- Email: [email protected]
Paxillin-associated focal adhesion involvement in perinatal pulmonary arterial remodelling
Matrix Biology 22 (2003) 193–205
Paxillin-associated focal adhesion involvement in perinatal pulmonary arterial remodelling Ibrahima Diagnea, Susan M. Halla, Shigetoyo Kogakia, Cay M. Kieltyb, Sheila G. Hawortha,* a Vascular Biology & Pharmacology Unit, Institute of Child Health, 30 Guilford Street, London WC1N 1EH, UK Wellcome Trust Centre for Cell-Matrix Research, The University of Manchester, 2.205 Stopford Building, Oxford Road, Manchester M13 9PT, UK
b
Received 15 July 2002; received in revised form 9 January 2003; accepted 15 January 2003
Abstract Birth is followed by remodelling of the actin cytoskeleton of pulmonary arterial smooth muscle cells, then by extracellular matrix deposition. Hypothesising that the cellymatrix adhesions would also be remodelled, we investigated the expression, localisation and biochemical characteristics of the focal adhesion protein paxillin in vivo, in vessels from normal and pulmonary hypertensive neonatal piglets. Initially we showed that in intact porcine pulmonary arteries exposed to cytochalasin D there was a reduction filamentous actin accompanied by a reduction in paxillin-associated focal adhesions, similar to that seen in cultured pulmonary arterial smooth muscle cells. Vessels from normal and hypoxic animals were found to have two isoforms of paxillin, of 60 and 66 kDa with pI values of 6.7–4.2. Transient changes occurred during the first 14 days of life. Between birth and 6 days there was a reduction in the amount of both paxillin isoforms, a shift to more acidic pI values and an increase in paxillin phosphorylation. Simultaneously, immunostaining showed a transient reduction in paxillin expression, a change temporally and spatially associated with a previously demonstrated reduction in actin. Findings are consistent with an immediate postnatal spatial reorganisation of paxillin-associated focal adhesions. Paxillin content and remodelling was abnormal in pulmonary hypertensive arteries, the response varying according to postnatal age. 䊚 2003 Elsevier Science B.V.yInternational Society of Matrix Biology. All rights reserved. Keywords: Paxillin; Phosphorylation; Focal adhesions; Pulmonary arteries
1. Introduction Immediately after birth the pulmonary arteries are remodelled as pulmonary vascular resistance falls. The remodelling process entails rapid changes in the shape and actin cytoskeleton of the smooth muscle cells, and the deposition of extracellular matrix (Hall and Haworth, 1987; Hall et al., 2000; Haworth et al., 1987; Allen and Haworth, 1988). It seems inevitable that this process involves changes in the activity and composition of the focal adhesion complexes linking these structures. One of the most important focal adhesion proteins is paxillin (Turner et al., 1990) which has an important role in focal adhesion assembly (Burridge and ChrzanowskaWodnicka, 1996), the formation of stress fibres, cytoskeletal remodelling and integrin-mediated signal *Corresponding author. Tel.: q44-20-7813-8459; fax: q44-207905-2321. E-mail address: [email protected] (S.G. Haworth).
transduction (Lewis and Schwartz, 1998). Thus, paxillin is involved in actin cytoskeletal assembly and cell adhesion to the extracellular matrix. In cultured cells it co-localises with talin and vinculin at the ends of actin stress fibres and can interact directly with the small vinculin fragment (Rod) (Turner et al., 1990; Turner and Miller, 1994; Wood et al., 1994), focal adhesion kinase (FAK) (Schaller and Parsons, 1994; Tachibana et al., 1995) and the b1 integrin tail (Turner and Miller, 1994; Schaller et al., 1995). Paxillin exists in multiple isoforms (Turner et al., 1990; Mazaki et al., 1997) has a molecular weight of 65–70 kDa (Turner et al., 1990) and can be phosphorylated on tyrosine, serine and threonine residues (De Nichilo and Yamada, 1996; Bellis et al., 1997; Slack, 1998). It exists in the cell in both phosphorylated and non-phosphorylated forms, but the tyrosine-phosphorylated form is preferentially localised to focal adhesions (Cattelino et al., 1997). Up-regulation of tyrosine phosphorylation of paxillin and cytoskeletal
0945-053X/03/$30.00 䊚 2003 Elsevier Science B.V.yInternational Society of Matrix Biology. All rights reserved. doi:10.1016/S0945-053X(03)00011-8
194
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
remodelling are known to occur when tracheal smooth muscle contracts, observations that have implications for the neonatal pulmonary arteries as pulmonary vascular resistance falls after birth (Wang et al., 1996; Mehta et al., 1998). We have previously described a transient reduction in the amount of smooth muscle-specific actin and transient disassembly of the actin cytoskeleton during the first week of life in the smooth muscle cells of large intrapulmonary arteries (Hall et al., 2000). We hypothesised that these changes, together with the simultaneous deposition of extracellular matrix (Hall and Haworth, 1987; Kitley et al., 2000), would be associated with temporal and spatial changes in paxillin protein expression and by heightened focal adhesion complex activity, as evidenced by changes in paxillin phosphorylation. We began by comparing the distribution of actin filaments and paxillin in the wall of porcine elastic pulmonary arteries and in cells cultured from the vessel wall before and after exposure to cytochalasin D. This substance promotes depolymerisation of the actin cytoskeleton. Having confirmed that actin cytoskeletal disassembly changed paxillin distribution in our tissue of interest, we then investigated changes in paxillin expression and phosphorylation over the period of normal postnatal adaptation in the pig, using biochemical and immunohistochemical techniques. In addition, we compared these results with those obtained from piglets exposed to chronic hypoxia (Haworth and Hislop, 1982; Tulloh et al., 1997). Previous studies showed that such animals develop pulmonary hypertension with right ventricular hypertrophy and pulmonary arterial medial hypertrophy, and those exposed from birth continue to shunt from right to left through foetal channels and have a systemic arterial oxygen saturation of 71"5% (Tulloh et al., 1997). The piglet was used as the experimental animal because of the structural and haemodynamic similarities between the normal porcine and human pulmonary circulation during early life (Haworth and Hislop, 1981). 2. Experimental procedures 2.1. Source of tissue Lung tissue was obtained from healthy Large White pigs killed 1 week preterm (foetal), from piglets aged 5 min (newborn) and 3, 6 and 14 days, and from 6month-old adult animals. Additional animals were exposed to hypobaric hypoxia (50.8 kPa) for 3 days, from birth to 3 days, from 3 to 6 days, and from 14 to 17 days (Tulloh et al., 1997). Each age group, both of normal and hypoxic pigs, included at least six animals. This work was carried out under British Home Office Project Licence number PPL 70y4455. The animals received humane care in compliance with British Home Office Regulations and with the Principles of Labora-
tory Animal Care formulated by the National Society of Medical Research and the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (Department of Health and Human Services Publication NIH 80-23, Revised 1996). All immature animals were killed with a lethal injection of sodium pentobarbital (100 mgykg) and adult tissue was obtained from an abbatoir. 2.2. Relationship between smooth muscle-cell actin cytoskeleton and paxillin in the pulmonary arterial wall and in cultured smooth muscle cells 2.2.1. Preparation and immunofluorescent staining of the pulmonary arterial wall A 1-cm length of the proximal half of the lower-lobe intrapulmonary artery was taken from three 3-day-old and two 14-day-old piglets. Each artery was cut into nine rings of approximately 1 mm in length. Three rings were fixed immediately and the remaining six were incubated in serum-free medium (199) (Gibco) for 1 h at 37 8C. Three of these rings were then incubated for 30 min in 199 with 1 mM cytochalasin D and 0.05% DMSO to promote actin filament dissociation, and three in 199q0.05% DMSO as controls. The rings were fixed in 1% paraformaldehyde for 4 h, then washed in phosphate-buffered saline (PBS) and prepared for examination by confocal microscopy. Each ring was cut open, the endothelium was scraped off and the adventitia stripped away. The resultant sheet of arterial smooth muscle was stripped into two or three layers. Strips were permeabilised and pre-incubated with 1% bovine serum albumen (BSA) in PBSq0.01% Triton-X 100 (blocking solution), then incubated with mouse monoclonal antipaxillin (1:100 clone Z035, Zymed Labs Inc, San Francisco, US). Paxillin staining was visualised by incubation with a fluorescein-conjugated goat antimouse secondary antibody (Sigma, Poole, UK) and the actin cytoskeleton by incubation with rhodamine-conjugated phalloidin (1:40) (Molecular Probes, Leiden, Netherlands). Nuclei were visualised by staining with ToPro-3 (Molecular Probes). Control tissue strips were prepared, in which either incubation with anti-paxillin or phalloidin was omitted or a fluorescein-conjugated anti-goat antibody was used as a secondary antibody. The tissues were mounted with Vectashield (Vector Laboratories, Peterborough, UK) and smooth muscle layers were examined en face by confocal microscopy (BioRad MRC-600) using a Nikon E1000 microscope with a 63= water immersion lens. Fluorescent images of sheets of smooth muscle cells were obtained by sequential exposure to excitation wavelengths of 488 nm for fluorescein-labelled paxillin (green), 568 nm for rhodamine-conjugated phalloidin, F-actin (red) and 645 nm for ToPro-3-labelled DNA (blue). Optical sections through the cell sheets were
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
195
acquired at steps of 0.1 mm, and two-dimensional projections of 0.3-mm-thick sections were used to visualise the relationship between paxillin and the actin cytoskeleton.
Sections were lightly counterstained with Ehrlich’s haematoxylin (blue) before examination.
2.2.2. Preparation and immunofluorescent labelling of cultured smooth muscle cells Using the same animals as those used to study the vessel wall, smooth muscle cells were dissociated from the same segment of the vessel in the contralateral lung, and dissociated in a collagenase- and elastase-containing medium (Hall and Haworth, 1996). Cells were plated in Medium 199q10% foetal calf serum onto uncoated glass coverslips and allowed to settle and spread for 72 h. The medium was then removed and replaced with 199q2% foetal calf serum for a further 24 h. At this time the coverslips remained subconfluent. Cells were then incubated for 1 h in 5=10y7 M cytochalasin D in HEPES-buffered 199 at 37 8C to enhance actin depolymerisation. Control cover slips were incubated at 37 8C in HEPES-buffered 199, with or without 0.05% DMSO. Cells were fixed with 4% paraformaldehyde, permeabilised and blocked with 1% BSA in PBS before incubation with mouse monoclonal anti-paxillin, and either rhodamine-conjugated phalloidin or polyclonal goat antivinculin (1:100; Santa-Cruz Biotechnology Inc, Santa Cruz, CA) as described above. Light-microscopic immunofluorescent images were acquired using a Zeiss Axio-
2.3.2.1. SDS gel analysis. Preparation of the tissue lysate. At all ages, intrapulmonary arteries were dissected from the proximal half of a lower lobe of at least six animals, snap frozen in liquid nitrogen and stored at y80 8C. Tissues were disrupted in liquid nitrogen, homogenised in lysis buffer w1% sodium dodecyl sulfate (SDS), 1.0 mM sodium orthovanadate, 10 mM Tris, pH 7.4x and heated to 100 8C for 5 min, then centrifuged. The protein concentration of the supernatants was measured for immunoprecipitation, using the dot METRIC娃 protein assay kit (Chemicon International Inc, Tenecula, CA). Preliminary Western blot analysis revealed that paxillin formed a very small proportion of the total protein in the tissue lysate, and therefore biochemical analysis was carried out using immunoprecipitated samples. Immunoprecipitation. A 300-mg sample of total lysate was made up to a volume of 500 ml with H2O. To this was added 2 mg of mouse anti-paxillin (clone 177, Transduction Laboratories, Lexington, KY) and 500 ml of immunoprecipitation buffer (1% Triton X-100, 150 mM NaCl, 10 mM Tris, pH 7.4, 1 mM EDTA, 1 mM EGTA, pH 8, 0.2 mM sodium orthovanadate, 0.2 mM PMSF (Phenylmethy Sulfonyl Fluoride), 0.5% NP-40) at twice the normal concentration. After 2 h of incubation at 4 8C, 6 mg of rabbit anti-mouse IgG antibody (Sigma M-6024) was added, and incubation was continued for 1 h before addition of 25 ml of protein Aagarose beads (Sigma P-2545). The mixtures were centrifuged and the supernatant removed. The immunoprecipitates were washed and centrifuged three times at 16 000=g in immunoprecipitation buffer at 4 8C. Paxillin was extracted by boiling each pellet for 5 min with 25 ml of sample buffer. Thus, the concentration of immunoprecipitated paxillin in the final isolate was directly related to the known soluble protein concentration of the initial vessel wall isolate, and therefore the paxillin concentration in the vessel wall could be compared in different animals. Western blotting. Initially, gels were loaded to compare paxillin in samples from porcine pulmonary arteries with a lysate of human endothelial cells derived from aortic endothelium cell lines (Transduction Laboratories). Control lanes were loaded with buffers, mouse anti-paxillin used for immunoprecipitation and molecular weight markers (Sigma). In later gels, equal volumes of immunoprecipitated samples from one animal at each age, normal and hypoxic, were compared. Samples were loaded onto 7.5% or 10% discontinuous polyacrylamide SDS gels, electrophoresed and the proteins transferred to nitro-cellulose membranes (Trans-Blot娃, BioRad
skop 2 microscope with a 63= oil immersion objective and a Hamamatsu C47442y95 monochrome digital camera (Hamamtsu Ltd, Japan) and OPENLAB 2.03 software (Improvision Ltd, Coventry, UK). 2.3. Studies on the intact vessel wall 2.3.1. Structural studies 2.3.1.1. Immunocytochemical staining of tissue sections. In both normal and hypoxic animals (ns3 animals at each age, normal and hypoxic) blocks of tissue were taken from regions of the mid-lung containing the axial artery and airway. Tissue was fixed in 10% formol saline and, following routine histological processing, 5-mm sections were cut and mounted on Vectabond-coated slides. Following antigen unmasking and blocking for both endogenous peroxidase activity and non-specific antibody binding, sections were incubated for 1 h at room temperature with either monoclonal anti-paxillin (1:1000; Zymed Labs Inc), monoclonal anti-vinculin (1:200; Sigma) or monoclonal anti-b1 integrin (clone 18, 1:750; Transduction Laboratories, Lexington, KY). Antibody binding was visualised using the Dako streptavidin AB complex HRP kit (Dako Ltd, Ely, UK) and diaminobenzidine, producing a brown positive signal.
