Tropomyosin variants describe distinct functional subcellular domains in differentiated vascular smooth muscle cells

Abstract

Tropomyosin (Tm) is known to be an important gatekeeper of actin function. Tm isoforms are encoded by four genes, and each gene produces several variants by alternative splicing, which have been proposed to play roles in motility, proliferation, and apoptosis. Smooth muscle studies have focused on gizzard smooth muscle, where a heterodimer of Tm from the α-gene (Tmsm-α) and from the β-gene (Tmsm-β) is associated with contractile filaments. In this study we examined Tm in differentiated mammalian vascular smooth muscle (dVSM). Liquid chromatography-tandem mass spectrometry (LC MS/MS) analysis and Western blot screening with variant-specific antibodies revealed that at least five different Tm proteins are expressed in this tissue: Tm6 (Tmsm-α) and Tm2 from the α-gene, Tm1 (Tmsm-β) from the β-gene, Tm5NM1 from the γ-gene, and Tm4 from the δ-gene. Tm6 is by far most abundant in dVSM followed by Tm1, Tm2, Tm5NM1, and Tm4. Coimmunoprecipitation and coimmunofluorescence studies demonstrate that Tm1 and Tm6 coassociate with different actin isoforms and display different intracellular localizations. Using an antibody specific for cytoplasmic γ-actin, we report here the presence of a γ-actin cortical cytoskeleton in dVSM cells. Tm1 colocalizes with cortical cytoplasmic γ-actin and coprecipitates with γ-actin. Tm6, on the other hand, is located on contractile bundles. These data indicate that Tm1 and Tm6 do not form a classical heterodimer in dVSM but rather describe different functional cellular compartments.

tropomyosin (Tm) is an abundant F-actin binding protein known to regulate actomyosin interactions in adult striated (overviews in Refs. 5, 17, 24, and 62) and smooth muscle (overview in Ref. 33). The four Tm genes (overview in Refs. 11 and 29), TPM1 (α), TPM2 (β) TPM3 (γ), and TPM4 (δ), lead to the production of multiple variants by alternative splicing (55). Smooth muscle Tm, based on work primarily using gizzard tissue, has been focused essentially exclusively on the heterodimers of the smooth muscle-specific Tmsm-α variant (from the α-gene) and the Tmsm-β variant (from the β-gene) (6, 59, 60), associated with the smooth muscle actin of the contractile filaments of gizzard. In contrast, >40 mammalian variants have been described and the majority of them are expressed in nonmuscle cells (12, 25, 55, 56, 58). It is known that the differentiation status of smooth muscle cells affects the expression pattern of Tm proteins. The TPM1 gene encodes for a smooth muscle-specific exon, exon 2a, which is only expressed in the differentiated state. Upon a switch to a synthetic phenotype, exon 2a is no longer utilized during splicing and instead exon 2b is used, creating the fibroblast-type Tm protein (9, 58).

In striated muscle, the function of Tm has been intensively studied and it is known that Tm proteins act in concert with troponin to regulate a Ca²⁺-dependent activation of actin filaments (overview in Ref. 5). In smooth muscle, however, there is no troponin, and contraction is mainly controlled by phosphorylation of the myosin light chain (LC20) (49) and the phosphorylation status of h-caldesmon, a smooth muscle-specific actin-binding protein (overview in Ref. 23). Here Tm might act as a gatekeeper for actin function by controlling the accessibility of other actin-binding proteins to actin filaments, and thereby regulating both actin polymerization as well as the interaction of actin and myosin.

Our lab and others have previously shown that actin net polymerization is important in regulating contractility of differentiated vascular smooth muscle cells (dVSMCs; 20–22). We have previously found that α-adrenergic agonist-induced contractility is primarily associated with a change in net polymerization of γ-actin-containing filaments, not with the α-actin filaments contained in the contractile filament bundles. On the basis of indirect evidence as well as work in nonmuscle cells, we predicted that γ-actin might be localized in the cell cortex. Unfortunately, we were not able to specifically image for γ-actin due to the lack of a specific antibody. A specific anti-cytoplasmic γ-actin antibody has now been developed (4), and we show here that cytoplasmic Beta, not γ-actin is localized in a cortical F-actin network in dVSMCs.

By antibody screening and liquid chromatography-tandem mass spectrometry (LC MS/MS), we have identified a total of five Tm variants in adult vascular smooth muscle. In addition to α-smooth muscle Tm6, Tm2 was identified from the α-gene. Additionally, from dVSM homogenates, evidence was found for Tm1 from the β-gene, Tm5NM1 from the γ-gene, and Tm4 from the δ-gene. We show here that Beta-Tm1 is less abundant than Tm6, is primarily localized in the cortex of differentiated smooth muscle cells, and coimmunoprecipitates with γ-actin. α-Smooth muscle Tm6 is not colocalized with Tm1, but does colocalize with α-smooth muscle actin and β-actin in contractile filament bundles.

MATERIALS AND METHODS

Tissue preparation.

