PAR-3/ASIP, PAR-6 AF-6/s-afadin CASK CAROM
2.1 TJ molecules
ZO-1 was the first molecule to be identified in TJs, as a phosphorprotein of 210-225 kDa in size (25-26) localised in the immediate vicinity of the plasma membrane of both endothelial and epithelial cells (22). ZO-1 is concentrated at the TJs although it is also found within the adherens junction, the nucleus and in cells that do not have distinct TJ structure (23). It is phosphorylated on serine residues under normal conditions but becomes phosphorylated on tyrosine residues after stimulatiuon. ZO-1 is a member of the MAGUK protein family (membrane associated and having the presence of a guanylate kinase or GUK domain) of which the members share the common domains of SH3, homologous to GUK, and have one or more PDZ domains (27-28). The PDZ domains mediate a reversible and regulated protein-protein interaction through contact with other PDZ domains or bu recognition of sequence motifs at the C-termini of integral proteins (27). The SH3 domain acts as a proteinprotein interaction molecule binding to the PXXP motif. The UGK domain is also involved in protein to protein interactions.
ZO-2 and ZO-3 are also members of this MAGUK family of 160 kDa and 130 kDa, respectively. ZO-2 is a phosphoprotein present at the TJs of epithelia and endothelia and at the adherens junctions of non-TJ containing cells. ZO-2 and ZO-1 co-precipitate as heterodimers through PDZ-2/ PDZ-2 interactions. Soluble ZO-1, -2 and -3 are found as independent ZO-1, ZO-2 and ZO-3 complexes. The N-terminal domain of ZO-2 directly binds to claudins, occludin, and alpha-catenin, while the C-terminal domain co-localises with actin filaments and interacts with the actin-binding protein 4.1 (28). ZO-3 interacts with both ZO-1 and ZO-2, sharing a high sequence homology with both (27).
ZO proteins also contain some unique motifs not shared by other MAGUK family members, including nuclear localisation and nuclear export signals and a leucine zipper-like sequence. Nuclear ZO-2 directly interacts with the DNA-binding protein SAF-B (scaffold attachment factor-B) (30). ZO-2 associates with SAF-B via its PDZ-1 domain, linking to the C-domain of SAF-B. SFA-B does not associate with ZO-1, supporting the idea that junctional MAGUK's serve non-redundant functions. ZO-3 directly interacts with ZO-1 and the cytoplasmic domain of occludin, but not with ZO-2. Increased nuclear staining of ZO-2 is observed in epithelial cells subjected to environmental stress conditions.
Sequence analysis of ZO-1 and ZO-2 revealed them to be homologous to members of the lethal discs large-1 (Dlg), PSD-95/SAP90 and p55 protein family indicating a role in signal transduction (27, 22). Evidence suggests that ZO-1 may well act as a tumour suppressor in mammals as mutations in the Dlg cause neoplastic overgrowth of imaginal discs in Drosophila (31).
MUPP-1, like ZO-1 may also function as a cross-linker between Claudin-based TJ strands and JAM oligomers in TJs. The difference is in the 13 PDZ domains in tandem repeat within a single MUPP-1 molecule (32). This may indicate that other integral membrane proteins can be recruited to the Claudin-based TJ through MUPP-1. It is interesting to note, that viral oncogene products bind to MUPP-1, and one can speculate that MUPP-1 is involved in the formation of macromolecular complexes beneath the plasma membranes at TJs which may play an important role in the regulation of the growth and/or differentiation of epithelial cells (32). Claudins generally have a valine residue at their COOH termini, suggesting that they strongly attract PDZ-containing proteins, such as ZO-1, -2, and -3. MUPP1 (multi-PDZ domain protein 1) is also a binding partner for claudins at the COOH termini (32). MUPP1 is not well characterised, but is exclusively concentrated at TJs of epithelial cells via its binding to claudins and JAM. It thus may play a role as a multivalent scaffold protein recruiting various proteins to the TJ (32).
The subfamily of MAGUKs termed MAGIs (MAGUKs with inverted domain structure) are also located at TJs. Two of the three known MAGI isoforms, MAGI-1 and MAGI-3 are present in the TJs of cultured epithelial cells (33); indeed, MAGI-1 is expressed in the TJs of all epithelial cell types examined. Human MAGI-1 transcripts are alternatively spliced at three sites, and two forms are expressed only in non-epithelial tissues, mainly the brain, although all are localised at the TJ. The major form expressed in colon cancer epithelial cell cultures contains an extended carboxy terminus encoding potential nuclear targeting signals. MAGI-1, ZO-1 and ZO-2 all col-ocalise in non-polarised epithelial cells, suggesting a pre-assembled structure incorporated into the TJ structure at polarisation.
