Dissociation of Crk-associated substrate from the vimentin network is regulated by p21-activated kinase on ACh activation of airway smooth muscle

Ruping Wang, Qing-Fen Li, Yana Anfinogenova, Dale D. Tang

Abstract

The intermediate filament protein vimentin has been shown to be required for smooth muscle contraction. The adapter protein p130 Crk-associated substrate (CAS) participates in the signaling processes that regulate force development in smooth muscle. However, the interaction of vimentin filaments with CAS has not been well elucidated. In the present study, ACh stimulation of tracheal smooth muscle strips increased the ratio of soluble to insoluble vimentin (an index of vimentin disassembly) in association with force development. ACh activation also induced vimentin phosphorylation at Ser56 as assessed by immunoblot analysis. More importantly, CAS was found in the cytoskeletal vimentin fraction, and the amount of CAS in cytoskeletal vimentin was reduced in smooth muscle strips on contractile stimulation. CAS redistributed from the myoplasm to the periphery during ACh activation of smooth muscle cells. The ACh-elicited decrease in CAS distribution in cytoskeletal vimentin was attenuated by the downregulation of p21-activated kinase (PAK) 1 with antisense oligodeoxynucleotides. Vimentin phosphorylation at this residue, the ratio of soluble to insoluble vimentin, and active force in smooth muscle strips induced by ACh were also reduced in PAK-depleted tissues. These results suggest that PAK may regulate CAS release from the vimentin intermediate filaments by mediating vimentin phosphorylation at Ser56 and the transition of cytoskeletal vimentin to soluble vimentin. The PAK-mediated dissociation of CAS from the vimentin network may participate in the cellular processes that affect active force development during ACh activation of tracheal smooth muscle tissues.

  • intermediate filaments
  • cytoskeleton
  • contraction

the cytoskeletal systems of smooth muscle cells comprise myofilaments, microtubules, and the intermediate filament network. It is well established that myofilaments containing thick (myosin) and thin (actin) filaments are major participants in force development (9). Disruption of microtubules by nocodazole potentiates high-potassium-mediated vascular smooth muscle contraction by increasing intracellular calcium concentration (18). We recently showed that the downregulation of vimentin (a major intermediate filament protein in airway and vascular smooth muscle) in tracheal smooth muscle tissues diminishes contractile responses to agonist stimulation (35). Flow-induced mechanical responses of mesenteric resistance arteries are also reduced in vimentin-knockout mice (10).

The adapter protein p130 Crk-associated substrate (CAS), a major member of the CAS family of proteins, has been proposed to serve as docking sites for other proteins in integrin-mediated signaling transduction (4). CAS has been shown to participate in the signaling cascades that regulate force development in smooth muscle (17, 29). The downregulation of CAS in smooth muscle depresses active force development by inhibiting actin polymerization, a cellular process that is important for contractile force development in smooth muscle (2, 11, 26, 28, 32, 37). However, the interaction of CAS with the vimentin framework in mammalian cells, in general, and in smooth muscle, in particular, is not well understood.

Vimentin intermediate filaments are capable of binding to certain signaling molecules in mammalian cells. In differentiated smooth muscle cells, calcium/calmodulin-dependent kinase IIγ (CamKII) associates with the vimentin network. Rho kinase colocalizes predominantly with the filamentous vimentin network in fibroblasts. Disassembly and/or reorganization of the vimentin cytoskeleton induces the redistribution of these molecules, which has been implicated in regulating cell functions (15, 20).

External stimulation induces vimentin filament disassembly/assembly in cultured smooth muscle cells, chromaffin cells, human fibroblasts, rat RVF-SM cells, and BHK-21 fibroblasts (5, 19, 22, 25). The physiological significance of vimentin disassembly in response to growth factor stimulation has been proposed to facilitate cell division, possibly via spatial reorganization of the vimentin network (16, 34). Remodeling of vimentin filaments has also been shown to be associated with the enhanced migration of promyelocytic leukemia cells during differentiation (1).

The vimentin network may be regulated by vimentin phosphorylation and dephosphorylation on serine/threonine residues (5). p21-Activated kinase (PAK, a serine/threonine kinase) has been shown to phosphorylate vimentin in in vitro studies (8, 25). In cultured smooth muscle cells, agonist stimulation induces vimentin disassembly and vimentin phosphorylation at Ser56, which are inhibited by silencing of PAK1, a dominant isoform in smooth muscle (3, 25).

The aim of this study was to evaluate the association of CAS with the vimentin network in smooth muscle in response to stimulation with ACh, a well-known airway smooth muscle contractile agonist. Our results suggest that ACh activation triggers the dissociation of CAS from the vimentin cytoskeleton, which may be mediated through phosphorylation and disassembly of vimentin by PAK.

MATERIALS AND METHODS

Preparation of smooth muscle tissues.

