Cingulin contains globular and coiled-coil domains and interacts with ZO-1, ZO-2, ZO-3, and myosin

We characterized the sequence and protein interactions of cingulin, an M(r) 140-160-kD phosphoprotein localized on the cytoplasmic surface of epithelial tight junctions (TJ). The derived amino acid sequence of a full-length Xenopus laevis cingulin cDNA shows globular head (residues 1-439) and tail (1,326-1,368) domains and a central alpha-helical rod domain (440-1,325). Sequence analysis, electron microscopy, and pull-down assays indicate that the cingulin rod is responsible for the formation of coiled-coil parallel dimers, which can further aggregate through intermolecular interactions. Pull-down assays from epithelial, insect cell, and reticulocyte lysates show that an NH(2)-terminal fragment of cingulin (1-378) interacts in vitro with ZO-1 (K(d) approximately 5 nM), ZO-2, ZO-3, myosin, and AF-6, but not with symplekin, and a COOH-terminal fragment (377-1,368) interacts with myosin and ZO-3. ZO-1 and ZO-2 immunoprecipitates contain cingulin, suggesting in vivo interactions. Full-length cingulin, but not NH(2)-terminal and COOH-terminal fragments, colocalizes with endogenous cingulin in transfected MDCK cells, indicating that sequences within both head and rod domains are required for TJ localization. We propose that cingulin is a functionally important component of TJ, linking the submembrane plaque domain of TJ to the actomyosin cytoskeleton.     

Evidence for a functional interaction between cingulin and ZO-1 in cultured cells

Cingulin, a protein component of the submembrane plaque of tight junctions (TJ), contains globular and coiled-coil domains and interacts in vitro with several TJ and cytoskeletal proteins, including the PDZ protein ZO-1. Overexpression of Xenopus cingulin in transfected Xenopus A6 cells resulted in the disruption of endogenous ZO-1 localization, suggesting that cingulin functionally interacts with ZO-1. Glutathione S-transferase pull-down experiments showed that a conserved ZO-1 interaction motif (ZIM) at the NH(2) terminus of cingulin is required for cingulin-ZO-1 interaction in vitro. An NH(2)-terminal region of cingulin, containing the ZIM, was sufficient, when fused to coiled-coil sequences, to target transfected cingulin to junctions. However, deletion of the ZIM did not abolish junctional localization of transfected cingulin in A6 cells, suggesting that cingulin can be recruited to TJ through multiple protein interactions. Interestingly, the ZIM was required for cingulin recruitment into ZO-1-containing adherens junctions of Rat-1 fibroblasts, indicating that cingulin junctional recruitment does not require the molecular context of TJ. Cingulin coiled-coil sequences enhanced the junctional accumulation of expressed cingulin head region in A6 cells, but purified recombinant cingulin did not form filaments under physiological conditions in vitro, suggesting that the cingulin coiled-coil domain acts primarily by promoting dimerization.     

Laminins and other strange proteins.

Laminins are large multidomain proteins of the extracellular matrix (ECM) with important functions in the development and maintenance of cellular organization and supramolecular structure, in particular in basement membranes. Each molecule is composed of three polypeptide chains, A (300-400 kDa) and B1 and B2 (180-200 kDa), which together form the characteristic cross-shaped laminin structure with three short arms and one long arm. Many different domains have been identified in laminin by sequence analysis, structural investigations, and functional studies. Each short arm is formed by homologous N-terminal portions of one of the three chains. Structurally, each short arm contains two or three globular domains which are connected by rows of manyfold-repeated Cys-rich "EGF-like" domains. In all three chains this region is followed by a long heptad repeat region similar to those found in many alpha-helical coiled-coil proteins. These parts of the three laminin chains constitute a triple-stranded coiled-coil domain, which forms the extended rodlike structure of the long arm. This is the only domain in the protein which is made up of more than one chain and consequently serves the function of chain assembly. The two B chains are terminated by the coiled-coil domain, but the A chain contains an additional C-terminal segment which accounts for five globular domains located at the tip of the long arm. Several important functions of laminin have been assigned to individual domains in either the short arms or terminal regions of the long arm.(ABSTRACT TRUNCATED AT 250 WORDS)     

Hendecad repeat in segment 2A and linker L2 of intermediate filament chains implies the possibility of a right-handed coiled-coil structure

