UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The role of Gliotactin in planar cell polarity Venema, Dennis Roy 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_2003-854140.pdf [ 10.48MB ]
JSON: 831-1.0091304.json
JSON-LD: 831-1.0091304-ld.json
RDF/XML (Pretty): 831-1.0091304-rdf.xml
RDF/JSON: 831-1.0091304-rdf.json
Turtle: 831-1.0091304-turtle.txt
N-Triples: 831-1.0091304-rdf-ntriples.txt
Original Record: 831-1.0091304-source.json
Full Text

Full Text

THE ROLE OF GLIOTACTIN IN PLANAR CELL POLARITY  by DENNIS ROY VENEMA B.Sc. (Honours), University of British Columbia, 1996  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS OF THE DEGREE OF  DOCTOR OF PHILOSOPHY  in THE FACULTY OF GRADUATE STUDIES Department of Zoology  We accept this thesis as corffbrming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA January, 2003 © Dennis Roy Venema, 2003  In presenting  this  degree at the  thesis  in partial fulfilment  University of  freely available for reference copying  of  department  this or  publication of  British Columbia, and study.  by  his  or  her  the  1 agree  requirements that the  representatives.  may be It  granted  is  this thesis for financial gain shall not be  Department The University of British Columbia Vancouver, Canada  for  an  advanced  Library shall make it  that permission for extensive  I further agree  thesis for scholarly purposes  permission.  DE-6 (2/88)  of  by the  understood  head  that  allowed without  of  my  copying  or  my written  ABSTRACT  Epithelial sheets exhibit apical-basal polarity and may also be polarized within the plane of the epithelium, a feature known as planar cell polarity, or P C P . In Drosophila, the pleated septate junction is a basolateral cell-cell junction that contributes to apicalbasal polarity. Gliotactin is a transmembrane protein with homology to serine esterases that localizes to the septate junction at the common vertex of three adjoining cells. Gliotactin is necessary for the function of the embryonic blood-nerve barrier which is formed by glial-glial septate junctions, as well as for septate junction formation in the embryonic epidermis. While septate junctions are a ubiquitous feature of epithelial sheets in Drosophila, the role of Gliotactin in post-embryonic tissues has not been addressed.  In this thesis the role of Gliotactin in the development of polarized post-embyronic epithelia was investigated. The Drosophila wing epithelium has both apical-basal and planar polarity. Through mutant analysis it was demonstrated that Gliotactin is required for P C P in this epithelium. The Gliotactin wing P C P phenotype was sensitive to enhancement by the septate-junction mutants scribble, discs-large, coracle and Neurexin IV, and by the putative septate-junction mutant vulcan. Through genetic tests these mutations were found to function in the Gliotactin P C P pathway in epistasis groups consistent with previous work in the embyonic epidermis. Combinatorial septate-junction mutations were found to cause blisters at the wing margin, identifying a role for Gliotactin and other septate-junction components in adhesion between wing epithelial  ii  layers. The  Gliotactin PCP pathway was shown to be independent of the frizzled PCP  signaling pathway through cell biological and genetic assays. Finally, a novel component of the  Gliotactin pathway was identified through a genetic screen for modifiers of  Gliotactin P C P phenotypes. A model of Gliotactin function in/h'zz/ed-independent PCP in the wing is presented.  Ill  Table of Contents ABSTRACT Table of contents  iv  List of figures  vi  List of tables  viii  List of abbreviations  x  I.  INTRODUCTION  1  II.  MUTAGENESIS OF GLIOTACTIN STRUCTURE A N D  III.  FUNCTION.  46  Introduction  47  Materials and methods  50  Results  54  Discussion  69  GLIOTACTIN IS NECESSARY FOR SEPTATE JUNCTIONMEDIATED P L A N A R C E L L POLARITY A N D ADHESION.  72  Introduction  73  Materials and methods  78  Results  83  Discussion  122  iv  IV.  V.  GLIOTACTIN DETERMINES PCP INDEPENDENTLY OF FRIZZLED SIGNALING.  134  Introduction  135  Materials and methods  137  Results  139  Discussion  145  A SCREEN FOR DOMINANT SUPPRESSORS OF GLIOTACTIN.  146  Introduction  147  Materials and methods  149  Results  152  Discussion  167  VI.  G E N E R A L DISCUSSION  169  VII.  REFERENCES  181  APPENDIX 1. DROSOPHILA  GENOTYPES  192  Gliotactin alleles  193  Septate junction and SJ-interacting mutations  195  Visible marker mutations  196  Balancer chromosomes  198  Planar cell polarity mutations  199  v  1  List of figures  Figure 1.  The Drosophila pleated septate junction  3  Figure 2.  The domain structure of Gliotactin  9  Figure 3.  Early events in prepupal wing morphogenesis  13  Figure 4.  Later events in prepupal and pupal wing morphogenesis  16  Figure 5.  The transalar cytoskeleton  19  Figure 6.  Planar cell polarity and prehair location in wild-type and fz mutant wings  24  Figure 7.  Fz, Dsh, Pk and Fmi localization during Fz PCP signaling  30  Figure 8.  A model of the frizzled PCP pathway  40  Figure 9.  An ethylmethane sulfonate (EMS) screen for novel Gli alleles  55  Figure 10.  Domain structure of Gli and molecular sites of Gli EMS alleles  60  Figure 11.  Gli mutants have altered leg morphogenesis and defects in PCP  63  Figure 12.  Wing PCP defects in Gli mutants  65  Figure 13.  Genetic interactions between SJ components in the embryonic epidermis Gli is expressed at the tricellular septate junction in the developing wing  Figure 14.  Figure 15.  74 84  Septate-junction mutations dominantly enhance the Gli wing PCP phenotype  86  Figure 16.  A cor hypomorph has defects in leg and wing PCP  89  Figure 17.  vie is a dominant enhancer of Gli  93  Figure 18.  cor does not enhance the phenotype of Afcx-enhanced Gli wings  99  vi  Figure 19.  dig dominantly enhances the phenotype of cor-enhanced Gli wings... 102  Figure 20.  Combinatorial mutant analysis between scrib and cor/Nrx  107  Figure 21.  Combinatorial mutant analysis between scrib and dig  109  Figure 22.  Combinatorial mutant analysis between vie and Nrx  113  Figure 23.  Combinatorial mutant analysis between vie and scrib  120  Figure 24.  Genetic interactions between SJ components in wing PCP  128  Figure 25.  A combined model of the septate junction  132  Figure 26.  Gli-mediated PCP is independent of Fz signaling  140  Figure 27.  Epistasis analysis between the Gli and/z-pathways  142  Figure 28.  An EMS screen for dominant suppressors of Gli lethality  153  Figure 29.  Mapping of Su(Gli)l  160  Figure 30.  The effect of Su(Gli)l on wing PCP and adhesion  164  Figure 31.  A model of/nzz/ed-independent PCP  177  vii  List of tables Table 1.  Interallelic complementation between Gli'' and homozygous10  lethal Gli alleles reveals an allelic series based on adult viability  58  Table 2.  Embryos heterozygous for Gli  59  Table 3.  Mutations in septate junction components dominantly enhance the  JvS  exit embryogenesis  wing P C P phenotype of specific Gliotactin genotypes Table 4.  88  Mutations in septate-junction components dominantly induce wing blisters in specific Gliotactin genotypes  92  Table 5.  vulcan is a dominant enhancer of the Gliotactin P C P phenotype  95  Table 6.  Interaction between and coracle and Neurexin IV in G//-mediated wing P C P  98  Interaction between discs-large and coracle in G//-mediated wing PCP  101  Interaction between and coracle and scribble in G//-mediated wing PCP ,  105  Interaction between Neurexin IV and scribble in G/i-mediated wing P C P  106  Interaction between discs-large and scribble in G//-mediated wing PCP  111  Interaction between and vulcan and discs-large in G/i-mediated wing P C P :  115  Interaction between and vulcan and coracle in G/;-mediated wing PCP  116  Interaction between and vulcan and Neurexin IV in G/j-mediated wing P C P  118  Interaction between and vulcan and scribble in G/j'-mediated wing PCP ,  119  Table 7.  Table 8.  Table 9.  Table 10.  Table 11.  Table 12.  Table 13.  Table 14.  viii  Table 15.  Table 16.  The dominant effect of second-chromosome suppressors on the viability of near-lethal Gli genotypes  155  The dominant effect of Su(Gli)! on the viability of semi-viable Gli genotypes  157  Table 17.  Su(Gli)l is a dominant suppressor that acts in trans to Gli  Table 18.  Possible candidates for Su(GU)]  162  Table 19.  Su(Gli)l is a dominant suppressor of septate-junction P C P and adhesion defects  163  ix  158  List of abbreviations  AP: anterior-posterior APF: after puparium formation Cot/cor: the Coracle protein and coracle locus, respectively CD: cytochalasin D, an actin-destabilizing drug Ck/ck: the Crinkled protein and the crinkled locus, respectively DEP: protein domain found in Dishevelled, Espinas and Testin Dgo/dgo: the Diego protein and die go locus, respectively D\g/dlg: the Discs-large protein and discs-large locus, respectively DN-Cdc42: a dominant-negative Cdc42 construct expressed using the GAL4:UAS system DN-Racl: a dominant-negative Racl construct expressed using the GAL4:UAS system DN-Wdb: a dominant-negative Widerborst construct expressed using the GAL4:UAS system Dxokldrok: the Drosophila rho-associated kinase protein and locus, respectively Dsh/dsh: the Dishevelled protein and dishevelled locus, respectively E C M : extracellular matrix EMS: ethyl methane sulfonate, a mutagenic, D N A alkylating agent Fm\/fmi: the Flamingo protein and flamingo locus, respectively; allelic to stan Fy/jfy: the Fuzzy protein and fuzzy locus, respectively  Fz/fz: the Frizzled protein and frizzled locus, respectively GFP: Green fluorescence protein Gli/G/i: the Gliotactin protein and Gliotactin locus, respectively H A J : hemiadherens junction if: inflated (see also PS2oc) In/in: the Inturned protein and inturned locus, respectively Lgl/lgl: the Lethal giant larvae protein and lethal giant larvae locus, respectively mys: lethal (I) myospheroid (see also PS(3) mew: multiple edematous wings (see also PS 1 a) mM: millimolar mwh: the multiple wing hairs locus Nrx/Nrx: the Neurexin IV protein and Neurexin IV locus, respectively PBS: phosphate-buffered saline PCP: planar cell polarity (also known as tissue polarity or merely planar polarity) PCR: polymerase chain reaction PD: proximal-distal PDZ: protein-protein interaction domain first characterized in PSD-95, Discs-large, and ZO-1. Pk/pk: the Prickle protein and prickle locus, respectively PP2A: protein phosphatase 2A P S l a : position-specific integrin al_ subunit, encoded in Drosophila by the mew locus  xi  PS2a: position-specific  integrin  a2 subunit,  encoded in Drosophila by the j/'locus  PSp: position-specific integrin ]3 subunit, encoded in Drosophila by the mys locus SJ: the  Drosophila pleated septate junction  Stan: see Fmi  Stbmlstbm: the Strabismus protein and strabismus locus, respectively; allelic to Vang Vang: see Stbm V B : vinblastine, a microtubule-destabilizing drug  Wdb/wdb: the Widerborst protein and the widerbprst locus, respectively Zip/zip: the Zipper protein and the zipper locus, respectively  xii  C H A P T E R I.  INTRODUCTION  Epithelial sheets in Drosophila are major contributors to morphogenesis and development. Within epithelia, cell-cell junctions maintain epithelial integrity, regulate proliferation, form diffusion barriers and contribute to apical-basal cell polarity (Tepass et al., 2001). Cellular junctions subdivide the lateral sides of epithelial cells into domains along the apical-basal axis (Figure I). The most apical lateral domain is the marginal zone, which lies apical to the zonula adherens. Basal to the zonula adherens is the pleated septate junction. Several molecular constituents of these lateral subdomains have been described (for a review see Tepass et al., 2001). While the roles of septate junction components have been investigated in embryonic tissues, less is known about this junction in subsequent stages of development. One protein that localizes to the pleated septate junction is Gliotactin. Embryonic phenotypes of Gliotactin mutants have been previously described (Auld et al., 1995), but postembryonic phenotypes have not. The major aim of this thesis is to test the hypothesis that Gliotactin is required for the development of polarized post-embryonic epithelia.  Septate junctions in  Drosophila  The Drosophila pleated septate junction (SJ) is a basolateral junction characterized by an interdigitated ladder of electron-dense septae between connected cells (Figure IB), (Fristrom, 1982). A specialized subdomain of SJs is the tricellular junction that occurs at a junction between three cells (Fristrom, 1982). A t the tripartate SJ specialized electron-translucent "diaphragms" aligned with septae are present, forming a "tricellular plug". This structure has been hypothesized to contribute to the transepithelial barrier of epithelial sheets (Fristrom, 1982). Septate junctions are present in the epidermis (Lamb et al., 1998) and epidermally-derived imaginal discs (Fristrom, 1982), as well as on glial membranes where they form the embryonic blood-nerve barrier (Auld et al., 1995; Baumgartner et al., 1996). Several proteins found at SJs have been characterized and mutagenized in Drosophila, including the tumor suppressors lethal (I) discs-large {dig) (Perrimon, 1988) and scribble (Bilder et al., 2000;  2  Figure 1. The  Drosophila pleated septate junction.  (A) A schematic diagram of cellular junctions in the embryonic epidermis. The pleated septate junction (SJ) is below the zonula adherens (ZA) on the basolateral face of each cell. MZ = marginal zone, GJ = gap junction, aHAJ = apical hemiadherens junction, bHAJ = basal hemiadherens junction. The thick line on the apical cell surface represents the embryonic cuticle. (Adapted from Tepass et al., 2001).  (B) An transmission electron micrograph of a thin section through a pleated septate junction. The characteristic, ladder-like septae are evident. (Adapted from Alberts et al., 1994).  3  4  Bilder and Perrimon, 2000), as well as Neurexin IV (Baumgartner et al., 1996) and coracle, an ortholog of protein 4.1 (Fehon et al., 1994). Mutations in dig and scrib result in loss of apical-basal polarity leading to epithelial overgrowth (Woods and Bryant, 1991; Bilder and Perrimon, 2000). Neurexin IV and coracle mutants do not lose apical-basal polarity but have defects in epithelial integrity (Baumgartner et al., 1996; Fehon et a l , 1994). The roles of these SJ components are discussed below.  The roles of Scribble and Discs-large  The first component of the SJ identified in Drosophila was Discs-large (Dig), which was also a founding member of the P D Z (PSD-95, Discs-large and ZO-1) protein family (Woods and Bryant, 1991). D i g is a cytoplasmic protein associated with SJs (Woods et al., 1996) by an unidentified proteinprotein interaction (Tepass et al., 2001), possibly through one or more of its three P D Z motifs, which are known to bind specific carboxyl-terminal epitopes of localized proteins (Songyang et al., 1997). Indeed, the second P D Z domain of D i g is necessary for its localization to the SJ (Hough et al., 1997). The most likely candidate to localize D i g to the SJ is a transmembrane protein with a PDZ-binding epitope (Tepass et al., 2001). D i g is maternally contributed; embryos lacking maternal and zygotic D i g function have defects in dorsal closure, the process by which the embryonic epithelium surrounds the yolk sac and merges dorsally (Perrimon, 1988). Zygotic loss of function of dig causes an overgrowth of imaginal discs as well as loss of apical-basal polarity in mutant cells (Woods and Bryant, 1991; Woods et al., 1996). In J/g-mutant imaginal discs, adherens junctions form at ectopic locations, tubulin and the apical circumferential actin ring are disrupted, and Coracle is membrane-associated but not restricted to the septate junction (Woods et al., 1996).  Mutations in scribble (scrib) also cause defects in the organization of epithelial sheets by altering apical-basolateral polarity: in the absence of functional Scrib at septate junctions, apical determinants  5  such as Crumbs (Wodarz et al., 1995) are mislocalized and spread into basolateral regions (Bilder and Perrimon, 2000). As was found for dig, circumferential actin and adherens junctions are not polarized in scrib mutants (Bilder and Perrimon, 2000). Failure to localize apical determinants such as Discs-lost and Crumbs are largely responsible for the scrib mutant phenotype; thus a SJ protein is necessary for delimiting the basal edge of the apical boundary and thereby maintaining the apical-basolateral polarity of the cell (Bilder and Perrimon, 2000). Scrib, like Dig, encodes a cytoplasmic protein with multiple PDZ domains (Bilder and Perrimon, 2000). Scrib was found to prefigure the future embryonic location of the SJ and is the earliest known marker that specifically localizes to this region (Bilder et al., 2000). Once SJs have formed, Scrib co-localizes with Coracle (see below) (Bilder and Perrimon, 2000). The SJ-localization of Scrib and Dig was found to be interdependent, as well as dependent on a third tumor suppressor gene, lethal (2) giant larvae (lgl) (Bilder et al., 2000). The subcellular distribution of Lgl protein overlaps with, but is not restricted to, septate junctions or even the membrane (Bilder et al., 2000); Lgl is also physically associated with non-muscle myosin (zipper in Drosophila, see below), which is a cytoskeletal component (Strand et al., 1994). Recent work has demonstrated that Lgl functions together with Dig and Scrib to regulate the function of Crumbs in determining the zonula adherens (Bilder et al., 2003; Tanentzapf and Tepass, 2003).  The roles of Coracle and Neurexin IV  A Drosophila protein similar to mammalian erythrocyte protein 4.1 is encoded by the coracle (cor) locus and is localized to the septate junction (Fehon et al., 1994). The Cor protein is cytoplasmic yet membrane-delimited, lacks the actin binding domain present in protein 4.1, and contains an aminoterminal region with shared homology to protein 4.1, Ezrin, Radixin and Moesin (the F E R M domain) (Ward et al., 1998; Ward et al., 2001). Cor localizes basal to the circumferential, filamentous actin ring at the apical membrane of epithelial cells (Eaton et al., 1996; Fehon et al., 1994). The formation of SJs is  abolished in cor mutants, adherens junctions, however, persist at the ultrastructural level and adherensjunction markers are localized correctly (Lamb et al., 1998), in direct contrast to scrib and dig mutants. Embryos lacking Cor also fail to complete dorsal closure (Fehon et al., 1994). The function of Cor at SJs has been shown to be interdependent with that of Neurexin I V (see below). Post-embryonic phenotypes of cor mutants include rough eyes, wing vein defects, missing ocelli and ocellar bristles, and unspecified leg abnormalities, indicating that C o r functions in post-embryonic stages (Lamb et al., 1998).  Neurexin I V (Nrx) is a single-pass transmembrane protein localized to SJs in Drosophila (Baumgartner et al., 1996) that shows homology to the mammalian Contactin-associated paranodal receptor (Caspr) localized to the glial-neuronal paranodal junction of myelinated axons (Einheber et al., 1997; Peles et al., 1997a; Peles et al., 1997b). Nrx mutants also fail to complete dorsal closure; additionally, Nrx embryos are paralyzed due to the absence of glial-glial septate junctions (Baumgartner et al., 1996) which form the Drosophila blood-nerve barrier (Auld et al., 1995). In Nrx mutants, Cor fails to localize to the SJ but spreads basolaterally in the membrane and is present in cytoplasmic puncta (Baumgartner et al., 1996; Ward et al., 1998). Conversely, Nrx fails to localize to the SJ in cor null mutants (Ward et al., 1998). The F E R M domain of Cor was necessary and sufficient to localize Nrx to the SJ in vivo, as well as necessary and sufficient to bind the intracellular domain of Nrx directly in biochemical assays, demonstrating a direct physical link between Cor and Nrx through these domains (Ward et al., 1998). Subsequent work demonstrated that the F E R M domain was sufficient to rescue all cor embryonic phenotypes including the restoration of SJs assayed by transmission electron microscopy (Ward et al., 2001). Consistent with the paralysis phenotype of Nrx mutants, cor mutants were also found to be paralyzed (Ward et al., 1998). Post-embryonic phenotypes of Nrx mutants somewhat parallel those of cor mutants: rough eyes, wing notching and vein defects, and leg malformations are evident, suggesting a post-embryonic function for Nrx (Baumgartner et al., 1996).  7  The role of Gliotactin  Gliotactin (Gli) was originally identified as a novel transmembrane protein expressed on peripheral glial membranes (Auld et al., 1995). Similar to Nrx mutant embyros, Gli null homozygotes fail to form the glial peripheral blood-nerve-barrier and are paralyzed at late embryogenesis (Auld et al., 1995). This similarity in phenotype between Gli and Nrx null mutants suggests that they function in the same pathway and that Gliotactin is a necessary component of the septate junction in embryonic peripheral glia. In contrast to Nrx mutants, Gli mutants complete dorsal closure (Auld et al., 1995). Gliotactin is a type 1, single-pass transmembrane protein whose major extracellular domain is a noncatalytic serine esterase-like motif (Figure 2) (Auld et al., 1995; Botti et al., 1998). The G l i serine esterase-like (SE) motif shares significant similarity with a functional serine esterase, the acetylcholinesterase of the electric ray Torpedo califorica ( T c A C h E ) (Botti et al., 1998; Sussman et al., 1991). The G l i S E motif is 30 percent identical and 61 percent similar to T c A C h E (Botti et al., 1998). The active site "gorge" of T c A C h E is surrounded by a ring of negative electostatic surface potentials, a feature that is conserved in G l i (Botti et al., 1998) (Figure 2). The intracellular domain of G l i is novel, although the extreme carboxyl terminus encodes a consensus PDZ-recognition peptide. Recently, G l i has been shown to localize to the septate junction of the embryonic epidermis, specifically to the tricellular septate junction (Schulte, 2003), which abuts the tricellular plug (Fristrom, 1982). G l i thus defines a subdomain of the SJ and may participate in the formation or maintenance of the tricellular plug (Schulte, 2003). Postembryonic phenotypes of Gli mutants have not been investigated as only embryonic-lethal (and presumed null) alleles have been identified (Ashburner et al., 1990; A u l d et al., 1995). Thus Gli mutant tissues have not been examined past the first absolute requirement for G l i function, the formation of the blood-nerve-barrier at late embryogenesis (Auld et al., 1995).  Figure 2. The domain structure of Gliotactin.  (A) A schematic representation of the Gliotactin protein product. The amino-terminus encodes a signal sequence (amino acids 1-28). A large portion of the extracellular domain (amino acids 175-723) comprises the serine esterase-like motif, which extends to the transmembrane domain (amino acids 723-746). The intracellular portion of Gli contains a PDZ-binding epitope at its extreme carboxyl terminus (amino acids 956-959).  (B-C) Three-dimensional diagrams showing electrostatic surface potentials in Torpedo acetylcholinesterase (TcAChE) (B) and Gliotactin (C). Red indicates a negative surface potential, white is neutral, and blue indicates a positive potential. The active site "gorge" in (TcAChE) is rimmed with negative electrostactic potentials. This annular negative charge is partially conserved in Gliotactin (C). (B and C were adapted from Botti et al., 1998).  9  10  The aim of this thesis is to test the hypothesis that Gli is required for the development of polarized post-embryonic epithelia. As a starting point for investigating post-embryonic roles for Gliotactin, a genetic screen was performed in Chapter II to identify adult-viable, hypomorphic Gli alleles. Several adult-viable Gli mutant genotypes were identified. Examination of adult epithelia in Gli mutants revealed defects in wing epithelial development. These defects were sensitive to enhancement by additional septate junction mutations. Wing development is well understood in Drosophila, however, the role of Gliotactin and other SJ components has not been addressed. Given that hypomorphic Gli genotypes were available, the role of Gli and other SJ proteins in wing development was investigated in Chapter III.  Wing development in  Drosophila  The adult wing of Drosophila is a cuticular structure derived from an imaginal disc monolayer that is ultimately derived from the embryonic epidermis (Tepass et al., 2001). Imaginally derived structures maintain epithelial apical-basal polarity throughout development and have fully elaborated cell-cell junctions including septate junctions (Fristrom, 1982; Tepass et al., 2001). Development following the third larval instar is divided into a 12-hour prepupal and an 84-hour pupal period (Fristrom and Fristrom, 1993). Wing development begins long before this period in the determination and morphogenesis of imaginal discs and ends post-eclosion when wing cells degenerate to leave only a cuticular bilayer with few living cells. For the purposes of this introduction, developmental events initiated at and following pupariation will be discussed. Current knowledge of these processes is greatly indebted to Dianne and James Fristrom, whose elegant descriptive work using confocal and transmission electron microscopy (Fristrom and Fristrom, 1993; Fristrom et al., 1993) consolidated and extended previous results (Brower and Jaffe, 1989; Morgensen and Tucker, 1988; Morgensen et al., 1989; Tucker et al., 1986; Wilcox et al., 1989; Zusman et al., 1990) and also preceded several studies of mutants with  11  abnormal wing development (Brabant et al., 1996; Fristrom et al., 1994; Prout et al., 1997), focusing on mutations that affect integrin-mediated epithelial adhesion (see below).  Prepupal wing development (0 - 12 hours after puparium formation)  The prepupal period extends from pupariation at the end of the third larval instar, marked by the eversion of larval spiracles, to 12 hours after puparium formation (APF), when the head is everted. During this period several dramatic morphogenic events take place. The wing epithelium (1) folds to become a bilayer, (2) increases in surface area through changes in cell shape, (3) everts, (4) adheres at apposed basal surfaces, and finally (5) is forced apart by the eversion of the head which marks the end of the prepupal period (Fristrom and Fristrom, 1993; Fristrom et al., 1993). At the start of the prepupal period (0 hours APF) the wing imaginal disk is a monolayer with a secreted basal lamina and several proximal folds (Figure 3), (Fristrom and Fristrom, 1993). During the early prepupal period the wing disc folds at the presumptive wing margin to give the primary basal-basal apposition of the dorsal and ventral wing epithelial sheets at approximately 4 hours APF (Figure 3), (Fristrom et al., 1993). The apposition of dorsal and ventral epithelia excludes the basal lamina from between the basal cell faces at the distal tip of the wing; a secreted extracellular matrix (ECM) replaces it (Fristrom and Fristrom, 1993). During and following this period (extending to 7 hours APF), cell elongation of the wing bilayer occurs without cell division: previously narrow columnar cells flatten out to extend the epithelial sheets (Figure 4), (Fristrom et al., 1993). This extension is a motive force in everting the wing (complete at 6 hours APF) from inside the larval epidermis into the space between epidermis and the larval cuticle (which persists as the prepupal/pupal case). Despite the close apposition of the dorsal and ventral epithelial sheets since four hours APF, basal junctions connecting the two bilayers are not evident until about nine hours APF, when basal hemiadherens junctions (HAJs) appear. At 11-12 hours APF transcellular microtubule arrays insert into basal HAJs to form the first wing-spanning, or "transalar" cytoskeleton (Figure 5) (Tucker et  12  Figure 3. Early events in prepupal wing morphogenesis.  V = presumptive ventral wing surface, D = presumptive dorsal wing surface, M = presumptive wing margin. The presumptive wing surfaces are shown in red, the basal lamina is shown in blue. Proximal is to the right and distal is to the left. All diagrams are longitudinal sections. (Adapted from Fristrom and Fristrom, 1993).  (A) At the start of the prepupal period (0 hours APF) the wing imaginal disc is a monolayer with several proximal folds. These folds flatten out as the presumptive ventral surface begins to fold under the presumptive dorsal surface. (B) At 2 hours APF the ventral surface has folded under the dorsal surface. This process excludes the basal lamina from between the presumptive wing surfaces. (C) At 4 hours APF the basal lamina is completely excluded and the presumptive wing surfaces are apposed for the first time.  13  al., 1986). With the exception of cells adjacent to the wing margins, transalar connections are broken during head eversion. The abdominal muscle contractions necessary for head eversion force hemolymph between the apposed wing epithelia, separating them (Fristrom and Fristrom, 1993; Fristrom et al., 1993).  ' •  Pupal wing development (12-96 hours after puparium formation)  The events of prepupal development are mirrored during pupal development as the developing wing undergoes another round of apposition, adhesion, expansion and separation during this period (Fristrom and Fristrom, 1993; Fristrom et al., 1993), (Figure 4). After the extreme separation induced by the force of everting the head, wing epithelia undergo final mitotic divisions from 15-24 hours APF. The pupal cuticle is secreted and separated from the wing epithelium at this stage. From about 20 hours APF, basal cell extensions bridge the ECM present between the wing layers and reestablish apposition of the two epithelia, although substantial extracellular spaces persist at this time (Fristrom et al., 1993). It is in this period of progressing apposition, from 24-36 hours APF, that wing epithelial cells are polarized in the plane of the epithelium (along the proximal-distal axis) and extend a single actin-rich prehair at a location specified by planar polarity determinants (see below). Extracellular spaces continue to be eliminated until an extensive ECM remains only basal to presumptive vein cells at 40 hours APF. Beginning at about 35 hours APF basal HAJs reappear, indicating the pupal adhesion stage. In contrast to prepupal adhesion, pupal adhesion is followed by expansion rather than preceded by it (Fristrom et al., 1993); this expansion doubles the surface area of the wing, which must fold inside the pupal cuticle during 45-60 hours APF. At the 60-hour mark pupal separation occurs, although this separation is minimal compared to prepupal separation. Microtubules spanning each monolayer from apical to basal HAJs arise from 60 to 84 hours APF and comprise the final transalar array (Fristrom et al., 1993; Tucker et al., 1986). Transalar microtubules nucleate at apical HAJs and extend basally to  15  Figure 4. Later events in prepupal and pupal wing morphogenesis.  All figures represent transverse sections through the developing wing. (Adapted from Fristrom et al., 1993).  (A) Prepupal apposition. (4 hours APF). Wing surfaces are apposed for the first time. (B) Prepupal expansion (6 hours APF). Flattening of columnar wing epithelial cells increases the surface area of the wing. (C) Prepupal adhesion (9 hours APF). The apposed wing epithelia elaborate basal cell-cell hemiadherens junctions. (D) Prepupal separation 1(11 hours APF). F-actin/microtubule bundles that span each monolayer from apex to base are inserted into basal hemiadherens junctions to give the prepupal transalar cytoskeleton. (E) .Prepupal separation II (13 hours APF). Head eversion has broken all basal cell-cell junctions with the exception of those at the wing margin. (F) Pupal appostion (26 hours APF). Basal cell extensions are evident and progressively re-appose the separated epithelia. It is during this period that the wing epithelial cells are polarized in the plane of the epithelium. (G) Pupal adhesion (40 hours APF). Ongoing appostion has brought the two epithelia together. Basal hemiadherens junctions reappear. (H) . Pupal expansion (50 hours APF). Further flattening of wing cells produces a doubling of wing size. (I) . Pupal separation (60 hours APF). The transalar cytoskeleton reappears and persists until after eclosion at 96 hours APF.  16  Prepupal development  Pupal development  basal HAJs (Figure 5B) (Morgensen et al., 1989). This array has been shown to include numerous actin filaments interspersed among the microtubule filaments (Morgensen and Tucker, 1988). The remainder of wing development consists of deposition of adult cuticle. After eclosion and inflation of the folded wings, all cells but those adjacent to veins degenerate to bring the dorsal and ventral cuticle sheets together to form the functional adult wing (Fristrom and Fristrom, 1993). The transalar array does not appear to be actively involved in this process (Morgensen and Tucker, 1988).  The role of integrins in wing development  Integrins are large transmembrane heterodimers composed of a and (3 subunits (Humphries, 2000) which are basally restricted during wing development (Fristrom et al., 1993) and required for adhesion between the dorsal and ventral wing surfaces (Brower and Jaffe, 1989; Wilcox et al., 1989; Zusman et al., 1990). Loss of integrin-mediated adhesion in the pupal wing produces adult wings with separated dorsal and ventral cuticle, or a wing "blister". In Drosophila, integrins were first identified as position-specific (PS) antigens present on the wing imaginal disc (Brower et al., 1984; Wilcox et al., 1981). Subsequent work identified the loci that encode the three PS integrin subunits: PS la is encoded by the multiple edematous wings (mew) locus (Brower et al, 1995); PS2 a is encoded by the inflated (if) locus; and PSp\ which is the common p subunit for PS1 and PS2, is encoded by the lethal (I) myospheroid (mys) locus (Brower and Jaffe, 1989; Wilcox et al., 1989). In late third instar larvae, PS integrins are expressed in distinct domains of the wing imaginal disc. PS la is expressed exclusively in the presumptive dorsal region of the disc. Consistent with this expression pattern, loss-of-function mew clones cause wing blistering only when clones are in the dorsal epithelium (Brower et al., 1995). Conversely, PS2a is expressed solely in the presumptive ventral epithelium and loss-of-function if clones induce blisters only if the clone is ventral (Brabant and Brower, 1993). PSP is expressed both  18  F i g u r e 5. The transalar cytoskeleton.  (A) The prepupal transalar cytoskeleton (11 hours APF). Bundles of F-actin and microtubules (parallel lines) nucleate at apical hemiadherens junctions (rectangles) and connect to interdigitated basal junctions through narrow cellular processes. The dark line at the apical surface of each cell represents cuticle.  (B) The pupal transalar cytoskeleton (60 hours APF). Apical wing hairs are present on the epithelial cells. Schematic representation is the same as in (A). (Adapted from Tucker et al., 1986).  19  20  ventrally and dorsally, consistent with its role as the p subunit for PS la and PS2a; similarly, mys clones cause blisters whether dorsal or ventral (Brower and Jaffe, 1989).  The subcellular location of PSP has been investigated during wing morphogenesis in wild-type wings (Fristrom et al., 1993) and wings containing mys clones (Brabant et al., 1996). In wild-type wings, PSp is diffusely expressed on the basal surface of dorsal and ventral epithelia prior to 8 hours APF, when it becomes localized to distinct basal foci, apparently prefiguring nascent HAJs (Fristrom et al., 1993). Actin does not co-localize with PSP immunoreactivity until 1 1 hours APF, at the time the first transalar cytoskeleton is assembled (Fristrom et al., 1993). Following head eversion and its concomitant dorsal-ventral separation, PSP reappears on the basal surfaces and remains there through 35 hours APF, the time basal HAJs reappear (Fristrom et al., 1993). PSP becomes concentrated at basal foci at 50 hours APF, preceding the arrival of apically originated microtubule/actin filaments at basal HAJs to reestablish the transalar array at 60 hours APF. In wings bearing clones lacking PSP (mys clones), prepupal apposition occurs normally at six hours APF, indicating that this apposition is integrinindependent (Brabant et al., 1996). In stark contrast to prepupal apposition, basal cytoplasmic extensions that initiate pupal apposition are absent in cells lacking PSp, and also in cells opposite a clone of cells lacking PSp. Additionally, basal HAJs fail to differentiate in wild-type cells apposed to a mys clone (Brabant et al., 1996). Numerous mutations that interact with integrin have also been identified (Prout et al., 1997; Walsh and Brown, 1998); among them are an intracellular integrin-actin cytoskeletal adaptor protein, Talin (Brown et al., 2002), and an ECM a-laminin that functions as an extracellular integrin ligand (Martin et al., 1999). Mutations in both genes produce wing blisters: the rhea locus (Prout et al., 1997) encodes Talin (Brown et al., 2002), whereas the wing blister locus encodes a-laminin in Drosophila (Martin et al., 1999).  21  Planar cell polarity in the Drosophila  wing epithelium  Presumptive wing epithelia maintain apical-basal polarity throughout their morphogenic movements, however, during the pupal apposition stage, wing epithelial cells also become polarized in the plane of the epithelium with respect to their proximal and distal sides and acquire what is variously referred to as tissue polarity, planar polarity, or planar cell polarity (PCP). Cells in several Drosophila epithelia undergo planar polarization. Cells comprising an ommatidia of the eye or a sensory bristle apparatus co-ordinate groups of cells to form polarized structures through asymmetric division, rotation and changes in cell fate (Mlodzik, 1999; Mlodzik, 2000). The wing epithelium, in contrast, is polarized within single cells in the absence of cell division. Many molecular determinants of PCP are utilized in both wing and eye/bristle PCP, but as these are fundamentally different processes, the role of a given determinant may not be directly comparable between them (Mlodzik, 1999). Although the specification of PCP in the wing occurs only during a brief portion of pupal wing development, numerous studies of altered wing PCP have been performed. As a result, the genetic and molecular basis of PCP is well understood in the pupal wing of Drosophila.  The vast majority of cells in the pupal wing produce a single microvillus-like, actin-rich prehair on their apical surface at about 33 hours APF (Wong and Adler, 1993). Prehairs are initiated at the distal vertex of each cell, oriented in a proximal to distal direction, and aligned with neighboring prehairs to form an ordered and parallel array of hairs on the adult wing (Wong and Adler, 1993). The location of prehair initiation is determined by the asymmetric localization of several polarity determinants in the plane of the epithelium that culminates around 30 hours APF. The accessibility of a flat epithelium such as the wing, combined with ease of forward genetic analysis in Drosophila, has lead to the identification of a large number of mutations that disrupt planar cell polarity in the wing. Polarity mutations have been isolated which alter hair orientation, the number of hairs per cell, or both (Adler, 2002). Mutations  22  which disrupt the parallel alignment of wing hairs have not been described (Wong and Adler, 1993; Adler, 2002).  Several genes necessary for correct PCP in the wing are thought to comprise a "core" set of genes that determine PCP in the eye and sensory bristles as well as in the wing (Shulman et al., 1998), (for a recent review see Tree et al., 2002). The "core" currently comprises frizzled (fz), dishevelled (dsh), prickle-spiny legs (pk-sple), flamingo (fmi), which is also called starry night (stan), possibly diego (dgo),  although its role in sensory bristle polarity is unclear (see below), and strabismus (strb), which is also known as Van Gogh (Vang). Other genes that are not necessary for PCP in all three tissues are thought to regulate polarity and hair number downstream of this core set in a tissue-specific manner (Tree et al., 2002).  The roles of Frizzled, Dishevelled and Prickle  The mutations/h'zz/ed (fz), dishevelled (dsh), and prickle (pk) (Gubb and Garcia-Bellido, 1982; Held et al., 1986; Vinson and Adler, 1987) were originally identified as mutants that produce wing hairs which follow curved vectors instead of the wild-type, linear proximal-distal vector (Figure 6A,C). Subsequent work demonstrated that prehairs produced in these mutants originate at the center of the apical cell surface instead of the distal vertex (Figure 6B,D); thus these mutants form a phenotypic group with respect to prehair initiation (Wong and Adler, 1993). Mutations \nfz, dsh and pk display similar epistatic interactions with other tissue polarity genes, leading to the suggestion that they form an epistasis group (Wong and Adler, 1993).  Thefz locus encodes a putative seven-pass, transmembrane receptor (Vinson et al., 1989), a ligand for which is unknown. Analysis of mutant/?, cells produced by mitotic recombination in a wild-  23  Figure 6. Planar cell polarity and prehair location in wild-type and fz mutant wings.  Distal is to the right and proximal is to the left for each image. Arrows indicate the polarity of hairs on the adult wing.  (A) In wild-type wings, hairs follow a proximal-distal vector that parallels the wing veins. (B) The subcellular location of the actin-rich prehair (black triangle) is restricted to the distal-most cell vertex in wild-type wings. (C) In the fz mutant, wing hairs exhibit aberrant polarity. Adjacent hairs remain essentially parallel, but follow curved vectors that can have completely reversed polarity. (D) . The subcellular location of the prehair in fz mutants is not restricted to the cell periphery. Prehairs initiate in the cell center and have altered polarity. (Adapted from Gubb and Garcia-Bellido, 1982).  24  25  type wing demonstrated that mutant/z cells can alter the polarity of wild-type cells distal to the fz mutant clone (Gubb and Garcia-Bellido, 1982; Vinson and Adler, 1987). Further studies demonstrated that certain alleles of fz do not display such non-autonomy and affect only mutant cells within the clone;/z thus has cell autonomous and non-cell autonomous functions (Vinson and Adler, 1987). The molecular lesions of non-autonomous alleles map to all regions of the Fz protein, whereas cell-autonomous alleles map to substitutions affecting a single proline residue in the presumptive first cytoplasmic loop (Jones et al., 1996). The subcellular distribution of Fz in pupal wings is restricted to the apical and apical-lateral cell surface, although the precision of this result is questionable due to poor staining obtained with antiFz antibodies (Park et al., 1994a). The non-autonomy of fz clones is directional: wild-type cells distal, and to a lesser extent, anterior and posterior, to fz cells reorient their hairs to point towards the fz clone (Adler et al., 2000; Vinson and Adler, 1987). In pk wings (see below) which have hairs with reversed (distal to proximal) polarity, wild-type cells proximal to a/z clone direct their hairs towards the clone. Thus fz non-autonomy follows the local polarity pattern, suggesting that/z non-autonomy is a locally propagated signal (Adler et al., 2000).  Unlike fz, dsh loss-of-function mutations are cell-autonomous and disrupt wingless (wg) signaling during embryogenesis (Klingensmith et al., 1994; Perrimon and Mahowald, 1987; Theisen et al., 1994). The original dsh allele does not affect wg signaling but has a wing planar polarity phenotype, 1  thus dsh functions both in the wg pathway and in wing PCP (Axelrod et al., 1998; Boutros et al., 1998). The Dishevelled protein is cytosolic and moldular: it contains an N-terminal DIX (Dishevelled. Axin) domain (Zeng et al., 1997); a central PDZ domain (Klingensmith et al., 1994; Theisen et al., 1994); a DEP (Dishevelled, Egl-10, Pleckstrin) domain (Ponting and Bork, 1996); and a proline-rich domain containing an SH3 motif (Penton et al., 2002). Expression of Drosophila Dsh and Fz in a heterologous Xenopus system demonstrated that Fz recruits Dsh from the cytosol to the membrane in a DEP domaindependent manner (Axelrod et al., 1998). Constructs harboring deletions of the DEP domain failed to  26  localize to the cell cortex. The DEP domain was also sufficient for cortical localization (Axelrod et al., 1998) . Sequence analysis of the dsh' allele in this study revealed that it is a point mutant in the DEPdomain, suggesting that this domain is dispensable for wg signaling (Axelrod et al., 1998). Subsequent mutagenesis of Dsh, however, recovered several dsh alleles that map to the DEP domain and alter wg signaling (Penton et al, 2002). Thus while dsh' remains an important allele for discriminating between wg and PCP signaling, the requirements for Dsh function in either pathway do not correlate strictly to identified protein domains.  The prickle locus is a complex locus spanning -70 kb that encodes three alternatively spliced transcripts: pk *, pk!*, and pk ' ' (Gubb et al., 1999). Mutations affecting only the pk" ' '' transcript were 1  11  1  1  1  classically identified as spiny legs and thought to be a distinct locus (Gubb and Garcia-Bellido, 1982). The three transcripts differ only in their 5' exon and each encodes a cytosolic protein (Gubb et al., 1999) . The 3' exons common to all three transcripts contain three LIM domains and a PET domain. LIM domains are zinc-finger, protein-protein interaction domains found in a wide variety of nuclear and cytoplasmic proteins (Khurana et al., 2002). A consensus LIM-binding motif has not been identified, however, several LIM proteins associate with the cytoskeleton (Khurana et al., 2002). The PET domain is a novel domain of unknown function found in Prickle as well as in Espinas and Testin, which also have LIM domains. Both the pkl' and pk"' '' transcripts (encoding the Pk and Sple proteins, respectively) k  1  are expressed in the wing and leg. Mutations affecting pk!' produce a strong planar polarity phenotype in k  the wing, whereas animals mutant for pk*'' '' have a strong leg phenotype. Mutations which affect all three 1  transcripts (pk'' ' '''' alleles) are homozygous viable and have a moderate planar polarity phenotype in k  S1  wings and legs. Over-expression of pk' ' '' in the wing phenocopies pk!' alleles; similarly, heterozygousity 1  1  k  for pk" ' '' suppresses the pk!' phenotype (i.e. in pk' lpkl' ' ' genotypes). The mutant phenotype of pk 1  1  k  k  k sl k  homozygotes is thus interpreted to result from an imbalance between functional Pk and Sple proteins (Gubb et al., 1999). The pk? transcript is expressed only during embryonic stages and has no known 4  27  pk  function, as homozygous pk!' ' '' ' mutants have no discernable embryonic phenotype (Gubb et al., 1999). k  s  h  Unlike fz but similar to dsh, clones of pk!' mutant cells are largely cell-autonomous (Adler et al., 2000; k  Gubb etal., 1999).  Recent improvements in reagents for Fz, Dsh and Pk have determined the functional relationship between these genes and revealed that asymmetric distribution of PCP determinants prefigures prehair location (reviewed in Strutt, 2002). The construction of a functional Fz-GFP transgene revealed asymmetric localization of Fz in the wing epithelium during PCP determination (Strutt, 2001a). Fz-GFP was localized apically to the distal side of pupal wing cells from 19-39 hours APF. Interestingly, distal localization of Fz-GFP was abolished in clones homozygous for a dsh amorphic allele. Thus Dsh is necessary for Fz localization and the previous model of Dsh acting solely downstream of Fz is incomplete (Strutt, 2001a). The connection between Fz subcellular localization and non-autonomous signaling was also investigated. Interestingly, non-autonomous Fz signaling was independent of both apical and proximal localization (Strutt, 2001a).  A functional Dsh-GFP transgene (Axelrod, 2001) and anti-Dsh antibodies (Shimada et al., 2001) have recently been developed. Both revealed adynamic distribution of Dsh during pupal wing development spanning the time of PCP determination. Dsh is distributed in the cytoplasm of third-instar larval wing disc cells, but is recruited to the apical circumference at 2 hours APF (Axelrod, 2001). Dsh remains circumferentially unpolarized until 18-24 hours APF, when redistribution of Dsh to proximaldistal cell boundaries becomes apparent (Axelrod, 2001; Shimada et al., 2001). Cells expressing DshGFP positioned distally to cells lacking the transgene did not accumulate the GFP signal at their proximal boundary (Axelrod, 2001); similarly, clones lacking endogenous Dsh did not have Dsh immunoreactivity on their distal edge (Shimada et al., 2001), indicating that Dsh is localized solely to the distal side of cells. Both groups demonstrated that the either the dsh' mutation or loss of functional  28  Fz prevents membrane association and redistribution of Dsh (Axelrod, 2001; Shimada et al., 2001), which was in agreement with previous work in a heterologous system (Axelrod et al., 1998). Thus the genetic relationship between dsh and/z is not a linear one, but involves a feedback mechanism (Axelrod, 2001; Strutt, 2001a; Tree et al., 2002).  Recently, an antibody produced against the common carboxyl-terminal domain of Pk and Sple (here referred to jointly as Pk) was used to determine the distribution of Pk protein isoforms in the developing wing (Tree et al., 2002). Pk is symmetrically distributed on the apical circumference in discrete puncta at 2.5 hours APF. At 30 hours APF, Pk was restricted to proximal-distal cell boundaries in a pattern that bordered Dsh distribution. Clonal analysis revealed that Pk is present only on proximal cell boundaries, where it non-autonomously promoted the accumulation of Dsh on the distal margin of the opposed cell. In dsh and/z cells Pk was membrane-associated but failed to polarize in cells 30 hours 1  APF. Conversely, Dsh failed to polarize in a 30 hour-APF clone lacking Pk (pk!' '"' '), even though early k  h  Fz-dependent recruitment of Dsh to the cell membrane was unaffected (Tree et al., 2002). The asymmetric distribution of Fz is also blocked in pk ' ""' ' cells (Strutt, 2001a). Thus Pk function is 1  1  1  dependent on and necessary for the asymmetric distribution of Fz and Dsh, indicating that it participates in the Fz/Dsh feedback pathway (Tree et al., 2002). Biochemical techniques showed that the DEP domain of Dsh and the PET/LIM domains of Pk physically interact; this interaction blocked the Fzinduced translocation of cytoplasmic Dsh to the membrane in a heterologous system (Tree et al., 2002). Taken together with the proximal localization of Pk in vivo, this activity may block membrane association of proximal Dsh, and by extension, Fz (Tree et al., 2002). A summary diagram of the actions of Fz, Dsh , Pk and Flamingo (see below) in determining PCP is given in Figure 7.  29  Figure 7. Fz, Dsh and Pk localization during Fz PCP signaling.  Fz/Dsh are shown in green, Pk in blue, Fmi in yellow and actin in red. (Adapted from Adler, 2002).  (A) At 30 hours APF, Fz signaling drives the accumulation of Fz and Dsh to the distal boundary of wing cells (green). This localization drives the accumulation of Pk on the adjacent proximal cell boundary (blue). Proximal Pk inhibits the proximal localization of Dsh, which in turn inhibits the proximal localization of Fz. The result is a postive feedback loop that drives asymmetric segregation of Fz, Dsh and Pk. Fmi (yellow) is present on proximal and distal boundaries. The asymmetric distribution of these PCP determinants determines the location and polarity of the actin-filled prehair that emerges at about 33 hours APF (B).  30  B  Proximal  Distal  Proximal  Distal  Fz/Dsh  31  The roles of Flamingo/Starry night, Diego, and Strabismus/Van Gogh  A large cadherin that functions as a core PCP determinant was identified almost simultaneously by two groups and named flamingo (fmi) and starry night (stan), respectively. Fmi/Stan (here referred to collectively as Fmi) encodes a seven-pass transmembrane protein with a large extracellular domain that contains cadherin, laminin and EGF repeats (Chae et al., 1999; Usui et al., 1999). Fmi is expressed in the embryonic nervous system, and fmi null mutants are embryonic lethal (Usui et al., 1999). Transgenic rescue of fmi mutants using embryo-specific promoters resulted in viable/m; adults lacking functional Fmi. Such animals exhibited PCP defects in sensory bristles, ommatidia, and wing hairs (Usui et al., 1999). Similarly, a wing-hair phenotype was obtained with a homozygous-viable//w allele (Chae et al., 1999). In rescued-null tissue,/m/-null clones and viable mutants, wing prehairs initiate at the cell center in a manner indistinguishable from/z or dsh mutants (Chae et al., 1999; Usui et al., 1999). Wing clones lacking Fmi were cell autonomous for polarity defects (Chae et al., 1999; Usui et al., 1999). Antibody localization of Fmi in the developing wing revealed dynamic redistribution during development prior to prehair initiation. Fmi was present at the cell cortex at 18 hours APF and polarized to the PD axis by 24 hours, with maximal polarity at 30 hours APF. Fmi asymmetry was lost by the time of hair elongation, at 36 hours APF (Usui et al., 1999). The PD polarization of Fmi was abolished in fz and dsh' mutant wings. In /mi-null clones, Fmi immunoreactivity was absent from all cells within and bordering the clone, suggesting that Fmi localizes to both the proximal and distal sides of wing cells, and that functional Fmi is required on both cells for localization in either. Consistent with this notion, Fmi promoted homophilic adhesion when expressed in tissue culture, and expression of Fmi in both cells was required for adhesion (Usui et al., 1999). Fmi was also found to accumulate preferentially at boundaries of cells expressing different levels of Fz. Cells lacking/z activity redistributed Fmi to the borders of the clone without preference for a PD boundary. A similar result was obtained for cells overexpressing Fmi: ectopic expression of Fz along the AP axis caused Fmi in neighboring wild-type cells to accumulate at  32  AP boundaries. Interestingly, the mislocalization of Fmi in wild-type cells distal to afz clone propagated over several cell diameters, suggesting a mechanism for the distal non-autonomy of fz clones (Usui et al., 1999). The PD pattern of Fmi was not disturbed in wild-type cells proximal to afz clone.  A PCP screen based on tissue-specific overexpression of random Drosophila genes has identified a potential core PCP gene, diego (Feiguin et al., 2001). Overexpression of Diego (Dgo) produced PCP phenotypes in sensory bristles, ommatidia and wing hairs, however, loss-of-function dgo mutants had defects only in wings and eyes (Feiguin et al., 2001). Unlike fmi, mutations in dgo are homozygous viable. Dgo co-localized with Fmi during wing development: it displayed an early cortical distribution and temporal PD relocalization that matched Fmi, with the exception that Dgo remained polarized during hair formation (Feiguin et al., 2001). Dgo immunoreactivity did not become cortical in/mi clones; conversely, Fmi was cortical in dgo clones, but failed to polarize. Wild-type cells do not localize Fmi on cell boundaries abutting/m/ clones (Usui et al., 1999); similarly, wild-type cells did not localize Dgo to the boundary of a fmi clone (Feiguin et al., 2001). Despite the necessity of Fmi to localize Dgo, Fmi was not sufficient to recruit Dgo to an ectopic location. Overexpressed Dgo, however, recruited Fmi. The dynamic distribution of Dgo was dependent on Fz in a manner analogous to Fmi: Dgo localization was abolished within fz clones but accumulated at the clone edges irrespective of the border axis. Consistent with the observed interdependence of Dgo and Fmi for correct localization, Dgo was reoriented in wild-type cells distal to afz clone. It is not known if Dgo is present on both cells at a PD boundary, as dgo clones were not examined in this study (Feiguin et al., 2001).  As was the case fox fmi and stun, separate mutations in another core PCP determinant were simultaneously identified and reported: strabismus and Van Gogh are allelic and affect the polarity of ommatidia and sensory bristles (Wolff and Rubin, 1998) as well as wing hairs (Taylor et al., 1998). The stbm/Vang locus encodes a protein that contains four putative transmembrane domains as well as a PDZ-  33  binding epitope; these features are conserved in proteins found in mice and humans (Wolff and Rubin, 1998) as well as in zebrafish and Xenopus (Park and Moon, 2002). In Vang mutants, prehairs initiated in the center of wing cells, phenocopying/z-pathway mutations (Taylor et al., 1998). Vang mutant clones display directional non-autonomy opposite to that of fz: Vang clones alter the polarity of wild-type cells proximal to the clone and reorient hairs to point away from the Vang cells (Taylor et al., 1998). The directional non-autonomy of Vang clones follows aberrant pk polarity with these properties intact, as was observed for fz (Adler et al., 2000). The directional non-autonomy of Vang required functional Fz; conversely, fz. non-autonomy required functional Vang, suggesting that Vang is required to transduce the Fz signal (Taylor et al., 1998). The mechanism by which Stbm/Vang does this is unclear, however, recent work in vertebrates has demonstrated that a vertebrate Stbm binds Dsh and promotes PCP-like signaling at the expense of the canonical Fz pathway (Axelrod, 2002; Park and Moon, 2002). The production of an antibody against Drosophila Stbm/Vang to determine its subcellular localization during PCP signaling and/or biochemical interactions with Dsh will resolve the question if these findings also apply to Drosophila epithelia.  The roles of Inturned, Fuzzy and Multiple wing hairs  In contrast to the (fz, dsh, pk) group first categorized by Wong and Adler (1993), the inturned (in)lfuzzy (fy) and multiple wing hairs (mwh) groups have received much less attention. Mutations in these two groups produce supernumerary wing hairs that initiate at the cell periphery and have aberrant polarity (Wong and Adler, 1993). Prehairs in these mutants are prefigured by multiple sites of distal actin accumulation; this phenotype persists in the absence of fz, dsh or pk signaling, indicating that these genes are epistatic to, and likely downstream of, the fz. group (Wong and Adler, 1993). Several alleles of in display a temperature-sensitive phenotype; temperature-shift experiments demonstrated that in function is required prior to, but not following, prehair initiation (Adler et al., 1994). Clonal analysis of  34  in found that mutant clones could express the in phenotype (Gubb and Garcia-Bellido, 1982), but subsequent work with marked clones determined that in is cell-autonomous in its effects on P C P (Park et al., 1996). The in locus was cloned and found to encode a protein with no similarity to known proteins or motifs (Park et al., 1996). Expression of epitope-tagged In constructs in tissue culture suggest that In is entirely cytoplasmic (Yun et al., 1999) despite earlier suggestions that In may have two transmembrane domains (Park et al., 1996). Attempts to determine the subcellular localization of In have met with problems similar to the early investigation of Fz. Anti-In antibodies do not detect the endogenous protein, but do detect an overexpressed heat-shock In transgene (hs-iri) (Yun et al., 1999). In flies carrying this construct the In protein is uniformly distributed around the cell periphery in pupal wing cells (Yun et al., 1999). While this transgene rescues the in phenotype (Park et al., 1996), it is unclear if the distribution of In during wing development has been completely described. Future work with improved antibodies or a functional GFP-tagged In protein, which have been instrumental for Fz and Dsh (see above), will be necessary to definitively address this question. A recent genetic study has upheld the notion that in functions downstream of/z (Lee and Adler, 2002). While clones simultaneously mutant for in and fz displayed typical fz non-autonomy over neighboring wild-type cells, in mutant cells did not alter their polarity in response to an adjacent/z clone. Thus in function is required for the Dsh-independent, non-autonomous function of F z only in the recipient cell (Lee and Adler, 2002). Similarly, cells mutant for in did not respond to gain-of-function phenotypes induced by overexpressed Pk, Sple or F m i , placing in downstream of these genes (Lee and Adler, 2002). Consistent with this result, the localization of F z and Fmi was not altered in wings homozygous for in (Strutt, 2001a).  The functions offy and mwh are even less well understood than for in. Mutations mfy are cell autonomous (Collier and Gubb, 1998). The fy locus encodes a protein with four putative transmembrane domains but no other obvious motifs (Collier and Gubb, 1998). Consistent with the results for in, F z  35  non-autonomy requires fy in the recipient cell (Lee and Adler, 2002). The subcellular location of Fy is unknown, as an antibody directed against Fy has not yet been produced. As was the case for in, the dynamic localization of Fz is unaffected by fy (Strutt, 2001a). Mutations in the mwh locus produce an average of about four hairs per cell, which exceeds the in and/y phenotype (Wong and Adler, 1993). As for in and fy, mwh is cell-autonomous and required for fz non-autonomy in the recipient cell (Lee and Adler, 2002). Additionally, the localization of Fz, Dsh and Fmi is not affected in mwh tissue (Axelrod, 2001; Strutt, 2001a; Usui et al., 1999). These results are consistent with the interpretation that mwh is downstream off?. (Wong and Adler, 1993). The mwh locus has not been cloned, precluding speculation as to its molecular function.  The cytoskeleton and wing P C P : the roles of Cdc42, R a c l , RhoA, Drok and Widerborst  Several lines of evidence implicate regulation of the actin/microtuble cytoskeleton in wing PCP. Racl, Cdc42 and RhoA are small GTPases involved in cytoskeletal regulation; these proteins have been identified in Drosophila (Luo et al., 1994; Strutt et al., 1997). Expression of a dominant-negative form of Cdc42 (DN-Cdc42) blocked actin polymerization in the forming prehair, leading to cells with stunted, malformed adult hairs or no hairs at all (Eaton et al., 1996). Expression of dominant-negative Racl (DNRacl), however, did not disrupt hair elongation, but produced duplicated prehairs (Eaton et al., 1996). In contrast to those on fy, in and mwh wings, duplicated adult hairs produced by DN-Racl-expressing cells did not have defects in polarity, but only in number, indicating that these processes are genetically distinct (Eaton et al., 1996). Investigation of the effect of DN-Racl on the cytoskeleton indicated that the circumferential band of actin localized to the apical junction region of wing epithelial cells was disrupted. Similarly, the apical, dome-shaped microtubule web present in wild-type cells was disorganized in cells expressing DN-Racl. In contrast to the specific effects of Racl on hair number, mitotic clones homozygous for hypomorphic RhoA mutations produced duplicated wing hairs that  36  exhibited non-distal polarity in a manner similar to in or fy mutants (Strutt et al., 1997). Thus three small GTPases have nonequivalent functions in hair outgrowth, polarity, and regulation of hair number (reviewed by Eaton, 1997).  