Mongrel dogs (20–25 kg body wt) were anesthetized with pentobarbital sodium (30 mg/kg iv) and quickly exsanguinated. All experimental protocols were approved by the Institutional Animal Care and Use Committee. A 15-cm segment of extrathoracic trachea was immediately removed and placed at room temperature in physiological saline solution (PSS) containing (in mM) 110 NaCl, 3.4 KCl, 2.4 CaCl2, 0.8 MgSO4, 25.8 NaHCO3, 1.2 KH2PO4, and 5.6 glucose. The solution was aerated with 95% O2-5% CO2 to maintain pH 7.4. Rectangular strips (0.6–0.8 mm wide, 0.2–0.3 mm thick, 9–10 mm long) were dissected from the trachea after removal of the epithelium and connective tissue layer.

Measurement of force development in smooth muscle.

Each muscle strip was placed in PSS at 37°C in a 25-ml organ bath and attached to a Grass force transducer that had been connected to a Gould recorder or a computer with analog-to-digital converter (Grass). At the beginning of each experiment, muscle strips were stretched to the reference muscle length (9–10 mm). After 10–20 min of equilibrium, they were stimulated with 10−5 M ACh repeatedly until contractile responses and passive tension stabilized. In untreated muscle tissues, an average passive tension was 0.2 g, while an average active force was 4 g.

For experiments associated with antisense, oligodeoxynucleotides (ODNs) dissolved in Tris-EDTA buffer were introduced into muscle strips according to experimental procedures described previously (27, 28, 31). Muscle strips were then incubated for 2 days with ODNs in DMEM. The strips were returned to PSS at 37°C in 25-ml organ baths and stretched to the corresponding reference muscle length. After repeated stimulation with ACh, each contractile response and passive tension were compared with the corresponding preincubation value. For analysis of protein expression and phosphorylation, muscle strips were frozen using liquid N2-cooled tongs and then pulverized under liquid N2 using a mortar and pestle.

Analysis of soluble-to-insoluble vimentin ratio.

The amount of soluble and insoluble vimentin was evaluated by modification of the method previously described (16, 25, 34). Briefly, smooth muscle strips were homogenized in a buffer containing 1% Nonidet P-40, 10% glycerol, 20 mM HEPES (pH 7.6), 150 mM NaCl, 2 mM sodium orthovanadate, 2 mM molybdate, 2 mM sodium pyrophosphate, and protease inhibitors (2 mM benzamidine, 0.5 mM aprotinin, and 1 mM phenylmethylsulfonyl fluoride). The homogenates were immediately incubated in the same buffer at 37°C for 30 min. The soluble (disassembled) and insoluble (assembled) fractions were collected after centrifugation at 5,200 rpm at 4°C for 30 min, separated by SDS-PAGE, and transferred to nitrocellulose membranes. The membranes were probed with vimentin antibody (1:10,000 dilution; BD Biosciences, clone RV202) and then with horseradish peroxidase (HRP)-conjugated anti-mouse Ig (Amersham Life Sciences) (25). Immunocomplex on the membranes was reacted with enhanced chemiluminescence (ECL) substrate (Supersignal West Dura Extended Duration Substrate, Pierce). The ECL signals on the immunoblots were detected and analyzed using a luminescent image system (model LAS3000, Fuji). The ratio of soluble to insoluble vimentin was determined after scanning densitometry of immunoblots.

Assessment of vimentin phosphorylation at Ser56.

Pulverized muscle strips were mixed with 100 μl of extraction buffer containing 20 mM Tris·HCl (pH 7.4), 2% Triton X-100, 0.2% SDS, 2 mM EDTA, phosphatase inhibitors (2 mM sodium orthovanadate, 2 mM molybdate, and 2 mM sodium pyrophosphate), and protease inhibitors (2 mM benzamidine, 0.5 mM aprotinin, and 1 mM phenylmethylsulfonyl fluoride). Each sample was kept on ice for 1 h and then centrifuged for the collection of supernatant. Muscle extracts were boiled in sample buffer [1.5% dithiothreitol, 2% SDS, 80 mM Tris·HCl (pH 6.8), 10% glycerol, and 0.01% bromphenol blue] for 4 min and separated by SDS-PAGE. Proteins were transferred to a nitrocellulose membrane, which was blocked with 2% gelatin for 1 h and probed with site-specific, state-dependent antibody for vimentin Ser56 (custom made by SynPep, Dublin, CA; synthetic phosphopeptide sequence: Ser-Leu-Tyr-Ala-Ser-phospho-Ser56-Pro-Gly-Gly-Ala-Tyr-Cys; antibody dilution 1:1,000) and then with HRP-conjugated anti-rabbit Ig (ICN Biomedicals, Irvine, CA) (25). Proteins were visualized by ECL. The membrane was stripped and reprobed with monoclonal vimentin antibody (BD Biosciences; 1:10,000 dilution) and then with HRP-conjugated anti-mouse Ig (Amersham Life Sciences) to normalize for minor differences in protein loading. Changes in protein phosphorylation were expressed as a magnitude increase over levels of phosphorylation in unstimulated muscle strips.

