REVIEWS

SIGNALLING TO AND FROM TIGHT JUNCTIONS Karl Matter and Maria S. Balda Tight junctions have long been regarded as simple barriers that separate compartments of different compositions, but recent research indicates that different types of signalling proteins and transduction pathways are associated with these junctions. They receive and convert signals from the cell interior to regulate junction assembly and function, and transmit signals to the cell interior to modulate gene expression and cell behaviour. APICAL DOMAIN

The domain of an epithelial cell that faces the lumen of a cavity or tube, or the outside of the organism. BASOLATERAL DOMAIN

The domain of an epithelial cell that adjoins underlying tissue. GAP JUNCTION

A communicating junction (permeant to molecules up to 1 kDa) between adjacent cells that is composed of 12 connexin protein subunits, 6 of which form a connexon or hemichannel that is contributed by each of the coupled cells. DESMOSOMES

These patch-like intercellular junctions are found in vertebrate tissue, and are particularly abundant in tissues undergoing mechanical stress. The central plaque contains adhesion molecules, and is an anchorage point for cytoskeletal filaments of the intermediate filament type.

Division of Cell Biology, Institute of Ophthalmology, University College London, London, UK. e-mails: [email protected] & [email protected] doi:10.1038/nrm1055

Multicellular organisms have specialized cells, epithelial and endothelial, that form selective barriers between tissues and different body compartments. These cells are polarized — that is, they have an APICAL DOMAIN and a BASOLATERAL DOMAIN — and they adhere to each other through complexes that form junctions between the cells. These intercellular junctions not only carry out adhesive functions but also contain crucial components of the signalling pathways that regulate epithelial proliferation and differentiation. This review focuses on the signalling mechanisms that are related to tight junctions (TJs), which are one type of intercellular junction. Although the identification of the molecular mechanisms that mediate the signal-transduction processes that occur at TJs is far from complete, many recent observations indicate that TJ proteins are important components of numerous signalling pathways that regulate various processes, which range from epithelial-barrier functions to cell proliferation and differentiation. In vertebrates, epithelial cells are joined to each other by a set of intercellular junctions that consists of GAP JUNCTIONS, DESMOSOMES, ADHERENS JUNCTIONS (AJs) and TJs or the zonula occludens (FIG. 1). The latter three junctions are often referred to as the ‘epithelial junctional complex’1. TJs generally encircle cells at the apical end of the lateral membrane. They form a paracellular diffusion barrier, or gate, that regulates epithelial permeability, and an intramembrane diffusion barrier, or fence, which restricts the apical–basolateral diffusion of membrane components. AJs and desmosomes are adhesive junctions and are linked to the actin cytoskeleton and INTERMEDIATE FILAMENTS, respectively. Depending

NATURE REVIEWS | MOLECUL AR CELL BIOLOGY

on the epithelium, AJ components can be concentrated close to TJs and colocalize with a prominent ACTIN BELT, whereas in other types of epithelial cells, they can be distributed over the entire lateral membrane as, for example, in cells of the mouse TROPHECTODERM2. Desmosomes are not restricted to a particular site but are distributed along the entire lateral membrane. Gap junctions form intercellular pores that allow the exchange of small hydrophilic molecules between cells. Gap junctions are generally distributed along the lateral membrane but, in some tissues, they can be intercalated with TJs. Structure and composition of epithelial TJs

TJs are detected easily in ultrathin sections of cells and tissues as very close approximations of two neighbouring plasma membranes (FIG. 1b), which, in certain types of preparations, might appear as focal hemifusions. In FREEZE-FRACTURE replicas, these sites of close contact are seen as a network of intramembrane fibrils. As the structure, composition and possible functions of TJs have been discussed in great detail in recent, comprehensive reviews3–7, we will only briefly summarize these data. Consequently, we apply the same criteria as the authors of these reviews for accepting whether the published data are sufficient to conclude that a protein is a TJ component. As TJs and AJs can be very close to each other, such conclusions are ideally based on immunoelectron microscopy data. This technique is often difficult to apply, however, as many antibodies do not work for this technique and/or the investigated protein is not expressed at sufficiently high levels (which is the case for

VOLUME 4 | MARCH 2003 | 2 2 5

© 2003 Nature Publishing Group

REVIEWS

ADHERENS JUNCTION

A cell–cell adhesion complex that contains cadherins and catenins that are attached to cytoplasmic actin filaments. INTERMEDIATE FILAMENT

A cytoskeletal filament, which is typically 10 nm in diameter, that is present in higher eukaryotic cells.

many signalling components, for example). Therefore, colocalization with TJ markers as viewed by confocal microscopy in cells with a clearly distinct distribution of AJs, such as, for example, mouse trophectoderm or Madin–Darby canine kidney (MDCK) cells, combined with biochemical interactions with known TJ proteins, is often taken as sufficient evidence for the association of proteins with TJs. Epithelial TJs are composed of at least three types of transmembrane protein — occludin, claudin and Apical

a

Tight junctions Actin filament Adherens junctions Intermediate filament Desmosomes

Lateral Nucleus Gap junctions

Basal

Hemidesmosomes

b

TJ

AJ

Figure 1 | Epithelial intercellular junctions. a | A schematic representation of a polarized absorptive epithelial cell. The different types of intercellular junctions, as well as hemidesmosomes, a type of cell–extracellular matrix adhesion, are shown. It should be noted that tight junctions (TJs) and adherens junctions (AJs) are linked to the actin cytoskeleton, and desmosomes and hemidesmosomes are linked to intermediate filaments. b | An electron micrograph of TJs and AJs. The electron micrograph shows an ultrathin frozen section of the junctional region of retinal pigment epithelial cells. The section was labelled with an antibody specific for the TJ component ZO-1 followed by a secondary antibody that was conjugated to colloidal gold. Indicated are TJs and AJs. The latter junction can be recognized by the prominent electron dense subjunctional material that indicates the presence of a well-established junctional actin ring. Note the specific labelling of TJs by the anti-ZO-1 antibody, illustrating the presence of biochemically distinct intercellular junctions. The electron micrograph was provided courtesy of C. Futter, Institute of Ophthalmology, University College London, UK.

226

| MARCH 2003 | VOLUME 4

adhesion proteins of the immunoglobulin superfamily, such as the junctional adhesion molecules (JAMs) — and a cytoplasmic ‘plaque’ consisting of many different proteins that form large complexes (FIG. 2). The transmembrane proteins mediate cell adhesion and are thought to constitute the intramembrane and paracellular diffusion barriers. The claudins and occludin are polytopic membrane proteins with four transmembrane domains. They are thought to assemble into heteropolymers that form the intramembrane strands that can be seen in freeze-fracture replicas. These strands have been proposed to contain fluctuating channels that allow the selective diffusion of ions and small hydrophilic molecules. In this model, the types and concentrations of claudins and occludin that are expressed by particular epithelia determine the ion and size selectivity of the paracellular diffusion pathway. The cytoplasmic plaque of TJs is formed by different types of proteins that include ADAPTORS, such as the zonula occludens (ZO) proteins and other proteins that contain PDZ DOMAINS, as well as regulatory and signalling components (FIG. 2). The adaptor proteins are thought to be connected to the transmembrane proteins and to recruit other cytosolic components, such as protein kinases, GTPases and transcription factors, to TJs. Several components of the cytoplasmic plaque, in addition to the transmembrane protein occludin, also function as cytoskeletal linkers that interact with the actin cytoskeleton. Although the physiological relevance of particular cytoskeletal interactions has not been determined, a large body of evidence indicates that signalling to the actin cytoskeleton is fundamentally important in the regulation of TJ assembly and function. Evolutionary conservation of TJ functions

Although the architecture of intercellular junctions differs between vertebrates and invertebrates, many of the functions that are associated with intercellular junctions are conserved. In Drosophila melanogaster, the junctional complex consists of SEPTATE JUNCTIONS, AJs, and the marginal zone or subapical complex8 (FIG. 3). As septate junctions often restrict paracellular permeability, it was assumed that they were the fly equivalent of vertebrate TJs. However, the composition of septate junctions is quite different, and the evolutionarily conserved components of vertebrate TJs are not associated with septate junctions but either with the subapical complex or AJs, which indicates that vertebrate TJs might have functions that are associated with different types of insect junction. Examples include ZO-1 (Tamou/Polychaetoid), which localizes to AJs in flies, or two recently discovered TJ-associated evolutionarily conserved signalling complexes (see below) that localize to the subapical complex/marginal zone in flies. So, although many junctional functions and signalling pathways have been conserved, they have been regrouped in different combinations during evolution; so, a component that is associated with TJs in vertebrates can be associated with AJs in flies.

www.nature.com/reviews/molcellbio

© 2003 Nature Publishing Group

REVIEWS

Signalling to TJs

Actin filaments

Transcriptional and post-transcriptional regulators For example, symplekin, ZONAB and huASH1

Regulatory proteins For example, Rab13, Rab3B, G proteins, aPKC, PP2A and PTEN Adaptors For example, ZO-1, -2, -3, MAGI-1, -2, -3, PAR3/6, cingulin, Pals1, PATJ and MUPP1

Transmembrane proteins For example, JAMs, claudins and occludin

Figure 2 | The composition of tight junctions. The biochemical composition of epithelial tight junctions (TJs) is outlined. For simplicity, the junctional components have been grouped into transmembrane proteins, adaptors, regulatory proteins, and transcriptional and posttranscriptional regulators. Although examples are given for each group, the list provided is not complete and only proteins that are discussed in the text are shown. A comprehensive list of TJ components can be found in recent reviews6,7. It should be noted that some of the components that are thought to become recruited to TJs do not localize exclusively to TJs, but can also localize to the nucleus (for example, huASH1 and ZONAB) or to other areas of the plasma membrane (for example, PTEN, PP2A and heterotrimeric G proteins). In addition, many TJ components interact directly or indirectly with actin filaments. aPKC, atypical protein kinase C; huASH1, human absent, small or homeotic discs 1; JAMs, junctional adhesion molecules; MAGI, Membrane-associated guanylate kinase inverted; MUPP1, multi-PDZ domain protein 1; Pals1, Protein associated with Lin-7 1; PAR, Partitioning defective; PATJ, Pals1-associated tight junction protein; PP2A, protein phosphatase 2A; PTEN, phosphatase and tensin homologue; ZO, zonula occludens; ZONAB, ZO-1associated nucleic-acid binding.

