Cortactin Associates with the Cell-Cell Junction Protein ZO-1 in both Drosophila and Mouse
1998; Elsevier BV; Volume: 273; Issue: 45 Linguagem: Inglês
10.1074/jbc.273.45.29672
ISSN1083-351X
AutoresTakanori Katsube, Manabu Takahisa, Ryu Ueda, Naoko Hashimoto, Mieko Kobayashi, Shin Togashi,
Tópico(s)Signaling Pathways in Disease
ResumoCortactin is an actin filament-binding protein localizing at cortical regions of cells and a prominent substrate for Src family protein-tyrosine kinases in response to multiple extracellular stimuli. Human cortactin has been identified as a protein product of a putative oncogene, EMS1. In this report, we describe the identification of a Drosophila homolog of cortactin as a molecule that interacts with Drosophila ZO-1 using yeast two-hybrid screening. Drosophila cortactin is a 559-amino acid protein highly expressed in embryos, larvae, and pupae but relatively underexpressed in adult flies. Deletion and substitution mutant analyses revealed that the SH3 domain of Drosophilacortactin binds to a PXXP motif in the proline-rich domain of Drosophila ZO-1. Colocalization of these proteins at cell-cell junction sites was evident under a confocal laser-scanning microscope. In vivo association was confirmed by coimmunoprecipitation of cortactin and ZO-1 from Drosophilaembryo lysates. We also demonstrate an association for each of the murine homologs by immunoprecipitation analyses of mouse tissue lysates. Our previous work has demonstrated the involvement of ZO-1 in a signaling pathway that regulates expression of the emcgene in Drosophila. The potential roles of the cortactin·ZO-1 complex in cell adhesion and cell signaling are discussed. Cortactin is an actin filament-binding protein localizing at cortical regions of cells and a prominent substrate for Src family protein-tyrosine kinases in response to multiple extracellular stimuli. Human cortactin has been identified as a protein product of a putative oncogene, EMS1. In this report, we describe the identification of a Drosophila homolog of cortactin as a molecule that interacts with Drosophila ZO-1 using yeast two-hybrid screening. Drosophila cortactin is a 559-amino acid protein highly expressed in embryos, larvae, and pupae but relatively underexpressed in adult flies. Deletion and substitution mutant analyses revealed that the SH3 domain of Drosophilacortactin binds to a PXXP motif in the proline-rich domain of Drosophila ZO-1. Colocalization of these proteins at cell-cell junction sites was evident under a confocal laser-scanning microscope. In vivo association was confirmed by coimmunoprecipitation of cortactin and ZO-1 from Drosophilaembryo lysates. We also demonstrate an association for each of the murine homologs by immunoprecipitation analyses of mouse tissue lysates. Our previous work has demonstrated the involvement of ZO-1 in a signaling pathway that regulates expression of the emcgene in Drosophila. The potential roles of the cortactin·ZO-1 complex in cell adhesion and cell signaling are discussed. membrane-associated guanylate kinase homologs guanylate kinase Drosophila ZO-1 extramacrochaetae Drosophila cortactin glutathioneS-transferase polyacrylamide gel electrophoresis kilobase(s). Cell-cell adhesions are essential for the development of the multicellular organisms. Among the proteins composing the cell-cell adhesion complexes, members of the membrane-associated guanylate kinase homologs (MAGUKs)1 are widely found in Hydra, Caenorhabditis elegans, Drosophila, and mammals (1Anderson J.M. Curr. Biol. 1996; 6: 382-384Abstract Full Text Full Text PDF PubMed Scopus (219) Google Scholar, 2Woods D.F. Bryant P.J. Mech. Dev. 1993; 44: 85-89Crossref PubMed Scopus (189) Google Scholar, 3Kim S.K. Curr. Opin. Cell Biol. 1995; 7: 641-649Crossref PubMed Scopus (99) Google Scholar). MAGUKs have distinctive domains including one or three copies of the PDZ domain, an SH3 domain, and a domain homologous to guanylate kinase (GUK) and implicated in both formation of cell-cell junctions and signal transduction. One of the most intensively characterized members of the MAGUKs is the mammalian ZO-1, which is known to associate with several cellular proteins including the components of cell-cell junctions (occludin, β-catenin, and ZO-2) and the components of cytoskeletal networks (α-spectrin and actin filaments (F-actin)) (4Stevenson B.R. Siliciano J.D. Mooseker M.S. Goodenough D.A. J. Cell