Endothelin converting enzyme is located on α-actin filaments in smooth muscle cells

Abstract

Objective: Endothelin converting enzyme is the key enzyme in the generation of endothelin-1 from big-endothelin-1. The mature endothelin-1 is a potent vasoconstrictor which also promotes mitogenesis and proliferation of smooth muscle cells. The objectives were to demonstrate in smooth muscle cells the presence of a phosphoramidon-sensitive endothelin converting enzyme activity, reveal the subcellular localization of the enzyme protein and determine the effects of the metalloproteinase inhibitor, phosphoramidon, the lysosomotrophic drug, chloroquine, and colchicine on the cycling pathway of the enzyme. Methods: Subcellular localization of endothelin converting enzyme on human smooth muscle cells and the rat cell line, A7r5, was by immunofluorescence and confocal microscopy or by biotinylation of cell cultures and immunoblotting, after treatment of cell cultures with cytochalasin D, colchicine, chloroquine and phosphoramidon. Converting enzyme activity was determined by high performance liquid chromatographic assay. Results: We detected phosphoramidon-sensitive endothelin converting enzyme activity in smooth muscle cells. In addition to its plasma membrane location, for the first time we revealed a striking co-localization of endothelin converting enzyme and α-actin filaments in smooth muscle cells. Colchicine treatment results in a perinuclear accumulation of endothelin converting enzyme. An increased level of endothelin converting enzyme protein was shown to be present in smooth muscle cells which had been grown in the presence of phosphoramidon or chloroquine. Conclusion: The 120 kDa endothelin converting enzyme co-localizes with α-actin in smooth muscle cells and resembles that found in endothelial cells in that it is present on both the plasma membrane and intracellularly.

Constitutive vesicles, Endothelin, Converting enzyme, Metalloproteinase, Human umbilical artery smooth muscle cells

CHO-K1, Chinese hamster ovary cells, ECE, endothelin converting enzyme, ECL, enhanced chemiluminescence, ET, endothelin, FITC, fluorescein isothiocyanate, HPLC, high performance liquid chromatography, TRITC, tetramethylrhodamine isothiocyanate, TRLEC-03, transformed rat lung endothelial cells, SDS-PAGE, sodium dodecylsulphate polyacrylamide gel electrophoresis, SMC, smooth muscle cells

Time for primary review 27 days.

1 Introduction

Although endothelins (ET) were initially isolated from vascular endothelium [1], ET-1 mRNA is also expressed in vascular smooth muscle cells (SMC) in culture and both ET-1 and ET-3 are released from such cells [2–4]. SMC express both ET_A and ET_B receptors [5,6] although the ET_A receptor protein is expressed more abundantly [3,4]. Conversely, on vascular endothelial cells, only the ET_B receptor mRNA is expressed [7]. ET-1 acts primarily through the ET_A receptors, which have a high affinity for ET-1 and ET-2 [8]. ET-1 regulates vascular tone, is a potent vasoconstrictor, and also promotes mitogenesis and proliferation of the SMC [9–13]. Furthermore, it has been suggested that ET-1 and angiotensin-II may act in concert to promote SMC growth through a common intracellular signalling mechanism involving stimulation of protein tyrosine phosphorylation as well as mitogen-activated protein kinases [14].

