- Email: [email protected]
Transmembrane proteins of tight junctions
seminars in CELL & DEVELOPMENTAL BIOLOGY, Vol. 11, 2000: pp. 281–289 doi: 10.1006/scdb.2000.0177, available online at http://www.idealibrary.com on
Transmembrane proteins of tight junctions Maria S. Baldaa and Karl Matterb
junctions directly contribute to the maintenance of cell surface polarity by forming an intramembrane diffusion barrier or fence that restricts the intermixing of apical and basolateral plasma membrane components. The formation of functional cellular barriers further requires the sealing of the intercellular space between neighboring cells. Tight junctions are responsible for the sealing of the paracellular pathway by forming a paracellular diffusion barrier or gate.2 The paracellular diffusion barrier is not an absolute barrier but is semipermeable because it allows the selective passage of certain solutes and ions. In freeze-fracture replicas, tight junctions appear as a fascinating network of fibrils that encircles the apical end of the lateral membrane.3 The composition of these fibrils was a mystery until the recent discovery of transmembrane proteins that localize to these strands and appear to be intimately involved in their formation. These strand-associated membrane proteins directly interact with submembrane proteins that function as adaptors and signaling proteins, as well as with the actin-based cytoskeleton (see reviews by Gonzalez-Mariscal, and Tuner and Mercer in this issue). This review will focus on the tight-junctionassociated membrane proteins and discuss the experimental evidence supporting their involvement in the barrier functions of tight junctions.
Tight junctions from a morphological and functional boundary between the apical and basolateral cell surface domains of epithelia and endothelia, and regulate selective diffusion along the paracellular space. Two types of fourspan transmembrane proteins, occludin and claudins, as well as the single-span protein JAM are associated with tight junctions. The functional analysis of these proteins starts to reveal how they are involved in the functions of tight junctions, which of their domains are important for these functions, and how they interact with each other to form the junctional diffusion barriers. Key words: tight junctions / zonula occludens / occludin / claudin / JAM c 2000 Academic Press
Introduction A hallmark in the development of multicellular organisms is the assembly of cellular sheets that separate compartments of different composition. This function is performed by epithelia and endothelia. To form such cellular barriers, epithelial and endothelial cells polarize and adhere to each other by forming different types of cell–cell junctions.1 Tight junctions (or zonula occludens) are the most apical junctions in epithelia and represent a morphological boundary between the two epithelial cell surface domains. These different cell surface domains are fundamental for the physiological functions of epithelial and endothelial tissues such as, for instance, selective transcellular transport of solutes and ions. Tight
Tight-junction-associated transmembrane proteins Occludin was the first transmembrane protein of tight junctions that was identified.4 It is widely expressed by essentially all epithelial and endothelial tissues and has also been reported to be expressed by neurons and astrocytes.4–6 Occludin is a 60–65 kDa protein that was predicted to span the membrane four times (see Figure 1 for a schematic drawing). Today, we have many experimental results to support this membrane topology. The C-terminal cytosolic domain of occludin interacts in vitro with several
From the Département de Biologie Cellulaire, Université de Genève Sciences III, 30, Quai Ernest-Ansermet, 1211 Genève-4, Switzerland. a E-mail: [email protected] b E-mail: [email protected] . c
2000 Academic Press 1084–9521 / 00 / 000281+ 09 / $35.00/0 / 0
281
M. S. Balda and K. Matter
proteins that associate with the cytosolic face of the junctional membrane (i.e. ZO-1, ZO-2, ZO-3 and cingulin) as well as with actin filaments.7–12 The topology of both extracellular loops was analyzed with probes specific for these extracellular domains (i.e. antibodies or peptides representing their sequence) that were shown to have access to occludin or could interfere with occludin localization and/or function in intact cells.13–15 This topology was further supported by the efficient glycosylation of N-linked glycosylation sites introduced into the sequences of the extracellular loops.16 Moreover, antibodies specific for the N-terminal cytosolic domain only have access to occludin in permeabilized and not in nonpermeabilized cells (Matter and Balda, unpublished). Occludin not only interacts with proteins of the subjunctional plaque but also with the membrane protein VAP-33, a protein of unknown function that was previously described to interact with proteins involved in membrane fusion during vesicular transport (i.e. SNAREs).17 Claudins were more recently discovered and are a family of tight junction proteins of 22–24 kDa. Claudin-1 and -2 were originally identified because they cofractionated with occludin from sonicated junctional fractions.18 There is no evidence, however, for a direct interaction between occludin and claudins, and claudins do not copurify with occludin from detergent-solubilized membranes. Although claudins do not exhibit significant sequence similarity with occludin, hydrophobicity plots suggest that they also span the membrane four times and expose both termini to the cytosol. At least 18 members of the claudin family have thus far been identified.18–27 Several claudins were cloned and sequenced prior to the description of claudin-1 and -2 but were not known to be associated with tight junctions. Examples are claudin-11/oligodendrocyte-specific protein (OSP), a protein associated with Sertoli cell and central nervous system myelin tight junctions,26–28 and claudin-5/transmembrane protein deleted in velo-cardio-facial syndrome (TMVCF), which is associated with endothelial tight junctions.29–31 In contrast to occludin, individual claudins are generally expressed in only a restricted number of specific cell types, suggesting that they are associated with tissue-specific functions of tight junctions. Besides their membrane topology, different claudin family membranes share other structural features. This is suggested by the observation that eight different claudins can bind to the three submembrane proteins ZO-1, -2, and -3 by an interaction involving
Figure 1. Occludin and the intramembrane diffusion barrier of tight junctions. Wild-type and transfected MDCK cells expressing mutant occludin were cultured on permeable tissue culture inserts. The structures of the expressed mutants are illustrated in the schematic drawings on the left, and the junctional distributions of transfected and endogenous occludin in a given cell line are indicated. The integrity of the intramembrane diffusion barriers of these monolayers was then tested by inserting fluorescently labeled sphingomyelin into the apical plasma membrane domain and diffusion was monitored by confocal microscopy (shown are XZ-sections; for a complete description of the assay, see Balda et al., 199637 ). Note, both termini of occludin have to be modified to induce a discontinuous distribution of transfected and endogenous occludin and to disrupt the intramembrane diffusion barrier.
the claudin’s C-terminal cytosolic domain and a PDZ domain of the submembrane plaque proteins.32 A third type of tight junction transmembrane protein is the junction-associated membrane protein JAM. It is a single-span membrane component belonging to the immunoglobulin superfamily.33 JAM was identified in mice with a monoclonal antibody generated against endothelial cells but is also expressed by epithelial cells of many different origins. A human orthologue has recently been described that, in contrast to the mouse protein, is also expressed in leukocytes.34
Tight junction structure The ultrastructural appearance of tight junctions has intrigued biologists for three decades. Tight 282
Transmembrane proteins of tight junctions
close-contact sites.4, 18 Occludin was first shown to be able to mediate cell–cell aggregation; however, its adhesive properties appear to be weak.13 At least some of the claudins are also able to mediate cell adhesion and they do so in a Ca+ -independent manner.40 Furthermore, expression of JAM in fibroblasts also induces cell–cell adhesion.33 Thus, cell–cell adhesion in tight junctions does not appear to rely on a single type of intercellular interaction but is mediated by different types of adhesive interactions.