The conformation adopted by intermediate filament chains (IF) has been described in terms of a central rod domain with four, alpha-helical, left-handed coiled-coil segments (1A, 1B, 2A, and 2B) joined by linkers (L1, L12, and L2, respectively). The rod domain is terminated at its N- and C-terminal ends by "globular" head and tail domains, respectively. This analysis, initially undertaken about 20-25 years ago, was based on the recognition of an underlying heptad substructure in the sequence of the rod domain, the presence of which can be directly associated with an alpha-helical coiled-coil structure. In this work, a hendecad sequence motif that is closely related to the heptad repeat but which is nonetheless significantly different from it has been recognized in the primary structure of segments 2A and linker L2. This motif, which is 11 residues long and structurally equivalent to a true heptad plus another heptad with an inclusive stutter, is consistent with the chains adopting a continuous right-handed coiled-coil structure with a long-period pitch length. It is therefore predicted that segment 2 as a whole may have a coiled-coil conformation with both right-handed (2A+L2) and left-handed (2B) regions. The changeover in handedness would be expected to occur at the C-terminal end of linker L2 and N-terminal end of segment 2B.     

Sequence and localization of human NASP: conservation of a Xenopus histone-binding protein.

In this study the sequence and localization of human testicular NASP (nuclear autoantigenic sperm protein) are reported. NASP cDNA contains 2561 nt encoding a protein of 787 amino acids. The open reading frame contains 2446 nt followed by an ochre stop codon (TAA) and 104 nucleotides of untranslated sequence containing a poly(A) addition signal 10 bases upstream of the poly(A) tail. Northern blot analysis of human testis poly(A) mRNA indicates a message of approximately 3.2 kb. Multiple sequence alignment (MSA) analysis of the encoded human NASP amino acid sequence with the sequence for the Xenopus histone-binding protein N1/N2 and the rabbit NASP amino acid sequence demonstrates that the human sequence and the Xenopus sequence have extensive amino acid homology upstream of the rabbit initiation codon. Significantly, there is an 85% identity between the human and the rabbit NASP sequences when the alignment starts at the N-terminal of the rabbit sequence and at amino acid 101 of the human sequence. The nuclear translocation signal found in N1/N2 and rabbit NASP is completely conserved in human NASP. The first histone-binding domain of Xenopus is 70% identical and 90% similar to the human NASP domain. The second histone-binding domain of Xenopus is 48% identical and 71% similar to the human NASP domain. MSA analysis of the three sequences generated an unrooted ancestral tree with two branches, indicating that fewer amino acid changes have occurred between the Xenopus and the human sequences than between the Xenopus and the rabbit sequences. In the human testis, NASP is localized predominantly in primary spermatocytes and round spermatids. Spermatogonia, Sertoli cells, Leydig cells, peritubular cells, and other somatic cells do not stain. Human spermatozoa contain NASP in the acrosomal region. Following the acrosome reaction, some NASP remains in the equatorial and postacrosomal regions. We propose that mammalian testes and sperm contain a histone-binding protein which may play a role in regulating the early events of spermatogenesis.     

Developmentally regulated cytokeratin gene in Xenopus laevis.

We have determined the sequence of cloned cDNAs derived from a 1,665-nucleotide mRNA which transiently accumulates during Xenopus laevis embryogenesis. Computer analysis of the deduced amino acid sequence revealed that this mRNA encodes a 47-kilodalton type I intermediate filament subunit, i.e., a cytokeratin. As is common to all intermediate filament subunits so far examined, the predicted polypeptide, named XK70, contains N- and C-terminal domains flanking a central alpha-helical rod domain. The overall amino acid homology between XK70 and a human 50-kilodalton type I keratin is 47%; homology within the alpha-helical domain is 57%. The N-terminal domain, which is not completely contained in our cDNAs, is basic, contains 42% serine plus alanine, and includes five copies of a six-amino-acid repeating unit. The C-terminal domain has a high alpha-helical content and contains a region with sequence homology to the C-terminal domains of other type I and type III intermediate filament proteins. We suggest that different keratin filament subtypes may have different functional roles during amphibian oogenesis and embryogenesis.     

