Functional expression of rat synapse-associated proteins SAP97 and SAP102 in Drosophila dlg-1 mutants: effects on tumor suppression and synaptic bouton structure

Mechanisms of Development 62 (1997) 161–174

Functional expression of rat synapse-associated proteins SAP97 and SAP102 in Drosophila dlg-1 mutants: effects on tumor suppression and synaptic bouton structure Ulrich Thomas a ,*, Bounpheng Phannavong a, Bettina Mu¨ller b, Craig C. Garner c, Eckart D. Gundelfinger a a

Federal Institute for Neurobiology, Department of Neurochemistry and Molecular Biology, P.O. Box 1860, D-39008 Magdeburg, Germany b Center for Molecular Neurobiology, University of Hamburg, D-20246 Hamburg, Germany c Neurobiology Research Center, University of Alabama at Birmingham, 1719 Sixth Avenue S., Birmingham, AL 35213-0021, USA Received 26 August 1996; revised version received 23 December 1996; accepted 6 January 1997

Abstract The synapse-associated proteins SAP97 and SAP102 are mammalian proteins that are structurally related to the Drosophila tumor suppressor protein DlgA. Previous analyses revealed that DlgA is essential for the integrity of epithelia and neuromuscular synapses. Here we show that synaptic bouton structure is severely affected in mutant larvae carrying the dlg-1XI-2 allele. We have tested SAP97 and SAP102 for functional homology to DlgA by heterologous expression in Drosophila. Both SAP97 and SAP102 can suppress tumor formation in dlg-1 mutant flies and mimic DlgA at larval neuromuscular junctions. Neuronal expression of SAP97 or SAP102 is required for morphological restoration of synaptic boutons, indicating that presynaptic DlgA function is essential for establishing structurally intact motor nerve terminals at larval neuromuscular junctions.  1997 Elsevier Science Ireland Ltd. Keywords: Membrane-associated guanylate kinases (MAGUKs); discs large; dlg-1 mutant; Tumor suppression; Neuromuscular junction; Synapse

1. Introduction Various approaches towards the molecular characterization of cellular junctions have recently led to the identification of several membrane-associated guanylate kinase homologs (MAGUKs; for review see Kim, 1995) Mammalian MAGUKs include the synapse-associated proteins PSD-95/SAP90 (Cho et al., 1992; Kistner et al., 1993), SAP97/hdlg (Lue et al., 1994; Mu¨ller et al., 1995), SAP102 (Mu¨ller et al., 1996) and PSD-93/Chapsyn-110 (Brenman et al., 1996a; Kim et al., 1996) as well as the tight junction proteins ZO-1 and ZO-2 (Itoh et al., 1993; Willott et al., 1993; Jesaitis and Goodenough, 1994). These molecules are structurally related to DlgA, the Drosophila prototype of this protein family (Woods and Bryant, 1991), i.e. they share a common three-partite domain * Corresponding author. Tel.: +49 391 6263224/6263202; fax: +49 391 6263229; e-mail: [email protected]

organization: three so-called PDZ-domains (also denoted as GLGF- or DHR-repeats) are followed by a SH3-domain and a region with similarity to yeast guanylate kinase (reviewed by Woods and Bryant, 1993a; Kim, 1995) (see also Fig. 1). A pivotal role of MAGUKs in the structural and functional organization of membrane specializations is indicated by several lines of evidence. For instance, various neuronally expressed MAGUKs can bind via their first two PDZ domains to neuronal NO-synthase (Brenman et al., 1996a; Brenman et al., 1996b) and to the cytoplasmic tails of NMDA-type glutamate receptors (Kornau et al., 1995; Mu¨ller et al., 1996; Niethammer et al., 1996) and Shakertype potassium channels (Kim et al., 1995); this binding may result in clustering of the respective membrane proteins (Kim et al., 1995, 1996). At tight junctions (TJs) ZO1 associates with the transmembrane protein occludin (Furuse et al., 1994) and with ZO-2 (Jesaitis and Goodenough, 1994). Moreover, regulated interactions of ZO-1

0925-4773/97/$17.00  1997 Elsevier Science Ireland Ltd. All rights reserved PII S0925-4773 (97 )0 0658-8

162

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174

Fig. 1. Comparative representation of DlgA, SAP97 and SAP102. The three proteins share five well conserved domains: three PDZ domains, a SH3 domain and a guanylate kinase-like domain (GuK). A 33aa sequence designated I3, which is subject to alternative splicing in SAP97 and hdlg (Lue et al., 1994; Mu¨ller et al., 1995), is present in DlgA and SAP97 but not in SAP102. Regions outside these domains do not show any significant sequence similarity.

with catenins have recently been shown to be involved in the assembly of TJs (Rajasekaran et al., 1996). Genetic and phenotypic analyses of the Drosophila tumor suppressor locus l(1) discs-large-1 (dlg-1), the gene encoding DlgA, have revealed functions for this gene in epithelia and at synapses. Removal of both maternal and zygotic dlg-1 expression results in embryonic lethality with severe defects in the epidermis and the developing nervous system (Perrimon, 1988; Woods and Bryant, 1991). Most strikingly, however, loss of zygotic DlgA function alone causes neoplastic overgrowth of imaginal disc epithelia and hyperplastic growth within larval brains (Stewart et al., 1972; Gateff, 1978; Woods and Bryant, 1989). Apico-basal polarity of epithelial cells is largely lost in the course of these tumorigenic events (Woods and Bryant, 1989; Abbott and Natzle, 1992; Woods et al., 1996). The continuous cell proliferation leads to a prolonged larval period which ends with the formation of pseudopupae and subsequent death. Furthermore, dlg-1 mutant larvae exhibit ultrastructural defects at the subsynaptic reticulum of certain neuromuscular junctions (NMJs), consistent with the localization of DlgA at these postsynaptic membrane specializations in the wildtype situation (Lahey et al., 1994; Budnik et al., 1996). In epithelia DlgA exerts its tumor suppressive function at septate junctions (Woods and Bryant, 1991; Woods et al., 1996), which are thought to be functional analogs of vertebrate TJs (Green and Bergquist, 1982; Wood, 1990; for a detailed description of Drosophila cellular junctions see Tepaß and Hartenstein, 1994). The similarity between ZO-1 and DlgA supports this view. PSD-95/SAP90, SAP97/hdlg and SAP102, however, are structurally more closely related to DlgA than ZO-1 or ZO-2. Moreover, similar to DlgA, SAP97/hdlg exhibits a widespread tissue distribution. In epithelial cells SAP97 has been found in association with lateral membranes, basal to adherens junctions (Mu¨ller et al., 1995); in neurons SAP97 has