Biol. 1986; 103: 755-766Crossref PubMed Scopus (1285) Google Scholar, 5Itoh M. Nagafuchi A. Yonemura S. Kitani-Yasuda T. Tsukita S. Tsukita S. J. Cell Biol. 1993; 121: 491-502Crossref PubMed Scopus (498) Google Scholar, 6Itoh M. Yonemura S. Nagafuchi A. Tsukita S. Tsukita S. J. Cell Biol. 1991; 115: 1449-1462Crossref PubMed Scopus (205) Google Scholar, 7Anderson J.M. Stevenson B.R. Jesaitis L.A. Goodenough D.A. Mooseker M.S. J. Cell Biol. 1988; 106: 1141-1149Crossref PubMed Scopus (285) Google Scholar, 8Furuse M. Itoh M. Hirase T. Nagafuchi A. Yonemura S. Tsukita S. Tsukita S. J. Cell Biol. 1994; 127: 1617-1626Crossref PubMed Scopus (808) Google Scholar, 9Itoh M. Nagafuchi A. Moroi S. Tsukita S. J. Cell Biol. 1997; 138: 181-192Crossref PubMed Scopus (570) Google Scholar). While ZO-1 has been considered as a homolog of a Drosophila tumor suppresser Dlg, its biological functions in the cell-cell junction and signal transduction remain obscure (10Hough C.D. Woods D.F. Park S. Bryant P.J. Genes Dev. 1997; 11: 3242-3253Crossref PubMed Scopus (123) Google Scholar, 11Woods D.F. Bryant P.J. Cell. 1991; 66: 451-464Abstract Full Text PDF PubMed Scopus (769) Google Scholar). We recently identified a new Drosophila MAGUK protein, Tamou, and reported its significant homology with ZO-1 (12Takahisa M. Togashi S. Suzuki T. Kobayashi M. Murayama A. Kondo K. Miyake T. Ueda R. Genes Dev. 1996; 10: 1783-1795Crossref PubMed Scopus (61) Google Scholar). We will refer to Tamou as Drosophila ZO-1 (DZO-1) because we also found that the transgenes of mouse ZO-1 could replace the tam gene function in Drosophila. 2S. Togashi, M. Takahisa, and R. Ueda, unpublished results. TheDZO-1 tam1 mutant flies exhibit the supernumerary mechanosensory organs. This is a similar phenotype to that of anextramacrochaetae (emc) mutation. Theemc gene encodes a helix-loop-helix type transcriptional regulator and negatively regulates specification of sensory organ precursor cells (13Ellis H. Spann D. Posakony J. Cell. 1990; 61: 27-38Abstract Full Text PDF PubMed Scopus (288) Google Scholar, 14Garrell J. Modolell J. Cell. 1990; 61: 39-48Abstract Full Text PDF PubMed Scopus (268) Google Scholar, 15Van Doren M. Ellis H.M. Posakony J.W. Development. 1991; 113: 245-255PubMed Google Scholar, 16Van Doren M. Powell P. Pasternak D. Singson A. Posakony J. Genes Dev. 1992; 6: 2592-2605Crossref PubMed Scopus (152) Google Scholar). We have previously shown thatDZO-1 locates at cell-cell junctions and is involved in the signaling pathway, which activates the transcription of emc(12Takahisa M. Togashi S. Suzuki T. Kobayashi M. Murayama A. Kondo K. Miyake T. Ueda R. Genes Dev. 1996; 10: 1783-1795Crossref PubMed Scopus (61) Google Scholar). Toward the elucidation of the DZO-1 functions in the signaling pathway, we performed a yeast two-hybrid screen to identify the Drosophila proteins that interact with DZO-1. One of the obtained cDNA clones is encoding a protein highly homologous to vertebrate cortactin. Cortactin is an F-actin binding protein initially discovered as a prominent substrate for Src protein-tyrosine kinase (17Wu H. Reynolds A.B. Kanner S.B. Vines R.R. Parsons J.T. Mol. Cell. Biol. 1991; 11: 5113-5124Crossref PubMed Scopus (375) Google Scholar, 18Wu H. Parsons J.T. J. Cell Biol. 1993; 120: 1417-1426Crossref PubMed Scopus (452) Google Scholar). A human homolog was identified as a protein product of a putative oncogene,EMS1, and has been implicated in both cell adhesion and cell signaling (19Schuuring E. Verhoeven E. Litvinov S. Michalides R.J. Mol. Cell. Biol. 1993; 13: 2891-2898Crossref PubMed Scopus (151) Google Scholar). We now report the primary structure of Drosophila cortactin (DCortactin) and show multiple lines of evidence suggesting that DCortactin interacts with DZO-1. We also present results demonstrating the association between cortactin and ZO-1 in mouse tissues. In association with ZO-1, cortactin may play important roles in the formation and/or regulation of cell-cell adhesion and communication during growth, differentiation, and tumorigenesis. pBD-DZO885–1367 was constructed as follows. The XhoI end of the 1.5-kb NcoI-XhoI fragment of the DZO-1 cDNA (12Takahisa M. Togashi S. Suzuki T. Kobayashi M. Murayama A. Kondo K. Miyake T. Ueda R. Genes Dev. 1996; 10: 1783-1795Crossref PubMed Scopus (61) Google Scholar) was blunted and ligated to pAS2–1 (CLONTECH) digested with NcoI and SmaI. Deletion