Sequence comparisons of intermediate filament chains: evidence of a unique functional/structural role for coiled-coil segment 1A and linker L1

A comprehensive analysis of the sequences of all types of intermediate filament chains has been undertaken with a particular emphasis on those of segment 1A and linker L1. This has been done to assess whether structural characteristics can be recognized in the sequences that would be consistent with the role of each region in the recently proposed "swinging head" hypothesis. The analyses show that linker L1 is the most flexible rod domain region, that it is the most elongated structure (on a per residue basis), and that it is the most variable region as regards sequence and length. Segment 1A has one of the two most highly conserved regions of sequence in the rod domain (the other being at the end of segment 2B), with seven particular residues conserved across all chain types. It also contains one of the very few potential interchain ionic interactions that could be conserved across all chain types. However, the aggregation of chains in segment 1A is specified less precisely overall by interchain ionic interactions than are the other coiled-coil segments. The apolar residue contents in positions a and d of the heptad substructure are the highest of any coiled-coil segment in the intermediate filament family. Segment 1A also displays an amino acid composition atypical of not only coiled-coil segments 1B and 2B, but indeed of two-stranded coiled coils in general. Nonetheless, molecular modeling based on the crystal structure of the monomeric 1A fragment from human vimentin shows that coiled-coil formation is plausible. The most extensive regions of apolar/aromatic residues lie at the C-terminal end of segment 2B in the helix termination motif and in segment 1A in and close to the helix initiation motif. The predicted stability of the individual alpha-helices in segment 1A is greater than in those comprising segments 1B and 2B, though potential intrachain ionic interactions are either lacking or are minimal in number. Analysis of the 1A sequence and those regions immediately N- and C-terminal to it has shown that the capping residues are near optimal close to the previously predicted ends, thus adding to the likely stability of the alpha-helical structure. However, a second terminating sequence is predicted in 1A (about 10 residues back from the C-terminus). This allows the possibility of some unwinding of the alpha-helical structure of 1A immediately adjacent to linker L1 when the head domains no longer stabilize the coiled-coil structure. All of these data are consistent with the concept of a flexible hinge at L1 and with the ability of the two alpha-helical coiled-coil strands to separate under appropriate conditions and partly unwind at their C-terminal ends to allow the head domains a greater degree of mobility, thus facilitating function.     

Zonula occludens-1 function in the assembly of tight junctions in Madin-Darby canine kidney epithelial cells

Zonula occludens (ZO)-1 was the first tight junction protein to be cloned and has been implicated as an important scaffold protein. It contains multiple domains that bind a diverse set of junction proteins. However, the molecular functions of ZO-1 and related proteins such as ZO-2 and ZO-3 have remained unclear. We now show that gene silencing of ZO-1 causes a delay of approximately 3 h in tight junction formation in Madin-Darby canine kidney (MDCK) epithelial cells, but mature junctions seem functionally normal even in the continuing absence of ZO-1. Depletion of ZO-2, cingulin, or occludin, proteins that can interact with ZO-1, had no discernible effects on tight junctions. Rescue of junction assembly using murine ZO-1 mutants demonstrated that the ZO-1 C terminus is neither necessary nor sufficient for normal assembly. Moreover, mutation of the PDZ1 domain did not block rescue. However, point mutations in the Src homology 3 (SH3) domain almost completely prevented rescue. Surprisingly, the isolated SH3 domain of ZO-1 could also rescue junction assembly. These data reveal an unexpected function for the SH3 domain of ZO-1 in regulating tight junction assembly in epithelial cells and show that cingulin, occludin, or ZO-2 are not limiting for junction assembly in MDCK monolayers.     

Xenopus laevis occludin. Identification of in vitro phosphorylation sites by protein kinase CK2 and association with cingulin.