junctions appear as a network of anastomosing intramembrane strands in electron micrographs of freeze-fracture replicas.3, 25 For many years, these intramembrane strands were attributed to either rows of transmembrane proteins or, alternatively, nonconventional lipid structures, such as inverted micelles. The latter model had its origin in the appearance of tight junctions in micrographs of thin sections, where tight junctions often look like apparent hemifusions between the two neighboring cell membranes. It is clear that the transmembrane proteins occludin and claudins are components of these strands. Several experimental results indicate, however, that their importance for the formation of the intramembrane strands is very different. Although immunoelectron microscopy demonstrated that occludin is directly associated with the intramembrane strands,35, 36 it is not critical for the formation of the strands. This was first suggested by the disruption of the continuous distribution of occludin along the cell periphery of MDCK cells by expression of a dominant negative mutant of occludin deficient in its interaction with the submembrane cytoskeleton: redistribution of occludin was not paralleled by changes in the appearance of the intramembrane strands.37 Furthermore, occludin-deficient embryonic stem cells form morphologically normal tight junctions.38 This does not mean that occludin may not contribute to the substance of the strands since overexpression of occludin produced intramembrane particles and very short fragments of filaments in membranes with high local concentrations of occludin.36 Claudins are also directly associated with the intramembrane strands18 and are clearly more important for their formation. Expression of several claudins in fibroblasts has been shown to result in the formation of adhesion planes containing an elaborate network of strands,39 and different claudins and occludin are able to coassemble into such induced tight junction strands (for a detailed review, see Reference 25). Moreover, at least for some tight junctions, the presence of a claudin family member appears to be necessary for the formation of intramembrane strands: OSP/claudin-11 null mice fail to assemble intramembrane strands in Sertoli cell and central nervous system myelin tight junctions.28 In electron micrographs of thin sections, tight junctions contain focal adhesion sites of very close contacts between the neighboring cells, indicating the presence of molecules mediating cell–cell adhesion. Occludin and claudins localize to these
Tight junction biogenesis Tight junctions are located between the apical and basolateral cell surface domains; hence, the question arises whether tight junction assembly under steady-state conditions occurs from the apical or the basolateral domain. This question has thus far only been studied using occludin as a tool. Chimeras consisting of the C-terminal domain of occludin and a normally apically sorted glycoprotein (i.e. the ectoand transmembrane domain of a mouse Fc receptor for IgG), were shown to be targeted directly to the basolateral membrane, indicating that the C-terminal domain of occludin contains a basolateral targeting determinant.16 This type of sorting signal is responsible for the basolateral expression of a wide variety of different basolateral membrane proteins.41 This suggested that transport of occludin to tight junctions involves transient passage through the basolateral membrane. This model is further supported by the observation of lateral occludin in certain tissues as well as in overexpressing cells.42, 43 The next question was how occludin becomes concentrated in tight junctions once it is inserted into the lateral membrane. Because the C-terminal domain of occludin interacts with submembrane proteins of tight junctions, it was proposed that this interaction is responsible for the junctional accumulation of occludin.7 Experiments with chimeric proteins containing the C-terminal cytosolic domain gave conflicting results. The above-mentioned chimera was not integrated into tight junctions to a significant extent;43 but a more recently described chimera in which the C-terminal cytosolic domain of occludin was fused to the N-terminal half of connexin-32 accumulated in tight junctions.44 Although there are many obvious structural differences between connexin-32 and an Fc receptor, it could be that these data mean that only proteins capable of mediating cell–cell interactions can be 283
M. S. Balda and K. Matter
that could be solubilized and immunoprecipitated,16, 53 suggesting that complex formation of mutant with endogenous occludin was responsible for co-clustering as well as for junctional transport in the absence of a C-terminal cytosolic domain. The integrity of the intramembrane diffusion barrier in cultured epithelial cells can be tested with a simple experiment in which a fluorescent lipid is inserted into the apical cell surface domain; the distribution of the lipid can then be followed by confocal microscopy.37 Figure 1 shows that expression of occludin or occludin mutants that do not affect the distribution of transfected and endogenous occludin (HAoccludin and occludinCT3) does not affect the intramembrane diffusion barrier. If a mutant is expressed in which both terminal cytosolic domains have been inactivated and which consequently disrupts the distribution of transfected and endogenous occludin (HAoccludinCT3), the intramembrane diffusion barrier is corrupted and the lipid can diffuse along the lateral membrane. It is not clear which submembrane proteins mediate the continuous distribution of occludin and, thereby, are responsible for the integrity of the intramembrane diffusion barrier. Possible candidates are ZO-1, ZO-2, ZO-3, and cingulin: all proteins shown to interact with the C-terminal domain of occludin, which can also bind directly to f-actin.7–12 It is thought that the N-terminal epitope interferes with the junctional distribution of occludin by blocking binding to a not yet identified submembrane plaque protein. Although no proteins have been identified that bind to the N-terminal domain of occludin, a structural and/or functional role of the N-terminal domain is also suggested by the recent discovery of an occludin isoform (occludin-1B) that possesses a different N-terminal amino acid sequence.54 The intramembrane strands of tight junctions were often proposed to represent the intramembrane diffusion barrier. Interestingly, induction of a discontinuous occludin distribution disrupted the intramembrane diffusion barrier but not the intramembrane strands,37 suggesting that the presence of such strands is not sufficient for tight junctions to fulfill their function as a fence. Occludin localizes to the strands, however, suggesting that intramembrane strands need to have the proper composition (e.g. containing occludin in the case of MDCK cells) to function as intramembrane diffusion barriers. Tight junctions regulate selective diffusion of ions and solutes throught the paracellular pathway.55 To do this, tight junctions need to establish a barrier that
concentrated in tight junctions by the C-terminal cytosolic domain of occludin. Nevertheless, it could also be that connexin-32 has an intrinsic affinity for tight junctions because it has recently been shown to coimmunoprecipitate with occludin from transfected hepatocytes.45 The specific accumulation of occludin in tight junctions appears to require other domains: expression of occludin with mutations in the extracellular loops was shown to result in reduced integration into tight junctions and lateral accumulation.16, 43 Although the functions of the extracellular loops in this process are not known, it could be that they mediate concentration in the junctions by interacting with molecules in the neighboring cells. Such a mechanism is supported by the depletion of occludin from tight junctions of epithelial cells cultured with a peptide representing the amino acid sequence of the second extracellular loop.14 These considerations suggest a model in which occludin is first transported to the basolateral membrane because of a basolateral sorting signal in its C-terminal cytosolic domain and then becomes concentrated into tight junctions because of interactions mediated by its extracellular loops. On arrival in tight junctions, interactions with the submembrane cytoskeleton (e.g. ZO-1) mediated by both terminal cytosolic domains are responsible for anchoring the protein in the junction and for a continuous junctional distribution of occludin.43 Furthermore, anchoring of occludin to the submembrane cytoskeleton may be regulated by phosphorylation.42, 46–50 It is not clear, however, whether phosphorylation indeed regulates the localization and/or is rather of functional relevance (e.g. regulation of paracellular permeability).
Tight junction functions Tight junctions form an intramembrane diffusion barrier that restricts the intermixing of apical and basolateral membrane components.51, 52 Thus far, there is only evidence for occludin to be involved in the formation of this intramembrane diffusion barrier or fence. Expression of mutant occludin with an N-terminal epitope and a deleted C-terminal cytosolic domain in MDCK cells resulted in a discontinuous distribution of transfected and endogenous occludin.37 Exogenous and endogenous occludin not only co-clustered but, in MDCK cells as well as Xenopus embryos, also formed oligomeric complexes 284
Transmembrane proteins of tight junctions
paracellular permeability. Paracellin-1 localizes to tight junctions of the thick ascending limb of Henle and is required for paracellular Mg2+ reabsorption.22 As in the case of occludin, it is not known whether paracellin-1 functions as a regulator (e.g. a Mg2+ sensor) or actually forms a Mg2+ -specific diffusion pathway. Nevertheless, recent experiments with dominant-negative occludin mutants in MDCK cells that inhibit selective paracellular diffusion suggest that occludin and claudins cooperate to allow selective paracellular diffusion.43 Given this experimental evidence and the existence of so many different claudin family members, it appears obvious to propose a tight junction model in which the basic framework of the intramembrane strands is formed by polymers of, depending on the tissue, one or different types of claudins (Figure 2). As an additive, these polymers contain occludin which changes their functional properties but not their physical appearance. These strands represent the paracellular barriers and their tissue-dependent composition determines the transepithelial electrical resistance. Selective paracellular diffusion across these barriers occurs by the opening and closing of channels that can be formed by one or different types of claudins as well as occludin. The composition of the channels determines their specificities and, therefore, varies from one tissue to another. The opening and closing of these channels is regulated or gated so that in the strands forming a compartment of the network, the channels are never open in more than one strand at once to avoid a breakdown of the barrier.
prevents unspecific diffusion of solutes and to assemble a pathway that allows selective diffusion of solutes along concentration gradients. The strand-associated transmembrane proteins occludin and claudins are involved in both of these functions. Occludin has been experimentally implicated in the formation of the diffusion barrier in a variety of different cell types.14, 37, 53, 56–58 Nevertheless, occludin is neither required for the formation of the paracellular barrier38 not does it affect the integrity of the barrier when the continuous distribution of occludin is disrupted.37 Hence, occludin may enhance the strength of the paracellular diffusion barrier and expression of mutant occludin may interfere with barrier formation, but occludin is not an essential component of the barrier. Claudins also appear to be involved in the formation of the paracellular barrier. In MDCK cells, overexpression of claudin-1 results in increased transepithelial electrical resistance59 and incubation with a fragment of Clostridium perfringens enterotoxin in the selective removal of claudin-4 and downregulation of the paracellular barrier.60 Although the first claudin null mice deficient in OSP/claudin-11 have been generated and the mutant mice exhibit strong phenotypes and lack intramembrane strands in Sertoli cells and central nervous system myelin, how the properties of these paracellular barriers are affected in these animals has not been determined.28 Occludin and at least one of the claudins, paracellin-1, are also involved in the pathway that allows selective paracellular diffusion. In the case of occludin, the evidence comes from expression experiments in MDCK cells. Overexpression of wildtype occludin resulted in upregulation of selective paracellular permeability.37, 56 Moreover, expression of mutant occludin lacking the C-terminal cytosolic domain enhanced this phenotype (independent of the distribution of occludin), suggesting that interactions mediated by this domain are important for the regulation of selective paracellular permeability.37 Given the importance of the actin cytoskeleton in the regulation of paracellular permeability,61–64 this may involve the direct interaction of the Cterminal domain of occludin with actin filaments or interactions with submembrane proteins of tight junctions.7–9, 11, 12 It is not clear whether occludin serves as a regulator of selective paracellular permeability and/or is directly involved in the formation of a pathway that allows selective diffusion. Paracellin-1 (claudin-16) is the only claudin family member that has been directly implicated in selective
Tight junctions and the transmigration of leukocytes Tight junctions also have to open and close in a highly coordinated and reversible manner to allow leukocytes to reach inflammatory sites. In response to chemoattractants, leukocytes emigrate from the blood by crossing the endothelium and, in the case of mucosal infections, cross the inflamed epithelium.65, 66 Transmigration occurs primarily along the paracellular route. Two of the tight-junction-associated transmembrane proteins have been implicated in this process. A monoclonal antibody against JAM inhibits the transmigration of leukocytes across endothelial cells in tissue culture assays as well as in an in vivo model of skin inflammatory reaction.33 Moreover, the 285
M. S. Balda and K. Matter
monoclonal anti-JAM antibody also interferes with the accumulation of leukocytes in the cerebrospinal fluid in cytokine-induced meningitis.67 How the anti-JAM monoclonal antibody inhibits transmigration of leukocytes is unknown, but may be due to an inhibition of a heterophilic interaction with a leukocyte receptor.67 Alternatively, the anti-JAM antibody may stabilize a homophilic JAM-mediated interaction between neighboring endothelial cells and thus inhibit the dissociation of the junctional complex. Occludin has been implicated in the transmigration of leukocytes across epithelial cells. Using wild-type and transfected MDCK cells, occludin was shown to modulate the efficiency of transepithelial migration of neutrophils.68 Occludin’s main function in transmigration is unrelated to its role in paracellular permeability and is only affected by a mutation in the N-terminal cytosolic domain. Because the N-terminal cytosolic domain is also sufficient to mediate a continuous junctional distribution (independent of the C-terminal cytosolic domain), it is likely that an interaction between the N-terminal domain of occludin and the submembrane cytoskeleton regulates the reversible opening of tight junctions during transmigration of leukocytes.
