Varicose: a MAGUK required for the maturation and function of Drosophila septate junctions
- 5k Downloads
Scaffolding proteins belonging to the membrane associated guanylate kinase (MAGUK) superfamily function as adapters linking cytoplasmic and cell surface proteins to the cytoskeleton to regulate cell-cell adhesion, cell-cell communication and signal transduction. We characterize here a Drosophila MAGUK member, Varicose (Vari), the homologue of vertebrate scaffolding protein PALS2.
Varicose localizes to pleated septate junctions (pSJs) of all embryonic, ectodermally-derived epithelia and peripheral glia. In vari mutants, essential SJ proteins NeurexinIV and FasciclinIII are mislocalized basally and epithelia develop a leaky paracellular seal. In addition, vari mutants display irregular tracheal tube diameters and have reduced lumenal protein accumulation, suggesting involvement in tracheal morphogenesis. We found that Vari is distributed in the cytoplasm of the optic lobe neuroepithelium, as well as in a subset of neuroblasts and differentiated neurons of the nervous system. We reduced vari function during the development of adult epithelia with a partial rescue, RNA interference and generation of genetically mosaic tissue. All three approaches demonstrate that vari is required for the patterning and morphogenesis of adult epithelial hairs and bristles.
Varicose is involved in scaffold assembly at the SJ and has a role in patterning and morphogenesis of adult epithelia.
KeywordsCora Optic Lobe Planar Cell Polarity Septate Junction Adult Nervous System
central nervous system
green fluorescent protein
membrane associated guanylate kinase
Flippase/Flippase recombination target
Proteins Associated with Lin-7
post synaptic density protein, disc large tumor suppressor, zonula occludens-1 protein
peripheral nervous system
pleated septate junction
src homology 3
The assembly of cellular junctions is pivotal for metazoans to maintain a homeostatic environment. Through these junctions, cells are able to communicate, synchronize function, and regulate the paracellular flow of molecules [1, 2, 3]. Epithelial cells are polarized along an apico-basal axis where the apical surface faces the exterior or lumen and the basal surface communicates with the extracellular matrix . In vertebrates, several epithelial intercellular junctions exist, the two most widely studied being tight junctions (TJs) and adherens junctions (AJs) [5, 6]. Invertebrate species, although lacking tight junctions possess the functionally analogous septate junction (SJ) . Despite a difference in lateral membrane location, TJs (apical to AJs) and SJs (basal to AJs) both form an intercellular barrier to regulate the transepithelial diffusion of solutes .
Ultrastructure and freeze-fracture analysis of cell junctions reveal that SJs maintain a fixed distance between epithelial cells through ladder-like septae that spiral on the outside of the cell and fill the intermembrane space . These encircling septae extend the travel distance for molecules to transverse the paracellular path, thereby regulating the flow of material . TJs however, appear as a series of contact points, or 'kissing sites'. Freeze-fracture analysis reveals that TJs consist of interconnecting mesh-works of fibrils forming a band-like structure around the cell. In spite of a diverged morphology, globular and transmembrane proteins are suspected to form the bridge between cells filling the intermembrane space of TJs and SJs, respectively [8, 10].
Two types of SJs have been observed in Drosophila, smooth (sSJs) and pleated (pSJs). Smooth SJs, which lack ladder-like septa are found in tissues such as the midgut and malpighian tubules. Pleated SJs are found in all ectodermally-derived epithelial tissues such as trachea, salivary glands, hindgut and epidermis as well as in glial cells . Glial SJs function to link ensheathing glial cells around peripheral nerves, forming the blood-brain barrier [11, 12]. Recently, SJ were indentified in the apical and basal regions between accessory cells, the cone cells and pigment cells of adult ommatidia . SJs and TJs, both in epithelia and neurons, share core components suggesting that barrier function is a conserved mechanism between vertebrate and invertebrate species .
