Journal of Cell Science, Supplem ent 17, 127-132 (1993)Printed in G reat Britain © The Company o f Biologists Limited 1993

127

The making of a tight junction

M. Cereijido, L. Gonzâlez-Mariscal, R. G. Contreras, J. M. Gallardo, R. Garcia-Villegas and J. ValdésCenter for Research and Advanced Studies, Apartado Postal 14-740, México, DF, 07000 México

SUMMARY

MDCK (epithelial cells from the dog kidney) plated at confluence, establish tight junctions in 12-15 hours through a process that requires protein synthesis, for­mation of a ring of actin filaments in close contact with the lateral membrane of the cells, calmodulin, and a Ca2+-dependent exocytic fusion of tight junction (TJ)- associated components. Monolayers incubated in the absence Ca2+ make no TJs. Yet, if Ca2+ is added under these circumstances, TJs are made with a faster kinet­ics. Ca2+ is needed mainly at a site located on the outer side of the cell membrane, where it activates uvomorulin and triggers the participation of the cellular components mentioned above, via G-proteins associated with phos-

pholipase C and protein kinase C. In principle, the sites of all these molecules and mechanisms involved in junc­tion formation may be where a variety of agents (hor­mones, drugs, metabolites) act to produce epithelia with a transepithelial electrical resistance (TER) ranging from 10 to 10,000 Q.cm2. This range may be also due to a variety of substances found in serum and in urine, that increase the TER in a reversible and dose-depen­dent manner.

Key words: tight junctions, epithelial cells, Ca2+, G-proteins, uvomorulin, phospholipase C, protein kinase C, calmodulin, urine extracts

INTRODUCTION

For a long while, roughly from the second half of the nine­teenth century to the middle of the present one, the obser­vation that epithelia mounted between two chambers with identical saline solutions maintain a spontaneous electrical potential appeared to be in violation of the laws of physics (Cereijido et al., 1989; Cereijido, 1992). However, in 1958 Koefoed-Johnson and Ussing put forward a model for the frog skin that accounted for such asymmetry, and gave rise to a successful period during which physiologists explained the movement of substances across epithelia, ranging from the sheep rumen to the human intestine, and from the cecropia midgut to the choroid plexus. The crucial features of these explanations were (a) tight junctions (TJs) that seal the interspace between epithelial cells, and (b) apical/baso- lateral polarity that gives rise to vectorial fluxes. By 1970, in spite of such accomplishments, the mechanisms that syn­thesize, assemble and seal TJs and generate apical/basolat- eral polarity were still unknown.

A suitable preparation was needed to learn about asym­metry, polarization and the formation of TJs. Oxender and Christensen (1959) devised a system consisting of Ehrlich ascites cells sandwiched between two Millipore filters, and mounted between two chambers, one with pyridoxal phos­phate and the other with an excess of K+ or alanine (Fig. 1). Although the cells were able to transport a net amount of amino acids from the first to the second chamber, the cells lacked an intrinsic asymmetry and junctional com­plexes, and remained randomly oriented within the multi­cellular layer.

We attempted another approach. We disassembled the epithelium of the frog skin with proteases and EGTA, made a suspension in which the cells had no junctions or polar­ity (Zylber et al., 1973, 1975; Rotunno et al., 1973) and re­plated them at confluence on Millipore filters, hoping that they would repolarize and restore TJs. Although the cells were able to keep their Na+, K+, Cl- or water balance for several hours, they where hard to culture, and did not polar­ize or establish TJs (Fig. 1).

Although epithelial cell lines are able to grow in vitro, nobody suspected that they could polarize and make TJs, until Leighton and his group (1969, 1970) attributed the blistering activity of monolayers of MDCK cells to trans­port and accumulation of fluid under the monolayer. There­fore, following this observation the next step was to cul­ture MDCK cells on permeable supports (Misfeld et al., 1976; Cereijido et al., 1978a,b; Rabito et al., 1978). When mounted as a flat sheet between two chambers, these mono­layers behave in several respects like natural epithelia. The procedure was soon adapted to form monolayers with cells derived from different epithelia (e.g. see Rabito and Ausiello, 1980; Handler et al., 1984) and endothelia (Shiv­ers et al., 1985).

