INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

VASCULAR ENDOTHELIAL CELLS recognize shear stress through unknown mechanosensing system(s) and respond both acutely and chronically by producing autocrine and paracrine factors (7). Through these endothelial responses, shear stress controls vascular tone, vessel wall remodeling, binding of blood cells to endothelium, and hemostasis (7). Shear stress selectively and differentially regulates expression of many genes that are important in the pathophysiology of vessel wall function (see Refs. 4 and 9 for reviews). Furthermore, a conserved cis-acting shear stress-response element has been identified in many shear-sensitive genes including platelet-derived growth factor-B, intracellular adhesion molecule-1, tissue plasminogen activator, and transforming growth factor -1, suggesting its broad implication in shear-dependent regulation of gene expression (4, 9, 20, 29). An additional cis-acting element, the phorbol ester 12-O-tetradecanoylphorbol 13-acetate-responsive element, has been found in the monocyte chemoattractant protein-1 gene (37). Shear stress also transiently activates transcription factors [nuclear factor-B (NF-B), AP-1, and early growth response-1] and immediate-early response genes (c-fos, c-jun, and c-myc) that are involved in the regulation of shear-dependent gene expression (13, 20, 21, 36, 37). Induction of specific genes by shear stress in endothelial cells is mediated by activation of mechanosensitive signaling pathways that include at least three members of the mitogen-activated protein (MAP) kinase family, extracellular signal-regulated kinase (ERK1/2), c-Jun NH₂-terminal kinase (JNK), and Big MAP kinase 1 (also called ERK5) (2, 18, 23, 36, 43, 47). MAP kinases are important signaling components linking extracellular stimuli to cellular responses such as cell growth, death, differentiation, and metabolic regulation (5, 19).

Shear stress activates ERK in a rapid and transient manner (maximum by 5 min and returning to basal levels by 30 min of shear exposure), whereas JNK activation occurs over a much slower and prolonged time course (requiring at least 30 min and returning to basal levels after 1 day of shear exposure) (18). Shear stress activates the two MAP kinases by distinct signaling pathways: activation of ERK is mediated by mechanisms involving G_i-2, protein kinases [Src, focal adhesion kinase (FAK), and protein kinase C-], and Ras, whereas JNK activation requires G/, phosphatidylinositol-3-kinase-, tyrosine kinases (Src and FAK), and Ras (11, 14, 16, 18, 22, 42). How do endothelial cells ensure the specific activation of each pathway when sharing the same common signaling molecules (e.g., Src, FAK, and Ras)? A likely hypothesis is that spatial compartmentalization of signaling molecules into microdomains provides the mechanism for the differential activation of ERK and JNK. In support of this hypothesis, we (28) and Rizzo et al. (30) have independently shown that the cholesterol-sensitive microdomains in the plasma membrane play a critical role in mechanosensitive activation of ERK, but not JNK, in bovine aortic endothelial cells (BAEC) and in perfused rat lung endothelium (28, 30). At present the underlying molecular mechanisms and morphological understanding by which cholesterol-sensitive microdomains segregate signaling molecules and elicit selective regulation of the MAP kinases in response to shear stress are not known. However, caveolae and caveolae-like domains including "raft" and "glycosphingolipid signaling domains" are the likely candidates (1, 15, 27, 30, 38).

Caveolae are noncoated micropatches of the plasma membrane with a variety of shapes (flat, invaginated, and tubular) and are found in most cell types including endothelial cells, fibroblasts, smooth muscle cells, and adipocytes (1, 27). Caveolae serve at least two different functions: 1) transport of large and small molecules (transcytosis of macromolecules and potocytosis of ions and folate) and 2) transmembrane signaling microdomains (1, 27). Caveolae are enriched with cholesterol, glycosphingolipids, lipid-anchored signaling molecules, and caveolin (1, 27).