Direct perturbation of the actin and microtubule cytoskeletons through application of cytoskeletal-disrupting drugs produced wing hair phenotypes consistent with the genetic manipulation of Cdc42 and Racl (Turner and Adler, 1998). Application of vinblastine (VB), which destabilizes microtubules, eliminated prehairs in wings treated before 28 hours APF. The effect of VB was partial when applied between 29-30 hours. After 31 hours, VB induced multiple prehairs with wild-type polarity, phenocopying DN-Racl expression (Turner and Adler, 1998). Exposure of pupal wings to the actin antagonist cytochalasin D (CD) produced bifurcated prehairs, which is a phenocopy of crinkled mutations (see below). Consistent with the expression of DN-Cdc42 (Eaton et al., 1996), higher doses of CD eliminated prehair elongation altogether (Turner and Adler, 1998). These results confirm the finding that actin and microtubules are required for hair elongation and restriction of prehair initiation to a single site.  A recent major study has further implicated the actin cytoskeleton in the Fz-dependent regulation of prehair number, and has separated prehair number and polarity through characterization of the Drosophila Rho-associated kinase, or Drok (Winter et al., 2001). Drok was shown to bind activated RhoA directly. The molecular structure of Drok includes a kinase domain, a RhoA binding-domain, and a pleckstrin homology domain containing a nested cysteine-rich motif. A loss-of-function Drok allele was obtained by screening for lethal mutations that could be complemented by a Drok transgene. Drok clones produce duplicated wing hairs with normal polarity; duplicated hairs were prefigured by multiple foci of distal actin in pupal wings (Winter et al., 2001). The Drok phenotype was cell-autonomous. A wild-type Drok transgene specifically suppressed multiple wing hair formation in dsh wings without +  1  37  affecting the dsh polarity pattern. This result was interpreted as Drok functioning only to regulate hair formation and not hair polarity (Winter et al., 2001). A major substrate of Drok was identified as the product of the spaghetti-squash locus (sqh) (Winter et al., 2001). Sqh is the Drosophila non-muscle myosin II regulatory light chain (MRLC) (Karess et al., 1991). Drok was shown to phosphorylate Sqh in vivo, and a Sqh transgene with substitutions mimicking phosphorylated serines (SqhE20E21) partially rescued Drok lethality and suppressed multiple hair formation in rescued adults (Winter et al., 2001). Phosphorylated Sqh was present at the cell cortex in wild-type developing wing cells and absent in Drok clones. The suppression of the dsh multiple hair phenotype by a Drok transgene was mimicked by the +  SqhE20E21 construct, indicating that Sqh is downstream of Fz and largely responsible for the effects of Drok on dsh mutants (Winter et al., 2001). This analysis was extended to include other Drosophila myosins that associate with Sqh: Zipper (Zip) and Crinkled (Ck). Zip is a non-muscle myosin heavy chain (Young et al., 1993), whereas Ck is the Drosophila myosin VILA (Turner and Adler, 1998). Clonal analysis of zip mutants revealed partial penetrance for a duplicated hair phenotype; Zip protein was also found to asymmetrically prefigure the site of prehair initiation in wild-type cells and relocalize to the cell center in dsh' mutant cells (Winter et al., 2001). Cells mutant for ck produce multiple and split hairs (Turner and Adler, 1998). Mutations in zip and ck had opposite effects on the dsh multiple hair phenotype: zip enhanced this aspect of the dsh phenotype but ck suppressed it. This result suggested that a balance of Zip and Ck activity regulates prehair number (Winter et al., 2001). The possibility that in and/y are downstream of RhoA/Drok was tested by attempting to modify in and/y hair number/polarity phenotypes with the activated SqhE20E21 transgene. That this construct had no effect on these genotypes, taken together with the observation that Drok specifically alters hair number and not polarity, seems to indicate that RhoA/Drok function separately from these PCP determinants (Winter et al., 2001). Thus the best evidence to date for actin-cytoskeletal participation in wing PCP does not support a role for In and Fy in this process, although divergent signaling through RhoA may still be upstream of in and/)' (Winter et al., 2001) and account for the hair polarity defects in RhoA mutants (for a review see  38  Strutt, 2001b), (Strutt et al., 1997). A summary genetic model of the frizzled PCP pathway is shown in Figure 8.  In an analysis complementary to the Drok study, a protein asymmetrically localized on the apical microtubule web that regulates hair polarity and number has recently been described. Widerborst (Wdb) is a B' regulatory subunit of protein phospatase 2A (PP2A) in Drosophila (Hannus et al., 2002). Wdb was found in the same overexpression screen that identifed Dgo, but overexpressed Wdb causes a multiple hair phenotype in the wing in contrast to the hair polarity defects observed for overexpressed Dgo (see above). Loss-of-function wdb escapers typically exhibited defects in prehair initiation and occasionally had polarity defects. A dominant-negative Wdb transgene (DN-Wdb) produced similar effects but with higher penetrance; DN-Wdb also acted non-autonomously to induce mutant phenotypes in cells adjacent to its domain of expression. The multiple wing hair phenotype of wdb escapers was dramatically suppressed by loss-of-function microtubule star (mts) mutations. The mts locus encodes the PP2A catalytic subunit; mfa-suppression of wdb suggests that Wdb induces wing hair formation by positively regulating its catalytic subunit (Hannus et al., 2002). At 18 hours APF, Wdb protein was found on the apical microtubule web previously described (Eaton et al., 1996). Wdb distribution on the microtubule web was asymmetric, with higher Wdb levels present distally (Hannus et al., 2002). The microtubule web was disrupted by expression of DN-Wdb, but junctional actin and Cor were not affected. Surprisingly, the localization of Wdb was unaffected in wings mutant for dsh and fmi, indicating that Wdb is targeted via a Fz-independent mechanism. Conversely, Dsh and Fmi failed to polarize in cells expressing DN-Wdb. Wdb thus appears to be upstream of frizzled signaling in wing cells and may function to selectively dephosphorylate PCP components (Hannus et al., 2002).  39  Figure 8. A model of the frizzled  P C P pathway.  Genetic interactions are indicated with solid arrows and/or boxes; the physical interaction between Pk and Dsh is indicated with a dashed arrow. The Fz group (Fz, Dsh, Pk, Fmi, Dgo and possibly Vang) are activated by an unknown signal, probably mediated through Fz or Fmi, to generate asymmetry. The wingless pathway also signals though Dsh, however, frizzled PCP signaling is independent of wingless, possibly through the effect of Vang on Dsh. Downstream of the Fz group is RhoA, Drok and the nonmuscle myosins Sqh, Zip and Ck, which modify the cytoskeleton to determine prehair number and polarity. The PCP determinants In, Fy and Mwh are epistatic to the Fz group, however, the relationship between these genes and the Fz signal is unclear. The small GTPases Rac and Cdc42 also modify the cytoskeleton to alter hair number, but their relationship to the Fz signal remains ambiguous.  40  I Fz Fmi  Wnt  Pk -Dsh Dgo Vang  - • Wg signal -) pathway  RhoA  In/Fy Mwh  /  Drok  I  Sqh/Zip/Ck Rac/Cdc42  F-actin/microtubules  Hair polarity/number  41  Upstream of Frizzled: the functions of Fat, Dachsous and Four-jointed  Three additional PCP determinants, dachsous, fat and four-jointed, have recently been found to function upstream of the frizzled pathway during wing PCP and may constitute a global signal which determines the orientation of/h'zz/ec/-pathway components. The dachsous (ds) and fat (ft) loci encode atypical cadherins that were identified as tumor-suppressors (Mahoney et al., 1991; Clark et al., 1995). Subsequent studies showed that mutations in these loci also disrupt wing PCP (Adler et al., 1998; Strutt and Strutt, 2003; Ma et al., 2003). The ds wing PCP phenotype is qualitatively distinct from the fz phenotype: large regions of hairs have reversed polarity and point from distal to proximal (Adler et al., 1998). Prehairs initiate at the cell center in large ds clones but retain correct polarity in small clones (Strutt and Strutt, 2003). Ds and Ft proteins are localized to the cell membrane apical to adherens junctions, but are not polarized during wing development (Ma et al, 2003). Several lines of evidence suggest that Ft and Ds interact physically in a heterotypic fashion. Clones lacking either protein fail to correctly localize the other (Ma et al., 2003). Additionally, Ds protein mft mutant cells on the border aft clone preferentially localizes to boundaries shared with adjacent wild-type cells. This same effect is observed for Ft protein at the border of a ds clone (Ma et al., 2003). Clones lacking both ds and ft simultaneously do not display this clone boundary effect, consistent with the interpretation that Ds and Ft are required to localize each other. There is controversy whether Ds is expressed in a proximal-distal gradient in pupal wings. Ma et al. (2003) report a proximal to distal gradient of Ds, however, Strutt and Strutt (2003) found no evidence for graded expression of Ds in the wing.  In ds mutants, Fz and Dsh remain polarized, however, the orientation of this polarization follows the ds mutant pattern (Ma et al., 2003). Taken together with their observation of a Ds gradient, Ma et al. (2003) propose that Ft, Ds and Fj provide global information which orients the Fz signaling loop to propagate directional information in a non-autonomous manner. While attractive, this model is  42  unconvincing and inconsistent with the cell autonomous nature of both fmi and dsh mutant clones (Adler, 2002; Strutt and Strutt, 2003). Specifically, the cell-autonomous polarization of Fz is abolished in fmi and dsh mutant clones, however, these clones do not extert non-autonomous effects on neighboring wild-type cells. Additionally, cell-autonomous alleles of fz disrupt the polarization of Fz and Dsh without disrupting neighboring cells. Thus the polarization of Fz, Dsh, Pk, and Fmi does not seem to be a primary mechanism for non-autonomous/z signaling (Adler, 2002). A more attractive model is that Ds, Ft and Fj regulate the polarization-independent, non-autonomous function of Fz (Strutt and Strutt, 2003). Consistent with this hypothesis, Strutt and Strutt (2003) devised a method to rescue only the cell-autonomous function of null/z wings: the resulting phenotype was significantly different than the fz null phenotype and closely resembled the ds phenotype. The non-autonomous fz signal was found to precede the cell-autonomous/z. signal, confirming that cell autonomous polarization of Fz is a result of the non-autonomous fz. signal, not its cause (Strutt and Strutt, 2003). The similarity in phenotype between the ds and non-autonomous fz phenotype is strong evidence that Ds is upstream of the non-autonomous function of fz. Given the reciprocal requirements between Ft and Ds for localization (Ma et al., 2003), Ft also functions upstream of the fz. non-autonomous signal. There is evidence that Four-jointed (Fj) also functions together with Ds and Ft. Ds and Ft localize preferentially to fj clone boundaries (Strutt and Strutt, 2003), suggesting that Fj negatively regulates the Ds -Ft interaction. Fj is a type II transmembrane protein that may be proteolytically cleaved to form a secreted signal (Zeidler et al., 2000). Expression of a//-reporter construct shows a graded expression in pupal wings with its peak at the distal margin (Zeidler et al., 2000; Ma et al., 2003). The localization of Fj protein in the wing or whether it is secreted or membrane associated has not yet been determined. Taken together, the action of Ds, Ft and Fj may regulate the non-autonomous component of the/z signal by providing a global cue based on possible Ds and Fj gradients (Strutt and Strutt, 2003; Ma et al., 2003).  43  Evidence for a/n'zz/ed-independent P C P pathway  The existence of a Fz-independent pathway acting in wing PCP has long been postulated to explain the nonrandom polarity of Fz-pathway mutants (Wong and Adler, 1993). Specifically, hairs produced by cells lacking functional fz signaling tend to be aligned to the hairs of neighboring mutant cells within the aberrant polarity pattern, indicating that the parallel alignment of wing hairs remains normal in/z-pathway mutants. Cell-cell cytoskeletal connections have been identified in the developing wing epithelium and postulated to mediate this Fz-independent signal (Turner and Adler, 1998). A component of this parallel alignment mechanism would display several characteristics. Firstly, mutations which disrupt the function of such a component should disrupt the parallel alignment of neighboring hairs. Secondly, the normal function of such a component should be independent of the frizzled pathway. Thirdly, a parallel alignment component should function to align hairs irrespective of the site of prehair initiation, since hairs remain largely parallel in mutants that alter the prehair initiation site. Additionally, a candidate for a Fz-independent PCP alignment mechanism able to co-ordinate hairs within a field of cells would be expected to have at least short-range non-autonomy in order to align neighboring hairs. Despite the progress made on/h'zz/ed-dependent PCP, no components of this frizzledindependent, parallel alignment PCP mechanism have been identified.  Do septate junctions function in PCP?  An interaction between apical-basal polarity and planar polarity has not been investigated, although several PCP determinants are known to localize to the apical-lateral membrane. Fz, Dsh, Pk, Fmi and Dgo are all membrane-delimited and apically restricted (Strutt, 2002). Additionally, the microtubule web on which Wdb is polarized is an apical structure. Of particular note, this microtubule web was found to have its base directly adjacent to Cor immunoreactivity in wild-type cells,  44  demonstrating at least a tight juxtaposition, i f not direct contact, of the web base to septate junctions (Eaton et al., 1996). This study also described a subtle dynamic localization for Cor during wing development. The distribution of C o r is uniformly cortical prior to prehair initiation. During prehair elongation, however, Cor immunoreactivity becomes less strictly localized to P D boundaries, but acquires a broader pattern. Cor remains tightly localized to A P cell boundaries, however (Eaton et al., 1996). These results are suggestive of a role for Cor and septate junctions in organizing or maintaining the microtubule web, but a mutational analysis of cor or other septate-junction mutants during wing development has not been pursued.  The goal of this thesis is to investigate the role of Gliotactin in the development of polarized post-embryonic epithelia. The possibility that Gliotactin contributes to planar polarity as well as apicalbasal polarity is readily testable using viable Gli mutants. To obtain the appropriate Gli genotypes, a mutagenesis screen was employed (Chapter II). Several hypomorphic Gli alleles were obtained which were adult-viable and had defects in wing P C P , specifically in the parallel alignment of neighboring hairs. The subcellular location of G l i protein in the wing epithelium was determined to be the tricellular septate junction. The contribution of other SJ components to P C P was then assayed by using SJ mutants to enhance the Gli P C P phenotype (Chapter III). SJ mutations enhanced the Gli wing P C P phenotype and were found to act in epistasis groups with respect to P C P determination. Additionally, combinatorial SJ mutants were found to have a wing blistering phenotype at the wing margin. SJ components thus mediate adhesion as well as P C P in the developing wing. The relationship between S J - P C P and the frizzled pathway was then determined (Chapter IV). SJ-PCP was found to be independent of, and parallel to, the/z pathway. Finally, to test the hypothesis that additional factors contribute to SJmediated P C P , a forward-genetics screen for Gli suppressors was performed (Chapter V ) . Several suppressors were obtained, at least one of which plays a role in SJ-mediated P C P and adhesion.  45  CHAPTER II. MUTAGENESIS OF GLIOTACTIN  STRUCTURE AND FUNCTION.  46  Introduction  Gliotactin is expressed in a wide variety of specialized epithelia during development including glia, the epidermis, gut and imaginal discs (Auld et al., 1995; Schulte, 2003). The primary embryonic phenotype of Gli mutants is paralysis at late embryogenesis due to leaky septate-junctions between glial membranes, exposing peripheral nerves to the hemolymph, which has a high concentration of potassium (Auld et al., 1995). The potassium in the hemolymph blocks action potentials in Gli embryos: bathing dissected Gli mutants in artificial hemolymph with low potassium allows muscle contractions to occur (Auld et al., 1995). The embryonic salivary gland epithelium is also compromised in Gli mutants (Schulte, 2003). The paralysis phenotype prevents Gli embryos from hatching into first instar larvae and is therefore lethal.  To address the function of Gliotactin in adult epithelia it is necessary to alter the function of Gli gene at developmental times that follow the lethal phase of Gli alleles at late embryogenesis. This could be accomplished in two ways: by inducing mitotic clones of existing Gli alleles late in development, or by screening for viable hypomorphic Gli alleles which survive to adult stages. The first strategy would demonstrate if Gli phenotypes were cell-autonomous, however, this approach would fail if wild-type cells surrounding the clone rescued the mutant phenotype. The second strategy also allows for correlation of Gli structure and function if novel Gli alleles are sequenced to reveal the underlying mutation altering the function of the protein. To gain information on Gli structure as well as function, a screen for Gli hypomorphs was performed and forms the basis for this chapter.  The primary structural feature of Gli is a large extracelluar domain with similarity to the serineesterase family of proteins (Botti et al., 1998); Gli, however, lacks the catalytic serine and thus is not enzymatically active (Auld et al., 1995). The atomic structure of acetlycholinesterase from the electric  47  ray Torpedo californica (TcAChE) has been determined with 2.8 angstrom resolution and found to contain a ring of negative charge around the active site cleft, or gorge (Botti et al., 1998; Sussman et al., 1991), (Figure 2). Computer models predict that this "annular" negative charge is a conserved feature of Gli structure: indeed, the distribution of surface potentials in Gli show exceptional similarity to TcAChE (Botti et al., 1998). Interestingly, Gli is more similar at the primary sequence level to TcAChE than to functional DrosophUia acetlycholinesterase (Botti et al., 1998). Taken together, these results indicate that the overall three-dimesional topology of the Gli serine esterase-like motif is likely similar to that of TcAChE.  The functional significance of the Gli serine esterase-like domain remains unclear, since mutational analysis of the Gliotactin protein has not been performed with domain specificity. Gli was originally cloned based on its proximity to a presumed peripheral glial-specific enhancer trap, AE2 (Auld et al., 1995). The P-element based AE2 enhancer trap contains a functional white* eye-color marker and lacZ reporter gene inserted into the 5' untranslated region of the Gli locus; additionally, AE2 flies are homozygous viable and display no mutant phenotype (Auld et al., 1995). Deletions obtained through imprecise excision of the AE2 P-element produced the first Gli mutations. All deletions obtained from imprecise excision of AE2 were directed towards the 3' end of the gene. The longest deletion obtained (Gli  ) extended close to the beginning of the coding region and is a recessive allele  AE2A4S  with the characteristic Gli phenotypes. A mouse monoclonal antibody raised against the extreme carboxyl-terminus of Gli (1F6) fails to detect protein in Gli  AE2A45  homozygotes (Auld et al., 1995).  Previous screens for ethylmethane sulfonate (EMS)-induced lethal mutations in the Gli region had identified several complementation groups close to Gli. Complementation tests between Gli excision lines and EMS lines mutant for the lethal (2) 35Dg locus (Ashburner et al., 1990) showed that Gli was allelic to lethal (2) 35Dg(Au\d et al., 1995). These EMS Gli alleles displayed the embryonic-lethal Gli phenotype and also failed to produce a Gli protein containing the 1F6 epitope, suggesting that these  48  alleles were not more informative than the original AE2 deletion lines. In addition to ignorance of the functional significance of Gli structure, Gliotactin function during larval, pupal and adult stages remained obscure in the absence of mutant Gli genotypes that exit embryogenesis (Auld et al., 1995).  To obtain genotypes suitable for addressing Gli post-embryonic phenotypes, a screen for EMSinduced Gli alleles was performed. EMS mutagenesis was chosen over against site-directed mutagenesis for several reasons: (1) EMS-induced alleles are isolated based on altered function (e.g. a failure to complement an existing allele) compared to site-directed mutants which are not guaranteed to alter function; (2) EMS mutagenesis does not require a priori hypotheses of Gli function, which would be inevitably incomplete at best and at worst incorrect; and (3) EMS mutagenesis is a preferred method for inducing point mutations in Drosophila (Ashburner, 1989). Point mutations are a potential source of hypomorphic Gli alleles amenable for studying post-embryonic phenotypes of Gli mutants.  49  Materials and methods  Fly strains and mutagenesis  Flies were grown on standard media at 22"C unless stated otherwise. An explanation of genetic symbols and genotypes of balancer chromosomes is given in Appendix 1. The Gli EMS alleles 1(2) 35Dg  (Gli  RAR7?  ) and 1(2) 35Dg  RAR77  1990),; the Gli  AE2  (Gli" ) were supplied by Michael Ashburner (Ashburner et al.,  PM  4  enhancer trap and imprecise excision allele Gli  ^ were from the Auld lab collection  Ah2A4  (Auld et al., 1995). Males homozygous for a viable Gliotactin enhancer trap (genotype w"" /Y; l  Gli IGli , AE2  AE2  (Auld et al., 1995)) were mutagenized with EMS according to standard procedures  (Ashburner, 1989) and mated to w"'7w' '; Gla/CyO females. Individual Fl w'"*/Y; Gli */CyO males m  were mated with w'"'Vw'" ; Gli s  AE2  YCyO females (Auld et al., 1995) at 25"C. Lines were established for  AE2A4:  Gli * chromosomes that failed to complement Gli AE2  . Each putative novel Gli chromosome was  AE2A4S  serially recombined with the flanking markers black and cinnabar in a w"" background to remove 1  linked mutations, using the lacW marker in the AE2 insert as a marker for the Gli locus. Following recombination putative Gli chromosomes were re-tested for complementation in trans to Gli  \ Only  AE2A4  lines that failed to complement Gli^' ^ after recombination were retained. 2  Complementation and embryonic lethal phase analysis  Adult complementation tests between novel Gli alleles, 1(2) 35Dg  RAR7?  (Gli  ) and 1(2) 35Dg  RAR7?  P34  (Gli ) (Ashburner et al., 1990) were performed with crosses between various balanced stocks at 22"C. P34  Escapers were scored as Cy and black (the Ashburner alleles were induced on a black, cinnabar +  chromosome). The percent viability of each Gli genotype was determined by comparing the frequency of escapers with a control cross using the mutagenesis chromosome. To determine the embryonic 50  lethality of various novel Gli allelic combinations, each Gli allele was balanced over a CyO balancer expressing the green fluorescent protein under the control of the actin promoter (CyO P{lacW, actin+  GFP}). Embryos resulting from crosses between various GFP-balanced Gli stocks were collected, aged to stage 16-17 and scored under a fluorescence dissecting microscope for the absence of the actin-GFP marker. GUI embryos were obtained for each cross and scored for the ability to hatch into first instar larvae.  Sequence analysis of novel Gli alleles  Genomic Gli sequences were amplified and directly sequenced from homozygous Gli embryos collected using the actin-GFP balancer described above. Genomic DNA was prepared from 25-30 homozygous embryos as follows: 25-30 embryos were homogenized in 50 pi of homogenization buffer (10 mM Tris-HCl pH 8, 25 mM NaCl, 200 pg/ml Proteinase K), incubated at 37"C for 30 minutes followed by 2 minutes at 94"C. Overlapping fragments spanning the Gli coding region were amplified using standard PCR conditions in 200 uM MgCf with one unit of Taq DNA polymerase (Amersham). Reactions were cycled 39 times as follows: 10 seconds at 94"C; 1 minute; at 55"C; 2 minutes at 72"C. A final extension for 10 minutes at 72"C was performed after the last cycle. PCR products were gel purified (Qiaex II gel purification kit, Qiagen) and directly sequenced on both strands using an ABI PRISM automated fluorescence sequencer (NAPS Facility, UBC). Changes relative to the AE2 mutagenesis chromosome were confirmed by sequencing an independent PCR product. Primer pairs utilized were as follows: pair 1: AE2 II: 5'-CGG GAA TTC CTT TTA AAT ACA TTT CCT AC-3' and AE2 Hindll: 5'GCC CTA GGC GAA TAC ATC CCG CTC CAG GG-3'; pair 2: AE2 III: 5'-GGA ATT CCT AAT GAT GTT ACC ATT TCC C-3' and AE2 IV: 5'-CGG GAT CCA CCA CGT CGT AGA ACG AGG C-3'; pair 3: AE2 V: 5'-CGG GAT CCA GCC TCC AAC CTA TTC CAG GG-3' and AE2 VI: 5' CGG GAT CCA CGG ATG AGG GTT TCC GGC G-3'; pair 4: AE2 VII: 5'-GGA ATT CCA  51  TTG GTA CGA GGG ATG GCG C-3' and AE2 VIII: 5'-CCT TAA GGA GAA CTC GGT GTC CAT GAA GG-3'; pair 5:AE2 IX: 5'-GGA ATT CCA TCG AAC.GTT ATT TCC TCA CCG-3' and AE2 X: 5'-CCT TAA GGA TTA TCC CTG GAG CGC AGG GC-3'; pair 6: AE2 XI: 5'-GGA ATT CCG CGT GGA GTG GAC AAT GCC C-3' and AE2 XIV: 5'-GCG ATC CTG CGC TTA GAG ACG-3'.  Scoring of mutant wings  Wings from relevant flies were rinsed in 95% ethanol and mounted in 85% lactic acid (BDH) or 1:1 Hoyer's mountant: 85% lactic acid, and examined using bright-field microscopy on a Ziess Axiskop compound microscope fitted with a Sony PowerHAD digital camera. Images were acquired using Northern Elite software and processed in Adobe Photoshop 5.5. The adult wing has seven main regions defined by veins (Ferris, 1950). The number of wing regions containing a region of mutant polarity was counted and averaged for 20 wings of each genotype. A region of mutant polarity was defined as at least two adjacent hairs that had defects in parallel alignment. Comparisons between genotypes were tested by ranking each wing according to the number of mutant regions observed and comparing cumulative rank data using the Mann-Whitney U-test.  Immunostaining of pupal wings  Pupae were aged to 30 hours APF at 25"C and fixed overnight in 4% formaldehyde in phosphatebuffered saline (PBS) +0.5% Tween-20 (PBTw) at 4"C. Fixed pupae were dissected in PBTw, washed and incubated with mouse anti-Gliotactin monoclonal antibody (1F6) (Auld et al., 1995) at a dilution of 1:1. Stained wings were labeled using a goat a mouse Alexa-488 secondary antibody at (Molecular Probes) 1:250, mounted in Vectashield (Vector Laboratories), and visualized using a Bioradiance Plus confocal microscope (Bioimaging Facility, UBC). Wings stained with Alexa-568 conjugated phalloidin  52  were fixed and dissected as described above. Alexa-568 conjugated phalloidin was used at 1:100 dilution for one hour.  53  Results  Isolation and characterization of novel Gliotactin alleles  Complementation analysis between the excision allele Gli  and previously identified EMS  AE2A45  Gli alleles (Gli  and Gli" ) (Ashburner et al., 1990; Auld et al., 1995) demonstrated that homozygotes  KAK77  4  for each allele and all possible heteroallelic combinations have the embryonic paralyzed phenotype. Western blot analysis using an anti-Gli monoclonal antibody also failed to detect Gli protein in homozygous mutant embryos suggesting that these four alleles are nulls or strong hypomorphs. In order to extend the analysis of Gliotactin structure and function, a screen was performed for novel EMS mutations that failed to complement Gli  (Figure 9). Several novel Gli alleles were identified among  AE2A45  a collection of approximately 5000 EMS-mutagenized chromosomes. In order to remove spurious second-site mutations linked to Gli, each novel allele was serially recombined with visible markers flanking the Gli locus. The alleles Gli'"", Gli' " , Gli' '' and Git* failed to complement Gli 1  2  1  3  5  AE2A45  recombination. One presumptive allele (Gli' ) complemented Gli lv4  after  after recombination and was  AE2A45  discarded. Gli' ' , Gli' , and Gli' '' homozygotes were paralyzed at late embryogenesis, however, the 1 1  lv2  1  3  Gli' ' allele was homozygous-viable: Gli' '' /Gli' '' escapers were observed at a frequency lower than 1 5  1  5  1  5  expected for a balanced cross of a wild-type Gli allele.  Complementation analysis between novel and previously identified alleles was performed to determine if additional adult-viable genotypes could be found. Of particular interest was the result that Gli  RAR7?  displayed significant interallelic complementation with Gli*' : Gli' ' IGli 5  h 5  RAR77  transheterozygotes  escaped at an almost wild-type frequency in contrast to the less viable Gli' ' /Gli' ' genotype or the 1 0  embryonic lethal Gli  IGli  RAR77  KAR7?  genotype. Similarly, the embryonic-lethal Gli' allele was viable in M  trans to Gli' " , and even the original presumptive null allele Gli 1  1 0  5  Ak2A45  54  was viable in trans to Gli' ' at a 1 0  Figure 9. A n ethylmethane sulfonate (EMS) screen for novel Gli alleles.  w = white, Gli = Gliotactin, Gla = Glazed, CyO = Curly of Oster. Genetic symbols and the genotypes  of balancer chromosomes are described in Appendix 1.  Isogenized males homozygous for the Gli AE2 enhancer trap and white were exposed to ethylmethane sulfonate (EMS) and mated to balanced Glazed virgin females. Single Fl males bearing a mutagenized chromosome balanced over CyO were selected and mated individually to virgin females carrying a balanced Gli loss-of-function allele. F2 progeny resulting from Fl single-pair matings were scored for the white* marker associated with the AE2 insert and for the dominant balancer marker Curly. Lines that failed to produce viable straight-winged progeny indicate that the C//7C/('  4i2zlJ,  genotype is lethal  for that single-pair mating, representing a putative Gli allele. Non-complementing Gli* chromosomes were distinguished from G//'  4i2iWl  by virtue of eye color and isogenized.  55  F o l cftf  Fi  cf  \,1118 Y  * AE2  *  Gli  '  1U8  *  G  /  *3  AE2  1118  w  1118  3  / CyO  o r Y  1118  1118  '  AE2A45  1118  w  '  CyO  = mutagenized chromosome  56  o  '  r  CvO  3  '3  3  3  o  r  3  3 *  '  * Gli  3  *  GUAE2A45  wUlSorY*  3  3  GU  o r Y  Gla  (jijAE2A45  1118'  *  G  w  w  wl"8 X  * nAE2  w  1118  w1118' CvO  3  CyO  F2  X  AE2 *  U  *Gli  W  w  5  3  3  o  r  3  Score for failure to complement  very low frequency (Table 1). Based on the frequency of transheterozygous adult escapers when paired with with Gli' ", interallelic complementation analysis resolved novel and previously identified Gli 1  alleles were into an allelic series: Gli  RAR77  > Gli' " > Gli' - Gli 1  5  lrl  AE2A45  > Gli' " ~ Gli" (Table 1). The adult 1  4  viability of these Gli genotypes may simply reflect the proportion of embryos that die at late embryogenesis, which is the lethal phase of null Gli embryos. To address this possibility, three Gli genotypes were tested for embryo to first instar hatching (Table 2). Surprisingly, even the null allele Gli' " in trans to Gli' ' was able to exit embryogenesis, indicating that the observed adult lethality of the 1  3  1 0  Gli' ' 1GU' genotype was not merely a function of embryonic lethality. This indicates that lethal phase 1 0  hi  of partially viable Gli genotypes occurs after late embryogenesis, suggesting that Gli functions later in development.  