All procedures in this study were performed according to protocols approved by the Institutional Care and Use Committee of Boston University. Ferrets (Marshall Farms, North Rose, NY) were euthanized by an overdose of isoflurane, and the aorta was removed and placed in an oxygenated physiological salt solution (PSS; in mM: 120 NaCl, 5.9 KCl, 1.2 NaH₂PO₄, 25 NaHCO₃, 11.5 dextrose, 1 CaCl₂, and 1.4 MgCl₂, pH 7.4). Aorta tissue was dissected under a microscope in oxygenated (95% O₂ and 5% CO₂) PSS to remove the adventitia and the endothelium. On the basis of previous studies showing that the normal intima and media of all mammalian arteries contain only smooth muscle cells (42), we assumed that the tissue we were working with was not contaminated with fibroblasts or other cell types. After 1 h of equilibration, tissue viability was tested by stimulation for 10 min with 51 mM KCl-PSS (51 mM of NaCl replaced with 51 mM KCl). After washing the tissue and relaxation for 1 h, tissue was used for further experiments. Finally, tissue was quick frozen in a dry ice-acetone slurry containing 10 mM DTT.

Identification of Tm variants in aorta via LC MS/MS analysis.

For purification of Tm, a protocol from Bretscher (1) was used. Briefly, smooth muscle tissue from freshly isolated ferret aorta was homogenized in extraction buffer (0.3 mM KCl, 1 mM EGTA, 0.5 mM MgCl₂, 50 mM imidazole-HCl, pH 6.9, and protease inhibitors) on ice. The homogenate was then heated in a boiling water bath for 10 min before chilled on ice for 10 min. The material was next cleared by centrifugation for 30 min at 47,000 g at 4°C. The supernatant, containing heat-stable proteins like Tm and caldesmon, was analyzed on a 10% SDS-Gel and proteins were detected via Coomassie staining. Five protein bands in the size range of 42 to 28 kDa (where Tm proteins were expected) were excised from the gel and subjected to in-gel reduction, alkylation, and enzymatic digestion (Roche Applied Science, Indianapolis, IN) in a HEPA-filtered hood to reduce keratin background. LC MS/MS analysis was performed on the in-gel digest extracts using an Agilent (Santa Clara, CA) 1100 binary pump directly coupled to a mass spectrometer. Six microliters of sample were injected on column using a LC Packings (Sunnyvale, CA) FAMOS autosampler. Nanobore electrospray columns were constructed from 360 μm outer diameter, 75 μm inner diameter fused silica capillary with the column tip tapered to a 10-μm opening (New Objective, Woburn, MA). The columns were packed with 200 Å 5 μm C18 beads (Michrom BioResources, Auburn, CA), a reverse-phase packing material, to a length of 10 cm. The flow through the column was split precolumn to achieve a flow rate of 320 nl/min. The mobile phase used for gradient elution consisted of 1) 0.3% acetic acid-99.7% water and 2) 0.3% acetic acid-99.7% acetonitrile. Tandem mass spectra (LC MS/MS) were acquired on a Thermo LTQ ion trap mass spectrometer (Thermo, San Jose, CA). Needle voltage was set to 3 kV, isolation width was 3 Da, relative collision energy was 30%, and dynamic exclusion was used to exclude recurring ions. Ion signals above a predetermined threshold automatically triggered the instrument to switch from MS to MS/MS mode for generating fragmentation spectra. The MS/MS spectra were searched against the National Center for Biotechnology Information nonredundant protein sequence database using the SEQUEST computer algorithm (61) to produce a list of proteins identified in each sample. LC MS/MS analysis was performed at both Tufts University and Boston University Core facilities.

Isolation of single differentiated vascular smooth muscle cells from aorta.

The method used for isolating dVSMCs was described previously (21). Briefly, ferret aorta tissue was cut into small pieces and incubated in Ca²⁺ and Mg²⁺-free HBSS containing 100 U/ml type II collagenase (Worthington), 1 U/ml elastase (Roche), and 0.2% BSA (Sigma). Tissue pieces were then filtered on a nylon mesh membrane (pore size 500 μm) and washed with Ca²⁺ and Mg²⁺-free HBSS and 0.2% BSA. The wash buffer containing the isolated differentiated vascular smooth muscle cells was poured onto glass coverslips on ice under a 100% oxygen atmosphere, stored in PSS, and used for immunofluorescence staining.

Generation of antibodies against different Tm variants.

Tm variant-specific polyclonal antibodies have been previously described (47): α9d (sheep polyclonal) recognizes the peptide sequence encoded by exon 9d from the TPM1 gene and detects Tm1, 2, 3, 5a, 5b; α2a (sheep polyclonal) recognizes the peptide sequence encoded by exon 2a from the TPM1 gene and detects Tm6; γ9d (sheep polyclonal) recognizes the peptide sequence encoded by exon 9d from the TPM3 gene (both human and mouse) corresponding to Tm5NM1 and Tm5NM2; δ9d (rabbit polyclonal) recognizes the peptide sequence encoded by exon 9d from the TPM4 gene and detects Tm4 (13).

Monoclonal antibodies, CG1 and CG-β6, were generated against chicken gizzard Tms and characterized previously (15, 16, 28, 30). In addition to chicken Tms, the CG1 specifically recognizes human Tm isoform, hTm1, whereas the CG-β6 recognizes the epitope in the COOH-terminal region of hTM2 and hTM3 (7, 29, 30). Monoclonal antibodies LC24 and LC1, generated against bacterially produced hTM5/4 (NH₂-terminal of hTM5 fused to the C-terminal of hTM4) fusion protein, specifically recognize the hTm4 and hTM5, respectively (29, 32, 53, 57).

Generation of Tm standards.