2.3.2. Biochemical studies
196
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
Laboratories Ltd, Hemel Hempstead, UK). Membranes were blocked for 1 h in PBS, pH 7.2, containing 0.1% Tween 20, and either 5% non-fat dried milk (antipaxillin), 1% BSA (anti-phosphotyrosine) or 5% BSA (anti-phosphoserine) before incubation with either antipaxillin (1:3000), anti-phosphotyrosine (1:2500) or antiphosphoserine (clone PSR-45, 1:150; Sigma) antibodies. Immunoreactive bands were visualised by chemiluminescence using horseradish peroxidase-conjugated antimouse IgG and ECL娃 reagent (RPN 2108, Amersham Pharmacia Biotech UK Ltd, Little Chalfont, UK). Quantification. Autoradiographic films of the immunoreacted blots were scanned using the Alpha Imager 1200 system (Flowgen, Staffordshire, UK). The digitised pictures were then analysed using SCION IMAGE for Windows (release beta 3b, Scion Corporation, www.scioncopr.com). The results of the densitometer analysis were transferred to EXCEL (Microsoft) and MINITAB (release 12, www.minitab.com) for statistical analysis. Data from the immature animals were standardised against the adult values, which were taken as 100%. Findings were compared in animals of different ages. In addition: (1) the total amount of paxillin and both the 66- and 60-kDa paxillin isoforms were compared in normal and hypoxic animals. (2) The amounts of phosphotyrosine- and phosphoserine-labelled paxillin were compared in normal and hypoxic animals. (3) The amounts of phosphotyrosine- and phosphoserine-labelled paxillin were related to total paxillin (in each sample) in both normal and hypoxic animals, and compared between normal and hypoxic animals. Statistical analyses were made using analysis of variance and the general linear model. Differences were taken as statistically significant at P-0.05. 2.3.2.2. Two-dimensional gel analysis. First dimension: isoelectric focusing. First-dimension isoelectric focusing (IEF) was carried out using a Multiphor horizontal electrophoresis system with a 7-cm immobilised pH gradient (IPG) strips (pH 4–7) (Amersham Pharmacia Biotech UK Ltd). The samples (ns2 normal animals per age and two 3-day-old hypoxic animals) were applied overnight using the in-gel rehydration method (Sanchez et al., 1997). The rehydration solution
contained 9 M urea, 4% CHAPS (3-w(3-Cholamidopropyl)-dimethyl-ammoniox-1-propane Sulphonate), 1% dithiothreitol, 1% ampholyte pH 4–6 and 1% Biolyte pH 5–7. A total sample volume of 125 ml, which contained 50 mg of immunoprecipitated protein, was loaded per strip. IEF was carried out at 20 8C for a maximum of 50 mAystrip (5 W limiting) and a maximum of 3500 V for a total of 9000 V h. After IEF was complete, the strips were frozen at y70 8C until used for the second dimension electrophoresis. Second dimension: SDS-PAGE. Before being subjected to SDS separation, the strips were equilibrated twice, and during the second equilibration step 4.8% (wyv) of iodoacetamide was added to the equilibration buffer. The strips were then placed on top of the vertical SDS gel and embedded in agarose. The second dimension was run, transferred and visualised as described above (Section 2.3.2.1.3). 2.3.3. Northern blotting A plasmid of human paxillin cDNA derived from kidney (GenBank accession no A1373645, IMAGE ID 2030636, clone 4997-g21) having high sequence similarity with paxillin derived from several mammalian species (CLUSTALX) was used as a template. The plasmid was cut using HindIII to linearise the probe and with AccIII to eliminate poly-A non-specific binding. The resulting fragment was used as a template to generate a probe. The radiolabelled probe was synthesised using random hexanucleotide. Northern blotting was carried out using pulmonary artery smooth-muscle mRNA with the expressHyb system (Clontech Laboratories UK Ltd, Basingstoke, UK). Hybridisation was carried out at 54 8C and the rinses at 60 8C. Bands were visualised using a phosphorimager (Typhoon, Amersham Pharmacia Biotech UK Ltd). 3. Results 3.1. Relationship between smooth-muscle-cell actin cytoskeleton and paxillin in the pulmonary artery wall and in cultured smooth muscle cells 3.1.1. Immunostaining of arterial wall Confocal microscopy of pulmonary arterial smooth muscle from arteries incubated at 37 8C in serum-free
Fig. 1. (A) Confocal images of intact pulmonary arterial smooth muscle at 14 days of age stained for paxillin (green) and F-actin (phalloidin; red); nuclei stained with ToPro-3 (blue) in merged image. In medium 199-incubated tissue (a–c) regions of paxillinyactin co-localisation are indicated by arrowheads. Following F-actin depolymerisation by cytochalasin D (d–f) paxillin is reduced at the cell margins and concentrated in the perinuclear cytoplasm (asterisks). (g–i) Unincubated tissue is indistinguishable from that of medium 199-incubated controls. (j) The specificity of paxillin staining is confirmed by the lack of green fluorescence following incubation with an anti-goat secondary antibody, shown by merged green and blue channels. All images taken at the same magnification. Bar, 20 mm. (B) Light-microscope immunofluorescent images of pulmonary arterial smooth muscle cells cultured from a 3-day-old animal: (a,d) paxillin, (b,e) phalloidin-bound F-actin, and (c,f) merged paxillin (green) and phalloidin (red) channels. In untreated controls (a–c) paxillin is present in large focal adhesions at the end of actin stress fibres, frequently co-localising with actin (arrowheads). Following incubation with cytochalasin D (d–f) actin stress fibres decrease and paxillin is concentrated in the perinuclear cytoplasm (asterisks). Bar, 25 mm.
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
Fig. 1.
197
198
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
medium from both 3- and 14-day-old animals and in control untreated vessels, double labelled for paxillin and F-actin, showed that patches of paxillin were present at or close to the cell periphery (Fig. 1A, a–c, g–i). Bundles of F-actin frequently terminated at, and colocalised with, these patches of paxillin. By contrast, in pulmonary arterial smooth muscle treated with cytochalasin D to promote actin depolymerisation, F-actin staining at the cell margin was weaker and paxillin was diffusely distributed throughout the cytoplasm (Fig. 1A, d–f). 3.1.2. Immunostaining of cultured smooth muscle cells The distribution of actin and paxillin in cells isolated from the large intrapulmonary arteries at 3 and 14 days was similar to that observed in intact tissue. In control untreated cells, paxillin co-labelled with vinculin (data not shown). Paxillin was concentrated in focal adhesion plaques at the end of actin stress fibres at the cell margin (Fig. 1B, a–c). Plaques were also present below the cell body. Following treatment with cytochalasin D, cells became more rounded in shape, paxillin-labelled focal adhesions were reduced in size and number, the actin stress fibres were thinner and fewer, and paxillin immunostaining became diffused throughout the cytoplasm, (Fig. 1B, d–f). Thus, in smooth muscle cells within the pulmonary artery wall and in cultured smooth muscle cells, a reduction in actin stress fibres was associated with a reduction in paxillin-labelled focal adhesions and redistribution of paxillin within the cell. 3.2. Characteristics of porcine paxillin 3.2.1. Biochemical studies At all ages, in both normal animals and in those exposed to chronic hypoxia, one-dimensional SDS gel electrophoresis of immunoprecipitated samples of pulmonary artery smooth muscle consistently demonstrated major bands at 60 and 66 kDa that reacted with paxillin antibody (Fig. 2a). An additional doublet was observed at 44–46 kDa. The 60- and 66-kDa porcine bands migrated at a similar level to the two paxillin immunoreactive bands from human endothelial lysate that was loaded for comparison (Fig. 2a). The 60-kDa band reacted with antibodies to both phosphotyrosine and phosphoserine, but the 66-kDa band did not react with the antibody to either (Fig. 2b,c). Two-dimensional electrophoresis of adult tissue showed a thick band of protein having isoelectric point (pI) values of between 5.5 and 4.2 (Fig. 2d). The 60- and 66-kDa isoforms could not be resolved within this band. Phosphotyrosine and phosphoserine co-localised with regions of the paxillin labelling, although more spots labelled for phosphoserine than phosphotyrosine (data not shown).
3.2.2. Paxillin mRNA analysis Northern blots of porcine paxillin revealed the presence of only a broad band of approximately 3.2 kbp of paxillin mRNA in all the pulmonary arterial tissue examined from foetal, 6- and 14-day-old and adult animals (Fig. 3). Northern blots of paxillin from piglets exposed to hypoxia from 3 to 6 days and from 14 to 17 days similarly showed a broad band of mRNA in the same position. 3.3. Developmental studies on porcine pulmonary arteries 3.3.1. Immunocytochemical studies on the intact pulmonary arterial wall 3.3.1.1. Normal. In all specimens examined at all ages, paxillin immunostaining was stronger in elastic arteries than in large and small muscular arteries, and was present in all the smooth muscle cells of the elastic arteries (Fig. 4). In the pre-term foetus and newborn animal, the smooth muscle cells in the sub-endothelial region were less intensely stained than those in the mid and outer media, although in adjacent sections of the same vessels the sub-endothelial region showed strong vinculin expression (Fig. 4). By contrast, at 3 and 6 days of age, both paxillin and vinculin immunostaining was weaker across the full width of the media than at birth and was weakest in the sub-endothelial region (Fig. 4). At 14 days and in the adult, paxillin and vinculin expression was greater than in the foetus–6day-old animals and was similar in all medial cell layers (Fig. 4). Expression of integrin-b1 in elastic pulmonary arteries showed postnatal changes in distribution similar to those of paxillin (data not shown). 3.3.1.2. Chronic hypoxia. The response to chronic hypoxia was dependent upon the age at which the animals were exposed to hypoxia. In the arteries of 3day-old animals exposed to chronic hypoxia from birth, the normal, post-natal transient reduction in paxillin, vinculin and integrin-b1 immunostaining did not occur and the appearance was similar to the normal at birth (Fig. 4). In vessels from 6-day-old animals exposed to chronic hypoxia from 3 days of age, the mid and outer medial cell layers were strongly immunostained for paxillin and integrin-b1, indicating an abnormal increase in expression at this site. Following exposure from 14 to 17 days of age, paxillin, vinculin and integrin-b1 expression remained normal for age. 3.3.2. Changes in paxillin protein expression during development Using Western blots from one-dimensional SDSPAGE gels (ns5 animals at each age) the density of immunostaining at each age was related to the adult
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
199
Fig. 2. (a) SDS-PAGE of immunoprecipitated paxillin from adult porcine intrapulmonary arteries (lanes 1 and 2) and a sample of lysate of human aortic endothelial cells (lane 4), showing the two porcine isoforms at 66 and 60 kDa with similar molecular weights to the two human isoforms (66 and 60 kDa), and the additional 44–46-kDa porcine doublet. Control (lane 3) loaded with immunoprecipitation buffer and mouse anti-paxillin. (b) A representative SDS-PAGE gel of immunoprecipitated paxillin from normal foetal (Ft), newborn (NB), 3-day (3D), 6-day (6D) 14-day (14D) and adult (AD) pigs and piglets exposed to chronic hypoxia from birth to 3 days (3H), from 3 to 6 days (6H) and from 14 to 17 days (17H) stained in the upper frame for paxillin and in the lower for phosphotyrosine. (c) A representative SDS-PAGE gel of immunoprecipitated paxillin from normal and chronically hypoxic animals as described in (b) stained in the upper frame for paxillin and in the lower for phosphoserine. (d) Immunoblots of two-dimensional electrophoresis gels of extracted pulmonary arterial tissue from normal newborn, 6-day-old and adult pigs. The gel margins are marked by arrows, molecular weights along the left-hand side and pI values at the bottom of each photograph. Paxillin is present over a range of pI values in a broad band at 60–66 and 45 kDa. There is a shift toward lower pI values between birth and adulthood.