Bidirectional signalling and TJs ACTIN BELT

A ring of actin filaments that circumvents many absorptive epithelial cells (for example, intestinal epithelial cells) at the level of the junctional complex. TROPHECTODERM

The outer epithelial layer of the blastocyst. FREEZE-FRACTURE

A method of visualizing the interior of cell membranes. Cells are frozen at the temperature of liquid nitrogen in the presence of antifreeze and the frozen block is then cracked with a knife blade. The fracture plane often passes through the hydrophobic middle of lipid bilayers, thereby exposing the interior of cell membranes. The resulting fracture faces are shadowed with platinum, the organic material is dissolved away and the replicas are floated off for electron microscopy.

TJs are involved in two principal types of signal-transduction process: signals are transduced from the cell interior towards forming or existing TJs to guide their assembly and to regulate their function; and signals are transmitted from the TJs to the cell interior to modulate gene expression, as well as cell proliferation and differentiation (FIG. 4). The molecular mechanisms that mediate these signal-transduction processes have only recently started to become unravelled; consequently, our knowledge of them is incomplete and entire pathways have not yet been identified. In the next sections, we summarize what is known and, as far as possible, try to connect these observations. For the purposes of this review, we have divided the known signalling components and pathways into two groups: those that have been preferentially associated with signalling to TJs and those that have been linked to signalling from TJs. Nevertheless, it is clear that these signalling processes are intertwined: a component that has been associated with signalling to the junction (for example, during assembly) might also influence the transmission of signals to the cell interior once the junctions are fully formed.

NATURE REVIEWS | MOLECUL AR CELL BIOLOGY

The intracellular signalling pathways that regulate TJ assembly and function have received much interest, but it is often difficult to differentiate between those pathways that regulate junction assembly and those that modulate selective paracellular permeability. This is due primarily to the fact that these two types of pathway overlap, and that the intensity of the stimulus seems to determine whether a pathway triggers selective increases in paracellular permeability or disassembly of intercellular junctions. Assembly of TJs can be studied in different experimental systems (BOX 1), but the processes and principles that govern epithelial junction assembly seem to be similar in all model systems. They have been most extensively studied in MDCK cells, in which calciumdependent E-cadherin-mediated cell–cell adhesion functions as the initial signal that triggers assembly of all intercellular junctions9–11. After E-cadherin-dependent intercellular adhesion has been established, the first step seems to be the assembly of a primordial junction that contains AJ and TJ components. The junctional complex then matures and establishes distinct AJs and TJs. Several different types of signalling pathway and protein have been implicated to participate in the stimulation of TJ assembly12,13. These include protein kinase A (PKA), monomeric and heterotrimeric G proteins and various protein kinase C (PKC) isotypes. Protein kinase A. Although stimulation of PKA prevents the disassembly of TJs after antibody-based neutralization of E-cadherin or the removal of calcium 9,14, it also inhibits TJ assembly in a calcium switch protocol, which is a model system for the de novo formation of intercellular junctions12 (BOX 1). Similarly, the effects of PKA on TJ permeability are controversial13, which indicates that various parameters might modify the response to PKA stimulation, such as strength of stimulus, cellular background and crosstalk with other signalling pathways. Heterotrimeric G proteins. Early pharmacological studies with PERTUSSIS TOXIN indicated that inhibition of inhibitory G-proteins (Gi) stimulates the assembly of TJs12, and Gαi2 was found at cell–cell contacts, where it colocalizes with ZO-1 (REFS 15–17). However, ectopically expressed αi proteins were shown to stimulate TJ assembly17. Similar to PKA, the reason for this contradiction between the pharmacological and transfection experiments is not clear, but it might be caused by the inhibition of several different αi proteins by pertussis toxin. The involvement of many G proteins in TJ regulation is supported by the recent finding that the expression of Gα12 increases paracellular permeability, which is an effect that might be mediated by ZO-1 (REF. 18). Protein kinase C. The PKC family of serine/threonine kinases has been divided into classical (cPKC), novel (nPKC) and atypical (aPKC) PKCs. The cPKC (α, β and γ) and nPKC isotypes (δ, ε, η and θ) are activated by diacylglycerol (DAG) and phosphoserine. Both of these isotype classes are also activated by phorbol esters.

VOLUME 4 | MARCH 2003 | 2 2 7

© 2003 Nature Publishing Group

REVIEWS

Vertebrates

Insects

Tight junctions Paracellular diffusion barrier PAR3–PAR6–aPKC Pals1–PATJ–Crumbs ZO-1

Subapical complex/ marginal zone Bazooka–PAR6–aPKC Stardust–Discs lost–Crumbs

Adherens junctions

Adherens junctions Tamou/Polychaetoid

Desmosomes

Septate junctions Paracellular diffusion barrier

Figure 3 | Intercellular junctions in vertebrates and insects. A schematic representation of epithelial intercellular junctions in vertebrates and insects (Drosophila). Only adherens junctions (AJs) are structurally and functionally conserved between vertebrates and insect cells. By contrast, the functions of vertebrate tight junctions (TJs) are associated with different types of intercellular junction in flies, which indicates that particular junctional functions and signalling pathways might have been conserved during evolution but have been regrouped and linked to different types of intercellular junctions. As both TJs and septate junctions act as paracellular diffusion barriers, they have generally been regarded as functional equivalents. On the basis of their composition, however, TJs and septate junctions are very different and known insect homologues of TJ components that are components of conserved signalling pathways localize to either AJs, such as Tamou/Polychaetoid (the Drosophila homologue of ZO-1), or the subapical complex/marginal zone, such as the Bazooka–PAR6–aPKC (the Drosophila equivalent of PAR3–PAR6–aPKC) and the Stardust–Discs lost–Crumbs (the Drosophila equivalent of Pals1–PATJ–Crumbs) complexes. Note that Drosophila Crumbs, as well as the only analysed mammalian Crumbs protein, can spread over the entire apical membrane of epithelial cells57,67. Another conserved junctional component is AF-6/afadin, which has been shown to localize to TJs in vertebrates and AJs in flies; however, the localization of AF-6/afadin to TJs is controversial, as it has also been shown to interact with nectins at AJs. aPKC, atypical protein kinase C; Pals1, Protein associated with Lin-7 1; PAR, Partitioning defective; PATJ, Pals1-associated tight junction protein; ZO, zonula occludens.

ADAPTOR PROTEINS

Proteins that augment cellular responses by recruiting other proteins to a complex. They usually have several protein–protein interaction domains. PDZ DOMAIN

A protein interaction domain that is often present in scaffolding proteins and is named after the founding members of this protein family (PSD-95, Discs-large A and ZO-1). SEPTATE JUNCTION

A junction that is basal to the zonula adherens in Drosophila epithelial cells.

228

cPKCs also require calcium for activation, whereas nPKCs do not. The aPKCs (ζ, λ, and PRK; for PKCrelated kinases) are calcium insensitive and do not respond to phorbol esters19. As potent activators of cPKC and nPKC signalling pathways, phorbol esters have been shown to induce the disassembly of TJs in many different cell lines13. However, inhibition of PKCs blocks both TJ assembly and disassembly, which implies that it is transient activation of PKCs that is important in junctional dynamics12,20,21. The importance of DAG-regulated PKCs in junction assembly is supported by experiments showing that the cell-permeable DAG analogue dioctanoylglycerol can stimulate partial junction assembly even when E-cadherin-mediated cell–cell adhesion is blocked22. Several PKC isoforms have been localized close to TJs, but little is known about the molecular mechanisms

| MARCH 2003 | VOLUME 4

by which different PKCs regulate junctional dynamics, with the exception of aPKCs, which are calcium- and DAG-independent. The aPKCs are part of an evolutionarily conserved signalling pathway that regulates junction assembly and cell polarity. They form a complex with PAR3 (which is also known as ASIP, for atypical PKC isotype-specific interacting protein) and PAR6 (REFS 23,24). PAR (partitioning defective) genes were identified originally in the nematode Caenorhabditis elegans, in which they regulate asymmetric cell division of the zygote25. It is now clear, however, that homologues of PAR3 and PAR6 are evolutionarily conserved regulators of cell polarity in Drosophila and mammals8,26. In mammals, PAR3 and PAR6 are PDZ domain proteins that localize to epithelial TJs. Recruitment to TJs is thought to be mediated by binding of PAR3 to the transmembrane protein JAM27,28. The successful assembly of active aPKC–PAR3–PAR6 complexes at TJs is important for TJ formation and epithelial polarization29–31. PAR6 also binds to GTP-bound cell-division control protein 42 (Cdc42), a Rho-family GTPase that is essential for epithelial-cell polarity32, and this is thought to result in the activation of aPKC29,30. So, one of the functions of Cdc42 during the formation of polarized epithelia is to regulate the activation of aPKC–PAR3–PAR6 complexes and, thereby, TJ assembly. Although the importance of recruitment of the aPKC–PAR3–PAR6 complex for the assembly of TJs has been established, it is not known how this complex actually regulates assembly. In wound-healing experiments, aPKC activity is not required for the initial assembly of junctional components at sites of new cell–cell adhesion to form a primordial junction, but for the maturation of the junctional complex into distinct TJs and AJs33. This indicates that aPKC might mediate junctional identity and that the aPKC–PAR3–PAR6 complex functions as a promoter for the assembly of functional TJs rather than cell–cell adhesion as such. This is supported by the observation that overexpression of PAR3 results in a faster formation of monolayers that have increased transepithelial electrical resistance (TER); that is, they have decreased ionic permeability 34. Moreover, aPKC activity is regulated negatively by PP2A (protein phosphatase 2A), which directly interacts with and dephosphorylates aPKC 35. In vitro, PP2A can also dephosphorylate TJ proteins that have been phosphorylated by aPKC, and the overexpression of its catalytic subunit in MDCK cells has been shown to result in dephosphorylation of TJ proteins and increased paracellular permeability, which indicates that PP2A is probably a negative regulator of TJs. These data indicate that the aPKC–PAR3–PAR6 complex and PP2A might represent opposing signalling pathways that regulate TJ assembly and disassembly. Taken together, these data favour a model for the regulation of intercellular junction assembly in which the aPKC–PAR3–PAR6 complex functions as a regulator at a late stage of junction biogenesis (FIG. 5). Assembly of intercellular junctions is generally triggered by the initiation of cell–cell adhesion by E-cadherin and nectins (BOX 2), which results in the recruitment of numerous