TJ FORMATION IN THE MONOLAYER OF MDCK CELLS

We use mainly two protocols (Fig. 2). The first one con­sists of plating the cells at confluence on a permeable sup­port, allowing 20-60 minutes for attachment, and transfer-

128 M. Cereijido and others

~MpF

Fig. 1. History o f the efforts to develop an asym m etric ‘epithelium ’.(A) Ehrlich ascites cells contained betw een two filters (M pF), and m ounted betw een two asym m etric solutions, one containing pyridoxal phosphate and the other a high concentration of K+ or alanine.A lthough the preparation did show a net flux o f am ino acids (aa), cells w ere not able to develop TJs o r asym metry.(B) Cells dissociated from frog skin and plated on filter paper w ere able to m aintain their own w ater and salt balance for a few hours, but they attached too weakly and failed to

develop TJs and polarity. (C) M D C K cells attach, grow for several days, form TJs and polarize on non-perm eable supports (glass, plastic). The accum ulation o f fluid under the m onolayer develops blisters. On perm eable supports like a nylon cloth coated with collagen (D) or a M illipore filter (E) they form a m onolayer that behaves in m any respects like a natural epithelium .

Fig. 2. (A) D evelopm ent o f transepithelial electrical resistance (TER) across m onolayers o f M D C K cells. The slow curve corresponds to cells plated and cultured in NC m edium (1.8 mM Ca2+), and the fast one to cells plated in 1.8 mM Ca2+ but transferred 20-40 m inutes later to LC m edium (1-5 (iM ), cultured in this m edium until the 20th hour, and then sw itched to NC.(B) Intracellular concentration o f C a2+ as m easured w ith the probe in do-l/A M in cells incubated in C D M EM and in LC m edium . After transfer to NC the intracellular concentration o f C a2+ shows an initial ‘sp ike’, and the concentration rises thereafter with some oscillations towards steady values.

ring this support with the monolayer to fresh medium. Cells form a continuous monolayer, polarize and establish TJs, a process that can be gauged through the development of a TER (transepithelial electrical resistance), and reaches a maximum in 12-15 hours (Fig. 2). The second protocol (‘Ca-switch’) consists of allowing 20-60 minutes for cell attachment in Ca2+-containing medium (CDMEM or NC, 1.8 mM), and then transferring the monolayers to low-cal- cium medium (LC, 1-5 |iM). Twenty hours later, while monolayers incubated in normal calcium exhibit 200-300 Q.cm2, those left in LC have a negligible TER. If at this time these monolayers are transferred to NC, they develop

a TER with a much faster kinetics (Fig. 2) (Gonzalez- Mariscal et al., 1985). For reviews of the properties of these monolayers see Cereijido (1992) and Schneeberger and Lynch (1992).

Using these approaches, it was found that, in order to make TJs, newly plated cells do not require synthesis of RNA, but do require synthesis of proteins, as the use of cycloheximide or puromycin during the first 4-5 hours after plating blocks the development of TER (Cereijido et al., 1978b, 1981). This synthesis may proceed in the absence of cell-cell contacts and of Ca2+ in the bathing media. Accordingly, the development of TER during the Ca-switch is not inhibited by cycloheximide (Gonzalez-Mariscal et al., 1985).

Although several peptides (ZO-1, ZO-2, cingulin, etc.) were found in close and specific association with the TJs (see Anderson and Stevenson, 1992), no information is available on whether the strands of the TJ that appear as ridges in the P face of freeze-fracture replicas, are them­selves made up of proteins. In fact, a model in which these strands consist of inverted cylindrical micelles made of lipids was put forward by Pinto da Silva and Kachar (1982). However, models comprising pure lipidic micelles do not seem to meet some expectations regarding the diffusion of lipid probes along the surface of the membrane (van Meer et al., 1986).

Observations performed with transmission electron microscopy indicate that junctional components are addressed polarizedly to the apical/basolateral limit (Vega- Salas et al., 1988). This possibility is in keeping with the observations made with freeze-fracture replicas, revealing that upon switching to Ca2+, strands start to develop through the formation of isolated segments that occupy from the beginning the place that they will fill in the mature TJ (Gonzalez-Mariscal et al., 1985).