The entire coding regions of hTm1, hTm2, hTm3, hTm4, or hTm5 cDNAs (38) were subcloned into the expression vector pET8c/s as described previously (39). The resulting plasmids were separately transformed to BL21(DE3)LysS cells for the production of respective Tm variants. The bacterially produced Tm was purified as described previously with FPLC monoQ and hydroxylapatite column chromatographs as the final steps. The purified Tm variants were used for quantification of Tms in dVSMCs. Chicken gizzard Tm, used as standard for quantifying Tm6 in dVSMCs, was prepared as described previously (19).

Immunofluorescence analysis.

For immunofluorescence analysis of Tm1, Tm6, α-actin, and β-actin (costaining with Tm6), differentiated vascular smooth muscle cells were fixed with 4% paraformaldehyde for 10 min. To remove residual paraformaldehyde, two washing steps for 5 min with 0.1 mM glycine and 1% BSA in HBSS followed. For permeabilizing the cells we used 0.1% Triton X-100 in HBSS for 10 min. After blocking with 10% goat serum (for Tm1, α-actin), respectively, donkey serum (for Tm6), 0.05% Triton X-100, and 1% BSA for 30 min, cells were incubated, depending on the experiment, with the sheep polyclonal anti-Tm6 antibody (Sh X Tropomyosin, Smooth Muscle; Millipore) at 1:200 dilution, the mouse monoclonal anti-Tm1 antibody [CG1; (7, 15, 16, 28)] at 1:500 dilution, the mouse monoclonal anti-α-actin antibody (Sigma) at 1:10,000 dilution, or the mouse monoclonal anti-β-actin antibody (Sigma) at 1:400 dilution overnight. For immunofluorescence staining of β-actin (single staining) and cytoplasmic γ-actin, differentiated vascular smooth muscle cells were fixed with 1% paraformaldehyde for 20 min followed by ice-cold methanol for 3 min. Cells were washed with HBSS containing 0.075% BSA and then permeabilized with 0.1% TX-100 for 10 min. Cells were then incubated with the mouse monoclonal anti-β-actin antibody (4) at 1:100 dilution or the mouse monoclonal anti-cytoplasmic γ-actin antibody (4) at 1:500 dilution over night. As secondary antibodies, Alexa488-conjugated goat anti-mouse IgG (Molecular Probes), Alexa568-conjugated goat anti mouse IgG (Molecular Probes), Alexa488-conjugated donkey anti-sheep IgG (Molecular Probes), or Alexa568-conjugated donkey anti-sheep or anti mouse IgG (Molecular Probes) was used at 1:400 dilution for 30 min. Total F-actin was stained with Alexa568-labeled phalloidin (Molecular Probes) at 1:5,000 dilution, and nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI; Sigma). Cells were examined with a Nikon TE 2000-E2 Perfect Focus EpiFluorescence Microscope and a ×60, 1.4 numerical aperture oil immersion objective. For deconvolution, three-dimensional images were acquired and recorded with a CoolSNAP HQ², Photometrics camera. Out-of-focus fluorescent blur was removed by deconvolution using the NIS-Elements AR 3.0 software (Richardson-Lucy algorithm, constrained iterative-maximum likelihood estimation algorithm). Images were transferred with Adobe Photoshop CS4 software.

SDS-PAGE and Western blot analysis.

Quick-frozen tissue was homogenized in homogenization buffer (20 mM MOPS, pH 6.8, 10% glycerol, 4% SDS, 1% Triton X-100, 3 mM DTT, 67 μM ZnCl₂, 29.6 mM β-glycerophosphate, protease inhibitors), and the homogenate was analyzed via SDS-PAGE according to standard procedures and electrophoretically transferred onto polyvinylidene difluoride membranes (Millipore). Residual protein-binding sites on the membrane were blocked with blocking buffer (Odyssey, LI-COR Biosciences, Lincoln, NB). The membranes were incubated either with the sheep polyclonal anti-Tm6 antibody (Sh X Tropomyosin) at 1:300 dilution, the mouse monoclonal anti-Tm1 antibody [CG1; (7, 15, 16, 28)] at 1:2,000 dilution, the mouse monoclonal anti-Tm2 and Tm3 antibody [CG-β6; (28, 30)] at 1:500 dilution, the mouse monoclonal anti-Tm4 antibody [LC24; (57)] at 1:500 dilution, the mouse monoclonal anti-Tm5NM1 antibody [LC1; (53, 57)] at 1:500 dilution, the sheep polyclonal α9d used at 1:400, the sheep polyclonal α2a used at 1:100, the sheep polyclonal γ9d used at 1:100, the rabbit polyclonal δ9d used at 1:250, the rabbit monoclonal anti-α-actin antibody (Epitomics) at 1:3,000 dilution, the mouse monoclonal anti-β-actin antibody (Sigma) at 1:100 dilution or the mouse monoclonal anti-γ-actin antibody (Santa Cruz) at 1:500 dilution. As secondary antibodies, IRDye680-conjugated donkey anti-sheep IgG (Odyssey, LI-COR Biosciences) or IRDye800CM-conjugated goat anti mouse IgG (Odyssey, LI-COR Biosciences) were used at 1:1,000 dilution. Bound antibodies were detected with an Odyssey 1400 Infrared Imager (Odyssey, LI-COR Biosciences).

Immunoprecipitation.