Generation of the mature ETs from big ET is catalyzed by the phosphoramidon-sensitive metalloprotease, endothelin-converting enzyme (ECE), which is an integral membrane protein existing as a homodimer [15–17]. Selective inhibition of ECE may represent a viable therapeutic strategy in certain cardiovascular and other disorders. An ECE-1 [18–22] and an ECE-2 [23] cDNA have been cloned. In addition, two isoforms of rat ECE-1 (ECE-1α and ECE-1β; ECE-1c and ECE-1a, respectively, in humans) which differ only in their N-terminal region have also been detected [24,25]. In humans, a minor third isoform, ECE-1b also exists [26]. Although endothelial ECE has been the focus of most previous studies, a phosphoramidon-sensitive ECE which converts big ET-1 to ET-1, has been detected in an homogenate of vascular SMC from bovine carotid arteries [27] and on the surface of cultured SMC from bovine pulmonary artery cells [28]. ECE-1 has also been detected by immunoblotting in a variety of tissues and on a transformed rat endothelial cell line, TRLEC-03, but was not detectable on cultured rat aortic SMC [29]. However, immunohistochemical analysis has located ECE-1 to endothelial cells and some secretory cells [29] and to neointimal but not medial SMC in rat balloon injured arteries, and in both SMC and macrophages in human coronary atherosclerotic lesions [30,31]. Concomitant with the detection of ECE-1 in the neointima, the levels of ET-1 also increase. Thus ECE-1 may contribute to the process of injury-induced neointimal formation and atherosclerosis particularly through the mitogenic and proliferative effects of ET-1.

The subcellular localization of ECE-1 on the cell surface and on constitutive vesicles of endothelial, neuroblastoma and glial cells has been well-documented by immunofluorescence and immunoelectron microscopical studies [29,32–35]. At the subcellular level, ECE-1 in endothelial cells has been shown to redistribute from the cell surface to an intracellular compartment and to increase in level after treatment with the metalloprotease inhibitors, phosphoramidon and thiorphan. This phenomenon may be due to the inhibition of activity of a novel metalloprotease involved in ECE-1 turnover [32]. In the present paper we have investigated the subcellular localization of ECE-1 on cultures of human SMC derived from umbilical artery and of a rat SMC line, A7r5. Our data suggest that the α-actin microfilaments of smooth muscle cells represent the terminal stage of the trafficking route of ECE-1 to the cell surface.

2 Methods

2.1 Materials

Phosphoramidon was from the Peptide Institute Inc., Osaka, Japan. The peptide substrate DIIWFNTPEHVVPYGLGNH₂ was synthesized by the Multiple Sclerosis Peptide Laboratory, Oxford Brookes University, UK. Cell culture reagents and Dulbecco’s phosphate buffered saline (DPBS) were from GIBCO BRL, Life Technologies Ltd., Paisley, Scotland, UK. The biotin/streptavidin fluorescein isothiocyanate (FITC) system, the tetramethylrhodamine isothiocyanate (TRITC) conjugate and the enhanced chemiluminescence (ECL) western blotting kit were from Amersham International plc, Amersham, Buckinghamshire, UK. The mouse monoclonal anti-transferrin receptor (human), B3/25 (2), was from Boehringer Mannheim Ltd., Lewes, Sussex, UK. All other reagents and the mouse monoclonal antibody, 1A4, specific to smooth muscle α-actin, were purchased from Sigma Chemical Co., Poole, Dorset, UK.

2.2 Cell culture

All cell types were cultured in 75 cm² or 175 cm² flasks (Nuclon, GIBCO.BRL). Human smooth muscle cells (human SMC) derived from umbilical artery were seeded into a low serum medium supplemented with growth factors and antibiotics according to the supplier’s instructions (Promocell, D-69120 Heidelberg, Germany), and used following one passage. The rat smooth muscle cell line (rat SMC), A7r5, was maintained in Dulbecco’s minimal essential medium plus 10% fetal calf serum supplemented with glutamine and antibiotics according to the supplier’s instructions (European Collection of Animal Cell Cultures, Centre for Applied Microbiology & Research, Salisbury, Wiltshire SP4 OJG, UK). Chinese hamster ovary (CHO-K1) cells were cultured in supplemented Glasgow minimal essential medium [36]. The expression vector pcDL-SRα296/rECE-1α was used to transfect the CHO-K1 cells [36]. All cells were maintained at 37°C in 5% CO₂ in air. For some experiments the rat and human SMC were incubated with 0.1 mmol/l phosphoramidon for 48 h or with 0.1 mmol/l chloroquine for 1 to 2 h, prior to preparation of membranes. Confluent cultures of cells were washed three times in Dulbecco’s phosphate buffered saline (DPBS) and harvested by scraping into 50 mmol/l Tris/HCl, 100 mmol/l NaCl, 18 mmol/l CaCl₂ pH 7.4, and homogenized in a Parr cell disruption bomb using N₂ (800 psi at 4°C for 10 min). The homogenate was centrifuged for 100,000×g for 90 min and the resultant membranes were solubilized in 1% Triton X-100 in 25 mmol/l 2-N[-morpholino] ethane sulphonic acid (Mes)/0.15 mol/l NaCl, pH 6.5 (v/v) for 45 min at 4°C and stored at −20°C until required for immunoblotting or assaying for ECE-1.