www.nature.com/reviews/molcellbio

© 2003 Nature Publishing Group

REVIEWS components of AJs (for example, catenins and afadin/AF-6) and TJs (for example, JAM, occludin, claudins, ZO-1 and aPKC–PAR3–PAR6) to these primordial junctions. Recruitment of these components is thought to be mediated by numerous protein–protein interactions not only between proteins of the same type of junction but also between proteins of different junction types, such as ZO-1 and α-catenin. In addition, the activation of E-cadherin or nectin-mediated adhesion regulates activation of the Cdc42-dependent polarity pathway36,37, which results in the stabilization of AJs as well as the activation of aPKC complexed to PAR3–PAR6 and, consequently, maturation of the junctional complex into distinct TJs and AJs. As PP2A counteracts aPKC function, it would be expected that initiation of cell–cell adhesion downregulates PP2A activity; however, such data have not yet been provided. Rho family GTPases. Three classes of small GTPases have been linked functionally or physically to TJs: Ras, Rabs and Rho-family GTPases. The involvement of Ras GTPases in TJ dynamics is discussed later, and the function of the TJ-associated Rab proteins is still relatively poorly understood38. However, as Rabs also regulate membrane transport, they might have a role either in TJ assembly or in regulating some aspect of cell–cell junctions in membrane transport39,40. Rho-family GTPases have been linked to both the regulation of junction assembly and the regulation of selective paracellular permeability. A role for GTPases in junction assembly was first implied by experiments using GTPγS. GTPγS inhibited the assembly of junctions during a calcium switch 12 (BOX 1). Furthermore, microinjection of C3 transferase, a toxin that inactivates Rho, inhibited the assembly of AJs and TJs41–43. When RhoA was activated in MDCK cells after transfection of a G-protein-coupled

Type of transmitted information TJ assembly and disassembly Regulation of paracellular permeability Tight junctions

Cell interior Gene expression Regulation of cell proliferation Regulation of differentiation

Figure 4 | Signal transduction to and from tight junctions. A schematic representation of the type of information that is transmitted from the cell interior to tight junctions (TJs) and from TJs to the cell interior. These signal-transduction pathways are a complex network of molecular mechanisms that we are only just beginning to understand. This review has been structured according to this scheme; it is clear, however, that certain signalling components might function in the transmission of signals in either direction.

prostaglandin receptor, an increase in TER — which indicates reduced ionic permeability — paralleled by an increase in paracellular tracer permeability, was obtained44. However, the expression of constitutively activated forms of RhoA and Rac1 has been reported to perturb TJ function in epithelia and endothelia45–47, which implicates these proteins in the disassembly of TJs. Inactivation of Rho-family GTPases in brain endothelial cells by C3 transferase inhibits the ability of lymphocytes to migrate across them, which indicates that junctional dynamics requires Rho activation48. Therefore, Rhofamily GTPase activities seem to be carefully balanced, and neither high nor low levels of Rho activity might be optimal for TJ integrity. Alternatively, the regulation of TJs by Rho GTPases might vary depending on the cell type and/or the ligand that induces the response, similar to the effect that Rho GTPases have on cadherin function in different cellular contexts47.

Box 1 | Model systems to study junction assembly and epithelial polarization

PERTUSSIS TOXIN

A mixture of proteins that is produced by Bordetella pertussis. It blocks the function of Gi proteins by catalysing ADP ribosylation of the α-subunit. GTPγS

A non-hydrolysable analogue of GTP.

Different experimental model systems and strategies have been developed to analyse TJ assembly and regulation. Many of these systems involve culturing an epithelial cell line on a permeable support115. This allows the measurement of parameters that are characteristic of the integrity and functional properties of the paracellular diffusion barrier that is formed by TJs. Two parameters that are often measured are the transepithelial electrical resistance (TER) — which is an instantaneous measurement of ionic conductivity that can be used to determine the integrity as well as ion selectivity of TJs — and the paracellular permeability of hydrophilic tracers, which is a measurement taken over a longer period of time that allows the quantification of slow diffusion across TJs and determination of size selectivity of the paracellular diffusion barrier. The use of permeable supports has been combined with different experimental protocols to allow the manipulation of junction assembly. In the calcium-switch protocol, cells are plated in medium that contains a low concentration of calcium and, therefore, cannot form cell–cell junctions; the addition of calcium to the medium then triggers junction assembly10. In an alternative protocol, ATP is depleted in cells that have already formed monolayers, which results in the dissociation of cell–cell junctions116. In vitro wound-healing assays, in which confluent epithelial monolayers are wounded by scratching the monolayers with a needle or pipette tip and then the reformation of the monolayer studied, are popular assays for studying the formation of intercellular junctions. To study junction dynamics during leukocyte transmigration, epithelial cells are plated on large-pore filters, which allows transmigrating neutrophils to cross from one side of the monolayer to the other117. Alternatively, de novo TJ assembly can be studied during early development2. Although experiments with this model are still at an early stage, in which the sequential recruitment of components to the forming junctional complex is correlated with the development of functional intercellular junctions and epithelial differentiation, this is a promising model system that will allow the relevance of data that are obtained from tissue-culture cell lines to the development of epithelial tissues to be determined.

NATURE REVIEWS | MOLECUL AR CELL BIOLOGY

VOLUME 4 | MARCH 2003 | 2 2 9

© 2003 Nature Publishing Group

REVIEWS

PP2A Recruitment of TJ components to primordial junctions (for example, PAR3–PAR6–aPKC, JAM, and so on)

Activation of aPKC

Initiation of cell–cell adhesion E-cadherin, nectin

Maturation of junctional complex Formation of distinct AJs and TJs Stimulation of Cdc42

Stabilization of cell–cell adhesion

Figure 5 | Signalling by aPKC and assembly of the junctional complex. A schematic representation of the sequence of events that lead to the assembly of distinct adherens junctions (AJs) and tight junctions (TJs). Initial assembly starts with the formation of a primordial junction that contains components of AJs and TJs. On adhesion-mediated activation of cell-division control protein 42 (Cdc42), this results in stabilization of cell adhesion and the activation of atypical protein kinase C (aPKC), which forms an evolutionarily conserved complex with PAR3 (which is also known as ASIP) and PAR6. Active aPKC then stimulates formation of distinct AJs and TJs. Protein phosphatase 2A (PP2A) seems to be able to affect aPKC as well as the phosphorylation of individual TJ components and can therefore inhibit aPKC function and trigger junction disassembly. JAM, junctional adhesion molecule; PAR, partitioning defective.

Rho effector proteins. Effector proteins that are activated by Rho GTPases and that regulate TJs have only recently started to be identified (FIG. 6). RhoA-dependent modification of paracellular permeability involves Rho-associated kinase (ROCK)49,50. The actin cytoskeleton has a crucial role in the regulation of TJs through direct interactions between the actin cytoskeleton and certain TJ components such as occludin, cingulin and the ZO proteins3,7,51. Given the importance of Rho-family GTPases in the regulation of the actin cytoskeleton52, it is therefore probable that they regulate TJs by inducing changes in the actin cytoskeleton. However, Rho might also affect TJ proteins directly. For example, it has been suggested that Rho-stimulated phosphorylation of the carboxy-terminal domain of occludin regulates the interaction of occludin with the submembrane cytoskeleton and thereby regulates its function in selective paracellular permeability50. As well as modulating the interactions between TJ proteins and the actin cytoskeleton, Rho might also influence TJs by mechanisms that involve the regulation of cortical actin contraction through the activation of the motor protein myosin53,54. The activity of myosin is regulated by the opposing actions of myosin lightchain phosphatase and myosin light-chain kinase. This regulatory system of myosin-mediated actin contraction is modulated by ROCK, and this might be important for regulating paracellular permeability (FIG. 6). Alternatively, permeability could also be controlled by other Rho effectors and pathways, such as the mouse Diaphanous protein homologue (Dia) or regulators of actin polymerization, the roles of which have not yet been analysed in TJ function.

GUANINE NUCLEOTIDE EXCHANGE FACTOR

(GEF). A protein that facilitates the exchange of GDP for GTP in the nucleotide-binding pocket of a GTP-binding protein.