Occludin is a protein component of the membrane domain of tight junctions, and has been shown to be phosphorylated in vivo in cultured cells and Xenopus laevis embryos. However, nothing is known about the identity of specific occludin kinase(s) and occludin phosphorylation site(s). Furthermore, nothing is known about the interaction of occludin with cingulin, a cytoplasmic plaque component of tight junctions. Here we report the isolation and sequencing of a complete X. laevis occludin cDNA, and experiments aimed at mapping X. laevis occludin in vitro phosphorylation site(s) and characterizing occludin interaction with cingulin. The sequence of Xenopus occludin is homologous to that of occludins from other species, with identities ranging from 41% to 58%. Bacterially expressed domain E of Xenopus occludin (amino acids 247-493) was a good substrate for protein kinase CK2 (stoichiometry 10.8%, Km 8.4 microM) but not for CK1 kinase, protein kinase A, cdc2 kinase, MAP kinase or syk kinase. Residues Thr375 and Ser379 were identified as potential CK2 phosphorylation sites in this region based on sequence analysis. Mutation of Ser379 to aspartic acid or alanine reduced phosphorylation by CK2 by approximately 50%, and double mutation of Ser379 into aspartic acid and Thr375 into aspartic acid essentially abolished phosphorylation. Glutathione S-transferase (GST) pull-down experiments using extracts of Xenopus A6 epithelial cells showed that constructs of GST fused to wild-type and mutant forms of the C-terminal region of X. laevis occludin associate with several polypeptides, and immunoblot analysis showed that one of these polypeptides is cingulin. GST pull-down experiments using in vitro translated, full-length Xenopus cingulin indicated that cingulin interacts directly with the C-terminal region of occludin.     

Conformational characteristics of the complete sequence of group A streptococcal M6 protein

M protein is considered a virulence determinant on the streptococcal cell wall by virtue of its ability to allow the organism to resist attack by human neutrophils. The complete DNA sequence of the M6 gene from streptococcal strain D471 has allowed, for the first time, the study of the structural characteristics of the amino acid sequence of an entire M protein molecule. Predictive secondary structural analysis revealed that the majority of this fibrillar molecule exhibits strong alpha-helical potential and that, except for the ends, nonpolar residues in the central region of the molecule exhibit the 7-residue periodicity typical for coiled-coil proteins. Differences in this heptad pattern of nonpolar residues allow this central rod region to be divided into three subdomains which correlate essentially with the repeat regions A, B, and C/D in the M6 protein sequence. Alignment of the N-terminal half of the M6 sequence with PepM5, the N-terminal half of the M5 protein, revealed that 42% of the amino acids were identical. The majority of the identities were "core" nonpolar residues of the heptad periodicity which are necessary for the maintenance of the coiled coil. Thus, conservation of structure in a sequence-variable region of these molecules may be biologically significant. Results suggest that serologically different M proteins may be built according to a basic scheme: an extended central coiled-coil rod domain (which may vary in size among strains) flanked by functional end domains.     

Distinct domains of paracingulin are involved in its targeting to the actin cytoskeleton and regulation of apical junction assembly

Paracingulin is an M(r) 150-160 kDa cytoplasmic protein of vertebrate epithelial tight and adherens junctions and comprises globular head, coiled-coil rod, and globular tail domains. Unlike its homologous tight junction protein cingulin, paracingulin has been implicated in the control of junction assembly and has been localized at extrajunctional sites in association with actin filaments. Here we analyze the role of paracingulin domains, and specific regions within the head and rod domains, in the function and localization of paracingulin by inducible overexpression of exogenous proteins in epithelial Madin Darby canine kidney (MDCK) cells and by expression of mutated and chimeric constructs in Rat1 fibroblasts and MDCK cells. The overexpression of the rod + tail domains of paracingulin perturbs the development of the tight junction barrier and Rac1 activation during junction assembly by the calcium switch, indicating that regulation of junction assembly by paracingulin is mediated by these domains. Conversely, only constructs containing the head domain target to junctions in MDCK cells and Rat1 fibroblasts. Furthermore, expression of chimeric cingulin and paracingulin constructs in Rat1 fibroblasts and MDCK cells identifies specific sequences within the head and rod domains of paracingulin as critical for targeting to actin filaments and regulation of junction assembly, respectively. In summary, we characterize the functionally important domains of paracingulin that distinguish it from cingulin.     

In vivo interaction of AF-6 with activated Ras and ZO-1.

AF-6 contains two putative Ras-associating domains (RA domains) which are seen in several Ras effectors such as RalGDS and RIN1. We previously showed that an AF-6 fragment containing the amino-terminal (N-terminal) RA domain directly binds to activated Ras and ZO-1 in vitro. In this study, we showed that a single amino acid mutation in the N-terminal RA domain of AF-6 abolished the interaction of AF-6 with activated Ras and that the sites of this critical amino acid residue were similar to those for Raf-1 and RalGDS. The overexpression of the N-terminal RA domain of AF-6 inhibited the Ras-dependent c-fos promoter/enhancer stimulation in NIH3T3 cells. Endogenous AF-6 was coimmunoprecipitated with activated Ras from Rat1 cells expressing activated Ras. Moreover, we showed that AF-6 was coimmunoprecipitated with ZO-1 from Rat1 cells. Taken together, these results indicate that the Ras-interacting region on AF-6 is structurally similar to that on Raf-1 and on RalGDS and that AF-6 interacts with activated Ras and ZO-1 in vivo.     