variants were constructed by double digesting pBD-DZO885–1367 with some pairs of restriction enzymes, one cut it within the DZO-1 coding region and another within the multi-cloning site, followed by blunt-ending before self-ligating. The SmaI, PinAI,Van91I, and AspI sites within theDZO-1 coding region and the BamHI,NcoI, and NdeI sites within the multi-cloning site were used for these construction. Point mutations were introduced into pBD-DZO1115–1253 by polymerase chain reaction-mediated site-directed mutagenesis (20Higuchi R. Krummel B. Saiki R.K. Nucleic Acids Res. 1988; 16: 7351-7367Crossref PubMed Scopus (2102) Google Scholar). The synthetic oligonucleotide primers used for mutagenesis were as follows: 5′-gggtgctggcaccgccttgaagggt-3′ and 5′-tcaaggcggtgccagcacccaaacc-3′ for pBD-DZO1115–1253M1, 5′-attcgcagaacgggccgaaccagag-3′ and 5′-gttcggcccgttctgcgaatggctt-3′ for pBD-DZO1115–1253M2, 5′-cgaagccgttggtgcataatagaag-3′ and 5′-attatgcaccaacggcttcgcatca-3′ for pBD-DZO1115–1253M3, 5′-cgaagcacgctgtgctgtgccatt-3′ and 5′-gcacagcacagcgtgcttcgaatct-3′ for pBD-DZO1115–1253M4, and 5′-caaagccttgcgtgcgctgctttcc-3′ and 5′-gcagcgcacgcaaggctttggtgga-3′ for pBD-DZO1115–1253M5. pGST407BtXm was constructed by inserting the blunt-ended 1.9-kbBstEII-XmnI fragment of the DCortactin cDNA into the SmaI site of pGEX-5X-3 (Pharmacia Biotech, Uppsala, Sweden). To construct pMBPMCort, mouse cortactin cDNA was amplified by reverse transcriptase-polymerase chain reaction and digested with PstI. The resulting 7.5-kb PstI fragment was blunt-ended and inserted into the XmnI site of pMAL-TER, which is a derivative of pMAL-c2 vector (New England Biolabs) containing stop codons for all three reading frames downstream from a multi-cloning site. The nucleotide sequences of the primers used for the reverse transcriptase-polymerase chain reaction were 5′-cattgatcatcgcagatgcc-3′ and 5′-tgctctgtagtgtgaaccct-3′. Two-hybrid screening was performed using a Matchmaker two-hybrid system (CLONTECH). Each of the bait constructs and a Drosophila melanogaster(Canton-S wild type) 0- to 18-h embryo cDNA library constructed in pGAD10 (CLONTECH) were co-introduced simultaneously into yeast strain CG1945. Transformants were selected for growth on plates lacking histidine and supplemented with 5 mm 3-aminotriazole. Plasmids were recovered and reintroduced into yeast strain Y187 to confirm the interaction by quantitative β-galactosidase liquid assay utilizing a luminescent β-galactosidase detection kit (CLONTECH) and a Lumat LB9501 luminometer (Berthold, Bad Wildbad, Germany). The interaction between the DZO-1 C-terminal domain variants and the DCortactin SH3 domain was also tested by the same assay. Northern blot analysis was done as described previously (12Takahisa M. Togashi S. Suzuki T. Kobayashi M. Murayama A. Kondo K. Miyake T. Ueda R. Genes Dev. 1996; 10: 1783-1795Crossref PubMed Scopus (61) Google Scholar). A Canton-S 4- to 8-h embryo cDNA library in a pNB40 plasmid vector was screened with a32P-labeled 1.8-kb EcoRI fragment from pcT407A. The nucleotide sequences of the plasmid clones were determined using an ABI Prism 377 DNA Sequencer (Perkin Elmer). Rabbit and rat antibodies againstDCortactin (amino acid residues 8–467) were directed against a bacterially expressed glutathione S-transferase (GST) fusion protein encoded by pGST407BtXm. Anti-DZO-1 polyclonal antibody, anti-mouse ZO-1 monoclonal antibody, and anti-DE-cadherin monoclonal antibody have been described previously (6Itoh M. Yonemura S. Nagafuchi A. Tsukita S. Tsukita S. J. Cell Biol. 1991; 115: 1449-1462Crossref PubMed Scopus (205) Google Scholar, 12Takahisa M. Togashi S. Suzuki T. Kobayashi M. Murayama A. Kondo K. Miyake T. Ueda R. Genes Dev. 1996; 10: 1783-1795Crossref PubMed Scopus (61) Google Scholar, 21Oda H. Uemura T. Harada Y. Iwai Y. Takeichi M. Dev. Biol. 1994; 165: 716-726Crossref PubMed Scopus (372) Google Scholar). Anti-chicken p80/85 (cortactin) monoclonal antibody was purchased from Upstate Biotechnology (Lake Placid, NY). Samples separated by SDS-PAGE were transferred electrophoretically to an Immobilon-P membrane (Millipore Corp.). Immunodetection was performed with horseradish peroxidase-conjugated secondary antibody and enhanced chemiluminescent substrate (SuperSignal substrate, Pierce). Immunostainings of wing discs from full grown Canton-S wild-type larvae were done as described previously (12Takahisa M. Togashi S. Suzuki