The observation that cytochalasin B, but not colchicine, prevents junction formation and opens already sealed TJs in a few hours, is taken to indicate that microfilaments but not microtubules are involved injunction formation. In fact, while microfilaments form a continuous ring under the lat­eral plasma membrane, in close contact with the zonula adherens, microtubules are distributed mainly in the vicin-

The making of a tight junction 129

2000£a.

<ocOD2UJs

Table 1. Effect of different cations on tight junction formation during a Ca2+-switch

p<O.OOI

-CDMEM-

20 25

TIME (hours)

Fig. 3. Membrane area of MDCK cells as measured by the membrane capacitance method during the Ca-switch procedure. Newly plated cells have a small area (first point on the left); 20 hours later the area increases in spite o f the low extracellular (1-5 |j.M) and intracellular (-20 nM) concentration of Ca2+. After the Ca-switch cells increase their surface by 22%, coincident with the process of junction formation. Prevention of this increase with 25 (xM chloroquine blocks the development of TJs. (Taken, with kind permission, from Gonzalez-Mariscal et al. (1990).)

ity of the nucleus (Meza et al., 1980, 1982). The possibil­ity exists that actin filaments form a scaffold that helps the cells to adapt their borders to each other so TJs can be established and sealed, and lead the peptide-containing vesi­cles to the points where junctions are to be assembled, etc. In the absence of Ca2+, vesicles containing TJ-associated peptides are retained in the cytoplasm. Upon switching the monolayers to Ca2+ these vesicles fuse to the plasma mem­brane, along with two other populations of vesicles that are vectorialy addressed and fused to the apical (Vega-Salas et al., 1988) or to the basolateral domain (Contreras et al., 1989a,b), a process that increases the surface area by some 22% (Fig. 3), as measured by the ‘whole cell clamp’ method (González-Mariscal et al., 1990).

THE ROLE OF CALCIUM

Ca2+, besides of triggering the assembly and sealing of the TJ (Fig. 2), is also needed for maintaining the sealed state, as junctions may be opened and resealed by removal and restoration of this ion (Sedar and Forte, 1964; Meldolesi et al., 1978; Cereijido et al., 1978a; Martinez-Palomo et al., 1980; Gumbiner and Simons, 1986).

We investigated the ability of Ca2+ to trigger junction formation in the presence of ions and drugs like La3+ and verapamil that inhibit its penetration into the cytoplasm, as indicated by the measurement of 45Ca influx and the use of intracellular Ca2+ probes (Gonzalez-Mariscal et al., 1990; Contreras et al., 1992a,b). As shown in Table 1, only Ca2+ among the ions tested is able to promote the development of TER. This was taken as an indication that Ca2+ acts on an extracellular site. La3+ for instance, is able to block Ca2+ penetration but does not substitute for this ion in trigger­ing junction formation. Interestingly, Cd2+ is able to block

CationsConcentration

(mM)TER

(Q ■ cm2) n

LC mediumNone 2±10 18Ca* 1.8 676+48 24Mg 2.0 6±4 7Ba* 1.0 3±1 6Sr* 2.0 37±13 7Mn* 1.0 3±0 6Cd* 1.0 1±6 6La* 1.0 3±1 6

NC bufferCa+Mg 694+131 13Ca+Ba 603±112 13Ca+Sr 623+121 13Ca+La 425±30 6Ca+Mn 186±57 13Ca+Cd 2±1 13

Values are means ± s.e.; n, no. of observations. Cells were plated on nitrocellulose filters in NC medium for 1 hour and then washed 3 times with phosphate-buffered saline without Ca2+ and transferred to low Ca2+ (LC) medium. After 20 hours medium was changed to either LC medium or NC buffer containing the indicated multivalent cations. NC buffer composition was (in mM): 140 NaCl, 5 KC1, 10 dextrose, 1.8 CaCh, and 20 Tris-HCl, pH 7.4, at room temperature. (Taken, with kind permission, from Contreras et al., 1992.)

*Mg was present at 1.0 mM.

Ca2+ penetration and sealing, but fails to trigger junction formation by itself. These results were interpreted in terms of two sites with different affinities: one used by Ca2+ to penetrate into the cytoplasm, where other multivalent ions may compete; and a second one, where this ion triggers junction formation, that has a much higher Ca2+ affinity.

