Annexin A2 heterotetramer: role in tight junction assembly

【摘要】  The tight junction has been characterized as a domain of focal fusions of the exoplasmic leaflets of the lipid bilayers from adjacent epithelial cells. Approximating membranes to within fusion distance is a thermodynamically unfavorable process and requires the participation of membrane-bridging or -fusion proteins. No known tight junction protein exhibits such activities. Annexin A2 (A2), in particular its heterotetramer (A2t), is known to form junctions between lipid bilayer structures through molecular bridging of their external leaflets. We demonstrate abundant A2 expression in Madin-Darby canine kidney II monolayers by two-dimensional gel electrophoresis. Confocal immunofluorescence microscopic analysis suggests the bulk of A2 is located along the apical and lateral plasma membrane in its tetrameric configuration, consisting of two A2 and two p11 (an 11-kDa calmodulin-related protein, S100A10) subunits. Immunocytochemistry and ultrastructural immunogold labeling demonstrate colocalization of the A2 subunit with bona fide tight junction proteins, zonula occludens-1, occludin, and claudin-1, at cell-cell contacts. The extracellular addition of a synthetic peptide, targeted to disrupt the binding between A2 and p11, completely aborts tight junction assembly in calcium chelation studies. We propose A2t as a member of a new class of tight junction proteins responsible for the long-observed convergence of adjacent exoplasmic lipid leaflets in tight junction assembly.

【关键词】  annexin II zonula occludens occludin claudin hemifusion

THE TIGHT JUNCTION forms a circumferential belt around an epithelial or endothelial cell, separating the plasma membrane into an apical and a basolateral domain. The belt from one cell adjoins belts from adjacent cells, thereby forming a sheet of cellular barrier between the external environment, e.g., the luminal content of the intestine and the renal tubule, from the regulated internal environment, the interstitial fluid ( 80 ). In blood-brain ( 38 ) and blood-testis ( 12 ) barriers, the tight junction also separates and maintains the integrity of specialized fluid compartments within the internal environment. As a boundary between the apical and basolateral plasma membrane, the tight junction demarcates the asymmetric distribution of protein and lipid molecules in these two domains, thereby generating polarity in the two-dimensional (2-D) plane of the plasma membrane. More recent studies indicate that the tight junction congregates polarity-regulating protein complexes that participate in polarized membrane trafficking and serves as spatial landmark for vesicle docking ( 87 ), suggesting that it also regulates polarity in the three-dimensional space of the cell cytoplasm. Intriguingly, tight junction proteins that regulate cell polarity also control cell proliferation and differentiation ( 9, 30 ) and participate in epithelial-mesenchymal transformation ( 29, 66 ) and carcinogenesis ( 52 ). In addition to their role as a platform for trafficking and signaling molecular complexes, the tight junction proteins also serve as viral receptors mediating viral budding and infection ( 7, 14, 84 ) and as targets for bacterial pathogens and their virulence factors ( 70 ).

Structurally, the tight junction is characterized by a linear fusion ( 20 ) or a series of focal fusions ( 20, 80 ) of juxtaposed exoplasmic lipid leaflets from adjacent epithelial or endothelial cells. Approximating cell membranes into molecular contact requires the displacement of water between the apposing hydrophilic surfaces of the interacting lipid bilayers and is a thermodynamically unfavorable process. Overcoming this high-activation energy barrier that opposes membrane fusion requires specialized fusion proteins ( 39, 81 ). Although a large number of tight junction proteins have been identified ( 27 ), many with adhesive properties, none is known to converge exoplasmic leaflets required in tight junction assembly. Thus the molecular basis for the original morphological observation of close approximation of or fusion between the exoplasmic lipid leaflets remains unknown. Annexin A2 (A2), 1 in particular its heterotetramer (A2t), anchors the cytoplasmic leaflets of secretory vesicles to each other or to the plasma membrane ( 18, 55, 72 ). A2t added to the suspension of chromaffin granules ( 18, 51 ) or liposomes ( 51 ) forms junctions between these lipid bilayer structures through molecular bridging of their external lipid leaflets. Here, we provide morphological and functional evidence for a role of A2t in epithelial tight junction assembly.

MATERIALS AND METHODS

Cell and Monolayer Preparation

Madin-Darby canine kidney (MDCK) II cell lines (a gift from Dr. R. Bacallao, Indiana University School of Medicine, Indianapolis, IN) were routinely cultured and propagated in T25 flasks with DMEM (Mediatech, Herndon, VA) plus 5% fetal calf serum, with penicillin and streptomycin, both at 100 U/ml. Cells were grown in 5% CO 2 at 37°C and refed every third day. Serum weaning was accomplished by passaging serum-grown MDCK II cells through serum-containing medium, progressively diluted by serum-free medium, supplemented with human insulin and transferrin (UltraMDCK, Bio-Whittaker, Walkersville, MD). Cell viability was monitored, and cells were considered completely serum-weaned after three consecutive passages in the UltraMDCK media alone. At 95% confluence, cells were transferred to propagate in new T25 flasks or to grow on optically transparent Transwell membrane inserts (Corning Costar, Cambridge, MA) or glass coverslips (seeding at 0.5 x 10 6 cells/ml). Cells grown on transparent inserts were monitored daily for cell morphology, percent confluence, and transepithelial resistance (TER) and were used for confocal and electron microscopic studies. Cells grown on glass coverslips were used for immunofluorescence microscopic studies.

Cell Lysate Preparation

Cell lysates were prepared according to the protocol of Kendrick Laboratories (Madison, WI). Confluent MDCK II monolayers, grown in T75 flasks, were rinsed three times with cold PBS. One milliliter of Osmotic Lysis Buffer (10 mM Tris, pH 7.4, and 0.3% SDS), preheated in boiling water, was added directly to each flask to give a protein concentration of 2-4 µg/µl. After cooling on ice, 10 µl of x 10 Nuclease Stock per 100 µl Osmotic Lysis Buffer were added and the cells were scraped from the flask using a rubber policeman, mixed with the solution, and incubated on ice for 30 min. An equal volume of SDS Boiling Buffer (5% SDS, 10% glycerol, and 60 mM Tris, pH 6.8) was then added and the mixture was heated in boiling water bath for 3 min. The samples were immediately quick-frozen in an ethanol/dry ice bath and stored at -70°C. The x 10 Nuclease Stock solution ( 25 ) contained 50 mM MgCl 2, 100 mM Tris, pH 7.0, 500 µg/ml RNase (Sigma R5125, Ribonuclease A from Bovine Pancreas Type IIIA), and 1,000 µg/ml DNase (Sigma D4527, Deoxyribonuclease I, Type II from bovine pancreas). The final concentration for these enzymes should be 50 µg/ml RNase and 100 µg/ml DNase in 5 mM MgCl 2 and 10 mM Tris·HCl, pH 7.0.

2-D SDS-Polyacrylamide Gel Electrophoresis

2-D electrophoresis was performed according to the method of O'Farrell ( 64 ) by Kendrick Laboratories. Isoelectric focusing was carried out in glass tubes of 2.0-mm inner diameter, using 2.0% Resolytes, pH 4-8 ampholines (BDH from Hoefer Scientific Instruments, San Francisco, CA) for 9,600 V/h. One microgram of an isoelectric focusing internal standard, tropomyosin protein, was added to the samples. Tropomyosin shows two polypeptide spots of similar pI; the lower spot of molecular mass 33,000 and pI 5.2 is marked with an arrow on the Coomassie blue-stained 2-D gels (S; Fig. 1 ). After equilibration for 10 min in buffer O (10% glycerol, 50 mM dithiothreitol, 2.3% SDS, and 0.0625 M Tris, pH 6.8), the tube gels were sealed to the top of stacking gels that were on top of 10% acrylamide slab gels (0.75-mm thick) and SDS slab gel electrophoresis was carried out for 4 h at 12.5 mA/gel. The following proteins (Sigma, St. Louis, MO) were added as molecular mass standards to a well in the agarose that sealed the tube gel to the slab gel: myosin (220,000), phosphorylase A (94,000), catalase (60,000), actin (43,000), carbonic anhydrase (29,000), and lysozyme (14,000). These standards appear as horizontal lines on the Coomassie Brilliant Blue R-250-stained 10% acrylamide gels ( Fig. 1 ). The Coomassie blue-stained gels were dried between sheets of cellophane with an acid edge to the left.