2.3 Biotinylation of human smooth muscle cells

The biotinylation procedure, for analysis of cell surface proteins was performed on confluent monolayers of human SMC in 75 cm² flasks at 4°C [34]. After removal of the medium, the cells were washed three times with 10 ml of DPBS, then 50 μl of biotinamidocaproate-N-hydroxysuccinimide ester in dimethylformamide (40 mmol/l) was added in 5 ml of bicarbonate buffer, pH 8.6. This solution was removed after 15 min and replaced by fresh biotinylation reagent for a further 15 min. The cells were washed twice with DPBS and each flask lysed with 0.5 or 1 ml of 1% Triton X-100 (vol/vol) in PBS which included 0.2 mmol/l PMSF, 0.002 mmol/l pepstatin A, and 0.01 mmol/l leupeptin. The cells were scraped from the flask immediately, left on ice for 20 min and the lysate plus debris centrifuged at 100,000×g for 30 min. The clear lysate (0.5 mg protein) was rotated for 2 h at 4°C with 200 μl of streptavidin–agarose beads. The beads were recovered by centrifugation for 1 min at 10,000×g and washed twice in PBS. The biotinylated cell surface proteins were eluted from the streptavidin–agarose beads in 25 μl of SDS–PAGE sample buffer then heated at 100°C for 4 min. The eluted bead samples plus the cleared lysates (which include the intracellular proteins) were stored at −20°C until required for immunoblotting. As a control, streptavidin–agarose beads were also incubated with non-biotinylated cells. To check that removal of the biotinylated cell surface proteins was complete, on some occasions a second aliquot of streptavidin–agarose beads was incubated with the lysate and removed for analysis (results not shown).

2.4 Electrophoresis and immunoblotting for ECE

SDS–PAGE sample buffer was added to samples of the thawed lysates from the biotinylation experiments. All samples were treated with 5% mercaptoethanol and separated by gel electrophoresis on 7.5% gels, prior to immunoblotting [32]. The Amersham enhanced chemiluminescence (ECL) Western blotting kit was used to visualize the proteins according to the manufacturer’s instructions. The primary mouse monoclonal antiserum used was AEC32-236 (4 mg/ml) raised to rat ECE, diluted 1/200 or 1/400. The specificity of the antiserum to ECE has been demonstrated previously [29]. The secondary anti-mouse horseradish peroxidase conjugate was diluted 1/5000.

2.5 Enzyme and protein assays

Protein determination was by the bicinchoninic acid method with bovine serum albumin (1 mg/l) as standard. An HPLC method was used to assay phosphoramidon-sensitive ECE using the synthetic peptide, DIIWFNTPEHVVPYGLGamide as substrate [37].

2.6 Immunofluorescence staining of smooth muscle cells

Monolayers of human SMC and rat SMC were cultured on glass coverslips and routinely fixed, except where stated, with methanol/acetone,1:1 (v/v). The cells were blocked with 1% normal goat serum in 0.02% gelatin diluted in Tris buffered saline (TBS: 50 mmol/l Tris, 150 mmol/lsodium chloride, pH 7.4), for 30 min and immunostained as in [32]. Two antibodies raised to rat ECE-1 were used for immunostaining of the smooth muscle cells: a mouse monoclonal, AEC32-236 (dilution 1/20) which recognizes the extracellular C-terminus [29] and an anti-peptide rabbit polyclonal raised to the N-terminal cytoplasmic tail of ECE-1, AS-66 (dilution 1/50) [38]. Visualization was by using the appropriate anti-mouse or anti-rabbit biotinylated conjugate (dilution 1/100) and streptavidin fluorescein isothiocyanate (FITC) (dilution 1/400). All antibodies and conjugates were diluted in the TBS blocking buffer. The use of two antisera of different origins and their application to two different cell lines limited the possibility that the immunopositive fluorescence was due to non-specific binding of the antibodies to the actin filaments.