Figure 2. Schematic model for the formation of a semipermeable diffusion barrier. Tight junctions are thought to consist of a series of diffusion barriers that are interconnected.71, 72 These barriers are formed by polymers containing one or several claudins (shown as black and gray squares) and occludin (shown as white squares). Although the diameter of the particles in intramembrane strands in freeze-fracture replicas suggests that they are oligomers themselves, this has not been considered in this scheme for the sake of simplicity. Pores within these strands are opened in a regulated fashion and allow selective diffusion along concentration gradients. The selectivity of diffusion is determined by the composition of the pores that may be formed by one or different types of molecules (i.e. claudins and/or occludin). This model implies that the claudin and occludin composition of a tight junction determine the specificity of the paracellular pathway; hence, the formation of such homo- and heteromeric channels would provide a molecular explanation for differences in selectivity of paracellular diffusion from one tissue to another. Importantly, the opening and closing of the pores is synchronized in such a way that within a segment only the pores in one strand are open at a given moment in time to avoid a breakdown of the barrier function. Therefore, solutes could cross such a junction in the same manner as ships overcome barriers in rivers using locks. When the pores of the first strand opens (Phase 1), a solute with the right physicochemical properties (small dots) can diffuse into the first chamber and remain there until the pores in the first strand close and those in the next strand open (Phase 2). Once the pores in the second strand are closed, the pores in the third strand open, and the tracer can exit the junctional lock system (Phase 3). Importantly, a solute that does not match the selectivity of the pores (large dots) does not enter the junction and is hence prevented from diffusing along the paracellular pathway.
Concluding remarks and future directions Several tight-junction-associated membrane proteins are now known, and several intermolecular interactions of these proteins with the submembrane plaque have been described. Although we are starting to understand for which functions of tight junctions these proteins are important, the molecular mechanisms involved in these functions and the functional relevance of essentially all the described intermolecular interactions are unknown. Hence, a major challenge will be to determine not only the interactions within the junctional protein network but to relate these interactions to tight junction functions. This task is likely to be a complicated one because many components and interactions may prove to be not essential but redundant, or only important under particular conditions or in specific tissues. Nevertheless, the molecular and functional analysis of the tight junction membrane proteins may finally help us to understand at a molecular level how tight junctions function as an intramembrane fence and form a semipermeable paracellular diffusion barrier. Furthermore, a recent study reported that occludin 286
Transmembrane proteins of tight junctions
participates in the regulation of signaling processes controlling epithelial transformation.69 This and the recent identification of a tight-junction-associated transcription factor that controls the expression of a proto-oncogene70 suggest that a complete functional analysis of tight junction membrane protein will require the inclusion of assays designed to determine effects on cell growth and differentiation.
13. Van Itallie CM, Anderson JM (1997) Occludin confers adhesiveness when expressed in fibroblasts. J Cell Sci 110:1113– 1121 14. Wong V, Gumbiner BM (1997) A synthetic peptide corresponding to the extracellular domain of occludin perturbs the tight junction permeability barrier. J Cell Biol 136:399–409 15. Lacaz-Vieira F, Jaeger MM, Farshori P, Kachar B (1999) Small synthetic peptides homologous to segments of the first external loop of occludin impair tight junction resealing. J Membr Biol 168:289–297 16. Matter K, Balda MS (1998) Biogenesis of tight junctions: The C-terminal domain of occludin mediates basolateral targeting. J Cell Sci 111:511–519 17. Lapierre LA, Tuma PL, Navarre J, Goldenring JR, Anderson JM (1999) VAP-33 localizes to both an intracellular vesicle population and with occludin at the tight junction. J Cell Sci 112:3723–3732 18. Furuse M, Fujita K, Hiiragi T, Fujimoto K, Tsukita S (1998a) Claudin-1 and -2: novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin. J Cell Biol 141:1539–1550 19. Paperna T, Peoples R, Wang YK, Kaplan P, Francke U (1998) Genes for the CPE receptor (CPETR1) and the human homolog of RVP1 (CPETR2) are localized within the Williams–Beuren syndrome deletion. Genomics 54:453– 459 20. Chen Z, Zandonatti M, Jakubowski D, Fox HS (1998) Brain capillary endothelial cells express MBEC1, a protein that is related to the Clostridium perfringens enterotoxin receptors. Lab Invest 78:353–363 21. Morita K, Furuse M, Fujimoto K, Tsukita S (1999) Claudin multigene family encoding four-transmembrane domain protein components of tight junction strands. Proc Natl Acad Sci USA 96:511–516 22. Simon DB et al. (1999) Paracellin-1, a renal tight junction protein required for paracellular Mg2+ resorption. Science 285:103–106 23. Katahira J, Sugiyama H, Inoue N, Horiguchi Y, Matsuda M, Sugimoto N (1997) Clostridium perfringens enterotoxin utilizes two structurally related membrane proteins as functional receptors in vivo. J Biol Chem 272:26652–26658 24. Swisshelm K, Machl A, Planitzer S, Robertson R, Kubbies M, Hosier S (1999) SEMP1, a senescence-associated cDNA isolated from human mammary epithelial cells, is a member of an epithelial membrane protein superfamily. Gene 226:285– 295 25. Tsukita S, Furuse M (1999) Occludin and claudins in tightjunction strands: leading or supporting players? Trends Cell Biol 9:268–273 26. Bronstein JM, Popper P, Micevych PE, Farber DB (1996) Isolation and characterization of a novel oligodendrocytespecific protein. Neurology 47:772–778 27. Morita K, Sasaki H, Fujimoto K, Furuse M, Tsukita S (1999) Claudin-11/OSP-based tight junctions of myelin sheaths in brain and Sertoli cells in testis. J Cell Biol 145:579–588 28. Gow A, Southwood CM, Li JS, Pariali M, Riordan GP, Brodie SE, Danias J, Bronstein JM, Kachar B, Lazzarini RA (1999) CNS myelin and sertoli cell tight junction strands are absent in Osp/claudin-11 null mice. Cell 99:649–659 29. Carlson C et al. (1997) Molecular definition of 22q11 deletions in 151 velo-cardio-facial syndrome patients. Am J Hum Genet 61:620–629 30. Morita K, Sasaki H, Furuse M, Tsukita S (1999) Endothelial claudin. Claudin-5/tmvcf constitutes tight junction strands in endothelial cells. J Cell Biol 147:185–194
Acknowledgements The authors are supported by the Swiss National Science Foundation and the Canton de Genève.