been localized at presynapses and within non-myelinated axons. Hdlg, the human counterpart of SAP97, binds to the cortico-cytoskeletal protein 4.1 (Lue et al., 1994, 1996; Marfatia et al., 1996), an observation that parallels the co-localization of DlgA and Coracle, a Drosophila homolog of protein 4.1, at septate junctions (Fehon et al., 1994). In contrast to SAP97/hdlg and DlgA, PSD-95/SAP90 and SAP102 have been detected exclusively in neuronal tissues. While SAP102 has been localized to postsynaptic structures (Mu¨ller et al., 1996), PSD-95/SAP90 is found pre- and postsynaptically (Cho et al., 1992; Kistner et al., 1993; Hunt et al., 1996) as well as at septate-like junctions in rat cerebellar Pinceaux (Laube et al., 1996). The sequence homology on one hand and the differences in overall and subcellular distribution on the other hand raises the question whether the various mammalian MAGUKs are functionally equivalent and to what extent DlgA function has been conserved in these proteins. This prompted us to test whether SAP97 and SAP102, the closest rat relatives of DlgA, can suppress the dlg-1 mutant phenotype. In fact, single splice variants of both proteins exhibit rescue activity with respect to both tumor suppression and synaptic bouton structure. This functional substitution is accompanied by subcellular distributions that largely mimic that of DlgA. In the course of this study we have also elucidated additional aspects of DlgA function; most notably, DlgA function was found to be important for the generation of morphologically and molecularly intact motor nerve endings.

2. Results 2.1. Generation of rat SAP97- and SAP102-expressing flies The rat proteins SAP97 and SAP102 share about 72%

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174

sequence identity and are both about 60% identical to DlgA (Fig. 1). Splice variation has been described for both SAP97 and SAP102 (Mu¨ller et al., 1995; Mu¨ller et al., 1996). The splice variants of SAP97 and SAP102 that were analyzed in this study differ with respect to the I3insertion between the SH3- and the GuK-domain. This splice domain, which has been shown to bind protein 4.1 (Lue et al., 1994; Lue et al., 1996; Marfatia et al., 1996), is present in the SAP97- but not in the SAP102-variant (Fig. 1). To test whether these proteins can functionally replace DlgA, we wanted to express them in a dlg-1 mutant background. We have therefore generated transgenic flies carrying cDNAs comprising the complete open reading frames for SAP97 or SAP102; in order to allow a regulated GAL4-driven expression, the cDNAs were placed downstream of UAS-elements (Brand and Perrimon, 1993). More than ten independent transgenic lines were obtained for each construct. Since DlgA is expressed in many tissues and at all developmental stages, we sought a GAL4 activator strain that could mimic this widespread expression. Strains carrying the GAL4 gene downstream of a promoter fragment from the daughterless gene, denoted as daG (Wodarz et al., 1995), proved to be well suited for this purpose. As visualized upon mating the activator strain to UAS-lacZ reporter strains, daG-mediated expression was detectable from embryonic stage 10 onwards throughout development and in tissues derived from all three germlayers. At all stages exceptionally high levels of expression were found in salivary glands and in the gut epithelium (Fig. 2A). A uniform but weak expression was detected in the developing mesoderm starting from stage 11. In the epidermis b-galactosidase activity was not apparent before stage 13. Expression levels increased in all tissues towards the end of embryogenesis. In the context of the present study several features of the daG-mediated expression should be considered: (i) Strong expression was found in leg, haltere and wing imaginal disc epithelia, i.e. in those tissues which are extensively tumorous in dlg-1 mutants. However, expression is not completely uniform in these tissues. As depicted in Fig. 2C, b-galactosidase-positive areas are interspersed with patches of non-expressing cells. These mosaics were obviously irregular and also appeared in other tissues. (ii) In body wall muscles increasing levels of expression were observed from late embryogenesis to 2nd instar larvae and considerable expression is maintained until the end of the 3rd larval stage (Fig. 2D). (iii) In the embryonic CNS, b-galactosidase was barely detectable up to stage 16 and expression was still low at the end of embryogenesis (Fig. 2A; compare with elavGAL4 driven lacZ expression in Fig. 2B). In larval brains b-galactosidase activity showed a dynamic pattern. From 1st to 3rd instar an increasing number of cells in the optic lobe anlagen showed enzyme activity. In the ventral nerve cord (VNC) of 1st and 2nd instar larvae expression was mainly restricted to cells in the ventral region (Fig. 2E), i.e.

163

beneath the layer where motoneuronal cell bodies are located. It is only at the 3rd instar that increasing numbers of cells in the VNC exhibit b-galactosidase activity (Fig. 2E). 2.2. Rescue of the dlg-1 mutant tumor phenotype In the rescue experiments we have primarily used the dlg-1XI-2 allele that has been described as a null allele by Perrimon (1988) and as a strong hypomorphic allele by Woods et al. (1996). Female flies carrying the balanced dlg-1XI-2 allele in combination with a UAS-SAP97 or UASSAP102 effector construct were mated to males contributing the daG activator (Fig. 3). Both SAP97 and SAP102 exhibited clear rescue activity with respect to the tumor phenotype of dlg-1XI-2 mutants in both imaginal discs and larval brains. The degree of rescue varied between the tested effector lines (reflecting position-specific effects of the genomic integration site) but was fairly constant for each given line (Fig. 4A–C and Table 1). Most strikingly, viable adult dlg-1 mutants expressing the correct genetic markers, i.e. yellow body color, singed bristles and normal eye size (Fig. 3), were obtained largely at mendelian numbers for eight of the SAP102 effector lines (Table 1). SAP97-bearing lines were less efficient in these rescue crosses than SAP102 effector lines. Viable mutant adults, however, were obtained when crosses were performed to express two copies of UAS-SAP97, indicating that the extent of rescue parallels the level of SAP97 expression and that there is no principal limit in the ability of SAP97 to replace DlgA as a tumor suppressor. In agreement with this notion, expression of only one copy of UAS-SAP97 was sufficient to obtain viable males hemizygous for hypomorphic alleles of dlg-1, i.e. dlg-1v55 and dlg-1HF321 (Perrimon, 1988). It should be noted that viable rescued males were sterile and unable to fly. 2.3. Rescue of dlg-1 mutants is associated with cell death Monitoring the daG-mediated activation of a lacZ reporter gene in imaginal discs revealed that not all cells express the target gene (Fig. 2C). Therefore, one should expect only partial rescue. However, many of the rescued flies lack any sign of tumors, whereas others do in fact still exhibit patches of tumor-like outgrowth, in particular on their legs or wings (Fig. 4D–G). Previous clonal analyses have revealed that only rather small dlg-1 mutant clones can survive and give raise to hyperplasia reminiscent to those described here (Woods and Bryant, 1991). On the other hand, the dlg-1 mutation does not always cause cell death as is obvious from the tumors. One explanation for this apparent contradiction could be that dlg-1 mutant cells become subjected to cell death if intact neighboring cells are present to induce apoptosis. In the rescued flies such apoptotic events could account for occasional phenotypes

164

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174

165

Fig. 3. Rescue genetics, exemplified for 3rd chromosomal UAS-SAP97 line 194.19 and daG-activator line 39. The genotype of the I. and III. chromosomes of rescued male progeny is emphasized by shading.