230

Rho activation. Although a large body of evidence indicates that Rho and — to a lesser extent — Rac might be involved in the regulation of TJs, little is known about the mechanisms that control these GTPases. Heterotrimeric G proteins have been implicated in the activation of Rho GTPases as well as the regulation of TJs12,17,18,55, but a direct link has yet to be found. Nevertheless, the activation of Rho by the prostaglandin E

| MARCH 2003 | VOLUME 4

receptor that was described above seems to be regulated by G proteins44. Activation of Rho GTPases generally requires a GUANINE NUCLEOTIDE EXCHANGE FACTOR (GEF), and a TJ-associated GEF for Rho has recently been identified. GEF-H1/Lfc has been shown to associate with TJs and to regulate paracellular permeability56. The observations that are outlined here indicate that Rho-mediated regulation of the actin cytoskeleton is important for regulating junctional dynamics and paracellular permeability, but also indicate a potential involvement for actin-independent mechanisms. Although the regulation of Rho involves at least one junction-associated GEF, the signalling systems that activate such regulators and trigger Rho activation have yet to be discovered. The molecular identification of such regulatory pathways that target Rho in a TJ-specific manner will be one of the main challenges of the near future. Signalling from TJs

Crumbs and epithelial organization. One of the most exciting recent developments is the recognition that components of the evolutionarily conserved Crumbs signalling pathway localize at, or close to, epithelial TJs. The Crumbs protein was identified in Drosophila as an apical membrane protein that is required for the polarized organization of ectodermal epithelia57. Crumbs is not specific for epithelia, however, it is also expressed by various types of neuron and has been linked to the differentiation of photoreceptors58–60. In Drosophila, Crumbs functions in the organization of the apical membrane, and its expression level seems to correlate with the extent of apical-membrane biogenesis, as well as with the formation and positioning of AJs61,62. The protein seems to be concentrated in the subapical complex/marginal zone, where it interacts with two adaptor proteins, Discs lost and Stardust63,64. In mammalian epithelial cells, the TJ-associated protein Pals1, which is a homologue of Stardust, binds to the mammalian homologue of Discs lost — PATJ — as well as to transfected chimaeras that contain the cytoplasmic domain of a mammalian Crumbs homologue65,66. Three human Crumbs homologues have been

www.nature.com/reviews/molcellbio

© 2003 Nature Publishing Group

REVIEWS

Box 2 | Interactions between tight and adherens junction components Although adherens junctions (AJs) and tight junctions (TJs) are distinct intercellular junctions, several of their components can interact with each other. The TJ component zonula occludens-1 (ZO-1) has the most variable distribution of the known junctional proteins and, depending on the cell type, has been localized to TJs, AJs and gap junctions. The association with AJs or gap junctions reflects the interaction of ZO-1 with α-catenin, a component of the cytoplasmic complex that interacts with E-cadherin, or with certain connexins, which are the transmembrane components of gap junctions118–121. Nevertheless, in cells that have TJs, ZO-1 specifically associates with TJs and not AJs or gap junctions. Another example is AF-6/afadin, which can interact with TJ proteins such as junctional adhesion molecule (JAM), but also with nectins, which are an important class of AJ protein83,84,122. Although these interactions might initially seem surprising, they could reflect the evolutionary origin of intercellular junctions from a single simple intercellular adhesion complex — such as the primordial junctions that are seen during junction assembly in in vitro model systems or early development — that evolved into different types of intercellular junctions. Even though many of the interactions that occur between TJ and AJ components seem to be important during junction assembly, some of them might also be relevant for the transmission of signals from intercellular junctions to the cell interior. The best characterized junction-associated signalling system is based on β-catenin, which activates T-cell factor (TCF)/Lef-mediated transcription in response to Wnt growth factors98,99. As ZO-1 can interact with αcatenin, a binding partner of junction-associated β-catenin, it is conceivable that ZO-1 influences β-catenin signalling. Indeed, expression of truncated ZO-1 in certain cell lines and manipulation of occludin, a transmembrane protein that interacts with ZO-1, can cause activation of β-catenin–TCF/Lef-mediated transcription91,93. However, the mechanisms of activation have not been identified; so, it is not clear whether these observations are due to direct or indirect effects. In addition, the domain of ZO-1 that interacts with α-catenin is not required for its anti-proliferative function113. On the other hand, activation of β-catenin-signalling has been shown to downregulate ZO-1 expression97. As both TJ components have been linked to the suppression of proliferation, it could be that β-catenin signalling inactivates the proliferation suppressive function of TJs.

identified, but so far only Crumbs 3 has been localized in epithelial cells and shown to be distributed along the entire apical membrane66,67. The three human Crumbs genes seem to be expressed in a tissue-specific manner, which indicates that Crumbs signalling in mammals might be more complex than in Drosophila, involving different types of activators and causing more variable cellular responses. However, these data do indicate that there is a well-conserved signalling pathway at the apical–lateral boundary and that this coincides with the subapical complex/marginal zone in Drosophila and TJs in vertebrates (FIG. 3). Little is known about the functional roles of Crumbs and Crumbs-associated proteins in mammalian epithelial cells. Nevertheless, overexpression of PATJ has been shown to interfere with the recruitment of ZO-1 and ZO-3 to forming TJs, and this probably involves a direct interaction between ZO-3 and PATJ66,68. This indicates that Crumbs signalling might regulate TJ assembly or participate in TJ-associated signalling systems that regulate epithelial proliferation and differentiation (for example, ZO-1-associated signalling systems, see below). An involvement of Crumbs in TJ assembly is supported by the recent finding that Pals1 links the Crumbs pathway to the PAR3–PAR6–aPKC complex, which is known to function in TJ assembly69. In addition, the network that is formed by TJ membrane proteins and associated adaptor proteins might function as a scaffold to recruit the Crumbs signalling complex. As the overexpression of Crumbs in Drosophila results in expansion of the apical domain61, the pathway might then function in the recruitment of components that are involved in apical-membrane biogenesis, such as proteins that regulate apical-membrane traffic or the assembly of the apical cytoskeleton.

NATURE REVIEWS | MOLECUL AR CELL BIOLOGY

Tight junctions and Ras signalling. Intercellular junctions are known to contribute to the regulation of differentiation. The loss of epithelial differentiation in cancers often correlates with mutations in the small GTPase Ras. Ras is known to mediate growth factor- and extracellular matrix signalling, but it also influences cell–cell adhesion70–72. Many signalling pathways, including the Raf pathway, are activated by Ras and regulate cell-cycle entry73. In addition, less well-characterized Ras effectors such as the junction-associated protein afadin /AF-6 are known to participate in differentiation74. TJs have been linked to two Ras effectors. Raf kinases are common Ras effectors, the activity of which leads to the sequential activation of the extracellular-signal regulated kinase/mitogen-activated protein kinase (ERK/MAPK) pathway and the transcription factor Elk1. Expression of activated Raf-1, or constitutively active MAPK and ERK kinase 1 (MEK1), is sufficient to transform epithelial cells75,76. Consequently, treatment of Ras-transformed MDCK cells with the selective MEK1 inhibitor PD98059 results in an increased expression of AJ proteins and the assembly of functional AJs and TJs77,78. TJs seem to be connected intimately with Raf-1 signalling, as overexpression of occludin is sufficient to reverse Raf-1-mediated transformation of a salivary-gland epithelial cell line76. The molecular mechanism by which TJs can counteract Raf signalling is not yet known, and it is not clear whether the inhibition of Raf-1-mediated transformation by occludin is direct or involves other signalling pathways. The PDZ domain protein AF-6 was originally identified as a Ras target but it seems to bind more tightly to the Ras-like GTPase Rap1 (REFS 79,80). It has been localized to TJs, but afadin, which is the rat homologue of AF-6, has also been reported to be recruited to

VOLUME 4 | MARCH 2003 | 2 3 1

© 2003 Nature Publishing Group

REVIEWS

MLCK Myosin

MLCP Actin Rho kinases

Rho

PKCs Submembrane adaptors GEF-H1/Lfc Transmembrane components

Stimulus

Paracellular pathway

Figure 6 | Regulation of paracellular permeability by Rho. A schematic representation of the signalling proteins that have so far been implicated in the regulation of tight junctions (TJs), and how they might regulate paracellular permeability. Transmembrane components of TJs are shown in red. Of the known transmembrane proteins, only claudins and occludin are thought to participate directly in the formation of the selective permeability barrier. The submembrane adaptors of the junctional plaque are indicated in purple. Several of the adaptors function as linkers to the actin-based cytoskeleton. Regulation of the actin cytoskeleton by changing the activities of myosin light-chain phosphatase (MLCP) and myosin light-chain kinase (MLCK) seems to be an important mechanism of regulation of paracellular permeability. Although protein kinase Cs (PKCs) are shown as a single symbol, different isoforms of PKC regulate TJ assembly and function. One of the main pathways of TJ regulation is signalling by the small GTPase Rho. This Rho-dependent pathway probably involves Rho effectors such as Rho kinases as well as Rho-specific guanine nucleotide exchange factors (GEFs), such as GEF-H1/Lfc. Rho-stimulated changes in paracellular permeability might involve modulation of myosin activity or the mechanisms that involve direct modifications of TJ proteins such as phosphorylation of occludin.

nectin-based AJs81,82, indicating that AF-6/afadin might not be a specific component of TJs. It is not yet clear what the reason for these controversial results is. As AF-6/afadin can interact with transmembrane components of AJs (such as nectins) as well as TJs (such as JAM)83,84, it might associate with more than one intercellular junction. It could also be, however, that the interactions with TJ proteins reflect transient associations that occur during junction assembly. Although the mechanisms by which AF-6/afadin mediates signalling are not known, this protein is known to be crucial for the establishment of intercellular junctions during mouse development74,85.

MAGUK PROTEINS

Scaffold proteins that contain PSD-95–Discs-large A–Zonula occludens-1 (PDZ), Srchomology-3 (SH3) and guanylate kinase domains.

