The AJC (apical junctional complex) of vertebrate epithelial cells orchestrates cell–cell adhesion and tissue barrier function. In addition, it plays a pivotal role in signalling. Several protein components of the AJC, e.g. the cytoplasmic proteins β-catenin, p120-catenin and ZO (Zonula Occludens)-2, can shuttle to the nucleus, where they interact with transcription factors to regulate gene expression and cell proliferation. Other junctional proteins, e.g. angiomotin, α-catenin and cingulin, are believed to act by sequestering either transcription factors, such as YAP (Yes-associated protein), or regulators of small GTPases, such as GEF (guanine-nucleotide-exchange factor)-H1, at junctions. The signalling activities of AJC proteins are triggered by different extracellular and intracellular cues, including cell density, and physiological or pathological activation of developmentally regulated pathways, such as the Wnt pathway. The interplay between junctional protein complexes, the actin cytoskeleton and signalling pathways is of crucial importance in the regulation of gene expression and cell proliferation.
Epithelial proliferation and gene expression must be tightly controlled during development and in adult tissues, and are dysregulated in cancer. In the present chapter, we review recent advances in understanding the role of cell–cell junctions, and specifically the AJC (apical junctional complex), in the regulation of cell proliferation and gene expression in vertebrate epithelial cells.
The AJC comprises the ZO (Zonula Occludens) [TJ (tight junction)] and the zonula adherens [AJ (adherens junction)]. Each of these junctions consists of a multi-molecular complex of transmembrane and cytoplasmic proteins, which are linked to the cytoskeleton. Claudins, occludin, tricellulin and JAM-A (junctional adhesion molecule A), and E-cadherin and nectin are the major transmembrane proteins of TJs and AJs respectively [1,2]. The cytoplasmic domains of TJ membrane proteins interact with scaffolding proteins (ZO-1, ZO-2, ZO-3 and others) and directly or indirectly with polarity complex proteins [PAR-3, PAR-6 and others (PAR is PARtioning defective)]. The cytoplasmic tail of E-cadherin interacts with β-catenin and p120-ctn (p120-catenin), whereas nectin binds to the afadin–ponsin complex. TJs and AJs contain several other proteins which serve as adaptors for signalling molecules, and provide linkages to the actin and microtubule cytoskeletons [1–3]. The cytoskeleton functions as a structural organizer of junctions and polarity, and induces cell shape changes through the development of tensile strength and its regulated assembly.
The functions of the AJC are critical in the development and physiology of vertebrate organisms. The TJ provides epithelial tissues with a semi-permeable barrier to the passage of solutes through the paracellular pathway, and contribute to maintaining apical–basal polarity [2,3]. The neighbouring AJ plays a fundamental role in initiating and maintaining cell–cell adhesion . However, besides these canonical functions, exciting new roles for TJs and AJs in signalling have been discovered in the last two decades. In this chapter, we describe the major proteins and mechanisms that are involved in the control of gene expression and cell proliferation by the AJC. The reader should be aware that the field is considerably more complex than presented here, and is referred to a review  for a more complete introduction to this field.
AJ proteins and nucleo-cytoplasmic signalling
The AJ proteins p120-ctn and β-catenin not only modulate the stability of E-cadherin and its interaction with the actin and microtubule cytoskeletons , but also have crucial signalling functions. Both proteins show a nuclear localization in proliferating cells, and interact with and tune the activity of transcription factors. A third catenin, α-catenin, associates with E-cadherin by binding to β-catenin, and has also been implicated in the control of gene expression and in the regulation of the nucleo-cytoplasmic shuttling of transcription factors (Figure 1).
Studies in Drosophila were the first to show that armadillo, the fly β-catenin homologue, is a downstream target of the Wnt signalling pathway. In the absence of Wnt ligand, β-catenin is bound to cadherin, and the free cytoplasmic β-catenin is rapidly degraded via phosphorylation by a destruction complex, which includes two scaffolding proteins, axin and APC (adenomatous polyposis coli), and two kinases, GSK (glycogen synthase kinase) 3β and CK1 (casein kinase 1) (Figure 1). The phosphorylated β-catenin is ubiquitinated, and then degraded by the proteasome. The interaction between the Wnt ligand and its receptors [frizzled and LRP (low-density-lipoprotein-receptor-related protein)] (Figure 1) inactivates the destruction complex and, consequently, β-catenin is not phosphorylated and accumulates in the cytoplasm. The stable cytoplasmic β-catenin translocates into the nucleus, where it acts as a co-activator of the transcription factor TCF (T-cell factor) to promote the expression of Wnt target genes , which include Myc and cyclin D1 (Figure 1). Wnt target genes in turn promote cell proliferation in developing tissues, and are important for the maintenance of stem cell properties in regenerating adult tissues. Significantly, many types of cancers are associated with mutations that lead to a pathological hyperactivation of the Wnt pathway. Wnt signalling is modulated not only by β-catenin, but also by other catenins. For example, α-catenin can negatively regulate β-catenin signalling, either by sequestering β-catenin in the cytoplasm or by inhibiting the β-catenin interaction with TCF (Figure 1).