such as reduced numbers of sex comb bristles, notches and holes in wing blades (Fig. 4D) and different sizes of wings or eyes on one and the same animal. To provide further support for cell death playing a beneficial role in the rescue experiment, we tested whether a reduction of the apoptosis-promoting genes reaper (rpr) and head involution defective (hid) (White et al., 1994; Grether et al., 1995) (Fig. 5A) affects the degree of rescue. No significant reduction of viability of SAP102-rescued males heterozygous for one of more than 20 other lethal mutations tested was observed (data not shown). In contrast, heterozygosity for Df(3L)H99, a small deletion which abolishes both rpr and hid, caused an 80% reduction in the number of viable rescued males (Fig. 5B) and those that reached adulthood were all found to bear tumors (Fig. 5C). 2.4. Expression of SAP97 and SAP102 in epithelial cells In epithelia DlgA has been localized to septate junctions (Woods and Bryant, 1991; Bryant et al., 1993). SAP97 is associated with the cytoskeleton underlying the lateral membranes between rat epithelial cells (Mu¨ller et al., 1995), whereas SAP102 has not been found in epithelia (Mu¨ller et al., 1996). To elucidate whether the tumor-sup-

pressive activity of the rat proteins goes along with a specific subcellular localization, epithelial tissues of rescued larvae were analyzed for SAP97 or SAP102 immunoreactivity. In sections of imaginal discs and in salivary gland cells SAP97 was detectable subapically, highly concentrated within a short range of the lateral membrane (Fig. 6A,E). In imaginal disc epithelia this became more obvious when SAP97 distribution was compared to that of the apical marker protein Crumbs (compare Fig. 6A and Fig. 6C). Imaginal discs from larvae that did not express SAP97 served as controls to confirm the specificity of the antibody (Fig. 6D). The subapical staining is consistent with an association of the protein with septate junctions. In support of this notion anti-DlgA antisera yielded very similar staining results in salivary glands (Woods et al., 1996) (compare also Fig. 6E and Fig. 6F). SAP102 immunoreactivity appeared somewhat more diffuse, but a subapical enrichment was also detectable (Fig. 6B,G). In order to verify if both subapical and diffuse immunoreactivity were specifically associated with SAP102 expression a ptc-GAL4 activator which drives target gene expression along the anterior-posterior border (Hinz et al., 1994) was employed to activate UAS-SAP102. This resulted in restriction of immunoreactivity to cells addressed by ptc-GAL4 (Fig. 6G).

Fig. 2. Expression of UAS-lacZ upon activation by daG39 (A,C–E) or elav-GAL4 (B). (A) Stage 16 embryo, ventro-lateral view. Note the strong staining in salivary glands (SG), the proventriculus (PV) and in the midgut epithelium (MG). The ventral nerve cord (VNC) is not visible due to barely detectable levels of enzyme activity. (B) Stage 16 embryo, lateral view. This embryo was stained in parallel to the one shown in (A); staining in the CNS is clearly visible due to activation of UAS-lacZ in postmitotic neurons through elav-GAL4. (C) Wing imaginal disc from a 3rd instar larva. Expression is high, but excluded from distinct groups of cells. (D) Body wall muscles from a 3rd instar larva. The present study has mainly focused onto NMJs at muscles 6/7. (E) Larval CNS, dissected from different stages: L1, L2, 1st and 2nd larval stages, respectively, lateral view; L3, 3rd larval stage, dorsal view. Note that expression in the optic lobe anlagen (OLA) is high during L1 and L2 and that expression in the VNC is mainly restricted to the ventral area; some stained cells visible in dorsal region of the L1 VNC belong to the midline. Such midline cells are still visible in L3. Scale bars: 100 mm in (C,D) and L3 in (E); 50 mm for L1, L2 CNS in (E).

166

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174

2.5. dlg-1XI-2 mutants exhibit abnormal NMJs Besides tumor suppression, another recently characterized function of DlgA concerns the ultrastructural organization of a subset of larval NMJs (Lahey et al., 1994; Guan et al., 1996). As employed previously, we used anti-HRP antibodies as a neuronal surface marker (Jan and Jan, 1982) to monitor defects at motor nerve endings. The distribution of HRP immunoreactivity was severely affected at NMJs of dlg-1XI-2 mutant larvae, i.e. virtually none of the

large type Ib boutons exhibited the typical smooth and round appearance. Instead, many boutons exhibited a bloated shape and frequently anti-HRP staining intensity was markedly reduced and irregular (compare Fig. 7A and Fig. 7B) (for a description of the various types of boutons see Johansen et al., 1989; Keshishian et al., 1993). Very similar results were obtained when antibodies against the synaptic vesicle protein synaptotagmin (Littleton et al., 1993) were used for immunohistochemistry (Fig. 7C,D). These clear presynaptic defects deviate from the more

Fig. 4. Phenotypical series of dlg-1XI-2 mutant and SAP97-rescued flies. (A) Final stage of a dlg-1XI-2 hemizygous giant larva; puparium formation has taken place but no signs of adult structures are detectable. (B) Partially rescued mutant: dlg-1XI-2 / Y; daG39 / UAS-SAP97194.6. This specimen has entered metamorphosis and was removed from its pupal case; the adult division into head, thorax and abdomen is visible. (C) Adult but poorly vital mutant: dlg1XI-2 / Y; daG39 / UAS-SAP97194.19. (D) Wing of a rescued, viable hemizygous male: dlg-1XI-2 / Y; daG39, UAS-SAP97194.19 / daG39, UAS-SAP97194.19. Small arrows mark patches of tumorous growth. At the same time loss of tissue is obvious at the posterior wing margin (arrowhead) and from a hole in the wing blade (large arrow). (E,F) Higher magnifications of tumors which preferentially arise at wing veins. (G) Outgrowth on a prothoracic leg. In agreement with previous clonal analyses of dlg-1XI-2 (Woods and Bryant, 1991) these hyperplasiae could generate hair-like structures (arrow) as cuticular differentiations.