Caveolin, a 21- to 24-kDa membrane protein, is a principal component of caveolae and also binds directly to cholesterol and interacts with such signaling molecules as G protein -subunits, Ras, Src family kinases, and the endothelial form of nitric oxide synthase (eNOS) (8, 25, 27). Three members (caveolin-1, -2, and -3), including two isoforms (caveolin-1 and -1), of the caveolin gene family in mammalian cells have been cloned (32, 33, 40). Caveolin-1 is abundantly expressed in endothelial cells (35), whereas caveolin-3 is expressed only in muscle cells (40). The tissue distribution of caveolin-2 has been shown to be similar to that of caveolin-1 (32). Caveolins are composed of NH₂- and COOH-terminal domains linked by a hairpinlike membrane-spanning domain. Therefore, both the NH₂ and COOH terminals face cytoplasm (see Fig. 1 for a schematic representation; see also Ref. 27). Caveolin-1 self-oligomerizes or binds caveolin-2 to form homo- or heterooligomers (27). The NH₂-terminal domain of caveolin-1 contains the oligomerization domain (residues 61-101) and the scaffolding domain (residues 81-101) (see Fig. 1 and Ref. 39). The oligomerization domain is the region involved in homooligomerization of caveolin-1, whereas the signaling molecules present in caveolae such as G_i-2, Ras, Src, and eNOS bind to the scaffolding domain of caveolin-1 and are held in the inactive state (27, 39).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Construction of glutathione-S-transferase (GST)-caveolin fusion proteins. Schematic diagram shows primary structure (NH2 and COOH terminals and membrane-spanning region) and functional domains (scaffolding domain: amino acids 82-101; oligomerization domain: amino acids 61-101) of caveolin-1 as previously proposed (25). Full-length (amino acids 1-178) and portions of caveolin-1 that were amplified by PCR and subcloned into the pGEX-4T-1 vector (29) were expressed in Escherichia coli, affinity purified using glutathione-agarose, and dialyzed against PBS.

In this study we examined the role of caveolin-1 in shear stress-dependent activation of ERK by delivering caveolin antibodies into mildly permeabilized endothelial cells in an attempt to block the shear response. Moreover, by using the recombinant glutathione-S-transferase (GST) fusion proteins containing the epitopes of the caveolin-1-recognizing caveolin antibodies, this study determined the critical role of the scaffolding and oligomerization domains (residues 61-101) of caveolin-1 in shear-dependent activation of ERK.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell culture and antibody treatment in permeabilized cells. BAEC obtained from descending thoracic aortas were maintained in a growth medium [DMEM (1 g/l glucose; Life Technologies) containing 20% fetal bovine serum (FBS; Atlanta Biologicals) without antibiotics] at 37°C and 5% CO₂ (17). Cells used in this study were between passages 3 and 10. For shear stress experiments, one million cells per glass slide (75 × 38 mm; Fisher) were seeded in growth medium. The next day the medium was changed to a shear medium (phenol red-free DMEM containing 0.5% FBS and 25 mM HEPES) and incubated for 18 h. Polyclonal (pCav-1) and monoclonal caveolin-1 antibodies (clones 2234, 2297, C060, and C20B) and monoclonal caveolin-2 antibodies (clone 65) were purchased from Transduction Laboratories. A polyclonal antibody raised against canine caveolin-2 was kindly provided by Dr. Kai Simons (31). BAEC were incubated with 0-2.5 µg/ml antibodies at 37°C in the starvation medium containing 0.01% Triton X-100 for 30 min. Treated cells were then exposed to shear stress or static control in fresh, detergent-free starvation medium. In some studies pCav-1 was preincubated for 2 h with the GST-caveolin fusion proteins containing caveolin residues 1-21 (GST-Cav^1-21) or 61-101 (GST-Cav^61-101) (see Fig. 1 and Ref. 33) to block the binding of the antibody to the epitopes of endogenous caveolin-1 in endothelial cells. The mixtures of pCav-1 antibody and GST-caveolins were then added to BAEC in the presence of Triton X-100 and subjected to shear stress.

Shear stress studies. The glass slide containing a confluent monolayer of BAEC was assembled into a parallel plate shear chamber forming a flow channel (220 µm high × 25 mm wide × 62 mm long) between the monolayer and a polycarbonate plate as described previously (18). Nonpulsatile laminar shear stress was controlled by changing the flow rate of the starvation medium delivered to the cells using the constant head flow loop or a syringe pump (KD Scientific) as described (18).

