DE-Cadherin, a Core Component of the Adherens Junction Complex Modifies Subcellular Localization of the Drosophila Gap Junction Protein Innexin2

Abstract

The Drosophila innexin multigene family of gap junction encoding proteins consists of eight family members whose function in epithelial morphogenesis is mostly unknown. We have recently shown that innexin2 plays a crucial role in the organization of embryonic epithelia. Innexin2 protein accumulates in the epidermis in the apico-lateral membrane domain and colocalizes with core proteins of adherens junctions, such as DE-cadherin and Armadillo, the β -catenin homolog. Innexin2 localization is altered in both armadillo and DE-cadherin mutants Biochemical interaction studies point to a direct interaction of DE-cadherin and Armadillo with innexin2 suggesting a close link between gap junction and adherens junction biogenesis. We have used the Drosophila Schneider cell tissue culture system to further study the interaction of innexin2 with DE-cadherin. Our results provide evidence that DE-cadherin may be a key component to control trafficking, and localization of Innexin2 to the plasma membrane.

Keywords:

• Drosophila
• gap junctions
• innexins
• adherens junctions
• DE-cadherin
• SL2 cells

INTRODUCTION

Gap Junctions contain clusters of intercellular channels, which allow the direct exchange of ions and small molecules among neighboring cells, thus enabling cells and tissues to integrate signaling activities (Goodenough et al. 1996; Wei et al. 2004). A functional gap junction channel is formed by two hexameric hemichannels, one contributed by each of the opposing cells, which dock head-to-head in the extracellular space to form a double membrane-spanning intercellular channel (Segretain and Falk 2004; Martin and Evans 2004). Two major gene families have evolved to construct gap junction channels, the connexins in vertebrates and the innexins in invertebrates (Söhl and Willecke 2004; Bauer et al. 2005; Phelan 2005). Both gene families encode four-pass membrane domains with two extracellular loops, cytoplasmic N-and C-termini and a cytoplasmic loop domain. Although innexins and connexins are structurally and functionally analogous, they show virtually no sequence similarity to each other. Distantly related innexin homologs, named pannexins, have recently also been identified in mouse and humans (Panchin et al. 2000; Bruzzone et al. 2003; Panchin 2005).

Genome sequencing projects identified eight innexin genes in the fruit fly Drosophila, however, only a few mutants exist that have been molecularly and functionally characterized. It has been shown that innexin1 (ogre) and innexin8 (shakingB) are involved in the physiology of the giant fibre and the adult visual systems (Watanabe and Kankel 1990; Curtin et al. 2002; Krishnan et al. 1993; Phelan et al. 1996, 2005; Shimohigashi and Meinertzhagen 1998; Zhang et al. 1999; Jacobs et al. 2000); innexin4 (zero population growth) controls germ cell differentiation (Tazuke et al. 2002; Gilboa et al. 2003) and we have recently shown that during foregut morphogenesis, innexin2 is a Wingless/WNT target gene that coordinates epithelial morphogenesis (Bauer et al. 2002). Mutants for innexin2 (kropf), which lack both their maternal and zygotic innexin2 contributions fail to develop any epithelial tissues, and overexpression of innexin2 in the epidermis results in multilayering and irregular cell shapes suggesting that innexin2 is a key factor to control epithelial organization during morphogenesis of tissues and organs in the body (Bauer et al. 2004).