In keeping with this interpretation, we found that using 0.1 mM Ca2+ to trigger the Ca-switch, elicits junction for­mation without a detectable increase in intracellular Ca2+ (Gonzalez-Mariscal et al., 1990).

These studies were done with indo-l/AM, measuring the overall Ca2+ concentration in the cytoplasm. Niggam et al. (1992) used instead a digital imaging procedure, and found some indications that this ion may accumulate in a sub- membranal intracellular compartment, small enough as to escape detection by the methods that we have used.

THE NEED OF A TRANSDUCTING MECHANISM

As shown by Gumbiner and Simons (1987), junction for­mation requires cell-cell attachment provided by Ca2+- dependent uvomorulin molecules, located in the lateral membrane of MDCK cells. Therefore, uvomorulin may constitute one of the extracellular sites where Ca2+ acts to promote the making of TJs. However, as mentioned above, Ca2+ touches off a series of intracellular events that result in the assembly and sealing of TJs, suggesting the existence of a mechanism that would convey the information from the extracellular site(s) to those involved in the rearrange­ment of actin filaments, exocytic fusion, assembly of the strands, etc.

Looking for this transducting mechanisms, we found that

130 M. Cereijido and others

G PROT PLC PCK CaM [cA M P ]

1 6 0 0

1 4 0 0

1200

£ 1000 (J

3 8 0 0

cry 6oo

4 0 0

200

0 i a _ U û QoO

junction formation may be stimulated by pertussis toxin (PTX) and prevented by AIF3 and carbamil choline, i.e. by

Fig. 4. Filled colum ns: values o f TER during a control C a-switch from 5 |lM to 1.8 m M in 5 hours. The rest o f the colum ns correspond to sim ilar C a-sw itches m ade in the presence o f 14 ng ml-1 PTX, which inhibits, and 2 mM AIF3, w hich activates the action o f inhibitory G proteins; 110 |lM neom ycin, w hich inhibits, and 2 |lM TRH. w hich stimulates PLC; 50 |aM H7, which inhibits, and 100 (ig m f 1 d iC 8 , w hich stim ulates PKC; 25 ¡aM TFPZ, w hich inhibits calm odulin; 2.5 m M dB-cA M P (a perm eable analog o f cAM P), 120 |aM forskolin (a stim ulator o f adenylate cyclase) and 120 |uM IBM X (a phosphodiesterase inhibitor). See text for abbreviations.

blockers and enhancers of an inhibitory G protein, respec­tively (Fig. 4). This is further suggested by the fact that GTPyS blocks the development of a cell border pattern of ZO-1 peptides, as observed by immunofluorescence in MDCK cells permeabilized with digitonin. At least two dif­ferent G proteins may participate in the formation of TJs: (1) an inhibitory G protein modulating phospholipase C (PLC) and protein kinase C (PKC) probably participate in these processes, because activation of PLC by thyrotropin- 1-releasing hormone (TRH) increases TER, and its inhibi­tion by neomycin blocks the development of this resistance. Also, 1,2-dioctanoylglycerol (diC8), an activator of PKC,

PARACELLULARPATHWAY

Fig. 5. A highly schematic view o f the process o f junction form ation during the Ca-switch. (A) Cells incubated in Ca2+-free m edium have a cytoplasm ic vesicular com partm ent (VC) where junctional com ponents m ight be stored. Extracellular C a2+ activates a uvom orulin-like m olecule and perm its attachm ent to neighboring cells. Cell-cell contact w ould stim ulate PLC that is connected to m em brane receptors via two G proteins and convert phosphatidylinositol (PIP2) into inositol trisphosphate (IP3) and diacylglycerol (DAG). IP3 m obilizes C a2+ from an internal reservoir (IR), D-m yo-inositol 4-m onophosphate (IP4) decreases Ca2+ perm eability, and DA G activates protein k inase C (PKC). (B) This cascade o f reactions provokes phosphorylation (P) and incorporation o f junctional com ponents through exocytic fusion of the VC and activation o f calm odulin (CaM ). (C) Activation of CaM causes actin filaments to form into a continuous ring that circles the cells (represented by two groups o f filled circles); the paracellular space is sealed, and TER develops.