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174 Table 1 Assessment of rescue activity of SAP97 and SAP102 transgenic linesa Number of lines promoting rescue to

UAS-SAP97 (1 copy) UAS-SAP97 (2 copies) UAS-SAP102

Giant larvae/ Pupaec pseudopupaeb

Pharate Adults adultsd

1 – –

5 – 3

7 – 1

– 4e 8

167

reactivity were basically devoid of presynaptic HRP antigen, again illustrating the incomplete rescue and demonstrating that the specific postsynaptic localization of SAPs does not depend on intact presynaptic nerve terminals. Interestingly, the flight inability of adult males rescued by daG-mediated expression of either SAP97 or SAP102 could be restored when elav-GAL4 was used in addition to daG to express SAP102, again demonstrating that the rat protein complements neuronal functions of DlgA.

a

For each line the number of rescued individuals that become arrested at the indicated developmental stage or became viable adults exceeded 70% of the calculated mendelian number. b Compare Fig. 4A. c Compare Fig. 4B. d Compare Fig. 4C. e Four tested lines that yielded pharate adults with one UAS-SAP97 copy gave rise to viable flies when homozygous or transheterozygous.

subtle alterations in HRP immunoreactivity reported for the other alleles of dlg-1 (Lahey et al., 1994; Budnik et al., 1996; Guan et al., 1996). 2.6. SAP97 and SAP102 mimic DlgA at NMJs Interestingly, when SAP97 or SAP102 were driven by daG, the NMJ presynaptic phenotype as judged by antiHRP staining was not restored to an appreciable extent in the otherwise rescued larvae (Fig. 7E). The bouton structure, however, appeared normal when the neuron-specific elav-GAL4 activator (Luo et al., 1994) was used instead of daG to drive expression in a dlg-1 mutant background (Fig. 7F,G). This finding indicates that (i) both SAP97 and SAP102 can replace DlgA, permitting type Ib boutons to acquire a normal morphology (concomitantly demonstrating that the defects are not due to second site mutations on the dlg-1XI-2-bearing X-chromosome), and (ii) the restoration is independent of postsynaptic expression of the SAPs in the innervated muscle. At type I boutons of 3rd instar larvae DlgA was localized to the subsynaptic reticulum and to a lesser extent to the presynaptic terminal (Lahey et al., 1994). Consistent with these results, daG-driven SAP97 and SAP102 were found around all type Ib boutons and with less intensity around type Is boutons (Fig. 7H), whereas no immunoreactivity was observed at type II and III boutons. In the dlg-1XI-2 mutant background both proteins also became localized to type I boutons upon daG activation. As judged from the distribution of SAP97 (Fig. 7H) and SAP102 (not shown) immunoreactivities the shape of some boutons appeared almost, though not completely, normal (five out of 54 analyzed NMJs at muscles 6/7). Most of the NMJs, however, exhibited arrays of irregularly shaped boutons despite considerable accumulation of SAP immunoreactivity. Double label confocal microscopy with antibodies against HRP and SAP97 revealed a principally postsynaptic localization of SAP97 (Fig. 7I). At the same time many boutons surrounded by SAP97 immuno-

3. Discussion A number of mammalian MAGUKs have been denoted as DlgA homologs on the basis of sequence similarity. However, even high degrees of sequence identity do not necessarily reflect functional equivalence. The Rac/Rho/ cdc42 subfamily of ras-like GTPases provides an example for this notion (for review see Chant and Stowers, 1995; Mackay et al., 1995). A common feature of the DlgA-like proteins is their close association with specializations of the plasma membrane. On the other hand, the mammalian members of this protein family exhibit remarkable differences with respect to both cell type specificity of expression and localization at particular kinds of membrane specialization. The data shown here demonstrate that, despite such differences, various functions of DlgA have been conserved in the rat MAGUKs SAP97 and SAP102. 3.1. Rat SAP97 and SAP102 function in a heterologous system at heterologous membrane specializations The mammalian MAGUKs SAP97 and SAP102 are able to replace DlgA at septate junctions, i.e. they become localized and act in a host-specific manner. This suggests that both proteins interact with native binding partners of DlgA. Besides DlgA the cell adhesion protein Fasciclin III and the cytocortical protein Coracle (Woods and Bryant, 1993b; Fehon et al., 1994) have been documented to specifically localize at septate junctions. The latter is a structural homolog of protein 4.1, which can bind via its Nterminal half (which exhibits the highest degree of conservation with Coracle) to the PDZ/DHR-repeats and to the I3-insertion domain of the human counterpart of SAP97, hdlg (Lue et al., 1994; Lue et al., 1996; Marfatia et al., 1996). Interestingly, the I3 domain was found to be dispensable in SAP102 for tumor suppression and restoration of the bouton phenotype in our in vivo assay. The absence of this domain could, however, account for the more dispersed distribution of SAP102 as compared to that of SAP97 (Fig. 6A,B). No genetic interaction between coracle and dlg-1 could be monitored in our experiments, i.e. the degree of rescue was unaffected by removal of one intact copy of coracle (i.e. heterozygosity for the allele cor1 or cor2; Fehon et al., 1994; data not shown).

168

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174

Fig. 5. Df(3L)H99 as a modifier of SAP102-mediated tumor suppression. (A) Genetic set-up to monitor the effect of H99 on the rescue. In principle, any other 2nd or 3rd chromosomal mutation instead of H99 can be tested for modification following this crossing scheme. The female contributes all genetic components that are needed for the rescue, i.e. the dlg-1 mutation, the UAS-SAP102 effector and the daG39 activator. (B) Male progeny, counted from three crosses. The number of viable dlg-1 mutant males bearing the H99 deficiency is clearly reduced as compared to their siblings. (C) All 17 survivors exhibited massive tumors; a wing of one such male is shown and only part of the numerous tumors are indicated by arrows.

The DlgA-like distribution of both rat proteins at NMJs again reflects their ability to share direct binding partners with DlgA. In light of recent findings that mammalian DlgA-like MAGUKs (including PSD-95/SAP90, SAP97

and SAP102) can physically interact with NMDA receptor subunits and Shaker-type potassium channels (Kim et al., 1995; Kornau et al., 1995; Mu¨ller et al., 1996; Niethammer et al., 1996), such binding partners may

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174

169

Fig. 6. Immunolocalization of transgenic SAP97 and SAP102 in larval epithelia. (A–D) Longitudinal 8-mm cryosections of imaginal discs; arrows mark the cleft where the apical sides of the columnar disc epithelium and the peripodial membrane face each other. (A) Disc from a SAP97-rescued 3rd instar larva, immunostained against SAP97. The restricted localization of SAP97 to the subapical regions within the columnar epithelium and the peripodial membrane results in two parallel bands of immunoreactivity. (B) SAP102 immunoreactivity in a SAP102-rescued disc. Although enriched subapically, staining is also detectable more basically. (C) Immunostaining for the apical marker Crumbs (Tepaß et al., 1990) in a wild-type disc is shown for comparison. (D) Control staining for SAP97 in a non-expressing disc epithelium from a UAS-SAP97194.19 larva. (E) SAP97 immunoreactivity in the salivary gland epithelium, demonstrating that the protein is associated with the plasma membrane (arrows); again immunoreactivity is restricted to subapical regions. Arrowheads mark the luminal face of epithelial cells (out of focus) which is devoid of immunoreactivity. (F) DlgA immunoreactivity in the salivary gland from a larva slightly younger than the one in (E). Again staining is restricted to a subapical area of the lateral membrane. (G) SAP102 expression in a cross-sectioned wing imaginal disc upon activation through ptc-GAL4 activator 559.1 (Hinz et al., 1994); restriction of staining to the ptcspecific expression domain demonstrates the specificity of the antibody. As in (B) staining is visible from apical to basal. Scale bars: 10 mm in (A–D), 20 mm in (B) and 50 mm in (F).