value, which was taken as 100%. Each gel contained one lane loaded with paxillin immunoprecipitated from elastic pulmonary arteries of one animal at each age, normal and hypoxic, and the pattern of change with age was consistent in all experiments. 3.3.2.1. Normal. The total amount of paxillin, that is the sum of the 66- and 60-kDa bands, changed with age (P-0.01, ANOVA). The paxillin content of foetal and
newborn tissue was similar to that in adult tissue, but there was a transient postnatal reduction in paxillin content that was maximal at 6 days (Fig. 5a, P-0.01). The content was again higher at 14 days (P-0.01), being similar to that in the adult. This pattern of change with age was observed in the expression of both the 66and 60-kDa paxillin isoforms (Fig. 5b, P-0.01) although there were some differences between the two isoforms. The amount of the 66-kDa isoform was greater
200
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
Fig. 3. Northern blot of paxillin mRNA from adult pulmonary artery, showing a broad mRNA band at ;3.2 kbp (arrow). Northern blotting was performed using the mRNA-enriched fraction from total RNA and digoxigenin-labelled riboprobes.
in the newborn than in the foetal tissue (P-0.02), unlike the 60-kDa isoform. The amount of the 60-kDa isoform increased more slowly from its nadir at 6 days, and there was significantly less protein at 14 days than in the adult (P-0.02). Two-dimensional gel electrophoresis on tissue samples from newborn, 3-, 6- and 14day-old animals showed a shift toward more acidic pI values with increasing age, the greatest shift occurring between birth and 6 days of age (Fig. 2d). At birth and 3 days, the pI values of the paxillin immunolabelled spots ranged from approximately 6.7 to 4.2 and labelling was evenly distributed over this range. At 6 and 14 days of age, as in the adult, most of the paxillin lay between pI 5.5 and 4.2 (Fig. 2d). The shift in pI may indicate an increase in the level of paxillin phosphorylation (Pavalko et al. 1995). However, since at all ages the thickness of the band staining for paxillin remained maximal at pI 5.0–5.5, only a portion of total paxillin may show this change in phosphorylation. Phosphotyrosine and phosphoserine co-localised with some paxillin spots, but in many neither phosphorylation was detectable. At all ages more spots labelled for phosphoserine than for phosphotyrosine.
Fig. 4. Expression of paxillin protein and vinculin in adjacent sections of the same intrapulmonary elastic arteries from foetal, 3- and 14-day normal piglets and a piglet exposed to hypoxia from birth to 3 days of age. Paxillin and vinculin immunostaining show a similar pattern of expression. In foetal and 14-day-old preparations both proteins are present in all medial cell layers. However, at 3 days all cell layers stain less strongly, with the least stain being observed in adlumenal layers. By contrast, in the 3-day-old hypoxic animal, paxillin and vinculin are uniformly strongly expressed across the media, as in the foetus. All micrographs taken at the same magnification; bar, 50 mm.
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
201
Fig. 5. Histograms showing developmental changes in paxillin expression in elastic pulmonary arteries. Paxillin content is expressed as a percentage of the adult value. (a) Change in total paxillin content in normal animals. (b) Change in expression of 66- and 60-kDa isoforms in normal animals. (c) Influence of hypoxia on expression of 66- and 60-kDa paxillin isoforms, comparing paxillin expression in hypoxic animals to that of normal animals at ages corresponding to the start and end of exposure. Ages: foetal (FT), newborn (NB), 3-day (3D), 6-day (6D) 14-day (14D) and adult (AD) pigs. Piglets exposed to chronic hypoxia from birth to 3 days (3H), form 3 to 6 days (6H) and from 14 to 17 day (17H). Mean value "standard error of mean. #P-0.05.
3.3.2.2. Chronic hypoxia. Tissue from both 3- and 6day-old hypoxic animals contained more paxillin, and more 60- and 66-kDa isoforms than age-matched controls (P-0.01 for all comparisons) (Fig. 5c). This increase was greatest in the 66-kDa isoform of 6-dayold hypoxic animals (Fig. 5c). Thus, exposure to chronic hypoxia from birth to 3 days resulted in abolition of the normal postnatal reduction in the amount of paxillin, and exposure from 3 to 6 days of age enhanced paxillin expression. By contrast, there was no change in the expression of paxillin or in either of its isoforms following exposure to hypoxia from 14 to 17 days of age (Fig. 5c). On two-dimensional gel electrophoresis, the range of pI values of the paxillin immunolabelled spots in tissue from 3-day-old animals exposed to hypoxia from birth was similar to the normal at birth and 3 days of age.
3.3.3.1. Normal. When the amount of phosphotyrosine and phosphoserine in each sample was related to the amount of 60-kDa paxillin (phosphorylated and unphosphorylated) in that sample, then the proportion of paxillin phosphorylated on both tyrosine and on serine increased, by 18.5 and 25.5%, respectively (P-0.05 for both comparisons), during the first 6 days of life (Fig. 6a). At 14 days, there was less phosphotyrosine and phosphoserine than in either 3-day-old (P-0.01 for both phosphorylations) or adult specimens (P-0.01 for phosphoserine) (Fig. 6a).
3.3.3. Changes in phosphorylated paxillin during development In all blots examined (ns7 animals at each age) both anti-phosphotyrosine and anti-phosphoserine reacted with the 60-kDa paxillin band, but not with the 66-kDa band (Fig. 2b,c)
4. Discussion
3.3.3.2. Chronic hypoxia. The amount of phosphotyrosine- and phosphoserine-labelled paxillin only became abnormal after exposure to chronic hypoxia from 14 to 17 days, when both increased (P-0.01 for both comparisons) (Fig. 6b,c).
We have investigated the in vivo distribution and phosphorylation of paxillin as evidence of activity and remodelling of the focal adhesions linking the pulmonary arterial smooth muscle cell cytoskeleton to the
202
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
Fig. 6. Histograms showing developmental changes in paxillin phosphorylation in elastic pulmonary arteries. Phosphorylation is expressed as the ratio of phosphotyrosine or phosphoserine staining to that of total paxillin at each age. (a) Phosphorylation of paxillin in normal animals. (b,c) Influence of hypoxia on phosphorylation of paxillin, comparing phosphorylated paxillin expression in hypoxic animals with that of normal animals at ages corresponding to the start and end of the period of chronic hypoxia for (b) phosphotyrosine and (c) phosphoserine. #P-0.05.
extracellular matrix during adaptation to extrauterine life. We describe a transient reduction in paxillin in the pulmonary arterial media immediately after birth biochemically and by immunohistochemistry, a change temporally and spatially associated with the transient reduction in actin previously described (Hall et al., 2000). We also found a simultaneous increase in paxillin phosphorylation compatible with enhanced focal adhesion activity. The amount of paxillin and the level of tyrosine and serine phosphorylation was similar in the foetus and adult, suggesting greater cell-matrix stability at these two extremes of life, despite the pulmonary arteries being exposed to different pressures. 4.1. Characteristics of porcine paxillin The biochemical and molecular characteristics of the paxillin found in porcine pulmonary arteries were generally similar to those previously reported in other tissues in other species (De Nichilo and Yamada, 1996; Bellis et al., 1997; Mazaki et al., 1997; Slack, 1998; Thomas et al., 1999). The molecular weights were similar to those of the two paxillin isoforms present in human aortic endothelial cells. We also found a paxillin doublet of 44–46 kDa similar to that previously described in the human cell line 32Dc13 (Salgia et al., 1995). Others have characterised a 50–51-kDa doublet
as HIC-5, a protein having extensive homology with paxillin (Thomas et al., 1999). Our two-dimensional electrophoresis revealed several isoforms of paxillin, with pI values ranging from 6.7 to 4.2. Chicken gizzard contains at least four paxillin isoforms, but with similar pI values (Turner et al., 1990), and cultured U937 human monocytes have yielded as many as eight isoforms (Mazaki et al., 1997). In addition, tyrosine and serine phosphorylation led to the generation of multiple isoforms in cultured U937 cells (Mazaki et al., 1997). The process of isoform formation is regulated by the number of genes encoded for the same protein andyor alternative splicing of the mRNA, and by post-translational modification. In the present study, Northern blot analysis did not unequivocally demonstrate more than one porcine paxillin mRNA in pulmonary arterial tissue at any age in normal or hypoxic animals. A single mRNA has been found in mouse tissue (Mazaki et al., 1998), and in human monocytes and cancer cells (Salgia et al., 1995; Mazaki et al., 1997). Our findings suggest that, as in the chicken (Turner et al., 1990), the multiple paxillin isoforms present in the porcine tissue are likely to be the result of post-translational processing only. Alternatively, different isoforms varying only by different pI values may reflect differing levels of phosphorylation of a single (60 kDa) isoform (Tang et al., 1999).