Analysis of CAS association with cytoskeletal vimentin.

Insoluble vimentin was collected from smooth muscle tissues as described above. An equal amount of cytoskeletal vimentin was separated by 10% SDS-PAGE and transferred to nitrocellulose membranes, which were cut into two parts: the upper part was probed with monoclonal p130CAS antibody (1:2,000 dilution; BD Biosciences, clone 24), and the lower part was blotted with vimentin antibody (1:10,000 dilution). The ratio of CAS to vimentin was calculated after densitometric analysis of immunoblots.

In vitro kinase assay.

Wild-type vimentin and vimentin S56A (alanine substitution at Ser56) were produced and purified as previously described (14, 25). Purified wild-type vimentin or the vimentin mutant S56A (0.1 mg/ml) was incubated with 2 μg/ml activated PAK (Upstate Biotechnology) for 30 min in a buffer containing 20 mM HEPES (pH 7.5), 60 mM NaCl, 2 mM MgCl2, 5 mM EGTA, and 100 μM ATP. High (150 mM) NaCl solution was then added to the reaction mixture, which was incubated at 37°C for 1 h to initiate filament formation. The mixture was treated with 0.1% glutaraldehyde at room temperature for 30 min to stabilize vimentin filaments (14, 25).

Immunoprecipitation.

CAS or vimentin was immunoprecipitated from muscle extracts as previously described with minor modification (23, 25). Briefly, muscle extracts containing equal amounts of protein were precleared for 30 min with 50 μl of 10% protein A-Sepharose. The precleared extracts were collected after centrifugation at 13,200 rpm for 2 min and incubated with monoclonal antibodies against CAS or vimentin overnight and then with 125 μl of a 10% suspension of protein A-Sepharose beads conjugated to rabbit anti-mouse IgG for 2.5 h. Immunocomplexes were washed four times in Tris-buffered saline [150 mM NaCl and 50 mM Tris (pH 7.6)]. All immunoprecipitation procedures were performed at 4°C.

Far-Western analysis.

CAS immunoprecipitates were resolved by SDS-PAGE and then transferred to membranes. The membranes were incubated with vimentin and its mutant, which had been treated with active PAK (see above), for 2 h at room temperature and then probed with vimentin antibody.

Cell dissociation and immunofluorescence analysis.

Smooth muscle cells were freshly dissociated from tracheal smooth muscle tissues according to the experimental procedures previously described (35). The cells were fixed for 10 min in 4% paraformaldehyde, washed three times in Tris-buffered saline (50 mM Tris, 150 mM NaCl, and 0.1% NaN3), and permeabilized with 0.2% Triton X-100 for 5 min. Cells were then incubated with CAS monoclonal antibody for 45–60 min at 37°C. Cells were then washed and incubated with a secondary antibody conjugated to Alexa 546 fluoroprobe (Molecular Probes, Eugene, OR) for 30 min at 37°C. The cellular localization of fluorescently labeled proteins was viewed under laser scanning confocal microscopy (510 Meta, Zeiss) using a ×63 oil immersion objective. The fluorescence of Alexa 546-labeled protein (red) was excited with a helium-neon laser at 543 nm, and emissions were collected at 565–615 nm.

Image analysis for protein localization was carried out using the previously described method with minor modification (35). With use of LSM5 analysis software (Zeiss), the pixel intensity was quantified for three to four line scans across the periphery of cells. Ratios of pixel intensity at the cell edge to pixel intensity at the cell interior were determined for each line scan as follows: average maximal pixel intensity at the cell periphery ÷ minimal pixel intensity in the cell interior. The ratios of pixel intensity at the cell border to that in the cell interior for all the line scans performed on a given cell were averaged to obtain a single value for the ratio of each cell.

Loading of ODNs and organ culture.

Antisense ODNs were used to selectively suppress PAK1 expression in canine tracheal smooth muscle on the basis of the cDNA sequence of human PAK1 [National Center for Biotechnology Information (NCBI) accession no. NM/20%002576]. The sequence of PAK1 is as follows: 5′-GGAGGGGCTGGGGGTTTGTC-3′ (antisense) and 5′-GACAAACCCCCAGCCCCTCC-3′ (sense). According to sequence-matching results obtained from NCBI, these sequences are not homologous to sequences of any other contractile proteins or cytoskeletal proteins. The antisense molecule targets to a region of mRNA that is unique to PAK1. The phosphorothioate ODNs were synthesized and purified by Invitrogen (Carlsbad, CA). The ODNs were introduced into the smooth muscle strips by chemical loading (also referred to as reversible permeabilization) using methods we have previously described (27, 30, 31).

Determination of protein expression.