Quick frozen tissue was homogenized in immunoprecipitation (IP) buffer (50 mM Tris·HCl pH 7.4, 1 mM EGTA, 50 mM NaCl, 3 mM MgCl₂, 1% Nonident P-40, 1% sodium deoxycholate, 3 mM ATP, 3 mM DTT, 67 μM ZnCl₂, 29.6 mM β-glycerophosphate, protease inhibitors) and then precleared with protein G-coupled dynabeads (Invitrogen). To reduce background, the antibodies used in this assay were cross-linked to protein G-coupled dynabeads (Invitrogen) using the cross-linking reagent Bis(sulfosuccinimidyl)suberate (BS₃) following the manufacturer's instructions. The precleared cell lysates were incubated for 16 h at room temperature with either the sheep polyclonal anti-Tm6 antibody (Sh X Tropomyosin) or the mouse monoclonal anti-Tm1 antibody [CG1; (7, 15, 16, 28)], cross-linked to protein G-coupled dynabeads. As negative control, the mouse monoclonal anti-green fluorescent protein antibody (Clontech) cross-linked to protein G-coupled dynabeads was used. Precipitated proteins were washed three times with IP buffer and eluted from the beads by boiling for 5 min in sample buffer. Samples were separated on SDS-PAGE and further analyzed by Western blotting.

Statistical analysis.

All values in the text are means ± SE. Differences between means were evaluated using a two-tailed Student's t-test. Significant differences were taken at the P < 0.05 level.

RESULTS

Protein expression levels of Tm variants in dVSMCs.

To understand if and how Tm proteins control actin functions in dVSMCs, we first investigated which Tm isoforms are expressed in this cell type. Therefore, smooth muscle tissue from freshly isolated aorta was homogenized and analyzed via Western blotting. A previously described set of polyclonal antibodies (47), raised against different regions of Tm proteins, was used as an initial screen for Tm isoforms present in aortic vascular smooth muscle tissue. As shown in Fig. 1A, an α9d antibody [raised against the peptide sequence encoded by exon 9d of the TPM1 (α) gene], detects Tm6, Tm1, Tm2, and Tm5a/5b. The antibody α2a raised against the peptide sequence encoded by exon 2a of the TPM1 (α) gene, detects Tm6 in aorta. The antibody γ9d raised against the peptide sequence encoded by exon 9d of the TPM3 (γ) gene, binds to three different Tm isoforms: Tm6, Tm1, and Tm5NM1/2. And finally, the δ9d antibody raised against the peptide sequence encoded by exon 9d of TPM4 (δ) gene, detects Tm4 in aorta. The antibody also recognizes a higher molecular weight of unknown identity.

Fig. 1.

Western blot analysis of tropomyosin (Tm) variants in differentiated vascular smooth muscle (dVSM). A: aorta tissue homogenates were separated on SDS-PAGE and further examined by Western blotting and staining with different polyclonal anti-Tm antibodies as indicated. Ba–Bd: 10 μg aortic homogenates were examined via Western blotting and probing with a set of anti-Tm monoclonal antibodies as indicated. For comparison, recombinant human Tm proteins were also loaded. The blank space between the lanes of Ba indicates that these are nonadjacent lanes of the same gel.

We also probed the aortic vascular smooth muscle homogenates with a set of previously described monoclonal antibodies (7, 28, 30, 32, 53). For comparison, we loaded recombinant Tm proteins. As can be seen in Fig. 1Ba, the antibody CG-β6 (raised against chicken gizzard Tm) reacts with recombinant Tm2 and Tm3 from the α-gene (7, 30) and detects a protein band in an aorta lysate (see arrow) running at about the same size as the Tm2 standard. The CG1 antibody (raised against chicken gizzard Tm), shown in Fig. 1Bb, is specific for recombinant Tm1 from the β-gene (7) and stains a specific band of the same size in aorta homogenates (see arrow). Figure 1Bc shows staining with the LC24 antibody [raised against a hTm5/4 fusion protein and specifically recognized hTm4 from nonmuscle human cells; (32, 57)]. This antibody detects both recombinant Tm1 from the β-gene and the short form Tm4 from the δ-gene. In aorta lysates, the antibody stains two bands of the same size as recombinant Tm1 and Tm4 (see arrows) although the intensity of staining of Tm4 is far less intense in aorta, suggesting a lesser abundance. In Fig. 1Bd, staining with the LC1 antibody [raised against a hTm5/4 fusion protein and specifically recognized hTm4 from nonmuscle human cells; (32, 57)] is pictured, showing that it detects a single band in aorta of ∼28 kDa (lane 1, see arrow) which likely represents the short form Tm5NM1 from the γ-gene since it runs at the same height as recombinant Tm5NM1 (lane 2).

LC MS/MS analysis of Tm variants in dVSMCs.