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
4.2. Paxillin expression and phosphorylation during normal development The findings in the present study indicate for the first time that birth is followed by remodelling of paxillinassociated focal adhesions. The transient reduction in paxillin accompanied a reduction in the amount of actin and in the proportion of filamentous to globular cytosolic actin previously described (Hall et al., 2000). The observations made on the intact tissue in vivo correspond to the reduction in paxillin-labelled focal adhesions resulting from cytochalasin D-induced disruption of the actin cytoskeleton in intact pulmonary artery and in the cultured pulmonary artery smooth muscle cells. The postnatal reduction in paxillin was associated with an increase in paxillin phosphorylation, possibly caused by the increased strain imposed on the vessel wall when the pulmonary arteries were abruptly distended with blood after birth. Mechanosensitive tyrosine phosphorylation of paxillin and focal adhesion kinase (FAK) has been described in tracheal smooth muscle strips (Tang et al., 1999), and cultured airway smooth muscle cells exposed to an increase in strain show a rapid increase in tyrosine phosphorylation of both these proteins (Smith et al., 1998). Thus, a strong mechanical stimulus acting on the smooth muscle cell–focal adhesion complexes at birth could have led to their activation and remodelling, triggering actin cytoskeletal remodelling. Focal adhesions are linked to the extracellular matrix through integrin receptors. The a1–a7b1 integrins are the key vascular integrins in cell matrix interactions (Hedin, 1994) and the cytoplasmic tail of b1 is known to associate with paxillin (Schaller and Parsons, 1994). Among the major ligands of the b1 integrin, fibrillar collagen deposition increases immediately after birth (Allen and Haworth, 1988; Kitley et al., 2000). Paxillin phosphorylation accompanies cell adhesion to the extracellular matrix following the binding of integrins to their cognate ligand (Cattelino et al., 1997; Brown et al., 1998; Lewis and Schwartz, 1998). This may help to explain the postnatal increase in tyrosine and serine phosphorylation of paxillin described here. Although tyrosine phosphorylation has been extensively implicated in cytoskeletal re-organisation and cell adhesion (Schoenwaelder and Burridge, 1999), in the present study the postnatal increase in serine phosphorylation was particularly marked. During macrophage adhesion to vitronectin and fibroblast adhesion to fibronectin, 96% of paxillin phosphorylation occurred on serine residues (De Nichilo and Yamada, 1996; Bellis et al., 1997). Binding serine to LIM domains increases cell adhesion to fibronectin, with accelerated fibronectininduced localisation of paxillin to focal contacts (Brown et al., 1998). Thus, the high level of paxillin phosphorylation, on both tyrosine and serine residues, at 3 and
203
6 days indicates remodelling of focal adhesions as they detach and re-attach to the extracellular matrix when the pulmonary arteries rapidly increase in size. By 14 days of age, paxillin and vinculin expression in the media resembled that in the adult, but the amount of the 60-kDa paxillin isoform, the phosphorylated isoform, remained lower than in the adult, and tyrosine and serine phosphorylation was less than at 3 or 6 days. This transient reduction in focal adhesion activity takes place during a period of relative cytoskeletal stability (Hall et al., 2000). In vitro studies have shown that a reduction in paxillin and FAK tyrosine phosphorylation does not reduce actin stress fibre and focal adhesion formation, as demonstrated following eNOS gene transfer into systemic arterial smooth-muscle cells (Fang et al., 1997). eNOS might also play a similar role in vivo. Porcine eNOS activity is extremely low at birth and then increases rapidly during the first 14 days of life (Tulloh et al., 1997; Arrigoni et al., 2002). 4.3. Paxillin expression and phosphorylation in pulmonary hypertension The response to chronic hypoxia depended on the age of the animals at the onset of exposure. Exposure from birth to 3 days of age appeared to prevent the normal postnatal down-regulation in total paxillin expression and increase in proportion of phosphorylated paxillin, indicating failure of the focal adhesion complexes to remodel after birth. By contrast, when hypoxic exposure commenced at 3 days, after the normal process of cytoskeletal remodelling was established, our findings indicated abnormal actin remodelling. The amount of both isoforms increased, but the proportion of phosphorylated paxillin was normal. These findings are consistent with previous studies showing that in this experimental porcine model of neonatal pulmonary hypertension there is no transient postnatal reduction in actin myofilaments, rather an increase in the volume density of both cytoskeletal filaments and surface dense bodies (focal adhesions) (Allen and Haworth, 1986). There may also be a functional correlate. Tang et al. (1999) have shown that paxillin phosphorylation in intact tracheal smooth muscle is sensitive to length rather than tension, and chronic hypoxic pulmonary arteries are vasoconstricted. In animals first exposed to hypoxia at 14 days of age, the expression of paxillin was unaffected, but the proportion of phosphoserineand phosphotyrosine-labelled paxillin increased. This may reflect hypertensive remodelling of a fully dilated vessel that had a relatively stable cytoskeleton before exposure to the insult (Hall et al., 2000). Our data indicate that with the abrupt distension of the vessel wall at birth, the increase in strain is associated with, or leads to, degradation of focal adhesionassociated paxillin and a concomitant increase in paxillin
204
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
phosphorylation. Having accommodated to the extrauterine environment, paxillin phosphorylation diminishes. This study is one of few that relate in vivo and in vitro paxillin and cytoskeletal re-arrangements. Our in vivo observations support the contention that paxillin may play a novel role in regulating phosphorylation of cellular proteins in both a spatial and temporal manner (Shen et al., 1998), possibly influencing outside-in signalling at the interface between cell and extracellular matrix and inside-out cytoskeleton–focal adhesion integrin-mediated signal transduction. The necessity of carrying out detailed studies on these signal transduction pathways in early life is now self-evident. Acknowledgments We would like to thank Dr Michael J Dunn, National Heart and Lung Institute, Harefield Hospital for his advice in the preparation of two-dimensional gels. This work was funded by a British Heart Foundation Programme Grant. References Allen, K., Haworth, S.G., 1988. Human postnatal pulmonary arterial remodeling. Ultrastructural studies of smooth muscle cell and connective tissue maturation. Lab. Invest. 59, 702–709. Allen, K.M., Haworth, S.G., 1986. Impaired adaptation of pulmonary circulation to extrauterine life in newborn pigs exposed to hypoxia: an ultrastructural study. J. Pathol. 150, 205–212. Arrigoni, F.I., Hislop, A.A., Pollock, J.S., Haworth, S.G., Mitchell, J.A., 2002. Birth upregulates nitric oxide synthase activity in the porcine lung. Life Sci. 70, 1609–1620. Bellis, S.L., Perrotta, J.A., Curtis, M.S., Turner, C.E., 1997. Adhesion of fibroblasts to fibronectin stimulates both serine and tyrosine phosphorylation of paxillin. Biochem. J. 325, 375–381. Brown, M.C., Perrotta, J.A., Turner, C.E., 1998. Serine and threonine phosphorylation of the paxillin LIM domains regulates paxillin focal adhesion localization and cell adhesion to fibronectin. Mol. Biol. Cell 9, 1803–1816. Burridge, K., Chrzanowska-Wodnicka, M., 1996. Focal adhesions, contractility, and signaling. Annu. Rev. Cell Dev. Biol. 12, 463–519. Cattelino, A., Cairo, S., Malanchini, B., de Curtis, I., 1997. Preferential localization of tyrosine-phosphorylated paxillin in focal adhesions. Cell Adhes. Commun. 4, 457–467. De Nichilo, M.O., Yamada, K.M., 1996. Integrin alpha v beta 5dependent serine phosphorylation of paxillin in cultured human macrophages adherent to vitronectin. J. Biol. Chem. 271, 11016–11022. Fang, S., Sharma, R.V., Bhalla, R.C., 1997. Endothelial nitric oxide synthase gene transfer inhibits platelet-derived growth factor-BB stimulated focal adhesion kinase and paxillin phosphorylation in vascular smooth muscle cells. Biochem. Biophys. Res. Commun. 236, 706–711. Hall, S.M., Gorenflo, M., Reader, J., Lawson, D., Haworth, S.G., 2000. Neonatal pulmonary hypertension prevents reorganisation of the pulmonary arterial smooth muscle cytoskeleton after birth. J. Anat. 196, 391–403. Hall, S.M., Haworth, S.G., 1987. Conducting pulmonary arteries: structural adaptation to extrauterine life. Cardiovasc. Res. 21, 208–216.
Hall, S.M., Haworth, S.G., 1996. Effect of cold preservation on pulmonary arterial smooth muscle cells. Am. J. Physiol. 270, L435–L445. Haworth, S.G., Hislop, A.A., 1981. Adaptation of the pulmonary circulation to extra-uterine life in the pig and its relevance to the human infant. Cardiovasc. Res. 15, 108–119. Haworth, S.G., Hislop, A.A., 1982. Effect of hypoxia on adaptation of the pulmonary circulation to extra-uterine life in the pig. Cardiovasc. Res. 16, 293–303. Haworth, S.G., Hall, S.M., Chew, M., Allen, K., 1987. Thinning of foetal pulmonary arterial wall and postnatal remodelling: ultrastructural studies on the respiratory unit arteries of the pig. Virchows Arch. A 411, 161–171. Hedin, U., 1994. Extracellular matrix components and integrins in the control of arterial smooth muscle cell structure and function. J. Atheroscler. Thromb. 1, S39–S46. Kitley, R.M., Hislop, A.A., Hall, S.M., Freemont, A.J., Haworth, S.G., 2000. Birth-associated changes in pulmonary arterial connective tissue gene expression in the normal and hypertensive lung. Cardiovasc. Res. 46, 332–345. Lewis, J.M., Schwartz, M.A., 1998. Integrins regulate the association and phosphorylation of paxillin by c-Abl. J. Biol. Chem. 273, 14225–14230. Mazaki, Y., Hashimoto, S., Sabe, H., 1997. Monocyte cells and cancer cells express novel paxillin isoforms with different binding properties to focal adhesion proteins. J. Biol. Chem. 272, 7437–7444. Mazaki, Y., Uchida, H., Hino, O., Hashimoto, S., Sabe, H., 1998. Paxillin isoforms in mouse. Lack of the gamma isoform and developmentally specific beta isoform expression. J. Biol. Chem. 273, 22435–22441. Mehta, D., Wang, Z., Wu, M.F., Gunst, S.J., 1998. Relationship between paxillin and myosin phosphorylation during muscarinic stimulation of smooth muscle. Am. J. Physiol. 274, C741–C747. Pavalko, F.M., Adam, L.P., Wu, M.F., Walker, T.L., Gunst, S.J., 1995. Phosphorylation of dense-plaque proteins talin and paxillin during tracheal smooth muscle contraction. Am. J. Physiol. 268, C563–C571. Salgia, R., Li, J.L., Lo, S.H., et al., 1995. Molecular cloning of human paxillin, a focal adhesion protein phosphorylated by P210BCRyABL. J. Biol. Chem. 270, 5039–5047. Sanchez, J.C., Rouge, V., Pisteur, M., et al., 1997. Improved and simplified in-gel sample application using reswelling of dry immobilized pH gradients. Electrophoresis 18, 324–327. Schaller, M.D., Parsons, J.T., 1994. Focal adhesion kinase and associated proteins. Curr. Opin. Cell Biol. 6, 705–710. Schaller, M.D., Otey, C.A., Hildebrand, J.D., Parsons, J.T., 1995. Focal adhesion kinase and paxillin bind to peptides mimicking beta integrin cytoplasmic domains. J. Cell Biol. 130, 1181–1187. Schoenwaelder, S.M., Burridge, K., 1999. Bidirectional signaling between the cytoskeleton and integrins. Curr. Opin. Cell Biol. 11, 274–286. Shen, Y., Schneider, G., Cloutier, J.F., Veillette, A., Schaller, M.D., 1998. Direct association of protein-tyrosine phosphatase PTP-PEST with paxillin. J. Biol. Chem. 273, 6474–6481. Slack, B.E., 1998. Tyrosine phosphorylation of paxillin and focal adhesion kinase by activation of muscarinic m3 receptors is dependent on integrin engagement by the extracellular matrix. Proc. Natl. Acad. Sci. USA 95, 7281–7286. Smith, P.G., Garcia, R., Kogerman, L., 1998. Mechanical strain increases protein tyrosine phosphorylation in airway smooth muscle cells. Exp. Cell Res. 239, 353–360. Tachibana, K., Sato, T., D’Avirro, N., Morimoto, C., 1995. Direct association of pp125FAK with paxillin, the focal adhesion- targeting mechanism of pp125FAK. J. Exp. Med. 182, 1089–1099.
I. Diagne et al. / Matrix Biology 22 (2003) 193–205 Tang, D., Mehta, D., Gunst, S.J., 1999. Mechanosensitive tyrosine phosphorylation of paxillin and focal adhesion kinase in tracheal smooth muscle. Am. J. Physiol. 276, C250–C258. Thomas, S.M., Hagel, M., Turner, C.E., 1999. Characterization of a focal adhesion protein, Hic-5, that shares extensive homology with paxillin. J. Cell Sci. 112, 181–190. Tulloh, R.M.R., Hislop, A.A., Boels, P.J., Deutsch, J., Haworth, S.G., 1997. Chronic hypoxia inhibits postnatal maturation of porcine intrapulmonary artery relaxation. Am. J. Physiol. 272, H2436–H2445. Turner, C.E., Miller, J.T., 1994. Primary sequence of paxillin contains putative SH2 and SH3 domain binding motifs and multiple LIM
205
domains: identification of a vinculin and pp125Fak-binding region. J. Cell Sci. 107, 1583–1591. Turner, C.E., Glenney Jr, J.R., Burridge, K., 1990. Paxillin: a new vinculin-binding protein present in focal adhesions. J. Cell. Biol. 111, 1059–1068. Wang, Z., Pavalko, F.M., Gunst, S.J., 1996. Tyrosine phosphorylation of the dense plaque protein paxillin is regulated during smooth muscle contraction. Am. J. Physiol. 271, C1594–C1602. Wood, C.K., Turner, C.E., Jackson, P., Critchley, D.R., 1994. Characterisation of the paxillin-binding site and the C-terminal focal adhesion targeting sequence in vinculin. J. Cell Sci. 107, 709–717.