Protein expression of ODN-treated muscle strips was assessed by immunoblot analysis. Briefly, muscle extracts were separated by SDS-PAGE and then transferred to membranes. The membrane was cut into two parts for immunoblotting of different proteins. The upper part of the membrane was blocked with 5% milk for 1 h and probed with monoclonal antibody against myosin light chain kinase (MLCK) (30) and then with HRP-conjugated anti-mouse IgG (Amersham). The lower part of the membrane was reacted with a polyclonal antibody against PAK1 (Cell Signaling, Beverly, MA) and then with HRP-conjugated anti-rabbit IgG (ICN). Proteins were visualized by chemiluminescence and quantified by scanning densitometry. Densitometric values of PAK and MLCK were determined for sense- and antisense-treated strips and normalized to those of non-ODN-treated strips. The ratios of these proteins were calculated to verify that changes in protein expression were selective for PAK.

Statistical analysis.

All statistical analysis was performed using Prism 4 software (GraphPad Software, San Diego, CA). Comparison among multiple groups was performed by one-way ANOVA followed by Tukey's multiple-comparison post test. Differences between pairs of groups were analyzed by Student-Newman-Keuls test or Dunn's method. Values of n refer to the number of experiments used to obtain each value. P < 0.05 was considered to be significant.

RESULTS

Ratio of soluble to insoluble vimentin is increased in smooth muscle strips in response to ACh stimulation.

To determine whether vimentin disassembly occurs during contractile activation, tracheal smooth muscle strips were stimulated with 10−5 M ACh for 1–10 min or left unstimulated. The ratio of soluble to insoluble vimentin of these muscle strips was assessed by the method of fractionation (16, 25, 34).

ACh stimulation resulted in vimentin partial disassembly in smooth muscle tissues. ACh stimulation increased the amount of soluble vimentin in smooth muscle strips, whereas the level of insoluble vimentin was decreased in response to contractile activation (Fig. 1A). The increase in the soluble-to-insoluble vimentin ratio was obvious 1–10 min after muscarinic activation (Fig. 1B; n = 5), which is closely associated with the increase in active force (Fig. 1C; n = 10).

Fig. 1.

ACh stimulation increases soluble-to-insoluble vimentin ratio in smooth muscle. A: effects of ACh stimulation on amount of soluble [supernatant (S)] and insoluble [pellet (P)] vimentin (Vim). Tracheal smooth muscle strips were stimulated with 10−5 M ACh for 1–10 min (1m, 5m, 10m) or left unstimulated, and supernatant and pellet fractions were separated and assessed by immunoblot analysis. C, unstimulated smooth muscle. B: ratio of soluble to insoluble vimentin as assessed by scanning densitometry of immunoblots for the fractions. Values are means ± SE (n = 5). C: increase in active force development stimulated by 10−5 M ACh. Force is expressed as percentage of contractile response to 5 min of stimulation by 10−5 M ACh. Values are means ± SE (n = 10). *Significantly different from No ACh (P < 0.05).

Vimentin undergoes phosphorylation at Ser56 in smooth muscle strips in response to ACh stimulation.

Vimentin dynamics may be regulated by its phosphorylation on serine/threonine positions. We previously showed that vimentin is phosphorylated on Ser56 in cultured smooth muscle cells in response to serotonin stimulation (25). To determine whether muscarinic activation induces vimentin phosphorylation in smooth muscle tissues, tracheal smooth muscle strips were stimulated with 10−5 M ACh for 1–10 min. Vimentin phosphorylation on Ser56 was determined by immunoblot analysis using phosphorylated vimentin (Ser56) antibody. Unstimulated muscle strips were also frozen for vimentin phosphorylation.

ACh stimulation of tracheal smooth muscle resulted in the enhancement of vimentin phosphorylation at Ser56. The phosphorylation levels in smooth muscle strips in response to 1–10 min of ACh stimulation were significantly higher than the phosphorylation level in unstimulated smooth muscle strips (Fig. 2; P < 0.05, n = 4).

Fig. 2.

ACh stimulates vimentin phosphorylation on Ser56 in smooth muscle. A: increase in vimentin phosphorylation at Ser56 during 1–10 min of ACh stimulation. Blots of extracts of unstimulated and ACh-stimulated muscle strips were probed using phosphorylated vimentin antibody [p-Vim (Ser56)], stripped, and reprobed with vimentin antibody (Vim). C, unstimulated smooth muscle. B: quantification of vimentin phosphorylation as multiples of phosphorylation levels in unstimulated muscle strips. Values are means ± SE (n = 4). *Significantly different from No ACh (P < 0.05).

CAS binds to unphosphorylated vimentin in vitro.

The vimentin framework binds to CamKII in smooth muscle cells and Rho kinase in fibroblasts (15, 20). The adapter protein CAS has been shown to regulate active force development in arterial smooth muscle (17, 29). Our previous studies demonstrated that Ser56 on vimentin is a major phosphorylation site mediated by PAK; phosphorylation at this residue triggers vimentin fiber disassembly (14, 25). To evaluate whether CAS associates with vimentin filaments and whether vimentin phosphorylation at Ser56 influences its binding to CAS in vitro, far-Western analysis was performed. Purified wild-type vimentin and the nonphosphorylatable vimentin S56A (alanine substitution at Ser56) mutant were treated with active PAK and then with high-sodium solution to facilitate filament assembly (14, 25). Blots of CAS immunoprecipitates from tracheal smooth muscle strips were reacted with the phosphorylated or unphosphorylated proteins and detected using vimentin antibody.