T. Kobayashi M. Murayama A. Kondo K. Miyake T. Ueda R. Genes Dev. 1996; 10: 1783-1795Crossref PubMed Scopus (61) Google Scholar). Rabbit anti-DZO-1 antiserum, rabbit anti-DCortactin antiserum, rat anti-DCortactin antiserum, and rat anti-DE-cadherin monoclonal antibody were used in 1:50–200 dilution. TRITC-conjugated phalloidin (Sigma) was used for staining F-actin. Visualization of antibodies was performed using fluorescein isothiocyanate-conjugated donkey anti-rabbit Ig antibody (Amersham Life Science, Buckinghamshire, United Kingdom), fluorescein isothiocyanate-conjugated goat anti-rat Ig antibody (Organon Teknika, West Chester, PA), and rhodamine-conjugated goat anti-rat Ig antibody (Organon Teknika). Specimens were observed with a Bio-Rad MRC-1024 laser scanning confocal microscope (Bio-Rad). Dechorionated 0- to 16-hCanton-S wild-type embryos or tissues of 1–3-day postnatal mice (ICR) were homogenized in 25 volumes of cold lysis buffer (50 mm Tris-HCl, pH 7.5, 150 mm NaCl, 1.0% Nonidet P-40, 1 mm EDTA, 10 μm benzamidine, 1 mm phenylmethylsulfonyl fluoride, and 10 μg/ml each of pepstatin A, leupeptin, and aprotinin). The homogenate was centrifuged at 15,000 × g for 15 min to pellet debris. The supernatant was then spun at 140,000 × g for 1 h at 4 °C. The resulting supernatant was precleared by incubating with a 250-μl settled volume of protein A-Sepharose beads (Pharmacia Biotech, Uppsala, Sweden) per ml of supernatant for 1 h at 4 °C. The beads were pelleted, and the supernatant was saved as precleared extract. 400 μl of precleared extracts were incubated with 1–4 μl of rabbit anti-DZO-1 antiserum, anti-DCortactin antiserum, or preimmune rabbit sera at 4 °C for 3–12 h. Immune complexes were recovered by adding 20 μl of protein A-Sepharose, followed by incubation at 4 °C for 2–8 h with rocking. The beads were washed three times with cold lysis buffer, boiled in SDS sample buffer, and processed for Western analysis. By a yeast two-hybrid screen, a cDNA clone pcT407A was isolated as a clone encoding a polypeptide that interacts specifically with the C-terminal 483 amino acids region of DZO-1 (Fig. 1). Developmental Northern analysis of the Canton-S wild type, using a 0.85-kb insert of pcT407A as a probe, represented a 2.7-kb transcript in all developmental stages (Fig. 2). To isolate a full protein-coding cDNA, we screened a Canton-S wild-type embryo cDNA library by the same probe. Several positive clones were isolated, and the longest insert was sequenced. This insert has a single long open reading frame encoding a deduced protein of 559 amino acids with a calculated molecular mass of 61 kDa (Fig. 3). A data base search found that the deduced protein is a new member of a protein family consisting of cortactin and HS1 (17Wu H. Reynolds A.B. Kanner S.B. Vines R.R. Parsons J.T. Mol. Cell. Biol. 1991; 11: 5113-5124Crossref PubMed Scopus (375) Google Scholar, 19Schuuring E. Verhoeven E. Litvinov S. Michalides R.J. Mol. Cell. Biol. 1993; 13: 2891-2898Crossref PubMed Scopus (151) Google Scholar, 22Kitamura D. Kaneko H. Miyagoe Y. Ariyasu T. Watanabe T. Nucleic Acids Res. 1989; 17: 9367-9379PubMed Google Scholar, 23Zhan X. Hu X. Hampton B. Burgess W.H. Friesel R. Maciag T. J. Biol. Chem. 1993; 268: 24427-24431Abstract Full Text PDF PubMed Google Scholar, 24Miglarese M.R. Mannion-Henderson J. Wu H. Parsons J.T. Bender T.P. Oncogene. 1994; 9: 1989-1997PubMed Google Scholar, 25Kitamura D. Kaneko H. Taniuchi I. Akagi K. Yamamura K. Watanabe T. Biochem. Biophys. Res. Commun. 1995; 208: 1137-1146Crossref PubMed Scopus (30) Google Scholar, 26Takemoto Y. Furuta M. Li X.K. Strong-Sparks W.J. Hashimoto Y. EMBO J. 1995; 14: 3403-3414Crossref PubMed Scopus (78) Google Scholar) and is more closely related to cortactin than to HS1 (Fig. 4). Thus, we concluded that the isolated cDNA encodes a Drosophilahomolog of cortactin (DCortactin).Figure 2Developmental Northern blot analysis of DCortactin transcript. 5 μg of poly(A)+RNA from 0–20-h embryos (lane 1), 3rd instar larvae (lane 2), pupae (lane 3), and adult flies (lane 4) of the Canton-S wild type were loaded. The blot was hybridized with the 32P-labeled 0.85-kbEcoRI fragment from pcT407A. The same blot probed with theras2 cDNA fragment is shown below. The positions of RNA markers are indicated on the left in kb.