For double immunostaining of the human and rat SMC two different protocols were used. The mouse monoclonal antibody, AEC32-236 to ECE-1 (diluted 1/20) was applied and the sections processed as described above. The cells were washed copiously in at least three changes of TBS and then overnight at 4°C in TBS, after which blocking buffer was applied for 30 min. The mouse monoclonal antibody to α-actin (dilution 1/400) was applied to the cells for 2 h, followed by washing in TBS and addition of the anti-mouse tetramethylrhodamine isothiocyanate (TRITC) conjugate (dilution 1/100), prior to washing and mounting. For controls the whole procedure was followed as above, except that one of the primary or secondary antibodies was omitted or replaced by equivalent dilutions of control mouse ascites fluid. Alternatively, the rabbit polyclonal antiserum to ECE-1 (dilution 1/50) and the relevant mouse monoclonal antisera to either α-actin (dilution 1/400) or to transferrin receptor (dilution 1/20) were applied. The secondary anti-rabbit biotinylated conjugate (dilution 1/100) was applied, followed by the addition of the streptavidin FITC (dilution 1/400) and anti-mouse TRITC (dilution 1/100) simultaneously. The mouse antiserum to ECE-1 and the rabbit antiserum to human trans-Golgi network protein, TGN46 (dilution 1/500), were also added concomitantly, followed by the appropriate fluorescent conjugates. Cells were washed between additions of antisera and conjugates as described above. In all cases, controls for individually labelled cell elements were carried out to exclude the possibility of signal crossover, or the primary antibody was replaced by a corresponding dilution of rabbit preimmune serum or mouse ascites. Coverslips were mounted in Vectashield (Vector Laboratories, Peterborough, UK) and examined by a Leitz confocal microscope.

2.7 Treatment of smooth muscle cells with chloroquine, colchicine and cytochalasin D

On some occasions human SMC cells which had been cultured on coverslips were pre-incubated for 3 h at 37°C with 0.1 mmol/l chloroquine prior to washing, fixation with methanol/acetone and immunostaining as described above. Rat and/or human SMC grown on coverslips were treated with 0.005 mmol/l colchicine for 30 min, or with 0.005 mmol/l cytochalasin D for 2 min or 5 min, in the culture medium at 37°C, prior to washing, fixation with methanol/acetone and double immunostaining as described above.

3 Results

3.1 Assay of phosphoramidon-sensitive ECE activity in human smooth muscle cells

To demonstrate the presence of a phosphoramidon-sensitive ECE activity in smooth muscle cells, membrane preparations of human SMC were assayed for ECE with a range of concentrations of phosphoramidon from 0.001 μmol/l to 0.1 mmol/l. The ECE activity was inhibited by phosphoramidon with an I₅₀ of approx. 0.01 mmol/l, although activity never fell below 18% of control even in the presence of 0.1 mmol/l inhibitor. The specific activity of the phosphoramidon-sensitive smooth muscle ECE was 1.8±0.01 nmol.min⁻¹.mg⁻¹ (3 experiments each in duplicate).