Preparation of cell lysates. After being exposed to shear stress, cells were washed in ice-cold PBS, scraped in 0.25 ml of lysis buffer [50 mM HEPES, pH 7.4, 150 mM NaCl, 1 mM vanadate, 1 mM dithiothreitol (DTT), 1.0% Triton X-100, and 0.1 mM phenylmethylsulfonyl fluoride], and solubilized for 15 min to prepare Triton-soluble lysate as described (18). Entire solubilization procedures were performed at 4°C, and the protein content of soluble cell lysates was measured using a Bio-Rad DC assay kit (Bio-Rad).

Determination of ERK activation by Western blot analysis. Soluble lysates (10 µg each) were resolved by 10% SDS-PAGE, transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore), and probed with an antibody specific to phosphorylated forms of ERK1/2 (pERK1/2) (New England Biolabs) to determine the activation status of the MAP kinase as described (18). As a control, the total amount of ERK1/2 was determined by Western blot analysis using an ERK1 antibody that also cross-reacts with ERK2 (UBI) (18). Goat anti-rabbit IgG conjugated to alkaline phosphatase was used as a secondary antibody, and the membrane was developed using a method for chemiluminescent detection (18).

Immune complex assay for JNK1. Activity of JNK1 was measured by an immune complex kinase assay using an antibody specific for JNK1 (clone G151-333; Pharmingen) and c-Jun (amino acids 5-89) fused to GST (GST-c-Jun) as the substrate, as described (18). Briefly, Triton-soluble cell lysates (100 µg each) were incubated with JNK1 antibody (0.25 µg) for 1 h at 4°C, followed by an additional 1 h of incubation with protein G-agarose. The immune complex was washed four times in the lysis buffer and twice in buffer A (20 mM HEPES, pH 7.6, 20 mM MgCl₂, 20 mM -glycerophosphate, 20 mM p-nitrophenyl phosphate, 0.1 mM vanadate, and 2 mM DTT). The washed immune complexes were incubated in 20 µl of buffer A containing GST-c-Jun (5 µg each), 20 µM ATP, and 5 µCi of [-³²P]ATP for 20 min at 30°C. The reaction was terminated by boiling in 5× Laemmli sample buffer, resolved by 10% SDS-PAGE, and electrotransferred to a PVDF membrane. Autoradiographs were obtained from the dried blot, and radioactivity incorporated into GST-c-Jun was quantitated by cutting and counting each band in a scintillation counter.

Purification and immunoprecipitation of GST-caveolin fusion proteins. Construction of GST-fusion proteins containing the full length or fragments of caveolin-1 (see Fig. 1) from Escherichia coli lysates and purification of the recombinant fusion proteins by glutathione-agarose affinity chromatography were described previously (33). For immunoprecipitation studies, 20 µg of GST-Cav (dialyzed in PBS before use) were incubated for 2 h at 4°C with 0.5 µg of pCav-1 in PBS, followed by an additional 1 h of incubation with protein A-agarose. Immune complexes were washed three times with the lysis buffer, boiled in Laemmli sample buffer, resolved by 12.5% SDS-PAGE, and analyzed by Coomassie blue staining and by Western blotting using a monoclonal anti-GST antibody (Santa Cruz).

Immunocytochemistry. BAEC grown on glass slides were incubated with 1 µg of pCav-1 or 1 µg of nonimmune (NI)-IgG in the presence or absence of 0.01% Triton X-100 for 30 min at 37°C as described in Cell culture and antibody treatment in permeabilized cells. Cells were washed three times in ice-cold PBS to remove unbound antibodies and were then fixed in 4% paraformaldehyde and 0.1% glutaraldehyde for 30 min and blocked (2% BSA, 10% goat serum, and 0.1% Triton) for 1 h. Bound antibodies were detected by using a secondary antibody (Cy3-conjugated goat anti-rabbit IgG) for 30 min. Cells were mounted with Slow-Fade and examined using an Olympus fluorescence microscope. Fluorescence intensity of stained cells was quantitated by an image analysis program (ESPRIT) as described (10).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