In the epidermis, innexin2 protein colocalizes with DE-cadherin and the β-catenin homolog Armadillo. Both proteins are core components of adherens junctions, which are multiprotein complexes mediating cell-cell adhesion and communication (Tepass 1999; Nagafuchi 2001, for reviews). Classic cadherins are major mediators of cell–cell adhesion. Their extracellular domain mediates calcium-dependent homophilic cell–cell adhesion, whereas their highly conserved intracellular domain is linked to the actin cytoskeleton (Tepass 1999). DE-cadherin possesses a highly conserved cytoplasmic domain that binds to the cytoplasmic Armadillo protein, thereby forming a cadherin-catenin complex. Armadillo, also serves as the transducer of the WNT signal/wingless signal (Tepass 1999; Nagafuchi 2001). Cadherins are required to maintain stable adhesion between epithelial cells (Larue et al. 1994; Riethmacher et al. 1995; Tepass et al. 1996; Uemura et al. 1996) but can also be used by migrating cells to adhere to a cellular substratum (Letourneau et al. 1990; Barami et al. 1994; Hazan et al. 2000; Li et al. 2001). In mutants for both zygotic armadillo and DE-cadherin, innexin2 localization is altered and in kropf mutants, components of adherens junctions are mislocalized (Bauer et al. 2004). Further evidence for a more direct interaction between innexin2 and adherens junction proteins was provided by yeast two hybrid analysis and coimmunoprecipitation experiments using embryonic extracts, which showed that innexin2 interacts via its cytoplasmic loop domain with the C-terminus of DE-cadherin (Bauer et al. 2004). These data suggest that the biogenesis of innexin2-containing gap junctions and of adherens junctions may be linked via the interaction of innexin2 and DE-cadherin/Armadillo. However, the molecular mechanism of how these proteins interact with each other is still rather elusive. We have used Drosophila tissue culture cells as a simpler model system to study the interactions of innexin2 and DE-cadherin. Our results provide evidence that DE-cadherin may be an important component to control trafficking and localization of innexin2 to the plasma membrane.

METHODS

Fly Strains

We used standard techniques for fly manipulation. The P allele kropf ^P16 that was used, is a transcript-null allele (Bauer et al. 2002).

Plasmids

pMT/V5-innexin2-GFP and pMT/V5-innexin2-Myc carry an innexin2 cDNA under the control of a copper-inducible promoter (pMT/V5 plasmid from; Invitrogen, Carlsbad, CA). innexin2-GFP cDNA was derived from pUAST-innexin2-GFP, which had been already used for the generation of transgenic flies (Bauer et al. 2004). innexin2-Mycwas generated by PCR reaction from LD11658 containing the entire innexin2cDNA using specific innexin2 primer pairs (innexin2 start-primer: 5′-GGAATTCCAAA ATGTTTGATGTCT TTGGGT -3′; innexin2 stop-primer containing the Myc-tag, which is printed in bold letters: 5′-CGCTCGAGCG GTTA CAGATCCTCTTCAGAGATGAGTTTC TGCTCGGCGTCGAAGGGCCGCTTG-3′). pR- mpRmHa3-DE-cadherin (DE-Cad wt) is also under control of a copper-inducible promoter, and was a kind gift of P. Rorth and A. Pacquelet (Pacquelet et al. 2003).

Cell Culture and DNA Transfection

Drosophila SL2 cells were maintained at 25°C in Drosophila Schneider Medium (Invitrogen) supplemented with 10% fetal bovine serum (heat inactivated), and 1% penicillin/streptomycin (Sigma, St Louis, USA). Cells were subcultured to 3 × 10⁶ cells/ml every week. Transfection was performed using Cellfectin reagent (Invitrogen) according to the manufacturer's instructions. Expression of pMT/V5-innexin2-GFP and pRmHa3 DE-cadherin was induced with 0.7 mM CuSO₄ for 72 h. Stable SL2 cell lines were generated by transfection of pMT/V5-innexin2-GFP (innexin2-GFP cells) or pMT/V5-innexin2-Myc (innexin2-Myc cells) together with a pAcHygro (Invitrogen) vector (ratio of transfected cDNAs; 2: 1 μg) using Cellfectin. Stable transfected cells were selected in SL2 medium containing hygromycin at 300 μg/ml. Subsequently, the cells were cultured continuously in hygromycin containing medium. Most of the cells (usually >70%) in the stable cell lines expressed innexin2-GFP or innexin2-Myc.