The making of a tight junction 131

stimulates TER development, while polymyxin B and l-(5- isoquinoline sulfonyl)-2-methyl-piperazine dihydrochloride (H7), which inhibits this enzyme, abolish TER (Baida et al., 1991). In turn, freeze-fracture electron microscopy reveals that diC8 triggers the formation of junctional strands. (2) Another G protein that may also participate in junctional events, is the one modulating the reactions involving adenylate cyclase. This stems from the fact that A,6,0 2 -dibutyryladenosine 3'-5'-cyclic monophosphate (dB- cAMP) a permeable analogue of cAMP, forskolin that stim­ulates adenylate cyclase, and 3-isobutyl- 1-methylxanthine (IBMX), which inhibits phosphodiesterase, produce signif­icant decreases in TER (Fig. 4).

Junction formation may be prevented by the use of tri­fluoperazine (TFPZ) and calmidazoline, two inhibitors of calmodulin (CaM) (Baida et al., 1991). This is a puzzling observation because, whereas activation of PLC and PKC may be elicited from an extracellular site through G pro­teins, calmodulin is an intracellular molecule that requires Ca2+, i.e. the mere information that this ion is present on the extracellular side does not suffice. This opens up at least two possibilities: (1) the signal transduced through the G proteins would cause the release of Ca2+ from an internal reservoir, presumably by the IP3 split from PIP2; or (2), as suggested by the observations of Niggam et al. (1992), there may be a small and localized increase in Ca2+ concentra­tion that would escape detection by most methods. Fig. 5 offers a very schematic summary of these results.

TER VARIES IN DIFFERENT EPITHELIA BY OVER THREE ORDERS OF MAGNITUDE

As the nephron progresses from the proximal to the col­lecting duct, the value of TER across the epithelia forming its walls increases by two orders of magnitude, a feature paralleled by an increase in the number of strands in the tight junctions (Claude and Goodenough, 1973). We inves­tigated whether the increase in TER may be attributed to a substance that, traveling with the filtrate through the lumen, would be finally eliminated in the urine. Accordingly, we prepared a dialyzed and lyophilized extract of urine (DLU) of normal dogs that, when applied to the basolateral (Fig. 6, 2nd column) or to the apical side (Fig. 6, 3rd column) of the MDCK monolayer, enhances TER (Gallardo et al., 1993). This effect is dose-dependent (not shown). When applied simultaneously to both sides of the monolayer the effect is significatively smaller (Fig. 6, 6th column), sug­gesting that DLU also has a TER-depressing component that acts mainly from the apical side. Only this second com­ponent is inactivated by heat, so that when DLU is boiled for 10 minutes before addition to both sides, the TER- enhancing effect prevails and reaches the level achieved with basolateral applications only (Fig. 6, 6th column). The effects of DLU can be completely abolished by incubation with protease type I (Fig. 6, 4th column). Since in its tran­sit from glomerulus to collecting duct the filtrate is con­centrated 100- to 200-fold, it is suggested that the TER- enhancing component plays a major role in the making of distal segments with tighter paracellular pathways.

In summary, the study of the TJ over more than a cen-

7 00

600

M5 00

£o 400

£

cn 300Ldf—

200

100

0

T

I

1

T

JL

CONT BL AP PROT BOTH

Fig. 6. Effects o f a dialyzed and lyophilized extract o f dog urine (DLU) on the value o f T ER in m onolayers o f M DCK cells. The effect is concentration dependent, but here is only shown at 30% DLU/70% DM EM . Filled colum n corresponds to control; (2nd) DLU added for 20 hours to the basolateral side; (3rd) to the apical; (4th) to the basolateral side after treatm ent with protease type I; (5th) to both sides; and (6th) to both sides after boiling the DLU solution for 10 minutes.

tury has shown that this structure is by no means a mere ‘terminal bar’, a ‘tight seal’, or a static structure. It may confer a TER of only 10 Q.cm2, as in the case of the prox­imal tubule of the kidney, or several thousands, as in the case of the mucosa of the urinary bladder, and can modify the permeability of an epithelium in response to physio­logical requirements and pharmacological challenges. It is expected that many of the factors involved in the forma­tion of a TJ (Fig. 5) will one day explain the wide range of TERs as well as the TJ’s dynamic behaviour.