pCav-1 is delivered into permeabilized BAEC and binds to caveolin-1. When BAEC were incubated for 30 min with pCav-1 in the presence of 0.01% Triton X-100 (minimal permeabilization), the antibody was detected inside virtually all treated cells (Fig. 2, B and C). Delivery of the pCav-1 antibody into the cells required the presence of the detergent, because the antibody was not detected in cells incubated without Triton X-100 (Fig. 2E). When NI-IgG was used as a control in Triton-permeabilized cells, the fluorescent signal was very faint, indicating that the IgG did not bind tightly to the cells and was mostly removed by extensive washing (Fig. 2D). The staining of pCav-1 showed a punctate pattern (Fig. 2C), which is consistent with its reported intracellular localization in the plasma membrane and Golgi apparatus (33). The characteristic punctate staining pattern of pCav-1 was also observed when BAEC were fixed first and subsequently permeabilized with the use of a high concentration of Triton X-100 (0.1%) before cells were incubated with pCav-1 followed by the secondary antibody (Fig. 2A, positive control).

View larger version (90K): [in this window] [in a new window]   Fig. 2.   Delivery of polyclonal caveolin-1 antibody (pCav-1) into permeabilized bovine aortic endothelial cells (BAEC). A: positive control. BAEC were first fixed in paraformaldehyde and glutaraldehyde, highly permeabilized (0.1% Triton X-100 for 30 min), and blocked for 1 h before being incubated with pCav-1. B-E: unlike control shown in A, cells were not fixed and highly permeabilized before antibody was added. Instead, cells were incubated with 1 µg/ml pCav-1 (B, C, and E) or nonimmune (NI)-IgG (D) in presence (B, C, and D) or absence (E) of 0.01% Triton X-100 for 30 min at 37°C. Cells were then washed in PBS and fixed in paraformaldehyde and glutaraldehyde. pCav-1 and NI-IgG were identified by incubation with Cy3-conjugated goat anti-rabbit IgG for 30 min. Fluorescence micrographs (A and B: original magnification, ×500; C-E: original magnification, ×2,5000; C is a higher magnification of B) are representative of at least 3 independent experiments.

View larger version (42K): [in this window] [in a new window]   Fig. 3.   Binding of pCav-1 to caveolin-1 in permeabilized BAEC. BAEC were treated with pCav-1 (2.5 µg/ml) in presence of 0.01% Triton X-100 for up to 30 min at 37°C. Cells were then washed with PBS and lysed in a buffer containing 1% Triton X-100. Cell lysates were incubated with protein A-agarose to immunoprecipitate the antibody-caveolin-1 complex. Top: immune complexes were resolved by 10% SDS-PAGE, transferred to a polyvinylidene difluoride membrane, cut in half, and analyzed by Western blots with a goat anti-rabbit antibody (top of membrane) to identify pCav-1 delivered into cells. Bottom of membrane was probed with a caveolin-1 antibody (m2297) and subsequently with a goat-anti-mouse antibody. Note that the heavy chain of pCav-1 (IgG-H at ~55 kDa) and caveolin-1 (Cav-1 at 22-24 kDa) were immunoprecipitated only from cells that were incubated for 30 min with antibody and Triton X-100. Results are representative of 3 independent experiments.

pCav-1 inhibits shear stress-dependent activation of ERK, but not JNK, in Triton-permeabilized BAEC. To investigate whether the pCav-1 antibody inhibits shear-dependent activation of ERK, cells were pretreated with the antibody in the presence of Triton before being subjected to shear stress. As we showed previously (28), permeabilization of BAEC by low concentrations of Triton has no significant effect on shear-dependent activation of ERK (Fig. 4, A and C). However, when cells were incubated with pCav-1 in the presence of Triton (minimally permeabilized), shear stress-dependent activation of ERK was inhibited in an antibody concentration-dependent manner, whereas NI-IgG had no effect (Fig. 4, A and C). The inhibitory effect of pCav-1 on the shear activation of ERK was not observed if cells were not Triton permeabilized (Fig. 4B). Although pCav-1 appeared to have a small effect in static controls of nonpermeabilized cells in some studies, its effect was not statistically significant (Fig. 4, B and C).