To corroborate the findings for Tm variants present in smooth muscle cells by using specific antibodies on a Western blot, we partially purified Tms by a modified version of the heat treatment method of Bretscher (1) (see materials and methods for details). Aorta tissue was homogenized and boiled to remove heat-unstable proteins. Heat-stable proteins (primarily caldesmon and Tms) were separated on a 10% SDS-gel and visualized via Coomassie staining, and five bands in the size range of 42 to 28 kDa were analyzed by LC MS/MS (see Fig. 2A). The peptides resulting from trypsin digestion were aligned to a human protein database because of the close similarity of the ferret to the human (2, 48, 54). The uppermost band at 37 kDa (band 2) was matched with a 326 aa long form of Tm6, also known as isoform 2 CRA_f of the TPM1 (α) gene (83% coverage). This was surprising, since this long Tm6 form has not been previously described in protein form and has only been previously reported as a transcript (AK131384). This match was based on peptides corresponding to both exon 2a and exon 2b from the α-gene; however, we did not find any peptide corresponding to the transition region of exon 2a/2b which would confirm that those two exons exist in the same protein. Moreover, previous studies have excluded the possibility that exons 2a and 2b from the α-gene can be spliced together, in theory, because the intron is thought to be too short to accommodate the splicing machinery (51). A sequence alignment of exons 2b of the α- and β-gene, compared with the peptide identified, moreover ruled out that the band contains a mix of Tm6 (α-gene) and Tm1 (β-gene) (Fig. 2B, underlined peptide; MS/MS fragmentation for this particular peptide is shown in Fig. 2D). Therefore, it is more likely that the band cut from this region of the gel contains a mix of two different Tm variants from the α-gene, containing exon 2a or exon 2b. Combined with our Western blot staining using antibodies α9d and CG-β6 (Fig. 1, A and Ba), it seems likely that this band contains Tm6 plus Tm2. Furthermore, the fact that we detected peptides only from exon 6b but not 6a from the α-gene makes it more likely that we have Tm2 and not Tm3 in the examined tissue.

Fig. 2.

LC MS/MS analysis of Tm variants in dVSM. A: Coomassie-stained SDS gel of the purified Tm fraction, showing the bands 1-5 that were excised from the gel and used for LC MS/MS analysis. The smaller box shows the top two bands of the gel at a lower exposure to make the two bands visible. LC MS/MS results for the 5 bands are indicated at right and molecular mass is indicated at the far left. B: protein sequence alignment of human exon 2b derived from the TPM1 (top peptide) or the TPM2 (bottom peptide) gene, via ClustalW. A peptide identified via LC MS/MS analysis of the uppermost 37-kDa band/band 2 (identified as a mix of Tm6 and Tm2; see also D) is underlined. C: Western blot staining of total aorta homogenate prepared with the extraction buffer from the Bretscher protocol (1) for Tm purification (“Input”, lane 1), insoluble fraction after high spin of the “Input” (“pellet”, lane 2) and soluble fraction after high spin of the “Input” (“sup”, lane 3) with antibodies (Ab) against Tm6 (α2a, top) and Tm1 (CG1, bottom). D: the MS/MS fragmentation of the doubly charged (m/z 574.3) form of the LKGTEDELDK peptide (1,146.6-Da monoisotopic mass). The gray letters in the protein sequence are the residues within the 81–122 residue region of interest (corresponding to exon 2b). The LKGTEDELDK peptide is underlined in the 81–122 residue region. The b and y fragment ions observed (not all shown in MS/MS spectrum for clarity) are from both single and doubly charged fragment ions. The evidence for the unambiguous determination of this peptide sequence (LKGTEDELDK) is the observance of singly charged fragment ions b₂, b₃, b₄, b₅, b₆, b₇, b₉, b₁₀ and doubly charged fragment ions y₃, y₄, y₅, y₆, y₇, y₈, y₉.

The next band (band 3), of slightly lower molecular mass, contained Tm1 [isoform 2 of the TPM2 (β) gene] (55% coverage), with peptides corresponding to exons 1a, 2b, 3, 4, 5, 6a, 7, 8, and 9d from the β-gene.

Band 5 was identified as a mix of two short Tm isoforms; Tm4 from the δ (38% coverage) and a variant from the γ-gene containing exons 1b and 6a (38% coverage). Since no peptides matching exon 9 from the γ-gene were detected by LC MS/MS, the specific Tm γ-variant is not determined by this method. Western blot staining with either the γ9d (Fig. 1A) or the LC1 antibody (Fig. 1Bd) however, indicates that the short Tm variant from the γ-gene is Tm5NM1/2. Since Tm5NM1 and Tm5NM2 differ only in the use of exon 6a versus 6b, we conclude, on the basis of the LC MS/MS data for exon 6a from the TPM3 (γ) gene, that Tm5NM1 is present in aortic smooth muscle.

Band 4 was identified as the smooth muscle isoform of calponin (h1 calponin, basic calponin), which was not surprising since this protein has been described previously to be heat stable (52).

The faint band at ∼42 kDa (band 1), which we expected to contain Tm6 because of the fact that the 2a antibody recognizes a specific band of that size in ferret aorta lysates (see Fig. 2C, lane 1 “Input”), was mainly composed of actin. Under the Tm purification conditions used here, it appears that the protein recognized by the Tm6 antibody is less heat-soluble than the other isoforms and was precipitated in the pellet (Fig. 2C, lane 2) during the purification process. To test this hypothesis, we also examined the 42-kDa band of a total aorta lysate via LC MS/MS to see whether Tm6 is present in this band. The top match for this band was again actin, but peptides matching the transition of exons 1a/2a and 2a/3 from the TPM1 (α) gene were also detected (data not shown), indicating the presence of Tm6. This indicates that 1) most of the Tm6 is insoluble with the Bretscher (1) protocol used here, and 2) Tm6 seems to shift to 37 kDa during the purification process since we were able to identify Tm6 in band 2 (upper band of 37 kDa doublet) of the purified Tm sample and the anti-Tm6 antibody does detect this 37-kDa band in the purified Tm fraction (Fig. 2C, lane 3, “sup”). Others have also reported that this 37-kDa Tm6 protein often runs at 40–42 kDa, indicating a gel shift which appears to be seen in a variable manner under different conditions (47). The fact that we found peptides matching the transition of exon 2a/3 in this 42-kDa band furthermore rules out the possibility that the putative long form of Tm6 (containing both exon 2a and 2b) is present in aortic smooth muscle.