Within the last decade, key molecular elements of pSJs have been identified and shown to be involved in processes such as establishing and maintaining cell polarity, cell adhesion and cell-cell interactions . Coracle (Cora), NeurexinIV (NrxIV), Neuroglian (Nrg), Na+/K+ ATPase (ATPα), and Discs Large (Dlg) have all been identified as SJ constituents [14, 15, 16, 17]. Drosophila Coracle, a member of the Protein 4.1 family, interacts with transmembrane proteins forming a link to the cytoplasmic surface of the plasma membrane . Cora localizes to epithelial pSJs, where it is required for SJ organization but is absent from the CNS and its derivatives [16, 19]. Interactions between Cora and transmembrane protein NrxIV are necessary in order to maintain their proper SJ localization . NrxIV, a member of the neurexin gene family, localizes to all pSJs of ectodermally-derived epithelia and the CNS . In both cora and nrxIV mutants, intermembrane septae are absent, resulting in a leaky paracellular seal. While Cora and NrxIV have been found to interact with ATPα at the SJ, they have also been shown to form a complex independently with Nrg. Nrg is an integral membrane glycoprotein that localizes to the lateral membrane of epithelial cells and to the surface of glial cells, regulating the adhesion between neurons and glial cells [20, 21]. Mutations in either ATPα or Nrg results in mislocalization of Cora and NrxIV and disrupts SJ structure and function. This suggests that interdependent protein complexes function to assemble the protein scaffold regulating paracellular movement. However, among these mutants, epithelial integrity, apico-basal polarization and the localization of SJ protein Dlg are unaffected .
Membrane Associated GUanylate Kinases, MAGUKs, are a class of scaffolding proteins that tether adhesion molecules at sites of cell-cell contact, such as septate and tight junctions . MAGUK proteins contain a core domain structure consisting of 1–3 PDZ domains (named after 3 founding MAGUK proteins PSD-95, Dlg and ZO-1), a src homology 3 (SH3) domain and a guanylate kinase domain (GUK) . In addition, some MAGUK members encode an N-terminal L27 domain (named after interacting proteins Lin-2 and Lin-7) which functions in protein-protein interactions . This multi-domain composition allows MAGUKs to function as the backbone onto which protein complexes can assemble . These complexes then bring together functionally dissimilar proteins to link transmembrane proteins with the cytoskeleton .
The PALS (Proteins Associated with Lin-7) subfamily of MAGUK proteins, PALS1 and PALS2, anchor scaffolding complexes at junctional regions [22, 27]. Drosophila Stardust (Sdt) and its homologue, vertebrate PALS1, function as adapter proteins linking two scaffolding complexes to establish epithelial polarity [28, 29]. The functional significance of PALS2 remains unclear. Vertebrate studies have shown that PALS2 interacts with the C-terminus of Nectin-like molecule 2 (Necl-2) through its PDZ domain at spot-like adhesion sites along the lateral plasma membrane. Furthermore, Necl-2 binds DAL1, a Protein 4.1 family member, implicating PALS2 in membrane organization and scaffold assembly . The evolutionary conservation of junctional proteins prompted our search for the invertebrate homologue of PALS2 to elucidate its molecular and genetic function. Our search identified Drosophila CG9326, also reported recently as the gene interrupted in the varicose (vari) mutation [31, 32]. Our previously reported senz'aria (szar) alleles  are thus renamed as alleles of vari.
We present here our findings of Drosophila CG9326, Varicose (Vari), a homologue of vertebrate PALS2. Previous independent studies identify a function for Vari in the function of epithelial SJs [32, 34]. Our studies show that Vari localizes to the SJ in all embryonic ectodermally-derived epithelial tissues as well as glial cells of the PNS. Mutations in vari result in mislocalized SJ markers such as NrxIV and compromise the seal of the transepithelial barrier. We have identified Vari expression in larval optic lobe neuroblasts and in the adult nervous system. Furthermore, Varicose is required for the patterning of adult epithelial structures. Adult wing hair extension and orientation, patterning of thoracic bristles, and the patterning of inter-ommatidial bristles are disrupted when vari function is reduced or removed. This study identifies both conserved and novel functions for MAGUK proteins in Drosophila development.
Varicose, a homologue of vertebrate PALS2/VAM-1
Varicose localizes to embryonic epithelial tissues
The embryonic expression of Vari has been previously described . We detected varicose transcripts from early stage 10 of embryogenesis until hatching. Protein expression was first detected during stage 13 of embryogenesis by immunolabeling with anti-Vari, and until late stage 17.