View larger version (3815K): [in this window] [in a new window]   Fig. 4.   pCav-1 inhibits shear-dependent activation of extracellular signal-regulated kinase (ERK). BAEC were incubated with increasing amounts of pCav-1 or NI-IgG in presence (A) or absence (B) of 0.01% Triton X-100 for 30 min at 37°C. Cells were then subjected to static control or shear stress (10 dyn/cm2 for 5 min) in fresh shear medium without Triton. Cell lysates were analyzed by Western blots with antibodies specific to total ERK1/2 or phospho-ERK1/2 (pERK1/2). C: data for line graph are means ± SE (n = 3) obtained from densitometric quantitation of pERK1 bands. Ab, antibody. *P < 0.05.

View larger version (33K): [in this window] [in a new window]   Fig. 5.   pCav-1 has no effect on shear-dependent activation of c-Jun NH2-terminal kinase (JNK). BAEC were pretreated without or with pCav-1 (2.5 µg/ml) or NI-IgG (2.5 µg/ml) and 0.01% Triton X-100 for 30 min, followed by exposure to static control or shear stress (10 dyn/cm2 for 1 h) and preparation of cell lysates. JNK activity was determined by phosphorylation of GST-cJun using immunoprecipitated JNK1. Top: results of Western blot obtained using a JNK antibody (JNK blot) showing that similar amounts of JNK1 were immunoprecipitated in each lane. A representative autoradiograph shows phosphorylation of GST-cJun. Bottom: radioactivity of each band was quantitated by scintillation counting, and means + SE (n = 3) are shown on bar graph.

The specific effect of pCav-1 on shear-dependent activation of ERK. To determine whether caveolin-1 antibodies other than pCav-1 could also block ERK activation in response to shear stress, we screened additional caveolin-1 antibodies. For this study, cells were incubated with three monoclonal antibodies specific to caveolin-1, m2297 (the epitope 61-71 residues, a part of the oligomerization domain), m2234 (the epitope 1-21 residues), and C20B (an antibody raised against the whole molecule), in minimally permeabilized cells as described in Fig. 3. The pCav-1 antibody and NI-IgG were also used as positive and negative controls, respectively. First, the ability of various caveolin-1 antibodies to bind caveolin-1 under our minimally permeabilized conditions for 30 min was determined by immunoprecipitation. For this study, minimally permeabilized and antibody-treated cells were lysed in 60 mM octylglucoside buffer (1 h at 4°C), and protein G-agarose was added to the cell lysates. Immunoprecipitates were subsequently probed by Western blot using the m2297 antibody as shown in Fig. 6A. To aid in quantifying the amount of caveolin-1 immunoprecipitated by each antibody, we directly used an aliquot of the lysate in the Western blot studies. As shown in Fig. 6A, the m2234 and pCav-1 antibodies immunoprecipitated 30 ± 2% (n = 3) and 2.7 ± 0.2% (n = 5) of total caveolin-1, respectively. These results suggest that both m2234 and pCav-1 bind to the native form of caveolin-1 and are consistent with previous reports (26, 33). In contrast, the m2297 and mC20B antibodies and NI-IgG did not immunoprecipitate any significant amount of caveolin-1 from cell lysate (Fig 6A). However, m2297 and mC20B work well in Western blot analysis (see Figs. 3 and 6 as well as unpublished results from Transduction Lab described in DISCUSSION), suggesting that these antibodies recognize denatured, but not native, caveolin-1.

View larger version (41K): [in this window] [in a new window]   Fig. 6.   Specificity of caveolin-1 antibodies on shear-dependent activation of ERK. BAEC were incubated without (none) or with 2.5 µg/ml caveolin-1 antibodies (m2297, m2234, C20B, or pCav-1) or NI in presence of 0.01% Triton X-100 for 30 min. A: to determine amount of caveolin-1 that can be immunoprecipitated by each antibody, cells were lysed in -octylglucoside (60 mM), and lysates (200 µg each) were used for immunoprecipitation with protein G-agarose for 1 h at 4°C. Each immunoprecipitate was analyzed by Western blot analysis with m2297. Also, lysate (10 µg) was used to determine total amount of caveolin-1. B: after treatment with caveolin antibodies and minimum permeabilization, cells were exposed to static control () or shear stress (+). Activation of ERK was then determined by Western blot analysis with a pERK antibody (pERK1/2) and quantitated densitometrically as described in Fig. 3. Data in bar graph are means ± SE (n = 3). *P < 0.05.