p120-ctn and α-catenin regulate nuclear signalling and proliferation through additional Wnt-independent mechanisms. p120-ctn shuttles to the nucleus, although it is not clear whether this depends on its phosphorylation state, like β-catenin, or is regulated by the levels of E-cadherin, which recruits p120-ctn to junctions. In the nucleus, p120-ctn interacts with the repressor KAISO, leading to dissociation of KAISO from DNA, and transcription of KAISO target genes , which also include some Wnt target genes (Figure 1). α-Catenin acts as a tumour suppressor, since its conditional ablation in the epidermis induces hyperproliferation, activation of the Ras-MAPK (mitogen-activated protein kinase) pathway, and activation of NF-κB (nuclear factor κB) and its pro-inflammatory target genes  (Figure 1). Recently, it has been shown that α-catenin controls the nucleo-cytoplasmic shuttling of the transcription factor YAP (Yes-associated protein), an effector of the Hippo pathway [8,9] (see also the section on YAP below).
The role of ZO proteins, symplekin and ZONAB (ZO-1-associated nucleic acid-binding protein) in the control of gene expression and cell proliferation
The ZO proteins (ZO-1, ZO-2 and ZO-3) are essential cytoplasmic components of TJs (Figure 2), since they form scaffolds to anchor TJ membrane proteins, and link them to actin filaments . However, they also play very important roles in signalling and in the control of gene expression and cell proliferation .
Some of the first evidence that ZO proteins were implicated in the control of gene expression was the observation that ZO-1 forms a complex with the transcription factor ZONAB, and that ZO-1 overexpression modulates the ZONAB-dependent transcription of the gene Erb-B2  (Figure 2). ZONAB is localized at TJs in confluent cells, in a complex with CDK4 (cyclin-dependent kinase 4), but it translocates to the nucleus in sparse proliferating cells [10,11] (Figure 2). Nuclear translocation of ZONAB results in increased expression of genes involved in cell proliferation, such as cyclin D1 and PCNA (proliferation cell nuclear antigen) (Figure 2). ZONAB also functionally interacts with symplekin , a RNA-processing factor that shuttles between the TJ and the nucleus, where it interacts with the transcription factor HSF1 (heat-shock factor protein 1) to regulate the polyadenylation of Hsp70 (heat-shock protein 70) mRNA  (Figure 2). The symplekin–ZONAB complex has multiple targets and, in intestinal cells, it inhibits cell differentiation by repressing the transcription factor AML1 (acute myeloid leukaemia 1)  (Figure 2).
ZO-2 is the only ZO protein that has unequivocally been shown to translocate to the nucleus. ZO-2 binds to the transcription factor c-Myc, thus inhibiting the transcription, translation and degradation of cyclin D1, and consequently blocking cell-cycle progression at G0/G1  (Figure 2). In addition, ZO-2 interacts with other transcription factors, including Jun/Fos and the scaffold attachment factor SAF-B, with as yet unknown effects [16,17] (Figure 2).
Unlike ZO-1 and ZO-2, the third ZO protein, ZO-3, appears to promote cell proliferation, since it recruits cyclin D1 to TJs during mitosis in intestinal cells, thus preventing its degradation, and thus promoting cell-cycle progression through S-phase . The junctional recruitment of cyclin D1 requires a conserved sequence motif, which mediates interaction with PDZ [PSD-95 (postsynaptic density 95), Dlg (discs large) and ZO-1] domains . PDZ domains are evolutionarily conserved protein modules, which mediate protein–protein interactions, and are found not only in ZO proteins, but in several other scaffolding proteins of TJs. Recruitment of transcription factors and signalling molecules to junctions through PDZ proteins might therefore be a general mechanism to control their subcellular localization and stability.