References 1. Farquhar MG, Palade GE (1963) Junctional complexes in various epithelia. J Cell Biol 17:375–412 2. Cereijido M (1991) Tight Junctions, CRC Press Inc., Boca Raton, FL 3. Staehelin LA (1973) Further observations of the fine structure of freeze-cleaved tight junctions. J Cell Sci 13:763–786 4. Furuse M, Hirase T, Itoh M, Nagafuchi A, Yonemura S, Tsukita S, Tsukita S (1993) Occludin: A novel integral membrane protein localizing at tight junctions. J Cell Biol 123:1777–1788 5. Hirase T, Staddon JM, Saitou M, Ando-Akatsuku Y, Itoh M, Furse W, Fujimoto K, Tsukita S, Rubin LL (1997) Occludin as a possible determinant of tight junction permeability in endothelial cells. J Cell Sci 110:1603–1613 6. Bauer H, Stelzhammer W, Fuchs R, Weiger TM, Danninger C, Probst G, Krizbai IA (1999) Astrocytes and neurons express the tight junction-specific protein occludin in vitro. Exp Cell Res 250:434–438 7. Furuse M, Itoh M, Hirase T, Nagafuchi A, Yonemura S, Tsukita S, Tsukita S (1994) Direct association of occludin with ZO-1 and its possible involvement in the localization of occludin at tight junctions. J Cell Biol 127:1617–1626 8. Fanning AS, Anderson JM (1998) The tight junction protein ZO-1 establishes a link between the membrane protein occludin and the actin cytoskeleton. J Biol Chem 273:29745– 29753 9. Haskins J, Gu L, Wittichen ES, Hibbard J, Stevenson BR (1998) ZO-3, a novel member of the MAGUK protein family found at the tight junction, interacts with ZO-1 and occludin. J Cell Biol 141:199–208 10. Cordenonsi M, Turco F, D’Atri F, Hammar E, Martinucci G, Meggio F, Citi S (1999) Xenopus laevis occludin. Identification of in vitro phosphorylation sites by protein kinase CK2 and association with cingulin. Eur J Biochem 264:374–384 11. Itoh M, Morita K, Tsukita S (1999) Characterization of ZO-2 as a MAGUK family member associated with tight and adherens junctions with a binding affinity to occludin and α catenin. J Biol Chem 274:5981–5986 12. Wittchen ES, Haskins J, Stevenson BR (1999) Protein interactions at the tight junction: actin has multiple binding partners, and ZO-1 forms independent complexes with ZO-2 and ZO-3. J Biol Chem 274:35179–35185
287
M. S. Balda and K. Matter
Physiol 273:C1856–C1867 47. Tsukamoto T, Nigam SK (1999) Role of tyrosine phosphorylation in the reassembly of occludin and other tight junction proteins. Am J Physiol 276:F737–750 48. Cordenonsi M, Mazzon E, De Rigo L, Baraldo S, Meggio F, Citi S (1997) Occludin dephosphorylation in early development of Xenopus laevis. J Cell Sci 110:3131–3139 49. Antonetti DA, Barber AJ, Hollinger LA, Wolpert EB, Gardner TW (1999) Vascular endothelial growth factor induces rapid phosphorylation of tight junction proteins occludin and zonula occluden 1. A potential mechanism for vascular permeability in diabetic retinopathy and tumors. J Biol Chem 274:23463–23467 50. Farshori P, Kachar B (1999) Redistribution and phosphorylation of occludin during opening and resealing of tight junctions in cultured epithelial cells. J Membr Biol 170:147–156 51. Dragsten PR, Blumenthal R, Handler JS (1981) Membrane asymmetry in epithelia: Is the tight junction a barrier to diffusion in the plasma membrane? Nature 294:718–722 52. van Meer G, Simons K (1986) The function of tight junctions in maintaining differences in lipid composition between the apical and the basolateral cell surface domains of MDCK cells. EMBO J 5:1455–1464 53. Chen Y, Merzdorf C, Paul DL, Goodenough DA (1997) COOH terminus of occludin is required for tight junction barrier function in early Xenopus embryos. J Cell Biol 138:891– 899 54. Muresan Z, Paul DL, Goodenough DA (2000) Occludin 1B, a variant of the tight junction protein occludin. Mol Biol Cell 11:627–634 55. Madara JL (1998) Paracellular permeability. Annu Rev Physiol 60:143–159 56. McCarthy KM, Skare IB, Stankewich MC, Furuse M, Tsukita S, Rogers RA, Lynch RD, Schneeberger EE (1996) Occludin is a functional component of the tight junction. J Cell Sci 109:2287–2298 57. Bamforth SD, Kniesel U, Wolburg H, Engelhardt B, Risau W (1999) A dominant mutant of occludin disrupts tight junction structure and function. J Cell Sci 112:1879–1888 58. Wan H et al. (1999) Der p 1 facilitates transepithelial allergen delivery by disruption of tight junctions [see comments]. J Clin Invest 104:123–133 59. Inai T, Kobayashi J, Shibata Y (1999) Claudin-1 contributes to the epithelial barrier function in MDCK cells. Eur J Cell Biol 78:849–855 60. Sonoda N, Furuse M, Sasaki H, Yonemura S, Katahira J, Horiguchi Y, Tsukita S (1999) Clostridium perfringens enterotoxin fragment removes specific claudins from tight junction strands: Evidence for direct involvement of claudins in tight junction barrier. J Cell Biol 147:195–204 61. Hecht G, Pestic L, Nicevic G, Koutouris A, Tripuraneni J, Lorimer DD, Grzegorz N, Guerriero V, Elson EL, De Lanerolle P (1996) Expression of the catalytic domain of myosin light chain kinase increases paracellular permeability. Am J Physiol 271:C1678–C1684 62. Turner JR, Rill BK, Carlson SL, Carnes D, Kerner R, Mrsny RJ, Madara JL (1997) Physiological regulation of epithelial tight junctions is associated with myosin light-chain phosphorylation. Am J Physiol 273:C1378–C1385 63. Citi S, Cordenonsi M (1998) Tight junction proteins. Biochim Biophys Acta 1448:1–11 64. Fanning AS, Mitic LL, Anderson JM (1999) Transmembrane proteins in the tight junction barrier. J Am Soc Nephrol 10:1337–1345 65. Dunon D, Piali L, Imhof BA (1996) To stick or not to stick:
31. Sirotkin H, Morrow B, Saint-Jore B, Puech A, Das Gupta R, Patanjali SR, Skoultchi A, Weissman SM, Kucherlapati R (1997) Identification, characterization, and precise mapping of a human gene encoding a novel membrane-spanning protein from the 22q11 region deleted in velo-cardio-facial syndrome. Genomics 42:245–251 32. Itoh M, Furuse M, Morita K, Kubota K, Saitou M, Tsukita S (1999) Direct binding of three tight junctionassociated MAGUKs, ZO-1, ZO-2, and ZO-3, with the COOH termini of claudins. J Cell Biol 147:1351–1363 33. Martin-Padura I et al. (1998) Junctional adhesion molecule, a novel member of the immunoglobulin superfamily that distributes at intercellular junctions and modulates monocyte transmigration. J Cell Biol 142:117–127 34. Williams LA, Martin-Padura I, Dejana E, Hogg N, Simmons DL (1999) Identification and characterisation of human Junctional Adhesion Molecule (JAM). Mol Immunol 36:1175– 1188 35. Fujimoto K (1995) Freeze-fracture replica electron microscopy combined with SDS digestion for cytochemical labeling of integral membrane proteins: application to the immunogold labeling of intercellular junctional complex. J Cell Sci 108:3443–3449 36. Furuse M, Fujimoto K, Sato N, Hirase T, Tsukita S, Tsukita S (1996) Overexpression of occludin, a tight junction integral membrane protein, induces the formation of intracellular multilamellar bodies bearing tight junction-like structures. J Cell Sci 109:429–435 37. Balda MS, Whitney JA, Flores C, González S, Cereijido M, Matter K (1996a) Functional dissociation of paracellular permeability and transepithelial electrical resistance and disruption of the apical-basolateral intramembrane diffusion barrier by expression of a mutant tight junction membrane protein. J Cell Biol 134:1031–1049 38. Saitou M, Fujimoto K, Doi Y, Itoh M, Fujimoto T, Furuse M, Takano H, Noda T, Tsukita S (1998) Occludin-deficient embryonic stem cells can differentiate into polarized epithelial cells bearing tight junctions. J Cell Biol 141:391–401 39. Furuse M, Sasaki H, Fujimoto K, Tsukita S (1998b) A single gene product, claudin-1 or -2, reconstitutes tight junction strands and recruits occludin in fibroblasts. J Cell Biol 143:397–408 40. Kubota K, Furuse M, Sasaki H, Sonoda N, Fujita K, Nagafuchi A, Tsukita S (1999) Ca(2+)-independent cell-adhesion activity of claudins, a family of integral membrane proteins localized at tight junctions. Curr Biol 9:1035–1038 41. Matter K (2000) Epithelial polarity: Sorting out the sorters. Curr Biol 10:R39–R42 42. Sakakibara A, Furuse M, Saitou M, Ando Akatsuka Y, Tsukita S (1997) Possible involvement of phosphorylation of occludin in tight junction formation. J Cell Biol 137:1393–1401 43. Balda MS, Flores-Maldonado C, Cereijido M, Matter K (2000a) Multiple domains of occludin participate in regulation of paracellular permeability. J Cell Biochem (in press) 44. Mitic LL, Schneeberger EE, Fanning AS, Anderson JM (1999) Connexin-occludin chimeras containing the ZO-binding domain of occludin localize at MDCK tight junctions and NRK cell contacts. J Cell Biol 146:683–693 45. Kojima T, Sawada N, Chiba H, Kokai Y, Yamamoto M, Urban M, Lee GH, Hertzberg EL, Mochizuki Y, Spray DC (1999) Induction of tight junctions in human connexin 32 (hC × 32)-transfered mouse hepatocytes: connexin 32 interacts with occludin. Biochem Biophys Res Commun 266:222–229 46. Wong V (1997) Phosphorylation of occludin correlates with occludin localization and function at the tight junction. Am J
288
Transmembrane proteins of tight junctions
66. 67.
68.
69.