The structure and organization of the human heavy neurofilament subunit (NF-H) and the gene encoding it.

Genomic clones for the largest human neurofilament protein (NF-H) were isolated, the intron/exon boundaries mapped and the entire protein-coding regions (exons) sequenced. The predicted protein contains a central region that obeys the structural criteria identified for alpha-helical 'rod' domains typically present in all IF protein components: it is approximately 310 amino acids long, shares amino acid sequence homology with other IF protein rod domains and displays the characteristic heptad repeats of apolar amino acids which facilitate coiled-coil interaction. Nevertheless, anomalies are noted in the structure of the NF-H rod which could explain observations of its poor homopolymeric assembly in vitro. The protein segment on the carboxy-terminal side of the human NF-H rod is uniquely long (greater than 600 amino acids) compared to other IF proteins and is highly charged (greater than 24% Glu, greater than 25% Lys), rich in proline (greater than 12%) and impoverished in cysteine, methionine and aromatic amino acids. Its most remarkable feature is a repetitive sequence that covers more than half its length and includes the sequence motif, Lys-Ser-Pro (KSP) greater than 40 times. Together with the recent identification of the serine in KSP as the main target for NF-directed protein kinases in vivo, this repetitive character explains the massive phosphorylation of the NF-H subunit that can occur in axons. The human NF-H gene has three introns, two of which interrupt the protein-coding sequence at identical points to introns in the genes for the two smaller NF proteins, NF-M and NF-L.(ABSTRACT TRUNCATED AT 250 WORDS)     

The laminin B2 chain has a multidomain structure homologous to the B1 chain

Laminin (Mr = 850,000) is a large basement membrane-specific glycoprotein composed of three chains: A, B1, and B2. Previously, we have reported the primary structure of the B1 chain of mouse laminin deduced from sequencing cDNA clones (Sasaki M., Kato, S., Kohno, K., Martin, G. R., and Yamada, Y. (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 935-939). Here we report the isolation of overlapping cDNA clones spanning 7642 bases which encode the entire B2 chain. The nucleotide sequence of the clones contains an open reading frame of 4821 bases coding for a protein of 1607 amino acids including 33 amino acids of a presumptive signal peptide. The mRNA for the B2 chain contains 2.5 kilobases of 3'-untranslated region. The deduced amino acid sequence indicates that the B2 chain consists of six distinct domains, including two domains with alpha-helical, coiled-coil structures, two domains with cysteine-rich homologous repeats, and two globular domains. These structural features of the B2 chain are similar to those of the B1 chain. In addition, the amino acid sequences of the B2 and B1 chains demonstrate considerable homology, suggesting that the genes for these two chains arose from a common ancestor.     

Drosophila laminin A chain sequence, interspecies comparison, and domain structure of a major carboxyl portion.

Recent studies ascribed some biological actions of cell adhesion and cell outgrowth to the carboxyl-most 1200 amino acids of vertebrate laminin A chains. Here we report a 6.1-kilobase pair nucleotide cDNA sequence encoding 1951 amino acids and the carboxyl end of a Drosophila laminin A chain. It corresponds to the mouse laminin A domains G, I, II, and III, but may represent a different type of laminin A chain. The arrangement of the cysteine-rich repeats of domain III resembles that of B2 chains. However, it has more amino acid identity with a portion of the mouse laminin A chain domain IIIb than with other laminin repeats. Domains I and II are consistent with an interrupted coiled-coil alpha-helical model of the long arm of laminin but are poorly conserved. The G domain contains five subdomains which are individually related to subdomains of vertebrate laminin A chains. The results indicate that laminin G subdomains should be considered individually, rather than merely as parts of a G-globule. A sequence of hydroxyamino acids contributes to a spacer between two of the subdomains. Stretches of hydroxyamino acids may be indicative of junctions between domains of extracellular Drosophila proteins.     