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 3A DCortactin cDNA and the predicted amino acid sequences. The 2413-base pair cDNA sequence includes a 173-base pair 5′-untranslated region, a 1677-base pair open reading frame, and a 548-base pair 3′-untranslated region. The nucleotide sequence of the 0.85-kb cDNA derived from the clone pcT407A matched the boxed region exactly. The deduced protein contains four direct repeats of a 37-amino acid unit (indicated by anunderline), an approximately 50-amino acid region similar to the predicted alpha-helical domain of vertebrate cortactin (broken underline), a proline-rich domain (21% proline content in 167 amino acids; dotted line), and an SH3 domain (double-underline). This feature is similar to vertebrate cortactin. The consensus polyadenylation signal sequence is shown aswhite letters on black background.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 4DCortactin is structurally related to vertebrate cortactin and HS1. a, schematic representation of the comparison of DCortactin (DCort) with human cortactin (H-Cort) and human HS1 (H-HS1). Vertebrate cortactin and DCortactin contains the 37-amino acid repeats (indicated by the shaded boxes), a predicted alpha-helical domain (hatched boxes), a proline-rich domain (dotted boxes), and a C-terminal SH3 domain (striped boxes). HS1 does not have a predicted alpha-helical domain but has a proline-glutamate repeat (closed box) just behind the proline-rich domain. This feature is not found in either vertebrate cortactin or DCortactin. The total numbers of amino acid residues for the three proteins are shown at the right. The percentage of identities of the N-terminal region, the 37-amino acid repeated units, and the SH3 domain of DCortactin with those of human cortactin or human HS1 are shown between the respective bars. Scores between human cortactin and human HS1 are indicated below the bar of HS1.b, the SH3 domain of the DCortactin is more closely related to that of cortactin than to that of HS1. The sequence of the C-terminal region of DCortactin (amino acid residues 498–559) was aligned with that of human cortactin (492–550) and human HS1 (428–486). Conserved amino acid residues are represented bydots.View Large Image Figure ViewerDownload Hi-res image Download (PPT) The cDNA clone pcT407A encodes the C-terminal 96 amino acids of DCortactin (Fig. 3, boxed region). An SH3 domain of approximately 60 amino acids dominates this region, and no other significant structural motif is found. The SH3 domain is known to bind to a PXXP motif often found in proline-rich regions (27Cohen G.B. Ren R. Baltimore D. Cell. 1995; 80: 237-248Abstract Full Text PDF PubMed Scopus (925) Google Scholar). To determine a region withinDZO-1 sufficient to interact with the C-terminal SH3 domain of DCortactin, we constructed a series of deletion mutants of DZO885–1367 (Fig. 5). Their ability to interact with the DCortactin SH3 domain in yeast was analyzed by monitoring β-galactosidase activity. The 139-amino acid region extending from amino acid 1115 to 1253 (DZO1115–1253) was sufficient for the interaction. This region contains four isolated PXXP motifs and three overlapping motifs (PFKPVPPPKP). To identify the PXXP motif necessary for the interaction, we introduced a series of point mutations into DZO1115–1253. The mutation that substitutes two proline residues of the central PXXP within the three overlapping motifs at amino acid residues 1192–1201 (PFKAVPAPKP,DZO1115–1253M1) abolished the interaction with theDCortactin SH3 domain, whereas mutations altering other isolated PXXP motifs did not. These results show that theDCortactin SH3 domain binds to the PXXP motif located at the center of the DZO-1 C-terminal proline-rich domain, where three PXXP motifs are clustered. Rabbit and rat antisera were raised against a GST-DCortactin fusion protein. Western blot analysis with the rabbit antiserum showed two prominent proteins with molecular masses of 105 and 110 kDa in all developmental stages (Fig. 6). The preimmune serum did not react with them. 3T. Katsube, M. Takahisa, R. Ueda, N. Hashimoto, M. Kobayashi, and S. Togashi, unpublished results. The expression of these proteins in adults was substantially less than those in other developmental stages. Western blotting with the rat antiserum gave almost the same profile.3 The molecular weights of these proteins determined by SDS-PAGE are significantly larger than the predicted value (61 kDa). Similar observations of two forms of protein products with anomalous electrophoretic mobility have also been reported for vertebrate cortactin (17Wu H. Reynolds A.B. Kanner S.B. Vines R.R. Parsons J.T. Mol. Cell. Biol. 1991; 11: 5113-5124Crossref PubMed Scopus (375) Google Scholar, 19Schuuring E. Verhoeven E. Litvinov S. Michalides R.J. Mol. Cell. Biol. 1993; 13: 2891-2898Crossref PubMed Scopus (151) Google Scholar, 24Miglarese M.R. Mannion-Henderson J. Wu H. Parsons J.T. Bender T.P. Oncogene. 