Susumu Hirabayashi

JAM4, a Junctional Cell Adhesion Molecule Interacting with a Tight Junction Protein, MAGI-1

ABSTRACT MAGI-1 is a membrane-associated guanylate kinase protein at tight junctions in epithelial cells. It interacts with various molecules and functions as a scaffold protein at cell junctions. We report here a novel MAGI-1-binding protein that we named junctional adhesion molecule 4 (JAM4). JAM4 belongs to an immunoglobulin protein family. JAM4 was colocalized with ZO-1 in kidney glomeruli and in intestinal epithelial cells. Biochemical in vitro studies revealed that JAM4 bound to MAGI-1 but not to ZO-1, whereas JAM1 did not bind to MAGI-1. JAM4 and MAGI-1 interacted with each other and formed clusters in COS-7 cells when coexpressed. JAM4 mediated calcium-independent homophilic adhesion and was accumulated at cell-cell contacts when expressed in L cells. MAGI-1, ZO-1, and occludin were recruited to JAM4-based cell contacts. JAM4 also reduced the permeability of CHO cell monolayers. MAGI-1 strengthened JAM4-mediated cell adhesion in L cells and sealing effects in CHO cells. These findings suggest that JAM4 together with MAGI-1 provides an adhesion machinery at tight junctions, which may regulate the permeability of kidney glomerulus and small intestinal epithelial cells.

Receptor for advanced glycation end-products is a marker of type I lung alveolar cells.

Lung alveolar epithelial cells are comprised of type I (ATI) and type II (ATII) cells. AΤΙ cells are polarized, although they have very flat morphology. The identification of marker proteins for apical and basolateral membranes of ATI cells is important to investigate into the differentiation of ATI cells. In this paper, we characterized receptor for advanced glycation end‐products (RAGE) as a marker for ATI cells. RAGE was localized on basolateral membranes of ATI cells in the immunoelectron microscopy and its expression was enhanced in a parallel manner to the differentiation of ATI cells in vivo and in primary cultures of ATII cells. RAGE and T1α, a well‐known ATI marker protein, were targeted to basolateral and apical membranes, respectively, when expressed in polarized Madine Darby canine kidney cells. Moreover, RAGE was expressed in ATI cells after T1αin vivo and in ex in vivo organ cultures. In conclusion, RAGE is a marker for basolateral membranes of well‐differentiated ATI cells. ATI cells require some signal provided by the in vivo environment to express RAGE.

Hippo Pathway–Dependent and –Independent Roles of RASSF6

RASSF6 is both an inhibitor and a promoter of apoptosis, and its proapoptotic activity is regulated by the mammalian kinase MST2, a Hippo homolog. Apoptosis, With or Without Hippo The Hippo signaling pathway, named after the kinase Hippo, limits organ size without affecting patterning, and the components of this pathway are conserved from Drosophila to mammals. Drosophila RASSF (Ras association domain family) not only inhibits Hippo signaling and thus counters Hippo-mediated apoptosis, but also exhibits tumor suppressor function. Of the 10 mammalian RASSF isoforms, the best-characterized isoform, RASSF1A, activates the mammalian homologs of Hippo (the MST kinases), which contrasts with the role of dRASSF. Ikeda et al. show that RASSF6 represents another twist in the RASSF protein family. RASSF6 and MST2 are mutual inhibitors: Activation of MST2 disrupted binding of RASSF6 to MST2, enabling MST2 to induce apoptosis through a process dependent on the Hippo pathway and allowing RASSF6 to mediate apoptosis through a Hippo-independent pathway. These results suggest that mammalian RASSF isoforms may have divergent functions and may link the Hippo pathway to other signal transduction cascades. The Hippo pathway restricts cell growth and proliferation and promotes apoptosis to control organ size. The Drosophila melanogaster isoform of RASSF (Ras association domain family; dRASSF) antagonizes proapoptotic Hippo signaling by inhibiting the binding of the adaptor protein Salvador to the kinase Hippo. Paradoxically, however, dRASSF also functions as a tumor suppressor. In mammals, RASSF1A induces apoptosis by stimulating the mammalian Ste20–like kinases (MSTs) 1 and 2, which are Hippo homologs. Here, we characterize the interaction between MST2 and another mammalian RASSF isoform, RASSF6. When bound to MST2, RASSF6 inhibited MST2 activity to antagonize Hippo signaling. However, RASSF6 caused apoptosis when released from activated MST2 in a manner dependent on WW45, the mammalian Salvador homolog. Thus, RASSF6 antagonizes Hippo signaling and mediates apoptosis through a pathway that is parallel to the canonical Hippo pathway. Our findings suggest that activation of MST2 causes apoptosis through the Hippo pathway, as well as through a RASSF6-mediated pathway.

Synaptic scaffolding molecule is involved in the synaptic clustering of neuroligin.