Antibodies and Immunohistochemistry

Antibody staining of Drosophila embryos was performed as described earlier (Fuss and Hoch 1998; Bauer et al. 2004). SL2 cells were fixed in 4% paraformaldehyde for 10 min at room temperature (RT), washed twice with PBS containing 1% BSA (PBSB), and permeabilised with PBS supplemented with 0, 2% Triton X-100 for 5 min. Following two washes with PBS and two washes with PBSB, SL2 cells were stained with the corresponding primary antibodies in PBSB for 30 min at RT. After rinsing four times for 5 min each time with PBSB at RT, the samples were processed for indirect immunofluorescence by incubation with the corresponding secondary fluorescent antibody for 30 min at RT. After rinsing three times for 5 min each time with PBSB and one time with PBS, the samples were mounted in Vectashield (hard set) mounting medium with DAPI. The following stains and primary antibodies were used: rabbit anti-Innexin 1, 2, 3 (1:75); goat anti-DE-cadherin (1:50; Santa Cruz); mouse anti-Crumbs (1:10; Hybridoma Bank); rabbit anti-PDI (Phospho-Disulfide Isomerase; 1:100, Calbiochem) as ER marker; monoclonal anti-c-Myc 9E10 (1: 100, Santa Cruz); Alexa Fluor 488 Phalloidin (1:1000, Molecular Probes). Secondary antibodies used were: Alexa Fluor 488 (1:200; MoBiTec); Alexa Fluor 546 (1:200; MoBiTec); and Cy5 (1:100, Dianova).

Microscopy and Image Processing

Fluorescent and DIC (Differential Image Contrast) pictures were taken on a Zeiss Axiovert 200M microscope with ApoTome using Zeiss Axiovision software (Carl Zeiss AG Jena, Germany). Otherwise fluorescent images were recorded using a Leica TSP2 confocal microscope (Leica, Wetzlar, Germany). Images of multilabeled samples were acquired sequentially on separate channels. All images were processed with Adobe Photoshop software.

Immunoprecipitation and Immunoblotting

For immunoprecipitation reactions Drosophila innexin2-Myc cells, were homogenized 72 h after induction with CuSO₄ in RIPA buffer (150 mM NaCl, 1% IGEPAL CA-630, 0.5% Sodium Deoxycholate, 0,1% SDS, 50 mM Tris (pH 8,0) with protease inhibitors (Roche)). Immunoprecipitation was performed using the Immunoprecipitation Starter Pack (Amersham Pharmacia Biotech) according to the manufacturer's instruction. To prevent unspecific binding of proteins to protein G Sepharose, preclearing of the SL2 cell lysate was performed in the presence of protein G Sepharose according the manufacturer's instructions. Cleared supernatant was incubated for 3 h with anti-innexin2 antibody or control IgG, followed by the addition 40 μl protein G-Sepharose to each mixture and incubation for 3 h. The beads were washed three times with excess RIPA-buffer and two times with excess PBS, resuspended in SDS-PAGE sample buffer, and boiled for 3 min. Samples were run on a 12.5% SDS-gel, and transferred to a PVDF-membrane (Immobilon P transfer-membrane; Millipore). The immunoblot membrane was blocked with nonfat dry milk in TBS plus 0.05% Tween (TBST) and then incubated with anti-DE-cadherin (1: 400, St Cruz). After washing, bound antibody was visualized with peroxidase–conjugated donkey antigoat IgG (1.15000) using the ECL system (Amersham Pharmacia).