The Coomassie-stained gels that were used for LC MS/MS analysis suggested initially that the proteins from the 37-kDa doublet (band 2: Tm6 and Tm2; band 3: Tm1) exist in about equal amounts. However, since the vast majority of the Tm6 is not soluble by this protocol, in the native tissue, Tm6 may be in excess of Tm1. Thus we approached the question of the relative abundance of Tm6 by calibration with standard proteins below.

Quantification of Tm variant protein levels.

Tms from the α- and β-gene are thought to exist in a 1:1 ratio in smooth muscle, on the basis of results from chicken gizzard (36, 40, 44). To examine this hypothesis, we quantified the amount of Tm1 and Tm6 in aorta tissue by using purified Tms as a standard. We loaded different amounts of human recombinant Tm1 on a gel, together with total lysate of aorta tissue. After densitometric measurements of the bands detected by the anti-Tm1 (CG1) antibody, we calculated the total amount of Tm1 in aortic smooth muscle cells as ∼1.4% of total protein in aortic smooth muscle cells (see Fig. 3A).

Fig. 3.

Protein calibration of Tm1 and Tm6 in dVSM. A: 25 μg of ferret aorta homogenate were separated on a 10% SDS-gel together with recombinant human Tm1 (0.05 μg, 0.2 μg, 0.4 μg, 0.8 μg) and examined by Western blotting (WB), using the anti-Tm1 (CG1) antibody. The gel image was digitally overexposed after background subtraction for visual display. Densitometric measurement of the Tm1 standards revealed that the total amount of Tm1 in the loaded aorta sample was 0.35 μg. This corresponds to 1.4% Tm1 of total protein in ferret aorta. B: 12.25 μg of ferret aorta were loaded on a 10% SDS-gel in parallel with a Tm6 standard (0.375 μg, 0.75 μg, 1.5 μg), purified from chicken gizzard. Densitometric analysis of the Western blot, probed with the anti-Tm6 (α2a) antibody, indicated that the loaded aorta sample contained 1.4 μg Tm6. This corresponds to 11.6% Tm6 of total protein. C: alignment of the peptide used to generate the anti-Tm6 antibody (“2a”) and exons 2a from chicken or human Tm6.

For the Tm6 calibration, we used purified chicken gizzard Tm as a standard, since to the best of our knowledge, no recombinant human Tmsm-α/Tm6 is available. Following densitometric measurement and creating a calibration curve with the Tm standard, we calculated Tm6 in aortic smooth muscle cells as 11.6% of total protein (see Fig. 3B). This is likely an overestimate, probably because the anti-Tm6 antibody used in this assay was raised against a mammalian (rat) sequence and likely has a lower affinity to the chicken Tmsm-α than to the ferret Tm6 (see sequence alignment of the peptide used for generating the α2a antibody and exons 2a from either chicken or human [close to ferret]; Fig. 3C), therefore likely artifactually increasing the calculated Tm6 amount in the aorta tissue. Therefore, we cannot give a precise molar ratio of Tm1 and Tm6: however, it is clear that Tm6 is present at higher levels in this tissue than is Tm1.

As can be estimated from the Western blots shown in Fig. 1B, Tm isoforms Tm2, Tm5NM1, and Tm4 were far less abundant than Tm1 or Tm6.

α-Smooth muscle Tm6 and β-smooth muscle Tm1 bind different actin isoforms.

The considerable difference in the amount of α-smooth muscle Tm6 and β-smooth muscle Tm1 present in these samples at the protein level makes it unlikely that these variants form 1:1 heterodimers as has been described for chicken gizzard smooth muscle (19, 27, 43). Thus we immunoprecipitated Tm6 or Tm1 from aorta lysates. Probing the samples for actin isoforms revealed that Tm1 preferentially binds to γ-actin while Tm6 binds preferentially to β-actin, previously shown to have a similar distribution as dense bodies of contractile filament bundles (41). A typical Western blot is shown in Fig. 4A for an anti-Tm1 coimmunoprecipitation and in Fig. 4B for an anti-Tm6 coimmunoprecipitation. Densitometry of Western blots, derived from independent experiments, confirms that the amount of γ-actin in Tm1 coimmunoprecipitations was significantly higher than in Tm6 coimmunoprecipitations (Fig. 4C), whereas the β-actin amount coprecipitated with Tm6 was significantly higher than in Tm1 coimmunoprecipitations (Fig. 4D).

Fig. 4.