232

Regulation of gene expression. There is increasing evidence that TJs have a role in the regulation of gene expression and proliferation. Much of this evidence comes from TJ-associated adaptor proteins. The identification of the primary structure of the TJ plaque protein ZO-1 showed that it belongs to the same protein family as the Drosophila Discs-large tumour suppressor (DlgA),

| MARCH 2003 | VOLUME 4

which indicates that it has a potential function in negatively regulating cell growth86,87. As members of the Discs-large protein family are associated with membranes and contain a domain that is homologous to yeast guanylate kinase, they were named MAGUKS (for membrane-associated guanylate kinases). In addition to ZO-1, several other MAGUK proteins, such as ZO-2, ZO-3 and the membrane-associated guanylate kinase inverted (MAGI) proteins, have now been identified at TJs6,7. A possible function of ZO-1 in the regulation of cell differentiation has been indicated by the finding that mutations in its Drosophila homologue, called Tamou/Polychaetoid, cause defects in dorsal closure, epithelial migration and cell-fate determination in sensory organs88,89. Interestingly, Tamou/Polychaetoid colocalizes with Canoe, which is the Drosophila homologue of AF-6, at AJs, and the two proteins interact physically and genetically in the regulation of dorsal closure90. Support for a function of TJ proteins in controlling epithelial proliferation and differentiation in vertebrates came from experiments in which TJs were manipulated either by the expression of truncated ZO-1 mutants or the addition of inhibitory peptides against an extracellular domain of occludin. Both treatments resulted in at least partial dedifferentiation91–93. Further evidence came from studies using tumour tissues that showed that the expression of ZO-1 and ZO-2 can be altered in certain tumours, and that ZO-1 is downregulated in colorectal cells when β-catenin is overexpressed94–97. β-catenin mediates Wingless/Wnt growth-factor signalling by regulating gene expression together with the T-cell factor (TCF)/Lef transcription factor, and activation of β-catenin signalling is seen in many cancers and generally promotes proliferation98,99. Therefore, β-catenin signalling seems to inactivate possible TJ-associated growth suppressors. Moreover, it has recently been shown that the TJassociated adaptor proteins ZO-2, MAGI-1 and MUPP1 (for multi-PDZ domain protein 1) can counteract viral oncogenes100–102. ZO-2 also interacts with the nuclear scaffolding factor, scaffold attachment factor B (SAF-B)103; however, it is not known whether these different interactions are functionally related. Examples of other TJ adaptors that inhibit signalling along a proliferationpromoting pathway are MAGI-2 and MAGI-3. These two proteins interact with the tumour suppressor phosphatase and tensin homologue (PTEN), which is a lipid phosphatase. This interaction has been proposed to recruit PTEN to TJs and, thereby, to dephosphorylate phosphatidylinositol phosphates, which therefore inhibits signalling by protein kinase B/Akt104,105. An involvement of TJ components in nuclear processes has also been proposed on the basis of reports indicating that some of them can localize to the nucleus. The nuclear localization of the MAGUK proteins ZO-1 and ZO-2 might be dependent on the proliferation state of the cells, as well as on other unknown parameters, as different laboratories disagree about the nuclear localization of ZO-1 (REFS 91,100,103,106–108). Furthermore, the only known effect of ZO-1 on transcription, which is regulation of the ErbB2 gene, does not require detectable nuclear pools of ZO-1, but instead, cytoplasmic sequestration of a transcriptional

www.nature.com/reviews/molcellbio

© 2003 Nature Publishing Group

REVIEWS Y-BOX transcription factor and binds to inverted CCAAT

Sequestration and inhibition of ZONAB and CDK4 Upregulation of TJ-associated suppressors of proliferation, assembly and stabilization of junctional components (for example, ZOs, occludin and MAGIs)

Recruitment of PTEN Inhibition of Raf signalling Inactivation of Rho

Inhibition of proliferation and upregulation of differentiation markers

Figure 7 | Assembly of epithelial tight junctions and regulation of proliferation and differentiation. The schematic shows low-density epithelial cells with low expression levels of tight junction (TJ)-associated proliferation suppressors and high-density cells with fully assembled TJs that no longer proliferate and have high expression levels of differentiation markers. To the left of the arrow are junctional processes that occur when cells reach full confluence, such as the upregulation of the junctional protein zonula occludens (ZO)-1. On the right side of the arrow are signalling pathways or signalling components that have been linked to, or have been proposed to be regulated by, TJs. The red symbol represents TJ-associated transmembrane proteins, the purple symbol represents junctional adaptors such as ZO-1, and TJ-associated transcription factors such as ZO-1-associated nucleic-acid binding protein (ZONAB) are indicated by the orange circles. The pink symbol represents a differentiation marker. It should be noted that the assembly of TJs results in the inhibition of several proliferation-stimulating pathways, which indicates that TJs might have a proliferation-suppressive function. The accumulation of TJ components such as ZO-1 with increasing cell densities could function as a mechanism to sense and titre cell density. Please note that only TJ-associated signalling systems and not pathways that originate at other sites of adhesion are shown. CDK4, cyclin-dependent kinase 4; MAGI, membrane-associated guanylate kinase inverted; PTEN, phosphatase and tensin homologue. Modified with permission from REF. 13 © CRC Press (2001).

SH3 DOMAINS

(Src-homology-3 domain). Protein sequences of about 50 amino acids that recognize and bind sequences that are rich in proline. Y-BOX

A promoter element that generally contains a central ATTGG sequence and interacts with a family of transcription factors that are known as Y-box binding proteins.

repressor108. Another TJ protein that can localize to the nucleus is symplekin109, which has been linked to messenger RNA processing110,111. Direct evidence for a role of TJs in the regulation of gene expression came with the discovery of two TJ-associated transcription factors. Human ASH1 had been cloned on the basis of its homology to Drosophila Ash1 (for absent, small or homeotic discs 1), a transcription factor that is involved in homeotic gene expression during development. huASH1 was found to colocalize with TJ proteins at cell–cell junctions112, but at present it is not clear how TJs contribute to the regulation of huASH1-mediated transcription. The second TJ-associated transcription factor is ZONAB (for ZO-1-associated nucleic-acid binding), which was isolated on the basis of its affinity for the Srchomology-3 (SH3) domain of ZO-1 (REF. 108). ZONAB is a

NATURE REVIEWS | MOLECUL AR CELL BIOLOGY

box sequences. ZONAB localizes to the nucleus, where it participates in the regulation of gene expression, and to intercellular junctions, where it binds to ZO-1. The relative distribution of ZONAB is determined by the cell density. ZONAB localizes primarily to junctions in cells that are grown to a high density, which contain high levels of ZO-1 and only a little ZONAB; however, in cells at a low density, which have low ZO-1 and high ZONAB expression levels, ZONAB accumulates in the nucleus108. So ZO-1 and ZONAB provide epithelial cells with a mechanism to regulate gene expression in a celldensity-dependent manner. The interaction between ZO-1 and ZONAB is of functional relevance, as the two proteins interact in the regulation of expression of the proto-oncogene ErbB2, as well as epithelial proliferation and cell density108,113. Regulation of ErbB2 expression does not require detectable nuclear accumulation of ZO-1, which indicates that ZO-1 does not participate directly in the regulation of transcription. Inhibition of ZONAB function by depletion, or cytoplasmic sequestration by overexpression of ZO-1, inhibits proliferation and reduces the cell density of mature MDCK monolayers113. This reflects a regulatory role for ZONAB and ZO-1 in the G1 to S-phase transition during the cell cycle. In accordance with this, ZONAB forms a complex with the G1 to S-phase regulator cyclin-dependent kinase 4 (CDK4); consequently, experimental reduction of nuclear ZONAB expression also resulted in a decrease of nuclear CDK4 (REF. 113). Therefore, ZO-1, ZONAB and CDK4 are part of a cell-density-dependent signalling mechanism that regulates gene expression and epithelial cell proliferation. Suppression of proliferation. Although there have recently been many exciting observations of TJs and the regulation of epithelial proliferation and differentiation, the picture is far from complete, and connections to other signalling pathways are largely unknown. Much of the evidence indicates that TJs, similar to the AJ-associated protein E-cadherin, suppress proliferation and stimulate differentiation. TJs affect several signalling pathways in a manner that favours growth arrest and differentiation. So it seems that accumulation of junctional proteins at forming TJs is used to regulate signalling pathways that are involved in proliferation and differentiation. Although at present this has not been shown for many TJ proteins, both the half-life and the expression level of ZO-1 increase with cell density108,114, and overexpression of ZO-1 in low-density cells inhibits proliferation113. We therefore propose a working model in which the assembly state of TJs is used as a sensor for cell density (FIG. 7). In subconfluent cells, expression levels of TJ proteins are low, and so the inhibition of signalling pathways that stimulate proliferation is weak; with increasing cell density, expression levels of TJ proteins increase as the proteins become stabilized at the forming TJs. As a consequence, signalling proteins such as ZONAB and PTEN become recruited to TJs. Whereas the transcription factor ZONAB and the

VOLUME 4 | MARCH 2003 | 2 3 3

© 2003 Nature Publishing Group

REVIEWS associated cell-cycle kinase CDK4 become sequestered at TJs, and thereby inhibited108, in the case of PTEN, this has been proposed to stimulate its suppressive function of proliferation by bringing it to its membrane-associated lipid substrates104. Raf-1 signalling might become inhibited owing to the accumulation of occludin at TJs76, and recruitment of a recently identified TJ-associated Rho exchange factor could contribute to the downregulation of active RhoA in confluent monolayers56. Consequently, the formation of fully assembled TJs suppresses numerous proliferation-promoting signalling pathways. Conclusions and perspectives

TJs are connected to different types of signalling pathway that function in the transmission of signals to TJs to regulate their function, and that transduce information towards the cell interior to modulate cell behaviour. Several signalling systems have been linked to these processes, but we are only starting to understand the underlying molecular mechanisms. It will now be important to determine how TJs are modified in response to signals that alter their functional properties. Such knowledge could then be used to study the physiological functions of these transduction systems in in vivo models. Recent exciting results indicate that TJs have a central role in processes that regulate epithelial proliferation and differentiation. Although these studies are at an early stage, the collective evidence indicates that fully formed TJs in high-density cells might function as suppressors of signalling pathways that stimulate proliferation and inhibit differentiation. Until now, most of these pathways have been studied in isolation and it is not clear how they interact with each other. Nevertheless, recent work

1.

2.

3.

4.

5.

6.

7. 8.

9.