Fig. 1. Coomassie stain of two-dimensional SDS-polyacrylamide gel electrophoresis of cell lysate from Madin-Darby canine kidney II monolayers. Peptides A and B were microsequenced and identified as annexin A2 (A2) and galectin-3, respectively. Spot S represents the internal standard tropomyosin (molecular mass 33,000, pI 5.2). Proteins added as molecular mass standards are myosin (220,000), phosphorylase A (94,000), catalase (60,000), actin (43,000), carbonic anhydrase (29,000), and lysozyme (14,000) and appear as light horizontal lines across the gel.

Laser Densitometric Scanning of the 2-D Gels

2-D gels were scanned with a Digital Instruments LGS 50 laser densitometer using an optical density range of 0-0.7 OD units. The images were further analyzed using Phoenix 2D Full software version 3.51, and the relative abundance of each polypeptide is expressed as a percentage of all the polypeptide spots scanned.

Identification of Polypeptides from 2-D Gels

This was performed at the Protein Chemistry Core Facility of Howard Hughes Medical Institute/Columbia University (New York, NY) based on published methods ( 22, 47 ). Gels containing the selected polypeptide spots (A and B in Fig. 1 ) were stained with 0.05% Coomassie Brilliant Blue G/0.5% acetic acid/20% methanol and destained with 30% methanol. The stained, excised polypeptide spots were soaked in water overnight and then transferred to a microcentrifuge tube and soaked in 500 µl of 0.1 M Tris, pH 9/50% acetonitrile for 30 min. The supernatant was discarded, and the wash was repeated four times. The washed gel pieces were placed on a clean glass plate to air dry for 5-10 min. To each gel piece was applied 2 µl of Lys-C (containing 0.075-0.2 µg enzyme depending on the concentration of protein in each piece). The enzyme solution was absorbed by the gel after several minutes. Two to five microliters of digestion buffer (0.1 M Tris, pH 9/0.01% Tween 20) were applied to each piece of gel and allowed to soak in. The gel pieces were then submerged in 50 µl of digestion buffer and incubated for 20 h at 30°C. When digestion was complete, the supernatant was transferred to a Hewlett-Packard HPLC injection vial. The gel pieces were washed three times with 100 µl of 60% acetonitrile/0.1% TFA, soaking 30 min each time, and the supernatants were transferred to the injection vial. The combined washes in the injection vial were dried in a Speed-Vac concentrator and redissolved in 200 µl of 0.1% TFA for injection onto a HPLC column (Vydac C18, 0.21 x 15 cm). Analysis was carried out using a HP 1090 instrument under the following conditions: absorbance 210 nm, flow rate 0.2 ml/min, buffer A 0.075% TFA, buffer B 0.65% TFA in 80% acetonitrile, gradient 2% buffer B (0-10 min) and 2-75% buffer B (10-120 min). Selected peaks were sequenced using an Applied Biosystems 477A sequencer (Applied Biosystems, Foster City, CA).

Immunofluorescence and Confocal Microscopy

Sample preparation is based on published protocols ( 5, 78 ) with modifications. In general, serum-weaned MDCK II cells grown into monolayers on optically transparent Transwell membrane inserts or glass coverslips were washed in PBS and fixed with 2% paraformaldehyde in PBS for 20 min at room temperature. After 30 min of blocking with PBS containing 3% nonfat milk (blocking buffer), primary antibodies were applied for 2 h at room temperature. The monolayers were then washed and incubated with secondary antibodies, either Cy3-conjugated anti-mouse or -rabbit IgG or FITC-conjugated anti-mouse or -rabbit IgG (Sigma BioSciences, St. Louis, MO) for 2 h at room temperature in the dark. The cells were then washed and mounted using SlowFade Antifade Kit (Molecular Probes, Eugene, OR). Blocking buffer was used for all washing and dilution of primary and secondary antibodies. For confocal microscopic localization of A2t subunits ( Fig. 2 ), 90-100% confluent monolayers on membrane inserts were washed in PBS before fixation with 2% paraformaldehyde in PBS for 20 min. The membranes were washed again with PBS, two times, and treated with 50 mM NH 4 Cl in PBS for 10 min at room temperature. After being washed in saponin buffer (0.075% saponin in PBS + 0.3% milk) for 10 min, the membranes were washed in PBS and treated with Triton X-100 (0.2% Triton X-100 in PBS) for 5 min at room temperature. The membrane inserts were then blocked with PBS and milk and stained with mouse monoclonal anti-A2 or mouse monoclonal anti-A2 light chain, p11 (BD Transduction Laboratories, Lexington, KY), both at a final concentration of 0.0167 mg/ml. The monolayers were then washed and incubated with secondary antibodies Cy3-conjugated anti-mouse IgG at 1:125 dilution. The images were analyzed using a Leica TCS Inverted Confocal Microscope at the Carol Moss Spivak Imaging Facility, University of California, Los Angeles. For colocalization of A2 and p11 subunits ( Fig. 3 ), monolayers on glass coverslips were preextracted with extraction buffer (0.2% Triton X-100, 100 mM KCl, 3 mM MgCl 2, 200 mM sucrose, 10 mM HEPES, pH 7.1), before fixing with 2% paraformaldehyde in PBS. The monolayers were then permeabilized with 0.1% Triton X-100 and washed according to the protocol above and stained with rabbit polyclonal anti-p11 (Biodesign International, Saco, ME) at 0.04 mg/ml concentration and mouse monoclonal anti-A2 at 0.008 mg/ml. Secondary antibodies FITC-conjugated anti-rabbit IgG and Cy3-conjugated anti-mouse IgG were applied at 1:30 and 1:250 dilutions, respectively. For the colocalization studies ( Fig. 4 ) involving A2 subunit with tight junction proteins zonula occludens-1 (ZO-1), occludin, and claudin-1, glass coverslips containing 70-100% confluent monolayers were stained with mouse monoclonal anti-A2 at a final concentration of 0.008 mg/ml and rabbit polyclonal anti-occludin, anti-ZO-1, or anti-claudin-1 (Zymed Laboratories, San Francisco, CA) at final concentrations of 0.005, 0.021, and 0.02 mg/ml, respectively. Secondary antibodies Cy3-conjugated anti-mouse IgG and FITC-conjugated anti-rabbit IgG were applied at 1:250 and 1:30 dilutions, respectively. Images of cell islands with well-established cell-cell contacts were obtained using an Olympus IX70 Microscope equipped with an RFC Reflected Light Fluorescence Attachment and visualized with the Olympus MagnaFire SP Digital Imaging System. Images were processed using Photoshop 6.0 (Adobe) software.

Fig. 2. Immunofluorescent confocal laser microscopy demonstrating similar plasma membrane staining of A2 and p11. In x - y sections ( top ), both A2 ( A ) and p11 ( B ) localize predominantly along cell-cell contacts. In x - z sections ( bottom ), both A2 and p11 staining is observed along the length of lateral and apical plasma membrane. Bar = 10 µm.

Fig. 3. Immunofluorescent colocalization of A2 and p11. Colocalization of A2 (red) and p11 (green) staining is confirmed by the sharp merged image (yellow). Top : images were obtained from cell aggregates with well-established cell-cell contacts in 50% confluent monolayers. Bottom : images are from 100% confluent monolayers. Note the loss of nuclear staining upon attainment of full confluence. Bar = 10 µm.