Zonulin may participate in the physiological regulation of intercellular TJs throughout a wide range of extraintestinal epithelia, as well as vascular endothelium, including the blood-brain barrier (34-35). Such disregulation may contribute to disordered intercellular communication, including inflammation, malignant transformation, and metastasis. Indeed, human brain plasma membrane contains a zonulin protein receptor of 45 kDa, which is a glycoprotein containing multiple sialic acid residues (36). This receptor has a striking similarity to MRP-8, a calcium-binding protein.
Membrane-associated guanykate kinase with inverted orientation (MAGI)-1/brain angiogenesis inhibitor 1-associated protein (BAP1) interacts with many transmembrane proteins, including receptors and channels through these domains (37). MAGI-1/BAP1 is ubiquitously expressed and localised at TJs in epithelial cells and is an isoform of the neurone-specific synaptic scaffolding molecule (S-SCAM), known to interact with NMDA erceptors and neuro-ligins. S-SCAm also interacts with a signalling molecule, a GDP/GTP exchange protein (GEP) that is specific for Rap 1 small G protein, Rap GEP. MAGI-1/BAP1 serves as a scaffolding molecule for Rap GEP at TJs in epithelial cells (37).
CASK (originally identified as a neurexin-interacting protein, a human homologue of Lin-2, (38) is a membrane-associated guanylate kinases of epithelial TJ. CASK is localised along the lateral membranes. Carom has a coiled-coil diamin in the middle region and two src homology 3 domains and a PSD-95/Dlg-A/ZO-1 (PDZ)-binding motif in the C-terminal region (39). Carom binds to the fifth PDZ domain of MAGI-1 and the calmodulin kinase domain of CASK in vitro. MAGI-1 and CASK bind to distinct sequences in the C-terminal region of Carom, but still compete with each other for Carom binding. MDCK cells expressing GFP-Carom revealed that Carom was partially over-lapped by MAGI-1 in MDCK cells which have not yet established mature cell junctions, but became separated from MAGI-1 and co-localised with CASK in polarised cells. Carom was highly resistant to Triton X-100 extractions and recruited CASK to the Triton X-100-insoluble structures. Carom is a binding partner for CASK, which interacts with CASK in polarised epithelial cells, and may link it to the cytoskeleton. CASK also interacts with syndecans, JAM-A, protein 4.1, hDLA, (40). CASK may be important for the proper targeting of junctional components and link them to the cytoskeleton (41). Moreover, CASK has been shown to be translocated to the nucleus and may be involved in the regulation of gene transcription (42).
The first transmembrane TJ protein identified was Occludin, 60-65 (82) kDa (43-44). It bears four transmembrane domains in its N-terminal half, with both the N- and C- termini located in the cytoplasm; the C-terminal (approximately, 150 amino acids) binding to ZO-1 (1, 22). The cytoplasmin domain (domain E) also interacts with both ZO-1 and ZO-2. The topology of occludin predicts two extracellular loops projecting into the paracellular space which interact with loops originating from occludin in the neighbouring cell or unidentified molecules to promote interaction and sealing of the paracellular space (45). The C-terminal of occludin is sufficient to mediate endocytosis, as the C-terminal governs intracellular transport of occludin (46). Occludin is a functional component of the TJ and widely expressed in both endothelial and epithelial cells, but not in cells and tissues without TJs (3).
The extracellular surface of occludin was found to be directly involved in cell-cell adhesion and the ability to confer adhesiveness correlated with the ability to co-localise with ZO-1 (47). The discrepancy in the size of occludin protein is a result of differential serine and threonine phosphorylation. The larger phosphorylated form of occludin is found localised within the TJ whereas the smaller less phosphrylated forms are found in the basolateral membrane and cytosol (45, 48). Thus its phosphorylation is directly related to its function. Small differences in the electrophoretic mobility of occludin were found to be distinct phosphorylated variants with altered membrane localisation, indicating that phosphorylation of occludin in an important step in TJ assembly (48-50). In endothelial cells it has been shown that selective proteolytic cleavage of occludin by metallo-proteinases after inhibition of protein tyrosine phosphatases raises para-cellular permeability (51).