234

Farquhar, M. G. & Palade, G. E. Junctional complexes in various epithelia. J. Cell Biol. 17, 375–412 (1963). Morphological definition of the epithelial junctional complex. Fleming, T. P., Sheth, B. & Fesenko, I. Cell adhesion in the preimplantation mammalian embryo and its role in trophectoderm differentiation and blastocyst morphogenesis. Front. Biosci. 6, D1000–D1007 (2001). Cereijido, M., Shoshani, L. & Contreras, R. G. Molecular physiology and pathophysiology of tight junctions. I. Biogenesis of tight junctions and epithelial polarity. Am. J. Physiol. Gastrointest. Liver Physiol. 279, G477–G482 (2000). Balda, M. S. & Matter, K. Transmembrane proteins of tight junctions. Semin. Cell Dev. Biol. 11, 281–289 (2000). Anderson, J. M. Molecular structure of tight junctions and their role in epithelial transport. News Physiol. Sci. 16, 126–130 (2001). Tsukita, S., Furuse, M. & Itoh, M. Multifunctional strands in tight junctions. Nature. Rev. Mol. Cell Biol. 2, 286–293 (2001). D’Atri, F. & Citi, S. Molecular complexity of vertebrate tight junctions. Mol. Membr. Biol. 19, 103–112 (2002). Tepass, U., Tanentzapf, G., Ward, R. & Fehon, R. Epithelial cell polarity and cell junctions in Drosophila. Annu. Rev. Genet. 35, 747–784 (2001). Behrens, J., Birchmeier, W., Goodman, S. L. & Imhof, B. A. Dissociation of Madin–Darby canine kidney epithelial cells by the monoclonal antibody anti-Arc-1: mechanistic aspects and identification of the antigen as a component related to uvomorulin. J. Cell Biol. 101, 1307–1315 (1985).

indicates that Crumbs-associated signalling molecules interact functionally and structurally with the ZO-1–ZO-3 complex as well as the PAR3–PAR6–aPKC complex66,68,69, which indicates that different TJ-associated signal-transduction systems might influence each other. Functional TJs assemble when epithelial cells reach high cell densities, which results in increased expression levels of TJ-associated proteins such as ZO-1 (REF. 108). This seems to be related to the sensing and regulation of cell density, because manipulating the expression of ZO-1 and the associated transcription factor ZONAB affects epithelial cell density113. It will therefore be important to determine how premature upregulation of specific TJ proteins in low-density cells affects proliferation and what types of signal-transduction pathway are affected. Furthermore, TJ-associated signalling systems are probably affected by, and influence, other pathways that control proliferation and differentiation, so it will be important to study crosstalk between TJs and wellestablished regulatory systems such as cadherin-based adhesion and signalling from the extracellular matrix. Not only have TJs been linked to the regulation of transcription, but TJ components or factors interacting with TJ proteins have been linked to RNA processing. The best studied example is symplekin, which has been shown to associate with the polyadenylation machinery110,111. The Y-box factor ZONAB, which can bind DNA and RNA (M.S.B. and K.M., unpublished observations), as well as the ZO-2-interacting protein SAF-B are also good candidate factors that are involved in RNA processing. So it is possible that the formation and assembly of TJs modulates several steps of protein biosynthesis and degradation, from the regulation of transcription and of mRNA processing to the regulation of protein turnover by stabilization.

10. Gonzalez-Mariscal, L., Chavez de Ramirez, B. & Cereijido, M. Tight junction formation in cultured epithelial cells (MDCK). J. Membr. Biol. 86, 113–125 (1985). 11. Gumbiner, B. & Simons, K. A functional assay for proteins involved in establishing an epithelial occluding barrier: identification of a uvomorulin-like polypeptide. J. Cell Biol. 102, 457–468 (1986). 12. Balda, M. S. et al. Assembly and sealing of tight junctions: possible participation of G-proteins, phospholipase C, protein kinase C and calmodulin. J. Membr. Biol. 122, 193–202 (1991). 13. Benais-Pont, G., Matter, K. & Balda, M. S. in Tight Junctions (eds Anderson, J. M. & Cereijido, M.) 367–394 (CRC Press, Boca Raton, USA, 2001). 14. Nilsson, M., Fagman, H. & Ericson, L. E. Ca2+dependent and Ca2+-independent regulation of the thyroid epithelial junction complex by protein kinases. Exp. Cell Res. 225, 1–11 (1996). 15. de Almeida, J. B. et al. Targeting of chimeric Gαi proteins to specific membrane domains. J. Cell Sci. 107, 507–515. (1994). 16. Dodane, V. & Kachar, B. Identification of isoforms of G proteins that colocalize with tight junctions. J. Membr. Biol. 149, 199–209 (1996). 17. Denker, B. M., Saha, C., Khawaja, S. & Nigam, S. K. Involvement of a heterotrimeric G protein α subunit in tight junction biogenesis. J. Biol. Chem. 271, 25750–25753 (1996). 18. Meyer, T. N., Schwesinger, C. & Denker, B. M. Zonula occludens-1 is a scaffolding protein for signaling molecules. Gα12 directly binds to the Src homology 3 domain and regulates paracellular permeability in epithelial cells. J. Biol. Chem. 277, 24855–24858 (2002).

| MARCH 2003 | VOLUME 4

19. Mellor, H. & Parker, P. J. The extended protein kinase C superfamily. Biochem. J. 332, 281–292 (1998). 20. Nigam, S. K., Denisenko, N., Rodriguez-Boulan, E. & Citi, S. The role of phosphorylation in development of tight junctions in cultured renal epithelial (MDCK) cells. Biochem. Biophys. Res. Commun. 181, 548–553 (1991). 21. Citi, S., Volberg, T., Bershadsky, A. D., Denisenko, N. & Geiger, B. Cytoskeletal involvement in the modulation of cell–cell junctions by the protein kinase inhibitor H-7. J. Cell Sci. 107, 683–692 (1994). 22. Balda, M. S., González-Mariscal, L., Matter, K., Cereijido, M. & Anderson, J. M. Assembly of tight junctions: the role of diacylglycerol. J. Cell Biol. 123, 293–302 (1993). 23. Tabuse, Y. et al. Atypical protein kinase C cooperates with PAR-3 to establish embryonic polarity in Caenorhabditis elegans. Development 125, 3607–3614 (1998). 24. Izumi, Y. et al. An atypical PKC directly associates and colocalizes at the epithelial tight junction with ASIP, a mammalian homologue of Caenorhabditis elegans polarity protein PAR-3. J. Cell Biol. 143, 95–106 (1998). This paper shows the association of evolutionarily conserved signalling components belonging to the PAR3–PAR6–aPKC complex with TJs. 25. Kemphues, K. Parsing embryonic polarity. Cell 101, 345–348 (2000). 26. Ohno, S. Intercellular junctions and cellular polarity: the PAR–aPKC complex, a conserved core cassette playing fundamental roles in cell polarity. Curr. Opin. Cell Biol. 13, 641–648 (2001).

www.nature.com/reviews/molcellbio

© 2003 Nature Publishing Group

REVIEWS 27. Itoh, M. et al. Junctional adhesion molecule (JAM) binds to PAR-3: a possible mechanism for the recruitment of PAR-3 to tight junctions. J. Cell Biol. 154, 491–497 (2001). 28. Ebnet, K. et al. The cell polarity protein ASIP/PAR-3 directly associates with junctional adhesion molecule (JAM). EMBO J. 20, 3738–3748 (2001). 29. Lin, D. et al. A mammalian PAR-3–PAR-6 complex implicated in Cdc42/Rac1 and aPKC signalling and cell polarity. Nature Cell Biol. 2, 540–547 (2000). Together with reference 30, this paper provides a connection between Cdc42, a central regulator of cell polarity, and activation of aPKC. 30. Joberty, G., Petersen, C., Gao, L. & Macara, I. G. The cell-polarity protein Par6 links Par3 and atypical protein kinase C to Cdc42. Nature Cell Biol. 2, 531–539 (2000). 31. Suzuki, A. et al. Atypical protein kinase C is involved in the evolutionarily conserved par protein complex and plays a critical role in establishing epithelia-specific junctional structures. J. Cell Biol. 152, 1183–1196 (2001). 32. Kroschewski, R., Hall, A. & Mellman, I. Cdc42 controls secretory and endocytic transport to the basolateral plasma membrane of MDCK cells. Nature Cell Biol. 1, 8–13 (1999). 33. Suzuki, A. et al. aPKC kinase activity is required for the asymmetric differentiation of the premature junctional complex during epithelial cell polarization. J. Cell Sci. 115, 3565–3573 (2002). 34. Hirose, T. et al. Involvement of ASIP/PAR-3 in the promotion of epithelial tight junction formation. J. Cell Sci. 115, 2485–2495 (2002). 35. Nunbhakdi-Craig, V. et al. Protein phosphatase 2A associates with and regulates atypical PKC and the epithelial tight junction complex. J. Cell Biol. 158, 967–978 (2002). 36. Fukata, M. & Kaibuchi, K. Rho-family GTPases in cadherin-mediated cell–cell adhesion. Nature Rev. Mol. Cell Biol. 2, 887–897 (2001). 37. Kawakatsu, T. et al. Trans-interactions of nectins induce formation of filopodia and lamellipodia through the respective activation of Cdc42 and Rac small G proteins. J. Biol. Chem. 11, 11 (2002). 38. Zahraoui, A., Louvard, D. & Galli, T. Tight junction, a platform for trafficking and signaling protein complexes. J. Cell Biol. 151, F31–36 (2000). 39. Yeaman, C., Grindstaff, K. K. & Nelson, W. J. New perspectives on mechanisms involved in generating epithelial cell polarity. Physiol. Rev. 79, 73–98 (1999). 40. Marzesco, A. M. et al. The small GTPase Rab13 regulates assembly of functional tight junctions in epithelial cells. Mol. Biol. Cell 13, 1819–1831 (2002). 41. Nusrat, A. et al. Rho protein regulates tight junctions and perijunctional actin organization in polarized epithelia. Proc. Natl Acad. Sci. USA 92, 10629–10633 (1995). 42. Zhong, C., Kinch, M. S. & Burridge, K. Rho-stimulated contractility contributes to the fibroblastic phenotype of Ras-transformed epithelial cells. Mol. Biol. Cell 8, 2329–2344 (1997). 43. Takaishi, K., Sasaki, T., Kotani, H., Nishioka, H. & Takai, Y. Regulation of cell–cell adhesion by rac and rho small G proteins in MDCK cells. J. Cell Biol. 139, 1047–1059 (1997). 44. Hasegawa, H. et al. Opposite regulation of transepithelial electrical resistance and paracellular permeability by Rho in Madin–Darby canine kidney cells. J. Biol. Chem. 274, 20982–20988 (1999). 45. Jou, T. S., Schneeberger, E. E. & Nelson, W. J. Structural and functional regulation of tight junctions by RhoA and Rac1 small GTPases. J. Cell Biol. 142, 101–115 (1998). 46. Wojciak-Stothard, B., Potempa, S., Eichholtz, T. & Ridley, A. J. Rho and Rac but not Cdc42 regulate endothelial cell permeability. J. Cell Sci. 114, 1343–1355 (2001). 47. Braga, V. M., Del Maschio, A., Machesky, L. & Dejana, E. Regulation of cadherin function by Rho and Rac: modulation by junction maturation and cellular context. Mol. Biol. Cell 10, 9–22 (1999). 48. Adamson, P., Etienne, S., Couraud, P. O., Calder, V. & Greenwood, J. Lymphocyte migration through brain endothelial cell monolayers involves signaling through endothelial ICAM-1 via a rho-dependent pathway. J. Immunol. 162, 2964–2973 (1999). 49. Fujita, H. et al. Molecular decipherment of Rho effector pathways regulating tight-junction permeability. Biochem. J. 346, 617–622 (2000).