The new leukocyte homing paradigm. Curr Opin Cell Biol 8:714–723 Huber D, Balda MS, Matter K (1999) Transepithelial migration of neutrophils. Invasion Metastasis 18:70–80 Del Maschio A et al. (1999) Leukocyte recruitment in the cerebrospinal fluid of mice with experimental meningitis is inhibited by an antibody to junctional adhesion molecule (JAM). J Exp Med 190:1351–1356 Huber D, Balda MS, Matter K (2000) Occludin modulates transepithelial migration of neutrophils. J Biol Chem 275:5773–5778 Li D, Mrsny K (2000) Oncogenic Raf-1 disrupts epithelial tight
junctions via downregulation of occludin. J Cell Biol 148:791– 800 70. Balda MS, Matter K (2000b) The tight junction protein ZO-1 and an associated transcription factor regulate ErbB-2 expression. EMBO J (in press) 71. Claude P (1978) Morphological factors influencing transepithelial permeability: A model for the resistance of the zonula occludens. J Membr Biol 39:219–232 72. Cereijido M, González-Mariscal L, Contreras G (1989) Tight junction: Barrier between higher organisms and environment. News Physiol Sci 4:72–74
289
Transmembrane proteins of tight junctions Maria S. Baldaa and Karl Matterb
junctions directly contribute to the maintenance of cell surface polarity by forming an intramembrane diffusion barrier or fence that restricts the intermixing of apical and basolateral plasma membrane components. The formation of functional cellular barriers further requires the sealing of the intercellular space between neighboring cells. Tight junctions are responsible for the sealing of the paracellular pathway by forming a paracellular diffusion barrier or gate.2 The paracellular diffusion barrier is not an absolute barrier but is semipermeable because it allows the selective passage of certain solutes and ions. In freeze-fracture replicas, tight junctions appear as a fascinating network of fibrils that encircles the apical end of the lateral membrane.3 The composition of these fibrils was a mystery until the recent discovery of transmembrane proteins that localize to these strands and appear to be intimately involved in their formation. These strand-associated membrane proteins directly interact with submembrane proteins that function as adaptors and signaling proteins, as well as with the actin-based cytoskeleton (see reviews by Gonzalez-Mariscal, and Tuner and Mercer in this issue). This review will focus on the tight-junctionassociated membrane proteins and discuss the experimental evidence supporting their involvement in the barrier functions of tight junctions.
Tight junctions from a morphological and functional boundary between the apical and basolateral cell surface domains of epithelia and endothelia, and regulate selective diffusion along the paracellular space. Two types of fourspan transmembrane proteins, occludin and claudins, as well as the single-span protein JAM are associated with tight junctions. The functional analysis of these proteins starts to reveal how they are involved in the functions of tight junctions, which of their domains are important for these functions, and how they interact with each other to form the junctional diffusion barriers. Key words: tight junctions / zonula occludens / occludin / claudin / JAM c 2000 Academic Press
Introduction A hallmark in the development of multicellular organisms is the assembly of cellular sheets that separate compartments of different composition. This function is performed by epithelia and endothelia. To form such cellular barriers, epithelial and endothelial cells polarize and adhere to each other by forming different types of cell–cell junctions.1 Tight junctions (or zonula occludens) are the most apical junctions in epithelia and represent a morphological boundary between the two epithelial cell surface domains. These different cell surface domains are fundamental for the physiological functions of epithelial and endothelial tissues such as, for instance, selective transcellular transport of solutes and ions. Tight
Tight-junction-associated transmembrane proteins Occludin was the first transmembrane protein of tight junctions that was identified.4 It is widely expressed by essentially all epithelial and endothelial tissues and has also been reported to be expressed by neurons and astrocytes.4–6 Occludin is a 60–65 kDa protein that was predicted to span the membrane four times (see Figure 1 for a schematic drawing). Today, we have many experimental results to support this membrane topology. The C-terminal cytosolic domain of occludin interacts in vitro with several
From the Département de Biologie Cellulaire, Université de Genève Sciences III, 30, Quai Ernest-Ansermet, 1211 Genève-4, Switzerland. a E-mail: [email protected] b E-mail: [email protected] . c
2000 Academic Press 1084–9521 / 00 / 000281+ 09 / $35.00/0 / 0
281
M. S. Balda and K. Matter
proteins that associate with the cytosolic face of the junctional membrane (i.e. ZO-1, ZO-2, ZO-3 and cingulin) as well as with actin filaments.7–12 The topology of both extracellular loops was analyzed with probes specific for these extracellular domains (i.e. antibodies or peptides representing their sequence) that were shown to have access to occludin or could interfere with occludin localization and/or function in intact cells.13–15 This topology was further supported by the efficient glycosylation of N-linked glycosylation sites introduced into the sequences of the extracellular loops.16 Moreover, antibodies specific for the N-terminal cytosolic domain only have access to occludin in permeabilized and not in nonpermeabilized cells (Matter and Balda, unpublished). Occludin not only interacts with proteins of the subjunctional plaque but also with the membrane protein VAP-33, a protein of unknown function that was previously described to interact with proteins involved in membrane fusion during vesicular transport (i.e. SNAREs).17 Claudins were more recently discovered and are a family of tight junction proteins of 22–24 kDa. Claudin-1 and -2 were originally identified because they cofractionated with occludin from sonicated junctional fractions.18 There is no evidence, however, for a direct interaction between occludin and claudins, and claudins do not copurify with occludin from detergent-solubilized membranes. Although claudins do not exhibit significant sequence similarity with occludin, hydrophobicity plots suggest that they also span the membrane four times and expose both termini to the cytosol. At least 18 members of the claudin family have thus far been identified.18–27 Several claudins were cloned and sequenced prior to the description of claudin-1 and -2 but were not known to be associated with tight junctions. Examples are claudin-11/oligodendrocyte-specific protein (OSP), a protein associated with Sertoli cell and central nervous system myelin tight junctions,26–28 and claudin-5/transmembrane protein deleted in velo-cardio-facial syndrome (TMVCF), which is associated with endothelial tight junctions.29–31 In contrast to occludin, individual claudins are generally expressed in only a restricted number of specific cell types, suggesting that they are associated with tissue-specific functions of tight junctions. Besides their membrane topology, different claudin family membranes share other structural features. This is suggested by the observation that eight different claudins can bind to the three submembrane proteins ZO-1, -2, and -3 by an interaction involving
Figure 1. Occludin and the intramembrane diffusion barrier of tight junctions. Wild-type and transfected MDCK cells expressing mutant occludin were cultured on permeable tissue culture inserts. The structures of the expressed mutants are illustrated in the schematic drawings on the left, and the junctional distributions of transfected and endogenous occludin in a given cell line are indicated. The integrity of the intramembrane diffusion barriers of these monolayers was then tested by inserting fluorescently labeled sphingomyelin into the apical plasma membrane domain and diffusion was monitored by confocal microscopy (shown are XZ-sections; for a complete description of the assay, see Balda et al., 199637 ). Note, both termini of occludin have to be modified to induce a discontinuous distribution of transfected and endogenous occludin and to disrupt the intramembrane diffusion barrier.
the claudin’s C-terminal cytosolic domain and a PDZ domain of the submembrane plaque proteins.32 A third type of tight junction transmembrane protein is the junction-associated membrane protein JAM. It is a single-span membrane component belonging to the immunoglobulin superfamily.33 JAM was identified in mice with a monoclonal antibody generated against endothelial cells but is also expressed by epithelial cells of many different origins. A human orthologue has recently been described that, in contrast to the mouse protein, is also expressed in leukocytes.34
Tight junction structure The ultrastructural appearance of tight junctions has intrigued biologists for three decades. Tight 282
Transmembrane proteins of tight junctions
close-contact sites.4, 18 Occludin was first shown to be able to mediate cell–cell aggregation; however, its adhesive properties appear to be weak.13 At least some of the claudins are also able to mediate cell adhesion and they do so in a Ca+ -independent manner.40 Furthermore, expression of JAM in fibroblasts also induces cell–cell adhesion.33 Thus, cell–cell adhesion in tight junctions does not appear to rely on a single type of intercellular interaction but is mediated by different types of adhesive interactions.