Fig. 4. Colocalization of the A2 subunit of A2t with tight junction proteins. Immunofluorescent micrographs of A2 (red: A, D, G ) and tight junction proteins (green): zonula occludens-1 (ZO-1; B ), occludin ( E ), and claudin-1 ( H ). Colocalization of A2 with each of the 3 tight junction proteins is demonstrated by the yellow merged images ( C, F, I ). Bar = 10 µm.

For confocal microscopic localization of A2t subunits ( Fig. 2 ), monolayers on membrane inserts were permeabilized as recommended ( 5 ). For colocalization of A2 and p11 on glass coverslips, monolayers were preextracted and permeabilized to enhance staining ( Fig. 3 ). Both A2 and p11 staining were visible but with lesser intensity in unextracted and unpermeabilized monolayers (data not shown). For colocalization of A2 with tight junction proteins, preextraction and permeabilization were performed only for A2/occludin dual staining ( Fig. 4, D-F ), as recommended by other investigators ( 6, 40 ). Neither preextraction nor permeabilization was used in A2/ZO-1 ( Fig. 4, A-C ) and A2/claudin-1 ( Fig. 4, G-I ) colocalization.

Electron Microscopy

Postembedding immunocytochemistry. Monolayers on Transwell membrane were first fixed with 1% paraformaldehyde in 0.1 M cacodylate buffer at room temperature and then placed on ice for 1-2 h. After being washed with cacodylate buffer, the preparations were partially dehydrated to 90% ethanol, infiltrated with LR White (London Resin White), and embedded in gelatin capsules at 50°C. Sections were cut with a DDK Diamond Knife on a Sorvall MT2-B Ultramicrotome and picked up on nickel grids. After citrate antigen retrieval, grids were stained as follows: blocked with 2.5% normal goat serum, 5% nonfat milk, and 0.1% cold fish gelatin in PBS, incubated with mouse monoclonal anti-A2, and rabbit polyclonal anti-occludin or anti-ZO-1, (all diluted 1:15 with PBS). Grids were then washed with PBS and incubated with goat anti-mouse IgG conjugated with 20 nm gold and goat anti-rabbit IgG conjugated with 10 nm gold (B B International) diluted 1:70 in PBS with cold fish gelatin. Grids were stained with uranyl acetate and lead citrate and viewed on a Philips 201c Electron Microscope.

Peptide Synthesis and Purification

Peptide A comprising the NH 2 -terminal 14 residues of A2 (AcSTVHEILCKLSLEG), including the N -acetyl group (Ac) of the terminal serine and Peptide S in which the sequence of the 14 residues was reversed, was synthesized on Wang resin via Fmoc chemistry. T-Butyl group was used for COOH (Glu/Asp) and NH 2 group (Ser/Thr) protection; Trityl group for Cys and Gln side chain and Boc for Trp; Pbf for Arg. All materials were obtained from American Peptide (Sunnyvale, CA). The peptide chain was assembled on resin by repetitive removal of the Fmoc protecting group and coupling of protected amino acid. HBTU was used as a coupling reagent, and N -methylmorphiline was used as a base. After removal of the last Fmoc protecting group, resin was treated with TFA cocktail for cleavage and removal of the side chain protecting groups. Crude peptide was precipitated from cold ether and collected by filtration. Purification of crude peptide was achieved via RP-HPLC using 100 x 300-mm column. Packing media were obtained from Waters group. Analytic HPLC column (C18, 4.6 x 250 nm) was obtained from Vydac Separation Groups. Purified peptide was isolated by lyophilization and verified by mass spectroscopic analysis.

Calcium Chelation Studies

Serum-weaned cells were grown into confluent monolayers on 6.5-mm-diameter, optically transparent membrane inserts in Costar Transwell cell culture chamber (Corning Costar) in UltraMDCK media (1.4 mM Ca 2+ concentration). Tight junction disassembly was achieved by transferring the inserts into Ca 2+ -free medium consisting of Minimum Essential Medium for Suspension (SMEM, Sigma BioSciences) plus 2 mM EGTA. Tight junction reassembly was induced by transferring the inserts back into UltraMDCK media. TER was monitored using a Millicell-ERS Electrical Resistance System (Millipore, Marlborough, MA). For details, see text and legends (see Fig. 7 ).

Fig. 7. Effect of specific A2t disruption on tight junction reassembly in calcium chelation studies. A : at minute 0, 2 groups of 3 confluent monolayers that attained peak transepithelial resistance (TER) were transferred into Ca 2+ -free medium (-Ca 2+ ), either with or without Peptide A (1 mM) in both the apical and basal compartments. At 240 min, both treatment groups were switched back to normal Ca 2+ (1.4 mM) medium (+Ca 2+ ). The anticipated recovery in TER observed in the control group was not seen in the Peptide A group. Replacement of fresh incubation medium without Peptide A at 1,470 min (arrow) was followed by prompt recovery in TER. B : study is repeated with an additional scrambled peptide ( Peptide S ) treatment group. Peptide S (1 mM) exerted no effect on TER recovery in monolayers following calcium restoration. Peptide A treatment again prevented TER recovery. In this experiment, periodic and complete replacement of incubation medium with fresh medium containing 1 mM Peptide A (arrowheads) was needed for sustained inhibition of TER recovery. Identical replacement of scrambled peptide (1 mM) in Peptide S monolayers or incubation medium (vehicle) in control monolayers did not influence TER recovery.

RESULTS

2-D gel electrophoresis of cell lysate from confluent MDCK II monolayers demonstrates 86 polypeptide spots between 17 and 93 kDa ( Fig. 1 ). Two major polypeptides (A and B in Fig. 1 ) in the 33-kDa range accounted for 5 and 6%, respectively, of all the polypeptides scanned by laser densitometry. They were microsequenced and identified as A2 (A in Fig. 1 ) and galectin-3 (B in Fig. 1 ). In epithelial cells, 90-95% of total cellular A2 is localized to the plasma membrane as a heterotetramer ( 26, 37 ), consisting of two subunits of 36-kDa monomeric A2 and two subunits of 11-kDa, calmodulin-related protein, S100A10 (p11) ( 26, 83 ). The plasma membrane localization of A2t in MDCK I monolayers is demonstrated by a similar membrane distribution pattern of its two subunits, A2 and p11 ( 37 ). In MDCK II monolayers, we also found the distribution of both subunits predominantly along the plasma membrane. Confocal microscopy demonstrates that both A2 and p11 stain cell-cell contacts in the x - y plane ( Fig. 2, top ) and the lateral and apical membrane domains in the x - z plane ( Fig. 2, bottom ). These findings are consistent with the notion that almost all A2 is localized to the plasma membrane as A2t in the intact epithelium ( 26, 37 ).

We then proceeded to address the questions whether A2t is a structural component of the tight junction and whether it has a direct functional role in its assembly. Because serum can modify tight junction structure and function ( 11, 16, 58, 60 ), all subsequent studies were conducted using serum-weaned cells. Because there is no antibody directed against the A2t molecule, its cellular distribution is conventionally demonstrated by the use of antibodies against its subunits, A2 or p11. To validate the use of the A2 subunit as a marker for A2t, we first confirmed the identical localization of A2 and p11 at cell-cell contacts. This is demonstrated in both preconfluent ( Fig. 3, top ) and confluent ( Fig. 3, bottom ) monolayers. Note that the nuclear staining in the preconfluent monolayer is lost upon attainment of full confluence. Confluent monolayers also give the appearance of cell crowding ( Fig. 3, bottom ) with focal areas of multilayered growth (data not shown). Because certain junction assembly events are not observed in postconfluent monolayers with cell crowding ( 35, 82 ), subsequent images were obtained in cell islands with established cell-cell contacts in immediately preconfluent monolayers (70-100% confluence).