50. Hirase, T. et al. Regulation of tight junction permeability and occludin phosphorylation by RhoA-p160ROCKdependent and-independent mechanisms. J. Biol. Chem. 276, 10423–10431 (2001). 51. Fanning, A. S. in Tight Junctions (eds Anderson, J. M. & Cereijido, M.) 265–284 (CRC Press, Boca Raton, USA, 2001). 52. Etienne-Manneville, S. & Hall, A. Rho GTPases in cell biology. Nature 420, 629–635 (2002). 53. Hecht, G. et al. Expression of the catalytic domain of myosin light chain kinase increases paracellular permeability. Am. J. Physiol. 271, C1678–C1684 (1996). 54. Turner, J. R. et al. Physiological regulation of epithelial tight junctions is associated with myosin light-chain phosphorylation. Am. J. Physiol. 273, C1378–C1385 (1997). 55. Somlyo, A. P. & Somlyo, A. V. Signal transduction by G-proteins, rho-kinase and protein phosphatase to smooth muscle and non-muscle myosin II. J. Physiol. (Lond.) 522, 177–185 (2000). 56. Benais-Pont, G. et al. Identification of a tight junctionassociated guanine nucleotide exchange factor that activates Rho and regulates paracellular permeability. J. Cell Biol. (in the press). The characterization of a junction-associated GEF that allows TJ-specific activation of Rho is reported here. 57. Tepass, U., Theres, C. & Knust, E. Crumbs encodes an EGF-like protein expressed on apical membranes of Drosophila epithelial cells and required for organization of epithelia. Cell 61, 787–799 (1990). This paper reports the identification of Drosophila Crumbs and its fundamental role in epithelial differentiation. 58. den Hollander, A. I. et al. Mutations in a human homologue of Drosophila crumbs cause retinitis pigmentosa (RP12). Nature Genet. 23, 217–221 (1999). 59. Izaddoost, S., Nam, S. C., Bhat, M. A., Bellen, H. J. & Choi, K. W. Drosophila Crumbs is a positional cue in photoreceptor adherens junctions and rhabdomeres. Nature 416, 178–183 (2002). 60. Pellikka, M. et al. Crumbs, the Drosophila homologue of human CRB1/RP12, is essential for photoreceptor morphogenesis. Nature 416, 143–149 (2002). 61. Wodarz, A., Hinz, U., Engelbert, M. & Knust, E. Expression of crumbs confers apical character on plasma membrane domains of ectodermal epithelia of Drosophila. Cell 82, 67–76 (1995). 62. Grawe, F., Wodarz, A., Lee, B., Knust, E. & Skaer, H. The Drosophila genes crumbs and stardust are involved in the biogenesis of adherens junctions. Development 122, 951–959 (1996). 63. Bachmann, A., Schneider, M., Theilenberg, E., Grawe, F. & Knust, E. Drosophila Stardust is a partner of Crumbs in the control of epithelial cell polarity. Nature 414, 638–643 (2001). 64. Hong, Y., Stronach, B., Perrimon, N., Jan, L. Y. & Jan, Y. N. Drosophila Stardust interacts with Crumbs to control polarity of epithelia but not neuroblasts. Nature 414, 634–638 (2001). 65. Roh, M. H. et al. The Maguk protein, Pals1, functions as an adapter, linking mammalian homologues of Crumbs and Discs Lost. J. Cell Biol. 157, 161–172 (2002). 66. Lemmers, C. et al. hINADl/PATJ, a homolog of Discs lost, interacts with crumbs and localizes to tight junctions in human epithelial cells. J. Biol. Chem. 277, 25408–25415 (2002). References 65 and 66 provide the first evidence for the association of components of the Crumbs pathway with epithelial TJs. 67. Makarova, O., Roh, M. H., Liu, C.-J., Laurinec, S. & Margolis, B. Mammalian Crumbs3 is a small transmembrane protein linked to protein associated with Lin-7 (Pals1). Gene 302, 21–29 (2002). 68. Roh, M. H., Liu, C.-J., Laurinec, S. & Margolis, B. The carboxy-terminus of zona occludens-3 binds and recruits a mammalian homologue of discs lost to tight junctions. J. Biol. Chem. 277, 27501–27509 (2002). 69. Hurd, T. W., Gao, L., Roh, M. H., Macara, I. G. & Margolis, B. Direct interaction of two polarity complexes implicated in epithelial tight junction assembly. Nature Cell Biol. 5, 137–142 (2003). This paper provides a structural and functional link between the Crumbs and the PAR3–PAR6–aPKC signalling pathways in mammalian epithelial cells. 70. Cary, L. A., Han, D. C. & Guan, J. L. Integrin-mediated signal transduction pathways. Histol. Histopathol. 14, 1001–1009 (1999).

NATURE REVIEWS | MOLECUL AR CELL BIOLOGY

71. Potempa, S. & Ridley, A. J. Activation of both MAP kinase and phosphatidylinositide 3-kinase by Ras is required for hepatocyte growth factor/scatter factorinduced adherens junction disassembly. Mol. Biol. Cell 9, 2185–2200 (1998). 72. Behrens, J., Mareel, M. M., Van Roy, F. M. & Birchmeier, W. Dissecting tumor cell invasion: epithelial cells acquire invasive properties after the loss of uvomorulinmediated cell–cell adhesion. J. Cell Biol. 108, 2435–2447 (1989). 73. Marshall, C. How do small GTPase signal transduction pathways regulate cell cycle entry? Curr. Opin. Cell Biol. 11, 732–736 (1999). 74. Zhadanov, A. B. et al. Absence of the tight junctional protein AF-6 disrupts epithelial cell–cell junctions and cell polarity during mouse development. Curr. Biol. 9, 880–888 (1999). 75. Schramek, H., Feifel, E., Healy, E. & Pollack, V. Constitutively active mutant of the mitogen-activated protein kinase kinase MEK1 induces epithelial dedifferentiation and growth inhibition in Madin–Darby canine kidney-C7 cells. J. Biol. Chem. 272, 11426–11433 (1997). 76. Li, D. & Mrsny, R. J. Oncogenic Raf-1 disrupts epithelial tight junctions via downregulation of occludin. J. Cell Biol. 148, 791–800 (2000). 77. Lu, Q., Paredes, M., Zhang, J. & Kosik, K. S. Basal extracellular signal-regulated kinase activity modulates cell–cell and cell–matrix interactions. Mol. Cell Biol. 18, 3257–3265 (1998). 78. Chen, Y., Lu, Q., Schneeberger, E. E. & Goodenough, D. A. Restoration of tight junction structure and barrier function by down-regulation of the mitogen-activated protein kinase pathway in ras-transformed Madin–Darby canine kidney cells. Mol. Biol. Cell 11, 849–862 (2000). 79. Linnemann, T. et al. Thermodynamic and kinetic characterization of the interaction between the Ras binding domain of AF6 and members of the Ras subfamily. J. Biol. Chem. 274, 13556–13562 (1999). 80. Boettner, B., Govek, E. E., Cross, J. & Van Aelst, L. The junctional multidomain protein AF-6 is a binding partner of the Rap1A GTPase and associates with the actin cytoskeletal regulator profilin. Proc. Natl Acad. Sci. USA 97, 9064–9069 (2000). 81. Yamamoto, T. et al. The ras target AF-6 interacts with ZO-1 and serves as a peripheral component of tight junctions in epithelial cells. J. Cell Biol. 139, 785–795 (1997). 82. Mandai, K. et al. Afadin: A novel actin filament-binding protein with one PDZ domain localized at cadherinbased cell-to-cell adherens junction. J. Cell Biol. 139, 517–528 (1997). 83. Takahashi, K. et al. Nectin/PRR: an immunoglobulinlike cell adhesion molecule recruited to cadherin-based adherens junctions through interaction with Afadin, a PDZ domain-containing protein. J. Cell Biol. 145, 539–549 (1999). 84. Ebnet, K., Schulz, C. U., Meyer Zu Brickwedde, M. K., Pendl, G. G. & Vestweber, D. Junctional adhesion molecule interacts with the PDZ domain-containing proteins AF-6 and ZO-1. J. Biol. Chem. 275, 27979–27988 (2000). 85. Ikeda, W. et al. Afadin: A key molecule essential for structural organization of cell–cell junctions of polarized epithelia during embryogenesis. J. Cell Biol. 146, 1117–1132 (1999). 86. Willott, E. et al. The tight junction protein ZO-1 is homologous to the Drosophila discs-large tumor suppressor protein of septate junctions. Proc. Natl Acad. Sci. USA 90, 7834–7838 (1993). 87. Tsukita, S., Itoh, M., Nagafuchi, A., Yonemura, S. & Tsukita, S. Submembranous junctional plaque proteins include potential tumor suppressor molecules. J. Cell Biol. 123, 1049–1053 (1993). This, and reference 86, are the first descriptions of homologies between a TJ-associated protein and a tumour suppressor. 88. Takahisa, M. et al. The Drosophila tamou gene, a component of the activating pathway of extramacrochaetae expression, encodes a protein homologous to mammalian cell–cell junctionassociated protein ZO-1. Genes Dev. 10, 1783–1795 (1996). 89. Chen, C. M. et al. Polychaetoid is required to restrict segregation of sensory organ precursors from proneural clusters in Drosophila. Mech. Dev. 57, 215–227 (1996).