S-SCAM has a similar molecular organization to PSD-95. Both of them interact with a cell adhesion molecule, neuroligin. We previously reported that beta-catenin binds S-SCAM and recruits it to synapses. We have here examined using rat primary cultured neurons whether neuroligin recruits S-SCAM to synapses or S-SCAM determines the localization of neuroligin. Overexpressed neuroligin formed larger clusters under co-expression of S-SCAM but not of PSD-95. Overexpressed neuroligin blocked synaptic accumulation of PSD-95 but not of S-SCAM. S-SCAM mutant containing the neuroligin-binding region interfered with synaptic accumulation of neuroligin and PSD-95, whereas the similar mutant of PSD-95 had no effect. Biochemical studies revealed that neuroligin forms a ternary complex with S-SCAM and PSD-95 through manifold interactions. These findings imply that S-SCAM is tethered by beta-catenin to synapses and induces synaptic accumulation of neuroligin, which subsequently recruits PSD-95 to synapses.

Synaptic scaffolding molecule (S-SCAM) membrane-associated guanylate kinase with inverted organization (MAGI)-2 is associated with cell adhesion molecules at inhibitory synapses in rat hippocampal neurons

Synaptic scaffolding molecule (S‐SCAM) is a synaptic protein, which harbors five or six PSD‐95/Discs large/ZO‐1 (PDZ), a guanylate kinase and two WW domains. It interacts with NMDA receptor subunits, neuroligin and β‐catenin, and is involved in the accumulation of neuroligin at excitatory synapses. In this study, we have demonstrated S‐SCAM is localized at inhibitory synapses in rat primary cultured hippocampal neurons. We have identified β‐dystroglycan (β‐DG) as a binding partner for S‐SCAM at inhibitory synapses. WW domains of S‐SCAM bind to three sequences of β‐DG. We have also revealed that S‐SCAM can interact with neuroligin 2, which is known to be exclusively localized at inhibitory synapses. The WW domains and the second PDZ domain of S‐SCAM are involved in the interaction with neuroligin 2. β‐DG, neuroligin 2 and S‐SCAM form a tripartite complex in vitro. Neuroligin 2 is detected in the immunoprecipitates by anti‐β‐DG antibody from rat brain. S‐SCAM, β‐DG and neuroligin 2 are partially co‐localized in rat hippocampal neurons. These data suggest that S‐SCAM is associated with β‐DG and neuroligin 2 at inhibitory synapses, and functions as a linker between the dystrophin glycoprotein complex and the neurexin–neuroligin complex.

MAGI-1 is a component of the glomerular slit diaphragm that is tightly associated with nephrin.

MAGUK with inverted domain structure-1 (MAGI-1) is a membrane-associated protein with one guanylate kinase, six PSD-95/Dlg-A/ZO-1 (PDZ), and two WW domains and is localized at tight junctions in epithelial cells. MAGI-1 interacts with various proteins and is proposed to function as a scaffold protein. In the previous study, we discovered a MAGI-1-interacting cell adhesion molecule junctional adhesion molecule 4 (JAM4). Both proteins are highly expressed in glomerular podocytes in the kidney and partially colocalized. In this study, we have further searched for a binding partner of MAGI-1 in the kidney through yeast two-hybrid screening and obtained nephrin. Nephrin is a cell adhesion molecule specifically localized at the slit diaphragm between neighboring foot processes of podocytes. Biochemical studies reveal that nephrin directly binds to the middle PDZ domains of MAGI-1 through its carboxyl terminus but does not bind to ZO-1. MAGI-1 forms a tripartite complex with nephrin and JAM4 in vitro. Immunoelectron microscopy shows that the localization of MAGI-1 is restricted to the slit diaphragm, whereas JAM4 is also distributed on apical membranes of podocytes. In puromycin aminonucleoside-induced nephrotic podocytes, MAGI-1 is localized with nephrin at the displaced slit diaphragm. These data indicate that MAGI-1 is a component of the slit diaphragm and tightly interacts with nephrin and JAM4 in vivo. MAGI-1 may play a role in determining the boundary between the apical and the bosolateral domain at the level of slit diaphragm.

Threonine 74 of MOB1 is a putative key phosphorylation site by MST2 to form the scaffold to activate nuclear Dbf2-related kinase 1.