Varicose localizes to the septate junction during embryogenesis but not in imaginal discs
In Drosophila, MAGUKs typically function as scaffolding proteins upon which multiprotein complexes form to regulate cell polarization and adhesion (reviewed in Funke et al., ). The restricted membrane localization of Vari suggested to us that it may act similarly. We compared Varicose expression with various lateral membrane markers in the hindgut of stage 15 embryos (Fig. 2E–G, see Additional file 1, panels B-G). Vari localizes adjacent to, yet fails to co-localize with, the sub-apical marker Crumbs  and the adherens junction marker Phosphotyrosine . Co-localization of Varicose and plasma membrane marker α-Spectrin  is seen in the apical region of the lateral membrane but Vari is not seen in the basal region, indicating Vari localizes to the apicolateral membrane, a region corresponding to the SJ. Septate junctions are characterized by the localization of proteins such as Discs-large, Coracle and NeurexinIV [14, 15, 16]. Double-labeling experiments with Varicose and SJ markers FasIII (Fig. 2E), NrxIV (Fig. 2F) and Na+K+ ATPase (Fig. 2G) reveal a complete overlap of expression, suggesting that Varicose is localized solely in the septate junction. Co-localization of Vari and Dlg is also seen in the trachea, salivary gland and proventriculus (data not shown). In concurrence with a previous study, our evidence indicates that Varicose expression is restricted to the SJ region in all embryonic ectodermally-derived epithelial tissues .
If Vari is required by epithelial SJ, then it should be expressed in imaginal epithelia. Although Bachmann et al.,, report immunolabeling of eye and wing discs, our antibody did not reveal a labeling pattern different from controls (see Additional file 3). Phenotype data explored below suggests that Vari does function in imaginal epithelia.
MAGUK function in the nervous system
In larvae, Varicose expression was observed in the neuroepithelium of the developing optic lobe. The optic lobe consists of two populations of cells, symmetrically dividing lateral neuroepithelial (NE) cells and asymmetrically dividing medial neuroblasts. NE cells possess similar properties as embryonic epithelial cells and express junctional markers at similar locations . To determine if Varicose localized to SJs in postembryonic epithelia, we labeled third instar larvae brains with Vari and Dlg. Dlg localizes to the SJ in NE cells (arrow, Fig. 3F) and to the cortex in neuroblasts (left of arrowheads, Fig. 3E–G). In contrast to what we observed in embryonic epithelia, Varicose has limited co-localization with Dlg (Fig. 3G). Varicose expression is found in the apical cytoplasm of NE cells (arrow, Fig. 3E) but is not found in neuroblasts (left of arrowheads, Fig. 3E). We were unable to detect neuroepithelial labeling with pre-immune sera suggesting the observed pattern is due to Varicose expression (see Additional file 2, panels B-D). If this expression pattern is true for other SJ markers, we would expect to see NrxIV at the SJ of NE cells. We did not observe any NrxIV expression in NE cells or in neuroblasts; we however did detect expression in neuroblast progeny (not shown). Expression of Varicose restricted to the neuroepithelium of the optic lobe suggests a role in the symmetrically dividing cell pool. Moreover, we have identified a MAGUK member that does not always associate with the plasma membrane, suggesting a novel role for this protein in the neuroepithelium.
Central neuroblasts found in third instar larvae brains also express Varicose. We performed various double-labeling experiments using several neuroblast markers. Varicose did not co-localize with Dlg, a cortex marker (Fig. 3H), Prospero, a ganglion mother cell marker (Fig. 3J), or Repo, a glial cell marker (Fig. 3K) [39, 41, 42]. Weak Varicose expression was observed in differentiated neurons labeled with Elav (Fig. 3I) . We concluded that Varicose expression at this stage remains in the cytoplasm of neuroblasts.
Identifying Varicose expression in neuroblasts prompted us to characterize Varicose expression in the adult nervous system. We immunolabeled pupal brains 50 hours after enclosion with Vari and either Dlg (Fig. 3L), FasII (Fig. 3M), Elav (Fig. 3N) or Repo (Fig. 3O). We did not observe Varicose expression in the mushroom bodies or in the antennal lobes. Varicose however, co-localized with Dlg in the cell bodies of neurons surrounding these neuropile regions (yellow; Fig. 3L). We deduced that these cell bodies belong to differentiated neurons as opposed to glial cells because Vari is localized with Elav (Fig. 3N) and not Repo labeled cells (Fig. 3O). Therefore, Varicose localizes to a subset of differentiated neurons surrounding neuropile regions of the adult nervous system.