Paxillin-associated focal adhesion involvement in perinatal pulmonary arterial remodelling Ibrahima Diagnea, Susan M. Halla, Shigetoyo Kogakia, Cay M. Kieltyb, Sheila G. Hawortha,* a Vascular Biology & Pharmacology Unit, Institute of Child Health, 30 Guilford Street, London WC1N 1EH, UK Wellcome Trust Centre for Cell-Matrix Research, The University of Manchester, 2.205 Stopford Building, Oxford Road, Manchester M13 9PT, UK
b
Received 15 July 2002; received in revised form 9 January 2003; accepted 15 January 2003
Abstract Birth is followed by remodelling of the actin cytoskeleton of pulmonary arterial smooth muscle cells, then by extracellular matrix deposition. Hypothesising that the cellymatrix adhesions would also be remodelled, we investigated the expression, localisation and biochemical characteristics of the focal adhesion protein paxillin in vivo, in vessels from normal and pulmonary hypertensive neonatal piglets. Initially we showed that in intact porcine pulmonary arteries exposed to cytochalasin D there was a reduction filamentous actin accompanied by a reduction in paxillin-associated focal adhesions, similar to that seen in cultured pulmonary arterial smooth muscle cells. Vessels from normal and hypoxic animals were found to have two isoforms of paxillin, of 60 and 66 kDa with pI values of 6.7–4.2. Transient changes occurred during the first 14 days of life. Between birth and 6 days there was a reduction in the amount of both paxillin isoforms, a shift to more acidic pI values and an increase in paxillin phosphorylation. Simultaneously, immunostaining showed a transient reduction in paxillin expression, a change temporally and spatially associated with a previously demonstrated reduction in actin. Findings are consistent with an immediate postnatal spatial reorganisation of paxillin-associated focal adhesions. Paxillin content and remodelling was abnormal in pulmonary hypertensive arteries, the response varying according to postnatal age. 䊚 2003 Elsevier Science B.V.yInternational Society of Matrix Biology. All rights reserved. Keywords: Paxillin; Phosphorylation; Focal adhesions; Pulmonary arteries
1. Introduction Immediately after birth the pulmonary arteries are remodelled as pulmonary vascular resistance falls. The remodelling process entails rapid changes in the shape and actin cytoskeleton of the smooth muscle cells, and the deposition of extracellular matrix (Hall and Haworth, 1987; Hall et al., 2000; Haworth et al., 1987; Allen and Haworth, 1988). It seems inevitable that this process involves changes in the activity and composition of the focal adhesion complexes linking these structures. One of the most important focal adhesion proteins is paxillin (Turner et al., 1990) which has an important role in focal adhesion assembly (Burridge and ChrzanowskaWodnicka, 1996), the formation of stress fibres, cytoskeletal remodelling and integrin-mediated signal *Corresponding author. Tel.: q44-20-7813-8459; fax: q44-207905-2321. E-mail address: [email protected] (S.G. Haworth).
transduction (Lewis and Schwartz, 1998). Thus, paxillin is involved in actin cytoskeletal assembly and cell adhesion to the extracellular matrix. In cultured cells it co-localises with talin and vinculin at the ends of actin stress fibres and can interact directly with the small vinculin fragment (Rod) (Turner et al., 1990; Turner and Miller, 1994; Wood et al., 1994), focal adhesion kinase (FAK) (Schaller and Parsons, 1994; Tachibana et al., 1995) and the b1 integrin tail (Turner and Miller, 1994; Schaller et al., 1995). Paxillin exists in multiple isoforms (Turner et al., 1990; Mazaki et al., 1997) has a molecular weight of 65–70 kDa (Turner et al., 1990) and can be phosphorylated on tyrosine, serine and threonine residues (De Nichilo and Yamada, 1996; Bellis et al., 1997; Slack, 1998). It exists in the cell in both phosphorylated and non-phosphorylated forms, but the tyrosine-phosphorylated form is preferentially localised to focal adhesions (Cattelino et al., 1997). Up-regulation of tyrosine phosphorylation of paxillin and cytoskeletal
0945-053X/03/$30.00 䊚 2003 Elsevier Science B.V.yInternational Society of Matrix Biology. All rights reserved. doi:10.1016/S0945-053X(03)00011-8
194
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
remodelling are known to occur when tracheal smooth muscle contracts, observations that have implications for the neonatal pulmonary arteries as pulmonary vascular resistance falls after birth (Wang et al., 1996; Mehta et al., 1998). We have previously described a transient reduction in the amount of smooth muscle-specific actin and transient disassembly of the actin cytoskeleton during the first week of life in the smooth muscle cells of large intrapulmonary arteries (Hall et al., 2000). We hypothesised that these changes, together with the simultaneous deposition of extracellular matrix (Hall and Haworth, 1987; Kitley et al., 2000), would be associated with temporal and spatial changes in paxillin protein expression and by heightened focal adhesion complex activity, as evidenced by changes in paxillin phosphorylation. We began by comparing the distribution of actin filaments and paxillin in the wall of porcine elastic pulmonary arteries and in cells cultured from the vessel wall before and after exposure to cytochalasin D. This substance promotes depolymerisation of the actin cytoskeleton. Having confirmed that actin cytoskeletal disassembly changed paxillin distribution in our tissue of interest, we then investigated changes in paxillin expression and phosphorylation over the period of normal postnatal adaptation in the pig, using biochemical and immunohistochemical techniques. In addition, we compared these results with those obtained from piglets exposed to chronic hypoxia (Haworth and Hislop, 1982; Tulloh et al., 1997). Previous studies showed that such animals develop pulmonary hypertension with right ventricular hypertrophy and pulmonary arterial medial hypertrophy, and those exposed from birth continue to shunt from right to left through foetal channels and have a systemic arterial oxygen saturation of 71"5% (Tulloh et al., 1997). The piglet was used as the experimental animal because of the structural and haemodynamic similarities between the normal porcine and human pulmonary circulation during early life (Haworth and Hislop, 1981). 2. Experimental procedures 2.1. Source of tissue Lung tissue was obtained from healthy Large White pigs killed 1 week preterm (foetal), from piglets aged 5 min (newborn) and 3, 6 and 14 days, and from 6month-old adult animals. Additional animals were exposed to hypobaric hypoxia (50.8 kPa) for 3 days, from birth to 3 days, from 3 to 6 days, and from 14 to 17 days (Tulloh et al., 1997). Each age group, both of normal and hypoxic pigs, included at least six animals. This work was carried out under British Home Office Project Licence number PPL 70y4455. The animals received humane care in compliance with British Home Office Regulations and with the Principles of Labora-
tory Animal Care formulated by the National Society of Medical Research and the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (Department of Health and Human Services Publication NIH 80-23, Revised 1996). All immature animals were killed with a lethal injection of sodium pentobarbital (100 mgykg) and adult tissue was obtained from an abbatoir. 2.2. Relationship between smooth muscle-cell actin cytoskeleton and paxillin in the pulmonary arterial wall and in cultured smooth muscle cells 2.2.1. Preparation and immunofluorescent staining of the pulmonary arterial wall A 1-cm length of the proximal half of the lower-lobe intrapulmonary artery was taken from three 3-day-old and two 14-day-old piglets. Each artery was cut into nine rings of approximately 1 mm in length. Three rings were fixed immediately and the remaining six were incubated in serum-free medium (199) (Gibco) for 1 h at 37 8C. Three of these rings were then incubated for 30 min in 199 with 1 mM cytochalasin D and 0.05% DMSO to promote actin filament dissociation, and three in 199q0.05% DMSO as controls. The rings were fixed in 1% paraformaldehyde for 4 h, then washed in phosphate-buffered saline (PBS) and prepared for examination by confocal microscopy. Each ring was cut open, the endothelium was scraped off and the adventitia stripped away. The resultant sheet of arterial smooth muscle was stripped into two or three layers. Strips were permeabilised and pre-incubated with 1% bovine serum albumen (BSA) in PBSq0.01% Triton-X 100 (blocking solution), then incubated with mouse monoclonal antipaxillin (1:100 clone Z035, Zymed Labs Inc, San Francisco, US). Paxillin staining was visualised by incubation with a fluorescein-conjugated goat antimouse secondary antibody (Sigma, Poole, UK) and the actin cytoskeleton by incubation with rhodamine-conjugated phalloidin (1:40) (Molecular Probes, Leiden, Netherlands). Nuclei were visualised by staining with ToPro-3 (Molecular Probes). Control tissue strips were prepared, in which either incubation with anti-paxillin or phalloidin was omitted or a fluorescein-conjugated anti-goat antibody was used as a secondary antibody. The tissues were mounted with Vectashield (Vector Laboratories, Peterborough, UK) and smooth muscle layers were examined en face by confocal microscopy (BioRad MRC-600) using a Nikon E1000 microscope with a 63= water immersion lens. Fluorescent images of sheets of smooth muscle cells were obtained by sequential exposure to excitation wavelengths of 488 nm for fluorescein-labelled paxillin (green), 568 nm for rhodamine-conjugated phalloidin, F-actin (red) and 645 nm for ToPro-3-labelled DNA (blue). Optical sections through the cell sheets were
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
195
acquired at steps of 0.1 mm, and two-dimensional projections of 0.3-mm-thick sections were used to visualise the relationship between paxillin and the actin cytoskeleton.
Sections were lightly counterstained with Ehrlich’s haematoxylin (blue) before examination.
2.2.2. Preparation and immunofluorescent labelling of cultured smooth muscle cells Using the same animals as those used to study the vessel wall, smooth muscle cells were dissociated from the same segment of the vessel in the contralateral lung, and dissociated in a collagenase- and elastase-containing medium (Hall and Haworth, 1996). Cells were plated in Medium 199q10% foetal calf serum onto uncoated glass coverslips and allowed to settle and spread for 72 h. The medium was then removed and replaced with 199q2% foetal calf serum for a further 24 h. At this time the coverslips remained subconfluent. Cells were then incubated for 1 h in 5=10y7 M cytochalasin D in HEPES-buffered 199 at 37 8C to enhance actin depolymerisation. Control cover slips were incubated at 37 8C in HEPES-buffered 199, with or without 0.05% DMSO. Cells were fixed with 4% paraformaldehyde, permeabilised and blocked with 1% BSA in PBS before incubation with mouse monoclonal anti-paxillin, and either rhodamine-conjugated phalloidin or polyclonal goat antivinculin (1:100; Santa-Cruz Biotechnology Inc, Santa Cruz, CA) as described above. Light-microscopic immunofluorescent images were acquired using a Zeiss Axio-
2.3.2.1. SDS gel analysis. Preparation of the tissue lysate. At all ages, intrapulmonary arteries were dissected from the proximal half of a lower lobe of at least six animals, snap frozen in liquid nitrogen and stored at y80 8C. Tissues were disrupted in liquid nitrogen, homogenised in lysis buffer w1% sodium dodecyl sulfate (SDS), 1.0 mM sodium orthovanadate, 10 mM Tris, pH 7.4x and heated to 100 8C for 5 min, then centrifuged. The protein concentration of the supernatants was measured for immunoprecipitation, using the dot METRIC娃 protein assay kit (Chemicon International Inc, Tenecula, CA). Preliminary Western blot analysis revealed that paxillin formed a very small proportion of the total protein in the tissue lysate, and therefore biochemical analysis was carried out using immunoprecipitated samples. Immunoprecipitation. A 300-mg sample of total lysate was made up to a volume of 500 ml with H2O. To this was added 2 mg of mouse anti-paxillin (clone 177, Transduction Laboratories, Lexington, KY) and 500 ml of immunoprecipitation buffer (1% Triton X-100, 150 mM NaCl, 10 mM Tris, pH 7.4, 1 mM EDTA, 1 mM EGTA, pH 8, 0.2 mM sodium orthovanadate, 0.2 mM PMSF (Phenylmethy Sulfonyl Fluoride), 0.5% NP-40) at twice the normal concentration. After 2 h of incubation at 4 8C, 6 mg of rabbit anti-mouse IgG antibody (Sigma M-6024) was added, and incubation was continued for 1 h before addition of 25 ml of protein Aagarose beads (Sigma P-2545). The mixtures were centrifuged and the supernatant removed. The immunoprecipitates were washed and centrifuged three times at 16 000=g in immunoprecipitation buffer at 4 8C. Paxillin was extracted by boiling each pellet for 5 min with 25 ml of sample buffer. Thus, the concentration of immunoprecipitated paxillin in the final isolate was directly related to the known soluble protein concentration of the initial vessel wall isolate, and therefore the paxillin concentration in the vessel wall could be compared in different animals. Western blotting. Initially, gels were loaded to compare paxillin in samples from porcine pulmonary arteries with a lysate of human endothelial cells derived from aortic endothelium cell lines (Transduction Laboratories). Control lanes were loaded with buffers, mouse anti-paxillin used for immunoprecipitation and molecular weight markers (Sigma). In later gels, equal volumes of immunoprecipitated samples from one animal at each age, normal and hypoxic, were compared. Samples were loaded onto 7.5% or 10% discontinuous polyacrylamide SDS gels, electrophoresed and the proteins transferred to nitro-cellulose membranes (Trans-Blot娃, BioRad
skop 2 microscope with a 63= oil immersion objective and a Hamamatsu C47442y95 monochrome digital camera (Hamamtsu Ltd, Japan) and OPENLAB 2.03 software (Improvision Ltd, Coventry, UK). 2.3. Studies on the intact vessel wall 2.3.1. Structural studies 2.3.1.1. Immunocytochemical staining of tissue sections. In both normal and hypoxic animals (ns3 animals at each age, normal and hypoxic) blocks of tissue were taken from regions of the mid-lung containing the axial artery and airway. Tissue was fixed in 10% formol saline and, following routine histological processing, 5-mm sections were cut and mounted on Vectabond-coated slides. Following antigen unmasking and blocking for both endogenous peroxidase activity and non-specific antibody binding, sections were incubated for 1 h at room temperature with either monoclonal anti-paxillin (1:1000; Zymed Labs Inc), monoclonal anti-vinculin (1:200; Sigma) or monoclonal anti-b1 integrin (clone 18, 1:750; Transduction Laboratories, Lexington, KY). Antibody binding was visualised using the Dako streptavidin AB complex HRP kit (Dako Ltd, Ely, UK) and diaminobenzidine, producing a brown positive signal.