1994; 9: 1989-1997PubMed Google Scholar). The proline-rich domains may be responsible for their anomalous electrophoretic mobility. To examine the cellular localization, DCortactin was immunostained in epithelial cells of imaginal discs. The typical honeycomb-like images indicated that the protein distributes in a cell-cell contact-associated manner. To clarify the subcellular localization, the double stainings of DCortactin withDZO-1, F-actin, and DE-cadherin were conducted using a laser-scanning confocal microscope. DE-cadherin is a component of the adherens junction and localizes at the apicolateral region of epithelial cell junctions (21Oda H. Uemura T. Harada Y. Iwai Y. Takeichi M. Dev. Biol. 1994; 165: 716-726Crossref PubMed Scopus (372) Google Scholar, 28Uemura T. Oda H. Kraut R. Hayashi S. Kotaoka Y. Takeichi M. Genes Dev. 1996; 10: 659-671Crossref PubMed Scopus (256) Google Scholar). The distribution of DZO-1 partially overlaps with that of DE-cadherin and extends to the slightly basal region corresponding to the site of the septate junction (12Takahisa M. Togashi S. Suzuki T. Kobayashi M. Murayama A. Kondo K. Miyake T. Ueda R. Genes Dev. 1996; 10: 1783-1795Crossref PubMed Scopus (61) Google Scholar). Colocalization of DCortactin,DZO-1, and DE-cadherin was evident, while the staining area of DCortactin in the periplasm seemed slightly broader than those of DZO-1 and DE-cadherin (Fig. 7, c and i). Colocalization of DCortactin and F-actin in a periplasmic region was also observed (Fig. 7 f). Regarding the apical-basal axis, the distribution of DCortactin extended from the basal half side of the adherens junction to the more baso-lateral region (Fig. 7, j, k, and l). To examine thein vivo association of DZO-1 and DCortactin, we immunoprecipitated Canton-Swild-type embryo lysates with the rabbit anti-DCortactin antiserum, the rabbit anti-DZO-1 antiserum, or the respective preimmune sera and analyzed the precipitates by Western blotting (Fig. 8). Western blotting with the affinity-purified anti-DZO-1 antibody detected a 160-kDa protein in the precipitate of the anti-DZO-1 antiserum (lane 1). This protein was coprecipitated by the anti-DCortactin antiserum (lane 3) but not by the corresponding preimmune serum (lane 2). An extra band of about 100 kDa, detected in lanes 2 and 3, was not precipitated by the anti-DZO-1 antiserum and is thought to be not specific for the anti-DZO-1 antibody. Western blotting with rat anti-DCortactin antiserum revealed that the DCortactin 105-kDa form was specifically coprecipitated by the anti-DZO-1 antiserum (compare lanes 5 and 6). These results clearly proved that DCortactin associates with DZO-1 in Drosophila embryo cells. The absence of the DCortactin 110-kDa form in the precipitate of the anti-DZO-1 antiserum is thought to be due to the instability of that form in lysate because it also disappeared in the diluted lysate (lane 4). However, we cannot rule out the possibility that the 110-kDa form cannot associate withDZO-1. We found that rabbit anti-DCortactin antibody could also react with a bacterially expressed protein containing the mouse cortactin 37-amino acid repeat domain fused to maltose binding protein.3 Using this antibody, Western blot analysis of tissue lysates from a 4-day postnatal mouse detected 80- and 85-kDa proteins (Fig. 9 a). These proteins are fairly abundant in brain and testis but not so in liver and kidney. Western blot analysis using the rat anti-DCortactin antiserum also yielded the same pattern as that using an anti-chicken p80/85 (cortactin) monoclonal antibody,3 which was reported to cross-react with mouse cortactin (17Wu H. Reynolds A.B. Kanner S.B. Vines R.R. Parsons J.T. Mol. Cell. Biol. 1991; 11: 5113-5124Crossref PubMed Scopus (375) Google Scholar, 24Miglarese M.R. Mannion-Henderson J. Wu H. Parsons J.T. Bender T.P. Oncogene. 