Epitope mapping of pCav-1. The pCav-1 antibody was raised against the NH₂-terminal domain (1-101) of caveolin-1, but its epitopes have not been determined. To determine the amino acid residues of caveolin-1 that are recognized by pCav-1 in a nondenatured state, we used the recombinant GST fusion proteins containing various fragments of caveolin-1 in immunoprecipitation studies. The recombinant GST-caveolins were produced in E. coli and purified by glutathione-agarose, and its purity was determined by protein staining (Fig. 7A). When pCav-1 was added to various GST-caveolins, only the fusion proteins containing either residues 1-21 or 61-101 were immunoprecipitated (Fig. 7B). In contrast, the m2234 antibody immunoprecipitated the fusion proteins containing residues 1-21 but not residues 61-101 (Fig. 7C). The epitope 1-21 that recognizes m2234 in a nondenatured state is the same as that previously reported using denatured and reduced caveolin-1 in Western blot analysis (33).

View larger version (40K): [in this window] [in a new window]   Fig. 7.   Epitope mapping of caveolin-1 antibodies. Full-length (1-178) and portions (amino acid residue numbers shown above each lane) of caveolin-1 were recombinantly expressed as GST-fusion proteins as described in Fig. 1. A: Coomassie blue staining of purified caveolin fusion proteins. B and C: GST-caveolin fusion proteins were incubated with either pCav-1 (B) or a monoclonal caveolin-1 antibody (m2234) (C) overnight at 4°C and subsequently with Protein A- or G-agarose for 1 h. Washed immunoprecipitates (IP) were then analyzed by Western blot analysis using an antibody specific to GST (GST). Note that pCav1 recognizes 2 epitopes within caveolin-1 residues 1-21 and 61-101, whereas m2234 binds to 1-21.

The inhibitory effect of pCav-1 can be prevented by preabsorption with the recombinant fusion protein GST-Cav^61-101. The study of epitope mapping of pCav-1 (Fig. 7) showed that two epitopes in caveolin-1 bind to pCav-1: 1) one at residues 1-21, which is at the NH₂ terminus of the caveolin-1 isoform, and 2) the other at residues 61-101, spanning both the oligomerization domain (residues 61-101) and the scaffolding domain (residues 81-101). To determine whether one or both epitopes could block the effect of pCav-1, we preincubated the antibody with GST-Cav^1-21, GST-Cav^61-101, or GST alone before adding the mixtures of antibodies and fusion proteins to permeabilized cells. After cells were exposed to static control or shear stress, activation of ERK was measured. As shown in Figs. 4 and 6, treatment of minimally permeabilized cells with pCav-1 inhibited shear-dependent activation of ERK (Fig. 8, compare lanes 1 and 2 with lanes 3 and 4). However, the inhibitory effect of pCav-1 completely reversed if the antibody was preincubated with GST-Cav^61-101 (Fig. 8, compare lanes 3 and 4 with lanes 9 and 10). In contrast, GST-Cav^1-21 or GST alone did not reverse the inhibitory effect of the pCav-1 antibody (Fig. 8, lanes 4-7). Although in some studies treatment of cells with GST-Cav^61-101 plus pCav-1 appeared to have a small effect on basal ERK activity (Fig. 8, lane 9), this effect was not statistically significant (n = 3). Furthermore, GST-caveolins or GST alone had no effect in the absence of the antibody on control or shear-dependent activation of ERK (Fig. 8, compare lanes 11 and 12 with lanes 13-18). These results show that the peptide 61-101 of caveolin-1 is the essential region that mediates the inhibitory effect of pCav-1 on shear-dependent activation of ERK.