2.3.2. Biochemical studies
196
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
Laboratories Ltd, Hemel Hempstead, UK). Membranes were blocked for 1 h in PBS, pH 7.2, containing 0.1% Tween 20, and either 5% non-fat dried milk (antipaxillin), 1% BSA (anti-phosphotyrosine) or 5% BSA (anti-phosphoserine) before incubation with either antipaxillin (1:3000), anti-phosphotyrosine (1:2500) or antiphosphoserine (clone PSR-45, 1:150; Sigma) antibodies. Immunoreactive bands were visualised by chemiluminescence using horseradish peroxidase-conjugated antimouse IgG and ECL娃 reagent (RPN 2108, Amersham Pharmacia Biotech UK Ltd, Little Chalfont, UK). Quantification. Autoradiographic films of the immunoreacted blots were scanned using the Alpha Imager 1200 system (Flowgen, Staffordshire, UK). The digitised pictures were then analysed using SCION IMAGE for Windows (release beta 3b, Scion Corporation, www.scioncopr.com). The results of the densitometer analysis were transferred to EXCEL (Microsoft) and MINITAB (release 12, www.minitab.com) for statistical analysis. Data from the immature animals were standardised against the adult values, which were taken as 100%. Findings were compared in animals of different ages. In addition: (1) the total amount of paxillin and both the 66- and 60-kDa paxillin isoforms were compared in normal and hypoxic animals. (2) The amounts of phosphotyrosine- and phosphoserine-labelled paxillin were compared in normal and hypoxic animals. (3) The amounts of phosphotyrosine- and phosphoserine-labelled paxillin were related to total paxillin (in each sample) in both normal and hypoxic animals, and compared between normal and hypoxic animals. Statistical analyses were made using analysis of variance and the general linear model. Differences were taken as statistically significant at P-0.05. 2.3.2.2. Two-dimensional gel analysis. First dimension: isoelectric focusing. First-dimension isoelectric focusing (IEF) was carried out using a Multiphor horizontal electrophoresis system with a 7-cm immobilised pH gradient (IPG) strips (pH 4–7) (Amersham Pharmacia Biotech UK Ltd). The samples (ns2 normal animals per age and two 3-day-old hypoxic animals) were applied overnight using the in-gel rehydration method (Sanchez et al., 1997). The rehydration solution
contained 9 M urea, 4% CHAPS (3-w(3-Cholamidopropyl)-dimethyl-ammoniox-1-propane Sulphonate), 1% dithiothreitol, 1% ampholyte pH 4–6 and 1% Biolyte pH 5–7. A total sample volume of 125 ml, which contained 50 mg of immunoprecipitated protein, was loaded per strip. IEF was carried out at 20 8C for a maximum of 50 mAystrip (5 W limiting) and a maximum of 3500 V for a total of 9000 V h. After IEF was complete, the strips were frozen at y70 8C until used for the second dimension electrophoresis. Second dimension: SDS-PAGE. Before being subjected to SDS separation, the strips were equilibrated twice, and during the second equilibration step 4.8% (wyv) of iodoacetamide was added to the equilibration buffer. The strips were then placed on top of the vertical SDS gel and embedded in agarose. The second dimension was run, transferred and visualised as described above (Section 2.3.2.1.3). 2.3.3. Northern blotting A plasmid of human paxillin cDNA derived from kidney (GenBank accession no A1373645, IMAGE ID 2030636, clone 4997-g21) having high sequence similarity with paxillin derived from several mammalian species (CLUSTALX) was used as a template. The plasmid was cut using HindIII to linearise the probe and with AccIII to eliminate poly-A non-specific binding. The resulting fragment was used as a template to generate a probe. The radiolabelled probe was synthesised using random hexanucleotide. Northern blotting was carried out using pulmonary artery smooth-muscle mRNA with the expressHyb system (Clontech Laboratories UK Ltd, Basingstoke, UK). Hybridisation was carried out at 54 8C and the rinses at 60 8C. Bands were visualised using a phosphorimager (Typhoon, Amersham Pharmacia Biotech UK Ltd). 3. Results 3.1. Relationship between smooth-muscle-cell actin cytoskeleton and paxillin in the pulmonary artery wall and in cultured smooth muscle cells 3.1.1. Immunostaining of arterial wall Confocal microscopy of pulmonary arterial smooth muscle from arteries incubated at 37 8C in serum-free
Fig. 1. (A) Confocal images of intact pulmonary arterial smooth muscle at 14 days of age stained for paxillin (green) and F-actin (phalloidin; red); nuclei stained with ToPro-3 (blue) in merged image. In medium 199-incubated tissue (a–c) regions of paxillinyactin co-localisation are indicated by arrowheads. Following F-actin depolymerisation by cytochalasin D (d–f) paxillin is reduced at the cell margins and concentrated in the perinuclear cytoplasm (asterisks). (g–i) Unincubated tissue is indistinguishable from that of medium 199-incubated controls. (j) The specificity of paxillin staining is confirmed by the lack of green fluorescence following incubation with an anti-goat secondary antibody, shown by merged green and blue channels. All images taken at the same magnification. Bar, 20 mm. (B) Light-microscope immunofluorescent images of pulmonary arterial smooth muscle cells cultured from a 3-day-old animal: (a,d) paxillin, (b,e) phalloidin-bound F-actin, and (c,f) merged paxillin (green) and phalloidin (red) channels. In untreated controls (a–c) paxillin is present in large focal adhesions at the end of actin stress fibres, frequently co-localising with actin (arrowheads). Following incubation with cytochalasin D (d–f) actin stress fibres decrease and paxillin is concentrated in the perinuclear cytoplasm (asterisks). Bar, 25 mm.
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
Fig. 1.
197
198
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
medium from both 3- and 14-day-old animals and in control untreated vessels, double labelled for paxillin and F-actin, showed that patches of paxillin were present at or close to the cell periphery (Fig. 1A, a–c, g–i). Bundles of F-actin frequently terminated at, and colocalised with, these patches of paxillin. By contrast, in pulmonary arterial smooth muscle treated with cytochalasin D to promote actin depolymerisation, F-actin staining at the cell margin was weaker and paxillin was diffusely distributed throughout the cytoplasm (Fig. 1A, d–f). 3.1.2. Immunostaining of cultured smooth muscle cells The distribution of actin and paxillin in cells isolated from the large intrapulmonary arteries at 3 and 14 days was similar to that observed in intact tissue. In control untreated cells, paxillin co-labelled with vinculin (data not shown). Paxillin was concentrated in focal adhesion plaques at the end of actin stress fibres at the cell margin (Fig. 1B, a–c). Plaques were also present below the cell body. Following treatment with cytochalasin D, cells became more rounded in shape, paxillin-labelled focal adhesions were reduced in size and number, the actin stress fibres were thinner and fewer, and paxillin immunostaining became diffused throughout the cytoplasm, (Fig. 1B, d–f). Thus, in smooth muscle cells within the pulmonary artery wall and in cultured smooth muscle cells, a reduction in actin stress fibres was associated with a reduction in paxillin-labelled focal adhesions and redistribution of paxillin within the cell. 3.2. Characteristics of porcine paxillin 3.2.1. Biochemical studies At all ages, in both normal animals and in those exposed to chronic hypoxia, one-dimensional SDS gel electrophoresis of immunoprecipitated samples of pulmonary artery smooth muscle consistently demonstrated major bands at 60 and 66 kDa that reacted with paxillin antibody (Fig. 2a). An additional doublet was observed at 44–46 kDa. The 60- and 66-kDa porcine bands migrated at a similar level to the two paxillin immunoreactive bands from human endothelial lysate that was loaded for comparison (Fig. 2a). The 60-kDa band reacted with antibodies to both phosphotyrosine and phosphoserine, but the 66-kDa band did not react with the antibody to either (Fig. 2b,c). Two-dimensional electrophoresis of adult tissue showed a thick band of protein having isoelectric point (pI) values of between 5.5 and 4.2 (Fig. 2d). The 60- and 66-kDa isoforms could not be resolved within this band. Phosphotyrosine and phosphoserine co-localised with regions of the paxillin labelling, although more spots labelled for phosphoserine than phosphotyrosine (data not shown).
3.2.2. Paxillin mRNA analysis Northern blots of porcine paxillin revealed the presence of only a broad band of approximately 3.2 kbp of paxillin mRNA in all the pulmonary arterial tissue examined from foetal, 6- and 14-day-old and adult animals (Fig. 3). Northern blots of paxillin from piglets exposed to hypoxia from 3 to 6 days and from 14 to 17 days similarly showed a broad band of mRNA in the same position. 3.3. Developmental studies on porcine pulmonary arteries 3.3.1. Immunocytochemical studies on the intact pulmonary arterial wall 3.3.1.1. Normal. In all specimens examined at all ages, paxillin immunostaining was stronger in elastic arteries than in large and small muscular arteries, and was present in all the smooth muscle cells of the elastic arteries (Fig. 4). In the pre-term foetus and newborn animal, the smooth muscle cells in the sub-endothelial region were less intensely stained than those in the mid and outer media, although in adjacent sections of the same vessels the sub-endothelial region showed strong vinculin expression (Fig. 4). By contrast, at 3 and 6 days of age, both paxillin and vinculin immunostaining was weaker across the full width of the media than at birth and was weakest in the sub-endothelial region (Fig. 4). At 14 days and in the adult, paxillin and vinculin expression was greater than in the foetus–6day-old animals and was similar in all medial cell layers (Fig. 4). Expression of integrin-b1 in elastic pulmonary arteries showed postnatal changes in distribution similar to those of paxillin (data not shown). 3.3.1.2. Chronic hypoxia. The response to chronic hypoxia was dependent upon the age at which the animals were exposed to hypoxia. In the arteries of 3day-old animals exposed to chronic hypoxia from birth, the normal, post-natal transient reduction in paxillin, vinculin and integrin-b1 immunostaining did not occur and the appearance was similar to the normal at birth (Fig. 4). In vessels from 6-day-old animals exposed to chronic hypoxia from 3 days of age, the mid and outer medial cell layers were strongly immunostained for paxillin and integrin-b1, indicating an abnormal increase in expression at this site. Following exposure from 14 to 17 days of age, paxillin, vinculin and integrin-b1 expression remained normal for age. 3.3.2. Changes in paxillin protein expression during development Using Western blots from one-dimensional SDSPAGE gels (ns5 animals at each age) the density of immunostaining at each age was related to the adult
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
199
Fig. 2. (a) SDS-PAGE of immunoprecipitated paxillin from adult porcine intrapulmonary arteries (lanes 1 and 2) and a sample of lysate of human aortic endothelial cells (lane 4), showing the two porcine isoforms at 66 and 60 kDa with similar molecular weights to the two human isoforms (66 and 60 kDa), and the additional 44–46-kDa porcine doublet. Control (lane 3) loaded with immunoprecipitation buffer and mouse anti-paxillin. (b) A representative SDS-PAGE gel of immunoprecipitated paxillin from normal foetal (Ft), newborn (NB), 3-day (3D), 6-day (6D) 14-day (14D) and adult (AD) pigs and piglets exposed to chronic hypoxia from birth to 3 days (3H), from 3 to 6 days (6H) and from 14 to 17 days (17H) stained in the upper frame for paxillin and in the lower for phosphotyrosine. (c) A representative SDS-PAGE gel of immunoprecipitated paxillin from normal and chronically hypoxic animals as described in (b) stained in the upper frame for paxillin and in the lower for phosphoserine. (d) Immunoblots of two-dimensional electrophoresis gels of extracted pulmonary arterial tissue from normal newborn, 6-day-old and adult pigs. The gel margins are marked by arrows, molecular weights along the left-hand side and pI values at the bottom of each photograph. Paxillin is present over a range of pI values in a broad band at 60–66 and 45 kDa. There is a shift toward lower pI values between birth and adulthood.
