We have studied caveolin-1 expression and the frequency and distribution of typical invaginated caveolae as they are identified by electron microscopy in the polarized epithelial cell lines MDCK II and Caco-2. In wild-type MDCK II cells caveolin expression is high and more than 400 caveolae/mm filter were observed at the basolateral membrane. No caveolae were found at the apical surface. By contrast, wild-type Caco-2 cells do not express caveolin-1 and have extremely few, if any caveolae. Caco-2 cells were stably transfected with the gene for caveolin-1 in order to investigate if the formation of caveolae is polarized also in these cells. We have isolated Caco-2 clones expressing different levels of caveolin-1, where the level of expression varies from 10-100% of the endogenous level in MDCK II cells. Caveolin-1 expression in Caco-2 cells gives rise to a marked immunofluorescense labeling mainly at the lateral plasma membrane. By electron microscopy an increase from less than 4 caveolae/mm filter in wild-type Caco-2 cells to 21-76 caveolae/mm filter in Caco-2 clones transfected with caveolin-1 was revealed and these caveolae were exclusively localized to the basolateral membrane. Thus expression of heterologous caveolin-1 in Caco-2 cells leads to polarized formation of caveolae, but there is a lack of correlation between the amount of caveolin expressed in the cells and the number of caveolae, suggesting that factors in addition to caveolin are required for generation of caveolae.

Caveolae are small, 50-80 nm omega-shaped invaginations of the plasma membrane defined on the basis of their morphological appearance in electron micrographs (invaginated caveolae) (Devine et al., 1971; Forbes et al., 1979; Gabella, 1978; Parton, 1996; Parton and Simons, 1995; Parton et al., 1997). They are not present in all cell types and their abundance varies greatly among cell types (Parton, 1996). Despite extensive research their function is still not clear. Evidence has been presented that they are involved in a number of cellular processes such as signal transduction (Lisanti et al., 1994), calcium signalling (Fujimoto, 1993), potocytosis (Anderson et al., 1992) and endocytosis (Schnitzer et al., 1996). Furthermore, some glycosylphosphatidylinositol (GPI)-linked receptors like the urokinase plasminogen receptor (Stahl and Mueller, 1995), the folate receptor (Mayor et al., 1994; Rothberg et al., 1990; Turek et al., 1993) and the membrane spanning tissue factor (Sevinsky et al., 1996), a key component of the coagulation cascade, have been reported to be localized to caveolae or organized in microdomains associated with caveolae (Schnitzer et al., 1995).

A major structural component of caveolae is the membrane protein caveolin. Caveolin-1, which is found in a variety of tissues, was both identified as a major component of the vesicular transport system in the trans-Golgi network (TGN) (Kurzchalia et al., 1992) and as a structural component of the caveolar coat (Dupree et al., 1993; Rothberg et al., 1992). It is a 21 kDa integral membrane protein and is found as a homooligomer in the cell (Monier et al., 1995). However, newly synthesized caveolin-1 forms oligomers in the endoplasmic reticulum (ER) indicating that oligomerization is a characteristic of caveolin-1 rather than an indication of caveolae localization (Monier et al., 1995). Caveolin-1 is inserted in the plasma membrane as a hair pin structure with both the N and the C terminal facing the cytosol (Monier et al., 1995), and it is a cholesterol binding protein (Murata et al., 1995). Recently, two more caveolins have been identified, caveolin-2 (Scherer et al., 1996) which is expressed in the same cells as caveolin-1, and caveolin-3 (Tang et al., 1996), which is expressed in muscle.

Little is known about caveolae biogenesis, except that caveolin and cholesterol is required (Parton, 1996). Fra et al. (1995) have shown that introduction of canine caveolin-1 into human lymphocytes, which do not normally form caveolae, was sufficient for caveolar formation. Likewise, Engelman et al. (1997) recently found that transfection of human cell lines with human caveolin resulted in the formation of caveolae. However, it is unclear whether some factors are required in addition to caveolin. Such factors could be present in some but not in other cell types, and caveolin expression in itself might therefore not necessarily lead to caveolae formation. In support of this, transfection of insect cells with caveolin-1 resulted in accumulation of caveolin-1 positive vesicles without leading to morphologically defined caveolae at the plasma membrane (Li et al., 1996).

In polarized epithelia it is an open question whether caveolae formation is polarized. Since caveolin has been reported to be involved in transport from the TGN to the apical surface in polarized MDCK cells (Kurzchalia et al., 1992) and since GPI-anchored proteins, which are sorted to the apical surface of polarized epithelial cells, have been proposed to associate with caveolin (Lisanti et al., 1993; Sargiacomo et al., 1993; Zurzolo et al., 1994), it is often implied that caveolae must be present at this surface (Joliot et al., 1997; Lisanti et al., 1993).

In the present study we have therefore asked the following questions: (1) are caveolae localized apically, basolaterally or both in the widely used canine cell line MDCK II which normally expresses caveolin-1? (2) does canine caveolin-1 expression lead, as in human lymphocytes, to generation of caveolae in polarized human Caco-2 cells, which normally do not express caveolin and have no caveolae (Mirre et al., 1996)? and (3) are such caveolae formed at random or localized either at the apical or basolateral surface? We show here that caveolae are localized basolaterally in MDCK II cells, that caveolin expression in Caco-2 cells at a level comparable to that seen in MDCK II cells leads to formation of caveolae exclusively at the basolateral surface and that the number of newly formed caveolae is much lower than the number of caveolae in MDCK II cells. Furthermore, the amount of caveolin in the transfected Caco-2 clones did not correlate with their ability to form caveolae. We therefore suggest that formation of caveolae in polarized epithelia is a basolateral phenomenon and that factors in addition to caveolin are involved.