W e thank the N ational Research Council o f M éxico (CONA- CYT) and the M exican Health Foundation for their econom ic sup­port.

REFERENCES

Anderson, J. M. and Stevenson, B. R. (1992). The molecular structure of the tight junction. In Tight Junctions (ed. M. Cereijido), pp. 77-90. CRC Press, Boca Raton, Florida.

Balda, M. S., Gonzàlez-Mariscal, L., Contreras, R. G., Macias-Silva, M. Torres-Marquez, M. E., Gareia-Sains, J. A. and Cereijido, M. (1991 ). Participation of a G protein modulation system in the assembly and sealing of tight junctions. J. Membr. Biol. 122, 193-202.

Cereijido, M. (1992). Tight Junctions. CRC Press, Boca Raton, Florida.Cereijido, M. and Martinez-Palomo, A. (1981). Occluding junctions in

cultured epithelial monolayers. Amer. J. Physiol. 240, C96-C102.Cereijido, M., Ponce, A. and Gonzàlez-Mariscal, L. (1989). Tight

junctions and apical/basolateral polarity. J. Membr. Biol. 110, 1-9.Cereijido, M., Robbins, E. S., Dolan, W. J., Rotunno, C. A. and Sabatini,

D. D. (1978a). Polarized monolayers formed by epithelial cells on a permeable and translucent support. J. Cell Biol. 77, 853-880.

Cereijido, M., Rotunno, C. A., Robbins, E. S. and Sabatini, D. D. ( 1978b). Polarized epithelial membranes produced in vitro. In Membrane Transport Processes (ed. J. F. Floffman), pp. 433-461. Raven Press, New York.

Claude, P. and Goodenough, D. A. (1973). Fracture faces of zonulae

132 M. Cereijido and othersoccludentes from ‘tight’ and ‘leaky’ epithelia. J. Cell Biol. 58, 390­400.

Contreras, R. G., Avila, G., Gutiérrez, C., Bolivar, J. J., González- Mariscal, L., Darzon, A., Beaty, G., Rodríguez-Boulán, E. and Cereijido, M. (1989a). Repolarization of Na-K-Pumps during establishment of epithelial monolayers. Amer. J. Physiol. 257, C896- C905.

Contreras, R. G., González-Mariscal, L., Balda, M. S., Garcia-Villegas, M. R. and Cereijido, M. (1992b). The role of calcium in the making of a transporting epithelium. News Physiol. Sci. 7, 105-108.

Contreras, R. G., Miller, J., Zamora, M., González-Mariscal, L. and Cereijido, M. (1992). The interaction of calcium with the plasma membrane of epithelial (MDCK) cells during junction formation. Amer. J. Physiol. 263, C313-C318.

Gallardo, J. M., Hernández, J. M., Garcia-Villegas, M. R., Contreras, R. G., González-Mariscal, L. and Cereijido, M. (1993). A urine extract that enhances the electrical resistance across monolayers of epithelial cells (MDCK). Amer. J. Physiol.

González-Mariscal, L., Chávez de Ramírez, B. and Cereijido, M. (1985). Tight junction formation in cultured epithelial cells (MDCK). J. Membr. Biol. 86, 113-125.

González-Mariscal, L., Contreras, R. G., Bolívar, J. J., Ponce, A., Chávez de Ramírez, B. and Cereijido, M. (1990). The role of calcium in tight junction formation between epithelial cells. Amer. J. Physiol. 259, C978-C986. '

Gumbiner, B. and Simons, K. (1986). A functional assay for proteins involved in establishing an epithelial occluding barrier: Identification of an uvomorulin-like peptide. J. Cell Biol. 102,457-468.

Gumbiner, B. and Simons, K. (1987). The role of uvomorulin in the formation of epithelial occluding junctions. In Junctional Complexes o f Epithelial Cells (ed. G. Bock and S. Clark), pp. 168-186. J. Wiley, Chichester.

Handler, J. S., Preston, A. S. and Steele, R. E. (1984). Factors affecting the differentiation of epithelial transport and responsiveness to hormones. Fed. Proc. Fed. Amer. Socs Exp. Biol. 43 ,2221-2224.

Koefoed-Johnson, V. and Ussing, H. H. (1958). The nature of the frog skin potential. Acta Physiol. Scand. 42, 298-308.