RT-PCR

Total RNA was prepared from SL2 cells with the Qiagen RNAeasy Kit (Qiagen) according to the manufacturer's instructions. RT-PCR was performed as described by Zinke et al. (2002) with minor modifications. Briefly, 1 μ g of total RNA was used for First strand cDNA reaction using oligo-d(T) primers (Amersham) and Super Script III reverse transcriptase (Gibco/Invitrogen) in a final volume of 20 μl at 50°C for 60 min. After inactivation of the Super Script III reverse transcriptase 80 μl H₂O were added for a final volume of 100 μl. PCR was done on aliquots of this reaction in a total volume of 25 μl with GoTaq™ DNA Polymerase (Promega). As loading control we used actin 5C. The following innexin primers were used in the PCR reactions depicted in Figure 1C: innexin1: 5′-AAACTCGCTCATCACGTCCT-3′, 5′-TGCAGAAGCACGACTCACTC-3′; innexin2: 5′-CCGCCATCTCCTACTCCGAG-3′, 5′-TGTGG GCGCTATGCTCATC-3′; innexin3: 5′-CGGTCAT GTCTCCGTCTCCTT-3′, 5′-CTATTCACTGGTG GTTATCA-3′.

Figure 1 Staining of DE-cadherin in Drosophila embryos, and mRNA expression levels of innexins 1, 2, and 3 in SL2 cells. (A) DE-cadherin (green) staining in a wildtype embryo. (B) Double staining of an innexin2 (kropf) mutant embryo monitoring expression of Crumbs (blue) and DE-cadherin (green). DE-cadherin is localized at the membrane in epithelial cells of wildtype embryos (A). In kropf embryos, the DE-cadherin signal is partially reduced and accumulates at the cortex or in the cytoplasm of epithelial cells. Most of the transmembrane protein Crumbs is still properly localized (B). (C, D) RT-PCR analysis of innexin 1, 2, and 3 mRNA expression level in SL2 cells. Actin 5C was used as loading control. (C) Innexins 1, and 3 are detectable after 30 PCR cycles. The mRNA innexin1 signal is slightly higher compared to the very weak innexin3 signal. Innexin2 clearly shows the highest expression level in SL2 cells compared to innexins 1, and 3. (E) iQ5 Real-Time PCR reaction using the iQ5 Real-Time PCR Detection System (BIO RAD) with innexin1, 2, and 3 primer pairs checked for primer-dimer artifacts and efficiencies confirmed the result depicted in (C) and (D); (SD: Standard deviation).

A second set of innexin1, 2, and 3 primer pairs was designed and tested for primer-dimer artefacts and efficiencies using the iQ5 Optical System Software (BIO RAD). Primer efficiencies determined for the different innexin primer pairs were: innexin1: 100%, (R²: 0.998); innexin2 : 98.3 %, (R²: 0.996); innexin3: 100%, (R²: 0.998).

PCR analysis with the second set of innexin1, 2, and 3 primer pairs was performed on aliquots of the reverse transcribed total RNA (described above) according to the manufacturer's instructions using the iQ5 Real-Time PCR Detection System (BIO RAD). PCR reactions were set up with the iQ™ SYBR^R Green Supermix (BIO RAD), in a final volume of 25 μ l. The following innexin primers were used in the iQ5 Real-Time PCR reaction: innexin1: 5′-AATATCACGATCTGCACCCGC-3′; 5′-CCGA ATGGCGTACAGCTTGT-3′; innexin2: 5′-CCTAC TCCGAGCCCGTTCC-3′; 5′-TGCCCAGCTGA TAGAGCAGG-3′; innexin3: 5′-GATCGGTCCA GAAACACGACA-3′; 5′-GGAGATGGTGGCCA AGATGAT-3′.