Loss-of-function varicose mutants are embryonic lethal
We have created a loss-of-function allelic series, vari48EP, variK4, variL4, variB4, and variB5, to determine whether Vari plays a role in septate junction assembly. The embryonic lethal P-element insertion line GE13049 (GenExel, Inc), contains an EP insertion 3507 bp downstream of the translation start site of VariL27Band VariL27D, and 1731 bp upstream of the translation start site of Vari (Fig. 1). Using standard procedures, we mobilized GE13049 and generated 5 mutant alleles by imprecise excision. Here we present allele vari48EP, an excision allele which removed 4717 bp of genomic sequence (20792456...20797173) leaving behind 416 bp of the P-element. We consider this allele is a true null, as no protein is been detected, and the excised sequence has removed all of the reading frame, except the predicted L27N domain . Our other alleles, variK4 and variL4 are clean excisions that removed 1964 bp and 2795 bp of genomic sequence, respectively. In addition, variB4 removed 3806 bp leaving behind 27 bp of P-element sequence, while variB5 removed 3313 bp of genomic sequence leaving 216 bp of P-element sequence behind. All vari alleles are late embryonic lethal, although variK4 has escapers that die during the second instar. All alleles fail to complement the lethality of GE13049 or the Df(2L)Exel7079 deficiency. The embryonic phenotype of GE13049 is indistinguishable from mutant vari48EPor trans-heterozygotes. Df(2L)Exel7079 is a molecularly characterized deficiency deleting chromosomal region 38E6-38F3. The 3' UTR of both varicose and CG9324/pomp, a 20S maturase, overlap . Our vari48EPexcision removed the 3' UTR of pomp as well as 3 carboxy-terminal amino acids. This vari allele complements a lethal allele, pompEY06518, indicating that our lethal phenotype is a result of disruption of vari. In addition, our sequenced revertant, vari34P, has a wildtype phenotype.
Septate junction assembly requires varicose
As previously mentioned, proper localization of SJ components is interdependent. We assessed the localization of Vari in several SJ mutants including dlgXl-2, nrg14, nrx4304, and coraK08713. Unexpectedly, Vari was properly localized in all SJ mutants examined, except nrg14. Varicose expression is severely reduced and mislocalized basally along the lateral membrane (Fig. 4E, 4F). While nrg14 mutants display reduced or absent transverse septa, the spacing between epithelial plasma membranes is maintained .
Septate junctions are the structural basis of the paracellular barrier in insect epithelia . To determine whether the transepithelial barrier was compromised in vari mutants we performed dye exclusion assays, as described by Lamb et al., . Rhodamine-conjugated dextran was injected into late stage wildtype embryos and dye was excluded from the lumens of the salivary glands and trachea beyond 90 minutes (Fig. 4G). In contrast, within 30 minutes of injection, dye could be detected in the tracheal lumen of vari48EPmutants (Fig. 4H).
Loss of variresults in dilated tracheal branches and reduced lumenal staining
To further understand the tracheal dilations, we examined tracheal cell ultrastructure using electron microscopy. Although vari mutants have abnormally large lumen diameters (arrow, Fig. 5B) compared to controls (Fig. 5A), the overall cell morphology is similar to wildtype (Fig. 5A and 5C). In addition, cuticle secretion was similar in both mutants and wildtype controls. Proper cuticle secretion, taenidial folds and normal cell morphology in vari mutants suggest that its role may be independent of apical secretion.
Varicose acts in adult morphogenesis
A role for vari in the eye and wing was suggested by a partial rescue with the vari transgene, tissue specific protein knockdown with an inverted repeat (IR) transgene , and by generation of tissue mosaic for vari function with the FLP/FRT technique . Ubiquitous expression of full-length Vari using daughterlessGAL4 rescues lethality of vari48EPnull embryos. Viability of the rescued animals ranges from late pupation (80%; n = 73) to viable adults (20%). Viable adults are unstable and unable to walk, and have a life span averaging 3 days. The ability to rescue vari mutants with a transgene lacking the L27 domain suggests that this domain is not essential for development, consistent with the finding of Bachmann et al., .
SJ components Gliotactin (Gli) and Cora are required for proper hair alignment in the adult wing . We examined overall wing morphology of adult flies overexpressing Vari in a null mutant background (inset, Fig. 7G) and found it to be similar to the imposed expression on a wildtype wing (inset, Fig. 7H). However, wing hair alignment is abnormal. Unlike the control wing (Fig. 7H), rescued adults show patches of wing hairs with abnormal alignment compared to their neighbouring hairs (arrowheads, Fig. 7G). The abnormal hair alignment was observed in unmounted wings, eliminating the possibility of a mounting artifact. Similar effects were seen in wings where Vari levels were reduced by expression of vari-IR (UAS-vari-IR; dpp-GAL4), and upon the generation of vari null patches in the wing with FLP/FRT. In both cases, domains of misaligned hairs are associated with patches of wing cells with incomplete hair extension (Fig. 7I, J).