Little is known about occludin kinases. However, a recombinant C-terminal fragment of occludin is a substrate for a kinase in crude extracts of brain. CK2 is a candidate kinase for regulation of occludin phosphorylation in vivo (52). Phosphorylation of serine residues on occludin will increase the formation of the TJs (53). Occludin is a Ca2+ -independent intercellular adhesion molecule that confers adhesiveness in proportion to the level of occludin expressed (24).
An alternatively spliced form of occludin, occludin 1B was identified in MDCK cells and cultured T84 human colon cancer cells (54). There are two gene products, the larger, predominant product corresponded to the canonical occludin (TM4+), whilst the smaller product exhibited a 162bp deletion encoding the entire TM4 and immediate C-terminal flanking region (TM4-) (55). The deleted section corresponded to exon 4, suggesting that TM4- is an alternatively spliced isoform. TM4- was also found in monkey epithelial cells, but not murine or canine. Staining of occludin in Caco-2 cells with a C-terminal occludin antibody revealed weak, discontinuous staining restricted to the periphery of subconfluent islands. A weak band at 58kDa (smaller than the predominant band at 65 kDa) corresponded to the predicted mass after blotting. The authors suggest that the TM4- isoform is upregulated in subconfluent cells, and that it is translated at low levels in specific conditions and may contribute to the regulation of occludin function; i.e., if occludin has no C-terminus, it cannot bind to ZO-1.
Occludin's function in the TJ is poorly defined (56). Suppression of occludin is associated with a decrease in claudin-1 and claudin-7 and an increase in claudin-3 and claudin-4. It is indicated that occludin transduces external (apoptotic cells) and intramembrane (rapid cholesterol depletion) signals via a Rho signalling pathway that, in turn, elicits reorganization of the actin cytoskeleton. Impaired signalling in the absence of occludin may also alter the dynamic behaviour of TJ strands, as reflected by an increase in permeability to large organic cations; the permeability of ion pores formed of claudins, however, is less affected.
Claudins are a family of integral trasnmembrane proteins located in the TJ (22). Claudin-1 and claudin-2 were the first described and are structurally related, are 23 kDa in size and have four membrane-spanning regions, although they share no homology with occludin and the other transmembrane proteins located within the TJ. These are found in cells with and without TJs, but are highly expressed in those that do. There are currently 25 claudins described, which often have distinct tissue-specific distributions, although this is debatable as previously claudins that were thougth only to be expressed in certain cells types have been found expressed-albeit at low levels-in disparate cells and tissues. Claudins are the primary seal-forming elements of the paracellular space when occludin is absent (24). They mediate calcium-independent cell-cell adhesion (22).
The claudin protein structure is predicted to consist of cytoplasmic N- and C-termini, four transmembrane domains, and two extracellular loops via which interactions with claudins on adjacent cells occur (57). These interactions can be homo- or heterotypic. The sealing function of claudins is mediated in part by phosphorylation events on the cytoplasmic C-terminus (58-59). In addition, the cytoplasmic C-terminal domain contains a PDZ-binging motif and thus claudins are able to bind to ZO-1, ZO-2, and ZO-3. Claudins are also able to bind to the other PDZ containing proteins such as PAR3 and PAR6. Claudins are also then, involved in cytoskeletal and cell signalling events via regulation of protein localisation in addition to their adhesive functions (61). In contrast to the other main constituents of the TJ, claudins have become an active area of research attempting to understand carcinogeneis and progression to metastasis as many claudins exhibit altered expression in cancer, which was noted shortly after their discovery (61). Claudins are usually over or under expressed in cancers.
Mutations in claudin genes give rise to a number of human hereditary diseases. Claudin-14 defects suffer autosomal deafness (62); Mutations in claudin-16 (paracellin-1) lead to hypomagnesemia syndrome (63). Claudin-16 was originally thought to be uniquely expressed in kidney tissues, but has been found to be expressed in low levels in normal breast tissues (15). Claudin-1 originally named senescence-associated epithelial membrane protein 1 (SEMP-1) was the first to be described and was found to be expressed in most tissue types (64). Moreover, it was the first TJ protein to be indicated as a tumour suppressor in human mammary epithelial cells (64). Although claudin-5 was originally described as being specifically expressed in endothelial cells, it has subsequently been detected in human epithelial cells also, albeit at low levels (6). It is believed that claudins are the paracellular channels which have selectivity for specific ions and play central roles in the regulation of paracellular permeability; the diversity of claudin expression contributes to the physiological homeostasis in response to a particular tissues requirement (65).