The sequence of several Gli alleles was determined by PCR amplification of the Gliotactin locus from homozygous mutant embryos and direct sequencing. In each case a single base-pair substitution altering the coding sequence was identified (Figure 10); each substitution was a GC-AT transition characteristic of EMS mutagenesis (Ashburner, 1989). The homozygous-viable Gli' ' allele is a missense 1 0  mutation at codon 525 within the serine esterase-like domain that alters a glycine codon (GGA) to a glutamic acid codon (GAA). The analogous glycine in Torpedo acetylcholinesterase functions as a hinge between two oc-helices (Sussman et al., 1991). Computer models of Gliotactin based on the crystal structure of Torpedo ACE (Botti et al., 1998) predict that a substitution of a glutamic acid at this position would induce a kink into the three-dimensional structure of the serine esterase-like domain by hindering the flexibility of this joint (S. Botti, personal communication). The remaining alleles were a series of nonsense mutations that paralleled the allelic series. The severe alleles Gli" and Gli' '' 4  1  3  contained premature stop codons early in the Gliotactin open reading frame at positions 108 and 203, respectively; both were glutamine (CAG)'to stop (UAG) transitions. These severely truncated products likely represent protein-null alleles. The Gli' "' allele contained a tryptophan (UGG) to stop (UGA) 1  57  Table 1. Interallelic complementation between Gli" -' v  and homozygous-lethal Gli alleles reveals an  allelic series based on adult viability  Gli/CyO Cross  Escaper genotype  Gli +/CyO x Gli /CyO  Gli l+  Gli  Gli  dv5  /CyO  dv5  x Gli ICyO  RAR77  dv5  Gli /CyO  x Gli /CyO  dv5  dv5  Gli ICyOxGli ICyO dv5  v  5  /Cli  dv5  dv5  x Gli /CyO dv5  Gli  d v 5  IGIi  Gli /CyO  x Gli /CyO dv5  Gli  d v 5  IGli  Gli  x Gli /CyO  Gli  d v 5  /Gli  /CyOx  Gli /CyO  AE2A45  dv3  /CyO  F34  dv5  * Only the Gli ^IGli ^ dv  dv  7  7  (males)*  dv5  /Gli  dv5  R  (females)*  dv5  d v 5  Gli  A  Gli IGIi  Gli  dv  R  Gli IGU  dv5  Gli '/CyO  d  d v l  A E 2 A 4 5  d v 3  R 3 4  Gli /Gli dv5  progeny  progeny  782  417  Frequency of escapers relative to Gli  d x  ^/+  i.00  681  291  0.80  1268  106  0.16  1384  66  0.09  998  17  0.03  750  8  0.02  1100  5  <0.01  1142  1  <0.01  genotype showed a significant difference in viability between males and  58  Table 2. Embryos heterozygous for Gli  Genotype  Unhatched  d v 5  e x i t embryogenesis  Hatched  Total  Embyronic viability  Giidv5, ijRAR77  0  56  56  1.00  Gli IGU  d  S  0  62  62  1.00  Gli IGIi  <  lv3  2  102  104  0.98  G  dv5  llv5  v  59  Figure 10. Domain structure of Gli and molecular sites of Gli E M S alleles.  Gli is a single-pass transmembrane protein with a signal sequence (Amino acids 1-28), a large extracellular motif similar to serine esterases (amino acids 175 - 723), and a consensus sequence for binding PDZ domains at the extreme carboxyl terminus (amino acids 956-959). The sequence changes present within several Gli alleles are indicated. A A indicates a nonsense mutation.  60  Signal sequence  GliPM =GliQ108A Glidv3 =GUQ2Q3A  GU dvl = GU W454A  Serine esterase-]ike  Glidv5 =GliG525E  Transmembrane  723^  746  PDZ-binding epitope  A  •Gli RAR77 =GljSH20A 956 959  61  nonsense mutation at codon 454, which is within the serine esterase-like domain. This allele was significantly more adult-viable in trans to Gli' " than Gli' " (Table 1, X test, P<0.001), indicating that 1  5  1  3  2  Gli' may retain partial function and thus represent a strong hypomorphic allele when in trans to Gli' . lvl  lv5  The Gli  RAR7?  allele is a nonsense mutation at position 820 where a serine codon (UCG) is replaced with a  stop (UAG). This mutation lies in the intracelluar domain of Gli and presumably produces a truncated protein that includes the transmembrane domain and a portion of the intracellular region. This allele retains significant function based on its viability in trans to Git ; it thus represents a weak hypomorphic 6  allele when paired with the hypomorphic Gli' but a null allele when homozygous or paired with alleles lv5  other than G/f'". 5  Postembryonic  Gliotactin phenotypes  The availability of adult-viable Gli genotypes allowed examination of adult tissues for mutant phenotypes to asses the role of Gli in these structures. Unlike the septate-junction mutant cor, Gli mutants do not have rough eyes or missing ocelli (not shown), however, Gli genotypes do have a moderate penetrance for malformed legs including ectopic joints, fused joints and shortened tibia (Figure 11). Tibia and tarsus segments were most commonly affected, as was the tibia-tarsus joint. Closer inspection of leg cuticle in Gli mutants revealed defects in planar cell polarity (PCP) of the leg epithelium (Figure 11). Specifically, leg hairs exhibited defects in parallel alignment and often converged in a chevron pattern. Leg hairs were misoriented in Gli genotypes, however, sensory bristle polarity appeared normal. Since the wing is a readily accessible structure with well-characterized PCP, and since genes have been identified that affect planar polarity in both the wing and the leg (Adler, 2002; Krasnow et al., 1995; Mlodzik, 2000), adult wings of Gli mutants were examined for PCP phenotypes (Figure 12). Wild-type wings have parallel hairs aligned along the proximal-distal axis (Figure 12A), indicating that the pupal wing epithelium correctly established planar polarity. The Gli  62  Figure 1 1 . Gli mutants have altered leg morphogenesis and defects in PCP.  (A) A wild-type leg with several segments indicated. (Adapted from Ferris, 1950). (B) A newly-eclosed Gli' IGli h5  dv3  adult. One metathoracic leg exhibits an ectopic joint in the tibia  (arrow). The contralateral leg is completely missing. (C) The tibia and tarsus of a Gli' ' /Gli' h 5  metathoracic leg. The first tarsal segment is shorter and  lvl  thicker than normal and has a higher bristle density. The joint between the tibia and tarsus is abnormal, as is the first tarsal joint. (D) The tibial-tarsal region of a Gli'^/Gli' ' metathoracic leg. As in (B), thickening of the first tarsal 1 0  segment is evident. Only two tarsal joints are present. (E) A high-magnification view of leg hairs on the femur of a wild-type (Gli IGU ) metathoracic leg. At2  AE2  Hairs and bristles have normal proximal to distal polarity. (F) The equivalent region of a Gli'^/Gli' *' metathoracic femur. Several hairs have altered polarity and 1  converge in a chevron pattern (arrows). Bristles and associated bracts have normal polarity. (G) A Gli ''IGli' dv:  lv3  metathoracic femur. Numerous hairs have altered polarity. Bristles and associated  bracts retain normal polarity.  Magnification: B = 25X, C-D = 100X, E-G = 400X.  63  64  Figure 12. Wing P C P defects in Gli mutants.  Anterior is up and distal is to the right for all images.  (A) A schematic diagram of a wild-type wing. Five longitudinal veins (LI-L5) and two crossveins are indicated. The anterior crossvein is indicated with an a; the posterior crossvein is indicated with a p. (B) A light micrograph of a wild-type wing showing distally-orientated hairs. (C) The distal margin of a Gli' ' IGli h 5  RAR7?  wing anterior to L3. Several adjacent cells have reciprocal  polarity defects causing hairs to converge in"a chevron pattern (arrowheads). (D) A Gli' /GW ' wing posterior to the posterior cross vein. Numerous hairs have disrupted planar ho  h s  polarity. (E) A Gli' ' IGli' h 5  lrl  wing anterior to the posterior cross vein. Numerous hairs are thinned and bent.  (F) The posterior margin of a GW'^/GW ^ wing. Numerous hairs have disrupted planar polarity and 1  chevrons in close proximity do not have aligned polarity (arrowheads). (G) A higher magnification of a Gli' IGli' ,v5  wing. Note the numerous hairs re-orientated at a consistent  lri  angle to both the anterior and posterior. (H) The subcellular location of Gli in the developing wing at 30 hours APF. Gli was localized to the tricellular junction. A higher magnification view is given as an inset (the magnified area is delimited by a white square). (I) Actin prehairs visualized in a Gli' ' /Gli''" wing using fluorescent phalloidin. Prehairs initiated at the h 5  5  distal vertex.  Magnification: B-F = 200X, G = 400X, H=l ,400X (inset = 2500X), I=2000X.  65  66  allelic series shows increasing severity of wing PCP defects in consistent with the lethality of the various genotypes (Figure 12C-G). Disruption of wing PCP in Gli mutants appeared essentially random with respect to the location of an affected area. Unlike mutations in classical wing PCP determinants such as frizzled or disheveled, which produce whorls of hairs, the orientation of wing hairs in Gli mutants was not stereotypic, but followed a stochastic pattern. Hairs with altered polarity were still directed distally, but miss-orientated towards both anterior and posterior. This reorientation was often at a consistent angle (Figure 12G) and usually occurred in patches, resulting in non-parallel neighboring hairs. Patches of deformed hairs were occasionally observed in all viable Gli genotypes, often at the anterior-posterior boundary (Figure 12E). Deformed bristles were thinner than normal and were often bent. A consistent feature of the Gli wing PCP pattern was adjacent cells with reciprocal polarity defects where the hairs of the two cells converge in a chevron pattern (arrows, Figure 12C.F). Hairs were never observed to reverse polarity from proximal-distal to distal-proximal, nor were multiple hairs per cell observed, even in near-lethal Gli genotypes (Figure 12G). The adult wing has seven major regions defined by wing veins (Figure 12A). Wings from several Gli genotypes were scored for the number of regions that contained a sector with mutant polarity. The Gli' IGli h0  RAR77  genotype had an average of 2.9 regions per  wing with a sector of mutant polarity (n= 20 wings scored). Similarly, the Gli' " /Gli' ' and 1  5  1 6  Gli' IGli h6  dvl  genotypes had averages of 3.4 and 4.1 regions that contained mutant sectors, respectively (n = 20 for each). While an increase in the average number of mutant regions correlated with the increasing lethality of these genotypes, these differences were not statistically significant from each other (Mann-Whitney U-test, P>0.05).  Given the effect of Gli mutants on the polarity of wing hairs, the subcellular distribution of Gli protein in the pupal wing epithelium was investigated using confocal microscopy. Consistent with the observed hair polarity defects in Gli mutants, Gli is localized to the tricellular corners at the common boundary between three epithelial cells in pupal wings at 30 hours APF formation, the time when PCP  67  signaling occurs (Figure 12H). This result is consistent with the expression pattern of Gli in the embryonic epidermis (Schulte, 2003). Gliotactin is thus localized directly subjacent to the site of prehair initiation at the distal vertex of each cell in this epithelium.  To test the possibility that the Gli non-parallel wing hair phenotype was the result of an altered site of prehair initiation, Gli''**I'Gli' " pupal wings were stained using fluorescent phallodin to visualize 1  5  the actin-rich prehairs (Figure 121). No alterations in the prehair initiation site were observed in GW'^/Gli' " wings. This suggests that the Gli mutant phenotype is not a consequence of an altered 1  prehair initiation site, but rather a consequence of altered hair orientation and parallel alignment.  68  Discussion  The purpose of this chapter was to identify and characterize hypomorphic Gli genotypes as a starting point for investigating the role of Gli in the development of adult polarized epithelia. Three hypomorphic and two novel null alleles were identified and sequenced. The only homozygous-viable allele was a point mutation in the extracellular serine esterase-like domain; all other alleles resulted from premature stop codons. These results produced Gli combinations that range in viability from near-lethal to almost complete viability. Consistent with the hypothesis that Gli is required for polarity in adult epithelia, adult-viable Gli animals had defects in the parallel alignment of hairs on both legs and wings. Gli mutants thus alter a component of PCP that has not been previously described.  Gli alleles exhibit intragenic complementation  The molecular analysis of novel and previously identified Gli alleles here described represents the first mutational analysis of Gli with domain specificity. Gli alleles with lesions in different regions of the protein could complement each other, suggesting that Gliotactin contains domains that function independently of one another. Notably, a homozygous lethal truncation within the intracellular domain (Gli  ) was able to rescue the lethality of an extracellular point mutation (Gli' ) to a great extent. The  RAR7?  h0  homozygous lethality of Gli  HAR7?  demonstrates the necessity of the intracellular portion of Gliotactin for  viability, whereas the Gli' * allele indicates that the esterase-like domain is also important for Gli 1  3  function. Gliotactin also appears to be dose-sensitive: the Gli' allele had a significantly higher viability lvS  when homozygous than when paired over null alleles. An interesting result is that an extracellular truncation within the esterase-like domain (Gli' " ) retained partial function when in trans to the Gli' ' 1  1  h 5  allele despite the absence of a transmembrane domain. That this protein retains some function suggests that it can be localized correctly and is not simply secreted. A possible explanation for this observed  69  interallelic complementation is that Gliotactin monomers bind via their serine esterase-like domains and function as a dimer, as is the case for Torpedo acetylcholineasterase (Sussman et al., 1991). Physical association between a Gli' monomer and a truncated Gli' '' monomer could retain the Gli' " product at h0  1  1  1  1  the membrane and explain its observed function. Similarly, physical association between Gli' ' and 1 0  Gli  RAR77  monomers would allow for a multimer with partial function in both extracellular and  intracellular regions, explaining the observed synergistic function of this transallelic combination. This model is also consistent with the observation that Gli' " homozygotes are less viable than Gli''" IGli 1  5  5  RAR77  transheterozygotes, as Gli' " dimers would not posses at least one serine esterase-like domain with 1  5  normal structure. Whether the observed interallelic interactions of Gli alleles arise through physical association or merely information/functional transfer between monomers, these results demonstrate the importance of both the intracellular and esterase-like domains in Gli function.  Gli mutations disrupt P C P in leg and wing epithelia  Adult-viable Gli genotypes had observable defects in the planar polarity of leg and wing hairs, but not ommatida or sensory bristles. Gli thus seems to function specifically in PCP signaling within a single cell, unlike classical PCP determinants such as Frizzled or Dishevelled, which function in all three tissues (Adler, 2002). A role for Gli in eye PCP cannot be ruled out, however, since subtle eye PCP defects may not produce a rough eye phenotype. The Gli PCP phenotype appears qualitatively different from the phenotypes of/z or dsh mutants in several ways. First, wing hairs in Gli mutants never reverse polarity, whereas fz. and dsh mutants have large regions of hairs oriented from distal to proximal (Wong and Adler, 1993). Second, the Gli phenotype appears generally in patches of hairs with reciprocal polarity defects, whereas mfz and dsh mutants adjacent wing hairs are usually parallel in their aberrant polarity (Wong and Adler, 1993). Third, fz and dsh have a low penetrance for a multiple wing hair phenotype (at least one or two cells per wing will produce two hairs) (Wong and Adler, 1993); Gli  70  wings, however, never had multiple hairs per cell. Fourthly, the subcellular location of prehair initiation in Gli mutants is at the distal vertex, the wild-type location. Taken together, these observations raise the possibility that Gli may be a founding member of a/z-independent parallel alignment mechanism, or conversely that Gli may act downstream of the frizzled pathway during the extension of the wing hair.  The subcellular location of Gli in the developing pupal wing matches the Gli expression pattern in the embryonic epidermis (Schulte, 2003). The distribution of Gli is unique among septate-junction proteins and may be of special importance in the wing. The role of other SJ proteins in PCP is unknown. The aim of the following chapter is to determine if the SJ proteins Dig, Scrib, Nrx and Cor participate with Gli in determining planar polarity in the developing wing. Also, the role of vulcan, a novel locus that may encode a SJ protein, is investigated.  71  C H A P T E R III.  GLIOTACTIN  IS N E C E S S A R Y F O R S E P T A T E  PLANAR CELL POLARITY AND  ADHESION.  72  JUNCTION-MEDIATED  Introduction  The finding that mutations in Gli affect the planar polarity of the developing wing raises the possibility that other known SJ components also play a role in wing epithelial PCP. The aim of this chapter is to investigate the roles of the SJ mutants discs-large, scribble, coracle and Neurexin IV in Glimediated wing PCP. To test the hypotheses that Discs-large, Scribble Neurexin IV and Coracle participate with Gliotactin in determining wing PCP interaction two possible approaches could be taken: (1) examine the phenotype of adult-viable mutants in each SJ gene, or (2) determine if strong alleles of each locus can dominantly enhance the Gli phenotype. The first approach is essentially correlative in nature, whereas the second approach would reveal if genetic interactions exist between Gli and other SJ mutations, which has not been attempted previously. For this reason the second strategy was pursued.  In the embryonic epidermis, complete loss of Dig function causes mislocalization of Cor and thus Nrx (Woods et al., 1996). Mutations in the dig locus also show strong genetic interactions with scrib, indicating that these two genes act independently of each other. Conversely, Cor and Nrx form a functional unit: each is necessary for the other's localization and function (Ward et al., 1998). Dig is not mislocalized in cor or Nrx mutants, indicating that Dig is targeted independently of Cor/Nrx (Ward et al., 1998). The work in the embryonic epidermis predicts a model of genetic interactions at the septate junction (Figure 13). This model is testable using PCP to determine if similar interactions occur in the wing. Additionally, interactions between Gli and SJ mutants can be assayed using this system.  In order to maximize potential enhancement of the Gli mutant phenotype null or strong alleles of dig, Nrx and cor were used. The dig null allele dig'" results from a nonsense mutation that produces a 52  truncated protein (Woods et al., 1996); this truncated product fails to localize to the septate junction but remains cytosolic (Woods and Bryant, 1991). Thus in the absence of post-transcriptional gene  73  Figure 13. Genetic interactions between SJ components in the embryonic epidermis.  Arrows indicate genetic interactions. The combined function of Lgl, Scrib and Dig is required to localize all three to the septate junction; as a result, strong genetic interations exist between all three components. The function of the Scrib/Lgl/DIg is required to localize Cor and Nrx to the septate junction. The mechanism by which Dig is targeted to the membrane at the level of the septate junction is unknown (question mark). Cor and Nrx are interdependent on each other for localization to the septate junction but are not required for the localization of other SJ components. They thus form an epistasis group (dashed box).  74  Apical  )  Dig  (Scrib)  ^ Nrx IV  Intracellular  Extracellular Basal  75  regulation, the dig" allele would reduce the amount of functional Dig at septate junctions. The cor 02  3  allele lacks detectable Cor protein and acts genetically as a null allele (Lamb et al., 1998), although this allele has not been sequenced. The Nrx allele is either null or strongly hypomorphic, additionally, the 46  Nrx " allele is hypomorphic (Baumgartner et al., 1996), permitting comparison between Nrx23  enhancement by alleles of two strengths. Sequence information for several EMS-induced Nrx alleles (including Nrx and Nrx ) has not been published and thus is not available. In lieu of a null allele of 46  2511  scrib, a hypomorphic P-element allele (scrib' ) was employed. This allele of scrib has a P-element 7B:i  insertion within the first intron of the Scrib transcribed region (Bilder and Perrimon, 2000) and is homozygous lethal.  In addition to previously confirmed SJ components, the novel protein Vulcan has recently been suggested to be localized to the septate junction (Gates and Thummel, 2000). Mutations in the vulcan (vie) locus are required for leg morphogeneis and are lethal during pupal stages. Wings of vie pupae have gross morphological defects that have not been carefully investigated. The Vulcan protein is a member of the SAPAP family, a group of proteins that has been shown to physically associate with membrane-associated guanylate-kinase (MAGUK) domains (Gates and Thummel, 2000). Based on the observations that dig mutants disrupt leg morphogenesis and Dig has a MAGUK domain (Woods and Bryant, 1991), Vulcan has been postulated to bind Dig at the septate junction. This proposition is purely speculative, since an antibody for Vulcan has not been produced and its subcellular location has not been determined (Gates and Thummel, 2000). Despite these shortcomings, vie may represent a SJ mutant and thus was included. To determine if vulcan interacts with Gli in wing PCP, the P-element allele vie" " was obtained and recombined into a Gli background. 7  22  Three Gli genotypes were chosen as candidates for dominant enhancement by SJ mutations: Gli' IGU h6  ,  RAR7?  Gli IGli' ~, and Gli' IGW . These three genotypes cover a range of viabilities and have dr5  h3  lv5  lvl  76  moderate wing PCP phenotypes that would allow for enhancement to be detected. Additionally, the specific mutations underlying these three alleles provide several diverse Gli structures that may detect allele-specific interactions that would be missed if only one Gli combination were employed.  77  Materials and methods  Fly stocks and alleles  Crosses were performed using standard medium at 22". Peter Byrant provided discs-large'" ; cor 52  and cor were supplied by Richard Fehon. The Neurexin IV alleles Nrx and Nrx * were provided by 13  46  2  11  Hugo Bellen. The Bloomington stock center provided the P-element lines scribble' and vulcan" " An 783  7  22  explanation of genetic symbols and genotypes of balancer chromosomes is given in Appendix 1.  Scoring of mutant wings  Dissection, mounting and scoring of mutant wings was performed as described in Chapter II.  Immunostaining of pupal wings  Pupae were aged to 30 hours APF at 25"C and fixed overnight in 4% formaldehyde in PBS at 4"C. Fixed pupae were dissected in PBS and incubated with rabbit anti-Gliotactin polyclonal antibody at a dilution of 1:50 and mouse anti-Coracle monoclonal antibody (9C and C615-16B cocktail) at 1:100 (Fehon et al., 1994). Stained wings were labeled using an Alexa 488-conjugated goat anti-rabbit, and goat anti-mouse Alexa 568-conjugated secondary antibodies (Molecular Probes) at 1:250 each. Wings were then mounted in Vectashield (Vector Laboratories), and visualized using a Zeiss Axioplan 2 deconvolution microscope.  78  5  Genetic interactions  discs-large: Males of the genotype {y, w, dlg" IY Dp{ 1 :Y dig*, y*}; b, Gli'"' , cnl+} were mated o2  to {w'"«; b, Gli  5  , cn/CyO}, {w'""; b, Gli' ' , cn/CyO} and {w"' ; b, Gli' ' , cn/CyO} females. Fl  RAR77  1 5  K  females of the genotypes {y, w, dlg lw'" ; b, Gli* , cnlb, Gli m52  s  5  1 1  , cn}, {y, w, dlg"' /w"'"; b, Gli' ' ,  RAR7?  52  1 5  cnlb, Gli' " , cn} and {y, w, dlg lw"' ; b, Gli' " , cnlb, Gli' "', cn} were identified as being black and 1  5  m52  s  1  5  1  cinnabar.  scribble: Males of the genotype [w" IY; b, Gli' " , cn/CyO; scrib ITWi Sb} were mated to lli  {w'"*; b, Gli  i7B3  1 5  {Gli''' /Gli ''" '"''"'; RAR77  5  , cn/CyO}, {w'"*; b, Gli' ' , cn/CyO} and {w"" ; b, Gli' ", cn/CyO} females. F l  RAR77  5  1  5  1  1  scrib' /+} progeny were identified as {b, cnlb, cn; Sb*l+}. 7H3  coracle: Females of the genotype w'" lw; b, Gli''"\ cnlb, cor were mated to w" /Y; b, cnlb, cn K  5  lfS  males and Fl {w"' IY; b, Gli' " , cn*lb, cn} males were selected by scoring the lacWinsert associated s  1  5  with Gli' " . A recombinant b, Gli' " , cn*, cor' chromosome was identified by complementation against b, 1  Gli  AE2MS  5  1  5  and b, coi chromosomes. {w'"'VY; b, Gli' ' , corVCyO} males were mated to [w'" ; b, Gli 1 5  s  cn/CyO}, {w"' ; b, Gli' ' , cn/CyO} and {w"' ; b, Gli' ", cn/CyO} females. Fl {Gli' ' , cor IGli s  1 5  s  1  1 5  ,  RAR7?  5  " "'  RAR77d  5  ''"'} progeny were identified as being homozygous for black.  Neurexin IV: Males of the genotype {w'"'7Y; b, Gli' ' , cnlCyO; Nrx 1 5  mated to {w" ; b, Gli lli  Fl {Gli' /Gli h5  RAR77,lh5  2511  L  "*7TM3  Sb} were  , cn/CyO}, {w" ; b, Gli' " , cn/CyO} and [w" ; b, Gli' "', cn/CyO} females.  RAR7?  lx  1  5  m  1  '"'''"; Nrx '"«•">/+} progeny were identified as {b, cnlb, cn; Sb*l+}. 251  vulcan: The ry* P-element vulcan allele vie" " (Gates and Thummel, 2000) was obtained from 7  22  the Bloomington stock center and crossed into a w"' background. Females of the genotype {w" /w" ; H  79  M  M  b, Git , cn\ coS/vlc" " , cn} were mated to w'"*fY; b, cnlb, cn males. Fl {w'"*/Y; b,Glt , cnlb, cn] 5  7  22  5  males were selected by scoring the lacW insert associated with Git . A recombinant F l , {b, Gli' ", cn, 5  1  vie" " , cor*} chromosome was identified by complementation against {b,Glt\ cn}, [cn, vie" " } and 7  22  7  22  {b, cor'} chromosomes using single pair matings and isogenized as a {w ; b, Gli' , cn, vlc" /CyO} inii  stock. Males from this stock were mated to {w"' ; b, Gli s  M  7022  , cn/CyO}, [w'" ; b,Glt , cn/CyO} and  RAR77  H  {w"" ; b,Gli' , cn/CyO} females. Fl {w"'*\ b,Glt , cn, vlc" " lb,Gli 1  lv5  5  7  22  5  ' '"' ",  RAR77 h5  lh  cn} progeny were  identified as being homozygous for black and cinnabar.  Epistasis analysis  Interaction between coracle and Neurexin IV: Males of the genotype {w "' /Y, b, /  s  Gli  "" \  RAR77  lv  c«/CyO; W / T M 3 Sb} were mated to {w"" / w" *; b, Git , corVCyO} females. F l {w"' ; b, 1  Git ,cor lb 0  5  ,Gli  5  s  cn; Nrx l+} progeny were identified as being Sb+ and homozygous for black.  "" \  RAR77  1  lv  46  Interaction between coracle and discs-large: Males of the genotype {y, w, dlg"' IY Dp{ 1 :Y 52  dlg\ y };b, Gli +  cn/+} were mated to { "' lw"'*; b, Git , corVCyO} females. F l females of  ,  RAR77mdv5  s  5  w  the genotype {y, w, dlg lw" ; b, Git ,cor lb, m52  ls  5  5  Gli  "" ,  RAR77  lv5  cn} were identified as homozygous for  black.  Interaction between scribble and Neurexin IV: Males of the genotype {w" '/Y, b, /s  Gli  °" ,  RAR77  lvS  cn/CyO; Nrx /TM3 Sb} were mated to {w " /w"''\ b, Git , cn/CyO; scrib /TM3 Sb} females. F l 46  {vv" '; b, Git ,cnlb, Gli  l  M  5  >i  5  l?B:1  ""' , cn; scrib' /+} progeny were identified as being Sb+ and homozygous  RAR77  v5  7Bi  for black and cinnabar.  80  Interaction between scribble and coracle:  Males of the genotype {w'" /Y, b, GU x  ''",  RAR77m  cn/CyO; scrib ITM3 Sb] were mated to {w lw" ; b, Git* , corVCyO} females. F l {w"" ; b, i?B3  IIIH  Gli' " ,cor lb, Gli 1  5  5  ',  lli  5  1  cn; scrib' "1+} progeny were identified as being Sb and homozygous for  RAR77m,h 5  7  +  black.  Interaction between scribble and Neurexin IV:  Males of the genotype {w'"' '/Y, b, Gli s  ""'"\  RAR77  cn/CyO; Nrx /TM3 Sb} were mated to {w"'*lw'"\ b, Gli' " , cn/CyO; scrib " /TM3 Sb) females. Fl 46  1  {w"' ; b, Gli' " , cn/b,Gli x  1  5  5  7  :  ""'" , cn; scrib "/Nrx } progeny were identified as being Sb* and  RAR77  5  17  46  homozygous for black and cinnabar.  Interaction between vulcan and discs-large:  dlg*,y*}; b, Gli  Males of the genotype {y, w, dlg'" /Y Dp{ 1: Y 52  ""'" , cn/+} were mated to {w" lw'" ; b, Gli' " , cn, vlc" " /CyO} females. F l females  RAR77  5  m  x  1  5  7  22  of the genotype {y, w, dlg"' /w " ; b, Gli' " , cn, vlc" " /b, Gli °"'" , cn} were identified as 52  l  s  1  5  7  22  RAR77  5  homozygous for black and cinnabar.  Interaction between vulcan and Neurexin IV:  Males of the genotype [w"" /Y; b, l  Gli  "" ,  RAR77  ho  cn/CyO; Nrx /TM3 Sb} were mated to {w"' lw" ; b, Gli' ' , cn, vlc" /CyO} females. Fl {w"' ; b, 46  x  Gli' " , cn, vlc" " /b, Gli 1  5  7  22  IH  1 5  7ll22  x  ""'" , cn; Nrx /+} progeny were identified as being Sb and homozygous for  RAR77  5  46  +  black and cinnabar.  Interaction between vulcan and coracle:  Females of the genotype {w'" /w" ; b, Gli' " , cn*, fl  ll<  1  5  cor" I vie" " , cn} were mated to w" IY; b, cnlb, cn males. F l {w" IY; b, Gli' ", cn*lb, cn} males were 7  22  lx  lx  1  selected by scoring the lacW insert associated with Gli' " . A recombinant F l , {b, Gli' " , cn, vie" " , cor } 1  5  1  5  7  22  5  chromosome was identified by complementation against {b, Gli' " , cn}, {cn, vie" " } and {b, cor'} 1  3  7  22  chromosomes using single pair matings and isogenized as a {w" ; b, Gli' " , cn, vie" " , corVCyO} stock. ls  81  1  5  7  22  Males from this stock were crossed to {w"" ; b, Gli t  vie" " , cor''lb, GU 7  22  , cn/CyO) females and Fl {w'" ; b, Git , cn,  RAR77  x  5  , cn} progeny were identified as homozygous for black and cinnabar. To generate  RAR7?  the {Git , vie" " 1 Git , cor'} genotype, {w'"'VY; b, Git , cn, vlc" " /CyO} males were crossed to 5  7  22  5  5  7  22  {w'" fw'"*; b, Git , corVCyO) females. Fl [w"' ; b, Git , cor lb, Git , cn, vie" " } progeny were s  5  s  5  5  5  7  22  identified as being homozygous for black.  Interaction between vulcan  and scribble:  Males of the genotype [w"' /Y, b, Gli s  ' '",  RAR77m 1  cnlCyO; scrib l1M3 Sb} were mated to {w"'*lw"'*\ b, Git , cn, vie" " ICyO} females. Fl {w'" ; b, i7B3  5  Git , cn, vlc" " /b, Gli 5  7  22  ,  RAR77m,k5  7  22  s  cn; scrib l+} progeny were identified as being Sb and homozygous i?B1  +  for black and cinnabar.  82  Results  A s a prelude to genetic studies of septate junction P C P , the subcellular location of Gliotactin was compared to that of Coracle in pupal wings aged to 30 hours A P F (Figure 14). A s previously determined, G l i was preferentially localized to tricellular corners (Figure 14A), while Cor was present around the circumference of the cell (Figure 14B). G l i and Cor overlap in their expression at the tricellular septate junction (Figure 14C).  To further test the hypotheses that Gliotactin is a functional component of septate junctions and that septate junctions are necessary for wing P C P , mutations in known septate junction components were tested for genetic interactions with the Gli wing P C P phenotype. A hypomorphic allele of Neurexin IV (Nrx ") dominantly enhanced all three viable Gli genotypes (Figure 15A,B; Table 3). The 23  Gli' /Gli' ; Nrx "l+ genotype was lethal. A null Nrx allele (Nrx ) enhanced the lr5  M  25  46  GW' IGli v5  RAR77  genotype to the same extent as Nrx ", indicating that the hypomorphic Nrx " allele (Baumgartner et 25  25  al., 1996) behaves as a null allele in this assay (Table 3). A null allele of discs-large (dlg"° ) dominantly 2  enhanced the wing P C P defect of the Gli' lGli lv5  RAR7?  and Gli' IGli lv5  dvl  genotypes (Figure 15C.D; Table 3),  but did not significantly affect the GW'^lGli'''' genotype (Table 3). Conversely, a null allele of coracle 5  (cor') dominantly enhanced the Gli " /Gli' d  or Gli' IGli' ly5  lrl  5  lv5  genotype, but did not significantly affect either Gli' * 1Gli 1  5  RAR7?  (Table 3). A n adult-viable, hypomorphic cor genotype (corVcor ) was also found to have 15  disrupted P C P (Figure 16). A s was found for Nrx, a strong hypomorphic allele of scrib (scrib ) 1783  dominantly enhanced all three viable Gli genotypes (Figure 15E.F; Table 3). The Gli' /Gli' ; scrib' 1+ lv5  lvl  783  genotype was lethal. Thus mutations affecting four SJ proteins act as potent enhancers of the Gli wing P C P phenotype, and homozygous cor mutants exhibit P C P phenotypes similar to Gli homozygotes.  