1994; 9: 1989-1997PubMed Google Scholar). These results indicate that both the rabbit and rat anti-DCortactin antibodies can cross-react with mouse cortactin. To examine the interaction between mouse cortactin and mouse ZO-1, we conducted immunoprecipitation analysis of brain and testis lysates with the rabbit anti-DCortactin antiserum. Western blotting of a mouse tissue lysate with anti-mouse ZO-1 monoclonal antibody showed one faint and two prominent bands with molecular masses of 210, 200, and 190 kDa, respectively (Fig. 9 b, lane 1). A 200-kDa protein was detected in the immunoprecipitates of the anti-DCortactin antiserum from brain and testis lysates of 1–3-day postnatal mice (lanes 3 and 5) but not in those of the pre-immune serum (lanes 2 and 4). These results clearly show that mouse cortactin associates with mouse ZO-1 in vivo. We have identified DCortactin as a protein that interacts with the C-terminal proline-rich domain of DZO-1 by a yeast two-hybrid system. Interaction of these proteins was supported by colocalization at cell-cell junction sites in wing disc epithelial cells and confirmed by coimmunoprecipitation from embryo lysates (Figs. 7 and 8). We showed that mouse cortactin associates with mouse ZO-1 in brain and testis of 1–3-day postnatal mice as DCortactin does with DZO-1 in Drosophila embryo (Fig. 9 b). This conservation suggests that the interaction of these proteins has a functional significance. Recently, the SH3 domain of vertebrate cortactin was found to bind preferentially to peptides sharing the consensus motif +PPΨPXKPXWL (+ and Ψ represent basic and aliphatic residues, respectively) (29Sparks A.B. Rider J.E. Hoffman N.G. Fowlkes D.M. Quillam L.A. Kay B.K. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 1540-1544Crossref PubMed Scopus (331) Google Scholar). A search of the proline-rich domain of DZO-1 revealed that the KPVPPPKPKNY sequence (amino acid residues 1194–1204) is the most likely ligand sequence for the SH3 domain of cortactin. This is consistent with our result showing that proline residues at amino acids 1195 and 1198 (PFKPVPPPKP) are necessary for the interaction with DCortactin in yeast (Fig. 5). We have previously shown that DZO-1 is involved in a signaling pathway that activates transcription of emc and proposed that mammalian ZO-1 is involved in a similar signaling pathway that activates transcription of Id genes (12Takahisa M. Togashi S. Suzuki T. Kobayashi M. Murayama A. Kondo K. Miyake T. Ueda R. Genes Dev. 1996; 10: 1783-1795Crossref PubMed Scopus (61) Google Scholar). Idgenes encode transcriptional regulators homologous to emcand are known to play an important role in the regulation of fate determination, proliferation, and transformation in several cell lineages (30Benezra R. Davis R.L. Lockshon D. Turner D.L. Weintraub H. Cell. 1990; 61: 49-59Abstract Full Text PDF PubMed Scopus (1804) Google Scholar, 31Sun X.H. Copeland N.G. Jenkins N.A. Baltimore D. Mol. Cell. Biol. 1991; 11: 5603-5611Crossref PubMed Scopus (532) Google Scholar, 32Jen Y. Weintraub H. Benezra R. Genes Dev. 1992; 6: 1466-1479Crossref PubMed Scopus (397) Google Scholar, 33Neuhold L.A. Wold B. Cell. 1993; 74: 1033-1042Abstract Full Text PDF PubMed Scopus (130) Google Scholar). Cortactin, a substrate for Src protein-tyrosine kinase, is phosphorylated in response to multiple extracellular stimuli and implicated in cell signaling (17Wu H. Reynolds A.B. Kanner S.B. Vines R.R. Parsons J.T. Mol. Cell. Biol. 1991; 11: 5113-5124Crossref PubMed Scopus (375) Google Scholar, 18Wu H. Parsons J.T. J. Cell Biol. 1993; 120: 1417-1426Crossref PubMed Scopus (452) Google Scholar, 23Zhan X. Hu X. Hampton B. Burgess W.H. Friesel R. Maciag T. J. Biol. Chem. 1993; 268: 24427-24431Abstract Full Text PDF PubMed Google Scholar, 34Maa M.C. Wilson L.K. Moyers J.S. Vines R.R. Parsons J.T. Parsons S.J. Oncogene. 1992; 7: 2429-2438PubMed Google Scholar, 35Wong S. Reynolds A.B. Papkoff J. Oncogene. 1992; 7: 2407-2415PubMed Google Scholar, 36Vuori K. Ruoslahti E. J. Biol. Chem. 1995; 270: 22259-22262Abstract Full Text Full Text PDF PubMed Scopus (269) Google Scholar, 37Dehio C. Prevost M.C. Sansonetti P.J. EMBO J. 1995; 14: 2471-2482Crossref PubMed Scopus (163) Google Scholar, 38Liu M. Qin Y. Liu J. Tanswell A.K. Post M. J. Biol. Chem. 1996; 271: 7066-7071Abstract Full Text Full Text PDF PubMed Scopus (118) Google Scholar). It is known that cortactin is up-regulated during osteoclast differentiation and megakaryocyte maturation (39Hiura K. Lim S.S. Little S.P. Lin S. Sato M. Cell Motil. Cytoskeleton. 