Cell culture

Caco-2 and MDCK II cells were grown in T25 flasks (Nunc, Roskilde, Denmark), on glass coverslips or on Transwell filters (Costar; pore size 0.4 μm, diameter 24.5 mm, cells seeded at a density of 106 per filter) in DMEM (Dulbecco’s modified Eagle’s MEM), 10% or 5% fetal calf serum, respectively, 2 mM glutamine, 100 units/ml penicillin, 100 μg/ml streptomycin. Caco-2 cells were further supplemented with nonessential amino acids. Transfected Caco-2 clones were also supplemented with 0.65 mg/ml G418 (Life Technologies).

For inhibition of protein kinase C, the Caco-2 cells were incubated in normal medium containing either 1 or 10 μM phorbol-12 myristate-13 acetate or 5 or 10 μM bisindolylmalemide for 24 hours before processing for electron microscopy (EM). Caco-2 cells were incubated for 3 days in growth medium containing 1.5 mM butyric acid to investigate whether differentiation could induce caveolae formation.

Construction of expression vectors

cDNA encoding caveolin/VIP21 isolated from MDCK II (Kurzchalia et al., 1992) inserted into the EcoRI-XhoI sites of pBluescript was a generous gift from K. Simons, EMBL, Heidelberg. The 890 bp EcoRI-XhoI fragment encoding the cDNA was excised and inserted into pUHD10-3 (obtained from H. Bujard; Gossen and Bujard, 1992) and excised as a EcoRI-BamHI fragment of similar size. This fragment was inserted into pTej4 or pTej8 resulting in pUV47 and pUV51, respectively (both pTej4 and pTej8 were generous gifts from T. E. Johansen; Johansen et al., 1990). pTej4 is an eukaryote expression vector where expression of the gene inserted in the multicloning site is driven by the strong, constitutive promotor, pUbc (Johansen et al., 1990). pTej8 is similar to pTej4, except that it also encodes the neomycin phosphotransferase gene, giving resistance to G418 (Johansen et al., 1990).

Caco-2 cells were either transfected with pUV47 alone or by cotransfection of pUV51 and pSV2-neo using the calcium precipitation method as described (Vogel et al., 1995). The transfected cells were selected using 0.65 mg/ml G418, and stable clones were isolated after 3 weeks using cloning cylinders.

Western blots

Cells were scraped off in PBS (phosphate buffer saline) and pelleted. The pellet was suspended in 1 volume PBS and 1 volume 2× sample buffer (Novex, San Diego) containing 100 mM DTT. The samples were heated to 100°C for 5 minutes, or incubated at 25°C for 30 minutes and subjected to standard SDS-PAGE either on 12% gels or on 4-12% Novex gradient gels. The proteins were immobilized on Hybond™ECL™ nitrocellulose membrane (Amersham, Buckinghamshire, UK) by semi-dry transfer, and processed for western blots. A polyclonal antibody against the N terminus of human caveolin (sc-894, Santa Cruz Biotechnology) was used as primary antibody (1:1,000 in 5% non-fat milk, 0.1% Tween-20, PBS). As secondary antibody, HRP-linked donkey anti-rabbit antibody (Amersham life science) was used. ECL™ (Amersham) was used as detection method.

Triton X-100 insolubility

Triton X-100 insolubility assay was performed as described by Danielsen (1995). Briefly, cells were lysed and incubated in 25 mM Hepes, 150 mM NaCl, pH 6.5, 1% Triton X-100 for 10 minutes on ice with frequent vortexing, and centrifuged at 48,000 g for 30 minutes. The pellet was solubilized in sample buffer, and equal volumes of solubilized pellet and supernatant were analysed by SDS-PAGE and western blotting as described.

Immunofluorescence

Cells grown in T25 flasks, on glass coverslips or on filters were washed in PBS and fixed in 2% formaldehyde in 0.1 M sodium phosphate buffer, pH 7.2, for 1 hour at room temperature (all subsequent incubations were performed at room temperature). The formaldehyde was quenched with 25 mM glycine in PBS, and the cells permeabilized with 0.2% saponin in PBS for 1 hour or by 0.1% Triton X-100 for 15 minutes. The cells were blocked for 15 minutes with 5% normal goat serum in PBS and incubated in the same buffer for one hour with the polyclonal anti-caveolin antibody sc-894 also used for western blots diluted 1:100. After 3 times 10 minute washes in PBS, the cells were incubated for 30 minutes with FITC-conjugated goat anti-rabbit antibody (Southern Biotechnology Associates, Inc. Birmingham, USA), diluted 1:50 in 5% normal goat serum in PBS. The cells were then washed 4 times 10 minutes in PBS and dried.

Electron microscopy

Confluent cultures of MDCK II and Caco-2 cells grown on filters were rinsed with PBS and fixed with 2% formaldehyde and 0.1% glutaraldehyde in 0.1 M sodium phosphate buffer, pH 7.2. Following fixation, the cells on filters were cut free of the plastic insert, postfixed in OsO4, contrasted en bloc with 1% uranyl acetate, dehydrated in a graded series of ethanols, and embedded in epon. Sections were cut perpendicular to the filter and collected on Formvar-coated mesh grids and examined in a Philips CM100 electron microscope. With respect to quantitative analysis, see Results.

Expression of caveolin-1 in Caco-2 cells

In agreement with previous studies (Mirre et al., 1996) we could not detect any caveolin-1 in wild-type (wt) Caco-2 cells, while MDCK II cells showed strong expression (Fig. 1). For transfection of Caco-2 cells we used 2 different plasmids expressing caveolin-1 cDNA isolated from MDCK II cells (Kurzchalia et al., 1992). The 2 plasmids both expressed caveolin-1 cDNA from the strong, constitutive promoter pUbc (Johansen et al., 1990). Thirteen of 36 isolated Caco-2 clones could be amplified. Nine of 13 analysed clones expressed detectable levels of caveolin. Seven of these clones are shown in Fig. 1. The different clones vary in the level of expression of caveolin-1 from 10% of the level seen in MDCK II cells to a level similar to that seen in MDCK II cells (Table 1). The recombinant caveolin has the same mobility as the native caveolin seen in MDCK II cells on denaturing acrylamide gels (Fig. 1).