value, which was taken as 100%. Each gel contained one lane loaded with paxillin immunoprecipitated from elastic pulmonary arteries of one animal at each age, normal and hypoxic, and the pattern of change with age was consistent in all experiments. 3.3.2.1. Normal. The total amount of paxillin, that is the sum of the 66- and 60-kDa bands, changed with age (P-0.01, ANOVA). The paxillin content of foetal and
newborn tissue was similar to that in adult tissue, but there was a transient postnatal reduction in paxillin content that was maximal at 6 days (Fig. 5a, P-0.01). The content was again higher at 14 days (P-0.01), being similar to that in the adult. This pattern of change with age was observed in the expression of both the 66and 60-kDa paxillin isoforms (Fig. 5b, P-0.01) although there were some differences between the two isoforms. The amount of the 66-kDa isoform was greater
200
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
Fig. 3. Northern blot of paxillin mRNA from adult pulmonary artery, showing a broad mRNA band at ;3.2 kbp (arrow). Northern blotting was performed using the mRNA-enriched fraction from total RNA and digoxigenin-labelled riboprobes.
in the newborn than in the foetal tissue (P-0.02), unlike the 60-kDa isoform. The amount of the 60-kDa isoform increased more slowly from its nadir at 6 days, and there was significantly less protein at 14 days than in the adult (P-0.02). Two-dimensional gel electrophoresis on tissue samples from newborn, 3-, 6- and 14day-old animals showed a shift toward more acidic pI values with increasing age, the greatest shift occurring between birth and 6 days of age (Fig. 2d). At birth and 3 days, the pI values of the paxillin immunolabelled spots ranged from approximately 6.7 to 4.2 and labelling was evenly distributed over this range. At 6 and 14 days of age, as in the adult, most of the paxillin lay between pI 5.5 and 4.2 (Fig. 2d). The shift in pI may indicate an increase in the level of paxillin phosphorylation (Pavalko et al. 1995). However, since at all ages the thickness of the band staining for paxillin remained maximal at pI 5.0–5.5, only a portion of total paxillin may show this change in phosphorylation. Phosphotyrosine and phosphoserine co-localised with some paxillin spots, but in many neither phosphorylation was detectable. At all ages more spots labelled for phosphoserine than for phosphotyrosine.
Fig. 4. Expression of paxillin protein and vinculin in adjacent sections of the same intrapulmonary elastic arteries from foetal, 3- and 14-day normal piglets and a piglet exposed to hypoxia from birth to 3 days of age. Paxillin and vinculin immunostaining show a similar pattern of expression. In foetal and 14-day-old preparations both proteins are present in all medial cell layers. However, at 3 days all cell layers stain less strongly, with the least stain being observed in adlumenal layers. By contrast, in the 3-day-old hypoxic animal, paxillin and vinculin are uniformly strongly expressed across the media, as in the foetus. All micrographs taken at the same magnification; bar, 50 mm.
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
201
Fig. 5. Histograms showing developmental changes in paxillin expression in elastic pulmonary arteries. Paxillin content is expressed as a percentage of the adult value. (a) Change in total paxillin content in normal animals. (b) Change in expression of 66- and 60-kDa isoforms in normal animals. (c) Influence of hypoxia on expression of 66- and 60-kDa paxillin isoforms, comparing paxillin expression in hypoxic animals to that of normal animals at ages corresponding to the start and end of exposure. Ages: foetal (FT), newborn (NB), 3-day (3D), 6-day (6D) 14-day (14D) and adult (AD) pigs. Piglets exposed to chronic hypoxia from birth to 3 days (3H), form 3 to 6 days (6H) and from 14 to 17 day (17H). Mean value "standard error of mean. #P-0.05.
3.3.2.2. Chronic hypoxia. Tissue from both 3- and 6day-old hypoxic animals contained more paxillin, and more 60- and 66-kDa isoforms than age-matched controls (P-0.01 for all comparisons) (Fig. 5c). This increase was greatest in the 66-kDa isoform of 6-dayold hypoxic animals (Fig. 5c). Thus, exposure to chronic hypoxia from birth to 3 days resulted in abolition of the normal postnatal reduction in the amount of paxillin, and exposure from 3 to 6 days of age enhanced paxillin expression. By contrast, there was no change in the expression of paxillin or in either of its isoforms following exposure to hypoxia from 14 to 17 days of age (Fig. 5c). On two-dimensional gel electrophoresis, the range of pI values of the paxillin immunolabelled spots in tissue from 3-day-old animals exposed to hypoxia from birth was similar to the normal at birth and 3 days of age.
3.3.3.1. Normal. When the amount of phosphotyrosine and phosphoserine in each sample was related to the amount of 60-kDa paxillin (phosphorylated and unphosphorylated) in that sample, then the proportion of paxillin phosphorylated on both tyrosine and on serine increased, by 18.5 and 25.5%, respectively (P-0.05 for both comparisons), during the first 6 days of life (Fig. 6a). At 14 days, there was less phosphotyrosine and phosphoserine than in either 3-day-old (P-0.01 for both phosphorylations) or adult specimens (P-0.01 for phosphoserine) (Fig. 6a).
3.3.3. Changes in phosphorylated paxillin during development In all blots examined (ns7 animals at each age) both anti-phosphotyrosine and anti-phosphoserine reacted with the 60-kDa paxillin band, but not with the 66-kDa band (Fig. 2b,c)
4. Discussion
3.3.3.2. Chronic hypoxia. The amount of phosphotyrosine- and phosphoserine-labelled paxillin only became abnormal after exposure to chronic hypoxia from 14 to 17 days, when both increased (P-0.01 for both comparisons) (Fig. 6b,c).
We have investigated the in vivo distribution and phosphorylation of paxillin as evidence of activity and remodelling of the focal adhesions linking the pulmonary arterial smooth muscle cell cytoskeleton to the
202
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
Fig. 6. Histograms showing developmental changes in paxillin phosphorylation in elastic pulmonary arteries. Phosphorylation is expressed as the ratio of phosphotyrosine or phosphoserine staining to that of total paxillin at each age. (a) Phosphorylation of paxillin in normal animals. (b,c) Influence of hypoxia on phosphorylation of paxillin, comparing phosphorylated paxillin expression in hypoxic animals with that of normal animals at ages corresponding to the start and end of the period of chronic hypoxia for (b) phosphotyrosine and (c) phosphoserine. #P-0.05.
extracellular matrix during adaptation to extrauterine life. We describe a transient reduction in paxillin in the pulmonary arterial media immediately after birth biochemically and by immunohistochemistry, a change temporally and spatially associated with the transient reduction in actin previously described (Hall et al., 2000). We also found a simultaneous increase in paxillin phosphorylation compatible with enhanced focal adhesion activity. The amount of paxillin and the level of tyrosine and serine phosphorylation was similar in the foetus and adult, suggesting greater cell-matrix stability at these two extremes of life, despite the pulmonary arteries being exposed to different pressures. 4.1. Characteristics of porcine paxillin The biochemical and molecular characteristics of the paxillin found in porcine pulmonary arteries were generally similar to those previously reported in other tissues in other species (De Nichilo and Yamada, 1996; Bellis et al., 1997; Mazaki et al., 1997; Slack, 1998; Thomas et al., 1999). The molecular weights were similar to those of the two paxillin isoforms present in human aortic endothelial cells. We also found a paxillin doublet of 44–46 kDa similar to that previously described in the human cell line 32Dc13 (Salgia et al., 1995). Others have characterised a 50–51-kDa doublet
as HIC-5, a protein having extensive homology with paxillin (Thomas et al., 1999). Our two-dimensional electrophoresis revealed several isoforms of paxillin, with pI values ranging from 6.7 to 4.2. Chicken gizzard contains at least four paxillin isoforms, but with similar pI values (Turner et al., 1990), and cultured U937 human monocytes have yielded as many as eight isoforms (Mazaki et al., 1997). In addition, tyrosine and serine phosphorylation led to the generation of multiple isoforms in cultured U937 cells (Mazaki et al., 1997). The process of isoform formation is regulated by the number of genes encoded for the same protein andyor alternative splicing of the mRNA, and by post-translational modification. In the present study, Northern blot analysis did not unequivocally demonstrate more than one porcine paxillin mRNA in pulmonary arterial tissue at any age in normal or hypoxic animals. A single mRNA has been found in mouse tissue (Mazaki et al., 1998), and in human monocytes and cancer cells (Salgia et al., 1995; Mazaki et al., 1997). Our findings suggest that, as in the chicken (Turner et al., 1990), the multiple paxillin isoforms present in the porcine tissue are likely to be the result of post-translational processing only. Alternatively, different isoforms varying only by different pI values may reflect differing levels of phosphorylation of a single (60 kDa) isoform (Tang et al., 1999).