1995; 30: 272-284Crossref PubMed Scopus (54) Google Scholar, 40Zhan X. Haudenschild C.C. Ni Y. Smith E. Huang C. Blood. 1997; 89: 457-464Crossref PubMed Google Scholar). DCortactin is not as abundant in adult flies as in embryos, larvae, and pupae, in which developmental stages cells are actively proliferating (Fig. 6). Further, the result of our preliminary experiment indicates that cortactin is abundant in highly proliferating cultured human cells but not in adult human tissues.3 These observations suggest that cortactin is involved in both cell differentiation and proliferation. Human cortactin encoding gene EMS1 is amplified and overexpressed in subsets of breast carcinomas (19Schuuring E. Verhoeven E. Litvinov S. Michalides R.J. Mol. Cell. Biol. 1993; 13: 2891-2898Crossref PubMed Scopus (151) Google Scholar, 41Schuuring E. Gene (Amst.). 1995; 159: 83-96Crossref PubMed Scopus (249) Google Scholar, 42Bringuier P.P. Tamimi Y. Schuuring E. Schalken J. Oncogene. 1996; 12: 1747-1753PubMed Google Scholar, 43Campbell D. deFazio A. Sutherland R. Daly R. Int. J. Cancer. 1996; 68: 485-492Crossref PubMed Scopus (33) Google Scholar). Oncogenic effects of Id-1 overexpression in mammary epithelial cells were demonstrated by the analysis using a cell line (44Desprez P.Y. Hara E. Bissell M.J. Campisi J. Mol. Cell. Biol. 1995; 15: 3398-3404Crossref PubMed Scopus (175) Google Scholar). We hypothesize that cortactin is included in the ZO-1 signaling pathway regulating the expression of emc/Id gene(s). Cortactin is an F-actin binding protein primarily localizing at cortical structures and thought to be important for microfilament-membrane interactions (18Wu H. Parsons J.T. J. Cell Biol. 1993; 120: 1417-1426Crossref PubMed Scopus (452) Google Scholar, 34Maa M.C. Wilson L.K. Moyers J.S. Vines R.R. Parsons J.T. Parsons S.J. Oncogene. 1992; 7: 2429-2438PubMed Google Scholar). Cortactin also has an F-actin cross-linking activity and is implicated in cytoskeleton reorganization (45Huang C. Ni Y. Wang T. Gao Y. Haudenschild C.C. Zhan X. J. Biol. Chem. 1997; 272: 13911-13915Abstract Full Text Full Text PDF PubMed Scopus (220) Google Scholar). On the other hand, ZO-1 is a component of cell-cell junctions (4Stevenson B.R. Siliciano J.D. Mooseker M.S. Goodenough D.A. J. Cell Biol. 1986; 103: 755-766Crossref PubMed Scopus (1285) Google Scholar, 5Itoh M. Nagafuchi A. Yonemura S. Kitani-Yasuda T. Tsukita S. Tsukita S. J. Cell Biol. 1993; 121: 491-502Crossref PubMed Scopus (498) Google Scholar, 7Anderson J.M. Stevenson B.R. Jesaitis L.A. Goodenough D.A. Mooseker M.S. J. Cell Biol. 1988; 106: 1141-1149Crossref PubMed Scopus (285) Google Scholar) and is thought to work in a structural linkage between cell-cell junctions and cytoskeletal networks (6Itoh M. Yonemura S. Nagafuchi A. Tsukita S. Tsukita S. J. Cell Biol. 1991; 115: 1449-1462Crossref PubMed Scopus (205) Google Scholar, 8Furuse M. Itoh M. Hirase T. Nagafuchi A. Yonemura S. Tsukita S. Tsukita S. J. Cell Biol. 1994; 127: 1617-1626Crossref PubMed Scopus (808) Google Scholar,9Itoh M. Nagafuchi A. Moroi S. Tsukita S. J. Cell Biol. 1997; 138: 181-192Crossref PubMed Scopus (570) Google Scholar). As we have shown, subcellular localization of DCortactin overlaps well with F-actin as well as with DZO-1 at cell-cell junction sites in wing disc cells (Fig. 7). Thus, cortactin may be included in the structural linkage in association with ZO-1. The tyrosine phosphorylation of cortactin by Src, which affects the F-actin binding and cross-linking activities, may regulate the cell-cell junction formation (45Huang C. Ni Y. Wang T. Gao Y. Haudenschild C.C. Zhan X. J. Biol. Chem. 1997; 272: 13911-13915Abstract Full Text Full Text PDF PubMed Scopus (220) Google Scholar). We found that the DCortactin gene is situated on the chromosome segment between 93B3 and 93B7.3The isolation of mutant alleles of DCortactin gene is currently under progress using deficiency lines encompassing this region. Molecular genetic studies of DCortactin will provide further information on the conserved roles of cortactin in cell adhesion and signaling required for cell fate determination and cell proliferation. We thank Ms. Y. Sasaki for assistance with the molecular analyses, S. Kamijo for mice. Drs. T. Uemura and M. Ito (Kyoto University, Kyoto, Japan) for the antibodies, and Y. Nishida (Nagoya University, Aichi, Japan) for the cDNA library.
Referência(s)