View larger version (37K): [in this window] [in a new window]   Fig. 8.   Residues 61-101 of caveolin-1 prevent the inhibitory effect of pCav-1 on shear-dependent activation of ERK. A: pCav-1 (2.5 µg/ml) was preincubated for 2 h without or with 1 µg/ml GST-Cav1-21 (1-21) and GST-Cav61-101 (61-101) or GST alone. Mixtures of pCav-1 and GST-fusion proteins were then added to BAEC in presence of Triton as described in Fig. 3. As a control (lanes 1 and 2), cells were incubated only with Triton X-100 without antibody. In some studies (lanes 13-18), BAEC were incubated with GST-fusion proteins (1-21 or 61-101) or GST alone with Triton but in absence of pCav-1. After minimum permeabilization and antibody treatment, cells were exposed to shear stress (10 dyn/cm2 for 5 min), and ERK activity was determined by Western blot using antibodies specific for pERK1/2 and total ERK (A) as described in Fig. 3. Representative Western blots with total ERK and pERK antibody, respectively, are shown. B: bar graphs represent means ± SE (n = 3) of densitometric quantitation of pERK1 bands. *P < 0.05.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The current study demonstrates that shear stress-dependent activation of ERK is inhibited by an antibody to caveolin-1, pCav-1, in Triton-permeabilized BAEC. In this study we have established a unique method to minimally permeabilize cells using a low concentration of Triton X-100. Using this method, we were able to show that pCav-1 can be delivered to virtually all BAEC and that the antibody bound to 3-7% of total caveolin-1 as determined by immunohistochemistry or immunoprecipitation (Figs. 2, 3, and 6A). Although only a small fraction of caveolin-1 was bound by the pCav-1 antibody, it was enough to inhibit shear stress-dependent activation of ERK by 65-90% of control in >12 different experiments. This finding raises the interesting possibility that endothelial cells may contain different pools of caveolin-1, one of which readily binds to the pCav-1 antibody and plays a crucial role in the mechanosensitive activation of ERK. Furthermore, the mechanisms by which the pCav-1 antibody inhibits shear-dependent activation of ERK involve the amino acid residues 61-101 of caveolin-1. These results suggest that the scaffolding and oligomerization domains (residues 61-101) of caveolin-1 are critical in regulating the mechanosensitive activation of ERK.

Caveolin-1 is composed of 178 amino acids, and its topology is thought to resemble a hairpinlike structure with its membrane-spanning domain (residues 102-134) flanked by the NH₂ terminus (residues 1-101) and COOH terminus (residues 134-178), both of which face the cytoplasm (27). Caveolin-1 forms oligomers through its oligomerization domain (residues 61-101) and the COOH terminus (39) and binds directly with many signaling molecules such as heterotrimeric G proteins, Ras, Src family kinases, and eNOS through the scaffolding domain (residues 81-101) (27). These caveolin-binding signaling molecules have been proposed to form signal transduction units in the inactive state within caveolae-like structures, ready to be activated by specific stimuli (27). The sequestration of preassembled signaling units may provide the mechanisms responsible for signaling specificity and efficiency in response to specific stimuli. Most caveolin-binding signaling proteins contain the two related motifs (XXXXX and XXXXXX, where  is the aromatic amino acid Trp, Phe, or Tyr) that bind to the scaffolding domain of caveolin-1 (38).

What is the mechanism by which pCav-1 inhibits activation of ERK in response to shear stress? Although pCav-1 binds to the two caveolin-1 epitopes (residues 1-21 and 61-101), only the addition of the competing fusion protein GST-Cav^61-101 prevents the inhibitory effect of pCav-1 on shear activation of ERK (Fig. 8). A simple interpretation of this finding is that the binding of pCav-1 to the residues 61-101 of caveolin-1 disrupts clustering of caveolin-1 (oligomerization domain at the residues 61-101) or the interaction between the scaffolding domain (residues 81-101) and signaling molecules (e.g., G_i-2, Src, and Ras) that are upstream regulators of the flow-sensitive ERK pathway. Current literature provides strong support for the possibility that binding of pCav-1 to the scaffolding domain would compete with the caveolin-binding signaling molecules and displace them from caveolin-1. This disruption would then prevent activation of the ERK pathway in response to shear stress. It was recently found that the scaffolding domain is both necessary and sufficient for membrane attachment of caveolin-1 (34). This provides an additional possibility that pCav-1 may disrupt attachment of caveolin-1 to membrane, resulting in its inability to mediate the mechanosensitive activation of ERK pathway. At present, we cannot exclude other possibilities, including the possibility that the binding of the pCav-1 antibody to caveolin-1 interferes with trafficking of cholesterol between caveolae and intracellular compartments such as endoplasmic reticulum (1). The exact mechanisms underlying the specific effect of pCav-1 await further studies.