Because the A2 subunit is distributed along the length of the lateral plasma membrane, including the tight junction area ( Fig. 2, bottom ), and because ZO-1 ( 15, 24, 68 ), occludin ( 15, 24, 68 ), or claudin-1 ( 23, 68 ) are established markers of the tight junction, colocalization of these proteins in x - z sections is predictable and will not be expected to yield additional information, especially at light microscopic resolutions. On the other hand, strong staining of A2 in the x - z sections, at the focal plane of known tight junction proteins, will provide additional support for the notion that A2 (although present along the whole length of the lateral membrane) is also present in quantity in the plane of the tight junction, in close proximity to bona fide tight junction markers. Figure 4 demonstrates the colocalization of A2 subunit with ZO-1 ( A - C ), occludin ( D - F ), or claudin-1 ( G - I ) in x - y sections at the focal plane of the tight junction. Note also the nuclear staining in all images, most apparent for A2 and ZO-1. The staining of ZO-1, a cytoplasmic protein, and claudin-1 using an antibody to a cytoplasmic epitope was accomplished in paraformaldehyde-fixed monolayers without postfixation permeabilization. Paraformaldehyde is known to alter membrane permeability, allowing antibodies and fluorochromes to reach cytoplasmic ( 34 ) and nuclear ( 54 ) targets. Although nuclear localization of A2 ( 19 ) and ZO-1 ( 28, 67 ) has been observed in preconfluent epithelial cells in the proliferative phase, we are not aware of any report on the nuclear presence of occludin and claudin-1. The basis for this observation is not addressed in this paper.

The predicted distribution of the tight junction proteins and their colocalization with A2t in longitudinal ( x - z ) sections of MDCK II monolayers is confirmed by immunoelectronmicroscopy. Figure 5 depicts ZO-1 ( a ) and occludin ( b ), both labeled with 10-nm gold particles, located in the most apical portion of the lateral plasma membrane, demarcating the tight junction. The A2 subunit of A2t (20-nm gold particles; c ) is not only present in the tight junction region but also extends basally along the length of the lateral membrane and apically into the luminal membrane. Although most of the A2 subunit is present along or in close proximity to the cytoplasmic face of the plasma membrane, confirming the well-described submembranous distribution of A2t ( 26, 48, 83 ), many gold labels are also present in the intercellular space and over the external face of the plasma membrane (open arrowheads; c - e ) consistent with prior observation of the presence of extracellular A2t ( 26, 44, 57, 83 ). Double labeling of A2, 20-nm gold particles (open arrowheads), with ZO-1 ( d ) or occludin ( e ), both 10-nm gold particles (filled arrowheads), confirms the close spatial association between A2t and the tight junction proteins. The colocalization of the A2 subunit with tight junction proteins, both at light microscopic ( Fig. 4 ) and ultrastructural ( Fig. 5 ) resolutions, leads to the question whether this molecule has a functional role in tight junction assembly.

Fig. 5. Ultrastructural immunogold localization of the A2 subunit of A2t and tight junction proteins. Longitudinal sections depicting lateral plasma membrane localization of ZO-1 ( a ), occludin ( b ), or A2 ( c ). ZO-1 and occludin (10-nm gold particles, between filled arrowheads) are located in the most apical pole demarcating the tight junction, whereas A2 (20-nm gold particles) extends beyond the tight junction: basally along the length of the lateral plasma membrane and apically into the luminal membrane. Most A2 labels are localized along or in proximity to the cytoplasmic face of the plasma membrane, although many are also seen over the intercellular space and the exoplasmic leaflet of the plasma membrane ( c - e, open arrowheads). Ultrastructural juxtapositioning of A2 with tight junction proteins is depicted by dual immunogold labeling of A2, 20-nm gold particles (open arrowheads) with ZO-1 ( d ) or occludin ( e ), both 10-nm gold particles (filled arrowheads). Sections are perpendicular to the plasma membrane with both apposing membranes clearly visualized. D, desmosomes. Bar = 0.1 µm.

This possibility is tested by examining the effect of molecular disruption of A2t, by targeted interference of the binding between A2 and p11, on tight junction assembly (as reflected by TER). The last 14 residues of the NH 2 -terminal domain (N-domain) of the A2 subunit, including the acetyl group (Ac), harbor the entire binding site for p11 ( Fig. 6 ) with the terminal acetyl group being critical for high binding specificity and affinity ( 45 ). Therefore, we constructed a peptide identical to the terminal 14 residues in the correctly acetylated form ( Peptide A ). Previous studies showed that this peptide, at micromolar concentrations, reversibly disrupts p11 binding to A2 in preformed A2t in living cells. In these studies, the peptide was loaded into cells using patch pipettes and the physiological consequence of intracellular A2t disruption was examined ( 45, 62 ). We used an incubation medium with Peptide A concentration of 1 mM to test for the possibility that intact, extracellular A2t is required for tight junction assembly. An extracellular concentration of 1 mM was used in a study on the effect of a synthetic peptide in disrupting the binding between A2 and the exoplasmic leaflet of endothelial cells ( 32 ). Also, in a pilot dose-response study we found 1 mM extracellular Peptide A exerted the greatest inhibitory effect on the attainment of maximum TER by MDCK II cells grown on membrane inserts (data not shown). Peptide A precipitates out of solution at concentrations greater than 1 mM.

Fig. 6. A2t-mediated junction formation between 2 lipid bilayers [modified from Lambert et al. ( 51 )]. A2-p11 complex (A2t) forms a symmetric bridge between 2 lipid bilayers ( right : only 1 of the 2 participating lipid bilayers). The 2 A2 subunits, each binding to a separate bilayer, are in turn linked via a p11 dimer. For details, see text.

Figure 7 summarizes the effect of Peptide A on tight junction assembly in calcium chelation studies. Confluent, serum-weaned MDCK II monolayers with stable TER were transferred into a calcium-free medium (-Ca 2+ ), either with or without Peptide A ( Fig. 7 A ). The anticipated tight junction disassembly is reflected by the rapid drop in TER to a steady level of 10% of its initial value. Synchronous reassembly of the tight junction was induced after 240 min by switching the monolayers back into normal calcium medium (+Ca 2+ ), again with or without Peptide A. In control monolayers, TER steadily returned toward its initial level. In Peptide A -treated monolayers, TER recovery, other than an initial rise of 7 /cm 2, was not observed. Prompt and full recovery of TER to levels in control monolayers followed the replacement of fresh incubation medium, but without Peptide A (arrow).

To exclude the possibility of a nonspecific "peptide effect," the study was repeated with an additional treatment group using an acetylated scrambled peptide ( Peptide S ) in which the sequence of the 14 amino acids was reversed. Monolayers treated with Peptide S (1 mM) behaved exactly as the control monolayers ( Fig. 7 B ). Again, except for a small rise of 5 /cm 2, TER in the Peptide A -treated monolayers failed to recover after Ca 2+ restoration. Intermittent replenishment with Peptide A (arrowheads), at 1 mM concentration, was needed to prevent TER recovery. No peptide replenishment was necessary in the first study ( Fig. 7 A ). In four additional studies, the need for Peptide A replenishment to prevent tight junction reassembly varied between these two extremes. This variability in the requirement for fresh Peptide A suggests that some monolayers may have greater peptide-lytic activities, or greater A2t synthetic activities, or a combination of both. What is clear from all the six experiments is that the disruption of A2-p11 binding using Peptide A aborts tight junction reassembly following recovery from Ca 2+ chelation.

DISCUSSION

Four decades ago, Farquhar and Palade ( 20 ) described the ultrastructural appearance of the tight junction as a linear or a series of punctate fusions of the exoplasmic leaflets of lipid bilayers from adjacent epithelial cells, completely obliterating the intervening extracellular space. Subsequent discovery of transmembrane tight junction proteins ( 27 ) supports earlier postulates that these fusions or "kisses" represent adhesive linkages between the extracellular domains of integral membrane proteins from adjacent cells. Linear polymerization of these tight junction proteins transforms individual protein particles into strands that criss-cross to form a tight junction network encircling the apical portion of each epithelial or endothelial cell ( 80 ). However, of the three major integral tight junction proteins identified to date, occludin, the claudins, and junctional adhesion molecule (JAM), none is known to exhibit membrane-converging activities that are necessary to account for the original observation of exoplasmic leaflet "kisses."