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
4.2. Paxillin expression and phosphorylation during normal development The findings in the present study indicate for the first time that birth is followed by remodelling of paxillinassociated focal adhesions. The transient reduction in paxillin accompanied a reduction in the amount of actin and in the proportion of filamentous to globular cytosolic actin previously described (Hall et al., 2000). The observations made on the intact tissue in vivo correspond to the reduction in paxillin-labelled focal adhesions resulting from cytochalasin D-induced disruption of the actin cytoskeleton in intact pulmonary artery and in the cultured pulmonary artery smooth muscle cells. The postnatal reduction in paxillin was associated with an increase in paxillin phosphorylation, possibly caused by the increased strain imposed on the vessel wall when the pulmonary arteries were abruptly distended with blood after birth. Mechanosensitive tyrosine phosphorylation of paxillin and focal adhesion kinase (FAK) has been described in tracheal smooth muscle strips (Tang et al., 1999), and cultured airway smooth muscle cells exposed to an increase in strain show a rapid increase in tyrosine phosphorylation of both these proteins (Smith et al., 1998). Thus, a strong mechanical stimulus acting on the smooth muscle cell–focal adhesion complexes at birth could have led to their activation and remodelling, triggering actin cytoskeletal remodelling. Focal adhesions are linked to the extracellular matrix through integrin receptors. The a1–a7b1 integrins are the key vascular integrins in cell matrix interactions (Hedin, 1994) and the cytoplasmic tail of b1 is known to associate with paxillin (Schaller and Parsons, 1994). Among the major ligands of the b1 integrin, fibrillar collagen deposition increases immediately after birth (Allen and Haworth, 1988; Kitley et al., 2000). Paxillin phosphorylation accompanies cell adhesion to the extracellular matrix following the binding of integrins to their cognate ligand (Cattelino et al., 1997; Brown et al., 1998; Lewis and Schwartz, 1998). This may help to explain the postnatal increase in tyrosine and serine phosphorylation of paxillin described here. Although tyrosine phosphorylation has been extensively implicated in cytoskeletal re-organisation and cell adhesion (Schoenwaelder and Burridge, 1999), in the present study the postnatal increase in serine phosphorylation was particularly marked. During macrophage adhesion to vitronectin and fibroblast adhesion to fibronectin, 96% of paxillin phosphorylation occurred on serine residues (De Nichilo and Yamada, 1996; Bellis et al., 1997). Binding serine to LIM domains increases cell adhesion to fibronectin, with accelerated fibronectininduced localisation of paxillin to focal contacts (Brown et al., 1998). Thus, the high level of paxillin phosphorylation, on both tyrosine and serine residues, at 3 and
203
6 days indicates remodelling of focal adhesions as they detach and re-attach to the extracellular matrix when the pulmonary arteries rapidly increase in size. By 14 days of age, paxillin and vinculin expression in the media resembled that in the adult, but the amount of the 60-kDa paxillin isoform, the phosphorylated isoform, remained lower than in the adult, and tyrosine and serine phosphorylation was less than at 3 or 6 days. This transient reduction in focal adhesion activity takes place during a period of relative cytoskeletal stability (Hall et al., 2000). In vitro studies have shown that a reduction in paxillin and FAK tyrosine phosphorylation does not reduce actin stress fibre and focal adhesion formation, as demonstrated following eNOS gene transfer into systemic arterial smooth-muscle cells (Fang et al., 1997). eNOS might also play a similar role in vivo. Porcine eNOS activity is extremely low at birth and then increases rapidly during the first 14 days of life (Tulloh et al., 1997; Arrigoni et al., 2002). 4.3. Paxillin expression and phosphorylation in pulmonary hypertension The response to chronic hypoxia depended on the age of the animals at the onset of exposure. Exposure from birth to 3 days of age appeared to prevent the normal postnatal down-regulation in total paxillin expression and increase in proportion of phosphorylated paxillin, indicating failure of the focal adhesion complexes to remodel after birth. By contrast, when hypoxic exposure commenced at 3 days, after the normal process of cytoskeletal remodelling was established, our findings indicated abnormal actin remodelling. The amount of both isoforms increased, but the proportion of phosphorylated paxillin was normal. These findings are consistent with previous studies showing that in this experimental porcine model of neonatal pulmonary hypertension there is no transient postnatal reduction in actin myofilaments, rather an increase in the volume density of both cytoskeletal filaments and surface dense bodies (focal adhesions) (Allen and Haworth, 1986). There may also be a functional correlate. Tang et al. (1999) have shown that paxillin phosphorylation in intact tracheal smooth muscle is sensitive to length rather than tension, and chronic hypoxic pulmonary arteries are vasoconstricted. In animals first exposed to hypoxia at 14 days of age, the expression of paxillin was unaffected, but the proportion of phosphoserineand phosphotyrosine-labelled paxillin increased. This may reflect hypertensive remodelling of a fully dilated vessel that had a relatively stable cytoskeleton before exposure to the insult (Hall et al., 2000). Our data indicate that with the abrupt distension of the vessel wall at birth, the increase in strain is associated with, or leads to, degradation of focal adhesionassociated paxillin and a concomitant increase in paxillin
204
I. Diagne et al. / Matrix Biology 22 (2003) 193–205
phosphorylation. Having accommodated to the extrauterine environment, paxillin phosphorylation diminishes. This study is one of few that relate in vivo and in vitro paxillin and cytoskeletal re-arrangements. Our in vivo observations support the contention that paxillin may play a novel role in regulating phosphorylation of cellular proteins in both a spatial and temporal manner (Shen et al., 1998), possibly influencing outside-in signalling at the interface between cell and extracellular matrix and inside-out cytoskeleton–focal adhesion integrin-mediated signal transduction. The necessity of carrying out detailed studies on these signal transduction pathways in early life is now self-evident. Acknowledgments We would like to thank Dr Michael J Dunn, National Heart and Lung Institute, Harefield Hospital for his advice in the preparation of two-dimensional gels. This work was funded by a British Heart Foundation Programme Grant. References Allen, K., Haworth, S.G., 1988. Human postnatal pulmonary arterial remodeling. Ultrastructural studies of smooth muscle cell and connective tissue maturation. Lab. Invest. 59, 702–709. Allen, K.M., Haworth, S.G., 1986. Impaired adaptation of pulmonary circulation to extrauterine life in newborn pigs exposed to hypoxia: an ultrastructural study. J. Pathol. 150, 205–212. Arrigoni, F.I., Hislop, A.A., Pollock, J.S., Haworth, S.G., Mitchell, J.A., 2002. Birth upregulates nitric oxide synthase activity in the porcine lung. Life Sci. 70, 1609–1620. Bellis, S.L., Perrotta, J.A., Curtis, M.S., Turner, C.E., 1997. Adhesion of fibroblasts to fibronectin stimulates both serine and tyrosine phosphorylation of paxillin. Biochem. J. 325, 375–381. Brown, M.C., Perrotta, J.A., Turner, C.E., 1998. Serine and threonine phosphorylation of the paxillin LIM domains regulates paxillin focal adhesion localization and cell adhesion to fibronectin. Mol. Biol. Cell 9, 1803–1816. Burridge, K., Chrzanowska-Wodnicka, M., 1996. Focal adhesions, contractility, and signaling. Annu. Rev. Cell Dev. Biol. 12, 463–519. Cattelino, A., Cairo, S., Malanchini, B., de Curtis, I., 1997. Preferential localization of tyrosine-phosphorylated paxillin in focal adhesions. Cell Adhes. Commun. 4, 457–467. De Nichilo, M.O., Yamada, K.M., 1996. Integrin alpha v beta 5dependent serine phosphorylation of paxillin in cultured human macrophages adherent to vitronectin. J. Biol. Chem. 271, 11016–11022. Fang, S., Sharma, R.V., Bhalla, R.C., 1997. Endothelial nitric oxide synthase gene transfer inhibits platelet-derived growth factor-BB stimulated focal adhesion kinase and paxillin phosphorylation in vascular smooth muscle cells. Biochem. Biophys. Res. Commun. 236, 706–711. Hall, S.M., Gorenflo, M., Reader, J., Lawson, D., Haworth, S.G., 2000. Neonatal pulmonary hypertension prevents reorganisation of the pulmonary arterial smooth muscle cytoskeleton after birth. J. Anat. 196, 391–403. Hall, S.M., Haworth, S.G., 1987. Conducting pulmonary arteries: structural adaptation to extrauterine life. Cardiovasc. Res. 21, 208–216.
Hall, S.M., Haworth, S.G., 1996. Effect of cold preservation on pulmonary arterial smooth muscle cells. Am. J. Physiol. 270, L435–L445. Haworth, S.G., Hislop, A.A., 1981. Adaptation of the pulmonary circulation to extra-uterine life in the pig and its relevance to the human infant. Cardiovasc. Res. 15, 108–119. Haworth, S.G., Hislop, A.A., 1982. Effect of hypoxia on adaptation of the pulmonary circulation to extra-uterine life in the pig. Cardiovasc. Res. 16, 293–303. Haworth, S.G., Hall, S.M., Chew, M., Allen, K., 1987. Thinning of foetal pulmonary arterial wall and postnatal remodelling: ultrastructural studies on the respiratory unit arteries of the pig. Virchows Arch. A 411, 161–171. Hedin, U., 1994. Extracellular matrix components and integrins in the control of arterial smooth muscle cell structure and function. J. Atheroscler. Thromb. 1, S39–S46. Kitley, R.M., Hislop, A.A., Hall, S.M., Freemont, A.J., Haworth, S.G., 2000. Birth-associated changes in pulmonary arterial connective tissue gene expression in the normal and hypertensive lung. Cardiovasc. Res. 46, 332–345. Lewis, J.M., Schwartz, M.A., 1998. Integrins regulate the association and phosphorylation of paxillin by c-Abl. J. Biol. Chem. 273, 14225–14230. Mazaki, Y., Hashimoto, S., Sabe, H., 1997. Monocyte cells and cancer cells express novel paxillin isoforms with different binding properties to focal adhesion proteins. J. Biol. Chem. 272, 7437–7444. Mazaki, Y., Uchida, H., Hino, O., Hashimoto, S., Sabe, H., 1998. Paxillin isoforms in mouse. Lack of the gamma isoform and developmentally specific beta isoform expression. J. Biol. Chem. 273, 22435–22441. Mehta, D., Wang, Z., Wu, M.F., Gunst, S.J., 1998. Relationship between paxillin and myosin phosphorylation during muscarinic stimulation of smooth muscle. Am. J. Physiol. 274, C741–C747. Pavalko, F.M., Adam, L.P., Wu, M.F., Walker, T.L., Gunst, S.J., 1995. Phosphorylation of dense-plaque proteins talin and paxillin during tracheal smooth muscle contraction. Am. J. Physiol. 268, C563–C571. Salgia, R., Li, J.L., Lo, S.H., et al., 1995. Molecular cloning of human paxillin, a focal adhesion protein phosphorylated by P210BCRyABL. J. Biol. Chem. 270, 5039–5047. Sanchez, J.C., Rouge, V., Pisteur, M., et al., 1997. Improved and simplified in-gel sample application using reswelling of dry immobilized pH gradients. Electrophoresis 18, 324–327. Schaller, M.D., Parsons, J.T., 1994. Focal adhesion kinase and associated proteins. Curr. Opin. Cell Biol. 6, 705–710. Schaller, M.D., Otey, C.A., Hildebrand, J.D., Parsons, J.T., 1995. Focal adhesion kinase and paxillin bind to peptides mimicking beta integrin cytoplasmic domains. J. Cell Biol. 130, 1181–1187. Schoenwaelder, S.M., Burridge, K., 1999. Bidirectional signaling between the cytoskeleton and integrins. Curr. Opin. Cell Biol. 11, 274–286. Shen, Y., Schneider, G., Cloutier, J.F., Veillette, A., Schaller, M.D., 1998. Direct association of protein-tyrosine phosphatase PTP-PEST with paxillin. J. Biol. Chem. 273, 6474–6481. Slack, B.E., 1998. Tyrosine phosphorylation of paxillin and focal adhesion kinase by activation of muscarinic m3 receptors is dependent on integrin engagement by the extracellular matrix. Proc. Natl. Acad. Sci. USA 95, 7281–7286. Smith, P.G., Garcia, R., Kogerman, L., 1998. Mechanical strain increases protein tyrosine phosphorylation in airway smooth muscle cells. Exp. Cell Res. 239, 353–360. Tachibana, K., Sato, T., D’Avirro, N., Morimoto, C., 1995. Direct association of pp125FAK with paxillin, the focal adhesion- targeting mechanism of pp125FAK. J. Exp. Med. 182, 1089–1099.
I. Diagne et al. / Matrix Biology 22 (2003) 193–205 Tang, D., Mehta, D., Gunst, S.J., 1999. Mechanosensitive tyrosine phosphorylation of paxillin and focal adhesion kinase in tracheal smooth muscle. Am. J. Physiol. 276, C250–C258. Thomas, S.M., Hagel, M., Turner, C.E., 1999. Characterization of a focal adhesion protein, Hic-5, that shares extensive homology with paxillin. J. Cell Sci. 112, 181–190. Tulloh, R.M.R., Hislop, A.A., Boels, P.J., Deutsch, J., Haworth, S.G., 1997. Chronic hypoxia inhibits postnatal maturation of porcine intrapulmonary artery relaxation. Am. J. Physiol. 272, H2436–H2445. Turner, C.E., Miller, J.T., 1994. Primary sequence of paxillin contains putative SH2 and SH3 domain binding motifs and multiple LIM
205
domains: identification of a vinculin and pp125Fak-binding region. J. Cell Sci. 107, 1583–1591. Turner, C.E., Glenney Jr, J.R., Burridge, K., 1990. Paxillin: a new vinculin-binding protein present in focal adhesions. J. Cell. Biol. 111, 1059–1068. Wang, Z., Pavalko, F.M., Gunst, S.J., 1996. Tyrosine phosphorylation of the dense plaque protein paxillin is regulated during smooth muscle contraction. Am. J. Physiol. 271, C1594–C1602. Wood, C.K., Turner, C.E., Jackson, P., Critchley, D.R., 1994. Characterisation of the paxillin-binding site and the C-terminal focal adhesion targeting sequence in vinculin. J. Cell Sci. 107, 709–717.