VOLUME 4 | MARCH 2003 | 2 3 5

© 2003 Nature Publishing Group

REVIEWS 90. Takahashi, K., Matsuo, T., Katsube, T., Ueda, R. & Yamamoto, D. Direct binding between two PDZ domain proteins Canoe and ZO-1 and their roles in regulation of the jun N-terminal kinase pathway in Drosophila morphogenesis. Mech. Dev. 78, 97–111 (1998). 91. Reichert, M., Muller, T. & Hunziker, W. The PDZ domains of zonula occludens-1 induce an epithelial to mesenchymal transition of Madin–Darby canine kidney I cells. Evidence for a role of β-catenin/Tcf/Lef signaling. J. Biol. Chem. 275, 9492–9500 (2000). 92. Ryeom, S. W., Paul, D. & Goodenough, D. A. Truncation mutants of the tight junction protein ZO-1 disrupt corneal epithelial cell morphology. Mol. Biol. Cell 11, 1687–1696 (2000). 93. Vietor, I., Bader, T., Paiha, K. & Huber, L. A. Perturbation of the tight junction permeability barrier by occludin loop peptides activates β-catenin/TCF/LEF-mediated transcription. EMBO Rep. 2, 306–312 (2001). 94. Hoover, K. B., Liao, S. Y. & Bryant, P. J. Loss of the tight junction MAGUK ZO-1 in breast cancer: relationship to glandular differentiation and loss of heterozygosity. Am. J. Pathol. 153, 1767–1773 (1998). 95. Chlenski, A. et al. Tight junction protein ZO-2 is differentially expressed in normal pancreatic ducts compared to human pancreatic adenocarcinoma. Int. J. Cancer 82, 137–144 (1999). 96. Kleeff, J. et al. Altered expression and localization of the tight junction protein ZO-1 in primary and metastatic pancreatic cancer. Pancreas 23, 259–265 (2001). 97. Mann, B. et al. Target genes of β-catenin–T cellfactor/lymphoid-enhancer-factor signaling in human colorectal carcinomas. Proc. Natl Acad. Sci. USA 96, 1603–1608 (1999). 98. Gottardi, C. J. & Gumbiner, B. M. Adhesion signaling: how β-catenin interacts with its partners. Curr. Biol. 11, R792–794 (2001). 99. Huelsken, J. & Birchmeier, W. New aspects of Wnt signaling pathways in higher vertebrates. Curr. Opin. Genet. Dev. 11, 547–553 (2001). 100. Glaunsinger, B. A., Weiss, R. S., Lee, S. S. & Javier, R. Link of the unique oncogenic properties of adenovirus type 9 E4-ORF1 to a select interaction with the candidate tumor suppressor protein ZO-2. EMBO J. 20, 5578–5586 (2001). This shows a structural and functional link between an oncogenic viral protein and one of the TJ-associated candidate tumour suppressors. 101. Lee, S. S., Glaunsinger, B., Mantovani, F., Banks, L. & Javier, R. T. Multi-PDZ domain protein MUPP1 is a cellular target for both adenovirus E4-ORF1 and highrisk papillomavirus type 18 E6 oncoproteins. J. Virol. 74, 9680–9693 (2000).

236

102. Glaunsinger, B. A., Lee, S. S., Thomas, M., Banks, L. & Javier, R. Interactions of the PDZ-protein MAGI-1 with adenovirus E4-ORF1 and high-risk papillomavirus E6 oncoproteins. Oncogene 19, 5270–5280 (2000). 103. Traweger, A. et al. The tight junction protein ZO-2 localizes to the nucleus and interacts with the heterogeneous nuclear ribonucleoprotein scaffold attachment factor-B. J. Biol. Chem. 278, 2692–2700 (2003). 104. Wu, X. et al. Evidence for regulation of the PTEN tumor suppressor by a membrane- localized multi-PDZ domain containing scaffold protein MAGI-2. Proc. Natl Acad. Sci. USA 97, 4233–4238 (2000). 105. Wu, Y. et al. Interaction of the tumor suppressor PTEN/MMAC with a PDZ domain of MAGI3, a novel membrane-associated guanylate kinase. J. Biol. Chem. 275, 21477–21485 (2000). 106. Gottardi, C. J., Arpin, M., Fanning, A. S. & Louvard, D. The junction-associated protein, zonula occludens-1, localizes to the nucleus before the maturation and during the remodeling of cell-cell contacts. Proc. Natl Acad. Sci. USA 93, 10779–10784 (1996). 107. Islas, S., Vega, J., Ponce, L. & Gonzalez-Mariscal, L. Nuclear localization of the tight junction protein ZO-2 in epithelial cells. Exp. Cell Res. 274, 138–148 (2002). 108. Balda, M. S. & Matter, K. The tight junction protein ZO-1 and an interacting transcription factor regulate ErbB-2 expression. EMBO J. 19, 2024–2033 (2000). 109. Keon, B. H., Schäfer, S., Kuhn, C., Grund, C. & Franke, W. W. Symplekin, a novel type of tight junction plaque protein. J. Cell Biol. 134, 1003–1018 (1996). 110. Takagaki, Y. & Manley, J. L. Complex protein interactions within the human polyadenylation machinery identify a novel component. Mol. Cell Biol. 20, 1515–1525 (2000). 111. Hofmann, I., Schnolzer, M., Kaufmann, I. & Franke, W. W. Symplekin, a constitutive protein of karyo- and cytoplasmic particles involved in mRNA biogenesis in Xenopus laevis oocytes. Mol. Biol. Cell 13, 1665–1676 (2002). 112. Nakamura, T. et al. huASH1 protein, a putative transcription factor encoded by a human homologue of the Drosophila ash1 gene, localizes to both nuclei and cell–cell tight junctions. Proc. Natl Acad. Sci. USA 97, 7284–7289 (2000). 113. Balda, M. S., Garrett, M. D. & Matter, K. The ZO-1 associated Y-box factor ZONAB regulates epithelial cell proliferation and cell density. J. Cell Biol. 160, 423–432 (2003). This paper provides a mechanism that links TJs to the regulation of proliferation and cell-cycle progression. 114. Gumbiner, B., Lowenkopf, T. & Apatira, D. Identification of a 160-kDa polypeptide that binds to the tight junction protein ZO-1. Proc. Natl Acad. Sci. USA 88, 3460–3464 (1991).

| MARCH 2003 | VOLUME 4

115. Cereijido, M., Robbins, E. S., Dolan, W. J., Rotunno, C. A. & Sabatini, D. D. Polarized monolayers formed by epithelial cells on a permeable and translucent support. J. Cell Biol. 77, 853–880 (1978). This establishes the most widely used model system to study TJ assembly and epithelial cell polarity. 116. Canfield, P. E., Geerdes, A. M. & Molitoris, B. A. Effect of reversible ATP depletion on tight-junction integrity in LLCPK1 cells. Am. J. Physiol. 261, F1038–F1045 (1991). 117. Cramer, E. B., Milks, L. C. & Ojakian, G. K. Transepithelial migration of human neutrophils: an in vitro model system. Proc. Natl Acad. Sci. USA 77, 4069–4073 (1980). 118. Rajasekaran, A. K., Hojo, M., Huima, T. & RodriguezBoulan, E. Catenins and zonula occludens-1 form a complex during early stages in the assembly of tight junctions. J. Cell Biol. 132, 451–463 (1996). 119. Itoh, M., Nagafuchi, A., Moroi, S. & Tsukita, S. Involvement of ZO-1 in cadherin-based cell adhesion through its direct binding to α-catenin and actin filaments. J. Cell Biol. 138, 181–192 (1997). 120. Giepmans, B. N. & Moolenaar, W. H. The gap junction protein connexin43 interacts with the second PDZ domain of the zona occludens-1 protein. Curr. Biol. 8, 931–934 (1998). 121. Toyofuku, T. et al. Direct association of the gap junction protein connexin-43 with ZO-1 in cardiac myocytes. J. Biol. Chem. 273, 12725–12731 (1998). 122. Yokoyama, S. et al. α-catenin-independent recruitment of ZO-1 to nectin-based cell–cell adhesion sites through afadin. Mol. Biol. Cell 12, 1595–1609 (2001).

Acknowledgements The research in the authors’ laboratories is supported by The Wellcome Trust, Cancer Research UK, the Biotechnology and Biological Sciences Research Council, the Medical Research Council and Fight for Sight.

Online links DATABASES The following terms in this article are linked online to: Swiss-Prot: http://www.expasy.ch afadin | ASH1 | Cdc42 | CDK4 | Crumbs | Crumbs 3 | Discs lost | E-cadherin | Elk1 | ErbB2 | GEF-H1 | MAGI-1 | MAGI-2 | MAGI-3 | MEK1 | MUPP1 | occludin | Pals1 | PAR3 | PAR6 | PTEN | SAF-B | Stardust | ZO-1 | ZO-2 | ZO-3 FURTHER INFORMATION Karl Matter’s laboratory: http://www.ucl.ac.uk/ioo/research/matter.htm Maria S. Balda’s laboratory: http://www.ucl.ac.uk/ioo/research/balda.htm Access to this interactive links box is free online.

www.nature.com/reviews/molcellbio

© 2003 Nature Publishing Group