Membrane fusion is an energetically unfavorable process and is accomplished in nature by special fusion molecules such as the SNARE proteins ( 69 ). Fusion between two separate phospholipid bilayer-enclosed aqueous compartments is accomplished first by the approximation and merger of the external leaflets of the two apposing phospholipid bilayers. This state of hemifusion is followed by the merger of the two internal lipid leaflets, thereby completing the fusion process and establishing a direct continuity between the two previously separate lipid bilayer-enclosed aqueous compartments ( 50 ). The tight junction, however, appears to represent a form of stable hemifusion that does not progress to complete fusion. This hemifusion may be mediated by proteins different from those of the SNARE family. Possible candidates include A2 and A2t, which are known to cross-link phospholipid membranes in a hemifusion state without progressing to full fusion ( 18, 51 ).

A2 is a member of the ubiquitous annexin protein family ( 26, 48, 83 ). All proteins of this evolutionary conserved multigene family exhibit the property of binding acidic, negatively charged phospholipids in a Ca 2+ -dependent fashion. A2 is expressed in three molecular configurations: a monomer (A2), a heterodimer (A2d), and a heterotetramer (A2t). A2 is distributed predominantly in the cytoplasm. The heterodimer, formerly known as the primer recognition protein, consists of one subunit each of A2 and 3-phosphoglycerate kinase, is associated with the nucleus, and regulates DNA polymerase -activities. The heterotetramer, made up of two copies each of A2 and p11, is localized to the plasma membrane ( 83 ). A2 and A2t not only bind to but also aggregate or cross-link phospholipid membranes without inducing complete fusion ( 18, 51 ). However, although A2 aggregates chromaffin granules with a K d [Ca 2+ ] {[Ca 2+ ] for half-maximal granule aggregation} of 1 mM, A2t initiates granule aggregation at a threshold [Ca 2+ ] of 0.7 µM and exhibits a K d [Ca 2+ ] of 1.8 µM ( 18 ). The low micromolar [Ca 2+ ] requirement for membrane aggregation renders A2t an ideal candidate for mediating membrane-membrane interactions in the intracellular environment.

Early studies confirmed A2t is an important intracellular membrane-bridging protein ( 18, 55, 72 ). Ultrastructural analysis in intact tissues reveals A2t anchors secretory vesicles/granules to each other or to the cytoplasmic face of the plasma membrane setting the stage, respectively, for vesicular fusion and vesicle-plasma membrane fusion and secretion ( 61, 71, 72 ). These studies were paralleled by in vitro studies in which the addition of exogenous A2t to suspension of isolated chromaffin granules ( 18, 51 ) or liposomes ( 51 ) induced aggregation and junction formation between these lipid bilayer-encased structures without inducing complete fusion. A2t-induced junction formation between liposomes (devoid of other proteins) provides conclusive evidence for its role in junction assembly.

A proposed molecular mechanism for bridging the external leaflets from adjacent lipid bilayers is depicted in Fig. 6. Similar to all annexins, the A2 molecule consists of a highly conserved COOH-terminal protein core domain (C-domain) and a NH 2 -terminal domain (N-domain) that is variable in length and composition in different annexins. The protein core is composed of a stretch of 70 amino acids, which is repeated four times (in the case of annexin VI, eight times). Each repeat folds into five -helices, which are, in turn, wound into a right-hand superhelix. The four repeats are arranged in a planar cyclic array, shaping the molecule into a slightly curved disk with a convex and a concave side ( 53 ). Ca 2+ -binding sites are located on the convex face, whereas the NH 2 - and COOH-terminus are located on the concave surface. Each A2 subunit of an A2t molecule binds to the outer phospholipid leaflet of one lipid bilayer along its convex face and to the p11 dimer along its concave face. The A2-outer phospholipid leaflet binding is mediated through a calcium-bridging mechanism in which a Ca 2+ molecule links the Ca 2+ -binding site of the A2 subunit to the negatively charged phospholipid leaflet ( 48, 76 ).

Our study suggests that A2t may assemble tight junctions in a similar fashion, i.e., by molecular linkage of the exoplasmic leaflets from adjacent lipid bilayers. This is supported by both structural and functional evidence. Based on 2-D gel analysis and microsequencing, we demonstrated A2 as a major protein in MDCK II monolayers ( Fig. 1 ). Confocal microscopic analysis revealed predominant localization of A2, as is p11, along the apical and lateral plasma membranes ( Fig. 2 ); consistent with prior reports that 90-95% of epithelial A2 is distributed along the plasma membrane as an A2 2 -p11 2 heterotetramer ( 26, 37 ). After confirming the identical localization of A2 and p11 at cell-cell contacts ( Fig. 3 ), we demonstrated the colocalization of A2t (using the A2 subunit as a marker) with tight junction proteins both at light microscopic ( Fig. 4 ) and ultrastructural ( Fig. 5 ) resolutions. Immunoelectronmicroscopy confirms the presence of ZO-1 ( Fig. 5 a ) and occludin ( Fig. 5 b ) in the tight junction domain and demonstrates the presence of many gold-labeled A2 subunits in the intercellular space and along the extracellular face of the lateral plasma membrane ( Fig. 5 c ). Ultrastructural proximity of A2t with ZO-1 ( Fig. 5 d ) and occludin ( Fig. 5 e ) further supports a role of this molecule in tight junction assembly. However, the morphological data up to this point cannot conclusively exclude the possibility that A2 by itself as a monomer, rather than A2t, is responsible for the assembly of the tight junction. This possibility is consistent with reports that A2 also induces junction formation between liposomes ( 51 ) and that, similar to A2t, it is present on the extracellular face of the plasma membrane. However, the observation that targeted disruption of A2t using Peptide A, which specifically dissociates p11 from A2, aborts tight junction reassembly in Ca 2+ -chelation studies ( Fig. 7 ) supports A2t, rather than A2, as the dominant player. In addition, the observation that the presence of Peptide A in the incubation medium aborts the anticipated TER recovery supports a critical role for intact, extracellular A2t in tight junction assembly.

Although the presence of A2 ( 13, 33, 56, 73, 79 ) and A2t ( 44, 46, 57, 86 ) on the extracellular face of the plasma membrane is well documented, an understanding of the export mechanism for these classical intracellular proteins has only begun to emerge. Although annexin proteins do not have the hydrophobic sequence required for the canonical secretory pathway ( 26, 83 ), externalization of A2 ( 21 ) and A2 and p11 ( 43 ) through nonclassical secretory pathways has been demonstrated. More recently, Zhao and associates ( 89 ) provided direct evidence for A2 secretion through the activation of the insulin receptor, the insulin-like growth factor receptor, and their signaling pathways.

The mechanism mediating the binding between the A2 subunit of A2t and the exoplasmic leaflet of the plasma membrane is also not well defined. The calcium-bridging mechanism ( 48, 76 ) discussed earlier is not directly applicable here because, unlike the external leaflets of the chromaffin granules and the liposomes or the cytoplasmic leaflet of the plasma membrane, all of which are enriched with negatively charged phosphatidylserine, the exoplasmic leaflet of the plasma membrane consists predominantly of phosphatidylcholine and sphingomyelin, which exhibit no net charge ( 75 ). However, recent studies indicate that in addition to the Ca 2+ -bridging mechanism, A2 and A2t also bind directly to cholesterol of the lipid membrane ( 3, 26, 88 ). Nusrat and associates ( 63 ) demonstrated that tight junctions are membrane microdomains that are cholesterol-sphyingolipid-enriched membrane structures ( 74 ). Moreover, in these cholesterol-rich membrane microdomains, Kunzelmann-Marche and associates ( 49 ), using human erythroleukemia cells, observed the externalization of phosphatidylserine from the cytoplasmic leaflet to the exoplasmic leaflet of the plasma membrane. This lipid asymmetry is dependent on microdomain integrity and is lost with its disruption. Thus in the tight junction microdomains, A2t may link the adjacent exoplasmic leaflets through cholesterol or Ca 2+ bridges, or a combination of both. In addition, Kassam et al. ( 44 ) noted a high affinity of A2t for heparin and speculated that A2t may bind with cell surface heparan sulfate glycosaminoglycan.