3.2 The ECE-1 in smooth muscle cells is an 120 kDa protein

To confirm the presence of ECE-1 protein in SMC, samples of solubilized membrane protein from smooth muscle cells (10 μg protein) and CHO-K1 cells transfected with rat ECE-1α (2.5 μg protein) were separated by SDS-PAGE and immunoblotted with the mouse monoclonal antibody to ECE-1. An 120 kDa ECE-1 protein band was observed when rat SMC (Fig. 1 lanes 2 to 4) and CHO-K1 cells transfected with ECE-1α (Fig. 1a, lane 5) were loaded on the same gel and immunoblotted. To examine whether the degradation of ECE-1 in SMC is inhibited by phosphoramidon, as proposed for endothelial cells [32], cultures of SMC were grown in the presence of the metalloprotease inhibitor, prior to preparation of membranes. An 120 kDa ECE-1 protein band was visualized when human SMC were immunoblotted (Fig. 1, lanes 6, 7). In the rat SMC line (A7r5) and in the human SMC, the level of ECE-1 protein increased after treatment with phosphoramidon (three experiments) (Fig. 1a: compare lanes 2 and 3, b: compare lanes 6 and 7). When rat SMC were incubated with 0.1 mmol/l of the lysosomotrophic drug, chloroquine, which perturbs endosomal luminal pH [40], (three experiments), there was also an increase in the level of ECE-1 protein (Fig. 1a: compare lanes 2 and 4).

Fig. 1

Immunoblot analysis of endothelin-converting enzyme-1 (ECE-1) in smooth muscle cells. Samples of 10 μg of solubilized rat or human smooth muscle cell membranes and 2.5 μg of solubilized ECE-1α-transfected CHO-K1 membrane proteins were separated by SDS-PAGE and, following transfer to PVDF membranes, were immunoblotted with the mouse monoclonal antiserum to ECE-1. Panels show in (a) lane 1: protein markers transferred to PVDF membranes and stained with Coomassie blue, lanes 2–4: rat smooth muscle cell line, A7r5, and lane 5: rat ECE-1α transfected CHO-K1 cells (b) lanes 6–7: human smooth muscle cells. All show ECE-1 to be a 120 kDa protein in smooth muscle cells. Cells, lanes (a) 3 and (b) 7, were treated with 0.1 mmol/l phosphoramidon for 48 h and cells, lane (a) 4, were treated with 0.1 mmol/l chloroquine for 3 h, prior to harvesting and immunoblotting. All treated cells show an increase in ECE-1 protein as compared with untreated cells.

3.3 The location of ECE-1 on human smooth muscle cells by surface biotinylation of cells

The subcellular distribution of ECE-1 was investigated by biotinylation of the cell surface of monolayers of human SMC, and isolation of the cell surface proteins binding to streptavidin–agarose beads, followed by Western blot analysis. The experiment was repeated three times on biotinylated and non-biotinylated flasks of cells. Cell surface ECE-1 protein was detectable after elution from streptavidin–agarose beads which had been incubated with extracts of biotinylated cells (Fig. 2, lane 2) but not from those which had been incubated with extracts of non-biotinylated monolayers (Fig. 2, lane 1). The ECE-1 visualized in the the lysate remaining after incubation with the streptavidin–agarose beads represents the intracellular enzyme (Fig. 2, lane 3). No immunopositive reaction was seen after a second aliquot of streptavidin–agarose beads was incubated with the cell lysates, indicating that there were no biotin labelled proteins remaining after incubation with the first aliquot of beads (results not shown).

Fig. 2

Immunoblot of biotinylated cell surface and intracellular fractions from human smooth muscle cells Solubilized, biotinylated cell surface protein from human smooth muscle cells was isolated on streptavidin–agarose beads, separated by SDS–PAGE and immunoblotted with the mouse monoclonal antiserum to endothelin converting enzyme-1 (ECE-1). Lane 1: eluant (25 μl≅30 μg protein) from streptavidin/agarose beads incubated with Triton X-100 solubilized non-biotinylated cells. Lane 2: eluant (25 μl≅30 μg protein) from streptavidin/agarose bead fraction incubated with solubilized biotinylated cells: cell surface ECE. Lane 3: 10 μg protein from the supernatant fraction: intracellular ECE.