Table 1.

Frequency of caveolae and relative amount of caveolin in the various cell lines and clones

Frequency of caveolae and relative amount of caveolin in the various cell lines and clones
Frequency of caveolae and relative amount of caveolin in the various cell lines and clones
Fig. 1.

Western blot of Caco-2 clones transfected with canine caveolin-1 cDNA. Equal amounts of protein (20 μg/lane) was loaded on a 4-12% gradient SDS-polyacrylamide gel. The proteins were transferred to a nitrocellulose membrane using semi-dry transfer. Caveolin-1 was visualised using ECL (Amersham). Lane 1: mock transfected Caco-2; lane 2: clone 33/2; lane 3: clone 33/4; lane 4: clone 50/1; lane 5: clone 50/2; lane 6: clone 50/3; lane 7: clone 50/4; lane 8: clone 50/5 and lane 9: MDCK II. Arrow, 21 kDa monomer.

Fig. 1.

Western blot of Caco-2 clones transfected with canine caveolin-1 cDNA. Equal amounts of protein (20 μg/lane) was loaded on a 4-12% gradient SDS-polyacrylamide gel. The proteins were transferred to a nitrocellulose membrane using semi-dry transfer. Caveolin-1 was visualised using ECL (Amersham). Lane 1: mock transfected Caco-2; lane 2: clone 33/2; lane 3: clone 33/4; lane 4: clone 50/1; lane 5: clone 50/2; lane 6: clone 50/3; lane 7: clone 50/4; lane 8: clone 50/5 and lane 9: MDCK II. Arrow, 21 kDa monomer.

Characterization of the caveolin-1 expressed in Caco-2 cells

Caveolin present in caveolae and in TGN transport vesicles has been shown to be insoluble in Triton X-100 at 0°C and it has been pointed out that this detergent insolubility probably reflects the biophysical properties of the protein (Fra et al., 1994; Kurzchalia et al., 1992; Mirre et al., 1996; Parton, 1996). It has also been shown that caveolin-1 is resistant to SDS at low temperatures (25°C), whereas it dissociates upon boiling (Monier et al., 1995). The caveolin-1 expressed by the highly expressing Caco-2 clone (33/4) and the low expressing Caco-2 clone (33/2) both turned out to be Triton X-100 insoluble (Fig. 2A) and form 200 kDa complexes on SDS-PAGE after incubation at 25°C (Fig. 2B) thus behaving like caveolin-1 from other sources.

Fig. 2.

Caveolin-1 expressed in Caco-2 cells is Triton X-100 resistant at 0°C (A) and forms SDS-resistant oligomers at 25°C (B). Equal amounts of protein were loaded on a 4-12% SDS-PAGE, immobilized on nitrocellulose and visualized with ECL (Amersham) using antibody sc-894 against caveolin-1. (A) Triton X-100 solubility of caveolin-1 at 0°C. Lanes 1 and 2 represent clone 33/4 (lane 1, soluble fraction; lane 2, insoluble fraction). Lanes 3 and 4 represent clone 33/2 (lane 3, soluble fraction; lane 4, insoluble fraction). Arrow: 21 kDa monomer. (B) SDS resistancy at 25°C. Cells were lysed in sample buffer containing 2% SDS and incubated either at 25°C for 30 minutes or at 100°C for 10 minutes. Lanes 1 and 2: clone 33/2 (lane 1 incubated at 100°C, lane 2 at 25°C). Lanes 3 and 4: clone 33/4 (lane 3 incubated at 100°C, lane 4 at 25°C). Arrow: 200 kDa homooligomer. The 21 kDa monomer appears at the bottom of the gel.

Fig. 2.

Caveolin-1 expressed in Caco-2 cells is Triton X-100 resistant at 0°C (A) and forms SDS-resistant oligomers at 25°C (B). Equal amounts of protein were loaded on a 4-12% SDS-PAGE, immobilized on nitrocellulose and visualized with ECL (Amersham) using antibody sc-894 against caveolin-1. (A) Triton X-100 solubility of caveolin-1 at 0°C. Lanes 1 and 2 represent clone 33/4 (lane 1, soluble fraction; lane 2, insoluble fraction). Lanes 3 and 4 represent clone 33/2 (lane 3, soluble fraction; lane 4, insoluble fraction). Arrow: 21 kDa monomer. (B) SDS resistancy at 25°C. Cells were lysed in sample buffer containing 2% SDS and incubated either at 25°C for 30 minutes or at 100°C for 10 minutes. Lanes 1 and 2: clone 33/2 (lane 1 incubated at 100°C, lane 2 at 25°C). Lanes 3 and 4: clone 33/4 (lane 3 incubated at 100°C, lane 4 at 25°C). Arrow: 200 kDa homooligomer. The 21 kDa monomer appears at the bottom of the gel.

We determined the localization of caveolin-1 in the Caco-2 cells by immunofluorescence microscopy using the sc-894 antibody which only seems to recognize caveolin in caveolae (Dupree et al., 1993). No caveolin could be detected in wt and mock transfected Caco-2 cells (Fig. 3A), confirming our observations made with western blotting (Fig. 1). In the transfected clones, labeling with sc-894 revealed a distinct, sometimes fine dotted fluorescence mainly of the lateral plasma membrane (Fig. 3B). No such labeling was seen when the focal plane was shifted to the top or bottom of the cell monolayer.

Fig. 3.

Immunofluorescense labeling of Caco-2 cells with the polyclonal antibody sc-894 to caveolae-associated caveolin-1. (A) Mock transfected Caco-2 cells. (B) Caveolin transfected Caco-2 clone 50/1 (focal plane in the middle of the cell monolayer). Note that caveolin in the transfected cells primarily appears as a fine dotted pattern associated with the lateral membrane. Bar, 10 μm.

Fig. 3.