Many studies demonstrated E-cadherin (formerly known as uvomorulin and a major structural and functional component of the adherens junction) is important in the formation and maintenance of the tight junction ( 8, 31, 85 ). Ando-Akatsuka et al. ( 2 ) observed colocalization of E-cadherin and ZO-1 to early cell-cell contacts. As the contact matures, ZO-1 recruits occludin to form the tight junction, whereas E-cadherin constitutes the more basal adherens junction. More detailed analyses of cadherin-induced cell-cell contacts demonstrate that cellular protrusions from one cell indent and embed into the plasma membrane of a second cell ( 1, 82 ). At the tip of each protrusion is a spot-like punctum formed by the clustering of transmembrane cadherin molecules and their accessory proteins. On tissue sections, the puncta at the tips of protrusions from each of the two junction-forming cells line up in a row and together form a two-row zipper-like structure. Stabilization and reorganization of this "adhesion zipper" eventually seal the opposing membranes ( 1, 82 ). However, one issue remains unresolved: although the adhesive action of the cadherins and their accessory proteins is well established, these proteins are not known to exhibit membrane-sealing or -fusion properties. We propose the heterotetramer of A2 is the missing sealing molecule in these studies. This is supported by the recent observation that A2 (as a subunit of its tetrameric moiety) accumulates at cadherin-driven cell-cell contacts ( 35 ) where, in earlier observations ( 1, 82 ), opposing membranes were seen to "zip" close.

The distribution of A2t along the length of the apical and lateral plasma membrane ( Figs. 2 and 5 ), outside the tight junction, is consistent with prior reports and with its other known biological actions, which include Ca 2+ -dependent exocytosis, endocytosis, and cell-cell adhesion ( 26, 48, 83 ). On the cytoplasmic side of the plasma membrane, A2t has been observed as ultrastructural strands of 7-8 nm (70-80 ) to as long as 80 nm (800 ) that link secretary granules to each other or to the plasma membrane ( 61, 72 ). Our observation of the basal extension of extracellular A2t beyond the tight junction ( Fig. 5 c, open arrowheads) raises the possibility that A2t may also link lateral plasma membranes from apposing cells and thereby regulate the dimension and configuration of the intercellular space beyond and basal to the tight junction. In addition, the possibility that the tight junction can dynamically extend its depth basally is suggested by the study of Diamond and Tormey ( 17 ), who reported the absence of intercellular space between adjacent cells in gall bladder epithelia in which water transport was arrested. Kachar and Pinto da Silva ( 41 ) reported that tight junction assembly could be induced along the entire length of the lateral plasma membrane in rat prostatic epithelial cells. Because tight junction strand formation was rapid (within 5 min) and occurred in the presence of protein synthesis inhibitor and metabolic uncoupler, they proposed that the assemblage was carried out by "preexisting components...of the lateral plasma membrane." We postulate A2t as a possible candidate for their proposed "preexisting components."

In conclusion, we propose that A2t plays a role in tight junction assembly possibly through linking juxtaposed exoplasmic leaflets to form a lipid platform across the intercellular space ( 42, 65 ). Into this lipid platform other tight junction proteins such as the occludin, the claudins, and the JAM insert and regulate permeability across epithelial or endothelial sheets. On the cytoplasmic face of the plasma membrane, A2t has been shown to participate in membrane microdomain assembly, on the one hand ( 4, 36 ), and to link the plasma membrane to the cytoskeleton, on the other hand ( 77, 90 ). Recent studies indicate that the tight junction lipid microdomain assembles a large number of macromolecules responsible for the regulation of cell polarity and cell growth and differentiation ( 63, 87 ). Thus A2t appears to play a role in all the known functions of the tight junction, i.e., permeability, polarity, and signal transduction as well as growth regulation.

GRANTS

This work is supported by grants from Public Health Service/National Institute of Diabetes and Digestive and Kidney Diseases (1RO1DK/HD-51948) and from the Department of Veterans Affairs Merit Review Board. I. Hale was the recipient, sequentially, of a research fellowship award from the National Kidney Foundation of Southern California and a training grant award from National Heart, Lung, and Blood Institute, National Institutes of Health, T32HL-07656. This work was presented in part at the 31st Annual Meeting of the American Society of Nephrology, October 25-28, 1998, Philadelphia, PA.

ACKNOWLEDGMENTS

The authors thank R. Clark, K. Mouchmouchian, K. Douglas, N. Kendrick (Kendrick Laboratories, Madison, WI), M. A. Gawinowicz (Protein Chemistry Core Facility, College of Physicians and Surgeons, New York, NY), M. Schibler (Carol Moss Spivak Imaging Facility, University of California, Los Angeles, CA), and G. Hu (American Peptide, Sunnyvale, CA) for technical assistance and P. W. Lee for editorial assistance.

【参考文献】
  Adams CL, Chen YT, Smith SJ, and Nelson WJ. Mechanisms of epithelial cell-cell adhesion and cell compaction revealed by high-resolution tracking of E-cadherin-green fluorescent protein. J Cell Biol 142: 1105-1119, 1998.

Ando-Akatsuka Y, Yonemura S, Itoh M, Furuse M, and Tsukita S. Differential behavior of E-cadherin and occludin in their colocalization with ZO-1 during the establishment of epithelial cell polarity. J Cell Physiol 179: 115-125, 1999. <a href="/cgi/external_ref?access_num=10.1002/(SICI)1097-4652(199905)179:2

Ayala-Sanmartin J, Henry JP, and Pradel LA. Cholesterol regulates membrane binding and aggregation by annexin 2 at submicromolar Ca 2+ concentration. Biochim Biophys Acta 1510: 18-28, 2001.

Babiychuk EB and Draeger A. Annexins in cell membrane dynamics. Ca 2+ -regulated association of lipid microdomains. J Cell Biol 150: 1113-1124, 2000.

Bacallao R, Kiai K, and Jesaitis L. Guiding principles of specimen preservation for confocal fluorescence microscopy. In: Handbook of Biological Confocal Microscopy (2nd ed.), edited by Pawley JB. New York: Plenum, 1995, p. 311-325.

Balda MS, Whitney JA, Flores C, Gonzalez S, Cereijido M, and Matter K. Functional dissociation of paracellular permeability and transepithelial electrical resistance and disruption of the apical- basolateral intramembrane diffusion barrier by expression of a mutant tight junction membrane protein. J Cell Biol 134: 1031-1049, 1996.

Barton ES, Forrest JC, Connolly JL, Chappell JD, Liu Y, Schnell FJ, Nusrat A, Parkos CA, and Dermody TS. Junction adhesion molecule is a receptor for reovirus. Cell 104: 441-451, 2001.

Behrens J, Birchmeier W, Goodman SL, and Imhof BA. 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.

Bilder D. PDZ proteins and polarity: functions from the fly. Trends Genet 17: 511-519, 2001.

Cereijido M, Robbins ES, Dolan WJ, Rotunno CA, and Sabatini DD. Polarized monolayers formed by epithelial cells on a permeable and translucent support. J Cell Biol 77: 853-880, 1978.

Chang C, Wang X, and Caldwell RB. Serum opens tight junctions and reduces ZO-1 protein in retinal epithelial cells. J Neurochem 69: 859-867, 1997.

