link between Fz signaling and asymmetric localization is not understood, and it remains a mystery how asym-metric localization is coupled to upstream elements, such as the atypical cadherins It appears likely that the extension of double mutant analysis will furnish additional valuable insights, however, it is equally clear that genetic and cell biological methods must be com-bined with biochemistry in order to reach the heart of these problems A detailed biochemical characteriza-tion of the core group may allow a funccharacteriza-tional dissec-tion of the process Moreover, we need a better understanding of the protein–protein interactions between the core elements, we need to describe the complexes formed in vivo, and we need to learn about their spatial and temporal regulation during PCP establishment

A second area of PCP research that has recently received considerable attention is the generation of long-range polarity cues While much has been learnt, the mysterious story of factor X continues, as the exact nature, source and mode of action of the polarity sig-nal are still unknown Interestingly, however, import-ant new results (not discussed in this minireview) have led to the identification of further upstream elements, one of which (widerborst) is asymmetrically localized

long before the redistribution of the core elements [53],

while the other (encoded by the grainy head

transcrip-tion factor) is indirectly required for the apical

local-ization of the core PCP proteins via the regulation of

fmitranscription [54] These findings tend to lead us to

question the view that asymmetric core PCP protein

localization is the key to proximodistal polarization of

the wing cells It rather appears that proximodistal

polarity is established much before these players act

Quite conceivably, Lawrence et al have proposed that

cells are perhaps polarized throughout all or most of

their development [55], and hence the core PCP

pro-teins should already be regarded as effector elements

that begin to execute the final steps of tissue

differenti-ation, resulting in an overt manifestation of planar

organization In the course of this process, they appear

to organize downstream acting asymmetric complexes,

as revealed by the analysis of In Research in this line

is very likely to give rise to changes in the prevailing

models on the downstream effectors Overall, the PCP

field continues to provide ample opportunities for hard

work to generate a new wave of exciting results, bright

ideas and fun for developmental biologists

Acknowledgements

We are grateful to Henrik Gyurkovics, Pe´ter Vilmos,

La´szlo´ Sipos, Gishnu Das, Marek Mlodzik, David

Durham and one of the anonymous reviewers for their

helpful comments and critical reading of the

manu-script J M is an EMBO⁄ HHMI Scientist and a

Bol-yai Ja´nos Research Scholar, and supported by an NIH

FIRCA grant, and EMBO⁄ HHMI

References

1 Vinson CR, Conover S & Adler PN (1989) A Drosophila

tissue polarity locus encodes a protein containing seven

potential transmembrane domains Nature 338, 262–264

2 Klingensmith J, Nusse R & Perrimon N (1994) The

Drosophilasegment polarity gene dishevelled encodes a

novel protein required for response to the wingless

sig-nal Genes Dev 8, 118–130

3 Theisen H, Purcell J, Bennett M, Kansagara D, Syed A

& Marsh JL (1994) Dishevelled is required during

wing-less signaling to establish both cell polarity and cell

identity Development 120, 347–360

4 Taylor J, Abramova N, Charlton J & Adler PN (1998)

Van Gogh: a new Drosophila tissue polarity gene

Genet-ics 150, 199–210

5 Wolff T & Rubin GM (1998) Strabismus, a novel gene

that regulates tissue polarity and cell fate decisions in

Drosophila Development 125, 1149–1159

6 Chae J, Kim MJ, Goo JH, Collier S, Gubb D, Charlton

J, Adler PN & Park WJ (1999) The Drosophila tissue polarity gene starry night encodes a member of the pro-tocadherin family Development 126, 5421–5429

7 Usui T, Shima Y, Shimada Y, Hirano S, Burgess RW, Schwarz TL, Takeichi M & Uemura T (1999) Flamingo,

a seven-pass transmembrane cadherin, regulates planar cell polarity under the control of Frizzled Cell 98, 585– 595

8 Gubb D, Green C, Huen D, Coulson D, Johnson G, Tree D, Collier S & Roote J (1999) The balance between isoforms of the prickle LIM domain protein is critical for planar polarity in Drosophila imaginal discs Genes Dev 13, 2315–2327

9 Feiguin F, Hannus M, Mlodzik M & Eaton S (2001) The ankyrin repeat protein Diego mediates Frizzled-dependent planar polarization Dev Cell 1, 93–101

10 Adler PN, Charlton J & Liu J (1998) Mutations in the cadherin superfamily member gene dachsous cause a tissue polarity phenotype by altering frizzled signaling Development 125, 959–968

11 Zeidler MP, Perrimon N & Strutt DI (1999) The four-jointed gene is required in the Drosophila eye for ommatidial polarity specification Curr Biol 9, 1363–1372

12 Zeidler MP, Perrimon N & Strutt DI (2000) Multiple roles for four-jointed in planar polarity and limb pat-terning Dev Biol 228, 181–196

13 Casal J, Struhl G & Lawrence PA (2002) Developmen-tal compartments and planar polarity in Drosophila Curr Biol 12, 1189–1198

14 Rawls AS, Guinto JB & Wolff T (2002) The cadherins fat and dachsous regulate dorsal⁄ ventral signaling in the Drosophilaeye Curr Biol 12, 1021–1026

15 Strutt H & Strutt D (2002) Nonautonomous planar polarity patterning in Drosophila: dishevelled-indepen-dent functions of frizzled Dev Cell 3, 851–863

16 Yang CH, Axelrod JD & Simon MA (2002) Regulation

of Frizzled by fat-like cadherins during planar polarity signaling in the Drosophila compound eye Cell 108, 675–688

17 Fanto M, Clayton L, Meredith J, Hardiman K, Char-roux B, Kerridge S & McNeill H (2003) The tumor-suppressor and cell adhesion molecule Fat controls planar polarity via physical interactions with Atro-phin, a transcriptional co-repressor Development 130, 763–774

18 Ma D, Yang CH, McNeill H, Simon MA & Axelrod

JD (2003) Fidelity in planar cell polarity signalling Nature 421, 543–547

19 Matakatsu H & Blair SS (2004) Interactions between Fat and Dachsous and the regulation of planar cell polarity in the Drosophila wing Development 131, 3785– 3794