3.4 Immunofluorescence localization of ECE-1

The distribution of ECE-1 in rat and human smooth muscle cells was analysed by immunofluorescence and by confocal microscopy. Two cell lines and two ECE-1 antisera raised in different species were used to determine that the observed fluorescence was specific. Rat and human SMC were routinely fixed and permeabilized with methanol/acetone, except where stated. Cells immunostained with either of the antisera revealed the enzyme to be localized as discrete punctate clusters, many of which were organized in a striking linear configuration (Fig. 3a and b, arrowheads), resembling actin filaments. In all controls only background levels of immunofluorescence were observed, as for example, when the primary antibody was replaced by pre-immune serum (Fig. 3c). After treatment of the rat and human SMC with the microtubule-disrupting agent, colchicine, predominantly immunopositive perinuclear staining for ECE-1 was observed (Fig. 3d). This further confirmed the specific nature of the immunopositive fluorescence for ECE-1, as the actin filaments remain intact but immunonegative for ECE-1 after treatment. When non-permeabilized cells fixed with 4% formaldehyde were immunostained with the monoclonal antibody to the extracellular C-terminus of ECE-1, only the clustered non-linear pattern of the cell surface plasma membrane ECE-1 was visualized (Fig. 1e). Dense patches of clustered ECE-1 were also observed (Fig. 3a and e, large arrows).

Fig. 3

Immunolabelling of smooth muscle cells (SMC) for endothelin converting enzyme-1 (ECE-1). Monolayers of rat SMC (a) and human SMC (b–e) were fixed with methanol/acetone (a–d) or 4% formaldehyde in phosphate buffer and immunostained for ECE-1 as described in the methods. Composite confocal microscope images (a),(b) (c) and (e) are formed from 5 focal planes: (d) is a composite image of three focal planes. When either the mouse monoclonal antiserum (a) or the rabbit polyclonal antiserum (b) to rat ECE-1 was used, the protein was observed as fluorescent punctate clusters, frequently organised in distinct rows (arrows). Only background levels of immunofluorescence staining were observed when the rabbit antiserum was replaced with the preimmune rabbit serum (c). Treatment of the cell monolayers with 0.005 mmol/l colchicine for 30 min. at 37° C and immunstaining with the rabbit antiserum to ECE-1 revealed ECE-1 in a perinuclear region (d: arrowheads). Clustered cell surface ECE-1 only was revealed in non-permeabilized cells, after fixation with 4% formaldehyde and immunostaining with the rabbit antiserum to ECE-1. The apparent reduction in ECE-1 staining over part of the nucleus is due to this region not being included in the composite image (e). Dense patches of clustered ECE-1 were observed in (a) and in (e): large arrows. Bars: all 10 μm.

3.5 Double immunostaining for ECE-1 and α-actin

To confirm whether the linear punctate staining for ECE-1 corresponded with the localization of smooth muscle α-actin, rat and human SMC were double immunostained for ECE-1 (Fig. 4a) and smooth muscle α-actin (Fig. 4b). Superimposition of the resulting confocal images showed a striking overlap and hence colocalization of the punctate ECE-1 staining and α-actin filaments (Fig. 4c). Perinuclear staining of ECE-1 was not associated with the α-actin (Fig. 4c).

Fig. 4

Confocal images of double-labelling of human smooth muscle cells for endothelin converting enzyme-1 (ECE-1) and α-actin. Monolayers of human smooth muscle cells grown on coverslips were sequentially immunostained with the mouse antisera to ECE-1 (a) and to smooth muscle α-actin (b) as described in the methods. ECE-1 is visualized with the anti-mouse biotin/streptavidin fluorescein isothiocyanate system, and α-actin with secondary anti-mouse antiserum conjugated to tetramethylrhodamine isothiocyanate. Micrograph (c) is a representation of the superimposition of the images (a) and (b): yellow indicates the co-localization of ECE-1 with α-actin microfilaments: not all ECE-1 is co-localized. Bar: 10 μm.

After treatment of the rat SMC with cytochalasin D, the organized α-actin filaments (Fig 5a: arrows) became disassembled (Fig. 5c: arrows) and the cells became spherical in shape. Concomitantly, ECE-1 which localized with the isotropic network of the α-actin filaments (Fig. 5b) before treatment of the cells, was localized after treatment with the disassembled focal accumulations of α-actin (Fig. 5d: arrows).