Immunofluorescense labeling of Caco-2 cells with the polyclonal antibody sc-894 to caveolae-associated caveolin-1. (A) Mock transfected Caco-2 cells. (B) Caveolin transfected Caco-2 clone 50/1 (focal plane in the middle of the cell monolayer). Note that caveolin in the transfected cells primarily appears as a fine dotted pattern associated with the lateral membrane. Bar, 10 μm.

Taken together, these data indicate that a large fraction of the caveolin-1 expressed in the Caco-2 clones behaves as endogeneous caveolae-associated caveolin-1. However, to which extent the expressed caveolin-1 actually gives rise to formation of caveolae can only be determined by electron microscopy (EM).

Frequency and distribution of caveolae in Caco-2 and MDCK II cells

The frequency and distribution of caveolae were studied by EM of MDCK II and Caco-2 cells grown on filters (Figs 4, 5). The term caveolae as used here is defined strictly morphologically as characteristic 50-80 nm invaginations of the plasma membrane which are readily identified by EM (Fig. 5) (Parton and Simons, 1995; Parton 1996). It is clear that caveolae which may exist in a flattened (non-invaginated) form and therefore cannot be distinguished from the other portions of the plasma membrane (Smart et al., 1996) are not included. Moreover, for quantification we only included invaginated caveolae which clearly opened to the cell surface (Fig. 5). This is likely to give an underestimate of the total number of caveolae since some must appear as ‘free’ vesicles in a random, single section. Caveolae which do not open to the cell surface in a given section can be accounted for by fixation with Ruthenium Red (van Deurs et al., 1996), but this approach, which gave roughly 2.5 times higher counts, makes the precise definition of caveolae more difficult and was therefore not used (data not shown).

Fig. 4.

Low magnification electron micrograph of MDCK II cells grown on a filter. The length of this portion of epithelial monolayer is approx. 110 μm, corresponding to about 12 epithelial cells. In such a portion there is on average about 50 caveolae which open towards the basolateral surface; no caveolae are found on the apical surface. Bar, 5 μm.

Fig. 4.

Low magnification electron micrograph of MDCK II cells grown on a filter. The length of this portion of epithelial monolayer is approx. 110 μm, corresponding to about 12 epithelial cells. In such a portion there is on average about 50 caveolae which open towards the basolateral surface; no caveolae are found on the apical surface. Bar, 5 μm.

Fig. 5.

Electron micrographs showing examples of caveolae in Caco-2 cells expressing caveolin-1 (clone 33/2, A-E), and in MDCK II cells (F). The arrows indicate various examples of caveolae opening to the cell surface as they are defined for quantification. Accordingly, profiles like those labeled with asterisks (C and F) are not included although they may represent caveolae. Note that in Caco-2 cells the caveolae are associated with the lateral plasma membrane (Lm) rather than with the basal plasma membrane (Bm). (A and B) Two clathrin-coated pits (Cp) to show that caveolae and clathrin coated pits are readily distinguished in our preparations. F shows the basal part of an MDCK II cell; here the caveolae are associated with the basal plasma membrane (Bm). Fil, filter. Bar, 100 nm.

Fig. 5.

Electron micrographs showing examples of caveolae in Caco-2 cells expressing caveolin-1 (clone 33/2, A-E), and in MDCK II cells (F). The arrows indicate various examples of caveolae opening to the cell surface as they are defined for quantification. Accordingly, profiles like those labeled with asterisks (C and F) are not included although they may represent caveolae. Note that in Caco-2 cells the caveolae are associated with the lateral plasma membrane (Lm) rather than with the basal plasma membrane (Bm). (A and B) Two clathrin-coated pits (Cp) to show that caveolae and clathrin coated pits are readily distinguished in our preparations. F shows the basal part of an MDCK II cell; here the caveolae are associated with the basal plasma membrane (Bm). Fil, filter. Bar, 100 nm.

For each cell line/clone two different portions of a filter grown, confluent epithelial monolayer cut perpendicular to the filter were selected at low magnification (Fig. 4), and then caveolae were identified and counted along the basal cell membrane facing the filter as well as at the lateral and apical surfaces at high magnification (Fig. 5). Finally the filter portions, typically 100-200 μm of filter corresponding to the distance between the grid bars, were photographed at low magnification (×700) to measure the length of filter examined (Fig. 4). The number of caveolae per mm of filter was then calculated as the mean of the two filter portions (n=1). The quantification revealed that there was on average 433 caveolae per mm filter associated with the basolateral membrane in wt MDCK II cells (Table 1). The caveolae were predominantly (>70%) associated with the basal cell membrane facing the filter, although the lateral part of the basolateral membrane represents by far the largest surface area due to folding or interdigitations. Moreover, the caveolae were not evenly distributed but often appeared as patches with 3-10 caveolae (as judged from a single thin section) (Fig. 5F). Accordingly, long stretches of basal plasma membrane were devoid of caveolae. Finally it should be stressed that no caveolae were identified at the apical surface.

In wt Caco-2 cells, less than four caveolae-like structures per mm filter were observed (Table 1). This is in agreement with the lack of caveolin-1 expression in these cells. However, the Caco-2 clones transfected with caveolin had 21-76 caveolae/mm filter, and thus significantly more caveolae than the wt Caco-2 cells (Table 1 and Fig. 5A-E). Control experiments showed that mock transfection of control cells did not significantly change the low frequency of caveolae in wt Caco-2 cells.