Treatment with PAK led to Ser56 phosphorylation of wild-type vimentin, but not vimentin S56A mutant (14, 25). CAS is able to bind to unphosphorylated (assembled), but not phosphorylated (disassembled), vimentin. Phosphorylation of wild-type vimentin by PAK was not associated with immobilized CAS immunoprecipitates on blots. In contrast, nonphosphorylatable vimentin S56A mutant interacted with CAS immunoprecipitates on the blots (Fig. 3).

Fig. 3.

Far-Western analysis of Crk-associated substrate (CAS)-vimentin interaction in vitro. CAS immunoprecipitates from muscle extracts were resolved by SDS-PAGE and transferred to membranes. Lane 1, immobilization of CAS immunoprecipitates on the membrane probed with CAS antibody. Lane 2, no association of vimentin with CAS immobilized on the membrane incubated with wild-type vimentin that had been treated with active p21-activated kinase (PAK) and then probed with vimentin antibody, indicating that phosphorylated vimentin is not able to bind CAS in vitro. Lane 3, no interaction of vimentin S56A mutant with CAS on membrane incubated with vimentin mutant S56A (alanine substitution at Ser56) that had been treated with active PAK and then probed with vimentin antibody, suggesting that nonphosphorylatable vimentin is able to bind CAS in vitro. Immunoblots (IB) are representative of 3 experiments.

Amount of CAS in insoluble vimentin fraction is reduced in ACh-stimulated smooth muscle strips.

We assessed the association of CAS with vimentin filaments in the context of smooth muscle tissues. Soluble and insoluble (cytoskeletal) vimentin from unstimulated smooth muscle strips were separated by SDS-PAGE, and blots of the fractions were probed with antibodies against CAS and vimentin. CAS was found in the fractions of insoluble and soluble vimentin (Fig. 4A). Quantification analysis showed that 25% of total CAS was associated with the insoluble fraction (Fig. 4B; n = 5). However, CAS was not associated with soluble vimentin; no CAS was found in vimentin immunoprecipitates from muscle extracts containing soluble vimentin. Nor did ACh stimulation change the amount of CAS in vimentin immunoprecipitates (Fig. 4C).

Fig. 4.

Adapter protein CAS can be found in the cytoskeletal vimentin fraction. A: presence of CAS in insoluble (cytoskeletal) vimentin. Blots of supernatant (S) and pellet (P) fractions from unstimulated smooth muscle strips were probed with CAS antibody and vimentin antibody. B: quantification of amount of CAS on immunoblots from each fraction after scanning densitometry analysis. Relative amount of CAS in soluble fraction is expressed as follows: soluble CAS ÷ total CAS (soluble CAS + insoluble CAS) × 100. Amount of insoluble CAS is expressed as follows: insoluble CAS ÷ total CAS × 100. Values are means ± SE (n = 5). C: CAS antibody and vimentin antibody detected in blots of vimentin immunoprecipitates from muscle extracts containing soluble vimentin. CAS was not detectable in vimentin immunoprecipitates in unstimulated and ACh-stimulated muscle extracts, indicating that CAS does not associate with soluble vimentin. Immunoblots are representative of 3 experiments.

We also assessed the effects of ACh stimulation on the association of CAS with cytoskeletal vimentin in smooth muscle. Treatment with ACh decreased the amount of CAS associated with cytoskeletal vimentin (Fig. 5A). The ratios of CAS to vimentin were statistically lower in ACh-stimulated than in unstimulated muscle strips (Fig. 5B; n = 4, P < 0.05).

Fig. 5.

Level of CAS in cytoskeletal vimentin in response to ACh stimulation is reduced in smooth muscle strips. Equal amounts of insoluble vimentin from unstimulated and ACh-stimulated smooth muscle strips were separated by SDS-PAGE and transferred to nitrocellulose membranes. Membranes were probed with CAS antibody and vimentin antibody. A: effects of ACh activation on amount of CAS in vimentin fraction. B: normalization of ratio of CAS to vimentin in stimulated muscles to that in unstimulated muscles. Values are means ± SE (n = 4). *Significantly different from No ACh (P < 0.05).

Peripheral localization of CAS is increased in freshly dissociated smooth muscle cells in response to ACh stimulation.

We also assessed the effects of ACh stimulation on the subcellular distribution of CAS. Unstimulated and ACh-stimulated (10 μM, 5 min) smooth muscle cells were immunostained for CAS and analyzed under a confocal fluorescence microscope.