Fig. 5

Double immunofluorescence-staining for endothelin converting enzyme-1 (ECE-1) and α-actin of untreated (a and b) and cytochalasin D-treated (c and d) rat smooth muscle cells (SMC). Composite images from 5 focal planes. Cultured rat SMC were treated with 0.005 mmol/l cytochalasin D for 5 min at 37°C and immunostained with the mouse antiserum to α-actin and the rabbit antiserum to ECE-1, followed by the relevant anti-mouse rhodamine (left hand panal) and anti-rabbit biotin/streptavidin fluorescein conjugates (right hand panal), respectively. ECE-1, co-localized with the isotropic α-actin, (a) and (b) (arrows), remains localized with the disrupted focal accumulations of α-actin, (c) and (d) (arrows). Bars: 10 μm.

3.6 Localization of ECE-1 to subcellular organelles

SMC were treated with chloroquine and immunostained for ECE-1. After treatment, the linear arrangement of ECE-1 (Fig. 6a: arrow) was to a large degree replaced by the concentration of ECE-1 in subcellular organelles (Fig. 6a). By double immunofluorescence staining, ECE-1 was localized consistently with transferrin receptors, a marker for early endosomes [39] (Fig. 6b and c, respectively) and with TGN46 (Fig. 6d and e, respectively). TGN46 is a protein which resides predominantly in the trans-Golgi network and cycles between this compartment and the cell surface. After chloroquine treatment TGN46 is localized to swollen endosomes [40].

Fig. 6

Analysis of co-localization of endothelin converting enzyme-1 (ECE-1) with intracellular organelle markers after chloroquine treatment. Human smooth muscle cell cultures were incubated with 0.1 mmol/l chloroquine for 3 h at 37°C, then fixed and immunostained. Composite confocal images are from 5 focal planes. Micrograph (a) was immunostained with the rabbit antiserum to ECE-1 followed by anti-rabbit biotin/streptavidin fluorescein. The enzyme is concentrated mainly in an intracellular organelle, although some microfilament staining remains (arrow). The images in (b) and (c) illustrate organelles double immunostained with the rabbit antiserum to ECE-1 (b) and the mouse antiserum to the transferrin receptor (c), followed by the appropriate fluorescein and rhodamine conjugates, respectively. Examples of the co-localization of ECE-1 and the transferrin receptor are marked by arrows. A pair of confocal images (d) and (e), depict intracellular organelles double immunostained with the mouse antiserum to ECE-1 and the anti-mouse biotin/streptavidin fluorescein system (d) and the rabbit antiserum to TGN46 and an anti-rabbit rhodamine conjugate (e). The colocalization of ECE-1 and TGN46 is marked by arrows. Bars: 10 μm.

4 Discussion

Endothelin-converting enzyme is the key enzyme in the processing of big ET-1 to ET-1, which has an important physiological role in the mitogenesis and proliferation of SMC. ET-1 is secreted mainly from endothelial cells, but significant amounts are also produced in vascular SMC isolated from human left ventricle [4]. Recently the significance of the role of ECE-1 in certain hyperproliferative vascular diseases has been emphasized with its localization in neointimal SMC and in atherosclerotic lesions [30,31]. Furthermore, the blockade of ECE-1 by the metalloprotease inhibitor, phosphoramidon, was effective in reducing the size of the intima after balloon injury, thus implicating the importance of the enzyme as a target for the development of more specific drug therapies. Here we present biochemical data which establish the presence of a phosphoramidon-sensitive ECE-1 on SMC membranes from umbilical artery and immunological data which clarify its subcellular localization on cultured SMC. The relatively widespread detection of ECE-1 in smooth muscle cells where ET-1 synthesis is low indicates other possible physiological roles for the enzyme in these cells. This is consistent with the ability of ECE to hydrolyse other peptide substrates, for example, bradykinin [36].