83  Figure 14. Gli is expressed at the tricellular septate junction in the developing wing.  (A) G l i staining (green) is preferentially at tricellular corners at 30 hours A P F . (B) Cor staining (red) indicates the level of the septate junction in this epithelium. (C) A merge of (A) and (B). G l i and Cor overlap at the tricellular septate junction (yellow). Magnification: A - C - 2500X. Images are representitive of at least 10 replicate wings.  84  85  Figure 15. Septate-junction mutations dominantly enhance the Gli wing P C P phenotype.  (A) A {Gli' /Gli lv5  ;  Nrx "/+] wing with an increased PCP phenotype relative to  RAR77  25  Gli' VGli h  RAR7?  wings. (B) An example of the posterior marginal blisters often observed in [Git 1 Git' ; /Vrx '7+} wings 0  5  2,  (dotted line). (C) A [dlg"' /+; Glt IGli 52  Gli' IGli h5  RAR77  5  }  RAR77  wing displaying a severe PCP phenotype which exceeds the  phenotype.  (D) The distal margin anterior to L3 for a {dlg"' /+; Gli' ' /Gli' '} wing. This genotype has severely 52  h 5  h  disrupted PCP and a high frequency of distal blisters (dotted line). (E) The anterior margin of a {Gli' /Gli lrS  ;  RAR77  scriV l+} wing displaying a severe PCP phenotype. 7Bi  (F) The posterior-distal margin of a {Gli' /Gli' ; scrib' /+} wing with characteristic posterior and lv5  ho  7B:>  distal blisters.  Magnification: A-F = 200X.  86  87  T a b l e 3. M u t a t i o n s in septate j u n c t i o n components d o m i n a n t l y enhance the w i n g P C P phenotype of specific Gliotactin  genotypes  Septate junction genotype  Gli genotype  +/+  cll /+  0.5  0.4  2.3*  2.0  1.2  1.8  2.9  5.0*  3.4  6.1*  6.2*  5.4*  Gli IGli  dv5  3.4  4.4  5.9*  6.7*  Gli IGli  dvl  4.1  6.7*  Gli S/+ dv  Gli  d v 5  dv5  dv5  IGIi  R A R 7 7  cor /+  m52  5  S  4.5  B  Nrx "/+  Nrx /+  25  L  scrib  46  B  6.3* L  B  J /+ 7B3  6.5*" L  The average number of wing sections containing a sector with abnormal polarity is indicated for various genotypes (n = 20 wings per genotype). * indicates a significant enhancement relative to the control genotype (Mann-Whitney U-test, P<0.001). A n L indicates the genotype is lethal, thus no wings could be scored. A  B  indicates wing blisters were present in more than  10% of wings scored.  88  Figure 16. A cor hypomorph has defects in leg and wing PCP.  (A) The posterior region of a corVcor wing. Numerous hairs have altered polarity and converge in a 15  chevron pattern (arrows). (B) A metathoracic femur from a corVcor animal. Bristles and bracts have normal polarity, however, 15  leg hairs have altered polarity (arrows).  Magnification: A = 200X, B = 400X.  89  90  Three genotypes with severely disrupted wing PCP also consistently displayed a novel phenotype: {dlg"' l+; Glt IGlt'}, 52  5  {GltVGlt ; 5  Nrx "/+) and {Glt IGlt ; 25  5  contained marginal blisters (Figure 15; Table 4). The Git , cor IGlt 5  5  5  5  scrib "1+) wings 17  genotype did not exhibit  blistering, consistent with the less severe PCP phenotype associated with this genotype (Table 4). Blisters were present in 11 of 20 Git 1 Git ; Nrx " wings scored, in 7 of 20 dlg"' l+; Glt IGlt' 5  wings, and in 20 of 20 Glt IGlt ; 5  5  5  25  52  5  scrib l+ wings scored (Table 4). Blisters in Glt IGlt ; i?B3  5  Nrx "l+  5  25  wings were always restricted to the posterior margin (Figure 15B), whereas blisters in dlg"' l+; 52  Git /Git' 5  wings were restricted mainly to the distal tip of the wing (Figure 15D). Strikingly,  1  Git 1 Git ; scrib 1+ wings had blisters that spanned both the posterior and distal margins (Figure 15 5  5  783  F). Taken together, these results suggest that a disruption of septate junctions can cause a failure in adhesion between wing epithelial layers at the wing margin, but not in the central regions of the wing.  Given the result that septate-junction components dominantly enhance the Gli wing PCP phenotype, the hypothesis that vulcan participates in the Gliotactin pathway of septate junctionmediated wing PCP was tested by evaluating the dominant effect of a strong hypomorphic vulcan mutation on several Gli genotypes (Figure 17, Table 5). This allele appeared to act in a dominant manner, significantly enhancing even the Glt l+ genotype (Figure 17, Table 5). vie" " significantly 5  7  22  enhanced all hypomorphic Gli genotypes tested. The severity of the v/c-enhanced Git 1 Git' and 5  Glt IGli 5  RAR77  Glt°IGli  RAR77  5  genotypes did not differ, despite the variation between the baseline Git 1 Git' and 5  5  phenotypes. Interestingly, the severe and consistent disruption of wing PCP in the {Git , 5  vie" " 1 Git'} genotype did not result in numerous wing blisters (Figure 17, Table 5). These results 7  22  indicate that vulcan is a component of the Gliotactin PCP pathway.  91  Table 4. Mutations in septate-junction components dominantly induce wing blisters in specific Gliotactin genotypes  Septate junction genotype  Gli genotype  dlg /+ m52  cor /+ 5  Nrx /+  GU IGli  0(25)  0(20)  0.05 (22)  Gli /Gli  0(25)  0(25)  0.55(20)  0.32(25)  0(25)  dv5  dv5  RAR77  dvS  Gli IGli dv5  dv!  Nrx /+  25ll  scrib  46  0(20)  0.15(20)  I  L  J I+ 7B3  0.05 (20)  1.0(20)  L  The frequency of blistered wings is indicated for each genotype. Genotypes exhibiting significant blistering are shown in bold type. A n L indicates that the genotype is lethal, thus no wings could be scored. The n value for each genotype is given in brackets.  92  Figure 17. vie is a dominant enhancer of Gli.  For all images anterior is up and distal is to the right.  (A) The anterior-distal margin of a wing doubly heterozygous for Gli' ' and vie" " . This genotype 1 5  7  22  displayed consistent disruption of wing hair polarity. (B) The effect of vie" " on a strong hypomorphic Gli genotype, vie" " strongly enhanced this 7  22  7  genotype but did not induce wing blisters.  Magnification: A-B = 2 0 0 X .  93  22  94  Table 5. vulcan is a dominant enhancer of the Gliotactin P C P phenotype  vulcan genotype"  Gli genotype  +/+  Gli /+ dv5  vie  07022  /+  0.5  4.7*  2.9  5.8*  Gli /Gli  3.4  5.5*  Gli /Gli  4.1  7.0*  dv5 RAR77  Gli  /ai  dv5  dv5  dv5  dvl  " The average number of wing regions with a section of altered polarity is indicated (n=20). * indicates statistical significance relative to the control genotype (Mann-Whitney U test, P<0.0l).  95  In order to place Gliotactin pathway members into epistasis groups, crosses to examine the simultaneous effect of two G//-enhancing mutations on hypomorphic Gliotactin genotypes were performed. Epistasis is defined as the ability of a mutation at one locus to mask the effects of a mutation at a second locus (Griffiths et al., 2003). There are three possible outcomes for simultaneous doubleenhancement of Gli genotypes by two SJ mutations. The effect of both mutations could exceed the phenotype of the strongest single enhancement, indicating that the two SJ mutations act in an additive manner, and thus act as enhancers of the other's phenotype. Alternatively, the effect of both mutations (  could be less severe than the phenotype of the weakest single enhancement, indicating that one of the two mutations acts to suppress the other's phenotype. Lastly, the effect of both mutations could match the phenotype of the strongest single enhancement, indicating that the ability of one of the mutations to enhance the Gli phenotype is masked by the other locus. This result would place these two SJ mutations into an epistasis group. Thus by definition, simultaneous enhancement of Gli by two septate-junction mutations in the same epistasis group would not enhance the Gli wing PCP phenotype beyond the enhancement seen for either of the two mutations singly. In contrast, simultaneous enhancement of Gliotactin by SJ mutations in separate epistasis groups would enhance the Gli wing PCP phenotype in an additive manner, exceeding the enhancement of either of the two mutations alone.  Neurexin IV and coracle: In the salivary gland epithelium of homozygous cor mutant embryos 1  Nrx is not localized to the septate junction, indicating that the cor" allele does not produce a Coracle protein capable of localizing Neurexin IV (Ward et al., 1998). The Neurexin IV allele Nrx is a null 46  allele (Baumgartner et al., 1996); homozygous Nrx embryos fail to localize Cor at the septate junction 46  (Ward et al., 1998). Biochemical studies have shown that Cor and Nrx physically interact (Ward et al., 1998). Since these proteins physically bind each other and each are required for the localization of the other to the septate junction, mutations in these genes would be predicted to form an epistasis group. To test this hypothesis, the wing PCP phenotype of {Gli' ' , cor IGli 1 5  96  5  ; Nrx l+] adults was scored and  RAR77  46  compared to the phenotype of the {Git , cor IGli 5  5  \ +/+} and {Glt'/Gli  RAR77  RAR77  The phenotype that resulted from cor' and Nrx together in a Glt' IGli 46  5  ; Nrx /+} genotypes. 46  background did not exceed  RAR77  the effect of Nrx"' alone (Table 6). The equivalent experiment in the more severe Glt'/Glt  5  genotype  was performed to confirm this result (Figure 18; Table 6). Similarly, the double mutant combination did not exceed the enhancement of either single mutant. Mutations in cor and Nrx are thus in the same epistasis group in the wing PCP pathway, as predicted by studies in the embryonic epidermis.  coracle  Discs-large is not mislocalized in homozygous Nrx ' or cor' mutant 4(  and discs-large:  embryos, but remains at the septate junction; therefore the localization of Dig is independent of Nrx and Cor (Ward et al., 1998). Mutations in dig and cor would thus be expected to fall into separate epistasis groups. To test this hypothesis, the phenotype of dlg"' l+; Gli' '' , cor /Gli 52  that of dlg" l+; Gli''" IGli ,52  5  RAR77  and Git , cor IGli 5  5  RAR77  1  5  5  RAR7?  wings was compared to  wings (Table 7). The dig'" allele is a truncation 52  allele that is not localized to the septate junction (Woods et al., 1996). Despite the fact that cor' had no dominant effect on the Glt IGli 5  genotype (Table 3), the combination of dig'" and cor' in this  RAR77  52  genotype dramatically enhanced its wing PCP phenotype compared to the effect of dig'" alone (Table 52  7). The phenotype of dlg"' l+; Git , cor'IGli 52  5  RAR7?  wings included wing blisters, which were not observed  in either single mutant enhancement (Table 7). To confirm this result the equivalent crosses were performed for the Git 1 Git' genotype (Figure 19; Table 7). Although dig'" does not significantly 5  enhance the Glt IGit' 5  5  5  52  genotype, the combined effect of dig'" and cor in this genotype showed strong 52  5  enhancement relative to Git , cor /Glt' wings, including wing blisters (Figure 19; Table 7). Taken 5  5  5  together, these results demonstrate that cor and dig mutations fall into separate epistasis groups in determining wing PCP and adhesion.  coraclelNeurexin  Wand scribble:  Genetic interactions and/or physical associations between  scriblScx'xb and either cor/Cor or /Vrx/Nrx have not been previously investigated. To determine the  97  Table 6. Interaction between and coracle and Neurexin IV in G/i'-mediated wing P C P  Septate junction genotype'  corr /+; 5  Gli genotype  cor  Nrx l+ 46  Nrx  Gli<lv5, RAR77 cli  Gli /Gli dv5  dv5  1+  46  3.4  6.2  6.2  5.9  6.3"  6.0  " The average number of wing regions with a section of altered polarity is indicated (n=20). " indicates wing blisters were present in greater than 10% of wings scored. * indicates statistical significance relative to both singlemutant enhancements (Mann-Whitney U test, P < 0.01).  98  Figure 18. cor does not enhance the phenotype of Afar-enhanced Gli wings.  For all images anterior is up and distal is to the right.  (A) The distal-anterior wing margin of a Gli' ' homozygote dominantly enhanced by the null allele h 5  Nrx . Numerous wing hairs deviate from wild-type proximal-distal orientation. 46  (B) The effect of adding a single copy of cor to the genotype in (A). The addition of cor did not 5  1  significantly alter the phenotype relative to the effect of Nrx alone. 46  Magnification: A-B = 200X.  99  100  Table 7. Interaction between discs-large and coracle in G//-mediated wing PCP  Septate junction genotype'  Gli genotype  dig  /+  m52  cor,.5/  dlg /+; m52  +  corr /+ 5  Glidv5/ciiMK77  5.0  3.4  6.7*"  Gli /Gli  AA  5.9  7.0*"  dv5  dv5  " The average number of wing regions with a section of altered polarity is indicated (n=20). " indicates wing blisters were present in greater than 10% of wings scored. * indicates statistical significance relative to both singlemutant enhancements (Mann-Whitney U test,  101  P  <  0.01).  Figure 19. dig dominantly enhances the phenotype of cor-enhanced Gli wings.  For all images anterior is up and distal is to the right. A dotted line indicates the edge of a wing blister.  (A) The effect of cor on the homozygous Gli' '' genotype. 5  1  5  (B-C) The effect of adding a single copy of dig'" to the genotype in (A). The combined effect of cor' 52  and dig'" greatly increased disruption of planar cell polarity compared to cor' alone and induced 32  blisters in the posterior margin (C). Magnification: A-C = 200X.  102  103  epistatic relationship between cor and scrib, the severity of {Gli''", cor'IGli  ; scrib' "1+] wings was  RAR77  compared to {Gli' " , corVGli 1  5  ; +/+} and {Gli''" IGli  RAR77  5  has no dominant effect on the Gli' IGli lv5  {Gli' IGli lv5  7  \ scrib' 1+) wings (Table 8). The cor' allele  RAR77  783  genotype (Table 3) but was able to significantly enhance the  RAR7?  ; scrib' /+} phenotype. The corresponding cross for the Gli''" /Gli''" genotype was not  RAR77  783  5  5  done, since wing PCP is severely disrupted in {Gli' " 1 Gli' " ; scrib 1+) animals and includes complete 1  5  1  5  1783  penetrance of the blistering phenotype (Tables 4, 5). To confirm the result that scrib and cor fall into separate epistasis groups the equivalent experiment with scrib and Nrx was performed (Figure 20; Table 9). Nrx and scrib' 46  783  both strongly enhance Gli IGli k5  RAR77  (Table 3), yet the double mutant was  significantly enhanced relative to both single mutant genotypes (Table 9), even inducing wing blisters (Figure 20). Taken together, these results place scrib outside the corlNrx epistasis group.  scribble and discs-large: Scribble is necessary to localize Dig and Lethal-giant-larvae (Lgl) at the septate junction, and there is a strong genetic interaction between scrib, lgl and dig mutations (Bilder et al., 2000). The current model of Dlg/LgL/Scrib function is that Dig is the first component targeted to the septate junction, which stabilizes Scrib; Scrib then recruits Lgl from the cytoplasm to the membrane (Bilder et al., 2000). To test if dig and scrib mutations fall into the same epistasis group, dig'" and 52  scrib'  783  were crossed into the Gli''" IGli 5  RAR77  background (Figure 21; Table 10). This double mutant  combination displayed a significantly stronger wing PCP phenotype than either mutant alone (Figure 21). This phenotype included a high frequency of marginal wing blisters which was not observed in either single mutant enhancement for this Gli genotype (Table 10). Additionally, blisters in {dlg"' /+; 52  Gli' IGli lv5  ;  RAR77  scrib /+] wings were observed on the anterior margin (Figure 2IB), posterior margin i783  (Figure 21 C), and distal margin (not shown). Rare wing hairs in blistered regions had reversed polarity and angled from distal to proximal (Figure 21C). This severe enhancement indicates that scribble and discs-large function in distinct epistasis groups to establish wing PCP. As before, these findings were not confirmed using the Gli' " 1 Gli''" background due to the severe {Gli' " 1 Gli' " ; scrib' /+} phenotype. 1  5  5  1  104  5  1  5  7B3  Table 8. Interaction between and coracle and scribble in G//-mediated wing P C P  Septate junction genotype"  cor Gli genotype  cor- /+ >  5/+;  scribJ'"-'/+ scrib J  Gli  d v 5  IGIi  R A R 7 7  3.4  5.4  7 B 3  6.1*  "The average number of wing regions with a section of altered polarity is indicated (n=20). * indicates statistical significance relative to both singlemutant enhancements (Mann-Whitney U test, P < 0.01).  105  h  Table 9. Interaction between Neurexin IV and scribble in G/i-mediated wing PCP  Septate junction genotype"  Nrx Gli genotype  Nrx /+ 46  scrib  /+;  46  J /+ 7B3  scribJ l+ 7B3  Gli  d v 5  IGli  R A R 7 7  6.2  5.4  6.8*  B  "The average number of wing regions with a section of altered polarity is indicated (n=20). B  indicates wing blisters were present in greater than 10%  of wings scored. * indicates statistical significance relative to both singlemutant enhancements (Mann-Whitney U test, P < 0.01).  106  Figure 20. C o m b i n a t o r i a l mutant analysis between scrib and cor/Nrx.  For all images anterior is up and distal is to the right. A dotted line indicates the edge of a wing blister.  (A) The anterior-distal margin of a heterozygous Gli' /Gli h0  RAR77  wing enhanced by a single copy of  scrib' . Numerous wing hairs deviate from a proximal-distal orientation. 783  (B-C) The dominant effect of Nrx on A*m'/>enhanced Gli wings included severe disruption of wing 46  hair polarity (B) and blistering of wing epithelia (C). Similar results were obtained for a null allele of coracle (not shown).  Magnification: A-C = 200X.  107  108  Figure 2 1 . Combinatorial mutant analysis between  scrib and dig.  For all images anterior is up and distal is to the right. A dotted line indicates the edge of a wing blister.  (A-C) Compare to Figure 14 A. dig enhancement of scrib-enhanced Gli wings included severe disruption of wing hair polarity (A) and blisters on the anterior (B) and posterior (C) margins, as well as on the distal margin (not shown). Rare wing hairs on blistered epithelia were angled from distal to proximal in this genotype (arrow, inset in C).  Magnification: A-C = 200X.  109  110  T a b l e 10.  Interaction between discs-large  scribble  and  in G / i - m e d i a t e d w i n g P C P  Septate junction genotype"  Gli genotype  dlg /+  scrib  m52  J /+ 7B3  scrib  Gli  d v 5  IGIi  R A R 7 7  5.0  5.4  J /+ 7B3  7.0*"  " The average number of wing regions with a section of altered polarity is indicated (n=20). " indicates wing blisters were present in greater than 10% of wings scored. * indicates statistical significance relative to both singlemutant enhancements (Mann-Whitney U test. P < 0.01).  Ul  Epistatic relationships between  vulcan and septate-junction mutations  Given that vie enhances Gli wing PCP phenotypes and that Vic has been suggested to be physically associated with Dig, epistasis analysis was extended to place vie in the Gli epistatic framework. To this end, double-mutant phenotypes for vie and various septate-junction mutations in hypomorphic Gli backgrounds were analyzed.  vulcan and discs-large: Given the speculation that Vic physically interacts with Dig, the primary experiment was to determine if vie and dig belonged to the same epistasis group. To determine their epistatic relationship, mutations in both genes (dig'" and vie" " ) were placed into the Gli' 'IGli 52  7  background and compared to the {dlg"' /+; Gli''" IGli 52  5  22  } and {Gli' " , vlc" " /Gli  RAR77  1  (Figure 22; Table 11). Interestingly, {dlg"' l+; Gli' " , vlc" " IGli 52  h:  1  5  7  22  5  7  22  RAR77  } genotypes  RAR77  } wings showed significant  RAR77  enhancement relative to both singly-enhanced genotypes. Despite a severe PCP phenotype, this genotype did not display wing blistering. To confirm this result, crosses to place dig" and vie" " into 02  the Gli' " /Gli' " genotype were performed (Table 11). Consistent with the Gli' IGli 1  5  1  5  h5  RAR77  7  22  results,  significant enhancement was observed in the double mutant. This genotype also did not exhibit blistering. Taken together, these results place vie and dig into separate epistasis groups in the Gli PCP pathway.  vulcan and coracle/Neurexin IV: Since vie did not fall into the dig epistasis group, additional crosses were performed to determine if vie grouped with other septate-junction mutations. Double mutant analysis placing cor and vie" " into the Gli''" IGli 5  7  22  enhancement relative to the [Gli' " , vlc" /Gli  5  1  5  7022  }  RAR77  background failed to show significant  RAR77  and [Gli' " , cor 1Gli 1  5  5  } genotypes (Table 12).  RAR77  This result was confirmed with equivalent crosses using the Gli' " /Gli' * genotype (Table 12). Since 1  5  1  5  Neurexin IV and coracle form an epistasis group, analogous crosses to place Nrx and vie" " into the 46  112  7  22  F i g u r e 22. C o m b i n a t o r i a l mutant analysis between vie and Nrx.  Anterior is up and distal is to the right for all images.  (A) The effect of vie" " on the Gli' /Gli 7  22  lv3  RAR7?  genotype.  (B) The simultaneous effect of vie and discs-large (dig) on the genotype in (A). The addition of dig enhanced the disruption of hair polarity but did not induce blistering. (C) The combined effect of vie and Nrx on the genotype in (A). The addition of Nrx to this genotype 16  did not significantly enhance the polarity phenotype. Similar results were obtained for cor (not shown).  Magnification: A-C = 200X.  113  114  Table 11. Interaction between and vulcan and discs-large in G/i-mediated wing PCP  Septate junction genotype"  dig Gli genotype  dig  m52,  +  07022  vlc  /+ v  Gli  d v 5  Gli IGli dv5  IGli  dv5  R A R 7 7  k  /+;  m52  07022  j+  5.0  5.8  6.9*  4.4  5.5  6.9*  " The average number of wing regions with a section of altered polarity is indicated (n=20). * indicates statistical significance relative lo both singlemutant enhancements (Mann-Whitney U test, P < 0.01).  115  Table 12. Interaction between and vulcan and coracle in G/i-mediated wing P C P  Seplate junction genotype"  07022 .  v]c  Gli genotype  l+  v , 07022,+ c  S,  cor  +  cor $/+  Gli  d v 5  Gli /Gli dv5  /Gli  dv5  R A R 7 7  5.8  3.4  6.1  5.5  5.9  6.0  " The average number of wing regions with a section of altered polarity is indicated (n=20). * indicates statistical significance relative to both singlemutant enhancements (Mann-Whitney U test. P < 0.01).  116  GW'" Gli 5  and Gli' ' Gli''" genotypes were performed (Figure 22; Table 13). Consistent with the  RAR77  h 5  5  results for coracle, significant enhancement between Nrx and vie was not observed in either Gli background. Taken together, these results place vulcan into the coraclelNeurexin IV epistasis group.  vulcan and scribble: Finally, the epistatic relationship between vie and scrib was evaluated by placing vie" " and scrib' " into the Gli' Gli 7  22  7  phenotype of {Gli' ,vlc" " lGli  lv5  lv5  {Gli' ,vlc" " IGli hs  7  22  7  22  ;  RAR77  RAR7?  genotype (Table 14). Interestingly, the wing PCP  scrib "/+] animals did not significantly exceed that of i7  } wings. Greater than 10% blistering was observed in this genotype, however, this  RAR77  frequency was not statistically significant (X , P>0.05) compared to the [Gli''" /Gli 2  5  ; scrib' "/+}  RAR77  7  genotype, which has a low frequency of wing blisters. This result suggested that scrib and vie belong to the same epistasis group, an observation which stands in contrast to the previous finding that scrib and corlNrx are separate groups, and that vie belongs to the corlNrx IV group. To resolve the issue, vie" " 7  22  and scrib' " were placed into the Gli' Gli' ' genotype. Interestingly, vie" " significantly suppressed the 7  lvS  h 5  7  22  {Gli' /Gli' ; scriV "l+\ blistering phenotype (Figure 23; Table 14), indicating that vie is epistatic to lv5  lv5  7  scrib in this Gli genotype. This result places vie and scrib into separate epistasis groups, and suggests that vie is a unique member of the corlNrx group in terms of its interaction with scrib.  117  Table 13. Interaction between vulcan and Neurexin IV in G/i-mediated wing PCP  Septate junction genotype'  Gli genotype  ,  lc  07022,•+  n r x  46  •Ic 07022,+. l +  46,  x  GlidvS, RAR77  5.8  6.2  6.7  Gli  5.5  6.3"  6.3  GU  d v 5  IGli  d v S  +  " The average number of wing regions with a section of altered polarity is indicated (n=20). B  indicates wing blisters were present in greater than 10%  of wings scored. * indicates statistical significance relative to both singlemutant enhancements (Mann-Whitney U test, P < 0.01).  118  Table 14. Interaction between vulcan and scribble in G/i'-mediated wing PCP  Septate junction genotype"  07022, .  vlc  Gli genotype  scrib J™3,  +  v  k  07022  1+  +  scribJ /+ 7B3  Gli  /Gli  d v 5  R A R 7 7  Gli IGIi dv5  dvS  5.4  6.5  B  5.8  6.1  5.5  6.7*''  11  " The average number of wing regions with a section of altered polarity is indicated (n=20). '' This genotype was significantly suppressed relative to the scrib single enhancement due to a reduction in blisters, " indicates wing blisters were present in greater than 10% of wings scored. * indicates statistical significance relative to both singlemutant enhancements (Mann-Whitney U test, P < 0.01).  119  Figure 23. Combinatorial mutant analysis between vie and scrib.  Anterior is up and distal is to the right for all images. A dotted line indicates the edge of a wing blister.  (A) The effect of vie" " on the homozygous Gli' " genotype. Blisters were not observed in this 7  27  1  5  genotype. (B) The effect of scrib on the homozygous Gli' ' genotype includes complete penetrance of the h 5  blistering phenotype and large posterior blisters. (C) The combined effect of vie and scrib on Gli' homozygotes significantly reduced both the size and lv5  frequency of wing blisters relative to Gli IGli' ' \scrib' l+ wings. dv5  h 5  7Bi  Magnification: A-C = 200X.  120  121  Discussion  Consistent with the hypothesis that the wing PCP phenotypes associated with hypomorphic Gliotactin genotypes is a result of altered function at the septate junction, the Gli PCP phenotype was dominantly enhanced by mutations in four SJ proteins: Neurexin IV, Coracle, Scribble, and Discs-large. Similar to cor mutants, viable dig genotypes have an eye PCP phenotype (Woods et al., 1996). The distribution of Discs-large within pupal wing epithelia has been investigated (Eaton et al., 1996), but wing PCP in pupal and adult dig tissues has not been studied. The dig" allele is a genetic null allele 02  resulting from a splice donor site mutation that leads to a stop codon prior to the third PDZ motif (Woods et al., 1996); additionally, the truncated protein produced by this allele fails to localize to the septate junction (Woods and Bryant, 1991). Interestingly, the dominant effect of the dlg"° allele on Gli 2  genotypes was selective to heteroallelic combinations where one Gli allele was truncated prior to the PDZ-binding epitope. Both the Glt IGli 5  RAR77  and Git 1 Git' genotypes were strongly enhanced by 5  dig" , despite the large difference in severity between the two Gli genotypes. The Glt IGlt° 02  5  genotype,  in which both alleles have a PDZ-binding epitope, was resilient to enhancement by dig" despite its 02  severity being intermediate relative to Glt IGli 5  RAR7?  and Glt IGlt'. 5  Enhancement by dig thus did not  correlate with the severity of the Gli genotype but with specific alterations of Gli structure. These results raise the possibility that Gli and Dig physically interact at septate junctions in wing epithelia via the Gliotactin PDZ-binding epitope and one of the PDZ domains of Discs-large. Another possibility is that Gli is recruited to the SJ through physical association with a PDZ domain of Scrib, however, there was no allele specificity observed in the Gli - scrib genetic interaction. In the absence of biochemical assays, and given the genetic interactions demonstrated for dig and scrib, it is uncertain if Gli associates with Dig, Scrib, or an unidentified PDZ protein at SJs. Recent work has shown that immunoprecipitation of an over-expressed Gli transgene also precipitates Dig (Schulte, 2003), however, this interaction was not demonstrated to be direct or to require the Gli PDZ-binding motif. Additionally, the presence of Scrib in  122  the immunoprecipitate was not tested (Schulte, 2003). While this result demonstrates that Gli participates in a complex at septate junctions, it does not identify specific binding partners of Gli. As per the Gli - scrib interaction, the genetic interaction between Gli and Nrx was strong in all genotypes tested. A physical interaction between Gliotactin and Neurexin IV would likely be through the extracellular domains of both proteins. Nrx is expressed in the developing wing with a pattern coincident with Cor (Ward et al., 1998), and adult Nrx animals have unspecified defects in eyes and wings (Baumgartner et al., 1996). In vertebrates, a class of serine esterase-like molecules called Neuroligins have been shown to bind (3-neurexins (Ichtchenko et al., 1996) and the Dig homologue PSD-95 (Irie et al., 1997). Neuroligin-3 is the vertebrate homologue of Gliotactin (Gilbert et al., 2001). (3-neurexins have not been identified in Drosophila, leading to the suggestion that Neurexin IV is a Gli receptor (Tepass et al., 2001).  If we accept the conjecture that Gli is in fact a ligand for Nrx, how might the observed genetic interactions between Gli and Nrx/Cor mutations be explained? The Gli' allele is a constant in all viable hS  Gli genotypes identified to date, and it represents a point mutation in the extracellular, serine esteraselike domain. Thus the observed Gli: Nrx interaction may reflect reduced binding between Nrx and a Gli multimer containing Gli' ' - derived monomers. A physical interaction between Nrx and Cor has been h 5  demonstrated: both are necessary for localizing the other to septate junctions (Ward et al., 1998). It might be predicted that a genetic interaction between Gli and cor would mirror the Gli: Nrx interaction, but a null cor allele enhanced only the intermediate Gli ' IGli' ' genotype. This genotype would posses lh 5  h 5  only monomers with an altered extracellular domain and thus would be predicted to interact strongly with Nrx, as was observed. In a wild-type Nrx background, cor did not enhance the Glf/Gl^"  or  77  Gli' IGli' h6  ht  genotypes, suggesting that reducing the dosage of Cor (and thus presumably Nrx) at septate  junctions could not destabilize a Gli:Nrx interaction that possesed a wild-type serine esterase-like domain, even if supplied by the truncated Git or Gli 1  RAR77  123  alleles. The resilience of the Gli' ' /Gli' " 1 6  1  genotype to enhancement by coracle suggests that the remaining serine esterase-like motif in the Gli'  M  monomer is sufficient to interact with Neurexin IV. This model will remain speculative until investigated at the biochemical level.  