2.1.4 Junctional Adhesion Molecules (JAM)/CTX molecules (CAR)
JAMs are members of the immunoglogulin superfamily of protein and are expressed in most cell types, including epithelial, endothelial cells, leukocytes, and platelets. The members of this family are approximately 40 kDa in size and are located at TJs in a similar distribution to ZO-1 (66). There are four members of the JAM protein family, which have recently been renamed; JAM-A, JAM-B, JAM-C, and JAML (67-68). JAM-A and JAM-C localise to the TJ in epithelia and JAM-B to the lateral membrane (69). JAMs have the structural and sequence conservation features of IgSF molecules with two extracellular Ig-like domains and sites for N-glycosylation (70). They are thus unlike occludin and the claudins in having a single transmembrane domain (71). The extracellular domains of JAM-A, -B and -C contain dimerisation motifs that play a role in their interactions (72). JAMs interact in both homo- and heterotypic fashion, as well as with integrins (73).
JAMs regulate both paracellular permeability and leukocyte transmigration via homphilic interaction (74-75). JAM has been suggested to play an important role in the regulation of TJ assembly in epithelia, and JAM-mediated effects may occur by direct or indirect interactions with occludin (76), as JAM is associated with occludin and not ZO-1 in reassembling the TJ structure. JAM's associate through their extracellular domains with the leukocyte beta2 integrins LFA-1 and Mac-1 as well as with the betal integrin alpha4beta1. All three integrins are involved in the regulation of leukocyte-endothelial cell interactions (77). Through their cytoplasmic domain JAMs directly associate with ZO-1, AF6, MUPP-1, and the cell polarity protein PAR-3. PAR-3 is part of a ternary protein complex containing PAR-3, atypical protein kinase C and PAR-6. This complex is highly conserved throughout evolution. This may suggest a dual function for JAMs; they appear to regulate leukocyte-platelet-endothelial cell interactions in the immune system, as well as TJ formation in epithelial and endothelial cells during the acquisition of cell polarity (77).
JAM-B, or VE-JAM, was originally believed to be a vascular molecule participating in interendothelial junctional complexes (69, 78). JAM-C is highly expressed during embryogenesis, in lymph nodes, stains darkly in endothelial venules, vascular structures in the kidney and in lymphatic vessels in lymphoid organs (69). JAM-B binds in a homotypic manner to JAM-B, but also has a receptor in JAM-C (73, 79) within numerous cell types, including endothelial cells. JAM-B adheres to T cells through heterotypic interactions with JAM-C. The engagement of a4p1 by JAMB is only enabled following prior adhesion of JAM-B with JAM-C (80).
There is a preferential expression of JAM-B mRNA in the endothelium in and around tumours and at sites of inflammation at tumour types such as breast, pulmonary squamous cell, pulmonary adenocarcinoma, prostate adenocarcinoma, and colonic carcinoma (73). JAM-1 is also an adhesion molecule for T-cell lines and some circulating lymphocytes and dendritic cells. JAM-2 and JAM-3 are an interacting pair in the A33/ JAM family of adhesion molecules. JAM is thought to be integrated into the TJ structure via its binding to the PDZ2 domain of ZO-1 (claudins bind to its PDZ1 domain). JAM also recruits PAR-3 (ASIP), a determinant of asymmetric cell division and polarised cell growth to TJs through binding to its COOH terminus (81). AF6, a PDZ domain protein in also an intracellular binding partner of JAM-1 via its C-terminus, which has a classical type II PDZ domain-binding motif (82). JAM also binds to the PDZ domains 2 and 3 of ZO-1.
JAML, a novel MAGI-1-binding protein co-localises with ZO-1 in kidney glomeruli and in intestinal epithelial cells (83). Biochemical in vitro studies revealed that JAML bound to MAGI-1 but not to ZO-2, whereas JAM-A did not bind to MAGI-1. They also found that MAGI-1, AO-1 and occludin were recruited to JAML-based cell contacts. JAML appeared to reduce the permeability of CHO cell monolayers. It is suggested that JAML and MAGI-1 provide an adhesion machinery at TJs, which may regulate the permeability of kidney glomeruli and small intestinal epithelial cells.