RESULTS

In Drosophila melanogaster the apical epithelial region is defined by the subapical complex (SAC) and the zonula adherens (ZA). The SAC complex harbors the transmembrane protein Crumbs (Crb), the four-PDZ domain protein Discs Lost and the MAGUGK protein Stardust (Tepass, 1996; Bhat et al. 1999; Klebes and Knust 2000; Bachmann et al. 2001; Hong et al. 2001). DE-cadherin, α-catenin, and β-catenin/Armadillo are major ZA components (Tepass 1996). Within this apico-lateral region DE-cadherin and β-catenin/Armadillo colocalize with innexin2 (Bauer et al. 2004). In kropf mutants, DE-cadherin (compare Figure 1A and 1B) and β-catenin/ Armadillo (Bauer et al. 2004) are mislocalized. Moreover, in both zygotic armadillo and DE-cadherin mutants, the innexin2 localization is altered (Bauer et al. 2004). To further study the interaction of innexin2 with key proteins of the adherens junction complex such as DE-cadherin, we established a tissue culture system using Drosophila Schneider cells. Most of the established Schneider cell lines such as S2 (Schneider 1972), SL2 (Bruckner et al. 2000), Kc₁₆₇ (Echalier and Ohanessian 1969), and S2R⁺ (Yanagawa et al. 1998) are derived from embryonic hemocytes. Whereas S2, SL2, and Kc₁₆₇ cells are small and round, S2R⁺ cells are large, flat and strongly adherent to glass, and extracellular matrix (Kiger et al. 2003). As we have shown already that Schneider cell lines such as S2 and SL2 express innexin2 and that Armadillo, a component of the adherens junction complex and also a mediator of the WNT/Wg signalling pathway regulates innexin2 expression (Bauer et al. 2002), we started our analysis of innexin2 interactions with key proteins of the adherens junction complex in SL2 cells. This line is derived from a primary culture of 20–24-h-old embryos. SL2 cells are hemocyte-like and phagocytic cells. Under routine culture conditions, SL2 cells display a roughly spherical morphology (diameter of 10 μ m) and a thin rim of cytoplasm surrounds the nucleus (Figure 2).

Figure 2 Localization of innexin 1, 2, and 3 within Drosophila SL2 cells. (A–C) Triple staining of SL2 cells using anti innexin 1, 2, and 3 (red), Phalloidin as actin marker (green), and Dapi as nuclear marker (blue). Innexin proteins are predominantly found in the cytoplasm of the SL2 cells. (D–F) Triple staining of innexin2-Myc cells displaying innexin2-Myc (green), PDI (red), an endoplasmic reticulum marker, and nuclear staining with Dapi (blue). (D) Most of innexin2-Myc is colocalized with the ER.

Towards further characterization of innexin expression levels in SL2 cells, we performed RT-PCR analysis using specific primers (see Methods) for the innexin genes 1, 2 and 3. All of the three innexin genes are expressed in SL2 cells (Figure 1C, D). The mRNA expression level is highest for innexin2 whereas innexin1 and 3 show a significantly lower expression. To exclude the possibility that different primer efficiencies of the different innexin primer pairs might influence RT-PCR analysis, we generated a second set of innexin primers. These primers were checked for primer-dimer artefacts and efficiencies (see Methods). We then used the tested primer pairs in repeat RT-PCR analysis of the relative abundance of innexins1, 2, and 3 in SL2 cells. As shown in Figure 1E, innexin2 expression is highest. An innexin2 signal is detected already after 25 PCR amplification cycles by the iQ5 Real-Time PCR Detection System (BIO RAD), whereas innexin1, and 3 signals are obtained after about 32 PCR amplification cycles. This experiment confirms the result depicted in Figure 1C and Figure 1D. To analyze the protein localization of endogenous innexin proteins in SL2 cells, we performed immunohistochemistry applying antibodies against the innexins 1, 2 and 3 (Bauer et al. 2001, 2003) in combination with DAPI for nuclear labeling, and phalloidin to visualize the actin cytoskeleton (Figure 2). All three innexin proteins are localized within the cytoplasm of SL2 cells, with high level of expression around the nucleus (Figure 2A–C); only a small fraction of the proteins is localized to the membrane. Coimmunostaining with the ER-marker Phospho-Disulfide-Isomerase (PDI; Lee et al. 2001; Urban et al. 2001) and an anti-Myc antibody in stable innexin2-Myc cells, indicates that most of the innexin2 protein is localized to the ER (Figure 2D Figure 2E Figure 2F). The cytoplasmic localization of endogenous innexins in SL2 cells is in contrast to their membrane localization in embryonic epithelia (Bauer et al. 2001, 2003, 2004) suggesting that key factors for trafficking, membrane insertion or for the stabilization of innexin hemichannels in the membranes may be lacking in SL2 cells. It is noteworthy in this context that SL2 cells are not polarized and also grow as single cells. Consistent with this, the cell adhesion molecule DE-cadherin is expressed in most of the SL2 cells at a low level (Figure 3B, Figure 3C); in addition, we often find a large amount of the protein localized to the cytoplasm (Figure 3B).