The involvement of septate junction components in wing PCP is not altogether surprising: Coracle and Discs-large have been implicated in PCP in the Drosophila eye (Lamb et al., 1998; Woods et al., 1996), and Discs-large, Coracle and Neurexin IV are highly expressed in the developing wing (Eaton et al., 1996; Ward et al., 1998; Woods et al., 1996). In contrast to previously identified septate junction components, Gliotactin is preferentially localized to the tricellular septate junction, or tricellular plug (Fristrom, 1982) at the junction of three epidermal cells. Thus Gli has apical/basolateral polarity and specific localization within the septate junction plane. Interestingly, the location of prehair initiation in the developing wing is directly adjacent to a tricellular junction (Wong and Adler, 1993). Thus Gliotatctin may be a key determinant in bridging PCP and apical/basolateral polarity by virtue of its localization within both apical/basal and planar axes. The specific localization of Gliotactin, taken together with its genetic interaction with general septate-junction components which are not restricted to tricellular junctions, may be responsible for the PCP phenotype of septate-junction mutants. Alternately, the SJ as a whole may contribute to the planar polarity of epithelial cells. It has long been postulated that a system independent from the frizzled pathway exists for PCP determination (Wong and Adler, 1993) based on the observation that cells of mutant polarity tend to resemble the polarity of neighboring cells. This system remains intact in all known tissue polarity mutants and is proposed to involve cell-cell interconnection via the actin cytoskeleton (Wong and Adler, 1993). Indeed, there are apparent lateral cell-cell cytoskeletal contacts in the developing wing (Turner and Adler, 1998) in addition to basal-basal contacts. Cells of mutant polarity arising from septate junction defects also are present in patches where neighboring cells have a polarity phenotype, except that adjacent cells often have a reciprocal phenotype. Recent work has linked the actin (Winter et al., 2001) and microtubule (Hannus et al., 2002)  124  cytoskeletons to PCP. The septate-junction PCP phenotype suggests that septate junctions mediate a component of this cell-cell signal, perhaps through the action of Lgl, Scrib and Dig, since Lgl has been demonstrated to be part of the cytoskeleton (Strand et al., 1994). The defects in hair morphology associated with Gli mutations are similar to those induced by expression of a dominant-negative Cdc42 construct in the wing (Eaton et al., 1996) and the application of cytoskeletal-disrupting drugs (Turner and Adler, 1998). Thus it is possible that Gli mutants have defects in actin or tubulin localization or polymerization. Both F-actin and tubulin are present at the septate junction (Tepass et al., 2001). The intriguing possibility that the septate junction contributes to/h'zzW-independent PCP remains to be tested by genetic and cell-biological assays, however, it is noteworthy that the microtubule-associated PCP determinant Widerborst acts in a/h'zz/ed-independent manner (Hannus et al., 2002).  The finding that combinatorial mutations in septate junction proteins disrupt wing adhesion to cause blisters was unexpected, although in retrospect coracle mutant embryos display a cuticular delamination phenotype (Lamb et al., 1998). Adhesion between wing epithelial layers is accomplished by the transalar array (Tucker et al., 1986), a cytoskeletal system that originates apically and terminates at the basal membrane (Fristrom et al., 1993), connecting the wing epithelial layers through basal-basal contact. If a connection exists between SJ proteins and the cytoskeleton, the observed adhesion defects in combinatorial septate-junction mutants may result from mislocalization of apical or basal transalar array components.  Two large screens for mutations affecting wing adhesion have not identified septate junction components, despite both screens relying on FRT-based somatic wing clones to bypass earlier vital requirements of potential adhesion genes (Prout et al., 1997; Walsh and Brown, 1998). The combination of both screens was estimated to have saturated the left arm of chromosome 2 (2L) which contains the Gli locus (Walsh and Brown, 1998), however, both screens failed to identify wing blister, which is also  125  on 2L and maps close to GU (Martin et al., 1999). The failure to identify any SJ component despite their varied map positions (dig maps to the X; Gli and cor map to 2L and 2R, respectively; Nrx maps to 3L and Scrib maps to 3R) suggests that a mutation in a single septate-junction component will not cause a blister in a somatic wing clone, or that such blisters are too small to be easily identified in a screen. Indeed, blisters present on {GW'^IGW'"; scrib/+} wings, which were always present and were the largest blisters observed, could not be detected under a dissecting microscope.  The specific location of blisters observed in combinatorial SJ mutants is of special interest. Large myospheroid clones (lacking PS(3 integrin) on the wing margin do not produce blisters, indicating that adhesion in wing perimarginal domain is integrin-independent (Brower and Jaffe, 1989). Large mys clones failed to blister at the anterior, posterior, and distal wing margins, even if the mys clone overlapped the margin, producing dorsal and ventral perimarginal cells lacking functional integrin directly apposed to each other (Zusman et al., 1990). While mutations have been described in which entire wing surfaces are separated (D'Avino and Thummel, 2000), the combinatorial SJ genotypes constructed in this thesis are the first described to blister specifically at the wing margin. The absence of central blisters in SJ mutants implies that these wing regions adhere in a SJ-independent manner, however, in contrast to the integrin work, the null phenotype of any given SJ mutant in the wing has not been addressed. Clonal analysis of SJ mutants in the wing will be required to bypass the essential embryonic functions of these genes and determine the wing SJ-null phenotype. Such studies may be hampered by the essential nature of the septate junction; cells that lose SJ function during larval stages (which is the timing required to produce a large mitotic clone) may not survive to be observed in the pupal or adult wing.  The results obtained from epistasis analysis of SJ mutations in wing PCP and adhesion are in agreement with and extend previous studies of SJ structure and function in the embryonic epidermis  126  (Figure 24). The use of viable Gli genotypes as sensitized genetic backgrounds allowed all currently identified SJ mutations to be placed into epistasis groups. The dominant effect of SJ mutations on the Gliotactin wing PCP phenotype was used to circumvent their recessive-lethal nature. This analysis also placed the proposed SJ-component Vulcan into the SJ epistasis model.  The dominant effect of a Neurexin IV mutation on Gli wing PCP could not be further exacerbated by a coracle allele, indicating that Nrx is epistatic to cor. in this assay system (Figure 24). This result is consistent with the interdependence between Cor and Nrx proteins for localization to septate junctions in the embryonic epidermis (Ward et al., 1998). A reduction in the amount of Nrx at the septate junction would also presumably reduce the amount of available Coracle, however, null mutations in both genes did not enhance the Gli wing PCP phenotype to the same degree or with the same allele-specificity. A null allele of cor enhanced Gli only when both Gli alleles carried an extracellular point mutation, whereas Nrx enhanced all Gli allele combinations. This suggests that the stability of Nrx at the wing septate junction is dependent not only on the presence of Cor, but also on the extracellular domain of Gli. These results are consistent with the suggestion that Nrx and Gli physically interact (Figure 24).  The finding that Nrx and cor mutations are epistatically separate from discs-large is also in agreement with embryonic studies. Dig is not mislocalized in the absence of Nrx or Cor in the embryonic epidermis (Lamb et al., 1998). Similarly, a null allele of dig retains its ability to enhance Gli genotypes already enhanced by Nrx or cor, indicating that Dig function remains these genotypes. Thus Dig and the Nrx/Cor complex function separately in the adult wing as well as in the embryonic epidermis (Figure 24). The agreement between these results and previous work serves as an important positive and negative control for the wing PCP assay system.  127  Figure 24. Genetic interactions between SJ components in wing PCP.  Arrows indicate genetic interactions at the wing epithelial septate junction that were demonstrated in this thesis. As in the embryonic epidermis, Scrib and Dig interact genetically, and Cor/Nrx form an epistasis group. The allele specificity of the Gli - Dig interaction suggests that Gli may bind Dig and restrict its localization to the septate junction. A Gli - Gli interaction was demonstrated through interallelic complementation. The interaction between Nrx and Gli is likely mediated through the extracellular domains of each protein. This interaction may occur between Nrx and Gli molecules in the same membrane, or between Gli and Nrx proteins on opposite membranes. For simplicity, the first possibility is illustrated. In the wing epithelium, Vic acts within the Cor/Nrx group to suppress Scrib.  128  129  Previous work has identified a strong genetic interaction between scribble and discs-large mutations in embryos. Homozygous scrib mutants fail to localize Dig to the septate junction (Bilder et al., 2000), an interaction similar to that observed for cor and Nrx. Unlike the epistasis between cor and Nrx mutations, considerable Dig function was retained in .sxn'6-enhanced Gli wings, placing scrib and dig into separate epistasis groups (Figure 24). This functional separation between Scrib and Dig may reflect differences in target molecules for these putative scaffolding proteins. From these results it seems unlikely that the Cor/Nrx complex is localized to the septate junction through a direct interaction with either Scrib or Dig.  The finding that vulcan acts in the Gli wing-PCP pathway is consistent with the grossmorphological wing defects observed in vie homozygotes (Gates and Thummel, 2000). Enhancement by vie of even severe Gli genotypes did not produce wing blisters, even though vie enhanced the Gli PCP phenotype as strongly and consistently as dig and Nrx. Thus there is not a linear correlation between disruption of wing PCP and the failure of adhesion between wing epithelial layers. This is consistent with the finding that scrib produced significantly more blisters than did Nrx in a given Gli genotype (Table 4), despite the equivalent enhancement of the Gli PCP phenotype by these two mutations (Table 3). Thus mutations that affect PCP have differing impacts on adhesion, suggesting that separate pathways emanate from the septate junction to impact both PCP and adhesion between wing layers.  Gli genotypes enhanced by dlg-were sensitive to further enhancement by vie, placing vie and dig into separate epistasis groups. Thus the effect of dig on Gli does not include a simultaneous equivalent effect on vie. Interestingly, double-dominant analysis between vie and septate-junction mutations placed vulcan into the corlNrx epistasis group and downstream of scrib with respect to blister formation. Lossof-function alleles of cor and Nrx significantly enhanced scrib, but a loss-of-function vie allele was a suppressor of scrib blister formation. Thus vulcan is a distinct member of the corlNrx epistasis group.  130  The finding that vie suppressed scrib blister formation is consistent with the absence of blisters in vlcenhancement of severe Gliotactin genotypes. Vulcan was postulated to be a septate-junction protein based on its possible association with Dig (Gates and Thummel, 2000). The vie allele used in this study has also been shown to interact with zipper, the Drosophila nonmuscle myosin heavy chain (Gates and Thummel, 2000), which has recently been implicated in/nzz/ed-mediated PCP downstream of Rhoassociated kinase in Drosophila (Winter et al., 2001). Zipper physically associates with Lethal-giantlarvae (Strand et al., 1994), which in turn genetically interacts with scribble in the embryonic epidermis (Bilder et al., 2000). Epistatic analysis placed vie with cor/Nrx but apart from dig and downstream of scrib. Thus it is possible that Vulcan acts as a link between the Coracle/Neurexin IV complex and the Scribble-cytoskeleton group. This function of Vulcan appears to be independent of Dig, suggesting that they do not physically interact in an interdependent manner analogous to Cor and Nrx. A model incorporating interactions between septate junction components in both the embryonic epidermis and the wing epithelium is given in Figure 25.  In summary, these results suggest two possibilities: (1) that septate junctions are an intermediary step between the Frizzled/Dishevelled signal and the actin cytoskeleton in determining prehair orientation, or that (2) Frizzled signaling and SJ function modify wing PCP by independent and parallel pathways. These possibilities could be resolved by (1) double-mutant analysis between SJ and frizzled pathway components, or by (2) investigating the dependence of SJ components on the frizzled pathway (or, conversely, the dependence of the frizzled pathway on SJ components). It is the aim of the next chapter to resolve this question using both genetic and cell-biological approaches.  131  Figure 25. A combined model of the septate junction.  Results from this thesis (blue arrows) are combined with interactions demonstrated in the embryonic epidermis (red arrows) and the leg epithelium (green arrow). Overlap between the wing and epidermis are indicated with purple arrows. The interaction between Scrib/DIg and Lgl in the epidermis suggests that the same may hold true in the wing, linking the Gli PCP pathway to the cytoskeleton. The interaction between Zip and Vic in the leg epithlium (green arrow) is also suggestive of a cytoskeletal link in SJ-mediated wing PCP.  132  133  C H A P T E R IV. GLIOTACTIN  DETERMINES PCP INDEPENDENTLY O F  SIGNALING.  134  FRIZZLED  Introduction  The finding that mutations in SJ components disrupt planar polarity immediately raises the issue of the relationship between SJ PCP and the frizzled pathway. Fz signaling transduces the asymmetry of apical PCP determinants to the cytoskeleton (Winter et al., 2001). The possibility that the Fz signal also impinges on the septate junction is an intriguing one. Equally intriguing is the possibility that septate junctions are a novel Fz-independent PCP determinant. The aim of this chapter is to determine if Gli functions downstream of the Fz signal, or if Gli-mediated PCP is Fz-independent.  Several lines of evidence suggest that SJ-mediated PCP functions independently of the Fz pathway. Fz-pathway mutants typically have large regions where wing hairs reverse polarity and point from distal to proximal (Gubb and Garcia-Bellido, 1982; Wong and Adler, 1993), (Figure 6). Hair polarity in SJ mutants, however, maintains overall proximal-distal polarity except for rare, isolated hairs in severely mutant wings. Secondly, adjacent cells tend to have parallel aberrant polarity in Fz pathway mutants, whereas SJ mutants typically have adjacent cells with reciprocal polarity defects, producing wing hairs that converge in a chevron pattern. Thirdly, all Fz-pathway members isolated to date produce at least some cells with multiple wing hairs, especially in regions where the polarity vector changes quickly (Wong and Adler, 1993). In contrast, multiple hairs per cell were not observed in Gli mutants, nor even in Gli wings severely enhanced by additional SJ mutations. Fourthly, by virtue of its use as a cell-boundary marker, it is known that the SJ component Coracle is not mislocalized in the Fz-pathway mutant diego (Feiguin et al., 2001). That Cor localization is independent of Fz signaling is not surprising given that the septate junction is basal to the apical cortical domain of Fz-pathway components. Given the differences between the/z and SJ-mutant PCP phenotypes and the/z-independence of Cor localization, it is likely that Gli mutants alter wing PCP through a/z-independent mechanism. The subcellular localization of Gli to the tricellular SJ, however, differs from that of Cor. Gli is polarized to  135  cell vertices in the proximal-distal plane was well as to the septate junction in the apical-basal plane. The subcellular localization of Gli makes it an ideal candidate to interact with the Fz pathway, if an interaction between septate junctions and Fz signaling exists.  To determine the relationship between the Fz pathway and SJ-mediated PCP, two approaches were taken. First, the subcellular localization of Gli was determined in a mutant that lacks Fz signaling. Second, the epistatic relationship between Gli and the Fz-pathway was evaluated using double mutants. The results obtained indicate that septate junction PCP is a parallel PCP pathway that functions independently of Fz signaling.  136  Materials and methods  Genetic interactions  Genetic symbols and the genotypes of balancer chromosomes is given in Appendix 1. Homozygous dsh' and pk' stocks were obtained from the Bloomington stock center and recombined with Git . The X-linked dsh mutation was doubly balanced with Git in a {FM7c/w, dsh ; h, Git , 5  1  5  cn/CyO) stock. Males hemizygous for dsh and homozygous for Git 1  5  1  5  were identified as lacking the  dominant marker associated with FM7c and being homozygous for black and cinnabar. The pk locus is on the second chromosome. Females of the genotype w" lw"' ; b, Git , cn/pk', cn were mated with IH  x  5  w" /Y; pk'lpk males. A recombinant {b, Git , pk', cn) chromosome was identified by scoring for a lf!  1  5  recombinant male carrying the w marker associated with Git that failed to complement pk'. This +  5  chromosome was isogenized and balanced over CyO. Flies simultaneously homozygous for Git and 5  pk' were identified in this stock as being homozygous for black and cinnabar. Wings from {dsh'lY; Git 1 Git } and [Glt ,pk'l 5  5  5  Glt ,pk'} flies were mounted and photographed as previously described. 5  Immunostaining of pupal wings  Homozygous wild-type [Gli ) and dsh (dsh ) pupae were aged to 30 hours APF at 25"C and AE1  1  fixed overnight in 4% formaldehyde in PBS at 4"C. Fixed pupae were dissected in PBS and incubated with rabbit anti-Gliotactin polyclonal antibody at a dilution of 1:50 and mouse anti-Coracle monoclonal antibody (9C and C615-16B cocktail) at 1:100 (Fehon et al., 1994). Stained wings were labeled using an Alexa 488-conjugated goat anti-mouse, and goat anti-rabbit Alexa 568-conjugated secondary antibodies (Molecular Probes) at 1:250 each. To compare the location of Gli to that of Fmi, wild-type {Glt ) E2  pupae were aged to 30 hours APF and stained for Gli using rabbit anti-Gliotactin polyclonal antibody at  137  1:50 and Fmi using mouse anti-Fmi monoclonal #74 at 1:10 (Usui et al., 1999). Stained wings were labeled using goat anti-rabbit, Alexa 488-conjugated and goat anti-mouse, Alexa 568-conjugated secondary antibodies (Molecular Probes) at 1:250 each. Wings were mounted in Vectashield (Vector Laboratories) and visualized using a Zeiss Axioplan 2 deconvolution microscope. Alexa-586 phalloid staining was performed as described in Chapter II.  138  Results  To determine if the subcellular localization of Gliotactin was dependent on Fz signaling, the location of Gli was assayed in dsh' wings (Figure 26). In dsh wings Dsh protein does not polarize to the 1  proximal-distal cell cortex but remains in the cytoplasm, thus abolishing Fz signaling (Axelrod, 2001). As previously shown, Gli localizes to the tricellular septate junction in wild-type pupal wings and overlaps with Cor (Figure 26A-C). The localization of Gli and Cor was not disrupted in a dsh' mutant wing (Figure 26D-F). Thus the subcellular localization of Gli is independent of Fz signaling, as has previously been shown for Cor (Feiguin et al., 2001).  To determine the extent of overlap between Gli and Fz-pathway components at the apical cortex, wild-type pupal wings were double-labeled for Gli and Flamingo. Fmi co-localizes with Dsh, Fz, and Dgo (Strutt, 2002). Even the most apical Gli immunoreactivity was basal to Fmi staining (Figure 26G-I). As has been demonstrated previously (Usui et al., 1999), Fmi staining was restricted to the apical cortex where it localized preferentially to proximal and distal cell boundaries (Figure 26J). Thus Gli does not physically contact members of the Fz pathway.  To confirm the finding that Gli is independent of Fz signaling, genetic interactions between Gli and the Fz-pathway mutants dsh and pk were investigated. In wings mutant for dsh', wing hairs follow curved vectors with areas of reversed polarity (Figure 27A). In dsh'/Y; Gli' " I Gli'" wings, the general 1  0  dsn' polarity pattern is maintained, however, the Gli pattern is evident as well (Figure 27B). Overall, wing hairs in this genotype follow curved vectors and have reversed polarity, as in dsh' wings. Interestingly, the Gli chevron pattern is superimposed onto the dsh' pattern: numerous adjacent hairs were found to converge in dsh'/Y; Gli' /Gli' h5  h0  wings (Figure 27B), even in regions of reversed polarity.  139  Figure 26. Gli-mediated P C P is independent of Fz signaling.  (A-C) Gli (green) and Cor (red) in a wild-type pupal wing at 30 hours APF. Gli is enriched at the tricellular junction (A). Cor immunoreactivity labels the entire septate junction domain evenly. The two proteins overlap in their distribution at the tricellular septate junction (C). (D-F) Gli (green) and Cor (red) in a dsh' pupal wing at 30 hours APF. The expression pattern of Gli (D) and Cor (E) is unaffected. Gli localizes with Cor at the tricellular septate junction (F). (G-H). Gli (green) and Fmi (red) in a wild-type pupal wing at 30 hours APF. The basalmost expression of Fmi does not contact Gli at the septate junction. (G) Fmi at the apical boundary of the Gli expression domain. No immunoreactivity was present. (H) Gli expression in the same optical section, and (I) the merged image. Fmi staining was apparent apical to the septate junction and was preferentially localized to proximal-distal cell boundaries (J).  Magnification: A-F = 2500X, G-J = 3800X.  140  141  Figure 27. Epistasis analysis between the Gli and /z-pathways.  (A) A dsh'/Y wing distal to the posterior crossvein. Hair polarity follows curved vectors that reverse polarity from distal to proximal. (B) A dsh'/Y; Gli' /Gli' ' wing wing distal to the posterior crossvein. The dsh' pattern is lv5  h 5  visible, yet numerous hairs converge with the typical Gli chevron phenotype (arrows). The Gli pattern follows the dsh' polarity vectors. (C) A pk'/pk wing distal to the posterior crossvein exhibiting the pk polarity pattern. 1  (D) A Gli' \ pk'/Gli' '' , pk' wing. The Gli mutant pattern is superimposed onto the pk pattern h  1  5  (arrows). (E) Prehair initiation in a Gli' '' , pk'/Gli' " , pk' wing At 33 hours APF visualized with 1  5  1  5  fluorescent phalloidin. Prehairs initiate at the cell center in this genotype. Magnification: A-D = 200X, E = 2500X.  142  143  Homozygosity for Gli' had the same effect on the pk' phenotype (Figure 27C,D). Hairs follow smooth lvS  curves in pk wings (Figure 27C); whereas in [Gli' , pk } homozygotes numerous chevrons are 1  h6  1  superimposed onto the pk pattern (Figure 27D). Taken together, these genetic results indicate that 1  neither the Gli nor the fz pathways are epistatic to the other, but rather that they operate in parallel. To confirm that the effect of Gli' homozygousity was independent of the site of prehair localization, lvS  {Gli'" , pk'} homozygotes were stained using phalloidin to visualize prehair initiation (Figure 27E). 6  Prehairs in {Gli' ' , pk'} double mutants initiate at the mutant pk position, at the cell center. Thus Gli 1 6  functions to align hairs whether the prehair initiates at the cell periphery or the cell center.  144  Discussion  The finding that G l i functions independently of the frizzled pathway to determine wing P C P identifies the septate junction as the first Fz-independent P C P determinant. The/z-independence of G l i function is consistent with the marked differences between the G l i and /z-pathway phenotypes and suggests that Fz signaling and septate junctions work in parallel to establish the polarity of wing hairs. Gli thus satisfies both the/z-independent and initiation site-independent predictions for a parallel alignment mechanism.  The physical location of G l i relative to Fz-pathway components indicates that G l i does not localize to the prehair initiation site on the apical surface, but rather below it at the level of the septate junction. This result strongly suggests that the effect of G l i on hair polarity is not direct, but rather mediated through another factor. The finding that Gli mutants disrupt P C P in dsh and pk mutant backgrounds is further evidence that Gli acts through an intermediate to determine hair polarity, since prehair initiation occurs at the center of wing cells in these/z-pathway mutants (Wong and Adler, 1993). Thus the action of G l i on hair polarity is independent of the site of prehair initiation and does not require the prehair to initiate at the cell periphery, which would be predicted for a direct effect of G l i on prehair polarity. A logical bridge between the septate junction and hair polarity is the cytoskeleton. F-actin and microtubules are present at the level of the septate junction (Tepass et al., 2001), and both the actin and microtubule cytoskeletons play a role in prehair elongation and polarity (Eaton, 1997; Eaton et al., 1996; Turner and Adler, 1998; Winter et al., 2001). Further experiments evaluating the cytoskeletal link between the septate junction and the prehair will be required to test the hypothesis that SJ P C P is mediated through the cytoskeleton.  145  CHAPTER V. A SCREEN FOR DOMINANT SUPPRESSORS OF GLIOTACTIN.  146  Introduction  The availability of a defined Gliotactin allelic series with a wide range in viabilities is an ideal starting point for second-site modifier screens to identify additional genes in the Gli pathway. The aim of this chapter is to identify putative components of the Gli wing P C P pathway by screening for dominant suppressors of Gli lethality that also suppress Gli P C P phenotypes.  Suppressor/enhancer screens are an effective method for identifying members of a genetic pathway (Karim et al., 1996; Raabe et al., 1996; Rogge et al., 1991). A s a forward-genetics approach it is unbiased in that it requires no preconceived hypothesis of gene interaction. Suppressor screens in Drosophila may recover intragenic suppressors (i.e. second-site mutations within the original mutant locus), in addition to genetic pathway members (extragenic suppressors). Since the characterization of gain-of-function, intragenic modifiers contributes to structure-function studies, suppressor screens can be a complementary approach to the induction and characterization of loss-of-function point mutations (Dellinger et al., 2000). Modifier screens rely on an appropriate phenotype that is easily scored, for example, a dominant eye phenotype induced by a transgene (Raabe et al., 1996). After screening, candidate modifiers are re-tested in a more appropriate context for similar effects. In the case of Gliotactin, screens for modifiers of Gli wing phenotypes are possible, but would be laborious in contrast to isolating modifiers of Gli lethality and subsequently testing these mutations for an effect on wing P C P . This approach, while easier, may indeed bias the obtained results; however, it has the advantage of identifying Gli interactors that do not participate in wing P C P .  Numerous screens for specific classes of C//-interacting mutations are possible, as second-site modifiers may be recessive or dominant, and may enhance or suppress Gli phenotypes. Perhaps the  147  simplest screen is for dominant suppressors of G7/-induced lethality. The Gli' "/Gli' '' genotype is ideal 1  1  3  for such a screen: this Gli genotype is essentially lethal, however, a small portion of Gli' "/Gli' '' animals 1  1  3  eclose, indicating that this genotype is viable in rare cases. Gli' '' /Gli' '' animals have severe motor 1  5  1  3  defects as well as leg malformations (Figure 1 1), generally die within 24 hours of eclosion, and are unable to mate. The viability and fertility of the Gli "IGli " combination allows for mutagenesis of a d  d  homozygous male genotype and subsequent screening for mutagenized Gli' " haplotypes that are viable 1  5  and fertile when paired with Gli' '' . Escapers can be easily identified by the recessive markers associated 1  3  with Gli mutant chromosomes and recrossed to Gli' '' heterozygotes to confirm the presence of a 1  3  suppressor. This strategy proved successful in identifying several dominant Suppressor of Gliotactin (Su(Gli)) loci, of which at least one exerts a dominant effect on the Gli wing PCP phenotype in addition to suppressing lethality.  148  Materials and methods  Fly stocks  An explanation of genetic symbols and the genotypes of balancer chromosomes is given in Appendix 1. The following stocks were from the Auld laboratory collection: {b, Gll  , cn/CyO), {b,  RAR77  Gli' " , cnlCyO), [b, Gli' ' , cn/CyO], [b, Gli' '\ cn/CyO], [b, Gli 1  5  1 1  h  ICyO], and [SplCyO; GllTM6b  AE2A45  Tb). The deficiency stocks Df(2L)El 10/CyO, Df(2L)BSC5/CyO, Df(2L)BSC6/CyO and Df(2L)BSC7/CyO were obtained from the Bloomington stock center, as were two P-element insertions on 2L (P{lacW}Gef26 /CyO at cytological band 26C2-3 and P{lacW}l(2)kl4206 VCyO at 26F3kl372l,  kl42(,f  5) and a CyO balancer carrying a P-insertion that contains the apoptosis gene hid under the control of the heat shock promoter (CyO, P{hs-hid}4).  Mutagenesis  Male or {b, Gli' " , cnlb, Gli' " , cn} or {b, Gli' , cnlCyO) flies were exposed to 25 mM ethyl 1  5  1  5  lvS  methanesulfonate (EMS) in a 4% sucrose solution for 24 hours at room temperature and then mated to virgin [b, Gli' '\ cnlCyO] females at 25"C. Males were discarded 72 hours after mating. h  The effect of second-chromosome Su(Gli) mutations on Gliotactin intragenic complementation  Crosses between heterozygotes were performed at 22"C. Each chromosome carrying a Gliotactin allele is marked with the recessive marker black. Adult {Su(Gli), Gli' }IGli escapers were scored as k5  149  black homozygotes. The proportion of escapers relative to the heterozygotes in each cross was determined and normalized to a control cross of Gli''" ICyO to Gli ICyO, 5  AE2  Scoring of mutant wings  Dissection, mounting and scoring of mutant wings was performed as described in Chapter II.  The effect of Su(Gli)l on wing PCP  To test the effect of Su(Gli)J on the Gli wing PCP phenotype, {b, Gli' " , Su(Gli)J recllCyO) 1  5  females were crossed to \b, Gli''"', cnlCyO) males; F l Gli' '' , Su(Gli)I rec2l Gli' "' progeny were 1  5  1  identified as black and scored for wing PCP. To test the effect of Su(Gli)l on wing PCP and blister formation, {b, Gli''" , Su(Gli)l rec2ICyO) females were crossed to {b, Gli' " , cn/CyO; scrib' /TM3, Sb] 5  1  5  7B3  and {b, Gli' " , cn/CyO; Nrx "/TM3, Sb} males. F l {Git* , Su(Gli)lIGlt'" ; scrib /+} or {Git*' , 1  5  25  5  5  i?B3  5  Su( Gli) 11 Gli' ' ; Nrx "/+} progeny were identified as black and Sb . 1 0  Mapping  23  +  Su(Gli)l  To map the Su(Gli)] locus to two white* P-elements on 2L, the l(2L)Su(Gli)l chromosome was balanced over a CyO derivative carrying an inducible death gene hid under the control of the heat shock promoter (hs-hid). Males from this stock were crossed to females heterozygous for l(2L)Su(Gli)l in trans to either P{ lac W}Gef26  kl 3 7 2 , 1  or PflacW }l(2)kl4206  kl42()fi  . F l third-instar larvae resulting from  these crosses were heat-shocked at 37"C to induce apoptosis in progeny carrying the hs-hid transgene. Surviving progeny were scored for the white* marker associated with the P-element. Viable F l Flies  150  carrying the white* marker were scored as parental; white-eyed flies were scored as recombinant. The l(2L)Su(Gli)l/l(2L)Su(Gli)l genotype is viable at a very low frequency, so putative recombinants were scored for a recessive motor phenotype associated with the l(2L)Su(GU)l/l(2L)Su(Gli)l genotype. Thus white-eyed, l(2L)Su(Gli)lll(2L)Su(Gli)l individuals were eliminated from the recombinant class.  151  Results  To screen for dominant male  Glt' IGlt 5  5  or  Su(Gli) mutations which suppressed the lethality of Glt°IGlt'  3  animals,  Glt' ICyO flies were exposed to E M S and mated to Glt' ICyO females (Figure 28). 5  3  Git' 1 Git' males are too weak to mate, thus only individuals carrying a novel suppressor of the 5  3  Glt°/Glt'  3  phenotype will generate progeny. Approximately 10,000 F l flies were scored and 24  Git' */Git' ; +*/+ males were obtained. O f these 24 flies, seven produced lines which contained a 5  3  heritable  Su(Gli) mutation. Lines were isogenized and balanced for the second and third chromosomes  (Figure 28).  