Several intriguing observations were made in relation to caveolin expression and the occurrence of caveolae in the Caco-2 clones. First, even though the level of caveolin expression in some Caco-2 clones was comparable to that of wt MDCK II cells there was no correlation between the amount of expressed caveolin and the number of caveolae. On the contrary, it seems that the Caco-2 clones which have medium and low expression of caveolin (clones 33/2 and 50/1, Table 1), have more caveolae than clones expressing high levels of caveolin (33/4 and 50/5). Overall, the number of caveolae were six to ten times lower than in wt MDCK II cells (Table 1). Second, the caveolae in the Caco-2 clones were localized to the lateral membranes (Fig. 5B-E); with a few exceptions caveolae were not observed at the basal membrane facing the filter, in contrast to the situation in MDCK II cells. On the other hand, clathrin-coated pits were sometimes seen at the basal membrane. Third, caveolae were not observed at the apical surface of Caco-2 clones although there appeared to be space enough between the highly differentiated microvilli for possible caveole-formation and moreover, clathrin coated pits were frequently seen at the apical surface of Caco-2 cells (not shown).

Since caveolin may not only be associated with plasmalemmal caveolae but also with various vesicular and other membrane structures within the cytosol, including TGN elements (Joliot et al., 1997; Kurzchalia et al., 1992; Li et al., 1996; Parton et al., 1997), we were attentive to the possible existence of such membrane material in particular in the vicinity of the lateral plasma membrane in the transfected Caco-2 cells. However, no increase in vesicles and other membrane structures was noticed in any clone. The only way to distinguish wt Caco-2 cells from transfected Caco-2 clones in blind experiments was to carefully count caveolae.

Clone 33/4 expresses a lot of caveolin, little of which thus seems to participate in formation of caveolae (Table 1). It is therefore possible that caveolae biogenesis is repressed in the clone, since it has been shown that a number of factors influence caveolae formation. For instance, activation of protein kinase C-α (PKC-α) has been shown to disrupt caveolae formation (Smart et al., 1994, 1995). If caveolae biogenesis is repressed by elevated PKC activity, inhibition of PKC-α might lead to caveolae formation. We have therefore treated clone 33/4 with a PKC-α inhibitor, bisindolylmalemide, and with an activator, phorbol-12 myristate-13 acetate (TPA), since prolonged stimulation of PKC by TPA has been shown to downregulate PKC (Young et al., 1987). Inhibition or downregulation of PKC did not, however, affect caveolae biogenesis. Furthermore, incubation with serum-free medium to avoid growth factor-induced stimulation of PKC did not increase the number of caveolae.

In some cell types, the number of caveolae is affected by differentation of the cells. In adipocytes the number of caveolae is increased upon differentiation (Fan et al., 1983) and also cultured myoepithelial cells which are allowed to differentiate have numerous caveolae (Petersen et al., 1989). Since butyric acid induces enhanced differentiation of intestinal epithelium (Nathan et al., 1990; Stoddart et al., 1989) we tested whether butyric acid treatment leads to increased caveolae formation in a caveolin-1 expressing Caco-2 clone. However, incubation with butyric acid had no influence on caveolae formation.

These results clearly show that there is not necessarily a direct correlation between the caveolin level and the number of caveolae present on the plasma membrane, and importantly the data also demonstrate that there is a polarized distribution of caveolae in the two epithelial cell lines studied.

The most important result of the present study is that the formation of invaginated caveolae is polarized both in MDCK II cells and in Caco-2 cells transfected with caveolin. In MDCK II cells such caveolae are predominantly present at the basal membrane, although some are observed at the lateral membrane. In contrast, caveolae were almost exclusively found at the lateral membrane in the Caco-2 clones expressing caveolin-1. One can speculate whether this indicates that the basolateral membrane is really two distinct domains, the basal and the lateral membrane, with different functional characteristics, varying among epithelial cells.

The polarized distribution of caveolae is not caused by an unequal distribution of cholesterol in the apical and the basolateral membrane, since the molar percentage of cholesterol in these 2 membranes is the same (30%) (Simons and van Meer, 1988).

The formation of invaginated caveolae exclusively at the basolateral side of epithelial cells studied so far is interesting, because there is a high amount of glycolipids at the apical side of epithelial cells, and glycolipid rich domains may be associated with caveolin (Lisanti et al., 1993; Sargiacomo et al., 1993; Zurzolo et al., 1994). Moreover, caveolin-1 is present in transport vesicles from the TGN which contains the apically destined influenza haemagglutinin (Kurzchalia et al., 1992). It has therefore been implied that caveolae are present at the apical membrane of polarized cells although it has not been shown by EM (Joliot et al., 1997; Lisanti et al., 1994). Importantly, caveolin is also found in vesicular stomatitis virus G protein-positive basolaterally destined transport vesicles in MDCK II cells (Kurzchalia et al., 1992). Whether only caveolin associated with these vesicles or also caveolin originally directed to the apical surface is involved in basolateral caveolae formation is still not known. Thus, it seems that caveolin-1 is transported to the apical surface in MDCK II cells without this leading to caveolae formation.

The polarized formation of caveolae is also of interest in relation to the sorting of GPI-anchored proteins. For exampel the urokinase plasminogen activator receptor has been localized to caveolae (Stahl and Mueller, 1995) and is apically sorted in MDCK cells (Limongi et al., 1995). Since caveolae are present only at the basolateral membrane of polarized MDCK cells, the urokinase plasminogen activator receptor cannot be localized to caveolae in such cells. One can therefore question the physiological significance of the caveolar localization of the receptor. It has been reported that apical sorting of other GPI-anchored proteins is also linked to their localization in caveolae in polarized MDCK cells (Lisanti et al., 1993; Sargiacomo et al., 1993). This now seems less likely. However, it is possible that caveolin present in the TGN is involved in sorting, since the GPI-anchored fusion protein gD1-DAF (Zurzolo et al., 1994) and the urokinase-type plasminogen activator (Canipari et al., 1992) are sorted differently in MDCK cells, which express caveolin, and in Fisher rat thyroid cells or Caco-2 cells, which do not express caveolin. It would certainly be interesting to study protein sorting in Caco-2 cells with and without caveolin expression. In this study we used the canine cDNA for caveolin-1 since this molecule has previously been used to induce caveolae in human lymphocytes (Fra et al., 1995) and since we are here comparing the localization of caveolae in Caco-2 cells with that in the widely used canine cell line MDCK II. We find it unlikely, however, that transfection of Caco-2 cells with human caveolin cDNA might change the distribution of caveolae since there are only minor differences between canine and human caveolin-1 (Scherer et al., 1996; Tang et al., 1996).