Caveolin-1 binds to itself as well as caveolin-2 to form homo- or heterooligomers (6, 31). Therefore, the inhibitory effect of pCav-1 on shear activation of ERK may involve caveolin-2 directly or indirectly. It is not clear, however, whether BAEC express caveolin-2. Unlike caveolin-1, the amino acid sequences of caveolin-2 show species-specific (human vs. canine) differences (31, 32). Two caveolin-2 antibodies are currently available: one is specific for human caveolin-2 (Transduction Lab) and the other for the canine form (31). However, both antibodies failed to recognize caveolin-2 by either immunoprecipitation or Western blot analysis in BAEC, whereas the mouse and canine antibodies recognized caveolin-2 expressed in RSV-3T3 cells (provided as a positive control by Transduction Lab) and Madin-Darby canine kidney cells, respectively (data not shown). Caveolin-2 antibodies recognizing the bovine form are not yet available to our knowledge. Although both caveolin-2 antibodies had no effect on the shear activation of ERK (results not shown), these studies need to be evaluated further when bovine-specific caveolin-2 antibodies become available.

Collectively, our current and previous (28) findings as well as the report by Schnitzer and colleagues (30) suggest that the cholesterol- and caveolin-sensitive platforms in the plasma membrane such as caveolae-like domains are responsible for mediating the activation of ERK in response to shear stress. These findings have implications in not only the mechanisms responsible for spatial sequestration of signaling units but also the elusive mechanosensor(s) that senses the changes in shear stress. Several candidates have been proposed as the potential mechanosensor, including the caveolae structure itself, G proteins, G protein-coupled receptors, integrins-cytoskeleton (tensegrity theory), and ion channels (3, 7, 12, 30, 44). It is interesting to note that many of these candidates, including caveolae, G protein-coupled receptors, G proteins, and integrins, have been shown to interact with caveolin-1 (3, 27, 44, 45, 46). Recently integrins, transmembrane glycoproteins that act as adhesion receptors for extracellular matrix, have been reported as caveolin-binding molecules (45, 46). Several lines of evidence provide support for the potential importance of caveolin-integrin interaction in the shear-dependent ERK pathway: 1) integrin-dependent cell adhesion is required for activation of ERK (41), 2) _v3- and ₁-integrins and FAK are involved in the mediation of shear-dependent ERK activation (14, 22), and 3) caveolin-1 binds to ₁-integrins and plays a critical role in integrin-dependent activation of ERK and cell adhesion (45, 46).

The minimal permeabilization method developed in this study using a non-cholesterol-specific detergent could be used as a general technique to deliver membrane-impermeable macro- or micromolecules (i.e., peptides, antibodies, oligonucleotides, etc.) inside the cells. This could be especially useful as an alternative to other wildly used membrane-permeabilizing agents such as digitonin and filipin, which work by chelating membrane cholesterol, thereby potentially interfering with caveolae-dependent cellular functions (28).

In summary, we report that the scaffolding and/or oligomerization domain of caveolin-1 plays an essential role in shear stress-dependent activation of ERK. On the basis of our current and previous (28) findings, we propose that the shear-sensitive signaling molecules that regulate the ERK pathway are assembled in caveolae-like domains by binding to the scaffolding domain of caveolin-1 in the inactive state. Changes in shear stress level may then specifically trigger rapid, organized, and compartmentalized signaling cascades to activate ERK, but not other pathways. This spatial compartmentalization model provides a plausible mechanism by which shear stress differentially activates MAP kinases and other mechanosensitive responses in endothelial cells.