We examined the polarity of bristles on the thorax, abdomen and legs of rescued adults (data not shown) and did not detect abnormal polarization, or multiple hairs per cell. From these results, we suspect that the misalignment phenotype is not due to a disruption in planar cell polarity, but rather due to altered cell polarization common to vari, cora and gli mutants . We rarely observed mutant patches of body cuticle of FLP/FRT mosaics. These patches were associated with loss of hairs, multiple bristles from fused sockets, and small breaches of the cuticle (Fig. 7K).
Our study, and others [32, 34] have characterised a MAGUK family member encoded by Drosophila CG9326, Varicose. We have shown that Varicose localizes to pSJs of all embryonic ectodermally-derived epithelial tissues as well as the pSJs of the PNS. We have detected Vari expression in the neuroepithelium of the developing optic lobe in a non-junction associated pattern, which is unique for a MAGUK member. Expression of Varicose in a subset of central brain neuroblasts and differentiated neurons of the adult nervous system emphasizes the importance and versatility of its function throughout development. Mutations in vari result in mislocalized SJ markers and disruption of the paracellular seal. Loss-of-function vari alleles display dilated and contorted tracheal tubes, implicating vari in tracheal morphogenesis. Furthermore, genetic mosaic and partial rescue phenotypes in the eye and wing suggests a role for vari during adult epithelial morphogenesis.
Varicose plays a role in septate junction assembly
We have presented here several lines of evidence demonstrating that Varicose is required for septate junction formation. First, Varicose co-localizes with known SJ protein NrxIV in all embryonic pSJs, and NrxIV is mislocalized in the absence of varicose activity. Second, in vari null mutants, SJ do not mature to the point of septa formation. Third, the transepithelial barrier of vari mutants is 'leaky' to tracer dyes.
Embryos mutant for varicose show mislocalization of SJ proteins like NrxIV, FasIII and the Na+K+ATPase basally along the plasma membrane of epithelia (this report and Wu et al., ). The localization of SJ protein Dlg was not affected however, indicating Vari is not required to establish epithelial polarity. This is not a surprising result as the onset of Varicose expression appears midway through embryogenesis, a time when polarity has already been established and SJs begin to assemble . Proper localization of Vari requires Nrg. Varicose has been shown to interact with NrxIV [32, 34] and all three proteins share an mutually-dependent relationship necessary for proper subcellular localization . Vari reduces the lateral mobility of Nrg and NrxIV , suggesting that in the absence of vari, the assembly of key SJ proteins is interrupted, disrupting the architecture of the junctional region and triggering a cascade of mislocalized proteins.
To date, all junctional proteins expressed in embryonic epithelia are also expressed in imaginal discs. While this work was in review, Bachmann and colleagues provided immunocytochemical evidence for Vari expression in the wing and eye discs .
Our ultrastructural analysis indicates that embryos null for vari die before intermembrane septa develop. Bachmann et al.,  establish that septa do not develop in hypomorphs that develop further as embryos. nrxIV and cora mutants also lack septa, which are proposed to have a sealing function in the transepithelial barrier [15, 18]. An affinity approach has identified NrxIV as a potential Vari binding partner . These data are consistent with the failure of vari mutants to exclude dye in embryonic trachea. In contrast, SJ mutants, gliotactin and sinuous (sinu) show defects in septa array and septa number, respectively [52, 53]. Mutations in vari enhance the sinu phenotype . Together these results suggest Vari, like NrxIV and Cora, functions in assembling septa strands. However, the low levels of Vari in larvae suggest that Vari is not essential to maintain SJ.
SJ integrity in Drosophila requires Megatrachea (Mega), a claudin that has a C-terminal PDZ binding domain . It has been suggested that a MAGUK member may act to tether Mega to the NrxIV/Cora complex to assemble the SJ [19, 54]. We propose Vari as a candidate for this function.
A role for MAGUKs in the nervous system?