Figure 3 Addition of DE-cadherin to Drosophihla SL2 cells alters the localization of innexin2. (A) DIC image of SL2 cells. (B) The same picture as depicted in (A) showing DE-cadherin (red) staining of SL2 cells. DE-cadherin is evenly distributed in the cytoplasm. Dapi (blue) was used as nuclear marker. (C) Different DE-cadherin levels of SL2 cells. Only a few cells express strongly DE-cadherin (red), whereas most of the SL2 cells show low DE-cadherin levels. (D, E) innexin2 localization before and after transfection of DE-cadherin. (D) Endogenous innexin2 in SL2 cells resides in the cytoplasm. (E) Addition of DE-cadherin results in an increased localization of innexin2 at the plasma membrane. Noteworthy, innexin2 is significantly enriched at cell-cell contact sites (arrows). (F–J) Distribution of innexin2-GFP and DE-cadherin in a single innexin2-GFP cell before and after the transfection of DE-cadherin. (F) Without addition of DE-cadherin innexin2-GFP is predominantly localized in the cytoplasm. (G–I) After transfection of DE-cadherin innexin2-GFP is predominantly and uniformly localized at the membrane. (G) Innexin2-GFP localization. (H) DE-cadherin distribution in the same cell as depicted in G. (I) Overlay of innexin2-GFP (G) and DE-cadherin (H). (J). 3-D view of the cell depicted in G–I rendered from different confocal sections. Note the high degree of DE-cadherin and innexin2-GFP colocalization. (K) Direct interaction of DE-cadherin and innexin2 in Drosophila SL2 cells. Innexin2-Myc cells either transfected (Lanes 1, 3, and 4) or not transfected with DE-cadherin (lane 2) were used for immunoprecipitation with an anti-innexin2 antibody (Lanes 1, 2, and 3). A control immunoprecipitation was performed using an IgG antibody (Lane 4). To accommodate the fact that innexin2-Myc cell extracts used for immunoprecipitation had either just some endogenous DE-cadherin or additionally transfected DE-Cadherin, endogenous DE-Cadherin and transfected DE-Cadherin are each identified by “+” and “+++”, respectively.

To test whether cell-to-cell contact may be required for correct innexin localization, we increased the amount of DE-cadherin in the cells by transfecting a DE-cad wt expression plasmid (see Methods). As has been previously observed, this causes SL2 cells to aggregate at a high rate (Pacquelet et al. 2003), although a considerable number of cells are still found that do not aggregate. Increasing the level of DE-cadherin, results in an altered localization of innexin2 protein. Whereas innexin2 is localized predominantly to the cytoplasm in untransfected cells (Figure 3D), it localizes to the membrane in transfected cells, which express high levels of DE-cadherin (Figure 3E, compare to D). Furthermore, we find a localized accumulation of innexin2 in membrane domains at cell-to-cell contact sites (arrows in Figure 3E), which are highly reminiscent of gap junction plaques induced by overexpression of connexins in HeLa cells (Lauf et al. 2002; Gaietta et al. 2002).