Coimmunoprecipitation analysis of actin isoforms with Tm1 and Tm6. A: aorta tissue homogenate was subjected to coimmunoprecipitation (IP) with the anti-Tm1 (CG1) antibody (lane 3). As a negative control, coimmunoprecipitation with an anti-green fluorescent protein (GFP) antibody was performed (lane 2). Two percent of total homogenate was loaded as Input control (lane 1). Precipitated proteins were separated via SDS-PAGE and transferred onto a membrane. Detection of Tm1, α-actin, β-actin, and γ-actin was performed using specific antibodies. The lines between the input and IP lanes indicate where empty lanes of the same gel were deleted. B: coimmunoprecipitation with ferret aorta homogenate using the anti-Tm6 (α2a) antibody (lane 3). The anti-GFP antibody was used as a negative control (lane 2). Two percent of the homogenate was loaded as Input (lane 1). Tm6, α-actin, β-actin, and γ-actin were detected on the membrane by specific antibodies (as indicated). The lines between the input and IP lanes indicate where empty lanes of the same gel were deleted. C: densitometric analysis of copurified γ-actin together with either Tm1 (right column, n = 7) or Tm6 (left column, n = 7). Shown on the y-axis is the relative γ-actin amount, normalized to Input. Furthermore, the γ-background signal from the GFP-control was subtracted from the γ-actin signal in the immunoprecipitation against either Tm1 or Tm6. To be able to compare different sets of immunoprecipitations, the relative copurified γ-actin amount was corrected for the amount of either Tm1 or Tm6 that was pulled down in each experiment. Ctr, GFP control. Note that the difference of coprecipitated γ-actin between Tm1 and Tm6 is highly statistically significant (**P = 0.002). D: graph showing copurified β-actin with either Tm1 (right column, n = 7) or Tm6 (left column, n = 6). The difference is highly statistically significant (**P = 0.004). E: coprecipitated α-actin with either Tm1 (right column, n = 7) or Tm6 (left column, n = 7).

α-Smooth muscle actin was also probed in these coimmunoprecipitation experiments, but we were unable to show a differential association with the Tm variants. The pulldown of α-actin with both Tm variants may indicate a connection of the α-actin filaments to both the nonmuscle γ- and β-actin networks although other interpretations are possible (Fig. 4E).

In summary, the immunoprecipitation experiments indicate that the two different Tm variants Tm1 and Tm6 tend to associate with different actin isoform populations in dVSMCs.

Isoform-specific localization of actin isoforms, Tm1, and Tm6 in dVSMCs.

Next, imaging studies on freshly dissociated cells from aorta were performed to demonstrate that the interactions shown in the coimmunoprecipitations are physiologically relevant and result in colocalization. We first used actin isoform-specific antibodies to locate the actin subpopulations. Previous studies in our lab have shown that γ-actin is the most dynamic population of actin in dVSMCs in response to an α-agonist stimulus (21). However, because of the lack of a specific antibody, the subcellular location of this dynamic subpopulation of γ-actin was not known.

In the present study we used a recently characterized specific cytoplasmic γ-actin antibody (4) for the first time in dVSMCs to determine the intracellular localization of this actin isoform. Deconvolution immunofluorescence microscopy revealed that, as shown in Fig. 5Aa, cytoplasmic γ-actin is mainly located cortical in dVSMCs. In contrast, α- and β-actin are predominately incorporated in stress fibers and short filaments, respectively, throughout the core of the cell (Fig. 5, Ab and Ac).

Fig. 5.

Immunofluorescence staining of Tm1, Tm6, and actin isoforms in dVSM cells (dVSMCs). A: freshly isolated dVSMCs from ferret aorta, attached to glass coverslips, were fixed and stained with specific antibodies against γ-actin (a), β-actin (b), or α-actin (c). Nuclei were visualized by 4,6-diamidino-2-phenylindole (DAPI). B: colabeling of Tm6 (α2a antibody, b) and α-actin (a) in dVSMCs. The nucleus was visualized using DAPI. A merged image is displayed in c. C: dVSMCs were colabeled for Tm6 (α2a antibody, b) and α-actin (a) in dVSMCs. The nucleus was visualized using DAPI. A merged image is displayed in c. C: dVSMCs were colabeled for Tm6 (α2a antibody, b) and β-actin (a). DAPI was used to stain nuclei. A merged image is shown in c. D: intracellular localization of Tm1 was determined in dVSMCs using the anti-Tm1 (CG1) antibody (b). Total F-actin was visualized by phalloidin (a), and colocalization is shown in c. The nucleus was stained with DAPI. E: colabeling of Tm6 with the α2a antibody (a) and Tm1 with the CG1 antibody (b). Colocalization is shown in c. Scale bar for all, 10 μm.

Since γ-actin is primarily associated with Tm1 and β-actin is primarily associated with Tm6 in coimmunoprecipitation experiments, we compared the subcellular location of these two Tm variants to that of the actin isoforms described above. We also wanted to examine if these actin binding proteins display a different intracellular localization. To visualize actin, the cells were either costained with α-actin or β-actin (for Tm6 labeling) or phalloidin (for Tm1 labeling, because of the incompatibility of the cytoplasmic γ-actin antibody with the conditions necessary for Tm1 staining). As can be seen in Fig. 5C, Tm6 partially colocalized with short bundles of β-actin (see arrows). Immunofluorescence staining of Tm6 also showed that the protein was found mainly at F-actin stress fibers, as shown by the colabeling with α-actin (Fig. 5B). In contrast, Tm1 colocalizes with cortical actin, and not with phalloidin-stained central, contractile actin filaments as can be seen in the merged image in Fig. 5D. This staining pattern is similar to the cytoplasmic γ-actin staining shown in Fig. 5Aa. The differences in localization of Tm6 and Tm1 were confirmed by colabeling for both proteins in the same cells (Fig. 5E). The colabeled cells demonstrated Tm6-containing central filaments and cortical Tm1 staining. An apparent insertion of the Tm6-containing filaments into the Tm1-containing cortical domain was also visible in some cases. Thus, these results are consistent with our conclusions from the above immunoprecipitation experiments.