Expression of Vari in peripheral glia, but not the perineural sheath or midline glia of the embryonic nervous system is consistent with function in the establishment of ectodermally-derived pSJs. Neural expression was not detected in embryos. However, the distribution of Vari in the late larval and adult central nervous system suggests non-junctional roles for this MAGUK. In the optic lobe NE, which does express Dlg, Vari expression overlaps, and extends into the apical cytoplasm. Vari is not expressed in NBs of the embryo and medial optic lobe, yet is expressed in the cytoplasm of some central NBs of late third instar, and in low levels in the soma of differentiated neurons. This pattern of expression is not typical of other junctional or cell polarity markers like Bazooka, Glaikit or Miranda [55, 56, 57], or of MAGUKs in general and must be clarified by further study. This issue may be approached with nervous system specific RNAi knockdown of Vari, which we found to be pupal lethal (data not shown).
Several independent lines of evidence suggest that our serum is specific to Varicose. First, our epithelial expression pattern observed during embyrogenesis is consistent with previous reports on varicose . Second, a null allele, vari48EP, lacks wildtype Varicose imunolabeling. Third, the NE expression pattern is absent when third instar larval brains are labeled with pre-immune sera. Fourth, our antibody detects Varicose expression in the ventral midline when UAS-vari is mis-expressed in the midline using single-minded GAL4. Fifth, although normal protein levels are at the threshold of detection, over-expression of UAS-vari using heat-shock GAL4 provides ample protein to be detected by western blotting (see Additional file 2).
Varicose is involved in regulating tube size
The Drosophila tracheal system is a well developed model for the dissection of pathways regulating tube formation . pSJ components are implicated in the regulation of tubule size. Genes regulating tube size fall into two phenotypic categories; those required to regulate tube length and those required for normal tube diameter (reviewed in ). Several lines of evidence have suggested that pSJ components are involved in regulating tube length. Mutations in genes for SJ proteins like mega, sinu and the Na+K+ATPase β subunit, nrv2 have tortuous and elongated tracheal trunks, without affecting tube diameter [52, 54, 59]. In contrast, mutations in vari do not appear to affect tube length. Tracheal tubes in vari mutants have irregular and enlarged tube diameters reminiscent of mutations affecting mmy/cystic and kkv, enzymes required for chitin synthesis [60, 61]. Epistatic analysis of vari and sinu reveals a tracheal phenotype in double mutants that is worse than either single mutant, suggesting these proteins function in different pathways .
The chitin matrix is secreted from the apical surface of tracheal cells and synthesis of the matrix has been linked to controlling tube diameter. During expansion of the dorsal trunk, the cylinder expands as the lumen dilates . It is suggested that formation of the chitin matrix is needed for the organized radial expansion of tracheal tubes . Tracheal enlargement in vari is similar to cystic and kkv mutants, and all three have reduced deposition of 2A12 antigen . Wu and colleagues  further show that vari mutants fail to secrete apical protein Serpentine and secrete variable amounts of Vermiform. Unexpectedly, cuticle ultrastructure and taenidial ridges appear normal in vari mutants. This result is unlike tracheal mutants affecting tube length, such as sinu, where taenidial folds are irregular . Our data suggests that lumenal protein secretion in vari mutants is sufficient to produce cuticle and regulate tube length, and that the SJ may also play a role in regulating tube diameter.
MAGUKs are involved in morphogenesis of adult tissues
Our knowledge of the function of septate junctions during metamorphosis is limited, however a role for SJs in the adult ommatidium has been described. The SJ component NrxIV was shown to localize to junctional regions in the pupal and adult eye. Loss of nrxIV disrupts SJ function, which leads to structural disorganization resulting from a loss of adhesion between cells of the adult ommatidia . Although nrxIV- clones survive in the adult eye , vari- clones do not, indicating that vari has functions other than localizing NrxIV. Partial restoration of vari (by transgene rescue) during adult morphogenesis results in missing ommatidia and irregular bristle patterning. Vari may spatially organise adhesions during this developmental process, and reduced levels of Vari disrupts epithelial patterning. Vari levels are much lower subsequent to vari-RNAi treatment described by Bachmann et al., . The retinal and wing epithelia survive, but epithelial patterning is more disorganised. We report that mosaic null clones of vari generate duplicated bristles in the eye, and clumped sensory hairs on the thorax. The failure to establish SJs in these clones may result in delaminating mutant cells adopting a neurogenic fate, and thereby generate extra sensory structures.