We here show that expression of canine caveolin-1 generates caveolae in Caco-2 cells which normally do not form caveolae. This is in agreement with a recent report by Fra et al. (1995), who found formation of caveolae upon transfection of lymphocytes with caveolin-1. However, Fra et al. (1995) found a positive correlation between the level of caveolin expression and the number of caveolae upon transfection of lymphocytes with caveolin-1. In the present study, we show that in Caco-2 clones there is no correlation between the level of caveolin expression and the number of caveolae (Table 1). There are a number of possible explanations for this somewhat surprising result. Mutations in the cDNA are unlikely, since the number of caveolae in the clones expressing low levels of caveolin correlates with the caveolin level, indicating the expression of a fully functional protein (clone 33/2 and 50/1, Table 1).

Alternatively, a high number of caveolae may be toxic to Caco-2 cells, so only clones that either express moderate amounts of caveolin or express high amounts of caveolin that do not cause formation of caveolae could be amplified. Since caveolin-1 has no designated function, it is hard to test if the protein is active. However, the caveolin expressed in clone 33/4 has the correct mobility on SDS-PAGE, forms oligomers in vivo and is Triton X-100 insoluble, like caveolin in other cell types.

It is possible that caveolae formation in Caco-2 cells may require other proteins in addition to caveolin-1 and that newly synthesized caveolin in the absence of such proteins associates with intracellular (non-caveolar) membranes. This seems to be true for some other expression systems for caveolin-1, since expression of caveolin-1 in insect cells leads to a huge accumulation of caveolin-1 positive intracellular vesicles (Li et al., 1996). Furthermore, according to these authors the caveolin-1 positive membrane structures identified as caveolae in the caveolin-1 transfected lymphocytes (Fra et al., 1995) were also predominantly intracellular vesicles. However, we did not observe accumulation of vesicular and other membrane structures by EM of the caveolin-expressing Caco-2 cells. Moreover, immunofluorescence labeling of caveolin-1 in the Caco-2 clones transfected with caveolin-1 indicates that caveolin-1 is predominantly localized to the lateral plasma membrane.

Caveolae were originally identified on the basis of their structure as invaginations on the plasma membrane with a characteristic shape and size. It has later been proposed by different groups that caveolae are able to close and form vesicles (Rothberg et al., 1990; Schnitzer et al., 1996). However, when caveolae are closed they are indistinguishable from other vesicles which happen to be of the same size. Consequently, cryo-immunogold labelling has been used to identify caveolin-positive membrane structures. This may be a problem since transport vesicles from the TGN also contain caveolin-1. In previous transfection studies there has apparently been no differentiation between morphologically defined caveolae and caveolin-1 positive vesicles (Joliot et al., 1997; Li et al., 1996). We believe it is important for future studies to distinguish between caveolin-1 positive vesicles of unknown origin and function and morphologically well-defined caveolae if caveolar function and biogenesis is to be elucidated. It will also be interesting to investigate to which extent other polarized cells have a polarized distribution of caveolae.

We are grateful to Ulla Hjortenberg, Mette Ohlsen, Kirsten Pedersen and Keld Ottosen for their excellent technical assistance. Drs K. Simons and T. E. Johansen are thanked for providing plasmids and Dr. L. Vogel for help with transfections. The present study is supported by the Danish Cancer Society, the Danish Medical Research Council, the Novo-Nordic Foundation, the Human Frontier Science Programme, and a NATO Collaborative Research Grant (CRG 900517).