To investigate whether it is merely cell aggregation that determines the membrane localization of innexin2, we analyzed DE-cadherin transfected innexin2-GFP cells that have not aggregated but still express DE-cadherin at a high level. In these cells, the cytoplasmic localization of innexin2-GFP is also altered. Both DE-cadherin and 2-GFP are predominantly localized to the membrane (Figure 3G–I) as compared to the distribution of 2-GFP in non-DE-cadherin transfected 2-GFP cells (Figure 3F). Noteworthy is also the high degree of 2-GFP and DE-cadherin colocalization in this cell (Figure 3I, and Figure 3J). Coimmunoprecipitation experiments in innexin2-Myc cells transfected with DE-cadherin provided strong evidence for a direct interaction of DE-cadherin and innexin2 (Figure 3K). Recently our genetic and biochemical interaction studies in Drosophila embryos have already shown a direct interaction of DE-cadherin and innexin2 (Bauer et al. 2004). Recapitulating, these data suggest that DE-cadherin may control trafficking and innexin2 localization to the plasma membrane in SL2 cells.

DISCUSSION

The assembly of the junctional complement of a particular cell is a dynamic process that is essential for the proper differentiation of cells and tissues during development(Tepass and Hartenstein 1994). It has been demonstrated previously that the spatial association between gap and adherens junctions is an important factor in the maturation of vertebrate and invertebrate tissues (Angst et al. 1997; Tepass and Hartenstein 1994). Numerous studies in mammalian and avian tissue culture cells have provided evidence for the importance of cell adhesion contacts for gap junction formation (see Wei et al. 2004, for an extensive review). It was demonstrated in embryonic chick lens cultures, in adult rat cardiomyocytes, and in reaggregated Novikoff cells that N-cadherin is required for gap junction assembly (Frenzel and Johnson 1996; Hertig et al. 1996; Meyer et al. 1992; Ko et al. 2000) and a direct interaction of the Armadillo homologue β-catenin with connexin43 was shown to occur in cardiac myocytes (Ai et al. 2000). Similarly, tissue culture studies and the analysis of neural crest cell migration in the mouse embryo have shown that the assembly and function of gap junctions is directly controlled by E-and N-cadherins, respectively (Jongen et al. 1991; Musil et al. 1990; Fujimoto et al. 1997; Xu et al. 2001). Inhibition of cadherin function, such as in N-cadherin-deficient myocytes, can disrupt gap junction formation and inhibit cell-cell coupling, suggesting that localization of cadherin to cell-cell contact sites may be a prerequisite for gap junction formation (Meyer et al. 1992; Luo and Radice 2003). Conversely, inhibition of connexin43 can block adherens junction formation (Zuppinger et al. 2000). The formation of connexin channels and the development of junctional communication is also promoted by NCAM-mediated adhesion in neuroectoderm of the chick (Keane et al. 1998).

Also in Drosophila, the biogenesis of innexin2-containing gap junctions and of adherens junctions may be linked via the interaction of innexin2 and DE-cadherin/Armadillo. It was demonstrated in the embryonic epidermis that innexin2 protein colocalizes with the adherens junction proteins Armadillo and DE-cadherin. Kropf mutant embryos lacking both the maternal and zygotic innexin2 contribution fail to develop any epithelial tissues (Bauer et al. 2004). Embryos in which innexin2 has been extensively overexpressed are embryonic lethal and display defects such as multilayering and irregular cell shapes (Bauer et al. 2004). In mutants for both zygotic armadillo and DE-cadherin innexin2 localization is altered and in innexin2 overexpression experiments, Armadillo and DE-cadherin are organized into the innexin2 pattern (Bauer et al. 2004). Further evidence for a more direct interaction between innexin2 and adherens junction proteins was provided by yeast two hybrid analysis and coimmunoprecipitation experiments using embryonic extracts, which showed that innexin2 interacts via its cytoplasmic loop domain with the C-terminus of DE-cadherin (Bauer et al. 2004).