include glutamate receptors, the principal excitatory neurotransmitter receptors at insect NMJs, and voltage-gated potassium channels. The successful rescue is consistent with the finding that multiple mammalian MAGUKs can share binding partners in vitro and, by extending this observation towards the more distantly related DlgA, demonstrates that this interchangeability is also valid with respect to physiological functions beyond direct binding capabilities. 3.2. DlgA and mammalian MAGUKs in tumor suppression In addition to DlgA several of the Drosophila tumor suppressor genes that have been identified upon mutational analysis encode components of the submembranous cytoskeleton (e.g. Boedigheimer and Laughon, 1993; Boedigheimer et al., 1993; Strand et al., 1994). A number of cytocortical proteins have also been demonstrated to

exert tumor suppressive activity in vertebrates (for review see Tsukita et al., 1993). Our results provide evidence that the MAGUKs SAP97 and SAP102 belong to this kind of tumor suppressors in mammals. In this context it is highly suggestive that the product of the human adenomatous polyposis coli tumor suppressor gene (APC) binds to hdlg/SAP97 (Matsumine et al., 1996). Woods and Bryant (1989) have worked out that, prior to overproliferation, loss of DlgA function is associated with an increased level of cell death within imaginal discs. In our experiments such cell death is likely to contribute to the successful rescue by overcoming the leakiness of the daG activator. Consequently, we have found that the rescue is sensitive to a reduction of the apoptosis-promoting genes hid and rpr. This sensitivity may also apply to further, yet unidentified loci, that are involved in apoptosis or other events that are of relevance for tumorigenesis. It could therefore contribute to make the multi-step process

170

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174

of carcinogenesis accessible to genetic analysis in Drosophila. 3.3. dlg-1XI-2 mutants exhibit a presynaptic phenotype that can be rescued by SAP97 and SAP102 A clear-cut observation illustrates a physiological role of DlgA in the nervous system that can be mimicked by rat MAGUKs: the neuron-specific activator elav-GAL4 supplements daG to restore flight ability of rescued mutants. The careful analysis of the dlg-1 mutant phenotype at type I glutamatergic NMJs has revealed a function for DlgA in the structural organization of the subsynaptic reticulum (Lahey et al., 1994; Budnik et al., 1996; Guan et al., 1996). Using various antibodies, we have found that there are also abnormalities concerning the presynaptic motor nerve terminals, including a fairly irregular and often bloated appearance of varicosities, which is visible at the light microscopic level. Since the NMJ phenotype could be rescued by elav-GAL4-driven SAP97 or SAP102, it cannot be attributed to second site mutations on the XI-2-bearing X-chromosome. This constellation also allows to conclude that the deformed NMJs are not a secondary effect of the tumorigenic events (or of prolonged larval life associated with it), since this rescue effect was observable in giant larvae. Moreover, since the presynaptic defect was present in mutant third instar larvae that developed in a normal time course upon daG-mediated SAP expression, it must have been established up to this stage. It should be noted that, despite these defects, the NMJs must be at least partially active in neurotransmission, allowing the larvae to move. The difference in the severity of the phenotype described here and the one described previously may simply be explained by the different alleles that were used in either study. For instance, in the v59 allele, that was used by Lahey et al. (1994), a small deletion leads to a disruption of the open reading frame in the guanylate kinase-like domain (Woods and Bryant, 1991). Thus, the conceptual gene product of this allele is a truncated version of DlgA that still carries the PDZ and SH3 domains and might therefore still exhibit some activity, e.g. binding to various proteins. The XI-2 allele has been classified as either a null allele (Perrimon, 1988) or as a strong hypomorphic allele

171

(Woods et al., 1996). The gene product of the XI-2 allele has recently been reported to be a truncated version that still bears all three PDZ domains and might therefore be partially active (Woods et al., 1996). In addition, however, it has also been documented that the transcript level is very low in dlg-1XI-2 mutant larvae (Woods and Bryant, 1989). Presynaptic expression of SAPs via elav-GAL4 SAPs was sufficient to restore type I boutons with respect to size, shape and HRP-immunoreactivity. It therefore appears unlikely that the here-described presynaptic phenotype simply constitutes a secondary effect caused by postsynaptic defects. This finding is in agreement with the ability of presynaptically activated UAS-dlg-1 to exert rescue activity in various dlg-1 alleles. Notably, presynaptic complementation ameliorates the postsynaptic phenotype (Budnik et al., 1996). Besides its effects on the shape of boutons, the dlg-1XI-2 mutation causes a misdistribution of HRP-epitope-bearing molecules (Desai et al., 1994; Sun and Salvaterra, 1995) and of the vesicle-associated protein synaptotagmin. daGactivated SAPs clearly failed to restore HRP-immunoreactivity in the motor nerve endings. This is likely to be due to insufficient activation of the target genes in the motor neurons. By means of the temperature-sensitive allele dlg-1HF321 Guan et al. (1996) have documented that the time around synapse formation during embryogenesis is particularly phenocritical with respect to synaptic bouton structure. Hence, the missing of any considerable daGdriven expression in the embryonic VNC and the lack of detectable expression in motor neurons of the 1st and 2nd larval instar provides a reasonable explanation for the inability of daG-driven SAPs to restore presynaptic bouton structures effectively. However, we cannot fully exclude that the phenotype and hence its rescue by neuron-specific activation of SAPs originates upstream from the nerve terminals, e.g. at the postsynapses of the motor nerves or within interneurons that innervate them. This possibility, however, appears unlikely, because mutations that generally affect excitability obviously do not alter bouton structure (Budnik et al., 1990). The NMJs of Drosophila are an attractive, genetically accessible model to study processes such as synaptogenesis and synaptic plasticity (for review see Keshishian et al., 1996; Martin and Kandel, 1996). The successful incor-

Fig. 7. dlg-1XI-2 mutant bouton phenotype, rescue effects and specific localization of SAP97 and SAP102 at larval NMJs. (A,B,E–G) HRP-specific DAB stainings of motor nerve endings at abdominal muscles 6/7 of genotypically different 3rd instar larvae. (A) Canton S (wild-type); type I boutons are of distinct, circular shape. (B) dlg-1XI-2 / Y; the contours of most boutons are blurred. (C,D) Anti-synaptotagmin stainings of wild-type and dlg-1XI-2 mutant boutons; again the shape of mutant boutons appears irregular and swollen. (E) dlg-1XI-2 / Y; daG39/ UAS-SAP97194.19; type Ib boutons appear only moderately restored (arrow), whereas some of the smaller type Is endings look fairly normal (arrowheads). (F,G) dlg-1XI-2 / Y; elav- GAL4 /UASSAP97194.19 (F) and dlg-1XI-2 / Y; elav- GAL4//UAS-SAP102 213.29 (G); the clear outlines of boutons are restored. Note that larvae of these genotypes still exhibit the tumor phenotype. (H) Same genotype as in (E); SAP97-specific staining at muscles 6/7, revealing that the protein is strictly localized around type I boutons. No crossreactions of the SAP97 mAb with endogenous fly proteins were observed in body wall preparations of daG39 larvae and the same specific localization was observed for SAP102 upon appropriate expression (data not shown). (I) Composite optical sections of bouton structures and their postsynaptic counterparts, double labeled with antibodies against HRP (red) and SAP97 (green). The distributions of both antigens are essentially disjunctive, demonstrating that SAP97 is predominantly postsynaptic upon activation by daG39. Arrowhead marks one of several coves, where considerable HRP staining would have been expected in case of a complete rescue. Scale bars: 10 mm.