Three  Su(Gli) lines have been investigated to date: Su(Gll) I, 7 and 8. (Line numbers correspond  to the number of original F l escaper male). In the  Su(Gli) J, 7 and 8 lines the Su(Gli) locus co-  Gliotactin locus in the absence of meiotic crossing over, indicating that the  segregates with the  suppressor mutation is on the second chromosome. The Su(Gli)7 locus has not been mapped further, whereas  Su(Gli)8 was mapped to Gliotactin itself. Recombination between Su(Gli)8 and Git' was not 5  observed (n= 50). Thus it was suspected that Su(Gli)8 represented an intragenic suppressor of the Git'  5  allele. Partial sequencing of the Gli locus on the Su(Gli)8 chromosome revealed that the original Git'  5  mutant site was indeed reverted to wild-type in this line. In order to asses the effect of on the viability of near-lethal chromosomes with the the lethality of three  5  5  AE2A45  AE2A45  and  Git' /Git' 5  3  and  Git' . Su(Gli)l, 7 and 8 significantly suppressed  (Table 15).  3  Su(Gli)l and 7 had comparable effects on the  Gliotactin genotypes tested (Table 15), whereas Su(Gli)8 consistently suppressed the lethality of  Git' 1 Git' and Glt' IGli 5  Gli genotypes, crosses were performed to pair these {Git' , Su(Gli)}  Gliotactin alleles Gli  Glt' /Gli  Su(Gli)l, 7 and 8  3  5  AE2A45  to a greater extent than either  Su(Gli)J or 7 (Table 15). The effect of  Su(Gli)] on the several semi-viable Gli genotypes was also determined. Su(Gli)] significantly  152  Figure 28. An EMS screen for dominant suppressors of Gli lethality.  An explanation of genetic symbols and the genotypes of balancer chromosomes is given in Appendix 1. The Gli' ' second chromosome and a wild-type third chromosome were mutagenized with EMS (F„) and 1 0  screened in the F, for viability in trans to Gli " . Surviving Gli' " /Gli' * males were recrossed to balanced d  J  1  5  1  3  Gli' " virgins to confirm the presence of a heritable Su(Gli) mutation. Gli' ' /Gli' * F males were mated 1  1  1 0  1  3  2  with doubly-balanced females. Several male progeny resulting from a given F male were serially 2  crossed to the same balancer stock and to balanced Gli' " virgins. Males that produced viable Gli' ' 1 Gli' " 1  3  1 5  progeny when crossed to Gli''" were identified and the corresponding balancer cross was used to 3  generate a doubly balanced Su(Gli) line.  153  1  3  Fo d t f  Fi  * dv5  W  *  GH  y  *CvO  1118  Cf  1118  3  dv3  w  *  CvO  1118 J H O  3  dv3 33 v / V Gijcivs  w  Gli  W  '  Y  w"J8 Y  G  l  i  3  GU  VI' 1118'  ' *3  *GHdv5  w  ^  F2  * JU8  dv3 ' 3  *Glidv5 ' G//^ ;  *  3  ' 3  ?  w>l 118  v  '|  S  G/  p  v i ^ # ' CvO '  TM6Tb  V ,  O  w  Y 5p  w  1118  X  ;  1118 '  1118  )  T  M  * dv5 C  y  0  6  *  GU  7 or wH18'  G  o  :  *  3 o  1118  b  o  U18  ;  W  r  r  3  Serial S P M s  TM6 Tb  w  X  T  3  udv3  r G  CvO  G/ '  CvC  * lidv5  ndv3  G  3  CvO  *Gli  dv5  3  Score for  Gli dv3  ' M 6 Tb r  Isogenized stock  154  9  o a.  «  z o  3  a.  0  o o  3*  o o  •o •o +  —  —  z  K>  o  O O  o b  A  o b  155  suppressed the lethality of the Gli' * 1 Gli' " and Gli' ' IGli 1  viability of Gli''" /Gli 5  RAR7?  5  1  5  h 5  lM  genotypes but did not significantly alter the  (Table 16).  The location of Su(Gli)I was determined to be the left arm of chromosome 2 (2L) by meiotic crossing over. Recombination between the Gli' " AE2 marker and black consistently separated Su(Gli)l 1  5  from Gli''* as assayed by the ability of recombinant chromosomes to escape over Gli''* (n=15). 5  5  Recombination between the AE2 P-element and cinnabar removed Su(Gli)I from Gli' * in 9 of 20 trials, 1  5  indicating that Su(Gli)l assorts independently of the cinnabar locus, or nearly so. The Gli''" , Su(Gli)l 5  chromosome is homozygous lethal in contrast to the Gli' * chromosome. Based on the approximate map 1  5  position of Su(Gli)] on 2L and the possible homozygous lethality of the Su(Gli)l mutation, a recombination event separating 2L from the {Gli' * , Su(Gli)J } chromosome was selected for and 1  5  tentatively named {lethal (2L) Su(Gli)l). This chromosome was 0.8% viable. In order to demonstrate the presence of Su(Gli)] on the {lethal (2L) Su(Gli)J) chromosome, {lethal (2L) Su(Gli)l) was allowed to recombine with a Gli' " chromosome to produce {lethal (2L) Su(Gli)l, Gli' * } recombinant 1  5  1  5  chromosomes that were scored for the suppressor phenotype in trans to Gli' * . Two independent lines 1  3  were recovered. Both recombinant lines failed to complement the {lethal (2L) Su(Gli)J} chromosome and suppressed the lethality of the Gli' " 1 Gli' " genotype (Table 17). The {lethal (2L) Su(Gli)!} locus 1  5  1  3  was also recombined onto a Gli' " chromosome and tested for complementation against both {lethal (2L) 1  3  Su(Gli)l} and Gli''" . The {lethal (2L) Su(Gli)], Gli' * } chromosome suppressed the lethality of the 5  1  3  Gli' " 1 Gli' " genotype (Table 17) and failed to complement {lethal (2L) Su(Gli)i}. These results 1  5  1  3  demonstrate that the Su(Gli)I locus can be recombined with the Gli locus, proving that Su(Gli)l is an extragenic suppressor of Gli. Additionally, these results prove that the {lethal (2L) Su(Gli)l} chromosome retains Su(Gli)].  156  In  cn  5?  H  2 o  a o  3 a. O  Q S3 S3  o O  o  o  +  2  2  .fc.  o  o  o O  a. O  o o  O  o  b  +  +  157  Table 17. Su(Gli)! is a dominant suppressor that acts in trans to Gli  in trans to Gli  Gli . dv5  Su(Gli)I genotype  N  o  n  -  Escapers  d  v  3  Escaper frequency  escapers  Gli ,Su(Gli)l  792  42  0.05  146  9  0.06  1465  93  0.06  dv5  Gli ,Su(Gli)l  reel  Gli ,Su(Gli)l  reel  dv5  dv5  dv5 i trans to GI.  Gli ,Su(Gli)l dv3  genotype  N  o  n  "  Escapers  Escaper frequency  escapers  Gli ,Su(Gli)l dv3  reel  839  54  158  0.06  Since the lethal (2L) Su(Gli)] mutation appeared in fact to be the Su(Gli)I locus, the lethal (2L) Su(Gli)] locus was mapped relative to two P-elements inserted on 2L: P{Gef26} is inserted at cytological region 26C2-3, and P{l(2)k 14206} is inserted proximal to P{Gef26} at 26F3-5 (FlyBase, 2003) (Figure 29). The lethal (2L) Su(Gli)] mutation mapped 0.7 map units away from P{Gef26} (n=268), whereas no recombination was observed between lethal (2L) Su(Gli)l and P{l(2)kl4206} (n=311). Thus lethal (2L) Su(Gli)] maps between these P-elements. Given this map location of the recessive lethal mutation on {lethal (2L) Su(Gli)!}, deficiency mapping of this region was undertaken in an effort to identify a deficiency which failed to complement the putative Su(Gli)l lethal. Deficiency stocks were obtained from the Bloomington stock center and tested for complementation against the Gli' '' , Su(Gli)] chromosome. No deficiency tested failed to complement Su(Gli)J (Figure 29). The exact 1  5  locations of the proximal breakpoint of Df(2L)El 10 and the distal breakpoint of Df(2L)BSC6 are uncertain (FlyBase, 2003). Using the most conservative estimates for each breakpoint, no deficiency exists for the region extending from 26D3 to 26E1 (Figure 29). Assuming that the Su(Gli)I mutation would fail to complement a deficiency for its wild-type locus, this small cytological region contains the Su(Gli)l gene. This region contains sixteen predicted open reading frames, none of which are known to function at the septate junction or in the wing (Table 18). Su(Gli)l thus represents a novel component of the Gli genetic pathway.  Since Su(Gli)J was definitively shown to be a novel participant in the Gli pathway, the effect of Su(Gli)l on septate junction-mediated wing PCP was determined. The Su(Gli)i mutation dominantly suppressed the wing PCP phenotype of the Gli' IGli '', h5  liAR7  Gli' " 1 Gli' " , Gli' " 1 Gli' *' and Gli' * 1 Gli' ' 1  5  1  5  1  5  1  1  5  1 3  genotypes (Table 19; Figure 30). Additionally, Su(Gli)] suppressed the PCP and adhesion defects of {GW IGli' ' ; Nrr l+) h5  h 5  5ll  and [Gli' " 1 Gli' ' ; scrib /+} animals (Table 19). Su(Gli)J prevented blister 1  5  1 5  i?B3  formation in [Gli' * 1 Gli' " ; Nrx /+} wings and significantly reduced the severity of the PCP phenotype 1  5  1  5  25ll  159  Figure 29. Mapping  ofSu(Gli)!.  Polytene chromosome segment 26 on 2L is shown aligned with the insertion sites of two P-elements (triangles) used to map Su(Gli)l and several deficiencies that complemented the lethality of the lethal (2L) Su(Gli)I locus. Grey bars represent uncertainty in deficiency breakpoints. Using the most conservative estimates for the breakpoints of Df(2L)El 10, Df(2L)BSC6 and Df(2L)BSC7, no deficiency exists which deletes cytological region 26D3-E1 (stippled bar).  160  P{Gef26}  P{l(2)k 14206}  V  V  Df(2L)BSC5 Df(2L)E.110  —  K^v. riv. v^ . v!V v^V 1  "^V  Df(2L)BSC6 Df(2L)BSC7 • • • • • • • • • • • • • • • I I  161  Table 18. Possible candidates for Su(Gli)l  Cytology  Proposed molecular function  Flybase I D  26D3-D4  CG9527  pristanoyl-CoA oxidase  26D4-D5  CG9528  phosphatidylinositol transporter  26D5  CG9526  unknown  26D5-D6  C G I 6947  unknown  26D7  CG9531  protoporphyrino oxidase  26D7  BcDNA:LD24639  UDP-N-acetylglucosamine pyrophosphorylase  26D7  CG9536  unknown  26D7-D8  Sec61a  translocon  26D8-D9  DLP  D a x x - l i k e protein, function unknown  26D9  CG9542  unknown  26D9  BcDNA:LD29885  C O P I vesicle coat  26D9  CG9548  unknown  26D9  CG31638  actin-binding  26D9  CG9547  glutaryl-CoA dehyrogenase  26D9  CG9550  chondroitin 6-sulfotransferase  26D9-E1  CG31637  '  sulfotransf'erase  162  Table 19. Su(Gli)l is a dominant supressor of septate-junction PCP and adhesion defects  Su(Gli)l genotype  Septate junction genotype  +/+  Su(Gli)l/+  dv5 RAR77  2.9  1.0*  3.4  1.6*  4.1  0.8*  6.7"  6.0*  6.5  3.6*  Gll  IGU  Gli /GIi dv5  Gli  dv5  IGIi  dv5  dvl  Gli IGli \Nrx l+ dv5  dv5  Gli IGli ; dv5  dv5  25n  scrib J /+ 7B3  B  The average number of wing sections containing a sector with abnormal polarity is indicated for various genotypes (n = 20 wings per genotype). * indicates a significant enhancement relative to the control genotype (Mann-Whitney U-tcst,  P<0.01). A "  indicates wing blisters were present in  more than 10% of wings scored.  163  Figure 30. The effect of Su(Gli)l on wing P C P and adhesion.  For all images anterior is up and distal is to the right.  (A) The posterior margin of a Gli' IGli' h6  M  wing. Numerous hairs show the Gli mutant polarity  phenotype. (B) The equivalent region in a Gli' ~\ Su(Gli)llGli lv  lM  wing. Wing hairs exhibit wild-type polarity.  (C) The posterior margin of a Gli' ' , Su(Gli)I/Gli " ; scriV l+ wing. No blistering is evident and the h 5  d  5  1B3  PCP phenotype is almost wild-type. (D) The anterior-distal margin of a lethal (2L) Su(Gli)IIlethal (2L) Su(Gli)l wing. Numerous hairs have a polarity pattern identical to severe SJ mutations.  164  165  for this genotype (Table 19). The effect ot'Su(Gli)! on the {Gli' /Gli''"~; scrib' "1+] genotype was h6  71  dramatic. The {Gli''* /Gli' ; scrib' l+) genotype has complete penetrance for the marginal blister 5  lv5  7Bi  phenotype and a severe PCP phenotype (Tables 4, 5). Su(Gli)l dominantly suppressed blister formation in this genotype with complete efficacy and simultaneously suppressed the PCP phenotype to a great extent (Figure 30C, Table 19).  To assess the impact of the Su(Gli)l mutation on wing PCP in a wild-type Gli context, flies homozygous for Su(Gli)] were collected from a balanced lethal (2L) Su(Gli)l stock and scored for wing PCP. Strikingly, these flies have a dramatic wing phenotype that phenocopies severe SJ-mutant combinations (Figure 30D). Taken together, these results indicate that the Su(Gli)l locus encodes a novel component of the Gli pathway that functions during SJ-mediated PCP and adhesion in the developing wing.  166  Discussion  The results obtained through this preliminary screen indicate that at least one Su(Gli) locus participates with Gliotactin in both its essential and wing PCP functions, demonstrating in principle that screening for suppressors of lethality can identify suppressors of septate-junction PCP. While the screen produced a low frequency of suppressor loci (approximately 0.07%), searching for dominant suppressors in the F l allowed for efficient identification of potentially interesting genotypes. Thus a large-scale screen for Gli suppressors using this approach seems feasible. Systematic screens have proved very useful in elucidating several pathways at the genetic and cell biological levels in Drosophila (Prout et al., 1997; Raabe et al., 1996; Rogge et al., 1991; Walsh and Brown, 1998). A similar approach using Gli and perhaps other hypomorphic SJ mutants should prove equally fruitful.  Su(Gli)l maps to a region of 2L that does not contain a known component of the septate junction, nor a participant in wing PCP or adhesion. Identification of the locus encoded by Su(Gli)l will therefore shed new light onto these three processes. One predicted open reading frame in this region (CG31638) has an actin-binding motif. A link between the septate junction and wing PCP/adhesion could be mediated by the cytoskeleton, thus this candidate for Su(Gli)I is perhaps the most promising based solely on sequence homology. Several lethal P-element insertions have been mapped cytologically to 26D-26E (FlyBase, 2003). Complementation tests between Su(Gli)l and these lethal P-element lines will hopefully identify which open reading frame in this region encodes Su(Gli)l.  Five additional Su(Gli) lines remain to be characterized. Su(Gli)7 maps to the second chromosome. Two G/('-interacting genes (excluding Gli itself) are on the second chromosome, namely vulcan and coracle. Su(Gli)7 thus may be a cor or vie allele, or it may be an intragenic suppressor as was  167  the case for  Su(Gli)8. Additionally, the lethal-giant larvae locus is on the second chromosome, raising  the possibility that Su(Gli)7 is allelic to lgl. Another possibility is that Su(Gli)7 is an additional allele of  Su(Gli)]. Complementation analysis of Su(Gli)7 with these candidate second-chromosome mutations would either rule them out as possibilities or identify the  Su(Gli)7 locus. The Gli' , Su(Gli)7 ho  chromosome is recessive lethal, suggesting that Su(Gli)7 will fail to complement a null allele of its locus. O f course, the lethality of the Gli' , Su(Gli)7 chromosome may reflect a random EMS-induced lv5  lethal mutation on this chromosome.  The effect of the remaining five  Su(Gli) loci on wing P C P and adhesion remains to be  investigated. A simple cross of the balanced  Su(Gli) stock to a suitable Gli allele (such as Gli' ) will hl  reveal if these suppressors function in wing P C P as well as suppressing Gli lethality; a cross to the C / ( ' / C y O ; scrib' ''/TM3 AJ  78  which the first six  stock would assess their role in perimarginal adhesion. Given the ease with  Su(Gli) loci were identified, this method of screening for SJ-interacting genes seems  suitable for a large-scale, saturating screen. B y identifying novel Gli interactors, such a screen would greatly advance our understanding of the structure and function of the septate junction in the wing. The role of any novel SJ-interacting proteins discovered could then be assessed in other tissues, such as the epidermis.  168  CHAPTER VI. GENERAL DISCUSSION  The aim of this thesis was to test the hypothesis that Gliotactin is necessary for the development of polarized post-embryonic epithelia. By inducing and characterizing adult-viable Gli genotypes, a role for Gliotactin in/n'zz/ed-independent PCP signaling in the pupal wing epithelium was identified and used as a means of testing genetic interactions between Gli and additional septate-junction mutants. Combinatorial SJ-mutant analysis also revealed that septate junctions function in wing adhesion. Finally, a suppressor screen identified that the Gli wing PCP pathway contains at least one additional gene not previously characterized as a participant in septate junctions, PCP or adhesion.  Gliotactin function is divisible into distinct protein domains  This work supplies the first evidence of the functional importance of two Gliotactin domains. The extracellular serine esterase-like domain is the largest motif of Gliotactin; this motif shares a high degree of similarity to a functional serine esterase as well as to other esterase-like proteins (Botti et al., 1998). The identification of a hypomorphic allele (Git ) as a point mutant in this domain demonstrates 5  the functional significance of this component of Gli structure. The predicted change in Gli conformation resulting from a glycine to glutamic acid substitution at this position is consistent with the Git  5  phenotype. While Git  5  demonstrates that the serine esterase-like domain is necessary for Gli function, a  second point mutant demonstrates that this motif is not sufficient. The Gli  HAK77  allele is a truncation that  retains the entire extracellular domain and a complete transmembrane domain, however, this allele is homozygous lethal at late embryogenesis and displays the Gli paralyzed phenotype. Therefore the intracellular domain of Gli is necessary for function in addition to the extracellular domain.  An additional finding was that the function of both intracellular and extracellular domains could be supplied in trans by separate Gli alleles. Animals simultaneously heterozygous for both the extracellular point mutant and the intracellular truncation survived at an almost wild-type frequency in  170  contrast to animals homozygous for the extracellular point mutant, or embryos homozygous for the intracellular truncation that died at late embryogenesis. This suggests that the protein encoded by the Gli  RAR7?  allele is not simply degraded but is correctly targeted to the septate junction, at least in the  presence of an allele with an intact intracellular domain. The subcellular location of Gll  RAR7?  protein is  not known. If indeed the intracellular PDZ-binding epitope is required for correct SJ-localization of Gli, the function of this allele would seem to be dependent on its physical association with a Gli monomer that possesses this epitope. A definitive test of this hypothesis requires an antibody capable of recognizing the truncated Gli  RAR7?  Gli  /Gli  RAR77  RAR7?  allele in the embryonic epidermis and comparing its localization in  homozygotes and Gli' 'lGli h  RAR77  heterozygotes.  Gliotactin has a unique distribution at the septate junction  The subcellular location of Gli in the developing wing was at the tricellular septate junction. This result is consistent with recent work in the embryonic epidermis (Schulte, 2003) and is the first description of Gli localization in an imaginally^dervied structure. Gli is the first protein described to specifically localize to the tricellular junction in any Drosophila epithelium, suggesting that its function may be distinct from that of other SJ proteins.  Septate junctions control P C P  While septate junction components impinge on the location and function of apical adherens junctions (Bilder and Perrimon, 2000; Bilder et al., 2003; Tanentzapf and Tepass, 2003), this thesis provides the first evidence that a junction in the apical-basal plane can alter cell polarity in the proximaldistal plane in the absence of cell division. While a novel result, the finding that septate junction mutants alter PCP in the developing wing is not completely unexpected. Mutants in coracle alter PCP in the eye  171  by interfering with ommatidial rotation (Lamb et al., 1998), and the rough eyes reported in Nrx and discs-large mutants may also be due to incorrect ommatidial polarity (Baumgartner et al., 1996; Woods et al., 1996). Several PCP determinants function in both eyes and wings despite the fundamental difference between these forms of PCP. Gliotactin seems to be specific to wing PCP: even severe Gli mutants do not have apparent defects in eye or sensory bristle PCP. Consistent with the presence of Cor, Nrx and Dig at the septate junction during wing development, mutations altering these proteins impinge upon SJ function in this epithelium. Epistasis analysis demonstrated that mutations in these SJ components are arranged into groups consistent with studies of the embryonic epidermis: cor and Nrx form an epistasis group, whereas scrib and dig interact genetically with each other and with Nrxlcor. Complementary studies using complete loss-of-function genotypes in the wing are not feasible, with the possible exception of clonal analysis. It is likely that mitotic clones of dig and scrib would overproliferate and cause loss of apical-basolateral polarity, precluding specific localization of any SJ marker. Indeed, recent work has demonstrated that clones of a null allele of dig causes overproduction of cuticle in the wing; the resulting clones have severely disrupted cuticle morphology but were not closely examined for hair polarity or adhesion (Bilder et al., 2003). Clones null for cor, Nrx or Gli alleles may be more informative.  Vulcan is required for SJ-mediated P C P in the wing  The work presented here is also the first experimental evidence that Vulcan plays a role at the septate junction. Vic is a member of the MAGUK-binding SAPAP family of proteins (Gates and Thummel, 2000); Vic thus may physically interact with Dig, which has a MAGUK domain (Woods and Bryant, 1991). Epistasis analysis placed vie within the cor/Nrx group and outside the dig group. Thus the interaction between Vic and Dig is not simply analogous to the Cor-Nrx interaction. The finding that vie groups epistatically with cor and Nrx suggests that Vic function at the SJ is closely linked with the  172  functions of these proteins. The lethal phase of the vie allele used in this work is pupal stages, thus it will be possible to evaluate the distribution of Cor, Nrx and Dig in vie mutant pupae. The epistasis experiments presented in this thesis predict that the subcellular location of Cor and Nrx may be affected in vie mutant cells, but that Dig would be correctly localized to the septate junction. The effect of vie on the SJ-mutant wing phenotype does not necessarily indicate that Vic has a SJ-specific localization. Mutations in lethal-giant larvae genetically interact with SJ-mutations such as dig and scrib (Bilder et al., 2000), yet the Lgl protein is not restricted to the septate junction (Strand et al., 1994; Tanentzapf and Tepass, 2003). The results obtained for vie in this thesis using the wing PCP assay argues strongly that the Vic protein at least overlaps with the septate junction in the wing epithelium, and demonstrates that Vic function is required for SJ-mediated PCP.  Severe SJ genotypes cause blisters at the wing margin  An additional phenotype described for the first time in this work was blistering at the wing margin caused by severe disruption of the septate junction during wing development. Viable Gli and cor mutants never exhibited wing blistering, but mutations in scrib, dig and Nrx dominantly induced marginal blisters in Gli backgrounds. An allele of cor also induced blistering in a <i/g-enhanced Gli genotype. Thus all four SJ components induced blistering in a SJ-mutant background, with varying severity. Null clones of dig in the wing have severely disrupted cuticle (Bilder et al., 2003), indicating that the phenotypes described in this thesis are less severe than the null SJ phenotype for this tissue.  The combinatorial SJ genotypes examined in this thesis are the first genotypes described which blister specifically at the wing margin, a region that adheres in an integrin-independent manner (Brower and Jaffe, 1989). During prepupal and pupal wing development, the wing margin is the only wing region that does not separate completely but remains apposed throughout development (Fristrom and Fristrom,  173  1993). Consistent with the failure of integrin null clones to blister at the margin, integrin expression in the third-instar wing disc is excluded from the margin (Fristrom et al., 1993). Taken together, it appears that septate junctions replace the function of integrins at the wing margins with respect to epithelial adhesion. Unfortunately, the possible formation of the transalar array in this wing subregion has not been investigated. A detailed analysis of the wing margin in wild-type and SJ-mutant pupal wings using transmission electron microscopy to evaluate the presence of septate junctions and their possible functions is required to investigate how septate junctions mediate adhesion in this area of the wing.  Do septate junctions interact with the cytoskeleton?  A common denominator between wing PCP and adhesion is the cytoskeleton. Microtubules and filamentous actin are dynamically regulated during wing development and impact the polarity, formation and number of apical wing hairs as well as the later transalar cytoskeleton required for adhesion. Given that SJ mutants affect both PCP and adhesion, it is possible that the SJ participates in organization or regulation of the apical and transalar cytoskeleton. Consistent with this interpretation, filamentous actin and microtubules are present at the SJ (Tepass et al., 2001). A detailed study of cytoskeletal dynamics at the wing margin using confocal and electron microscopy in wild-type and SJmutant animals would be necessary to fully test this hypothesis. Additionally, testing candidate cytoskeletal components for interactions with SJ mutations may be informative. Possible interacting mutations could include lethal (2) giant larvae and the non-muscle myosin mutants crinkled, zipper and spaghetti-squash. A genetic interaction between vulcan and zipper has previously been reported (Gates and Thummel, 2000), suggesting a possible link between the Gli PCP pathway and the cytoskeleton.  174  SJ-mediated P C P is independent of Fz signaling  The demonstration in this thesis that septate junctions function independently of the frizzled pathway to determine PCP is the first reported instance of a Fz-independent PCP mechanism in the Drosophila wing. Septate junctions are an attractive model for Fz-independent PCP for several reasons. First, the septate junction prefigures Fz-dependent PCP signaling in the wing, since septate junctions are present in the wing imaginal disc from pupariation through 30 hours APF, when Fz-pathway determinants are asymmetrically localized (Fristrom and Fristrom, 1993; Strutt, 2002). Thus the septate junction and associated cytoskeleton may serve as a cellular "template" for the formation of the wing hair. Secondly, the apical microtubule web present in wing epithelial cells is closely apposed to the septate junction (Eaton et al., 1996); this web is the subcellular structure on which the PCP determinant, WdB, is localized (Hannus et al., 2002). Thirdly, the expression of Gli at the septate junction is enriched at tricellular corners, placing a SJ component directly subjacent to the site of prehair initiation at the apical surface.  The possibility that Fz signaling is dependent on the septate junction has not yet been addressed. The /z-pathway determinants Fmi and Dsh are mislocalized in cells mutant for Wdb, suggesting that SJ function may also be required for correct localization of Fz signaling components (Hannus et al., 2002). This hypothesis is readily testable by assaying the subcellular distribution of Fz-pathway components in Gli mutant backgrounds. Additionally, possible interactions between Gli and wdb should be investigated to determine if these two PCP mechanisms are in fact components of the same pathway. These studies would also be an appropriate test of the hypothesis that Gli and septate junctions determine PCP through the cytoskeleton. Possible tests for a Gli wdb interaction would be to assess if the microtubule web/Wdb complex is disrupted in Gli mutants, and conversely, to determine if Gli and other SJ components are  175  mislocalized in wdb tissue. A possible mechanism of Fz-independent PCP signaling incorporating septate junctions, Gli and Wdb is modeled in Figure 31.  Is Gliotactin function conserved in vertebrates?  The septate junction is not a feature of epithelial sheets in vertebrates, although evidence suggests that a septate-like junction exists between axons and glia in the vertebrate nervous system (Einheberet al., 1997). Myelinating Schwann cells wrap peripheral axons and shield them from the surrounding environment (for a review see Peles and Salzer, 2000); this process is analogous to peripheral glia in Drosophda (Auld et al., 1995; Baumgartner et al., 1996). At the borders of nodes of Ravier in myelinated nerves, Schwann cell membrane loops contact the axonal membrane in a series of paranodal junctions which closely resemble septate junctions at the ultrastructural level (Einheber etal., 1997). Interestingly, a vertebrate homologue of Neurexin IV, the Contactin-associated paranodal receptor, or Caspr, is a component of this junction (Einheber et al., 1997). Loss of Caspr function in knockout mice produces several phenotypes associated with disruption of the paranode including decreased motor function and disrupted paranodal myelin loops (Bhat et al., 2001). Similar phenotypes have been reported for the Caspr co-receptor, Contactin (Boyle et al., 2001). The Caspr/Contactin complex functions as a receptor for Neurofascin, a transmembrane protein present at the paranode on myelinating Schwann cells (Charles et al., 2002). The mutant phenotype of Neurofascin has not yet been evaluated.  The murine homolog of Gliotactin is Neuroligin 3, a serine-esterase domain-containing, transmembrane protein expressed in Schwann cells, astrocytes, and other types of glia in the developing mouse embryo (Gilbert et al., 2001). The expression of Neuroligin 3 in Schwann cells raises the possibility that Neuroligin 3 functions at the paranodal junction, however, this hypothesis has not yet  176  Figure 31. A model of/n'zz/erf-independent PCP.  Prior to prehair initiation, components of the apical Fz-dependent PCP signaling pathway are asymmetrically distributed to proximal and distal cell boundaries. A microtubule web (MT) is present basal to the Fz-dependent pathway and contains a distally localized protein phosphatase regulatory subunit, Widerborst (Wdb). Coracle (Cor) and Gliotactin (Gli) overlap in their expression at the tricellular septate junction just basal to the domain of Wdb expression. The sepate-junction/Gli pathway impinges on prehair polarity in a Fz-independent manner, whereas Wdb functions upstream of Fz.  177  Distal  Proximal  178  been tested. Thus despite the recent demonstration of a septate-like junction in vertebrates, it is not known if the vertebrate Gliotactin homologue is localized to, or necessary for, this junction.  Several recent studies have explored the function of vertebrate proteins homologous to Drosophila PCP determinants in zebrafish and Xenopus (Darken et al., 2002; Dijane et al., 2000; Jessen et al., 2002; Park and Moon, 2002). The emerging picture is that convergent extension during vertebrate gastrulation is regulated by a pathway similar to the Drosophila frizzled PCP pathway (for a review see Mlodzik, 2002). Convergent extension is a morphogenic process by which cells intercalate to narrow the lateral axis and extend the anterior posterior axis; this process does not involve cell division or changes in cell fate and thus may correlate more closely with PCP in the Drosophila wing than in the eye or sensory bristle (Axelrod, 2002). While it is possible that Fz-independent PCP signaling occurs in vertebrates, however, nothing is known about the possible function of any septate-junction homologues in this process.  Genetic screens to identify Gli pathway members  The relationship of the Gli PCP pathway to the paranodal junction or/h'zz/ec/-independent convergent extension in vertebrates may be revealed by identifying additional components of the Gli pathway. The preliminary screen performed in this work demonstrated that at least one additional factor participates with Gli in its essential and wing PCP functions, and that this factor is a unknown component of these processes. The ease of forward genetics in Drosophila makes the modifier-screen approach a logical one to identify G//-interacting loci. The hypothesis that the Gli PCP pathway interacts with the cytoskeleton would also be supported if cytoskeletal components were identified as Gli interactors. Consistent with this possibility, one candidate locus for the first Gli suppressor encodes a putative actin-binding motif.  