Proteinchemical characterization of three structurally distinct domains along the protofilament unit of desmin 10 nm filaments

Limited chymotryptic cleavage of soluble chicken gizzard desmin protofilaments allows the characterization of three structurally distinct domains. A surface-exposed very basic amino-terminal region (the headpiece) with an amino acid sequence excluding a-helical organization (7.5 kd) is separated from the perhaps globular carboxy-terminal 48 residues (the tailpiece) by a distinctly different middle domain of approximately 330 residues. This 38 kd domain is very rich in α-helix (at least 83%), and electron microscopy reveals a thin rod with a length of 500 ± 50 Å. Amino acid sequence data also show that the rod domain is interrupted by a nonhelical portion. An a-helical array is able to form a coiled-coil spanning the carboxy-terminal half of the 38 kd domain. The a-type diffraction pattern of 10 nm filaments arises from a coiled-coil conformation displayed through most but not all of the middle domain of the protofilaments.     

Cloning and sequencing of rat plectin indicates a 466-kD polypeptide chain with a three-domain structure based on a central alpha-helical coiled coil

We have determined the complete cDNA sequence of rat plectin from a number of well-characterized overlapping lambda gt11 clones. The 4,140-residue predicted amino acid sequence (466,481 D) is consistent with a three-domain structural model in which a long central rod domain, having mainly an alpha-helical coiled coil conformation, is flanked by globular NH2- and COOH-terminal domains. The plectin sequence has a number of repeating motifs. The rod domain has five subregions approximately 200-residues long in which there is a strong repeat in the charged amino acids at 10.4 residues that may be involved in association between plectin molecules. The globular COOH-terminal domain has a prominent six-fold tandem repeat, with each repeat having a strongly conserved central region based on nine tandem repeats of a 19-residue motif. The plectin sequence has several marked similarities to that of desmoplakin (Green, K. J., D. A. D. Parry, P. M. Steinert, M. L. A. Virata, R. M. Wagner, B. D. Angst, and L.A. Nilles. 1990. J. Biol. Chem. 265:2,603-2,612), which has a shorter coiled-coil rod domain with a similar 10.4 residue charge periodicity and a COOH-terminal globular domain with three tandem repeats homologous to the six found in plectin. The plectin sequence also has homologies to that of the bullous pemphigoid antigen. Northern blot analysis indicated that there is a significant degree of conservation of plectin genes between rat, human, and chicken and that, as shown previously at the protein level, plectin has a wide tissue distribution. There appeared to be a single rat plectin gene that gave rise to a 15-kb message. Expression of polypeptides encoded by defined fragments of plectin cDNA in E. coli has also been used to localize the epitopes of a range of monoclonal and serum antibodies. This enabled us to tentatively map a sequence involved in plectin-vimentin and plectin-lamin B interactions to a restricted region of the rod domain.     

ZO-3, a Novel Member of the MAGUK Protein Family Found at the Tight Junction, Interacts with ZO-1 and Occludin

A 130-kD protein that coimmunoprecipitates with the tight junction protein ZO-1 was bulk purified from Madin-Darby canine kidney (MDCK) cells and subjected to partial endopeptidase digestion and amino acid sequencing. A resulting 19–amino acid sequence provided the basis for screening canine cDNA libraries. Five overlapping clones contained a single open reading frame of 2,694 bp coding for a protein of 898 amino acids with a predicted molecular mass of 98,414 daltons. Sequence analysis showed that this protein contains three PSD-95/SAP90, discs-large, ZO-1 (PDZ) domains, a src homology (SH3) domain, and a region similar to guanylate kinase, making it homologous to ZO-1, ZO-2, the discs large tumor suppressor gene product of Drosophila, and other members of the MAGUK family of proteins. Like ZO-1 and ZO-2, the novel protein contains a COOH-terminal acidic domain and a basic region between the first and second PDZ domains. Unlike ZO-1 and ZO-2, this protein displays a proline-rich region between PDZ2 and PDZ3 and apparently contains no alternatively spliced domain. MDCK cells stably transfected with an epitope-tagged construct expressed the exogenous polypeptide at an apparent molecular mass of ~130 kD. Moreover, this protein colocalized with ZO-1 at tight junctions by immunofluorescence and immunoelectron microscopy. In vitro affinity analyses demonstrated that recombinant 130-kD protein directly interacts with ZO-1 and the cytoplasmic domain of occludin, but not with ZO-2. We propose that this protein be named ZO-3.