DISCUSSION

In this study we have, for the first time, demonstrated that at least five different Tm variants are expressed at the protein level in one vascular cell type: Tm6 and Tm2 from the α-gene, Tm1 from the β-gene, Tm5NM1 from the γ-gene, and Tm4 from the δ-gene. Comparison with protein standards demonstrate that, in aorta, Tm6 is the most abundant form, followed by Tm1. Tm2, Tm4, and Tm5NM1 are far less abundant.

We have also demonstrated the unexpected finding that Tm6 and Tm1 can be found in different subcellular locations as detected by immunofluorescence staining of freshly isolated dVSMCs. Tm1 colocalizes with the cortical actin network, whereas Tm6 localizes at α-actin-containing contractile filament bundles in the cells. Thus, even though we are not ruling out the possibility of in vitro heterodimerization, it appears that, in this dVSM cell type, the majority of Tm6 and Tm1 do not form heterodimers. This is in contrast with the heterodimers of smooth muscle α- and β-isoforms that have been reported for gizzard smooth muscle (19, 43). It is known that the properties of visceral smooth muscle cells differ from those of vascular smooth muscle cells, as has been shown for actin isoforms (6, 14); thus, it is not surprising that a heterodimer of Tmsm-α/Tmsm-β is present in gizzard but not in aorta.

Owing to the lack of specific antibodies against Tm2, we were not able to determine the intracellular localization of this Tm isoform and cannot say whether Tm2 is located at the stress fibers, at cortical actin, or both. However, the low level of Tm2 compared with Tm6 and Tm1 and the preference of Tm2 to form homodimers in other tissues (8, 31) indicate that any heterodimerization is likely to be at no more than trace levels. Given that Tm proteins need to be of the same length to be able to dimerize (8), the possibility exists that the two short 248 aa long isoforms Tm5NM1 and Tm4 may form heterodimers in this cell type.

Another new finding from this study is that cytoplasmic γ-actin is located in a different cellular compartment, the cell cortex. This has not been previously reported for dVSM, and the current working model for smooth muscle actin compartmentalization comes from gizzard studies and does not include a cortical actin component (37). However, previous studies on fibroblasts showed that cytoplasmic γ-actin is mainly organized as a meshwork in cortical and lamellipodial structures (4). The cortical actin localization in dVSM is also of interest in comparison to our past studies on this same tissue type. We had previously shown that stimulation of dVSMCs with the α-agonist phenylephrine increases the ratio of F- to G-actin, indicating that contractility of dVSM is associated with changes in actin net polymerization. Furthermore, we showed that γ-actin is the most dynamic isoform in the cells since a cell-permeant γ-actin decoy peptide construct selectively inhibits phenylephrine-stimulated smooth muscle cell contraction and two-dimensional gels of differential centrifugation samples show greater changes in net polymerization of γ-actin (21). Thus, the finding that this dynamic γ-actin is localized in the cell cortex, an area of high actin dynamics in nonmuscle cells, is of considerable interest.

It had previously been found that an overexpression of γ-actin in mouse C2 myoblasts can drive Tm2, but not TM5NM1, out of α-actin-containing stress fibers. These results were interpreted to indicate a preferred association between actin and Tm isoforms (46). However, to the best of our knowledge, the current study is the first to clearly show preferred association of endogenous Tms with separate actin isoforms, both by imaging and biochemical immunoprecipitation. Tm has been described as a gatekeeper, regulating the access of actin to actin-binding proteins (11, 18). This concept is likely to apply to smooth muscle as well (10, 26). The fact that Tm1 associates predominately with γ-actin at the cell cortex, whereas Tm6 is located at α/β-actin stress fibers, points to a model where Tm variants, as gatekeepers of actin subpopulations, might differentially control the access of different actin-binding proteins, thereby also defining different functional cellular actin compartments.

The question remains as to how Tm variants are targeted to different intracellular locations. Tm variants might be targeted to different cellular compartments during the translation process via a signal sequence in the mRNA, also called “ZIP-code.” This mechanism was described previously for β-actin (50). Other studies, however, indicate that Tm isoforms do not possess a sorting signal that is responsible for their different intracellular localization. Cytochalasin D treatment, for example, disintegrates the isoform-specific Tm localization pattern in several cell types (3, 45). Therefore it is more likely that Tm proteins display a variant-specific intracellular localization via either their differential affinities for actin-binding proteins or actin isoforms (34). The latter is supported by previous studies, showing that overexpression of nonmuscle actin isoforms alters the organization of Tm variants (46). Moreover, a recent study suggests that the entire Tm molecule serves as the unit of sorting (35).

In summary, vascular smooth muscle not only express the “classical” Tmsm-α (Tm6) and Tmsm-β (Tm1) Tm proteins, but also at least three other Tm variants: Tm2 and the two short forms Tm5NM1 and Tm4. Our studies further indicate that Tm6 and Tm1 do not form heterodimers in this tissue, since Tm6 is more far abundant than Tm1 and moreover, Tm6 and Tm1 associate with different actin isoforms and are located in different intracellular locations. The multitude of different Tm variants in dVSM opens the possibility for complex regulation and organization of specific intracellular actin subpopulations and associated actin-binding proteins into separate functional cellular subcompartments.

GRANTS

This study was supported by National Heart, Lung, and Blood Institute Grants HL-80003, HL-31704, and HL-86655 (to K. G. Morgan) and Swiss National Foundation Grant 310030-125320 (to C. Chaponnier).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).