172

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174

poration of two synaptic MAGUK proteins from the vertebrate CNS into the Drosophila NMJ demonstrates the suitability of this model system also for the study of components of vertebrate excitatory synapses.

4. Experimental procedures 4.1. Fly stocks, germ line transformation and GAL4mediated expression of transgenes Flies were grown on standard medium and crosses were performed at room temperature or at 25°C. Genetic markers and balancer chromosomes are described by Lindsley and Zimm (1992). The allele dlg-1XI-2 (Stewart et al., 1972; kindly provided by E. Gateff, Mainz) is linked on the Xchromosome to y, w and sn and was propagated as a Bascbalanced stock. The y marker allowed identification of dlg-1XI-2 mutants throughout development whereas the sn bristle phenotype could be monitored from late pupal stages onwards. Strains bearing hypomorphic, temperature-sensitive alleles and a dlg-1 comprising duplication, i.e. FM7/dlg-1v55 and FM3/dlg-1HF321/Dp (1; Y) v+ Y y+ were described by Perrimon (1988). dlg-1 mutant females could be generated by means of the mentioned duplication that allows transmission of dlg-1 alleles through males. GAL4-responsive SAP97 and SAP102 minigenes were obtained upon cloning of SAP97 cDNA 31f2 (B.M. and C.C.G., unpublished) or SAP102 cDNA 2d3 (Mu¨ller et al., 1996), respectively, into the EcoRI-site of the pUAST vector (Brand and Perrimon, 1993). The 31f2 cDNA comprises a 225 bp 5′-untranslated region (UTR), a 2697 bp open reading frame (ORF) and about 1.3 kb 3′-UTR; cDNA 2d3 is composed of 340 bp 5′-UTR, 2547 bp ORF and about 1.6 kb 3′-UTR. Germ line transformation of these constructs into the w1118 strain was carried out essentially as described by Spradling (1986). Transgenic stocks were established over appropriate balancers or kept as homozygotes. Several strains bearing dlg-1XI-2 recombined with various lines of either P[w + , UAS-SAP97 ], P[w + , UAS-SAP102] and/or daG39 were generated by standard genetic procedures. The GAL4 activator line daG39 is based on the same construct as line daG32, that was described by Wodarz et al. (1995). The elav-GAL4 activator strain was generated and described by Luo et al. (1994). Three independent UAS-lacZ reporter strains (Brand and Perrimon, 1993; K. Itoh and J. Urban, unpublished) were used and yielded similar results.

dehydrated in isopropanol and mounted in Hoyer’s medium. Flight ability was tested by placing flies singly into a fly cage and irritating them continuously with a needle for at least 3 min or until they escaped by flying away. None of the more than 50 males that were rescued upon exclusively daG-mediated expression of SAP97 or SAP102 exhibited flight ability, although most of these flies were able to jump and did not show held-up wings which are associated with other mutations affecting flight ability. In contrast, 12 out of 30 rescued males, expressing SAP102 through activation by both daG and elav-GAL4, were able to fly. 4.3. Antibodies and histochemical stainings A monoclonal antibody, mAb-77.4, directed against the 163 amino terminal residues of SAP97, was generated by the UAB hybridoma facility following immunization of mice with an appropriate GST fusion protein (Mu¨ller et al., 1995). This antibody was used diluted 1:50 to 1:100. To detect SAP102, a polyclonal rabbit antiserum directed against the 119 amino terminal residues of this protein (Mu¨ller et al., 1996) was used diluted 1:100. Anti-crumbs antibody Cq4 (Tepaß et al., 1990) was diluted 1:1 and antisynaptotagmin serum DSYT2 (Littleton et al., 1993) and rabbit anti-HRP serum (Sigma) were used at dilutions of 1:100 to 1:400. DlgA antiserum (Woods et al., 1996) was used at a 250-fold dilution. Antibody-mediated 3,3′-diaminobenzidine (DAB, Sigma) staining and cryosectioning of imaginal discs were carried out essentially as described, using peroxidase-anti-peroxidase complexes (PAP, Dianova) to increase signal intensity (Schuster et al., 1993). Larval body walls were prepared for immunocytochemistry according to Lin and Goodman (1994). At least 15 body wall preparations, each comprising six pairs of the various types of abdominal muscle fibers, were analyzed for each genotype. Similar effects as those monitored at NMJs of muscles 6/7 were also observed for type Ib boutons at muscles 12/13 and 15/16. Secondary antibodies labeled with FITC and Cy5 (Dianova) were used for confocal microscopy on a Leica confocal microscope. To monitor b-galactosidase activity, tissues were fixed for 4–5 min in 3.7% formaldehyde in PBS at 22°C, washed in PBS and incubated for 12–24 h at 32°C in Xgal staining solution (Bellen et al., 1989) containing 0.04% X-gal.

Note added in proof 4.2. Morphology and assessment of flight ability Prepupae and pupae were fixed for several hours in 10% formaldehyde, air-dried and visualized by phase-contrast dark-field microscopy; adult flies were photographed on a stereo-microscope. Adult cuticle structures were

Shortly after submission of the revised version of this paper, Tejedor et al. [(1997) J. Neurosci. 17, 152–159] published that DlgA and Shaker-type channels interact at NMJs in Drosophila, consistent with assumptions made in Section 3.

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174

Acknowledgements We are grateful to Uwe Hinz, Ulrich Tepaß, Liquon Luo, Norbert Perrimon, Elisabeth Gateff, Rick Fehon, Joachim Urban and the fly stock centers at Umea (Sweden) and Bloomington (IN) for providing GAL4 activator lines and mutant strains. We would also like to thank Eli Knust, Karen Schulze, Hugo Bellen and Dan Woods for antibodies, Werner Zuschratter for help with confocal microscopy and Nadja Oellers and Dorothea Godt for helpful comments on the manuscript. This work was supported by the BMBF and the Fonds der Chemischen Industrie.