YAP, a regulator of cell growth, is controlled by junctional proteins
In mammals, cell and organ growth and size are regulated by organ-extrinsic factors, such as nutrition, growth factors and hormones, and by organ-intrinsic mechanisms. The Hippo pathway, which was originally described in Drosophila, is an organ-intrinsic mechanism, which acts by controlling two opposite processes: cell proliferation and apoptosis (see also Chapter 9 in this volume). The Hippo pathway consists of a cascade of kinases, which ultimately phosphorylate downstream effectors, such as the transcription factors YAP and its paralogue TAZ (transcriptional coactivator with PDZ-binding motif). When YAP and TAZ are not phosphorylated, they migrate into the nucleus, interact with the transcription factor TEAD, and promote the expression of proliferation-associated genes. When they are phosphorylated by the Hippo pathway kinases MST [mammalian Ste20 (sterile 20)-like kinase] 1/2 and LATS (large tumour suppressor), YAP and TAZ remain in the cytoplasm and either interact with proteins such as 14-3-3 (Figure 1) or are degraded (see also Chapter 9 in this volume).
Very importantly, experiments by different laboratories have highlighted a role for junctional proteins as upstream regulators of YAP localization. First, AMOT (angiomotin), a protein localized in the cytoplasmic domain of TJs, modulates the Hippo signalling pathway by interacting with YAP and its related protein TAZ [19,20] (Figure 2). AMOT recruits YAP to TJs, and stabilizes the phosphorylated form of YAP, thus inhibiting its nuclear translocation (Figure 2). Interestingly, the YAP paralogue TAZ interacts with the basal polarity complex protein Scribble to regulate cancer stem-cell related traits .
Another cytoplasmic TJ protein, ZO-2, interacts with YAP, and has been implicated in driving YAP to the nucleus, independently of phosphorylation  (Figure 2). Moreover, the AJ protein α-catenin is a major regulator of YAP localization and activity in the skin. In basal keratinocytes and stem cells of hair follicles α-catenin interacts with YAP through the 14-3-3 protein (Figure 1), and the loss of α-catenin leads to a nuclear re-localization of YAP, together with an increase in cell proliferation and tumour progression [8,9]. α-Catenin helps to maintain the phosphorylated state of YAP by preventing its association with the phosphatase PP2A (protein phosphatase 2A)  (Figure 1). Finally, experiments on cultured cancer cells reveal that Hippo signalling pathway components are required for E-cadherin-dependent contact inhibition of proliferation, and that expression of E-cadherin in breast cancer cells restores the density-dependent regulation of YAP nuclear exclusion . In summary both TJ and AJ protein complexes, depending on the cellular context and experimental approach, have been shown to play a crucial role in controlling the subcellular localization and activity of YAP, through both Hippo pathway-dependent and -independent mechanisms.
RhoA-dependent control of gene expression: interplay with junctional proteins?
The notion that actin dynamics affects the activity of transcription factors, and hence gene expression, was first raised by experiments showing that the polymerization state of actin controls the activity of SRF (serum response factor) . Actin polymerization and dynamics is regulated at multiple levels, and a major role is played by Rho family GTPases: RhoA, Rac1 and Cdc42 (cell division cycle 42). Indeed, SRF activity and subcellular localization was eventually shown to depend on RhoA activity. Intriguingly, Rho GTPases are regulated by several junctional proteins, including p120-ctn, E-cadherin, afadin, ZO-3, cingulin, paracingulin, AMOT and PAR-6, through different mechanisms, including binding to activators [GEFs (guanine-nucleotide-exchange factors)] and inhibitors [GAPs (GTPase-activating proteins)], and post-translational modifications . This raises the possibility that the junction-dependent regulation of Rho family GTPases may be an additional mechanism through which junctional proteins control gene expression and cell proliferation.
This hypothesis has been validated by studies on cingulin, a cytoplasmic TJ protein that recruits and inactivates the RhoA activator GEF-H1 at TJs, thus resulting in a down-regulation of RhoA activity in confluent monolayers  (Figure 2). In different types of cultured cell models, knockout or depletion of cingulin affects gene expression, for example the expression of claudin-2, and cell proliferation [27,28]. The proliferation phenotype and the up-regulation of claudin-2 expression in cingulin-depleted kidney cells are rescued by inhibition of RhoA, indicating that some of the effects of cingulin depletion on gene expression are due to its GEF-H1-dependent regulation of RhoA activity  (Figure 2). Interestingly, ZONAB activity is also regulated by GEF-H1, indicating that GEF-H1 and ZONAB form a signalling module that mediates Rho-regulated cyclin D1 promoter activation and expression  (Figure 2).