junctions appear as a network of anastomosing intramembrane strands in electron micrographs of freeze-fracture replicas.3, 25 For many years, these intramembrane strands were attributed to either rows of transmembrane proteins or, alternatively, nonconventional lipid structures, such as inverted micelles. The latter model had its origin in the appearance of tight junctions in micrographs of thin sections, where tight junctions often look like apparent hemifusions between the two neighboring cell membranes. It is clear that the transmembrane proteins occludin and claudins are components of these strands. Several experimental results indicate, however, that their importance for the formation of the intramembrane strands is very different. Although immunoelectron microscopy demonstrated that occludin is directly associated with the intramembrane strands,35, 36 it is not critical for the formation of the strands. This was first suggested by the disruption of the continuous distribution of occludin along the cell periphery of MDCK cells by expression of a dominant negative mutant of occludin deficient in its interaction with the submembrane cytoskeleton: redistribution of occludin was not paralleled by changes in the appearance of the intramembrane strands.37 Furthermore, occludin-deficient embryonic stem cells form morphologically normal tight junctions.38 This does not mean that occludin may not contribute to the substance of the strands since overexpression of occludin produced intramembrane particles and very short fragments of filaments in membranes with high local concentrations of occludin.36 Claudins are also directly associated with the intramembrane strands18 and are clearly more important for their formation. Expression of several claudins in fibroblasts has been shown to result in the formation of adhesion planes containing an elaborate network of strands,39 and different claudins and occludin are able to coassemble into such induced tight junction strands (for a detailed review, see Reference 25). Moreover, at least for some tight junctions, the presence of a claudin family member appears to be necessary for the formation of intramembrane strands: OSP/claudin-11 null mice fail to assemble intramembrane strands in Sertoli cell and central nervous system myelin tight junctions.28 In electron micrographs of thin sections, tight junctions contain focal adhesion sites of very close contacts between the neighboring cells, indicating the presence of molecules mediating cell–cell adhesion. Occludin and claudins localize to these
Tight junction biogenesis Tight junctions are located between the apical and basolateral cell surface domains; hence, the question arises whether tight junction assembly under steady-state conditions occurs from the apical or the basolateral domain. This question has thus far only been studied using occludin as a tool. Chimeras consisting of the C-terminal domain of occludin and a normally apically sorted glycoprotein (i.e. the ectoand transmembrane domain of a mouse Fc receptor for IgG), were shown to be targeted directly to the basolateral membrane, indicating that the C-terminal domain of occludin contains a basolateral targeting determinant.16 This type of sorting signal is responsible for the basolateral expression of a wide variety of different basolateral membrane proteins.41 This suggested that transport of occludin to tight junctions involves transient passage through the basolateral membrane. This model is further supported by the observation of lateral occludin in certain tissues as well as in overexpressing cells.42, 43 The next question was how occludin becomes concentrated in tight junctions once it is inserted into the lateral membrane. Because the C-terminal domain of occludin interacts with submembrane proteins of tight junctions, it was proposed that this interaction is responsible for the junctional accumulation of occludin.7 Experiments with chimeric proteins containing the C-terminal cytosolic domain gave conflicting results. The above-mentioned chimera was not integrated into tight junctions to a significant extent;43 but a more recently described chimera in which the C-terminal cytosolic domain of occludin was fused to the N-terminal half of connexin-32 accumulated in tight junctions.44 Although there are many obvious structural differences between connexin-32 and an Fc receptor, it could be that these data mean that only proteins capable of mediating cell–cell interactions can be 283
M. S. Balda and K. Matter
that could be solubilized and immunoprecipitated,16, 53 suggesting that complex formation of mutant with endogenous occludin was responsible for co-clustering as well as for junctional transport in the absence of a C-terminal cytosolic domain. The integrity of the intramembrane diffusion barrier in cultured epithelial cells can be tested with a simple experiment in which a fluorescent lipid is inserted into the apical cell surface domain; the distribution of the lipid can then be followed by confocal microscopy.37 Figure 1 shows that expression of occludin or occludin mutants that do not affect the distribution of transfected and endogenous occludin (HAoccludin and occludinCT3) does not affect the intramembrane diffusion barrier. If a mutant is expressed in which both terminal cytosolic domains have been inactivated and which consequently disrupts the distribution of transfected and endogenous occludin (HAoccludinCT3), the intramembrane diffusion barrier is corrupted and the lipid can diffuse along the lateral membrane. It is not clear which submembrane proteins mediate the continuous distribution of occludin and, thereby, are responsible for the integrity of the intramembrane diffusion barrier. Possible candidates are ZO-1, ZO-2, ZO-3, and cingulin: all proteins shown to interact with the C-terminal domain of occludin, which can also bind directly to f-actin.7–12 It is thought that the N-terminal epitope interferes with the junctional distribution of occludin by blocking binding to a not yet identified submembrane plaque protein. Although no proteins have been identified that bind to the N-terminal domain of occludin, a structural and/or functional role of the N-terminal domain is also suggested by the recent discovery of an occludin isoform (occludin-1B) that possesses a different N-terminal amino acid sequence.54 The intramembrane strands of tight junctions were often proposed to represent the intramembrane diffusion barrier. Interestingly, induction of a discontinuous occludin distribution disrupted the intramembrane diffusion barrier but not the intramembrane strands,37 suggesting that the presence of such strands is not sufficient for tight junctions to fulfill their function as a fence. Occludin localizes to the strands, however, suggesting that intramembrane strands need to have the proper composition (e.g. containing occludin in the case of MDCK cells) to function as intramembrane diffusion barriers. Tight junctions regulate selective diffusion of ions and solutes throught the paracellular pathway.55 To do this, tight junctions need to establish a barrier that
concentrated in tight junctions by the C-terminal cytosolic domain of occludin. Nevertheless, it could also be that connexin-32 has an intrinsic affinity for tight junctions because it has recently been shown to coimmunoprecipitate with occludin from transfected hepatocytes.45 The specific accumulation of occludin in tight junctions appears to require other domains: expression of occludin with mutations in the extracellular loops was shown to result in reduced integration into tight junctions and lateral accumulation.16, 43 Although the functions of the extracellular loops in this process are not known, it could be that they mediate concentration in the junctions by interacting with molecules in the neighboring cells. Such a mechanism is supported by the depletion of occludin from tight junctions of epithelial cells cultured with a peptide representing the amino acid sequence of the second extracellular loop.14 These considerations suggest a model in which occludin is first transported to the basolateral membrane because of a basolateral sorting signal in its C-terminal cytosolic domain and then becomes concentrated into tight junctions because of interactions mediated by its extracellular loops. On arrival in tight junctions, interactions with the submembrane cytoskeleton (e.g. ZO-1) mediated by both terminal cytosolic domains are responsible for anchoring the protein in the junction and for a continuous junctional distribution of occludin.43 Furthermore, anchoring of occludin to the submembrane cytoskeleton may be regulated by phosphorylation.42, 46–50 It is not clear, however, whether phosphorylation indeed regulates the localization and/or is rather of functional relevance (e.g. regulation of paracellular permeability).
Tight junction functions Tight junctions form an intramembrane diffusion barrier that restricts the intermixing of apical and basolateral membrane components.51, 52 Thus far, there is only evidence for occludin to be involved in the formation of this intramembrane diffusion barrier or fence. Expression of mutant occludin with an N-terminal epitope and a deleted C-terminal cytosolic domain in MDCK cells resulted in a discontinuous distribution of transfected and endogenous occludin.37 Exogenous and endogenous occludin not only co-clustered but, in MDCK cells as well as Xenopus embryos, also formed oligomeric complexes 284
Transmembrane proteins of tight junctions
paracellular permeability. Paracellin-1 localizes to tight junctions of the thick ascending limb of Henle and is required for paracellular Mg2+ reabsorption.22 As in the case of occludin, it is not known whether paracellin-1 functions as a regulator (e.g. a Mg2+ sensor) or actually forms a Mg2+ -specific diffusion pathway. Nevertheless, recent experiments with dominant-negative occludin mutants in MDCK cells that inhibit selective paracellular diffusion suggest that occludin and claudins cooperate to allow selective paracellular diffusion.43 Given this experimental evidence and the existence of so many different claudin family members, it appears obvious to propose a tight junction model in which the basic framework of the intramembrane strands is formed by polymers of, depending on the tissue, one or different types of claudins (Figure 2). As an additive, these polymers contain occludin which changes their functional properties but not their physical appearance. These strands represent the paracellular barriers and their tissue-dependent composition determines the transepithelial electrical resistance. Selective paracellular diffusion across these barriers occurs by the opening and closing of channels that can be formed by one or different types of claudins as well as occludin. The composition of the channels determines their specificities and, therefore, varies from one tissue to another. The opening and closing of these channels is regulated or gated so that in the strands forming a compartment of the network, the channels are never open in more than one strand at once to avoid a breakdown of the barrier.
prevents unspecific diffusion of solutes and to assemble a pathway that allows selective diffusion of solutes along concentration gradients. The strand-associated transmembrane proteins occludin and claudins are involved in both of these functions. Occludin has been experimentally implicated in the formation of the diffusion barrier in a variety of different cell types.14, 37, 53, 56–58 Nevertheless, occludin is neither required for the formation of the paracellular barrier38 not does it affect the integrity of the barrier when the continuous distribution of occludin is disrupted.37 Hence, occludin may enhance the strength of the paracellular diffusion barrier and expression of mutant occludin may interfere with barrier formation, but occludin is not an essential component of the barrier. Claudins also appear to be involved in the formation of the paracellular barrier. In MDCK cells, overexpression of claudin-1 results in increased transepithelial electrical resistance59 and incubation with a fragment of Clostridium perfringens enterotoxin in the selective removal of claudin-4 and downregulation of the paracellular barrier.60 Although the first claudin null mice deficient in OSP/claudin-11 have been generated and the mutant mice exhibit strong phenotypes and lack intramembrane strands in Sertoli cells and central nervous system myelin, how the properties of these paracellular barriers are affected in these animals has not been determined.28 Occludin and at least one of the claudins, paracellin-1, are also involved in the pathway that allows selective paracellular diffusion. In the case of occludin, the evidence comes from expression experiments in MDCK cells. Overexpression of wildtype occludin resulted in upregulation of selective paracellular permeability.37, 56 Moreover, expression of mutant occludin lacking the C-terminal cytosolic domain enhanced this phenotype (independent of the distribution of occludin), suggesting that interactions mediated by this domain are important for the regulation of selective paracellular permeability.37 Given the importance of the actin cytoskeleton in the regulation of paracellular permeability,61–64 this may involve the direct interaction of the Cterminal domain of occludin with actin filaments or interactions with submembrane proteins of tight junctions.7–9, 11, 12 It is not clear whether occludin serves as a regulator of selective paracellular permeability and/or is directly involved in the formation of a pathway that allows selective diffusion. Paracellin-1 (claudin-16) is the only claudin family member that has been directly implicated in selective
Tight junctions and the transmigration of leukocytes Tight junctions also have to open and close in a highly coordinated and reversible manner to allow leukocytes to reach inflammatory sites. In response to chemoattractants, leukocytes emigrate from the blood by crossing the endothelium and, in the case of mucosal infections, cross the inflamed epithelium.65, 66 Transmigration occurs primarily along the paracellular route. Two of the tight-junction-associated transmembrane proteins have been implicated in this process. A monoclonal antibody against JAM inhibits the transmigration of leukocytes across endothelial cells in tissue culture assays as well as in an in vivo model of skin inflammatory reaction.33 Moreover, the 285
M. S. Balda and K. Matter
monoclonal anti-JAM antibody also interferes with the accumulation of leukocytes in the cerebrospinal fluid in cytokine-induced meningitis.67 How the anti-JAM monoclonal antibody inhibits transmigration of leukocytes is unknown, but may be due to an inhibition of a heterophilic interaction with a leukocyte receptor.67 Alternatively, the anti-JAM antibody may stabilize a homophilic JAM-mediated interaction between neighboring endothelial cells and thus inhibit the dissociation of the junctional complex. Occludin has been implicated in the transmigration of leukocytes across epithelial cells. Using wild-type and transfected MDCK cells, occludin was shown to modulate the efficiency of transepithelial migration of neutrophils.68 Occludin’s main function in transmigration is unrelated to its role in paracellular permeability and is only affected by a mutation in the N-terminal cytosolic domain. Because the N-terminal cytosolic domain is also sufficient to mediate a continuous junctional distribution (independent of the C-terminal cytosolic domain), it is likely that an interaction between the N-terminal domain of occludin and the submembrane cytoskeleton regulates the reversible opening of tight junctions during transmigration of leukocytes.