It has been suggested that the heterogeneous occurrence of ECE-1 on subsets of SMC may be related to phenotypic differences of the contractile and synthetic smooth muscle cell types [31] and that this may explain why ECE-1 has not been universally detected on these cells in previous studies (for example, see [29]). Here we demonstrate for the first time by confocal immunofluorescence microscopy the striking localization of α-actin and ECE-1 in human SMC and rat SMC. Co-localization of ECE-1 and α-actin remains even after treatment of SMC with cytochalalasin D, which induces dramatic changes in the organization of actin microfilaments of bovine tracheal SMC [41] This co-localization, together with the detection of an 120 kDa protein in the intracellular fraction from the biotinylation experiment, indicates the presence of an intracellular pool of ECE-1 in smooth muscle cells, similar to that observed in endothelial cells [34]. These data are in accordance with the ultrastructural studies showing an intracellular siting for ECE-1 on vesicles in pig-lung endothelial cells, and with previous electron microscopical data locating ET-1 and the ECE substrate, big ET-1, on vesicles prepared from bovine aortic endothelial cells [34,42].

The occurrence of ECE-1 at the plasma membrane of SMC was confirmed by the presence of ECE-1 in the biotinylated cell surface fraction, where a phosphoramidon-sensitive ECE has been previously detected [28]. The clustered appearance of the plasma membrane ECE-1 observed by the immunofluorescence analysis in non-permeabilized SMC was first noted in endothelial cells [29,32]. It has been proposed [34] that endothelial cell ECE-1 cycles from the cell surface to an intracellular compartment. Endothelial ECE-1 was localized with the trans-Golgi network protein, TGN46, in neutralized swollen endosomes, after treatment of cell monolayers with the lysosomotrophic drug, chloroquine [34]. TGN46 recycles from the cell surface via endosomes to the trans-Golgi network [43]. After chloroquine treatment, SMC ECE-1 was also localized with TGN46 and with an early endosomal marker, the transferrin receptor. We suggest that, like the endothelial cell ECE-1, the SMC ECE-1 cycles from the cell surface to endosomes.

When endothelial cells are incubated with the metalloprotease inhibitors, phosphoramidon and thiorphan, ECE-1 relocates from the cell surface to a perinuclear Golgi region [32]. Concomitantly, there is an increase in the level of functionally active ECE-1 protein, similar to that demonstrated in the the rat and human SMC. This increase in ECE-1 protein may indicate a disruption in the turnover of the enzyme, which it has been suggested is due to inhibition of a metalloprotease which degrades ECE-1 [32,34,44]. Phosphoramidon treatment following balloon injury leads to an eventual decrease in the constitutive production of ET-1 and subsequent reduction in the size of the intima [31]. An accumulation of ECE-1 in a perinuclear position in SMC also occurs after disruption of the microtubule network by colchicine. This perinuclear accumulation may be explained by the role that microtubules and microfilaments have in organelle transport, for example, in neurons [45]. More recently it has also been suggested that myosin II is involved in the transport of particular basolateral proteins, for example, Vesicular Stomatitis G protein, via the constitutive pathway to the cell surface [46]. ET-1 is secreted constitutively and big ET-1 and ECE-1 have been located together on vesicles by immunoelectron microscopy [34]. Thus, one explanation for the striking co-localization of ECE-1 and α-actin in SMC is that the microfilaments provide a transport route for ECE-1 to the cell surface. We propose that ECE-1 sited on vesicles is trafficking in SMC initially on the microtubules and subsequently via the microfilaments to the cell surface.

Acknowledgements

We thank the British Heart Foundation for financial support, Dr Kazuhiko Tanzawa for the monoclonal antibody and cDNA to ECE-1 and Carolyn D. Brown for preparation of the polyclonal antibody to ECE-1. The rabbit polyclonal antibody, P12, to human TGN 46 was kindly donated by Dr Sreenivasan Ponnambalam. We thank Dr John H. Walker, University of Leeds, for helpful discussions and for reading the manuscript.