It should be noted that the relationship between junctions and Rho family GTPases is dual. Not only do AJC proteins control Rho GTPases, but the AJC is itself a major target of Rho family GTPases, which are crucially required for its formation and maintenance, through remodelling of the actin cytoskeleton, and activation of the PAR-3–PAR-6–aPKC (atypical protein kinase C) polarity complex. Furthermore, extrinsic cues act on junctions through Rho GTPases. For example, TGFβ (transforming growth factor β)-dependent EMT (epithelial–mesenchymal transition) leads to the disruption of junctions through the PAR-6-dependent ubiquitination and degradation of RhoA .
Finally, another example illustrating the multiple routes through which Rho GTPases control cell physiology was the recent demonstration that RhoA/ROCK (Rho-associated kinase)-dependent cytoskeletal tension and stress fibre formation control the nucleo-cytoplasmic shuttling of the transcription factor YAP, independently of the Hippo pathway . Since YAP and TAZ also interact with and are regulated by junctional and polarity proteins (see above), these observations suggest the existence of a complex intriguing interplay between junctions, the actin cytoskeleton and transcriptional regulation.
The cytoplasmic domain of TJs and AJs is a hot spot for signalling, since it provides a platform to recruit transcription factors and other essential signalling molecules, such as regulators of GTPases. Moreover, several junctional proteins can shuttle to the nucleus, and bind to and tune the activity of transcription factors. The signalling functions of some junctional proteins have been validated by studies on animal-knockout models and diseases, demonstrating that the AJC is essential to control gene expression and cell proliferation in epithelial cells at the organism level.
How do junctions regulate gene expression and cell proliferation? A likely scenario, on the basis of the information currently available, is that when differentiated epithelial cells form stable junctions within confluent monolayers, cell proliferation is down-regulated through the sequestration of signalling molecules at junctions. Conversely, disruption of junctions following intrinsic or extrinsic stimuli, for example the activation of the Wnt pathway, leads to a redistribution of junctional proteins, and transcriptional activation of proliferation-related genes. Phosphorylation appears to be a general mechanism that promotes the cytoplasmic retention of junction-associated signalling proteins. Since the configuration and regulation of signalling pathways may depend on cell type and differentiated/pathological cell properties, such a scenario may not be always applicable, underlining the importance of hypothesis validation in different cell culture and animal models, whenever possible. Moreover, how the organization of the actin cytoskeleton controls the localization of transcription factors remains an interesting question to be investigated.
In summary, the AJC responds to a variety of extracellular and intracellular signalling cues through many of its protein components, and functions to integrate these signals into co-ordinated changes in cell shape, adhesion and proliferation.
• The AJC (TJs and AJs) has canonical roles in cell adhesion and barrier function of epithelia, but also plays a major role in signalling.
• Several TJ and AJ proteins can shuttle between the cytoplasm and the nucleus, depending on cell density, culture conditions and different extrinsic or intrinsic stimuli.
• TJ and AJ proteins can bind to and sequester transcription factors and signalling proteins at cell–cell junctions.
• The subcellular localization of the transcription factors ZONAB, YAP and other nuclear proteins correlates with cell density, and is regulated by junctional and cytoskeletal proteins.
• Transcription factors that shuttle between the AJC and the nucleus, such as ZONAB and YAP, are critical modulators of cell proliferation, cell differentiation and growth.
• The small GTPase RhoA regulates cell proliferation and gene expression by controlling the activity and nuclear localization of the transcription factors SRF, ZONAB and YAP.
• The ability of AJC proteins, such as cingulin, to control the activity of Rho family GTPases is one of the mechanisms through which the AJC regulates gene expression and cell proliferation.
We thank the Swiss National Foundation, the State of Geneva and the Section of Biology of the Faculty of Sciences of the University of Geneva for support, and Serge Paschoud for comments on the chapter. We apologize for not citing a large number of relevant publications, due to a limitation in the number of references.
- © The Authors Journal compilation © 2012 Biochemical Society