Leighton, J. L., Brada, Z., Estes, L. W. and Justh, G. (1969). Secretory activity and oncogenicity of a cell line (MDCK) derived from canine kidney. Science 163,472-473.

Leighton, J. L., Estes, L. W., Mansukhani, S. and Brada, Z. (1970). A cell line derived from a normal dog kidney (MDCK) exhibiting qualities of papillary adenocarcinoma and renal tubular epithelium. Cancer 26, 1022-1028.

Martínez-Palomo, A., Meza, I., Beaty, G. and Cereijido, M. (1980). Experimental modulation of occluding junctions in a cultured transporting epithelium. J. Cell Biol. 8 7 ,736-745.

Meldolesi, J., Castiglioni, G., Parma, R., Nassivera, N. and De-Camilli,P. (1978). Ca++-dependent disassembly and reassembly of occluding junctions in guinea pig pancreatic acinar cells. J. Cell Biol. 7 9 ,156-172.

Meza, I., Ibarra, G., Sabanero, M., Martinez-Palomo, A. and Cereijido,M. (1980). Occluding junctions and cytoskeletal components in a cultured transporting epithelium. J. Cell Biol. 87 ,746-754.

Meza, I., Sabanero, M., Stefani, E. and Cereijido, M. (1982). Occluding junctions in MDCK cells: Modulation of transepithelial permeability by the cytoskeleton. J. Cell. Biochem. 18,407-421.

Misfeld, D. S., Hamamoto, S. T. and Pitelka, D. R. (1976). Transepithelial transport in cell culture. Proc. Nat. Acad. Sci. USA 73,1212-1216.

Niggam, S. K., Rodriguez-Boulan, E. and Silver, R. B. (1992). Changes in intracellular calcium during the development of epithelial polarity and junctions. Proc. Nat. Acad. Sci. USA 89, 6162-6166.

Oxender, D. L. and Christensen, H. N. (1959). Transcellular concentration as a consequence of intracellular accumulation. J. Biol. Chem. 234, 2321-2324.

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

Rabito, C. A. and Ausiello, D. A. (1980). Na-dependent sugar transport in a cultured epithelial cell line from pig kidney. Amer. J. Physiol. 250, F734-F743.

Rabito, C. A., Tchao, R. Valentich, J. and Leighton, J. (1978). Distribution and characteristics of the occluding junctions in a monolayer of a cell line (MDCK) derived from a canine kidney. J. Membr. Biol. 43, 351-365.

Rotunno, C. A., Zylber, E. and Cereijido, M. (1973). Ion and water balance in the epithelium of the abdominal skin of the frog Leptodactylus ocellatus. J. Membr. Biol. 13, 217-232.

Schneeberger, E. and Lynch, R. D. (1992). Structure, function, and regulation of cellular tight junctions. Amer. J. Physiol. 262, L647- L661.

Sedar, A. W. and Forte, J. G. (1964). Effects of calcium depletion on the junctional complex between oxyntic cells of gastric glands. J. Cell Biol. 2 2 ,173-188.

Shivers, R., Bowman, P. and Martin, K. (1985). A model for the novo synthesis and assembly of tight intercellular junctions. Ultraestructural correlates and experimental verification of the model revealed by freeze- fracture. Tissue & Cell 17,417-440.

van Meer, G., Gumbiner, B. and Simons, K. (1986). The tight junction does not allow lipid molecules to diffuse from one epithelial cell to the next. Nature 322,639-641.

Vega-Salas, D. E., Salas, P. J. I. and Rodriguez-Boulan, E. (1988). Exocytosis of a vacuolar apical comparment (VAC): a cell-cell contact controlled mechanism for the establishment of the apical plasma membrane domain in epithelial cells. J. Cell Biol. 107, 1717­1728.

Zylber, E., Rotunno, C. A. and Cereijido, M. (1973). Ion and water balance in the isolated epithelial cells of abdominal skin of the frog Leptodatylus ocellatus. J. Membr. Biol. 1 3 ,199-216.

Zylber, E., Rotunno, C. A. and Cereijido, M. (1975). Ionic fluxes in isolated epithelial cells of the abdominal skin of the frog Leptodatylus ocellatus. J. Membr. Biol. 22,265-284.