Anderson
,
R. G. W.
,
Kamen
,
B. A.
,
Rothberg
,
K. G.
and
Lacey
,
S. W.
(
1992
).
Potocytosis: Sequestration and transport of small molecules by caveolae
.
Science
255
,
410
411
.
Canipari
,
R.
,
Zurzolo
,
C.
,
Polistina
,
C.
,
Garbi
,
C.
,
Aloj
,
L.
,
Cali
,
G.
,
Gentile
,
R.
and
Nitsch
,
L.
(
1992
).
Polarized secretion of plasminogen activators by epithelial cell monolayers
.
Biochem. Biophys. Acta
1175
,
1
6
.
Danielsen
,
E. M.
(
1995
).
Involvement of detergent-insoluble complexes in the intracellular transport of intestinal brush border enzymes
.
Biochemistry
34
,
1596
1605
.
Devine
,
C. E.
,
Simpson
,
F. O.
and
Bertaud
,
W. S.
(
1971
).
Surface features of smooth muscle cells from the mesenteric artery and vas deferens
.
J. Cell Sci
.
8
,
427
443
.
Dupree
,
P.
,
Parton
,
R. G.
,
Raposo
,
G.
,
Kurzchalia
,
T. V.
and
Simons
,
K.
(
1993
).
Caveolae and sorting in the trans-Golgi network of epithelial cells
.
EMBO J
.
12
,
1597
1605
.
Engelman
,
J. A.
,
Wykoff
,
C. C.
,
Yasuhara
,
S.
,
Song
,
K. S.
,
Okamoto
,
T.
and
Lisanti
,
M. P.
(
1997
).
Recombinant expression of caveolin-1 in oncogenically transformed cells abrogates anchorage-independent growth
.
J. Biol. Chem
.
272
,
16374
16381
.
Fan
,
J. Y.
,
Carpentier
,
J.-L.
,
Obberghen
,
E. v.
,
Grunfeld
,
C.
,
Gorden
,
P.
and
Orci
,
L.
(
1983
).
Morphological changes of the 3T3-L1 fibroblast plasma membrane upon differentiation to the adipocyte form
.
J. Cell Sci
.
61
,
219
230
.
Forbes
,
M. S.
,
Rennels
,
M. L.
and
Nelson
,
E.
(
1979
).
Caveolar system and sarcoplasmic reticulum in conorary smooth muscle cells of the mouse
.
J. Ultrastruct. Res
.
67
,
325
339
.
Fra
,
A. M.
,
Williamson
,
E.
,
Simons
,
K.
and
Parton
,
R.
(
1994
).
Detergent- insoluble glycolipid microdomains in lymphocytes in the absence of caveolae
.
J. Biol. Chem
.
269
,
30745
30748
.
Fra
,
A. M.
,
Williamson
,
E.
,
Simons
,
K.
and
Parton
,
R. G.
(
1995
).
De novo formation of caveolae in lymphocytes by expression of VIP21-caveolin
.
Proc. Nat. Acad. Sci. USA
92
,
8655
8659
.
Fujimoto
,
T.
(
1993
).
Calcium pump of the plasma membrane is localized in caveolae
.
J. Cell Biol
.
120
,
1147
1157
.
Gabella
,
G.
(
1978
).
Inpocketings of the cell membrane (caveolae) in the rat myocardium
.
J. Ultrastruct. Res
.
65
,
135
147
.
Gossen
,
M.
and
Bujard
,
H.
(
1992
).
Tight control of gene expression in mammalian cells by tetracycline-responsive promoters
.
Proc. Nat. Acad. Sci. USA
89
,
5547
5551
.
Johansen
,
T. E.
,
Scholler
,
M. S.
,
Tolstoy
,
S.
and
Schwartz
,
T. W.
(
1990
).
Biosynthesis of peptide precursors and protease inhibitors using new constitutive and inducible eukaryotic expression vectors
.
FEBS Lett
.
267
,
289
294
.
Joliot
,
A.
,
Trembleau
,
A.
,
Raposo
,
G.
,
Calvet
,
S.
,
Volovitch
,
M.
and
Prochiantz
,
A.
(
1997
).
Association of engrailed homeoproteins with vesicles presenting caveolae-like properties
.
Development
124
,
1865
-
1875
.
Kurzchalia
,
T. V.
,
Dupree
,
P.
,
Parton
,
R. G.
,
Kellner
,
R.
,
Virta
,
H.
,
Lehnert
,
M.
and
Simons
,
K.
(
1992
).
VIP21, a 21-kD membrane protein is an integral component of trans-Golgi-network-derived transport vesicles
.
J. Cell Biol
.
118
,
1003
1014
.
Li
,
S.
,
Song
,
K. S.
,
Koh
,
S. S.
,
Kikuchi
,
A.
and
Lisanti
,
M. P.
(
1996
).
Baculovirus-based expression of mammalian caveolin in Sf21 insect cells
.
J. Biol.Chem
.
271
,
28647
28654
.
Limongi
,
P.
,
Resnati
,
M.
,
Hernandez-Marrero
,
L.
,
Cremona
,
O.
,
Blasi
,
F.
and
Fazioli
,
F.
(
1995
).
Biosynthesis and apical localization of the urokinase receptor in polarized MDCK epithelial cells
.
FEBS Lett
.
369
,
207
211
.
Lisanti
,
M. P.
,
Tang
,
Z.
and
Sargiacomo
,
M.
(
1993
).
Caveolin forms a hetero- oligomeric protein complex that interacts with an apical GPI-linked protein: implications for the biogenesis of caveolae
.
J. Cell Biol
.
123
,
595
604
.
Lisanti
,
M. P.
,
Scherer
,
P. E.
,
Tang
,
Z. L.
and
Sargiacomo
,
M.
(
1994
).
Caveolae, caveolin and caveolin-rich membrane domains: a signalling hypothesis
.
Trends Cell Biol
.
4
,
321
235
.
Mayor
,
S.
,
Rothberg
,
K.
and
Maxfield
,
F.
(
1994
).
Sequestration of GPI- anchored proteins in caveolae triggered by crosslinking
.
Science
264
,
1948
1951
.
Mirre
,
C.
,
Monlauzeur
,
L.
,
Garcia
,
M.
,
Delgrossi
,
M. H.
and
Bivic
,
A. L.
(
1996
).
Detergent-resistant membrane microdomains from Caco-2 cells do not contain caveolin
.
Am. J. Physiol
.
271
,
C887
894
.
Monier
,
S.
,
Parton
,
R. G.
,
Vogel
,
F.
,
Behlke
,
J.
,
Henske
,
A.
and
Kurzchalia
,
T. V.
(
1995
).
VIP21-caveolin, a membrane protein constituent of the caveolar coat, oligomerizes in vivo and in vitro
.
Mol. Biol. Cell
6
,
911
-
927
.
Murata
,
M.
,
Paranen
,
J.
,
Schreiner
,
R.
,
Wieland
,
F.
,
Kurzchalia
,
T.
and
Simons
,
K.
(
1995
).
VIP21/caveolin is a cholesterol-binding protein