References Abbott, L.A. and Natzle, J.E. (1992) Epithelial polarity and cell separation in the neoplastic l(1)dlg-1 mutant of Drosophila. Mech. Dev. 37, 43–56. Bellen, H.J., O’Kane, C.J., Wilson, C., Grossniklaus, U., Pearson, R.K. and Gehring, W.J. (1989) P-element-mediated enhancer detection: a versatile method to study development in Drosophila. Genes Dev. 3, 1288–1300. Boedigheimer, M. and Laughon, A. (1993) Expanded: a gene involved in the control of cell proliferation in imaginal discs. Development 118, 1291–1301. Boedigheimer, M., Bryant, P. and Laughon, A. (1993) Expanded, a negative regulator of cell proliferation in Drosophila, shows homology to the NF2 tumor suppressor. Mech. Dev. 44, 83–84. Brand, A.H. and Perrimon, N. (1993) Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 118, 401–415. Brenman, J.E., Chao, D.S., Gee, S.H., McGee, A.W., Craven, S.E., Santillano, D.R., Wu, Z., Huang, F., Xia, H., Peters, M.F., Froehner, S.C. and Bredt, D. (1996a) Interaction of nitric oxide synthase with the postsynaptic density protein PSD-95 and a1-syntrophin mediated by PDZ domains. Cell 84, 757–767. Brenman, J.E., Christopherson, K.S., Craven, S.E., McGee, A.W. and Bredt, D.S. (1996b) Cloning and characterization of Postsynaptic Density 93, a nitric oxide synthase interacting protein. J. Neurosci. 16, 7407–7415. Bryant, P.J., Watson, K.L., Justice, R.W. and Woods, D.F. (1993) Tumor suppressor genes encoding proteins required for cell interactions and signal transduction in Drosophila. Development Suppl., 239–249. Budnik, V., Zhong, Y. and Wu, C.F. (1990) Morphological plasticity of motor axons in Drosophila mutants with altered excitability. J. Neurosci. 10, 3754–3768. Budnik, V., Koh, Y.-H, Guan, B., Hartmann, B., Hough, C., Woods, D. and Gorczyca, M. (1996) Regulation of synapse structure and function by the Drosophila tumor suppressor gene dlg. Neuron 17, 627–640. Chant, J. and Stowers, L. (1995) GTPases cascades choreographing cellular behavior: movement, morphogenesis, and more. Cell 81, 1–4. Cho, K.O., Hunt, C.A. and Kennedy, M.B. (1992) The rat brain postsynaptic density fraction contains a homolog of the Drosophila discslarge tumor suppressor protein. Neuron 9, 929–942. Desai, C.J., Popova, E. and Zinn, K. (1994) A Drosophila receptor tyrosine phosphatase expressed in the embryonic CNS and larval optic lobes is a member of the set of proteins bearing the ‘HRP’ carbohydrate epitope. J. Neurosci. 14, 7272–7283. 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 120, 545–557. Furuse, M., Itoh, M., Hirase, T., Nagafuchi, A., Yonemura, S., Tsukita,

173

S. and Tsukita, S. (1994) Direct association of occludin with ZO-1 and its possible involvement in the localization of occludin at tight junctions. J. Cell Biol. 127, 1617–1626. Gateff, E. (1978) Malignant neoplasms of genetic origin in Drosophila melanogaster. Science 200, 1448–1459. Green, C.R. and Bergquist, P.R. (1982) Phylogenetic relationships within the invertebrata in relation to the structure of septate junctions and the development of ‘occluding’ junctional types. J. Cell Sci. 53, 279–305. Grether, M.E., Abrams, J.M., Agapite, J., White, K. and Steller, H. (1995) The head involution defective gene of Drosophila melanogaster functions in programmed cell death. Genes Dev. 9, 1694–1708. Guan, B., Hartmann, B., Kho, Y.-H., Gorczyca, M. and Budnik, V. (1996) The Drosophila tumor suppressor gene, dlg, is involved in structural plasticity at a glutamatergic synapse. Curr. Biol. 6, 695– 706. Hinz, U., Giebel, B. and Campos-Ortega, J.A. (1994) The basic-helixloop-helix domain of Drosophila lethal of scute protein is sufficient for proneural function and activates neurogenic genes. Cell 76, 77–87. Hunt, C.A., Schenker, L.J. and Kennedy, M.B. (1996) PSD-95 is associated with the postsynaptic density and not with the presynaptic membrane at forebrain synapses. J. Neurosci. 16, 1380–1388. Itoh, M., Nagafuchi, A., Yonemura S., Kitani-Yasuda, T., Tsukita, S. and Tsukita, S. (1993) The 220-kD protein colocalizing with cadherins in non-epithelial cells is identical to ZO-1, a tight junction-associated protein in epithelial cells: cDNA cloning and immunoelectron microscopy. J. Cell Biol. 121, 491–502. Jan, L.Y. and Jan, Y.N. (1982) Antibodies to horseradish-peroxidase as specific neuronal markers in Drosophila and in grasshopper embryos. Proc. Natl. Acad. Sci. USA 79, 2700–2704. Jesaitis, L.A. and Goodenough, D.A. (1994) Molecular characterization and tissue distribution of ZO-2, a tight junction protein homologous to ZO-1 and the Drosophila discs-large tumor suppressor protein. J. Cell Biol. 124, 949–961. Johansen, J., Halpern, M.E., Johansen, K.M. and Keshishian, H. (1989) Stereotypic morphology of glutamatergic synapses on identified muscle cells of Drosophila larvae. J. Neurosci. 9, 710–725. Keshishian, H., Chiba, A., Chang, T.N., Halfon, M.S., Harkins, E.W., Jarecki, J., Wang, L., Anderson, M., Cash, S., Halpern, M.E. and Johansen, J. (1993) Cellular mechanisms governing synaptic development in Drosophila melanogaster. J. Neurobiol. 24, 757–787. Keshishian, H., Broadie, K., Chiba, A. and Bate, M. (1996) The Drosophila neuromuscular junction: a model system for studying synaptic development and function. Annu. Rev. Neurobiol. 19, 545–575. Kim, S.K. (1995) Tight junctions, membrane-associated guanylate kinases and cell signaling. Curr. Opin. Cell Biol. 5, 772–778. Kim, E., Niethammer, M., Rothschild, A., Jan, Y.N. and Sheng, M. (1995) Clustering of Shaker-type K + channels by interaction with a family of membrane-associated guanylate kinases. Nature 378, 85–88. Kim, E., Cho, K.-O., Rothschild, A. and Sheng, M. (1996) Heteromultimerization and NMDA receptor-clustering activity of chapsyn-110, a member of the PSD-95 family of proteins. Neuron 17, 103–113. Kistner, U., Wenzel, B.M., Veh, R.W., Cases-Langhoff, C., Garner, A.M., Appeltauer, U., Voss, B., Gundelfinger, E.D. and Garner, C.C. (1993) SAP90, a rat presynaptic protein related to the product of the Drosophila tumor suppressor gene dlg-A. J. Biol. Chem. 268, 4580–4583. Kornau, H.-C., Schenker, L.T., Kennedy, M.B. and Seeburg, P.H. (1995) Domain interaction between NMDA receptor subunits and the postsynaptic density protein PSD-95. Science 269, 1737–1740. Lahey, T., Gorczyca, M., Jia, X.X. and Budnik, V. (1994) The Drosophila tumor suppressor gene dlg is required for normal synaptic bouton structure. Neuron 13, 823–835. Laube, G., Ro¨per, J., Pitt, J.C., Sewing, S., Kistner, U., Garner, C.C., Pongs, O. and Veh, R.W. (1996) Ultrastructural localization of Shaker-related potassium channel subunits and synapse-associated protein 90 to septate-like junctions in rat cerebellar Pinceaux. Mol. Brain Res. 42, 51–61.