Mammalian nuclear Dbf2-related (NDR) kinases (LATS1 and 2, NDR1 and 2) play a role in cell proliferation, apoptosis and morphological changes. These kinases are regulated by mammalian sterile 20-like kinases (MSTs) and Mps one binder (MOB) 1. Okadaic acid (OA), which activates MST2, facilitates the complex formation of MOB1, MST2 and NDR1 in HEK293FT cells. The in vitro biochemical study demonstrates the phosphorylation of MOB1 by MST2. The phosphorylated MOB1 alone is capable to partially activate NDR1 in vitro, but MST2 is also required for the full activation. The knockdown of MOB1 or MST2 abolishes the OA-induced NDR1 activation in HEK293FT cells. Among MOB1 mutants, in which each serine or threonine residue is replaced with alanine, MOB1 T74A and T181A mutants fail to activate NDR1. Thr74, but not Thr181, is phosphorylated by MST2 in vitro, although MOB1 is also phosphorylated by MST2 at other site(s). The interaction of MOB1 T74A with NDR1 is barely enhanced by OA treatment. These findings indicate that the phosphorylation of MOB1 at Thr74 by MST2 is essential to make a complex of MOB1, MST2 and NDR1, and to fully activate NDR1.

Localization of p0071-interacting proteins, plakophilin-related armadillo-repeat protein-interacting protein (PAPIN) and ERBIN, in epithelial cells.

PAPIN has six PDZ domains and interacts with p0071, a catenin-related protein. Recent studies have revealed that catenins determine the subcellular localization of some PDZ proteins. We have examined whether the localization of PAPIN is determined by p0071 in epithelial cells. PAPIN was localized not only on the lateral membrane but also on the apical membrane, where p0071 was absent. The targeting to both membranes was mediated by the middle region of PAPIN and did not require the p0071-interacting PDZ domain. In cells that came into contact, PAPIN was diffusely distributed on the plasma membrane, while p0071 was concentrated at immature cell–cell contacts. When epithelial cells were exposed to the low concentration of calcium, p0071 was internalized, whereas PAPIN remained on the plasma membrane. We also confirmed that the interaction with p0071 was not essential for the membrane targeting of ERBIN, a recently identified p0071- and ErbB2-binding protein. PAPIN, p0071, and ERBIN formed a complex in 293T cells. Furthermore, ERBIN and ErbB2 were colocalized with PAPIN on the lateral membrane. These findings suggest that PAPIN, p0071, and ERBIN come to the cell–cell contacts independently and interact with each other on the lateral membrane.

Ligand-of-Numb protein X is an endocytic scaffold for junctional adhesion molecule 4

Junctional adhesion molecule 4 (JAM4) is a cell adhesion molecule that interacts with a tight junction protein, membrane-associated guanylate kinase inverted 1 (MAGI-1). Our previous studies suggest that JAM4 is implicated in the regulation of paracellular permeability and the signalings of hepatocyte growth factor. In this study, we performed yeast two-hybrid screening to search for an unidentified JAM4-binding protein and obtained one isoform of Ligand-of-Numb protein X1 (LNX1), LNXp70, that is an interactor of Numb. Ligand-of-Numb protein X1 is expressed in kidney glomeruli and intestinal epithelial cells, where JAM4 is also detected. Immunoprecipitation from kidney lysates supports the in vivo interaction of proteins. Biochemical studies reveal that JAM4 directly binds the second PDZ domain of LNX1 through its carboxyl terminus. Junctional adhesion molecule 4, LNX1 and Numb form a tripartite complex in vitro and are partially colocalized in heterologous cells. Ligand-of-Numb protein X1 facilitates endocytosis of JAM4 and is involved in transforming growth factor β -induced redistribution of JAM4 in mammary epithelial cells. Experiments using dominant-negative constructs and RNA interference insure that Numb is necessary for the LNX1-mediated endocytosis of JAM4. All these findings indicate that LNX1 provides an endocytic scaffold for JAM4 that is implicated in the reorganization of cell junctions.

A CTX Family Cell Adhesion Molecule, JAM4, Is Expressed in Stem Cell and Progenitor Cell Populations of both Male Germ Cell and Hematopoietic Cell Lineages

ABSTRACT Stem cells are maintained in an undifferentiated state by interacting with a microenvironment known as the “niche,” which is comprised of various secreted and membrane proteins. Our goal was to identify niche molecules participating in stem cell-stem cell and/or stem cell-supporting cell interactions. Here, we isolated genes encoding secreted and membrane proteins from purified male germ stem cells using a signal sequence trap approach. Among the genes identified, we focused on the junctional adhesion molecule 4 (JAM4), an immunoglobulin type cell adhesion molecule. JAM4 protein was actually localized to the plasma membrane in male germ cells. JAM4 expression was downregulated as cells differentiated in both germ cell and hematopoietic cell lineages. To analyze function in vivo, we generated JAM4-deficient mice. Histological analysis of testes from homozygous nulls did not show obvious abnormalities, nor did liver and kidney tissues, both of which strongly express JAM4. The numbers of hematopoietic stem cells in bone marrow were indistinguishable between wild-type and mutant mice, as was male germ cell development. These results suggest that JAM4 is expressed in stem cells and progenitor cells but that other cell adhesion molecules may substitute for JAM4 function in JAM4-deficient mice both in male germ cell and hematopoietic lineages.