Exposure of smooth muscle cells to ACh caused a redistribution of CAS from the cytoplasm to the cortical region. CAS staining was found primarily in the myoplasm of unstimulated smooth muscle cells (Fig. 6A,a). In response to ACh stimulation, peripheral distribution of CAS was increased, while fluorescence intensity of the protein was decreased in the cell interior (Fig. 6A,b). However, a fraction of total CAS was still observed in the cytoplasm of stimulated cells (Fig. 6A,b). Ratios of pixel intensity at the cell periphery to pixel intensity at the cell interior were twofold higher in ACh-stimulated than in unstimulated cells (Fig. 6B; n = 18, P < 0.05).

Fig. 6.

ACh stimulation increases peripheral distribution of CAS in freshly dissociated smooth muscle cells. Unstimulated cells and cells stimulated with 10 μM ACh for 5 min were freshly dissociated from tracheal smooth muscle tissues, immunostained for CAS, and analyzed under a confocal fluorescence microscope. A: CAS in myoplasm, as well as at the cell border, without ACh stimulation (a). Cell border-associated distribution of CAS was increased in cells in response to ACh stimulation (b). Arrow indicates a single line scan to quantify pixel intensity for each cell. B: protein distribution in cells expressed as ratio of pixel intensity at the cell periphery to ratio of pixel intensity at cell interior. Each mean value was obtained from an average of 3–4 line scans in each of 18 unstimulated and 18 stimulated cells from 3 experiments. *Significantly different from No ACh (P < 0.05).

Downregulation of PAK with antisense inhibits the decrease in association of CAS with cytoskeletal vimentin during ACh stimulation.

PAK has been shown to be an upstream regulator of the vimentin network (25). To evaluate whether PAK affects the interaction of CAS with cytoskeletal vimentin, antisense ODN-, sense ODN-, or non-ODN-treated tracheal smooth muscle strips were stimulated with 10−5 M ACh for 5 min or left unstimulated, and the association of CAS with insoluble vimentin was assessed.

We first verified the effectiveness of PAK antisense treatment. PAK expression was lower in antisense ODN- than in sense ODN- or non-ODN-treated muscle strips (Fig. 7A). MLCK was similar in ODN antisense-, ODN sense-, and non-ODN-treated muscle strips (Fig. 7A). The normalized ratio of PAK to MLCK in antisense-treated tissues was 0.28 ± 0.04, which was significantly lower than that in sense-treated (1.07 ± 0.07) and non-ODN-treated (1.00 ± 0.00) muscle strips (n = 4, P < 0.01).

Fig. 7.

PAK downregulation attenuates dissociation of CAS from insoluble vimentin elicited by ACh. A: PAK1 downregulation by antisense in smooth muscle strips. Blots of protein extracts from tracheal smooth muscle tissues that had been treated with antisense or sense oligodeoxynucleotides (ODNs) or with no ODNs were detected with antibodies against PAK1 and myosin light kinase (MLCK). Amount of PAK1 was lower in antisense- than in non-ODN- or sense ODN-treated muscle strips. Similar amounts of MLCK, a key enzyme in smooth muscle, were detected in all 3 groups of muscle strips. B: CAS-to-vimentin ratio in smooth muscle strips treated with PAK antisense (AS) or sense (S) or with no ODNs and stimulated with 10−5 M ACh for 5 min (solid bars) or left stimulated (open bars). CAS-to-vimentin ratio in smooth muscle strips with various treatments was normalized to ratio obtained from unstimulated non-ODN-treated muscle strips. Values are means ± SE (n = 4). *Significantly different from corresponding unstimulated value (P < 0.05).

The downregulation of PAK inhibited the ACh-induced decrease in the CAS-to-vimentin ratio in muscle tissues. Compared with non-ODN-treated muscle strips, treatment with PAK antisense or sense ODNs did not affect the ratio of CAS to insoluble vimentin in unstimulated muscle strips. In non-ODN- and sense ODN-treated muscle strips, ACh stimulation resulted in a significant decrease in the CAS-to-vimentin ratio (Fig. 7B; n = 4, P < 0.05). However, the ratio of CAS to vimentin on stimulation with ACh was not decreased in antisense-treated tissues compared with corresponding unstimulated muscles (Fig. 7B; n = 4).

Depletion of PAK attenuates the ACh-induced increase in vimentin phosphorylation at Ser56, disassembly, and active force in smooth muscle strips.

We speculated that the effects of PAK on the association of CAS with cytoskeletal vimentin stem from vimentin phosphorylation and disassembly mediated by PAK. PAK antisense ODN-, sense ODN-, or non-ODN-treated tracheal smooth muscle strips were stimulated with 10−5 M ACh for 5 min or left unstimulated, and vimentin phosphorylation at Ser56 and the soluble-to-insoluble vimentin ratios were evaluated.

Although vimentin phosphorylation in unstimulated muscle strips was not affected by treatment with PAK antisense ODNs, vimentin phosphorylation at Ser56 in response to ACh stimulation was significantly lower in PAK antisense- than sense- or non-ODN-treated muscle strips (Fig. 8, A and B; n = 4, P < 0.05).