179  Conclusions  Through the characterization of Gli point mutants the functional significance of the Gli serine esterase-like and intracellular domains was demonstrated, including the ability of these domains to act synergistically even when present in different Gli monomers. Gli mutants were examined in adult and pupal stages and found to exhibit defects in the planar cell polarity of the wing and leg epithelia; this phenotype was sensitive to enhancement by additional septate-junction mutations. Combinatorial septate-junction mutants have a blister phenotype at the wing margin, a region that adheres in an integrin-independent fashion. Epistasis analysis using the wing PCP assay placed cor, Nrx and vie into an epistasis group distinct from the dig group and the scrib group. The localization of Gli was found to be independent of frizzled signaling, identifying the septate junction as only the second Fz-independent PCP mechanism reported. Finally, strong hypomorphic Gli genotypes were shown to be a suitable reagent for isolating novel mutations that participate in SJ-mediated PCP and adhesion in the Drosophila wing.  180  VII. R E F E R E N C E S  181  Adler, P. N. (2002). Planar signaling and morphogenesis in Drosophila. Developmental Cell 2, 525-535. Adler, P. N., Charlton, J., and Lui, J. (1998). Mutations in the cadherin superfamily member gene dachsous cause a tissue polarity phenotype by ahtr'mg frizzled signaling. Development 125, 959-968. Adler, P. N., Charlton, J., and Park, W. J. (1994). The Drosophila tissue polarity gene inturned functions prior to wing hair morphogenesis in the regulation of hair polarity and number. Genetics 137, 829-836. Adler, P. N., Krasnow, R. E., and Lui, J. (1997). Tissue polarity points from cells that have higher Frizzled levels towards cells that have lower Frizzled levels. Current Biology 7, 940-949. Adler, P. N., Taylor, J., and Charlton, J. (2000). The domineering non-autonomy of frizzled and Van Gogh clones in the Drosophila wing is a consequence of a disruption in local signaling. Mechanisms of Development 96, 197-207. Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K., and Watson, J. D. (1994). Molecular Biology of the Cell, 3 edn (New York, Garland Publishing). Ashburner, M. (1989). Drosophila: a laboratory handbook (Cold Spring Harbor, Cold Spring Harbor Press). Ashburner, M., Thompson, P., Roote, J. L., F. , Grua, M., El Messal, M., Roth, S., and Simpson, P. (1990). The genetics of a small autosomal region of Drosophila melanogaster containing the structural gene for alcohol dehydrogenase. VII. Characterization of the region around the snail and cactus loci. Genetics 126, 679-694. Auld, V. J., Fetter, R. D., Broadie, K., and Goodman, C. S. (1995). Gliotactin, a novel transmembrane protein on peripheral glia, is required to form the blood-nerve barrier in Drosophila. Cell 81, 757-767. Axelrod, J. D. (2001). Unipolar membrane association of Dishevelled mediates Frizzled planar cell polarity signaling. Genes & Development 15, 1182-1187. Axelrod, J. D. (2002). Strabismus comes into focus. Nature Cell Biology 4, E6-E8. Axelrod, J. D., Miller, J. R., Shulman, J. M., Moon, R. T., and Perrimon, N. (1998). Differential recruitment of Dishevelled provides signaling specificity in the planar cell polarity and Wingless signaling pathways. Genes & Development 12, 2610-2622. Basu, J., Bousbaa, H., Logarinho, E., Li, Z., Williams, B. C., Lopes, C., Sunkel, C. E., and Goldberg, M. L. (1999). Mutations in the essential spindle checkpoint gene bubl cause chromosome missegregation and fail to block apoptosis in Drosophila. The Journal of Cell Biology 146, 13-28. Baumgartner, S., Littleton, J. T., Broadie, K., Bhat, M. A., Harbecke, R., Lengyel, J. A., ChiquetEhrismann, R., Prokop, A., and Bellen, H. J. (1996). A Drosophila neurexin is required for septate junction and blood-nerve barrier formation and function. Cell 87, 1059-1068. Bellen, H. J., Lu, Y., Beckstead, R., and Bhat, M. A. (1998). Neurexin IV, caspr and paranodin-novel members of the neurexin family: encounters of axons and glia. Trends Neurosci 21, 444-449.  182  Bhat, M. A., Rios, J. C, Lu, Y., Garcia-Fresco, G. P., Ching, W., St. Martin, M., J., L., Einheber, S., Chester, M., Rosenbluth, J., et al. (2001). Axon-glia interactions and the domain organization of myelinated axons requires Neurexin IV/Caspr/Paranodin. Neuron 30, 369-383. Bilder, D., Li, M., and Perrimon, N. (2000). Cooperative regulation of cell polarity and growth by Drosophila tumor suppressors. Science 289, 113-116. Bilder, D., and Perrimon, N. (2000). Localization of apical epithelial determinants by the basolateral PDZ protein Scribble. Nature 403, 676-680. Bilder, D., Schober, M., and Perrimon, N. (2003). Integrated activity of PDZ protein complexes regulates epithelial polarity. Nature Cell Biology 5, 53-58. Botti, S. A., Felder, C, Sussman, J. L., and Silman, I. (1998). Electrotactins: a class of adhesion proteins with conserved electrostatic and structural motifs. Protein Engineering 77, 415-420. Boutros, M., Paricio, N., Strutt, D. I., and Mlodzik, M. (1998). Dishevelled activates JNK and discriminates between JNK pathways in planar polarity and wingless signaling. Cell 94, 109-118. Boyle, M. E. T., Berglund, E. O., Murai, K. K., Weber, L., Peles, E., and Ranscht, B. (2001). Contactin orchestrates assembly of the septate-like junctions at the paranode in myelinated peripheral nerve. Neuron 30, 385-397. Brabant, M. C, and Brower, D. L. (1993). PS2 integrin requirements in Drosophila embryo and wing morphogenesis. Dev Biol 757, 49-59. Brabant, M. C, Fristrom, D., Bunch, T. A., and Brower, D. L. (1996). Distinct spatial and temporal functions for PS integrins during Drosophila wing morphogenesis. Development 122, 3307-3317. Brower, D. L., Bunch, T. A., Mukai, L., Adamson, T. E., Wehrli, M., Lam, S., Firedlander, E., Roote, C, and Zusman, S. (1995). Nonequivalent requirements for PS1 and PS2 integrin at cell attachments in Drosophila: genetic analysis of the a ] integrin subunit. Development 72/, 1311-1320. PS  Brower, D. L., and Jaffe, S. M. (1989). Requirement of integrins during Drosophila wing development. Nature 342, 285-287. Brower, D. L., Wilcox, M., Piovant, M., Smith, R. J., and Reger, L. A. (1984). Related cell-surface antigens expressed with positional specificity in Drosophila imaginal discs. Proc Natl Acad Sci USA 81, 7485-7489. Brown, N. H., Gregory, S. L., L., R. W., Fessler, L. I., Prout, M., White, R. A. Ft., and Fristrom, J. W. (2002). Talin is essential for integrin function in Drosophila. Dev Cell 3, 569-579. Chae, J., Kim, M.-J., Goo, J. H„ Collier, S., Gubb, D., Charlton, J., Adler, P. N., and Park, W. J. (1999). The Drosophila tissue polarity gene starry night encodes a member of the protocadherin family. Development 126, 5421-5429.  183  Charles, P., Tait, S., Faivre-Sarrailh, C, Barbin, G., Gunn-Moore, F., Denisenko-Nehrbass, N., Guennoc, A., Girault, J., Brophy, P. J., and Lubetzki, C. (2002). Neurofascin is a glial receptor for .the Paranodin/Caspr-Contactin axonal complex at the axoglial junction. Current Biology 12, 217-220. Collier, S., and Gubb, D. (1998). Drosophila tissue polarity requires the cell-autonomous activity of the fuzzy gene, which encodes a novel transmembrane protein. Development 124, 4029-4037. DAvino, P. P., and Thummel, C. S. (2000). The ecdysone regulatory pathway controls wing morphogenesis and integrin expression during Drosophila metamorphosis. Dev Biol 220, 211-224. Darken, R. S., Scola, A. M., Rakeman, A. S., Das, G., Mlodzik, M., and A., W. P. (2002). The planar polarity gene strabismus regulates convergent extension movements in Xenopus. EMBO J 27, 976-985. Dellinger, B., Felling, R., and Ordway, R. W. (2000). Genetic modifiers of the Drosophila NSF mutant, comatose, include a temperature-sensitive paralytic allele of the calcium channel otl-subunit gene, cacophony. Genetics 755, 203-211. Dijane, A., Riou, J., Umbhauer, M., Boucaut, J., and Shi, D. (2000). Role of Frizzled 7 in the regulation of convergent extension movements during gastrulation in Xenopus laevis. Development 127, 30913100. Eaton, S. (1997). Planar polarity in Drosophila and vertebrate epithelia. Curr Opin Cell Biol 9, 860-866. Eaton, S., Wepf, R., and Simons, K. (1996). Roles for Racl and Cdc42 in planar polarization and hair outgrowth in the wing of Drosophila. J Cell Biol J35, 1277-1289. Edwards, K. A., and Kiehart, D. P. (1996). Drosophila non-muscle myosin II has multiple essential roles in imaginal disc and egg chamber morphogenesis. Development 722, 1499-1511. Einheber, S., Zanazzi, G., Ching, W., Scherer, S., Milner, T. A., Peles, E., and Salzer, J. L. (1997). The axonal membrane protein Caspr, a homologue of neurexin IV, is a component of the septate-like paranodal junctions that assemble during myelination. J Cell Biol 739, 1495-1506. Fehon, R. G., Dawson, I. A., and Artavanis-Tsakonas, S. (1994). A Drosophila homologue of membrane-skeleton protein 4.1 is associated with septate junctions and is encoded by the coracle gene. Development 720, 545-557. Fehon, R. G., LaJeunesse, D., Lamb, R., McCartney, B. M., Schweizer, L., and Ward, R. E. (1997). Functional studies of the protein 4.1 family of junctional proteins in Drosophila. Soc Gen Physiol Ser 52, 149-159. Feiguin, F., Hannus, M., Mlodzik, M., and Eaton, S. (2001). The ankyrin repeat protein Diego mediates Frizzled-dependent planar polarization. Dev Cell / , 93-101. Ferris, G. F. (1950). External morphology of the adult. In Biology of Drosophila, M. Demerec, ed. (New York, John Wiley & Sons), pp. 368-419. FlyBase (2003). The Fly Base database of the Drosophila genome projects and community literature. Available from http://flybase.bio.indiana.edu/. Nucleic Acids Res 31, 172-175.  184  Fogerty, F. J., Fessler, L. I., Bunch, T. A., Yaron, Y., Parker, C. G., Nelson, E. E., Brower, D. L., Gullberg, D., and Fessler, J. H. (1994). Tiggrin, a novel Drosophila extracellular matrix protein that functions as a ligand for Drosophila OC PPS intregrins. Development 120, 1747-1758. Fristrom, D., and Fristrom, J. W. (1993). The metamorphic development of the adult epidermis. In The Development of Drosophila, A. Martinas-Arias, and B. M., eds. (Cold Spring Harbor, New York, Cold Spring Harbor Press), pp. 843-897. PS2  Fristrom, D., Gotwals, P., Eaton, E., Romberg, T. B., Sturtevant, M., Bier, E., and Fristrom, J. W. (1994). blistered: a gene required for vein/intervein formation in wings of Drosophila. Development 120, 2661-2671. Fristrom, D., Wilcox, M., and Fristrom, J. W. (1993). The distribution of PS integrins, laminin A and Factin during key stages in Drosophila wing development. Development 117, 509-523. Fristrom, D. K. (1982). Septate junctions in imaginal disks of Drosophila: a model for the redistribution of septa during cell rearrangement. The Journal of Cell Biology 94, 77-87. Gates, J., and Thummel, C. S. (2000). An enhancer trap screen for ecdysone-inducible genes required for Drosophila adult leg morphogenesis. Genetics 156, 1765-1776. Gilbert, M., Smith, J., Roskams, A. J., and Auld, V. J. (2001). Neuroligin 3 is a vertebrate gliotactin expressed in olfactory ensheathing glia, a growth-promoting class of macroglia. Glia 34, 151 -164. Graner, M. W., Bunch, T. A., Baumgartner, S., Kerschen, A., and Brower, D. L. (1998). Splice variants of the Drosophila PS2 integrins differentially interact with RGD-containing fragments of the extracellular proteins Tiggrin, Ten-m, and D-Laminin a2. Journal of Biological Chemistry 273, 1823518241. Griffiths, A. J. F., Gelbart, W. M., Lewontin, R. C, and Miller, J. H. (2002). Modern Genetic Analysis, Second Edition edn (New York, W. H. Freeman and Company). Gubb, D., and Garcia-Bellido, A. (1982). A genetic analysis of the determination of cuticular polarity during development in Drosophila melanogaster. J Embryol Exp Morphol 68, 37-57. Gubb, D., Green, C., Huen, D., Coulson, D., Johnson, G., Tree, D., Collier, S., and Roote, J. (1999). The balance between isoforms of the Prickle LIM domain protein is critical for planar polarity in Drosophila imaginal discs. Genes & Development 13, 2315-2327. Hannus, M., Feiguin, F., Heisenberg, C.-B., and Eaton, S. (2002). Planar cell polarization requires Wiberborst, a B' regulatory subunit of protein phosphatase 2A. Development 129, 3493-3503. Held, L. I., Duarte, C. M., and Derakhshanian, K. (1986). Extra tarsal joints and abnormal cuticular polarities in various mutants of Drosophila melanogaster. Roux Arch Dev Biol 195, 145-157. Hough, C. D., Woods, D. F., Park, S., and Bryant, P. J. (1997). Organizing a functional junctional complex requires specific domains of the Drosophila MAGUK Discs large. Genes Dev 11, 3242-3253. Humphries, M. J. (2000). Integrin structure. Biochem Soc Trans 28, 311-339.  185  Ichtchenko, K., Nguyen, T., and Sudhof, T. C. (1996). Structures, alternative splicing, and neurexin binding of multiple neuroligins. J Biol Chem 271, 2676-2682. Irie, M., Hata, Y., Takeuchi, M., Ichtchenko, K., Toyoda, A., Hirao, K., Takai, Y., Rosahl, T. W., and Sudhof, T. C. (1997). Binding of neuroligins to PSD-95. Science 277, 1511-1515. lessen, J. R., Topczewski, J., Bingham, S., Sepich, D. S., Marlow, F., Chandrasekhar, A., and SolnicaKrezel, L. (2002). Zebrafish trilobite identifies new roles for Stabismus in gastrulation and neuronoal movements. Nat Cell Biol 4, 610-615. Jones, K. H., Liu, J., and Adler, P. N. (1996). Molecular analysis of EMS-induced frizzled mutations in Drosophila melanogaster. Genetics 142, 205-215. Karess, R. E., Chang, X. J., Edwards, K. A., Kulkarni, S., Aguilera, I., and Kiehart, D. P. (1991). The regulatory light chain of nonmuscle myosin is encoded by spaghetti-squash, a gene required for cytokinesis in Drosophila. Cell 65, 1177-1189. Khurana, T., Khurana, B., and Noegel, A. A. (2002). LIM domains: association with the actin cytoskeleton. Protoplasma 219, 1-12. Klingensmith, J., Nusse, R., and Perrimon, N. (1994). The Drosophila segment polarity gene dishevelled encodes a novel protein required for response to the wingless signal. Genes & Development 8. Krasnow, R. E., Wong, L. L., and Adler, P. N. (1995). dishevelled is a component of the frizzled signaling pathway in Drosophila. Development 727, 4095-4102. LaJeunesse, D. R., McCartney, B. M., and Fehon, R. G. (2001). A systematic screen for dominant second-site modifiers of Merlin/NF2 phenoptypes reveals an interaction with blistered/DSRF and scribbler. Genetics 158, 667-679. Lamb, R. S., Ward, R. E., Schweizer, L., and Fehon, R. G. (1998). Drosophila coracle, a member of the protein 4.1 superfamily, has essential structural functions in the septate junctions and developmental functions in embryonic and adult epithelial cells. Mol Biol Cell 9, 3505-3519. Lee, H., and Adler, P. N. (2002). The function of the frizzled pathway in the Drosophila wing is dependent on inturned and fuzzy'• Genetics 160, 1535-1547. Lindsey, D. L., and Zimm, G. G. (1992). The genome of Drosophila melanogaster (New York, Academic Press). Luo, L., Liao, Y. J., Jan, L. Y., and Jan, Y. N. (1994). Distinct morphogenic functions of similar small GTPases: Drosophila Dracl is involved in axonal outgrowth and myoblast fusion. Genes & Developments, 1787-1802. Ma, D., Yang, C, McNeill, H., Simon, M. A., and Axelrod, J. D. (2003). Fidelity in planar polarity signalling. Nature 421, 543-547. Mackrell, A. J., Blumberg, B., Yaynes, S. R., and Fessler, J. H. (1988). The lethal myosperoid gene of Drosophila encodes a membrane protein homologous to vertebrate integrinftsubunits. Proc Natl Acad Sci USA 85, 2633-2637. 186  Manfruellin, P., Arquier, N., Hanratty, W. P., and Semeriva, M. (1996). The tumor suppressor gene, lethal(2)giant larvae (l(2)gl), is required for cell shape change of epithelial cells during Drosophila development. Development 122, 2283-2294. Mansfield, S. G., Al-Shirawi, D. Y., Ketchum, A. S., Newbern, E. C.^and Kiehart, D. P. (1996). Molecular organization and alternative slpicling in zipper, the gene that enocdes the Drosophila nonmuscle myosin II heavy chain. J Mol Biol 255, 98-109. Marfatia, S. M., Lue, R. A., Branton, D., and Chishti, A. H. (1994). In vitro binding studies suggest a membrane-associated complex between erythroid p55, protein 4.1, and glycophorin C. J Biol Chem 269, 8631-8634. Martin, D., Zusman, S., Li, X., Williams, E. L., Khare, N., DaRocha, S., Chiquet-Ehrismann, R., and Baumgartner, S. (1999). wing blister, a new Drosophila laminin a chain required for cell adhesion and migration during embryonic and imaginal development. The Journal of Cell Biology 145, 191-201. Mitchell, H. K., Roach, J., and Petersen, N. S. (1983). The morphogenesis of cell hairs on Drosophila wings. Dev Biol 95, 387-398. Mlodzik, M. (1999). Planar polarity in the Drosophila eye: a multifaceted view of signaling specificity and cross-talk. EMBO Journal 18, 6873-6879. Mlodzik, M. (2000). Spiny legs and prickled bodies: new insights and complexities in planar polarity establishment. BioEssays 22, 311-315. Mlodzik, M. (2002). Planar cell polarization: do the same mechanisms regulate Drosophila tissue polarity and vertebrate gastrulation? Trends in Genetics 18, 564-571. Morgensen, M. M., and Tucker, J. B. (1988). Intermicrotubular actin filaments in the transalar cytoskeletal arrays of Drosophila. J Cell Sci 91, 431-438. Morgensen, M. M., Tucker, J. B., and Stebbings, H. (1989). Microtubule polarities indicate that nucleation and capture of microtubules occurs at cell surfaces in Drosophila. J Cell Biol 108, 14451452. Park, M., and Moon, R. T. (2002). The planar cell-polarity gene stbm regulates cell behaviour and cell fate in vertebrate embryos. Nature Cell Biology 4, 20-25. Park, W.-J., Liu, J., and Adler, P. N. (1994a). The frizzled gene of Drosophila encodes a membrane protein with an odd number of transmembrane domains. Mech Dev 45, 127-137. Park, W.-J., Liu, J., and Adler, P. N. (1994b).,frizzled gene expression and the developmnet of tissue polarity in the Drosophila wing. Dev Genet 15, 383-389. Park, W. J., Lui, J., Sharp, E. J., and Adler, P. N. (1996). The Drosophila tissue polarity gene inturned acts autonomously and encodes a novel protein. Development 722, 961-969. Peifer, M. (2000). Cell biology. Travel bulletin - traffic jams cause tumors. Science 289, 67-69.  187  Peifer, M., and Tepass, U. (2000). Cell biology. Which way is up? Nature 403, 611-612. Peles, E., Joho, K., Plowman, G. D., and Schlessinger, J. (1997a). Close similarity between Drosophila neurexin IV and mammalian Caspr protein suggests a conserved mechanism for cellular interactions. Cell 88, 745-746. Peles, E., Nativ, M., Lustig, M., Grumet, M., Schilling, J., Martinez, R., Plowman, G. D., and Schlessinger, J. (1997b). Identification of a novel contactin-associated transmembrane receptor with multiple domains implicated in protein-protein interactions. EMBO Journal 16, 978-988. Peles, E., and Salzer, J. L. (2000). Molecular domains of myelinated axons. Curr Opin Neurobiol 10, 558-565. Peng, C.-H., Manning, L., Albertson, R., and Doe, C. Q. (2000). The tumor-supressor genes lgl and dig regulate basal protein targeting in Drosophila neuroblasts. Nature 408, 596-600. Penton, A., Wodarz, A., and Nusse, R. (2002). A mutational analysis of dishevelled in Drosophila defines novel domains in the Dishevelled protein as well as novel suppressing alleles of axin. Genetics 161, 747-762. Perrimon, N. (1988). The maternal effect of lethal( 1 )discs-large-1: a recessive oncogene of Drosophila melanogaster. Develpomental Biology 727, 392-407. Perrimon, N., and Mahowald, A. P. (1987). Multiple functions of segment polarity genes in Drosophila. Dev Biol 7/9,587-600. Ponting, C. P., and Bork, P. (1996). Plekstrin's repeat performance: A novel domain in G-protein signalling? Trends Biochem Sci 27, 245-246. Prout, M., Damania, Z., Soong, J., Fristrom, D., and Fristrom, J. W. (1997). Autosomal mutations affecting adhesion between wing surfaces in Drosophila melanogaster. Genetics 146, 275-285. Roote, C. E., and Zusman, S. (1996). Alternatively spliced forms of the Drosophila aPS2 subunit of integrin are sufficient for viability and can replace the function of the aPS 1 subunit of integrin in the retina. Development 122, 1985-1994. Schulte, J. (2003) The role of Gliotactin in the formation of Drosophila septate junctions, University of British Columbia, Vancouver. Shimada, Y., Usui, T., Yanagawa, S., Takeichi, M., and Uemura, T. (2001). Asymmetric colocalization of Flamingo, a seven-pass transmembrane cadherin, and Dishevelled in planar cell polarization. Current Biology 77, 859-863. Shulman, J. M., Perrimon, N., and Axelrod, J. D. (1998). Frizzled signaling and the developmental control of cell polarity. Trends in Genetics 14, 452-458. Songyang, Z., Fanning, A. S., Fu, C, Xu, J., Marfatia, S. M., Chishti, A. H., Crompton, A., Chan, A. C, Anderson, J. M., and Cantley, L. C. (1997). Recognition of Unique Carboxyl-Terminal Motifs by Distinct PDZ Domains. Science 275, 73-77.  188  Strand, D., Raska, I., and Mechler, B. M. (1994). The Drosophila lethal (2) giant larvae tumor suppressor protein is a component of the cytoskeleton. J Cell Biol 127, 1345-1360. Strutt, D. I. (2001a). Asymmetric localization of Frizzled and the establishment of cell polarity in the Drosophila wing. Molecular Cell 7. Strutt, D. I. (2001b). Planar polarity: getting ready to ROCK. Current Biology / / , R506-R509. Strutt, D. I. (2002). The asymmetric subcellular localisation of components of the planar polarity pathway. Semin Cell Dev Biol 13, 225-231. Strutt, D. I., Weber, U., and Mlodzik, M. (1997). The role of RhoA in tissue polarity and frizzled signalling. Nature 387, 292-295. Strutt, H., and Strutt, D. (2002). Nonautonomous planar polarity patterning in Drosophila: Dishevelledindependent functions of Frizzled. Developmental Cell 3, 851-863. Sussman, J. L., Harel, M., Frolow, F., Oefner, C, Goldman, A., Toker, L., and Silman, I. (1991). Atomic structure of acetylcholinesterase from Torpedo californica: a prototypic acetylcholine-binding protein. Science 253, 872-879. Tanentzapf, G., and Tepass, U. (2003). Interactions between the crumbs, lethal giant larvae and bazooka pathways in epithelial polarization. Nature Cell Biology 5, 46-52. Taylor, J., Abravoma, N., Charlton, J., and Adler, P. N. (1998). Van Gogh: a new tissue Drosophila tissue polarity gene. Genetics 150, 199-210. Tepass, U., Tanentzapf, G., Ward, R., and Fehon, R. (2001). Epithelial cell polarity and cell junctions in Drosophila. Annu Rev Genet 35, 747-784. Theisen, H., Purcell, J., Bennett, M., Kansagara, D., Syed, A., and Marsh, J. L. (1994). dishevelled is required during wingless signaling to establish both cell polarity and cell identity. Development 120, 347-360. Tree, D. R. P., Ma, D., and Axelrod, J. D. (2002a). A three-tiered mechanism for regulation of planar cell polarity. Semin Cell Dev Biol 13, 217-224. Tree, D. R. P., Shulman, J. M., Rousset, R., Scott, M. P., Gubb, D., and Axelrod, J. D. (2002b). Prickle mediates feedback amplification to generate asymmetric planar cell polarity signaling. Cell 709, 371381. Tsunoda, S., Sierralta, J., Sun, Y., Bodner, R., Suzuki, E., Becker, A., Socolich, M., and Zuker, C. S. (1997). A multivalent PDZ-domain protein assembles signaling complexes in a G-protein-coupled cascade. Nature 388, 243-249. Tucker, J. B., Milner, M. J., Currie, D. A., Muir, J. W., Forrest, D. A., and Spencer, M. (1986). Centrosomal microtubule-organizing centres and a switch in the control of protofilament number for cell surface-associated microtubules during Drosophila wing morphogenesis. Eur J Cell Biol 41, 279-289. Turner, C. M., and Adler, P. N. (1998). Distinct roles for the actin and microtubule cytoskeletons in the morphogenesis of epidermal hairs during wing development in Drosophila. Mech Dev 70, 181-192. 189  Usui^ T., Shima, Y., Shimada, Y., Hirano, S., Burgess, R. W., Schwarz, T. L., Takeichi, M., and Uemura, T. (1999). Flamingo, a seven-pass transmembrane cadherin, regulates planar cell polarity under the control of Frizzled. Cell 98, 585-595. Vinson, C. R., and Adler, P. N. (1987). Directional non-cell autonomy and transmission of polarity information by the frizzled gene of Drosophila. Nature 329, 549-551. Vinson, C. R., Conover, S., and Adler, P. N. (1989). A Drosophila tissue polarity locus encodes a protein containing seven potential transmembrane domains. Nature 338, 263-264. Walsh, E. P., and Brown, N. H. (1998). A screen to identify Drosophila genes required for integrinmediated adhesion. Genetics 150, 791-805. Ward, R. E., Lamb, R. S., and Fehon, R. G. (1998). A conserved functional domain of Drosophila coracle is required for localization at the septate junction and has membrane-organizing activity. J Cell Biol 140, 1463-1473. Ward, R. E., Schweizer, L., Lamb, R. E., and Fehon, R. G. (2001). The Protein 4.1, Ezrin, Radixin, Moesin (FERM) domain of Drosophila coracle, a cytoplasmic component of the septate junction, provides functions essential for embryonic development and imaginal cell proliferation. Genetics 759, 219-228. Wilcox, M., Brower, D. L., and Smith, R. J. (1981). A position-specific cell surface antigen in the Drosophila wing imaginal disc. Cell 25, 159-164. Wilcox, M., DiAntonio, A., and Leptin, M. (1989). The function of PS integrins in Drosophila wing morphogenesis. Development 707, 891-897. Willott, E., Balda, M. S., Fanning, A. S., Jameson, B., Van Itallie, C, and Anderson, J. M. (1993). The tight junction protein ZO-1 is homologous to the Drosophila discs-large tumor suppressor protein of septate junctions. Proc Natl Acad Sci USA 90, 7834-7838. Winter, C. G., Wang, B., Ballew, A., Royou, A., Karess, R., Axelrod, J. D., and Luo, L. (2001). Drosophila Rho-associated kinase (Drok) links Frizzled-mediated planar cell polarity signalling to the actin cytoskeleton. Cell 105, 81-91. Wodarz, A. (2000). Tumor suppressors: linking cell polarity and growth control. Current Biology 10, R624-R626. Wodarz, A., Hinz, U., Engelbert, M., and Kunst, E. (1995). Expression of crumbs confers apical character on plasma membrane domains of ectodermal epithelia. Cell 82, 67-76. Wolff, T., and Rubin, G. M. (1998). strabismus, a novel gene that regulates tissue polarity and cell fate decisions in Drosophila. Development 725, 1149-1159. Wong, L. L., and Adler, P. N. (1993). Tissue polarity genes of Drosophila regulate the subcellular location for prehair initiation in pupal wing cells. The Journal of Cell Biology 123, 209-221.  190  Woods, D. F., and Bryant, P. J. (1991). The discs-large tumor suppressor gene of Drosophila encodes a guanylate kinase homolog localized at septate junctions. Cell 66, 451-464. Woods, D. F., Hough, C, Peel, D., Callaini, G., and Bryant, P. J. (1996). Dig protein is required for junction structure, cell polarity, and proliferation control in Drosophila epithelia. J Cell Biol 134, 14691482. Woods, D. F., Wu, J. W., and Bryant, P. J. (1997). Localization of proteins to the apico-lateral junctions of Drosophila epithelia. Dev Genet 20, 111-118. Young, P. E., Richman, A. M., Ketchum, A. S., and Kiehart, D. P. (1993). Morphogenesis in Drosophila requires nonmuscle myosin heavy chain function. Genes Dev 7, 29-41. Yun, U. J., Kim, S. Y., Lui, J., Adler, P. N., Bae, E., Kim, J., and Park, W. J. (1999). The Inturned protein of Drosophila melanogaster is a ctyoplasmic protein located at the cell periphery in wing cells. Developmental Genetics 25, 297-305. Zeidler, M. P., Perrimon, N., and Strutt, D. I. (2000). Multiple roles for four-jointed in planar polarity and limb patterning. Developmental Biology 228, 181-196. Zeng, L., Fagatto, F., Zhang, T., Hsu, W., Vasicek, T. J., Perry III, W. L., Lee, J. J., Tilghman, S. M., Gumbiner, B. M., and Costantini, F. (1997). The mouse Fused locus encodes Axin, an inhibitor of the Wnt signaling pathway that regulates embryonic axis formation. Cell 90, 181-192. Zusman, S., Patel-King, R. S., Ffrench-Constant, C., and Hynes, R. O. (1990). Requirements for integrins during Drosophila development. Development 108, 391-402.  191  A P P E N D I X I. DROSOPHILA  GENOTYPES.  192  Gliotactin alleles  Abbreviation Locus name  Comments  Reference  Gli  Gliotactin  viable P-element insertion into 5' untranslated region, carries white* marker  (Auld etal., 1995)  Gliotactin  null, imprecise P-element excision of the Gli " P-element, 5' untranslated region, white* marker deleted  (Auld etal., 1995)  strong hypomorph, EMS-induced point mutation induced on the Gli chromosome, premature stop codon within extracellular serine esterase-like motif (W454A)  Chapter II  AE2  AE2A45  GU  Ar  Gli  lhl  Gliotactin  AE1  Gli' '  h 3  Gliotactin  null EMS-induced point Chapter II mutation induced on the Gli chromosome, premature stop codon prior to the serine esteraselike motif (Q203A) AE1  Gli'  h0  Gliotactin  hypomorph, EMSinduced point mutation induced on t\\tGli chromosome, nonconservative substitution within the serine esteraselike motif (G525E)  Chapter II  AE2  Gli  P34  Gliotactin  (Ashburner et al., 1990) null, EMS-induced point mutation, premature stop codon prior to the serine esterase-like motif (Q108A)  193  Gliotactin alleles (continued) Abbreviation Locus name  Comments  Reference  Gli  hypomorph, EMS-  (Ashburner et al., 1990)  Gliotactin  induced point mutation, premature stop codon after the transmembrane domain (S820A)  194  Septate junction and SJ-interacting mutations Abbreviation Locus name  Comments  Reference  cor  15  coracle  EMS- induced hypomorph  (Lambetal., 1998)  cor"  coracle  EMS- induced null  (Lambetal., 1998)  dig"  discs-large  null allele, contains premature stop codon, truncated protein remains cytoplasmic  (Woods et al., 1996)  02  Nrx "  Neurexin IV EMS- induced hypomorph  (Baumgartner et al., 1996)  Nrx'"''  Neurexin IV EMS- induced null  (Baumgartner et al., 1996)  25  scrib' " 7  scribble  strong hypomorph induced by P-element insertion into noncoding transcribed region  (Bilder and Perrimon, 2000)  vulcan  strong hypomorph induced by P-element insertion  (Gates and Thummel, 2000)  195  Visible marker mutations Abbreviation Locus name  Comments  Reference  black  recessive body color marker, homozygotes have dark adult cuticle, distal to Gli on 2L  (FlyBase, 2003)  cn  cinnabar  recessive eye color marker, homozygotes have light orange eyes, requires white* to score, on 2R  (FlyBase, 2003)  Cy  Curly  dominant wing morphology marker, on chromosome 2, heterozygotes have curled wings, present on CyO balancer  (FlyBase, 2003)  lacW  P-element  white* transgene used to marker P-element insertions in a w"" background  (FlyBase, 2003)  dominant eye morphology marker, on chromosome 2, heterozygotes have fused ommatidia  (FlyBase, 2003)  P-element with green fluorescence protein expressed under the control of actin regulatory sequences, inserted onto CyO  (FlyBase, 2003)  1  Gla  Glazed  P{actin-GFP} P-element  196  Visible marker mutations (continued) Abbreviation Locus name  Comments  Reference  Sp  Sternopleural dominant bristle morphology (FlyBase, 2003) marker, on chromosome 2, heterozygotes have duplicated sternopleural bristles  Sb  Stubble  dominant bristle marker on chromosome 3, heterozygotes have shortened thoracic bristles, present on TM3 balancer  (FlyBase, 2003)  Tb  Tubby  dominant larval marker on chromosome 3, larvae are shortened relative to wild-type, present on TM6  (FlyBase, 2003)  w"' (or w) 8  white  recessive eye color marker on the X chromosome, homozygotes and hemizygotes have white eyes, epistatic to cn, complemented by a lacW P-element marker such as Gli  (FlyBase, 2003)  recessive cuticle color marker on the X chromosome, homozygotes and hemizygotes have yellow adult cuticle  (FlyBase, 2003)  AL2  y  yellow  197  Balancer chromosomes Name  Chromosome  Dominant marker(s)  Reference  YDp{y\ dig*}  1  y in a y background, (FlyBase, 2003) has a duplication of dig* from the X on the Y, used to complement dig mutants in males  CyO  2  Cy  (FlyBase, 2003)  CyO(actin-GFP)  2  Cy, lacW, P{actin-GFP}  (FlyBase, 2003)  TM3  3  Sb  (FlyBase, 2003)  TM6  3  Tb  (FlyBase, 2003)  +  198  Planar cell polarity loci Abbreviation Locus name  Comments  Reference  dsh  point mutation within the DEP domain, specific for the winglessindependent PCP pathway  (Axelrod, 2001)  spontaneous allele specific to Pk transcript  (Gubb etal., 1999)  1  pk'  dishevelled  prickle  pk  199  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items