The involvement of SJs during wing imaginal disc to adult wing morphogenesis is also unclear. However, two known SJ components, Gli and Cora are required for survival of pre-hair cells during pupation [18, 50]. Similar to mutations in gli and cora, our partial rescue of vari, and RNAi of vari resulted in patches of wing hairs that fail to point distally . Although reminiscent of a Frizzled (Fz) planar cell polarity phenotype, the mechanism regulating hair alignment acts independently of Fz. Patches of wing hairs, although not pointing distally, retain a parallel alignment with neighbouring hairs in Fz mutants [62, 63]. This is unlike the random alignment seen in vari, gli, and cora mutants where polarity of neighbouring wing hairs is different. During pupal development, the position and orientation of wing prehairs are determined and then stabilized during later stages . Vari and other SJ proteins play a role in prehair patterning.
The MAGUK protein PALS2 has been proposed to act in scaffold formation at the basolateral membrane of mammalian epithelia . Here we show that a Drosophila homologue, Vari, is similarly distributed, and is required in ectodermally-derived epithelia to elaborate pSJs and establish a paracellular barrier. Embryos lacking vari function display mislocalization of essential pSJ membrane proteins, including NrxIV, Na+K+ATPase and FasIII, and are unable to control the permeability of the tracheal membrane. As a result, the trachea fail to fill with air, and the embryos die in early stage 17. The function of SJs in the morphogenesis of the wing and eye is less well characterised, yet imaginal epithelia lacking vari, cora or gli do not survive to the adult. The eye and wing phenotypes of reduced vari function overlaps with patterning defects of mutations in SJ genes nrxIV, gli and cora. Together, they indicate an uncharacterised role for SJs in establishing pattern in epithelial sheets. Vari is not expressed in the embryonic central nervous system, but is expressed apically in the neuroepithelium of the optic lobes and in neuronal cell bodies. These structures do not have pSJs, and indicate that there are uncharacterised functions of Vari, distinct from a role in the assembly of cell junctions.
Canton-S P was used as a wildtype control. EP-element insertion line GE13049 was obtained from GenExel, Inc. Stocks were backcrossed to ensure clean backgrounds. vari48EP, variB4, variK4, variL4 and variB5 generated by imprecise excision of GE13049. vari48EPis a protein null allele (this study). Df(2L)Exel7079 deficiency deletes 19 genes in addition to vari. Single-minded GAL4 was obtained from John Nambu. UAS-vari-IR (transformant 24156) was obtained from the Vienna Drosophila RNAi Centre. All other stocks were obtained from the Bloomington Drosophila Stock Centre.
Rat polyclonal antibody was produced against Vari. Total RNA was extracted from Canton-S P adults using TRIzol (Invitrogen). cDNA synthesis was performed using Ready-To-Go RT-PCR Beads (Amersham). A 909 bp fragment was amplified using the sense primer 5'-GCAAGATCTAGTGGACGACGAATAATCAAG-3' and the antisense primer 5'-GATGGATTCCGGTTGGAGCCCGTGG-3'. This cDNA fragment, corresponding to amino acid residues 83–386, was fused to a C-terminal his-tag in pET29b(+) (Novagen) and expressed in BL21DE3. Following induction, the fusion protein was purified under denaturing conditions by affinity chromatography using His-Select Nickel Affinity Gel (Sigma) and used for rat immunization. Polyclonal antiserum was affinity purified using CNBr-activated sepharose according to manufacturer's protocol (Amersham).
Immunohistochemistry techniques were adapted from Patel . Embryos were collected, decorionated, fixed and incubated in primary antibody diluted in phosphate-buffered saline (PBS) containing 0.5% Triton X-100 and 10% normal goat serum (NGS). Primary antibodies were used at the following dilutions: rat anti-Vari (1:15) (this study), rabbit anti-NrxIV (1:300) (gift from H.J. Bellen, Baylor College of Medicine, Houston, TX), chicken anti-βgalactosidase (1:150) , and anti-Phosphotyrosine (1:300) (Millipore). The following monoclonal antibodies were obtained from the Developmental Studies Hybridoma Bank: anti-Crumbs (1:30), anti-αSpectrin (1:30), anti-Discs Large (1:30), anti-Na+K+ATPase (1:300), MAb2A12 (1:30), anti-Repo (1:7), anti-Elav (1:75), anti-Prospero (1:4), and anti-FasIII (1:30). Embryos were incubated in fluorescent secondary (1:150 dilution, Alexa 488, Alex 594; Molecular Probes). Anti-βgal was detected using biotinylated secondary antibody (1:150) (Vector Laboratories) followed by incubation with Vector Laboratories Elite ABC and 3, 3-Diaminobenzidine Tetra hydrochloride (DAB, Gibco-BRL). Embryos were visualized by confocal microscopy using Zeiss LSM510 or Zeiss Axioskop microscope. Images were processed using ImageJ and Adobe Photoshop.