To study further the interaction of innexin2 with key proteins of the adherens junction complex such as DE-cadherin, we established a tissue culture system using Drosophila SL2 cells. Our RT-PCR experiments and our antibody staining show that innexin1, 2, and 3-mRNAs and -proteins are expressed in SL2 cells. However, the endogenous innexin proteins are localized within the cytoplasm of SL2 cells with high level of expression around the nucleus and only a small fraction of the proteins is localized to the membrane (Figure 2). SL2 cells lack cell polarity and grow mostly as single cells. The cell adhesion molecule DE-cadherin is expressed in most of the SL2 cells at a quite low level (Figure 3C). Increasing the amount of DE-cadherin in the cells causes SL2 cells to aggregate and results in the displacement of innexin2 from the cytoplasm to the membrane (Figure 3E); innexin2 accumulates locally at cell-to-cell contact sites, a pattern which is highly reminiscent of gap junction plaques induced by overexpression of connexins in HeLa cells (Lauf et al. 2002; Gaietta et al. 2002). The membrane localization of innexin2 is also achieved in single DE-cadherin expressing cells (Figure 3F, Figure 3G, Figure 3H, Figure 3I, Figure 3J), which are nonaggregated and lack cell-to-cell contact sites. innexin2 shows a uniform distribution in the membrane of these cells (Figure 3F, Figure 3G, Figure 3H, Figure 3I, Figure 3J) suggesting that DE-cadherin may control trafficking of innexin2 to the plasma membrane in SL2 cells; cell-to-cell contact seems to be required for the accumulation of innexin2 in gap junction plaques (Figure 3E). It is known that E-cadherin is trafficked to and from the cell surface by exocytic and multiple endocytic pathways (Bryant and Stow 2005, for review).

During trafficking to and from the cell surface, cadherin–catenin complexes are dynamically assembled and disassembled. β-catenin binds to E-cadherin early in the biosynthetic pathway and the two proteins are trafficked and sorted as a complex for delivery to the epithelial-cell surface, as demonstrated in Madin–Darby canine kidney cells (Chen et al. 1999; Miranda et al. 2003). Since E-cadherin is intimately involved in the establishment and maintenance of epithelial-cell polarity, it is sorted and delivered in a polarized manner to the lateral cell surface. Putative sorting signals have been characterized within the cytoplasmic tail of E-cadherin (Bryant and Stow 2005, for review). The direct interactions of innexin2 with DE-cadherin, which we have shown previously by yeast two hybrid and immunoprecipitation analysis, (Bauer et al. 2004) suggest the possibility that innexin2 may be part of the cadherin-catenin complexes during trafficking. Experiments in rat cardiomyocytes have indicated that the formation of a catenin-ZO1-connexin43 complex is required for connexin43 transport to the plasma membrane during the assembly of gap junctions (Wu et al. 2003). Furthermore, it was recently shown that the mammalian connexin43α 1 coassembles in a multiprotein complex containing N-cadherin and various N-cadherin associated proteins in NIH3T3 cells (Wei et al. 2005). By using siRNA knock down, it was demonstrated that cell surface expression of connexin43α 1 requires N-cadherin and conversely, that N-cadherin surface expression requires connexin43α 1. Pulse chase labeling and biotinylation experiments indicated that in the absence of N-cadherin, connexin43α 1 trafficking is blocked, suggesting that the intracellular coassembly of connexins and cadherin is required for gap junction and adherens junction formation (Wei et al. 2005). Thus in summary, there seems an intimate linkage between the assembly of connexin-and innexin-containing gap junctions and adherens junctions in both vertebrates and in Drosophila. Our results suggest that also in the Drosophila epidermis, common trafficking routes of adherens junction and gap junction proteins within cells may ensure the positioning of innexin-containing hemichannels in membrane domains close to adherens junctions.

ACKNOWLEDGMENTS

We thank P. Rorth and A. Paquelet for the pRmHa3-DE-cadherin construct, F. Eckardt for technical assistance, and F. Josten for his assistance with the microscopes. Also we want to thank Prof. Waldemar Kolanus for helpful comments. The work was supported by a DFG grant to M.H. (SFB 645).