Figure 2. Schematic model for the formation of a semipermeable diffusion barrier. Tight junctions are thought to consist of a series of diffusion barriers that are interconnected.71, 72 These barriers are formed by polymers containing one or several claudins (shown as black and gray squares) and occludin (shown as white squares). Although the diameter of the particles in intramembrane strands in freeze-fracture replicas suggests that they are oligomers themselves, this has not been considered in this scheme for the sake of simplicity. Pores within these strands are opened in a regulated fashion and allow selective diffusion along concentration gradients. The selectivity of diffusion is determined by the composition of the pores that may be formed by one or different types of molecules (i.e. claudins and/or occludin). This model implies that the claudin and occludin composition of a tight junction determine the specificity of the paracellular pathway; hence, the formation of such homo- and heteromeric channels would provide a molecular explanation for differences in selectivity of paracellular diffusion from one tissue to another. Importantly, the opening and closing of the pores is synchronized in such a way that within a segment only the pores in one strand are open at a given moment in time to avoid a breakdown of the barrier function. Therefore, solutes could cross such a junction in the same manner as ships overcome barriers in rivers using locks. When the pores of the first strand opens (Phase 1), a solute with the right physicochemical properties (small dots) can diffuse into the first chamber and remain there until the pores in the first strand close and those in the next strand open (Phase 2). Once the pores in the second strand are closed, the pores in the third strand open, and the tracer can exit the junctional lock system (Phase 3). Importantly, a solute that does not match the selectivity of the pores (large dots) does not enter the junction and is hence prevented from diffusing along the paracellular pathway.
Concluding remarks and future directions Several tight-junction-associated membrane proteins are now known, and several intermolecular interactions of these proteins with the submembrane plaque have been described. Although we are starting to understand for which functions of tight junctions these proteins are important, the molecular mechanisms involved in these functions and the functional relevance of essentially all the described intermolecular interactions are unknown. Hence, a major challenge will be to determine not only the interactions within the junctional protein network but to relate these interactions to tight junction functions. This task is likely to be a complicated one because many components and interactions may prove to be not essential but redundant, or only important under particular conditions or in specific tissues. Nevertheless, the molecular and functional analysis of the tight junction membrane proteins may finally help us to understand at a molecular level how tight junctions function as an intramembrane fence and form a semipermeable paracellular diffusion barrier. Furthermore, a recent study reported that occludin 286
Transmembrane proteins of tight junctions
participates in the regulation of signaling processes controlling epithelial transformation.69 This and the recent identification of a tight-junction-associated transcription factor that controls the expression of a proto-oncogene70 suggest that a complete functional analysis of tight junction membrane protein will require the inclusion of assays designed to determine effects on cell growth and differentiation.
13. Van Itallie CM, Anderson JM (1997) Occludin confers adhesiveness when expressed in fibroblasts. J Cell Sci 110:1113– 1121 14. Wong V, Gumbiner BM (1997) A synthetic peptide corresponding to the extracellular domain of occludin perturbs the tight junction permeability barrier. J Cell Biol 136:399–409 15. Lacaz-Vieira F, Jaeger MM, Farshori P, Kachar B (1999) Small synthetic peptides homologous to segments of the first external loop of occludin impair tight junction resealing. J Membr Biol 168:289–297 16. Matter K, Balda MS (1998) Biogenesis of tight junctions: The C-terminal domain of occludin mediates basolateral targeting. J Cell Sci 111:511–519 17. Lapierre LA, Tuma PL, Navarre J, Goldenring JR, Anderson JM (1999) VAP-33 localizes to both an intracellular vesicle population and with occludin at the tight junction. J Cell Sci 112:3723–3732 18. Furuse M, Fujita K, Hiiragi T, Fujimoto K, Tsukita S (1998a) Claudin-1 and -2: novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin. J Cell Biol 141:1539–1550 19. Paperna T, Peoples R, Wang YK, Kaplan P, Francke U (1998) Genes for the CPE receptor (CPETR1) and the human homolog of RVP1 (CPETR2) are localized within the Williams–Beuren syndrome deletion. Genomics 54:453– 459 20. Chen Z, Zandonatti M, Jakubowski D, Fox HS (1998) Brain capillary endothelial cells express MBEC1, a protein that is related to the Clostridium perfringens enterotoxin receptors. Lab Invest 78:353–363 21. Morita K, Furuse M, Fujimoto K, Tsukita S (1999) Claudin multigene family encoding four-transmembrane domain protein components of tight junction strands. Proc Natl Acad Sci USA 96:511–516 22. Simon DB et al. (1999) Paracellin-1, a renal tight junction protein required for paracellular Mg2+ resorption. Science 285:103–106 23. Katahira J, Sugiyama H, Inoue N, Horiguchi Y, Matsuda M, Sugimoto N (1997) Clostridium perfringens enterotoxin utilizes two structurally related membrane proteins as functional receptors in vivo. J Biol Chem 272:26652–26658 24. Swisshelm K, Machl A, Planitzer S, Robertson R, Kubbies M, Hosier S (1999) SEMP1, a senescence-associated cDNA isolated from human mammary epithelial cells, is a member of an epithelial membrane protein superfamily. Gene 226:285– 295 25. Tsukita S, Furuse M (1999) Occludin and claudins in tightjunction strands: leading or supporting players? Trends Cell Biol 9:268–273 26. Bronstein JM, Popper P, Micevych PE, Farber DB (1996) Isolation and characterization of a novel oligodendrocytespecific protein. Neurology 47:772–778 27. Morita K, Sasaki H, Fujimoto K, Furuse M, Tsukita S (1999) Claudin-11/OSP-based tight junctions of myelin sheaths in brain and Sertoli cells in testis. J Cell Biol 145:579–588 28. Gow A, Southwood CM, Li JS, Pariali M, Riordan GP, Brodie SE, Danias J, Bronstein JM, Kachar B, Lazzarini RA (1999) CNS myelin and sertoli cell tight junction strands are absent in Osp/claudin-11 null mice. Cell 99:649–659 29. Carlson C et al. (1997) Molecular definition of 22q11 deletions in 151 velo-cardio-facial syndrome patients. Am J Hum Genet 61:620–629 30. Morita K, Sasaki H, Furuse M, Tsukita S (1999) Endothelial claudin. Claudin-5/tmvcf constitutes tight junction strands in endothelial cells. J Cell Biol 147:185–194
Acknowledgements The authors are supported by the Swiss National Science Foundation and the Canton de Genève.
References 1. Farquhar MG, Palade GE (1963) Junctional complexes in various epithelia. J Cell Biol 17:375–412 2. Cereijido M (1991) Tight Junctions, CRC Press Inc., Boca Raton, FL 3. Staehelin LA (1973) Further observations of the fine structure of freeze-cleaved tight junctions. J Cell Sci 13:763–786 4. Furuse M, Hirase T, Itoh M, Nagafuchi A, Yonemura S, Tsukita S, Tsukita S (1993) Occludin: A novel integral membrane protein localizing at tight junctions. J Cell Biol 123:1777–1788 5. Hirase T, Staddon JM, Saitou M, Ando-Akatsuku Y, Itoh M, Furse W, Fujimoto K, Tsukita S, Rubin LL (1997) Occludin as a possible determinant of tight junction permeability in endothelial cells. J Cell Sci 110:1603–1613 6. Bauer H, Stelzhammer W, Fuchs R, Weiger TM, Danninger C, Probst G, Krizbai IA (1999) Astrocytes and neurons express the tight junction-specific protein occludin in vitro. Exp Cell Res 250:434–438 7. Furuse M, Itoh M, Hirase T, Nagafuchi A, Yonemura S, Tsukita S, Tsukita S (1994) Direct association of occludin with ZO-1 and its possible involvement in the localization of occludin at tight junctions. J Cell Biol 127:1617–1626 8. Fanning AS, Anderson JM (1998) The tight junction protein ZO-1 establishes a link between the membrane protein occludin and the actin cytoskeleton. J Biol Chem 273:29745– 29753 9. Haskins J, Gu L, Wittichen ES, Hibbard J, Stevenson BR (1998) ZO-3, a novel member of the MAGUK protein family found at the tight junction, interacts with ZO-1 and occludin. J Cell Biol 141:199–208 10. Cordenonsi M, Turco F, D’Atri F, Hammar E, Martinucci G, Meggio F, Citi S (1999) Xenopus laevis occludin. Identification of in vitro phosphorylation sites by protein kinase CK2 and association with cingulin. Eur J Biochem 264:374–384 11. Itoh M, Morita K, Tsukita S (1999) Characterization of ZO-2 as a MAGUK family member associated with tight and adherens junctions with a binding affinity to occludin and α catenin. J Biol Chem 274:5981–5986 12. Wittchen ES, Haskins J, Stevenson BR (1999) Protein interactions at the tight junction: actin has multiple binding partners, and ZO-1 forms independent complexes with ZO-2 and ZO-3. J Biol Chem 274:35179–35185
287
M. S. Balda and K. Matter
Physiol 273:C1856–C1867 47. Tsukamoto T, Nigam SK (1999) Role of tyrosine phosphorylation in the reassembly of occludin and other tight junction proteins. Am J Physiol 276:F737–750 48. Cordenonsi M, Mazzon E, De Rigo L, Baraldo S, Meggio F, Citi S (1997) Occludin dephosphorylation in early development of Xenopus laevis. J Cell Sci 110:3131–3139 49. Antonetti DA, Barber AJ, Hollinger LA, Wolpert EB, Gardner TW (1999) Vascular endothelial growth factor induces rapid phosphorylation of tight junction proteins occludin and zonula occluden 1. A potential mechanism for vascular permeability in diabetic retinopathy and tumors. J Biol Chem 274:23463–23467 50. Farshori P, Kachar B (1999) Redistribution and phosphorylation of occludin during opening and resealing of tight junctions in cultured epithelial cells. J Membr Biol 170:147–156 51. Dragsten PR, Blumenthal R, Handler JS (1981) Membrane asymmetry in epithelia: Is the tight junction a barrier to diffusion in the plasma membrane? Nature 294:718–722 52. van Meer G, Simons K (1986) The function of tight junctions in maintaining differences in lipid composition between the apical and the basolateral cell surface domains of MDCK cells. EMBO J 5:1455–1464 53. Chen Y, Merzdorf C, Paul DL, Goodenough DA (1997) COOH terminus of occludin is required for tight junction barrier function in early Xenopus embryos. J Cell Biol 138:891– 899 54. Muresan Z, Paul DL, Goodenough DA (2000) Occludin 1B, a variant of the tight junction protein occludin. Mol Biol Cell 11:627–634 55. Madara JL (1998) Paracellular permeability. Annu Rev Physiol 60:143–159 56. McCarthy KM, Skare IB, Stankewich MC, Furuse M, Tsukita S, Rogers RA, Lynch RD, Schneeberger EE (1996) Occludin is a functional component of the tight junction. J Cell Sci 109:2287–2298 57. Bamforth SD, Kniesel U, Wolburg H, Engelhardt B, Risau W (1999) A dominant mutant of occludin disrupts tight junction structure and function. J Cell Sci 112:1879–1888 58. Wan H et al. (1999) Der p 1 facilitates transepithelial allergen delivery by disruption of tight junctions [see comments]. J Clin Invest 104:123–133 59. Inai T, Kobayashi J, Shibata Y (1999) Claudin-1 contributes to the epithelial barrier function in MDCK cells. Eur J Cell Biol 78:849–855 60. Sonoda N, Furuse M, Sasaki H, Yonemura S, Katahira J, Horiguchi Y, Tsukita S (1999) Clostridium perfringens enterotoxin fragment removes specific claudins from tight junction strands: Evidence for direct involvement of claudins in tight junction barrier. J Cell Biol 147:195–204 61. Hecht G, Pestic L, Nicevic G, Koutouris A, Tripuraneni J, Lorimer DD, Grzegorz N, Guerriero V, Elson EL, De Lanerolle P (1996) Expression of the catalytic domain of myosin light chain kinase increases paracellular permeability. Am J Physiol 271:C1678–C1684 62. Turner JR, Rill BK, Carlson SL, Carnes D, Kerner R, Mrsny RJ, Madara JL (1997) Physiological regulation of epithelial tight junctions is associated with myosin light-chain phosphorylation. Am J Physiol 273:C1378–C1385 63. Citi S, Cordenonsi M (1998) Tight junction proteins. Biochim Biophys Acta 1448:1–11 64. Fanning AS, Mitic LL, Anderson JM (1999) Transmembrane proteins in the tight junction barrier. J Am Soc Nephrol 10:1337–1345 65. Dunon D, Piali L, Imhof BA (1996) To stick or not to stick:
31. Sirotkin H, Morrow B, Saint-Jore B, Puech A, Das Gupta R, Patanjali SR, Skoultchi A, Weissman SM, Kucherlapati R (1997) Identification, characterization, and precise mapping of a human gene encoding a novel membrane-spanning protein from the 22q11 region deleted in velo-cardio-facial syndrome. Genomics 42:245–251 32. Itoh M, Furuse M, Morita K, Kubota K, Saitou M, Tsukita S (1999) Direct binding of three tight junctionassociated MAGUKs, ZO-1, ZO-2, and ZO-3, with the COOH termini of claudins. J Cell Biol 147:1351–1363 33. Martin-Padura I et al. (1998) Junctional adhesion molecule, a novel member of the immunoglobulin superfamily that distributes at intercellular junctions and modulates monocyte transmigration. J Cell Biol 142:117–127 34. Williams LA, Martin-Padura I, Dejana E, Hogg N, Simmons DL (1999) Identification and characterisation of human Junctional Adhesion Molecule (JAM). Mol Immunol 36:1175– 1188 35. Fujimoto K (1995) Freeze-fracture replica electron microscopy combined with SDS digestion for cytochemical labeling of integral membrane proteins: application to the immunogold labeling of intercellular junctional complex. J Cell Sci 108:3443–3449 36. Furuse M, Fujimoto K, Sato N, Hirase T, Tsukita S, Tsukita S (1996) Overexpression of occludin, a tight junction integral membrane protein, induces the formation of intracellular multilamellar bodies bearing tight junction-like structures. J Cell Sci 109:429–435 37. Balda MS, Whitney JA, Flores C, González S, Cereijido M, Matter K (1996a) Functional dissociation of paracellular permeability and transepithelial electrical resistance and disruption of the apical-basolateral intramembrane diffusion barrier by expression of a mutant tight junction membrane protein. J Cell Biol 134:1031–1049 38. Saitou M, Fujimoto K, Doi Y, Itoh M, Fujimoto T, Furuse M, Takano H, Noda T, Tsukita S (1998) Occludin-deficient embryonic stem cells can differentiate into polarized epithelial cells bearing tight junctions. J Cell Biol 141:391–401 39. Furuse M, Sasaki H, Fujimoto K, Tsukita S (1998b) A single gene product, claudin-1 or -2, reconstitutes tight junction strands and recruits occludin in fibroblasts. J Cell Biol 143:397–408 40. Kubota K, Furuse M, Sasaki H, Sonoda N, Fujita K, Nagafuchi A, Tsukita S (1999) Ca(2+)-independent cell-adhesion activity of claudins, a family of integral membrane proteins localized at tight junctions. Curr Biol 9:1035–1038 41. Matter K (2000) Epithelial polarity: Sorting out the sorters. Curr Biol 10:R39–R42 42. Sakakibara A, Furuse M, Saitou M, Ando Akatsuka Y, Tsukita S (1997) Possible involvement of phosphorylation of occludin in tight junction formation. J Cell Biol 137:1393–1401 43. Balda MS, Flores-Maldonado C, Cereijido M, Matter K (2000a) Multiple domains of occludin participate in regulation of paracellular permeability. J Cell Biochem (in press) 44. Mitic LL, Schneeberger EE, Fanning AS, Anderson JM (1999) Connexin-occludin chimeras containing the ZO-binding domain of occludin localize at MDCK tight junctions and NRK cell contacts. J Cell Biol 146:683–693 45. Kojima T, Sawada N, Chiba H, Kokai Y, Yamamoto M, Urban M, Lee GH, Hertzberg EL, Mochizuki Y, Spray DC (1999) Induction of tight junctions in human connexin 32 (hC × 32)-transfered mouse hepatocytes: connexin 32 interacts with occludin. Biochem Biophys Res Commun 266:222–229 46. Wong V (1997) Phosphorylation of occludin correlates with occludin localization and function at the tight junction. Am J
288
Transmembrane proteins of tight junctions
66. 67.
68.
69.
The new leukocyte homing paradigm. Curr Opin Cell Biol 8:714–723 Huber D, Balda MS, Matter K (1999) Transepithelial migration of neutrophils. Invasion Metastasis 18:70–80 Del Maschio A et al. (1999) Leukocyte recruitment in the cerebrospinal fluid of mice with experimental meningitis is inhibited by an antibody to junctional adhesion molecule (JAM). J Exp Med 190:1351–1356 Huber D, Balda MS, Matter K (2000) Occludin modulates transepithelial migration of neutrophils. J Biol Chem 275:5773–5778 Li D, Mrsny K (2000) Oncogenic Raf-1 disrupts epithelial tight
junctions via downregulation of occludin. J Cell Biol 148:791– 800 70. Balda MS, Matter K (2000b) The tight junction protein ZO-1 and an associated transcription factor regulate ErbB-2 expression. EMBO J (in press) 71. Claude P (1978) Morphological factors influencing transepithelial permeability: A model for the resistance of the zonula occludens. J Membr Biol 39:219–232 72. Cereijido M, González-Mariscal L, Contreras G (1989) Tight junction: Barrier between higher organisms and environment. News Physiol Sci 4:72–74
289