Fig. 8.

Vimentin phosphorylation at Ser56, soluble-to-insoluble vimentin ratio, and ACh-induced active force are diminished in PAK-deficient muscle tissues. PAK antisense-, sense-, or non-ODN-treated smooth muscle strips were stimulated with 10−5 M ACh for 5 min or left unstimulated (US). A: effects of PAK downregulation on vimentin phosphorylation at Ser56. B: vimentin phosphorylation quantified as multiples of phosphorylation levels obtained in non-ODN-treated unstimulated muscle strips. Open bars, unstimulated; solid bars, ACh stimulated. Values are means ± SE (n = 4). *Significantly different from corresponding unstimulated value (P < 0.05). **Significantly different from No ODNs and PAK S (P < 0.05). C: effects of PAK depletion on soluble-to-insoluble vimentin ratio. Open bars, unstimulated; solid bars, ACh stimulated. Values are means ± SE (n = 4). *Significantly different from corresponding unstimulated value (P < 0.05). **Significantly different from No ODNs and PAK S (P < 0.05). D: mean active force in response to 10−5 M ACh as percentage of ACh-induced force in each strip before incubation. Values are means ± SE (n = 8). *Significantly different from No ODNs or PAK S (P < 0.05).

Similarly, treatment with PAK antisense or sense ODNs did not affect the ratio of soluble to insoluble vimentin in unstimulated muscle strips; however, the soluble-to-insoluble vimentin ratio on stimulation with ACh was significantly lower in antisense- than in non-ODN- or sense-treated muscle strips (Fig. 8C; n = 4, P < 0.05).

PAK downregulation also inhibited active force in response to ACh stimulation. Contractile force of tracheal smooth muscle strips was compared before and after 2 days of incubation with PAK sense or antisense (26, 32, 35). Although active force in sense ODN- and non-ODN-treated muscle strips was not suppressed after the 2-day incubation, contractile response in antisense-treated tissues was significantly reduced to 21% of preincubation force (n = 8, P < 0.01; Fig. 8D). There were no differences in passive tension among the three groups of strips.

DISCUSSION

The type III intermediate filament protein vimentin is required for active force development in smooth muscle (10, 35). The adapter protein CAS participates in the signaling cascades that regulate smooth muscle contraction (17, 29). Our present results demonstrate that ∼25% of total CAS binds to the vimentin framework in smooth muscle tissues and that ACh induces the disassociation of CAS from the vimentin cytoskeleton, which may be regulated through vimentin phosphorylation and disassembly by PAK. We propose that the PAK-mediated dissociation of CAS from cytoskeletal vimentin may be an important part of the cellular mechanisms that regulate active force development in smooth muscle.

Recent studies indicate that vimentin filaments continuously exchange between a small disassembled faction and a large assembled fraction in certain cell types. External stimulation has been shown to induce vimentin disassembly in a variety of cultured cells, including smooth muscle cells (5, 19, 25). Vimentin disassembly has been implicated in the regulation of cell mitosis and migration (1, 16). In this report, ACh stimulation increased the ratio of soluble to insoluble vimentin and active force development, suggesting that vimentin undergoes disassembly in association with contractile force in smooth muscle in response to agonist activation. The ratio of soluble to insoluble vimentin in unstimulated muscles was ∼0.1, indicating that 10% of total vimentin is in disassembled form. The ratio of soluble to insoluble vimentin during 5–10 min of ACh stimulation was increased to ∼0.25, implying that 20% of total vimentin is soluble (25).

How the dynamic change in vimentin filaments may affect force development in smooth muscle tissues is unknown. The adapter protein CAS has been reported to participate in the signaling process that mediates smooth muscle contractility (17, 29). In this report, we found that CAS was able to bind to cytoskeletal vimentin in in vitro studies, as assessed by far-Western analysis, as well as in smooth muscle tissues, as evaluated by the fractionation method. The amount of CAS associated with cytoskeletal vimentin was attenuated during muscarinic activation. These results indicate that vimentin disassembly during contractile activation may induce the release of CAS from the vimentin cytoskeleton, which may allow involvement of more CAS in the signaling cascades that mediate smooth muscle contraction (17, 29).

Our previous studies showed that CAS is able to regulate the activation of the actin-regulatory protein profilin and actin dynamics in smooth muscle (17, 29). It has been well documented that actin polymerization occurs in tracheal smooth muscle strips in response to ACh stimulation. Remodeling of the actin cytoskeleton has emerged as an essential event during contractile stimulation of smooth muscle (2, 26, 30, 32, 37). Actin dynamics in smooth muscle tissues are believed to occur in the cortical region of cells (37). In the present study, ACh activation of smooth muscle initiated the redistribution of CAS from the myoplasm to the peripheral area. Thus it is probable that the CAS released from the vimentin cytoskeleton may translocate to the cell border, facilitating cortical actin polymerization and force development in smooth muscle.