Third instar larvae brains and pupae brains (50 hours after inclusion) were dissected in PBS and fixed in 4% paraformaldehyde. Following several washes in PBS with 0.3% Triton X-100, tissues were incubated in primary antibody as stated above. Anti-FasII was diluted 1:30 in PBS with 0.5% Triton X-100 and 10% NGS.
Wings were prepared as described in Settle, et al. .
Total RNA was extracted from Canton-S P adults using TRIzol (Invitrogen). Reverse transcription was performed using M-MLV Reverse Transcriptase and random decamers (Ambion). Full-length Vari was amplified using the sense primer 5'-CCGAGGACGTCCTCTAGACCAAGATGCCAG-3' and the antisense primer 5'-CCCCGGAGGGCGCATCTAGACTTATACAAACATTGC-3' to amplify a cDNA fragment corresponding to amino acid residues 1–469. This cDNA was cloned into pUASt and injected into embryos using standard techniques.
Mosaic clones of vari mutant cells were generated using a FLP/FRT-mediated technique . Mitotic recombination of our null allele, vari48EP, was induced by treating yw-, hsFLP; vari48EP, FRT40A/ubi-GFP, FRT40A larvae 24, 48 and 96 hours after egg laying to a single heat shock at 37°C for 1 hour. Flies were raised at 25°C prior to and following heat shock treatment. Mutant clones in the adult eye and thorax were visualized by scanning electron microscopy as described below. To visualize wing defects, whole flies were dehydrated using an ethanol gradient and stored in methyl salicylate. Wings were mounted in D.P.X. on microscope slides and sealed with coverslips. Photomicrographs were processed using OpenLab and Adobe Photoshop® 7.0.
Stage 17 embryos were injected with 5% glutaraldehyde in 0.05 M Cacodylate buffer (pH 7.2) as described , post-fixed in 1.0% Osmium tetroxide in dH2O and stained en bloc with aqueous 2% uranyl acetate. Embryos were dehydrated, embedded and sectioned with established methods . Four embryos of each genotype were sectioned for analysis.
Scanning electron microscopy
Imaging of compound eyes and thorax was performed as described in Settle et al., . In brief, cold anaesthetised adults were imaged at 3.0 Torr in an Electroscan 2020 Environmental Scanning Electron Microscope.
Dye permeability assay
Fluorescent dye injection was performed as described . Stage 17 embryos were examined within 30 minutes of injection on a Zeiss LSM510. Mutants were identified by lack of GFP expression from the balancer (CyO, Kr-GAL4, UAS-GFP).
A special thanks to Kelly Teal for generating our Vari antibody and Shirley Liang for help with mosaics. Xiao-Li Zhao helped generate our transgenic stocks. Noor Hossain and Leena Patel generated the compound eye micrographs. Leena Patel, Allison MacMullin and Verónica Rodriguez Moncalvo provided technical assistance. Thanks to Dr. Juliet Daniel and Dr. Ana Campos for their helpful suggestions and encouragement. We thank Hugo Bellen for providing NrxIV antibody. Antibodies obtained from the Developmental Studies Hybridoma Bank: Cq4 (Crumbs, provided by E. Knust), 3A9 (αSpectrin, provided by D. Branton and R. Dubreuil), 4F3 (Discs Large, provided by C. Goodman), α5 (Na+K+ATPase, provided by D. Fambrough), MAb2A12 (provided by M. Krasnow), 8D12 (Repo, provided by C. Goodman), 9F8A9 (Elav, provided by G.M. Rubin), MR1A (Prospero, provided by C. Doe), 7G10 (FasIII, provided by C. Goodman). This work supported by NSERC.
- 18.Lamb RS, Ward RE, Schweizer L, Fehon RG: 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. 1998, 9: 3505-3519.CrossRefPubMedCentralPubMedGoogle Scholar
- 30.Shingai T, Ikeda W, Kakunaga S, Morimoto K, Takekuni K, Itoh S, Satoh K, Takeuchi M, Imai T, Monden M, Takai Y: Implications of nectin-like molecule-2/IGSF4/RA175/SgIGSF/TSLC1/SynCAM1 in cell-cell adhesion and transmembrane protein localization in epithelial cells. J Biol Chem. 2003, 278: 35421-35427.CrossRefPubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.