The Coxsackie-Adenovirus Receptor (CAR) is a 46 kDa transmembrane protein enabling the attachment of virus via the interaction of the adenovirus finger-knob (84). It is expressed ubiquitously in most benign epithelial tissues and although its role is poorly understood has been suggested to be associated with the TJ structure in normal cells (84) and loss of CAR expression can reduce infectivity. Earlier studies have reported a frequent reduction in CAR expression in highly malignant bladder and prostate tumours (85-88). CAR has been much studied due to its importance as a means of entry to cancer cells regarding adenovirus-based cancer therapies. The expression is reported to be often low in a number of cancer types, including ovarian, colorectal, lung, prostate, head and neck tumours, and breast (89-95). CAR expression may correlate inversely with tumour progression (96).
There is a downregulation of CAR gene expression in invasive transitional cell carcinoma in bladder cancer (97). This low expression may have an impact on developing adenoviral-based gene therapies, and they proposed that loss of CAR expression could decrease rigid cell adhesion, possibly increasing migratory potential. Loss of CAR correlates with invasive bladder cancer.
There is localisation of CAR at cell-cell adhesions in several human cancer cell lines, with disruption of cell-cell contacts increasing adenoviral gene transfer into human cancer cells (98). Moreover, TNF alpha increases CAR expression in HeLa and ovarian cancer cells, but decreases CAR expression in U87MG glioblastoma cells. Dexamethasone downregulates CAR expression in both cell types.
The Raf-MEK-ERK pathway is suggested to be involved in regulating ZO-1 expression at the cell surface (99); ZO-1 is restored after inhibition of MEK. CAR expression in pancreatic and colorectal cancer cell lines is upregulated by inhibition of MEK, accompanied by increased CAR protein at the cell surface (100). They conclude that CAR expression loss in cancer cells is at least in part mediated through the Raf-MEK-ERK signal transduction pathway.
Nectins, Ca(2+)-independent immunoglobulin-like cell adhesion molecules (CAMs), first form cell-cell adhesion where cadherins are recruited, forming adherens junctions in epithelial cells and fibroblasts. In addition, nectins recruit claudins, occludin, and JAMs to the apical side of adherens junctions, forming TJs in epithelial cells. Nectins are associated with these CAMs through peripheral membrane proteins (PMPs), many of which are actin filament-binding proteins. The nectin-1-based cell-cell adhesion is formed and maintained irrespective of the presence and absence of the actin filament-disrupting agents, such as cytochalasin D and latrunculin A (101). In the presence of these agents, only afadin remains at the nectin-1-based cell-cell adhesion sites, whereas E-cadherin and other PMPs at adherens junctions, a-catenin, P-catenin, vinculin, a-actinin, ADIP, and LMO7, are not concentrated there. Claudin-1, occludin, and JAM-A, or the PMPs at TJs, ZO-1, and MAGI-1, are not concentrated there, either. These results indicated that the actin cyto-skeleton is required for the association of the nectin-afadin unit with other CAMs and PMPs at adherens and TJs.
Although nectin was initially thought to be only localised at adherens junctions, recent studies have suggested that a role in the formation or organisation of TJs may be found. Nectin-3 (PRR3) interacts with afadin by interaction of their C-terminal to the PDZ domain of afadin (102). The nectin-afadin system is able to recruit ZO-1 to the nectin-based cell-cell adhesion sites in non-epithelial calls that have no TJs (103).
There is a nectin trans-hetero-interaction network; nectin-3 binds to nectin-1, nectin-2, and PVR (poliovirus receptor); nectin-1 also binds to nectin-4 (104). Nectin-1/nectin-3 and nectin-1/nectin-4 trans-hetero-interactions are mediated through trans- V - V domain interactions, whereas C domains contribute to increase the affinity of the interaction. Nectin-3 and nectin-4 share a common binding region in the nectin-1 V domain: (i) nectin-3 strongly competed with nectin-4 binding, (ii) nectin-3 and nectin-4 binding to nectin-1 was reduced by monoclonal antibodies directed towards the nectin-1 V domain, (iii) the glycoprotein D of HSV-1 that binds to the V domain of nectin-1 reduced nectin-3 and nectin-4 binding.
All four nectin family members have one extracellular region with three Ig-like loops, one transmembrane segment, and one cytoplasmic tail (106). The formation of cis-dimers is necessary for the formation of nectin trans-dimers. The authors noted that the first Ig-like loop of nectin-3 is essential and sufficient for the formation of trans-dimers with nectin-1, but that the second Ig-like loop of nectin-3 was furthermore necessary for its cell-cell adhesion activity.
An increasing number of TJ-associated molecules has been revealed over recent years. They are too numerous to detail fully here and the reader is directed to the numerous reviews available (65, 106-110).
Was this article helpful?
Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.