Cheng CY and Mruk DD. Cell junction dynamics in the testis: sertoli-germ cell interactions and male contraceptive development. Physiol Rev 82: 825-874, 2002.

Chung CY, Murphy-Ullrich JE, and Erickson HP. Mitogenesis, cell migration, and loss of focal adhesions induced by tenascin-C interacting with its cell surface receptor, annexin II. Mol Biol Cell 7: 883-892, 1996.

Cohen CJ, Shieh JT, Pickles RJ, Okegawa T, Hsieh JT, and Bergelson JM. The coxsackievirus and adenovirus receptor is a transmembrane component of the tight junction. Proc Natl Acad Sci USA 98: 15191-15196, 2001.

Colosetti P, Tunwell RE, Cruttwell C, Arsanto JP, Mauger JP, and Cassio D. The type 3 inositol 1,4,5-trisphosphate receptor is concentrated at the tight junction level in polarized MDCK cells. J Cell Sci 116: 2791-2803, 2003.

Conyers G, Milks L, Conklyn M, Showell H, and Cramer E. A factor in serum lowers resistance and opens tight junctions of MDCK cells. Am J Physiol Cell Physiol 259: C577-C585, 1990.

Diamond JM and Tormey JM. Role of long extracellular channels in fluid transport across epithelia. Nature 210: 817-820, 1966.

Drust DS and Creutz CE. Aggregation of chromaffin granules by calpactin at micromolar levels of calcium. Nature 331: 88-91, 1988.

Eberhard DA, Karns LR, VandenBerg SR, and Creutz CE. Control of the nuclear-cytoplasmic partitioning of annexin II by a nuclear export signal and by p11 binding. J Cell Sci 114: 3155-3166, 2001.

Farquhar MG and Palade GE. Junctional complexes in various epithelia. J Cell Biol 17: 375-412, 1963.

Faure AV, Migne C, Devilliers G, and Ayala-Sanmartin J. Annexin 2 "secretion" accompanying exocytosis of chromaffin cells: possible mechanisms of annexin release. Exp Cell Res 276: 79-89, 2002.

Ferrara P, Rosenfeld J, Guillemot J, and Capdeville J. Techniques in Protein Chemistry, edited by Angeletti RH. San Diego, CA: Academic, 1993, p. 379-387.

Furuse M, Fujita K, Hiiragi T, Fujimoto K, and Tsukita S. Claudin-1 and -2: novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin. J Cell Biol 141: 1539-1550, 1998.

Furuse M, Itoh M, Hirase T, Nagafuchi A, Yonemura S, and Tsukita S. Direct association of occludin with ZO-1 and its possible involvement in the localization of occludin at tight junctions. J Cell Biol 127: 1617-1626, 1994.

Garrels JI. Quantitative two-dimensional gel electrophoresis of proteins. Methods Enzymol 100: 411-423, 1983.

Gerke V and Moss SE. Annexins: from structure to function. Physiol Rev 82: 331-371, 2002.

Gonzalez-Mariscal L, Betanzos A, Nava P, and Jaramillo BE. Tight junction proteins. Prog Biophys Mol Biol 81: 1-44, 2003.

Gottardi CJ, Arpin M, Fanning AS, and 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.

Grande M, Franzen A, Karlsson JO, Ericson LE, Heldin NE, and Nilsson M. Transforming growth factor- and epidermal growth factor synergistically stimulate epithelial to mesenchymal transition (EMT) through a MEK-dependent mechanism in primary cultured pig thyrocytes. J Cell Sci 115: 4227-4236, 2002.

Greaves S. Growth and polarity: the case for scribble. Nat Cell Biol 2: E140, 2000.

Gumbiner B, Stevenson B, and Grimaldi A. The role of the cell adhesion molecule uvomorulin in the formation and maintenance of the epithelial junctional complex. J Cell Biol 107: 1575-1587, 1988.

Hajjar KA, Guevara CA, Lev E, Dowling K, and Chacko J. Interaction of the fibrinolytic receptor, annexin II, with the endothelial cell surface. Essential role of endonexin repeat 2. J Biol Chem 271: 21652-21659, 1996.

Hajjar KA, Jacovina AT, and Chacko J. An endothelial cell receptor for plasminogen/tissue plasminogen activator. I. Identity with annexin II. J Biol Chem 269: 21191-21197, 1994.

Hannah MJ, Weiss U, and Huttner WB. Differential extraction of proteins from paraformaldehyde-fixed cells: lessons from synaptophysin and other membrane proteins. Methods 16: 170-181, 1998.

Hansen MD, Ehrlich JS, and Nelson WJ. Molecular mechanism for orienting membrane and actin dynamics to nascent cell-cell contacts in epithelial cells. J Biol Chem 277: 45371-45376, 2002.

Harder T and Gerke V. The annexin II2p11(2) complex is the major protein component of the Triton X-100-insoluble low-density fraction prepared from MDCK cells in the presence of Ca 2+. Biochim Biophys Acta 1223: 375-382, 1994.

Harder T and Gerke V. The subcellular distribution of early endosomes is affected by the annexin II2p11(2) complex. J Cell Biol 123: 1119-1132, 1993.

Huber JD, Egleton RD, and Davis TP. Molecular physiology and pathophysiology of tight junctions in the blood-brain barrier. Trends Neurosci 24: 719-725, 2001.

Jahn R and Sudhof TC. Membrane fusion and exocytosis. Annu Rev Biochem 68: 863-911, 1999.

Jou TS, Schneeberger EE, and Nelson WJ. Structural and functional regulation of tight junctions by RhoA and Rac1 small GTPases. J Cell Biol 142: 101-115, 1998.

Kachar B and Pinto da Silva P. Rapid massive assembly of tight junction strands. Science 213: 541-544, 1981.

Kan FW. Cytochemical evidence for the presence of phospholipids in epithelial tight junction strands. J Histochem Cytochem 41: 649-656, 1993.

Karimi-Busheri F, Marcoux Y, Tredget EE, Li L, Zheng J, Ghoreishi M, Weinfeld M, and Ghahary A. Expression of a releasable form of annexin II by human keratinocytes. J Cell Biochem 86: 737-747, 2002.

Kassam G, Choi KS, Ghuman J, Kang HM, Fitzpatrick SL, Zackson T, Zackson S, Toba M, Shinomiya A, and Waisman DM. The role of annexin II tetramer in the activation of plasminogen. J Biol Chem 273: 4790-4799, 1998.

Konig J, Prenen J, Nilius B, and Gerke V. The annexin II-p11 complex is involved in regulated exocytosis in bovine pulmonary artery endothelial cells. J Biol Chem 273: 19679-19684, 1998.

Kristoffersen EK and Matre R. Surface annexin II on placental membranes of the fetomaternal interface. Am J Reprod Immunol 36: 141-149, 1996.

Krutzsch HC and Inman JK. N -isopropyliodoacetamide in the reduction and alkylation of proteins: use in microsequence analysis. Anal Biochem 209: 109-116, 1993.

Kubista H, Sacre S, and Moss SE. Annexins and membrane fusion. Subcell Biochem 94: 73-131, 2000.

Kunzelmann-Marche C, Freyssinet JM, and Martinez MC. Loss of plasma membrane phospholipid asymmetry requires raft integrity. Role of transient receptor potential channels and ERK pathway. J Biol Chem 277: 19876-19881, 2002.

Kuzmin PI, Zimmerberg J, Chizmadzhev YA, and Cohen FS. A quantitative model for membrane fusion based on low-energy intermediates. Proc Natl Acad Sci USA 98: 7235-7240, 2001.

Lambert O, Gerke V, Bader MF, Porte F, and Brisson A. Structural analysis of junctions formed between lipid membranes and several annexins by cryo-electron microscopy. J Mol Biol 272: 42-55, 1997.

Li D and Mrsny RJ. Oncogenic Raf-1 disrupts epithelial tight junctions via downregulation of occludin. J Cell Biol 148: 791-800, 2000.