.
Proc. Nat. Acad. Sci. USA
92
,
10339
10343
.
Nathan
,
D. F.
,
Burkhart
,
S. R.
and
Morin
,
M. J.
(
1990
).
Increased cell surface EGF receptor expression during the butyrate-induced differentiation of human HCT-116 colon tumor cell clones
.
Exp. Cell Res
.
190
,
76
84
.
Parton
,
R. G.
and
Simons
,
K.
(
1995
).
Digging into caveolae
.
Science
269
,
1398
1399
.
Parton
,
R. G.
(
1996
).
Caveolae and caveolins
.
Curr. Opin. Cell Biol
.
8
,
542
548
.
Parton
,
R. G.
,
Way
,
M.
,
Zorzi
,
N.
and
Stang
,
E.
(
1997
).
Caveolin-3 associates with developing t-tubules during muscle differentiation
.
J. Cell Biol
.
136
,
137
154
.
Petersen
,
O. W.
,
Hansen
,
S. H.
,
Laursen
,
I.
and
van Deurs
,
B.
(
1989
).
Effect of insulin on growth and expression of smooth muscle isoactin in human breast gland myoepithelial cells in a chemically defined culture system
.
Eur. J. Cell Biol
.
50
,
500
509
.
Rothberg
,
K. G.
,
Ying
,
Y.
,
Kolhouse
,
J. F.
,
Kamen
,
B. A.
and
Anderson
,
R. G. W.
(
1990
).
The glycophospholipid-linked folate receptor internalizes folate without endtering the clathrin-coated pit endocytic pathway
.
J. Cell Biol
.
110
,
637
649
.
Rothberg
,
K. G.
,
Heuser
,
J. E.
,
Donzell
,
W. C.
,
Ying
,
Y.-S.
,
Glenney
,
J. R.
and
Anderson
,
R. G. W.
(
1992
).
Caveolin, a protein component of caveolae membrane coats
.
Cell
68
,
673
682
.
Sargiacomo
,
M.
,
Sudol
,
M.
,
Tang
,
Z.
and
Lisanti
,
M. P.
(
1993
).
Signal transducing molecules and glycosyl-phosphatidylinositol-linked proteins form a caveolin-rich insoluble complex in MDCK cells
.
J. Cell Biol
.
122
,
789
807
.
Scherer
,
P. E.
,
Okamoto
,
T.
,
Chun
,
M.
,
Lodish
,
H. F.
and
Lisanti
,
M. P.
(
1996
).
Identification, sequence, and expression of caveolin-2 defines a caveolin gene family
.
Proc. Nat. Acad..Sci. USA
93
,
131
135
.
Schnitzer
,
J. E.
,
McIntosh
,
D. P.
,
Dvorak
,
A. M.
,
Liu
,
J.
and
Oh
,
P.
(
1995
).
Separation of caveolae from associated microdomains of GPI-anchored proteins
.
Science
269
,
1435
1439
.
Schnitzer
,
J. E.
,
Oh
,
P.
and
MacIntosh
,
D. P.
(
1996
).
Role of GTP hydrolysis in fission of caveolae directly from plasma membranes
.
Science
274
,
239
242
.
Sevinsky
,
J. R.
,
Rao
,
L. V. M.
and
Ruf
,
W.
(
1996
).
Ligand-induced protease receptor translocation into caveolae: A mechanism for regulating cell surface proteolysis of the tissue factor-dependent coagulation pathway
.
J. Cell Biol
.
133
,
293
304
.
Simons
,
K.
and
van Meer
,
G.
(
1988
).
Lipid sorting in epithelial cells
.
Biochemistry
27
,
6197
6202
.
Smart
,
E. J.
,
Foster
,
D. C.
,
Ying
,
Y.-S.
,
Kamen
,
B. A.
and
Anderson
,
R. G. W.
(
1994
).
Protein kinase C activators inhibit receptor-mediated potocytosis by preventing internalization of caveolae
.
J. Cell Biol
.
124
,
307
313
.
Smart
,
E. J.
,
Ying
,
Y.-S.
and
Anderson
,
R. G. W.
(
1995
).
Hormonal regulation of caveolae internalization
.
J. Cell Biol
.
131
,
929
938
.
Smart
,
E. J.
,
Mineo
,
C.
and
Anderson
,
R. G. W.
(
1996
).
Clustered folate receptors deliver 5-methyltetrahydrofolate to cytoplasm of MA104 cells
.
J. Cell Biol
.
134
,
1169
1177
.
Stahl
,
A.
and
Mueller
,
B. M.
(
1995
).
The urokinase-type plasminogen activator receptor, a GPI-linked protein, is localized in caveolae
.
J. Cell Biol
.
129
,
335
344
.
Stoddart
,
J. H.
,
Lane
,
M. A.
and
Niles
,
R. M.
(
1989
).
Sodium butyrate supresses the transforming activity of an activated N-ras oncogene in human colon carcinoma cells
.
Exp. Cell Res
.
184
,
16
27
.
Tang
,
Z.
,
Scherer
,
P. E.
,
Okamoto
,
T.
,
Song
,
K.
,
Chu
,
C.
,
Kohtz
,
D. S.
,
Nishimoto
,
I.
,
Lodish
,
H. F.
and
Lisanti
,
M. P.
(
1996
).
Molecular cloning of caveolin-3, a novel member of the caveolin gene family expressed predominantly in muscle
.
J. Biol. Chem
.
271
,
2255
2261
.
Turek
,
J. J.
,
Leamon
,
C. P.
and
Low
,
P. S.
(
1993
).
Endocytosis of folate- protein conjugates: ultrastructural localization in KB cells
.
J. Cell Sci
.
106
,
423
430
.
van Deurs
,
B.
,
von Bulow
,
F.
,
Vilhardt
,
F.
,
Holm
,
P. K.
and
Sandvig
,
K.
(
1996
).
Destabilization of plasma membrane structure by prevention of actin polymerization. Microtubule-dependent tubulation of the plasma membrame
.
J. Cell Sci
.
109
,
1655
1665
.
Vogel
,
L. K.
,
Noren
,
O.
and
Sjostrom
,
H.
(
1995
).
Transcytosis of aminopeptidase N in Caco-2 cells is mediated by a non-cytoplasmic signal
.
J. Biol. Chem
.
270
,
22933
22938
.
Young
,
S.
,
Parker
,
P. J.
,
Ullrich
,
A.
and
Stabel
,
S.
(
1987
).
Down-regulation of protein kinase C is due to an increased rate of degradation
.
Biochem. J
.
244
,
775
779
.
Zurzolo
,
C.
,
vant Hof
,
W.
,
van Meer
,
G.
and
Rodriguez-Boulan
,
E.
(
1994
).
VIP21/caveolin, glycosphingolipid clusters and the sorting of glycosylphosphatidylinositol-anchored proteins in epithelial cells
.
EMBO J
.
13
,
42
53
.