174

U. Thomas et al. / Mechanisms of Development 62 (1997) 161–174

Lin, D.M. and Goodman, C.S. (1994) Ectopic and increased expression of fasciclin II alters motoneuron growth cone guidance. Neuron 13, 507–523. Lindsley, D.L. and Zimm, G.G. (1992) The Genome of Drosophila melanogaster. Academic Press, New York. Littleton, J.T., Bellen, H.J. and Perin, M.S. (1993) Expression of synaptotagmin in Drosophila reveals transport and localization of synaptic vesicles to the synapse. Development 118, 1077–1088. Lue, R.A., Marfatia, S.M., Branton, D. and Chishti, A.H. (1994) Cloning and characterization of hdlg: the human homologue of the Drosophila discs large tumor suppressor binds to protein 4.1. Proc. Natl. Acad. Sci. USA 91, 9818–9822. Lue, R.A., Brandin, E., Chan, E.P. and Branton, D. (1996) Two independent domains of hDlg are sufficient for subcellular targeting: the PDZ1-2 conformational unit and an alternatively spliced domain. J. Cell Biol. 135, 1125–1137. Luo, L., Liao, Y.J., Jan, L.Y. and Jan, Y.N. (1994) Distinct morphogenetic functions of similar small GTPases: Drosophila Drac1 is involved in axonal outgrowth and myoblast fusion. Genes Dev. 8, 1787–1802. Mackay, D.J.G., Nobes, C.D. and Hall, A. (1995) The Rho’s progress: a potential role during neuritogenesis for the Rho family of GTPases. Trends Neurosci. 18, 496–501. Marfatia, S.M., Morais Cabral, J.H., Lin, L., Hough, C., Bryant, P.J., Stolz, L. and Chishti, A.H. (1996) Modular organization of the PDZ domains in the human Discs Large protein suggests a mechanism for coupling PDZ domain-binding proteins to ATP and the membrane cytoskeleton. J. Cell Biol. 135, 753–766. Martin, K.C. and Kandel, E.R. (1996) Cell adhesion molecules, CREB, and the formation of new synaptic connections. Neuron 17, 567–570. Matsumine, A., Ogai, A., Senda, T., Okumura, N., Satoh, K., Baeg, G.H., Kawahara, T., Kobayashi, S., Okada, M., Toyoshima, K. and Akiyama, T. (1996) Binding of APC to the human homolog of the Drosophila discs large tumor suppressor protein. Science 272, 1020– 1023. Mu¨ller, B.M., Kistner, U., Veh, R.W., Cases-Langhoff, C., Becker, B., Gundelfinger, E.D. and Garner, C.C. (1995) Molecular characterization and spatial distribution of SAP97, a novel presynaptic protein homologous to SAP90 and the Drosophila discs-large tumor suppressor protein. J. Neurosci. 15, 2354–2366. Mu¨ller, B.M., Kistner, U., Kindler, S., Chung, W.J., Kuhlendahl, S., Lau, L.-F., Veh, R.W., Huganir, R.L., Gundelfinger, E.D. and Garner, C.C. (1996) SAP102, a novel postsynaptic protein that interacts with the cytoplasmic tail of the NMDA receptor subunit NR2B. Neuron 17, 255–265. Niethammer, M., Kim, E. and Sheng, M. (1996) Interaction between the C terminus of NMDA receptor subunits and multiple members of the PSD-95 family of membrane-associated guanylate kinases. J. Neurosci. 16, 2157–2163. Perrimon, N. (1988) The maternal effect of lethal(1) discs-large-1: a recessive oncogene of Drosophila melanogaster. Dev. Biol. 127, 392–407. Rajasekaran, A.K., Hojo, M., Huima, T. and Rodriguez-Boulon, E. (1996) Catenins and Zonula Occludens-1 form a complex during

early stages in the assembly of tight junctions. J. Cell Biol. 132, 451–463. Schuster, R., Phannavong, B., Schroeder, C. and Gundelfinger, E.D. (1993) Immunohistochemical localization of a ligand-binding and a structural subunit of nicotinic acetylcholine receptors in the central nervous system of Drosophila melanogaster. J. Comp. Neurol. 335, 149–162. Spradling, A.C. (1986) P element-mediated transformation. In Roberts, D.B. (ed.), Drosophila: a Practical Approach, IRL Press, Oxford, pp. 175–197. Stewart, M., Murphy, C. and Fristrom, J. (1972) The recovery and preliminary characterization of X chromosome mutants affecting imaginal discs of Drosophila melanogaster. Dev. Biol. 27, 71–83. 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. Sun, B. and Salvaterra, P.M. (1995) Two Drosophila nervous system antigens, Nervana 1 and 2, are homologous to the b subunit of Na+, K+-ATPase. Proc. Natl. Acad. Sci. USA 92, 5396–5400. Tepaß, U. and Hartenstein, V. (1994) The development of cellular junctions in the Drosophila embryo. Dev. Biol. 161, 563–596. Tepaß, U., Theres, C. and Knust, E. (1990) crumbs encodes an EGF-like protein expressed on apical membranes of Drosophila epithelial cells and required for organization of epithelia. Cell 61, 787–799. Tsukita, S., Itoh, M., Nagafuchi, A., Yonemura, S. and Tsukita, S. (1993) Submembranous junctional plaque proteins include potential tumor suppressor molecules. J. Cell Biol., 123, 1049–1053. White, K., Grether, M.E., Abrams, J.M., Young, L., Farrell, K. and Steller, H. (1994) Genetic control of programmed cell death in Drosophila. Science 264, 677–683. Willott, E., Balda, S.B., 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. Wodarz, A., Hinz, U., Engelbert, M. and Knust, E. (1995) Expression of Crumbs confers apical character on plasma membrane domains of ectodermal epithelia of Drosophila. Cell 82, 67–76. Wood, R.L. (1990) The septate junction limits mobility of lipophilic markers in the plasma membranes of Hydra vulgaris (attenuata). Cell Tissue Res. 259, 61–68. Woods, D.F. and Bryant, P.J. (1989) Molecular cloning of the lethal(1) discs large-1 oncogene of Drosophila. Dev. Biol. 134, 222–235. 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. and Bryant, P.J. (1993a) ZO-1, DlgA and PSD-95/SAP90: homologous proteins in tight, septate and synaptic cell junctions. Mech. Dev. 44, 85–89. Woods, D.F. and Bryant, P.J. (1993b) Apical junctions and cell signaling in epithelia. J. Cell Sci. Suppl. 17, 171–181. Woods, D.F., Hough, C., Peel, D., Callaini, G. and Bryant, P.J. (1996) Dlg protein is required for junction structure, cell polarity and proliferation control in Drosophila epithelia. J. Cell Biol. 134, 1469–1482.