Liemann S and Huber R. Three-dimensional structure of annexins. Cell Mol Life Sci 53: 516-521, 1997.

Linden E, Skoglund P, and Rundquist I. Accessibility of 7-aminoactinomycin D to lymphocyte nuclei after paraformaldehyde fixation. Cytometry 27: 92-95, 1997. <a href="/cgi/external_ref?access_num=10.1002/(SICI)1097-0320(19970101)27:1

Liu L, Fisher AB, and Zimmerman UJ. Lung annexin II promotes fusion of isolated lamellar bodies with liposomes. Biochim Biophys Acta 1259: 166-172, 1995.

Ma AS, Bell DJ, Mittal AA, and Harrison HH. Immunocytochemical detection of extracellular annexin II in cultured human skin keratinocytes and isolation of annexin II isoforms enriched in the extracellular pool. J Cell Sci 107: 1973-1984, 1994.

Mai J, Waisman DM, and Sloane BF. Cell surface complex of cathepsin B/annexin II tetramer in malignant progression. Biochim Biophys Acta 1477: 215-230, 2000.

Marmorstein AD, Mortell KH, Ratcliffe DR, and Cramer EB. Epithelial permeability factor: a serum protein that condenses actin and opens tight junctions. Am J Physiol Cell Physiol 262: C1403-C1410, 1992.

Martinez-Palomo A, Meza I, Beaty G, and Cereijido M. Experimental modulation of occluding junctions in a cultured transporting epithelium. J Cell Biol 87: 736-745, 1980.

Mortell KH, Marmorstein AD, and Cramer EB. Fetal bovine serum and other sera used in tissue culture increase epithelial permeability. In Vitro Cell Dev Biol 29A: 235-238, 1993.

Nakata T, Sobue K, and Hirokawa N. Conformational change and localization of calpactin I complex involved in exocytosis as revealed by quick-freeze, deep-etch electron microscopy and immunocytochemistry. J Cell Biol 110: 13-25, 1990.

Nilius B, Gerke V, Prenen J, Szucs G, Heinke S, Weber K, and Droogmans G. Annexin II modulates volume-activated chloride currents in vascular endothelial cells. J Biol Chem 271: 30631-30636, 1996.

Nusrat A, Parkos CA, Verkade P, Foley CS, Liang TW, Innis-Whitehouse W, Eastburn KK, and Madara JL. Tight junctions are membrane microdomains. J Cell Sci 113: 1771-1781, 2000.

O'Farrell PH. High resolution two-dimensional electrophoresis of proteins. J Biol Chem 250: 4007-4021, 1975.

Pinto da Silva P and Kachar B. On tight-junction structure. Cell 28: 441-450, 1982.

Reichert M, Muller T, and 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 -catenin/Tcf/Lef signaling. J Biol Chem 275: 9492-9500, 2000.

Riesen FK, Rothen-Rutishauser B, and Wunderli-Allenspach H. A ZO1-GFP fusion protein to study the dynamics of tight junctions in living cells. Histochem Cell Biol 117: 307-315, 2002.

Rothen-Rutishauser B, Riesen FK, Braun A, Gunthert M, and Wunderli-Allenspach H. Dynamics of tight and adherens junctions under EGTA treatment. J Membr Biol 188: 151-162, 2002.

Scales SJ, Bock JB, and Scheller RH. The specifics of membrane fusion. Nature 407: 144-146, 2000.

Sears CL. Molecular physiology and pathophysiology of tight junctions V. assault of the tight junction by enteric pathogens. Am J Physiol Gastrointest Liver Physiol 279: G1129-G1134, 2000.

Senda T, Okabe T, Matsuda M, and Fujita H. Quick-freeze, deep-etch visualization of exocytosis in anterior pituitary secretory cells: localization and possible roles of actin and annexin II. Cell Tissue Res 277: 51-60, 1994.

Senda T, Yamashita K, Okabe T, Sugimoto N, and Matsuda M. Intergranular bridges in the anterior pituitary cell and their possible involvement in Ca 2+ -induced granule-granule fusion. Cell Tissue Res 292: 513-519, 1998.

Siever DA and Erickson HP. Extracellular annexin II. Int J Biochem Cell Biol 29: 1219-1223, 1997.

Simons K and Ikonen E. Functional rafts in cell membranes. Nature 387: 569-572, 1997.

Simons K and van Meer G. Lipid sorting in epithelial cells. Biochemistry 27: 6197-6202, 1988.

Swairjo MA, Concha NO, Kaetzel MA, Dedman JR, and Seaton BA. Ca 2+ -bridging mechanism and phospholipid head group recognition in the membrane-binding protein annexin V. Nat Struct Biol 2: 968-974, 1995.

Thiel C, Osborn M, and Gerke V. The tight association of the tyrosine kinase substrate annexin II with the submembranous cytoskeleton depends on intact p11- and Ca 2+ -binding sites. J Cell Sci 103: 733-742, 1992.

Tidball JG and Spencer MJ. PDGF stimulation induces phosphorylation of talin and cytoskeletal reorganization in skeletal muscle. J Cell Biol 123: 627-635, 1993.

Tressler RJ, Updyke TV, Yeatman T, and Nicolson GL. Extracellular annexin II is associated with divalent cation-dependent tumor cell-endothelial cell adhesion of metastatic RAW117 large-cell lymphoma cells. J Cell Biochem 53: 265-276, 1993.

Tsukita S, Furuse M, and Itoh M. Multifunctional strands in tight junctions. Nat Rev Mol Cell Biol 2: 285-293, 2001.

Van Oss CJ. Energetics of cell-cell and cell-biopolymer interactions. Cell Biophys 14: 1-16, 1989.

Vasioukhin V, Bauer C, Yin M, and Fuchs E. Directed actin polymerization is the driving force for epithelial cell-cell adhesion. Cell 100: 209-219, 2000.

Waisman DM. Annexin II tetramer: structure and function. Mol Cell Biochem 149-150: 301-322, 1995.

Walters RW, Freimuth P, Moninger TO, Ganske I, Zabner J, and Welsh MJ. Adenovirus fiber disrupts CAR-mediated intercellular adhesion allowing virus escape. Cell 110: 789-799, 2002.

Watabe M, Nagafuchi A, Tsukita S, and Takeichi M. Induction of polarized cell-cell association and retardation of growth by activation of the E-cadherin-catenin adhesion system in a dispersed carcinoma line. J Cell Biol 127: 247-256, 1994.

Yeatman TJ, Updyke TV, Kaetzel MA, Dedman JR, and Nicolson GL. Expression of annexins on the surfaces of nonmetastatic and metastatic human and rodent tumor cells. Clin Exp Metastasis 11: 37-44, 1993.

Zahraoui A, Louvard D, and Galli T. Tight junction, a platform for trafficking and signaling protein complexes. J Cell Biol 151: F31-F36, 2000.

Zeuschner D, Stoorvogel W, and Gerke V. Association of annexin 2 with recycling endosomes requires either calcium- or cholesterol-stabilized membrane domains. Eur J Cell Biol 80: 499-507, 2001.

Zhao WQ, Chen GH, Chen H, Pascale A, Ravindranath L, Quon MJ, and Alkon DL. Secretion of annexin II via activation of insulin receptor and insulin-like growth factor receptor. J Biol Chem 278: 4205-4215, 2003.

Zobiack N, Gerke V, and Rescher U. Complex formation and submembranous localization of annexin 2 and S100A10 in live HepG2 cells. FEBS Lett 500: 137-140, 2001.

作者单位：1 The Epithelial Transport Laboratory, Veteran‘s Affairs Greater Los Angeles Healthcare System (VISN 22), Sepulveda 91343; UCLA School of Medicine, Los Angeles, California 90024; and 2 Department of Anatomy and Cell Biology, Queen‘s University, Kingston, Ontario K7L 3N Canada