Partial disassembly of vimentin may also mediate changes in the cytoskeletal systems, which might modulate contractile responses. Vimentin filaments in differentiated smooth muscle cells display a well-spread network that extends from the nuclear membrane to the plasma membrane (15, 25). The membrane-associated anchoring of intermediate filaments is believed to occur at the desmosome, an intercellular junction (6, 13, 21). Vimentin filament disassembly in response to stimulation with soluble factors is related to the spatial reorganization of the vimentin network in various cell types (16, 25, 34). Vimentin downregulation disrupts desmosomal organization and suppresses smooth muscle contraction (35). Therefore, the spatial reorientation of vimentin filaments in smooth muscle could strengthen the linkage of vimentin filaments to the desmosome, which may promote intercellular mechanical force transmission (13, 25).

We then sought to understand the mechanisms that regulate the unique association of CAS with the vimentin network. In vitro studies have shown that PAK directly catalyzes vimentin phosphorylation (8, 25). The expression of constitutively active PAK induces vimentin phosphorylation in COS-7 cells (8). The silencing of PAK attenuates vimentin phosphorylation in cultured smooth muscle cells during serotonin stimulation (25). In the present study, PAK downregulation depressed the dissociation of CAS from cytoskeletal vimentin in response to ACh activation. The results lead us to suggest that PAK is a pivotal upstream regulator for the association of CAS with the vimentin network on contractile activation of smooth muscle.

PAK may regulate the interaction of CAS with cytoskeletal vimentin and force development by affecting vimentin phosphorylation and disassembly in smooth muscle tissues. In BHK-21 fibroblasts, treatment with the protein phosphatase inhibitor calyculin-A induces vimentin phosphorylation in concert with the disassembly of vimentin polymers into soluble tetramers (5). In COS-7 cells expressing the active form of PAK, the vimentin framework displays a granulate-like structure (disassembly) (8). In this report, PAK depletion diminished vimentin phosphorylation at Ser56 and the transition of cytoskeletal vimentin to soluble vimentin during muscarinic activation. Active force development was also inhibited in the PAK-deficient tissues. As mentioned above, the ACh-stimulated dissociation of CAS from cytoskeletal vimentin was attenuated in the PAK-deficient tissues. Thus we propose that the contractile stimulation may induce vimentin phosphorylation at Ser56 and disassembly via PAK. The partial disassembly of the vimentin framework may facilitate the release of vimentin-associated CAS, which may be involved in the induction of actin polymerization and force development (2, 26, 29, 32, 37) (Fig. 9).

Fig. 9.

Proposed mechanism. Contractile stimulation with ACh may trigger vimentin phosphorylation at Ser56 and vimentin partial disassembly via PAK, which may initiate dissociation of CAS from the vimentin network. “Released” CAS may participate in the signaling processes (e.g., actin polymerization) that regulate force development in smooth muscle.

Vimentin phosphorylation may be mediated by several other kinases in vitro and/or other cell types. Although protein kinase A (PKA) is able to phosphorylate vimentin in in vitro biochemical studies (5), it is unlikely that ACh exposure activates PKA in smooth muscle. The activation of PKA by β-agonists leads to airway smooth muscle relaxation (11, 24), whereas ACh stimulation induces force generation in the smooth muscle. Protein kinase C and Rho kinase may be activated in smooth muscle in response to ACh stimulation. However, inhibition of protein kinase C or Rho kinase does not inhibit vimentin phosphorylation at Ser56 (a critical site for smooth muscle function), although several other phosphorylation sites may be mediated by these two kinases (7, 12, 25, 33). In addition, CamKII may mediate phosphorylation of vimentin on several residues other than Ser56 in vitro (5, 33). Finally, cyclin-dependent kinase 1 and polo-like kinase 1 have been reported to mediate vimentin phosphorylation during mitosis of U251 and T24 cells (36). Although there is no evidence that these two kinases can be activated in differentiated cells, we do not rule out the possibility that ACh stimulation may activate these two kinases, which might mediate vimentin phosphorylation at Ser56 in smooth muscle tissues.

In summary, ACh stimulation of tracheal smooth muscle strips induces disassembly of cytoskeletal vimentin and vimentin phosphorylation at Ser56. More importantly, ACh activation leads to the decrease in the distribution of CAS in the vimentin cytoskeleton in concert with active force development. CAS translocates to the peripheral area from the myoplasm on contractile stimulation. The downregulation of PAK inhibits the decrease in the association of CAS with cytoskeletal vimentin and active force, probably by attenuating vimentin phosphorylation at this residue and its disassembly. We conclude that the PAK-mediated CAS release from the vimentin network may be an essential cellular process on muscarinic activation of smooth muscle tissues.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant HL-75388, an American Heart Association Scientist Development Grant, and the Indiana Showalter Foundation (to D. D. Tang).

Acknowledgments

The authors thank Amy M. Spinelli and Taoying Huang for assistance.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract