Translocation of telokin by cGMP signaling in smooth muscle cells

Satoshi Komatsu, Koji Miyazaki, Richard A. Tuft, Mitsuo Ikebe

Abstract

Telokin is an acidic protein with a sequence identical to the COOH-terminal domain of myosin light chain kinase (MLCK) produced by an alternate promoter of the MLCK gene. Although it is abundantly expressed in smooth muscle, its physiological function is not understood. In the present study, we attempted to clarify the function of telokin by analyzing its spatial and temporal localization in living single smooth muscle cells. Primary cultured smooth muscle cells were transfected with green fluorescent protein (GFP)-tagged telokin. The telokin-GFP localized mostly diffusely in cytosol. Stimulation with both sodium nitroprusside (SNP) and 8-bromo-cyclic GMP induced translocation of GFP-tagged telokin to near plasma membrane in living single smooth muscle cells. The translocation was slow, and it took more than 10 min at room temperature. Mutation of the phosphorylation sites of telokin (S13A, S19A, and S13A/S19A) significantly attenuated SNP-induced translocation. Both KT-5823 (cGMP-dependent protein kinase inhibitor) and PD-98059 (mitogen-activated protein kinase inhibitor) diminished the telokin-GFP translocation. These results suggest that telokin changes its intracellular localization because of phosphorylation at Ser13 and/or Ser19 via the cGMP-signaling pathway.

  • green fluorescent protein
  • myosin
  • myosin light chain kinase
  • guanosine 3′,5′-cyclic monophosphate-dependent protein kinase
  • phosphorylation

it is well known that smooth muscle contraction is initiated by an increase in cytoplasmic free calcium concentration ([Ca2+]i) induced by external stimuli and consequent phosphorylation of myosin regulatory light chain (MLC) by a Ca2+-calmodulin (CaM)-dependent MLC kinase (MLCK) (18, 51, 52).

Recently, evidence has accumulated that the level of myosin phosphorylation in smooth muscle is not solely dictated by [Ca2+]i, and it has been suggested that other signaling systems play a role in controlling MLC phosphorylation level. Recent studies revealed that MLC phosphatase (MLCP) activity changes upon agonist stimulation in smooth muscle, and this changes the myosin phosphorylation and smooth muscle contraction (50, 51).

MLCP consists of three subunits, a myosin-binding large subunit (MBS), a 20-kDa small subunit, and a catalytic subunit of the type 1 protein serine/threonine phosphatase family (1, 45, 46). It has been shown that MBS can be phosphorylated by Rho-dependent kinase, which results in a decrease in MLCP activity (29). Therefore, it has been postulated that an activation of the Rho signaling pathway would downregulate MLCP and thus increase contraction. It was also shown that the activation of protein kinase C (PKC) inhibits MLCP, thus inducing contraction (21, 37), suggesting that MLCP is regulated by a membrane-associated large G protein signaling pathway acting via activation of phospholipase C to produce diacylglycerol (DAG), an activator of PKC. Subsequently, it was suggested that the PKC-induced increase in contraction involves an activation of MLCP-specific inhibitor protein CPI17 by PKC-mediated phosphorylation (8, 9). On the other hand, arachidonic acid can directly interact with MLCP to induce the dissociation of the MBS from the MLCP holoenzyme, which results in a decrease in phosphatase activity (15).

Recently, it was reported that telokin accelerates dephosphorylation of MLC and induces relaxation of permeabilized smooth muscle strips (55). Although the mechanism underlying telokin-induced relaxation is not yet understood, this raised the possibility of a new regulatory mechanism of MLC phosphorylation in smooth muscle.

Telokin is an acidic protein with a sequence identical to the carboxyl-terminal domain of MLCK (26), and it is expressed abundantly in smooth muscle because of an alternate promoter mechanism in the MLCK gene (13, 19, 58). However, its function has been unclear, because telokin contains neither a kinase domain nor a CaM-binding domain of MLCK. Shirinsky et al. (47) reported that telokin stabilizes the filament formation of dephosphorylated myosin in the presence of Mg2+-ATP in vitro. Subsequently, it was found that the COOH-terminal acidic cluster is responsible for the stabilization of myosin filament formation (47). These findings raised the hypothesis that telokin functions as a thick filament stabilizing factor in smooth muscle, although the physiological significance of telokin in myosin filament formation is obscure because other proteins such as caldesmon (28) and a 38-kDa protein (41) also stabilize smooth muscle myosin filaments. Recently, Wu et al. (55) reported that exogenously added telokin accelerates dephosphorylation in Triton X-100-skinned smooth muscle strips. They also found that telokin is phosphorylated in intact muscle and α-toxin-skinned fiber in response to forskolin and 8-bromo-cGMP (8-BrcGMP), respectively. The findings raised a possibility that telokin plays a role in regulating the extent of MLC phosphorylation in smooth muscle due to a change in MLCP activity. However, the mechanism by which telokin can influence myosin phosphorylation level is unknown.

On the other hand, little is known about the localization of telokin in smooth muscle cells. This is at least partly due to the fact that telokin shares its entire amino acid residues with MLCK; thus telokin antibodies recognize both telokin and MLCK. This makes it difficult to distinguish telokin localization from that of MLCK by immunocytochemical techniques. MLCK can bind to actin in vitro (44) and localizes at thin filaments in vivo (7,16). However, the actin-binding region of MLCK at the NH2-terminal end of the molecule is not present in telokin (14, 57). Because the localization of proteins in cells can dictate the physiological function of proteins, one way to determine the function of telokin in smooth muscle cells is to clarify the spatial and temporal localization of telokin in cells.

In this study, we succeeded in visualizing the distribution of telokin in living smooth muscle cells by using green fluorescence protein (GFP)-chimeric telokin as a probe. Interestingly, we found that telokin translocates to the near plasma membrane region after activation of the cGMP signaling pathway in single living smooth muscle cells. The results provide further insight in an understanding of the physiological roles of telokin in smooth muscle.

MATERIALS AND METHODS

Materials.

Restriction enzymes and modifying enzymes were purchased from New England Biolabs (Beverly, MA). A polyclonal antibody against cGMP-dependent protein kinase (PKG)Iα, KT-5823, and PD-98059 were purchased from Calbiochem (San Diego, CA). Smooth muscle myosin II and MLCK were prepared from frozen turkey gizzards (22, 24).Xenopus oocyte calmodulin (3) was expressed inEscherichia coli and purified as described previously (23).

Preparation of smooth muscle cells and cell cultures.

Porcine tracheal smooth muscle cells and COS7 cells were used in the present study. COS7 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% FBS. Smooth muscle cell primary culture was performed as previously described (17) with a slight modification. Briefly, cleaned tracheal muscle was obtained by dissection and minced with scissors. Myocytes were enzymatically dispersed for 40 min at 37°C in Hanks' balanced salt solution containing 600 U/ml of collagenase, 10 U/ml of elastase, and 2 U/ml of protease (Nagarse type XXVII, Sigma). Porcine tracheal smooth muscle cells were maintained in DMEM-Ham's F-12 (1:1 vol/vol) medium supplemented with 10% FBS (GIBCO-BRL) and 0.1 mM nonessential amino acids (NEAAs). Cells from passage 3 were used in this study. Transfection of plasmids was carried out by electroporation with a Gene Pulser II (BioRad Laboratories). Transfected cells were seeded onto glass coverslips or culture dishes. To induce the differentiated phenotype, porcine tracheal smooth muscle cells were cultured in differentiation medium, which is serum-free DMEM-Ham's F-12 (1:1 vol/vol) medium supplemented with 0.1 mM NEAAs and insulin-transferrin-selenium medium (final concentrations: 5 μg/ml of insulin, 5 μg/ml of transferrin, and 5 ng/ml of selenium) (17).

Production of various telokin constructs.

For making the telokin-GFP chimeric construct, a cDNA fragment of MLCK subcloned into pT7-7 vector (56) was used as a template. To create the restriction sites for subcloning, two uniqueNheI and XhoI sites were created at the initiation codon of telokin and at the termination codon of MLCK, respectively, with site-directed mutagenesis strategy (56). Telokin and pEGFP-N1 mammalian expression vector (Invitrogen) were digested with NheI/XhoI and then in-frame ligated to produce telokin/pEGFP-N1. A linker, LELKLRILQSTVPRARDPPVAT, was introduced to segregate the GFP and telokin domains. It should be noted that the presence of the linker sequence did not affect the diffuse localization of control GFP. A baculovirus transfer vector for telokin in pFastBacHTa (GIBCO-BRL) was produced as follows. A unique NdeI site was created at the initiation codon of telokin and was excised withNdeI/EcoRI digestion and in-frame ligated to a pFastBacHTa baculovirus transfer vector containing a hexahistidine tag sequence with a unique NdeI site generated by site-directed mutagenesis strategy (56). Telokin mutants (S13A, S19A, and S13A/S19A) were produced by site-directed mutagenesis with telokin-GFP as a template.

Binding of GFP-chimeric telokin to smooth muscle myosin.

Telokin-GFP and control vector GFP were expressed in COS7 cells. After the transfection was confirmed under fluorescence microscopy, cells (1.0 × 106 cells/ml) were harvested and lysed by sonication with buffer A [in mM: 0.2Nα-p-tosyl-l-lysine chloromethyl ketone (TLCK), 0.2 N-tosyl-l-phenylalanine chloromethyl ketone (TPCK), 2 phenylmethylsulfonyl fluoride, 2 dithiothreitol (DTT), 50 NaCl, 0.1 EGTA, and 30 Tris · HCl, pH 7.0, with 10 μg/ml leupeptin and 10 μg/ml soybean trypsin inhibitor]. After sonication the samples were centrifuged at 10,000g for 10 min, and then the supernatant was dialyzed againstbuffer B (in mM: 5 NaCl, 2 DTT, 0.1 EGTA, 5 MgCl2, and 10 Tris · HCl, pH 7.0) with or without 3 mg/ml myosin filaments overnight at 4°C. After dialysis, the reaction solutions were centrifuged at 10,000 g for 10 min and then subjected to determination of telokin-GFP myosin binding capacity. SDS-polyacrylamide gel electrophoresis (PAGE) was carried out on a 7.5–20% polyacrylamide gradient slab gel with the discontinuous buffer system of Laemmli (32). Western blotting was carried out as described previously (20, 56) with the polyclonal antibody against GFP (MBL). Bound antibodies were detected by the enhanced chemiluminescence method (Amersham).

Binding of telokin to phosphorylated and dephosphorylated smooth muscle myosin.

To express His-tagged telokin, 80 ml of Sf9 cells (∼1.0 × 109) were infected with viruses expressing the telokin. The cells were cultured at 28°C in 175-cm2 flasks and harvested after 3 days. Cells were lysed with sonication in 5 ml of lysis buffer (in mM: 60 KCl, 2 MgCl2, 0.2 EGTA, 2 phenylmethylsulfonyl fluoride, 0.2 TLCK, 10 β-mercaptoethanol, 10 imidazole, and 30 Tris · HCl, pH 7.5, with 20 μg/ml leupeptin and 10 μg/ml soybean trypsin inhibitor). After centrifugation at 10,000 g for 20 min, the supernatant was applied to a nickel-nitrilotriacetic acid-agarose (Qiagen, Hilden, Germany) column (1 × 10 cm) and washed with 20-fold volume of buffer (0.3 M NaCl, 0.2 mM EGTA, 10 μg/ml leupeptin, 10 mM β-mercaptoethanol, 10 mM imidazole, and 30 mM Tris · HCl, pH 7.5). Telokin was eluted with buffer C (0.1 M NaCl, 0.2 mM EGTA, 10 μg/ml leupeptin, 10 mM β-mercaptoethanol, and 0.2 M imidazole, pH 7.5). After SDS-PAGE analysis, fractions containing telokin were pooled and dialyzed against (in mM) 50 NaCl, 0.2 EGTA, 2 MgCl2, 2 DTT, and 30 Tris · HCl, pH 7.5. Purified telokin was used in an in vitro cosedimentation assay. Telokin was phosphorylated by PKGIα (10,000 U/reaction; Calbiochem). Protein concentration was determined by the method of Bradford (2) using BSA as a standard. The reaction was carried out in buffer D [in mM: 30 Tris · HCl, pH 7.5, 1 MgCl2, 0.2 EGTA, and 0.1 [γ-32P]ATP (0.1 mCi/ml) with 5 μM cGMP] with 0.3 mg/ml of telokin substrate for 20 min at 30°C. Purified smooth muscle myosin filaments were phosphorylated by MLCK [in mM: 30 Tris · HCl, pH 7.5, 1 MgCl2, 0.2 CaCl, and 0.1 [γ-32P]ATP (0.05 mCi/ml) with 5 μM cGMP] for 20 min at 25°C. Either phosphorylated or unphosphorylated telokin and myosin filaments were incubated with 20 U/ml hexokinase, 20 mM glucose, and 10 mM MgCl2 at 25°C for 30 min to completely hydrolyze residual ATP and to filament myosin. After the reaction solutions were centrifuged at 10,000 g for 5 min, pellets were washed withbuffer E (in mM: 25 NaCl, 10 MgCl2, 2 DTT, and 30 Tris · HCl, pH 7.5). To determine phosphorylated telokin myosin binding capacity, samples were subjected to SDS-PAGE followed by autoradiography.

Immunofluorescence staining and image processing.

Immunocytochemistry was performed as described previously (31). Transfected cells were washed twice with PBS, placed in fixation solution I (4% formaldehyde, 2 mM MgCl2, and 1 mM EGTA in PBS), and, after extensive washing, permeabilized with 0.1% Triton X-100 in PBS for 10 min. After permeabilization in all cases, the coverslips were washed twice with PBS and incubated for 30 min with 1% BSA in PBS. GFP-tagged telokin was visualized with GFP fluorescence signal or with polyclonal anti-GFP antibody staining followed by FITC-conjugated anti-rabbit antibodies (Jackson Immuno Research Lab). The latter method was used to obtain a higher fluorescence signal because of the quenching of GFP signal during the fixation.

Cells transfected with the wild-type or mutant telokin-GFP were treated with 20 μM sodium nitroprusside (SNP) or 1 mM 8-BrcGMP in culture medium containing 25 mM HEPES (pH 7.2) for 18 min at room temperature and fixed as described above. In living cell images, the translocation of telokin-GFP was induced by SNP in the presence of 3-isobutyl-1-methylxanthine (IBMX), the phosphodiesterase inhibitor to prevent degradation of cGMP. To examine the effect of protein kinase inhibitors on telokin translocation, the cells were pretreated with 1 μM KT-5823 or 50 μM PD-98059 for 10 min at room temperature. After pretreatment with the inhibitors, 20 μM SNP or 1 mM 8-BrcGMP was added to the cells. After 18 min, the cells were fixed as described above and the cells showing an elongated tubelike shape were examined. The cells showing typical submembrane localization as shown in Fig.4 A,2 were only counted as “translocated cells,” and the number was divided by the total transfected cell number.

Images were observed with a Leica confocal microscope (Leica Microsystems). Differential interference contrast (DIC) and fluorescence images were viewed with a Leica DM IRB laser scanning confocal microscope controlled by Leica TCS SP II systems (Leica Microsystems). For three-dimensional reconstruction, a series of optical sections obtained by confocal microscope were collected at 0.2-μm intervals from the bottom to the top of the cell. The images were reconstructed with LCS 3D software from Leica Microsystems. Images were processed with Adobe Photoshop 5.5 software (Adobe Systems).

Translocation of telokin-GFP in living smooth muscle cells.

Transfected cells were identified by GFP signal with a fluorescence microscope, and images were recorded by imaging systems as described above. During the stimulation with external stimuli, digital images of the transfected cells were monitored in culture medium containing 25 mM HEPES (pH 7.2) at room temperature.

Statistics.

Data are expressed as means ± SE and were evaluated by one-way ANOVA. Statistical significance was determined by Fisher's protected least significant difference (PLSD) post hoc test.

RESULTS

Expression of GFP-tagged telokin and its myosin binding capacity in vitro.

To determine the localization of telokin in smooth muscle cells, a GFP-chimeric telokin expression vector was produced. The telokin derived from chicken gizzard MLCK cDNA was inserted into the pEGFP-N1 mammalian expression vector downstream of its cytomegalovirus (CMV) promoter as described in materials and methods. GFP was introduced at the COOH-terminal end of telokin, because the NH2-terminal end of telokin (Ser13 and Ser19) (36) is phosphorylated in vivo. A long, flexible linker sequence was placed between the two components to avoid an influence of the GFP moiety on telokin function. Because the COOH-terminal acidic region is responsible for the binding to smooth muscle myosin (48), we examined whether or not the GFP linked to the COOH-terminal side of telokin interferes with the normal binding activity of telokin to myosin.

COS7 cells were transfected with telokin-GFP expression vector as described in materials and methods. After the transfection was confirmed under fluorescence microscopy, the cells were harvested and the lysates were subjected to determination of the binding activity of telokin-GFP to myosin. To assess the binding activity of telokin-GFP to myosin, an in vitro cosedimentation assay was performed. Telokin-GFP was mixed with smooth muscle myosin at a low ionic strength to form thick filaments, and the mixture was centrifuged to precipitate the myosin filaments (see materials and methods). As shown in Fig. 1, expressed telokin-GFP chimeric proteins were detected as a single band showing a molecular mass of 47 kDa, which is consistent with the sum of the molecular mass of chimeric protein composed of telokin (20 kDa) and GFP (27 kDa). Although control GFP was not precipitated with myosin filaments (Fig. 1, lane 5), expressed telokin-GFP was precipitated with myosin filaments (Fig. 1, lane 3) but not without myosin filaments (Fig. 1,lane 1). The binding activity was relatively weak, and approximately one-third of telokin-GFP coprecipitated with myosin filaments. A similar result was reported for the natural isolated telokin (47). The affinity of telokin for myosin is relatively weak (47), and the result is consistent with earlier findings with natural telokin. Because the region responsible for the binding of telokin to myosin is the COOH-terminal end of the molecule (48), the result also suggests that the addition of GFP to the COOH-terminal end of telokin does not alter the authenticity of telokin.

Fig. 1.

Association of telokin-green fluorescent protein (GFP) with myosin filaments. To assess the binding capacity of telokin-GFP to myosin, an in vitro cosedimentation assay was performed. COS7 cells were transiently transfected with telokin-GFP expression vector. Mixtures of telokin-GFP- or vector GFP-transfected cell lysates with (+) or without (−) myosin filaments were centrifuged, and then the pellets (lanes 1, 3, and 5; P) and the supernatants (lanes 2, 4, and 6; S) were analyzed by Western blot analysis with anti-GFP antibody.

Translocation of telokin-GFP in living single smooth muscle cells.

Localization of telokin-GFP in living single smooth muscle cells was monitored. It was previously reported that expression levels of telokin vary between different smooth muscle tissues and that telokin is expressed at relatively high levels in tracheal smooth muscle tissue (13, 19). Therefore, tracheal smooth muscle cells were used for the experiment. Of particular interest is the effect of cGMP signaling in telokin localization, because it has been shown that telokin is phosphorylated by PKG (cGMP-dependent protein kinase) in smooth muscle fiber, which might influence the relaxation of smooth muscle (55). Previously, it was reported that the expression level of PKG in cultured smooth muscle cells decreased with successive passages, decreasing to 60% and 10% of the original level at passages 4 and 6, respectively (5). Therefore, we used the cells within three passages to minimize loss of PKG expression. The expression level of PKG in cultured tracheal cells was ∼60% of the fresh tracheal tissues based on Western blot analysis with specific anti-PKG antibodies (not shown).

Telokin-GFP-transfected cells were stimulated with a relaxant agonist, the nitric oxide (NO) donor SNP. As shown in Fig. 2,A–G, telokin-GFP translocated to the near membrane with a prolonged time course after 20 μM SNP stimulation, but the localization of control GFP remained unchanged (Fig. 2, H and I). Fluorescence distribution of the transverse section of the images is also shown in Fig. 2. The data clearly demonstrated that the fluorescence intensity near plasma membrane increased significantly after SNP stimulation. Because SNP, a donor of NO, activates soluble guanylate cyclase, thus increasing the cellular levels of cGMP, the question is whether the translocation of telokin-GFP is due to the increase in cGMP in cells. A membrane-permeant cGMP analog, 8-BrcGMP, was added to the tracheal cells expressing telokin-GFP. Translocation of telokin-GFP to the near membrane was observed in response to 1 mM 8-BrcGMP (Fig.3, A–G) with a time course similar to the stimulation with SNP (Fig. 2,A–G). Fluorescence distribution of the transverse section of the images demonstrated a marked increase in fluorescence near plasma membrane after 8-BrcGMP stimulation. It should be noted that although the expression level of telokin-GFP varies in transiently transfected tracheal cells, the translocation of telokin-GFP was observed irrespective of the telokin-GFP expression level; the cells expressing a low level of telokin-GFP showed good translocation. Therefore, it is unlikely that the observed translocation is due to the potential binding of overexpressed protein to the nonspecific low-affinity sites in the cells. It should also be mentioned that the apparent total fluorescence level decreased with prolonged exposure time because of photobleaching (Figs. 2 and 3), and this did not occur when continuous exposure was avoided (not shown).

Fig. 2.

Translocation of telokin-GFP in primary cultured porcine tracheal smooth muscle cells stimulated with sodium nitroprusside (SNP). Porcine tracheal smooth muscle cells were transiently transfected with telokin-GFP expression vector. After the transfection, the cells were cultured in differentiation medium for 48 h and then stimulated with 20 μM SNP and 500 μM 3-isobutyl-1-methylxanthine (IBMX). Cells were viewed by confocal microscopy. The focal plane is close to the middle of the cell acquired at the times indicated at bottom leftin each panel. Top: time course of the GFP-telokin localization. A–G: signals of telokin-GFP. Hand I: signals of vector GFP. Translocation of telokin-GFP to the near membrane was observed with a prolonged time course after SNP stimulation (arrows), whereas the localization of control vector GFP was unchanged. Bars, 10 μm. Bottom: fluorescence distribution of the transverse section of the images before (A) and 18 min after (G) stimulation.

Fig. 3.

Translocation of telokin-GFP under the influence of 8-bromo-cyclic GMP (8-BrcGMP). Primary cultured porcine tracheal smooth muscle cells were transiently transfected with telokin-GFP expression vector. After being cultured in differentiation medium for 48 h, the cells transfected with telokin-GFP were treated with 1 mM 8-BrcGMP. Cells were viewed by confocal microscopy. The focal plane is close to the middle of the cell acquired at the times indicated at bottom left of panels. Top: time course of the GFP-telokin localization. A–G: signal of telokin-GFP. Hand I: signal of vector GFP. a–g: differential interference contrast (DIC) images obtained from same cells in A–G, respectively. A change in localization of telokin-GFP to the near membrane was observed after treatment with 8-BrcGMP with a prolonged time course (arrows). Bars, 6 μm.Bottom: fluorescence distribution of the transverse section of the images before (A) and 18 min after (G) stimulation.

The number of the cells showing translocation of telokin-GFP was counted (Fig. 4). GFP-telokin showed diffuse cytosolic localization in a majority of cells before stimulation, and the number of the cells showing submembrane localization of telokin-GFP significantly increased after SNP or 8-BrcGMP stimulation. The cells transfected with GFP alone showed a diffuse cytosolic GFP signal regardless of the stimulation (not shown).

Fig. 4.

Enhancement of telokin-GFP translocation by cGMP agonists. Porcine tracheal cells were transfected with telokin-GFP and visualized as described in Figs. 2 and 3. A: localization of telokin GFP wild type before (1) and 18 min after (2) SNP treatment. Bar, 20 μm. B: increase in the number of the cells showing submembrane localization of telokin-GFP. The number of the cells that show the fluorescence intensity of the cell peripheral region is more than 2 times the average intensity in the line plot (A,2); they were counted as “translocated cells,” and the number was divided by the total transfected cell number. The values were compared with those of nonstimulated cells. Values shown are means ± SE from 3 independent experiments and were evaluated by 1-way ANOVA. Statistical significance was then determined by Fisher's protected least significant difference (PLSD) post hoc test. * P < 0.002 vs. control by ANOVA. A majority of cells showed diffuse cytosolic localization of telokin-GFP before treatment (A,1), and the cells showing submembrane localization (A,2) were significantly increased by cGMP agonists. n, No. of cells.

Figure 5 shows a three-dimensional reconstruction image of telokin-GFP localized to the near membrane after stimulation with SNP. On the other hand, control GFP did not show a particular localization in cell and had a diffuse cytosolic distribution (Fig. 5 B). Similar three-dimensional localization was obtained with the cells treated with 8-BrcGMP (not shown). These results suggest that telokin changes its cellular distribution after stimulation affecting the cGMP signaling pathway.

Fig. 5.

Three-dimensional reconstruction of telokin-GFP. Telokin-GFP-transfected porcine tracheal smooth muscle cells were stimulated with 20 μM SNP. After 18 min of SNP stimulation, cells were fixed as described in materials and methods. Cells were viewed by confocal microscope. Series of optical sections were collected at 0.2-μm intervals from the bottom to the top of the cell and reconstructed with software from Leica Microsystems. A: telokin-GFP. Inset: telokin-GFP before treatment with SNP.B: GFP alone.

Effect of phosphorylation of telokin on its translocation.

It was shown previously that PKG phosphorylates telokin at Ser13 (36) and 8-BrcGMP increases telokin phosphorylation at Ser19 in smooth muscle fibers (55). Therefore, it is plausible that the phosphorylation at these sites is involved in the translocation mechanism of telokin. To address this notion, we produced telokin-GFP mutants in which Ser13/Ser19 is substituted by Ala. These constructs were expressed in tracheal smooth muscle cells, and the translocation of telokin-GFP after SNP stimulation was monitored (Fig.6 A). The translocation was markedly diminished by the mutation of Ser13/Ser19 of telokin. The result clearly indicates that phosphorylation at Ser13/Ser19 is critical for the translocation of telokin, which is presumably mediated by PKG signaling.

Fig. 6.

Effect of kinase inhibitors and phosphorylation site mutations on the translocation of telokin-GFP. Primary culture porcine tracheal smooth muscle cells were transiently transfected with various telokin-GFP constructs. After culture in differentiation medium for 48 h, the cells were treated with 20 μM SNP or 1 mM 8-BrcGMP. For inhibitor treatment, cells were preexposed to 1 μM KT-5823 or 50 μM PD-98059 for 10 min at room temperature and then the stimulants were added to the medium. After 18 min, the cells were fixed and observed under the fluorescence microscope. Cells were viewed by confocal microscope. A: effect of mutation of the phosphorylation sites on the translocation of telokin-GFP. Values shown are means ± SE from 3 independent experiments. AA, S13A/S19A double mutant. B: effect of inhibitors on the translocation of telokin-GFP. Transfected cells with telokin-GFP in the presence and absence of the inhibitors were examined under the fluorescence microscope. The number of cells that show the fluorescence intensity of the cell peripheral region is more than 2 times the average intensity in the line plot (see Fig. 4 A,2); they were counted as translocated cells, and the number was divided by the total transfected cell number. Values shown are means ± SE from 3 independent experiments. Statistical analysis was done by 1-way ANOVA. Statistical significance was then determined by Fisher's PLSD post hoc test. * P < 0.002 vs. control by ANOVA.n, No. of cells.

To further confirm the involvement of PKG in the telokin translocation mechanism, tracheal cells were treated with KT-5823, a PKG-specific inhibitor, before SNP or 8-BrcGMP stimulation. As shown in Fig.6 B, KT-5823 significantly attenuated the translocation of telokin-GFP, indicating that PKG plays a role in telokin translocation. Interestingly, PD-98059, a mitogen-activated protein kinase (MAPK) inhibitor, also attenuated the telokin-GFP translocation, suggesting the involvement of MAPK in telokin translocation. These results strongly suggest that PKG-induced phosphorylation of telokin triggers the translocation of telokin to a near membrane structure. It should be noted that these observations were obtained with cells at passage 3 but not with cells beyond passage 5, and PKG level gradually diminished with increased passage of the cells (not shown). The cells showed that the translocation of telokin-GFP was ∼30% of the transfected cells. This is not very surprising because of the heterogeneity of the primary cultured cells. However, because the phosphorylation of telokin, presumably by a PKG pathway, is involved in the translocation, it is reasonable to assume that PKG expression level in the cells is the key factor for telokin translocation. This view is consistent with the fact that the translocation was not observed with cells beyond passage 5 and the expression level of PKG diminishes with increased passage of the cells. The expression level of PKG in the cultured cells (<3 passages) was examined by Western blot of the cultured cells and compared with the fresh tracheal tissues. The expression level was normalized by the total protein. The expression level of PKG in the cultured smooth muscle cells was ∼60% of that of the fresh tracheal tissues (not shown).

We further examined the expression level of PKG in each cell by antibody staining. The cells were stained with anti-PKG antibodies, and the signal intensity was analyzed by digital fluorescence microscopy. The results showed that the PKG expression level of the cultured cells was 93–38% of the level of the fresh isolates. Approximately 30% of the cultured cells expressed PKG above 70% of that of fresh isolates (not shown). Therefore, it is plausible that the cells that expressed relatively high levels of PKG showed the translocation of telokin.

Effect of phosphorylation of telokin by PKGIα on its myosin binding activity.

Although the biochemical function of telokin is not understood, it is known that telokin can bind to smooth muscle myosin that stabilizes myosin filament formation in vitro (47). We examined whether PKG-induced phosphorylation affects the binding of telokin to myosin. Telokin was first phosphorylated by PKGIα with [γ-32P]ATP and then mixed with either phosphorylated or dephosphorylated smooth muscle myosin filaments. The mixture was then centrifuged to precipitate myosin filaments, and the bound telokin was estimated by SDS-PAGE analysis. Telokin phosphorylated by PKGIα bound to unphosphorylated myosin (Fig. 7,lane 8) but only poorly to phosphorylated myosin (Fig. 7,lane 12). On the other hand, phosphorylation of telokin did not dramatically alter its binding activity to myosin (Fig. 7,lanes 4 and 8, respectively). This result clearly indicates that phosphorylation of telokin by PKG had no significant effect on its myosin binding activity.

Fig. 7.

Effect of cGMP-dependent protein kinase (PKG) on the cosedimentation of telokin and myosin. A: recombinant telokin was phosphorylated by PKGIα and incubated with or without either phosphorylated or unphosphorylated myosin filaments by myosin light chain kinase (MLCK). Mixtures were centrifuged and after the pellet was washed, the supernatants (lanes 1, 3,5, 7, 9, and 11; S) and the pellets (lanes 2, 4, 6, 8,10, and 12; P) were subjected to SDS-polyacrylamide gel electrophoresis (top) followed by autoradiography (bottom). Lanes 5–8,11, and 12 were telokin phosphorylated by PKGIα. LC20, regulatory light chain; LC17, essential light chain; P-LC20, phosphorylated LC20. B: amount of cosedimented telokin with myosin was quantitatively determined by scanning densitometry (NIH Image program). The densities of cosedimented telokin with myosin were normalized by those of total telokin bands. Values shown are means ± SE from 3 independent experiments. Statistical analysis was done by 1-way ANOVA. Statistical significance was then determined by Fisher's PLSD post hoc test. * P < 0.05 vs. control by ANOVA. CaM, calmodulin.

DISCUSSION

It has been a while since telokin was first identified in smooth muscle (6), yet its physiological function is still unknown. This is in part because telokin does not have functional or structural motifs in its primary structure. Another problem is that there is little information about its cellular localization that can suggest the physiological relevance of this protein. The lack of information on telokin localization results from the fact that the telokin sequence is entirely a subsequence of MLCK, and the lack of a unique telokin sequence makes it difficult to produce specific antibodies as localization probes. In the present study, we used GFP-tagged telokin as a probe to determine localization in smooth muscle cells. GFP moiety was placed at the COOH-terminal end of the telokin molecule to avoid the effect of phosphorylation of telokin on its localization in cells, because it has been shown that the NH2-terminal end of telokin can be phosphorylated in vitro and in vivo (36, 55). The COOH-terminal end of telokin is involved in the binding of telokin to myosin in vitro (48). To avoid the obstruction of this binding by COOH-terminal GFP, we introduced a long linker sequence between the two moieties. In fact, telokin-GFP retained myosin binding activity in vitro. This result suggests that telokin-GFP retains the authentic properties of telokin.

The major findings of this study are that telokin translocates to near the plasma membrane in smooth muscle cells and that this is triggered by the activation of cGMP signaling pathway. Both SNP, a donor of NO that activates guanylate cyclase, and 8-brcGMP, a cell-permeant, metabolically resistant cGMP analog, initiated translocation of telokin-GFP. Because telokin can be phosphorylated by PKG, the results suggest that PKG is involved in the telokin translocation mechanism. To this end, we mutated the phosphorylatable residues of telokin and examined the effect of mutation on the translocation. The result clearly indicated that the phosphorylation of Ser13/Ser19 is critical for cGMP signaling-dependent translocation. It was reported recently that although Ser13 is the site phosphorylated by PKG in vitro (36), Ser19 in addition to Ser13 is phosphorylated in smooth muscle fibers in response to 8-BrcGMP (55). The present result is consistent with these reports. Although the phosphorylation at Ser19 is enhanced by cGMP agonist in fiber, this site is not a PKG consensus site (55). Consistently, PKG does not phosphorylate Ser19 in vitro (36). Therefore, one question is which protein kinase phosphorylates Ser19. To answer this question, we tested the effect of protein kinase inhibitors on telokin translocation. Both KT-5823, a PKG inhibitor, and PD-98059, a MAPK inhibitor, abolished the translocation of telokin-GFP, suggesting that both PKG and MAPK are involved in the telokin translocation mechanism. It was shown previously that cGMP and PKG activate MAPK in vascular smooth muscle cells (30). Furthermore, the sequence near Ser19 constitutes a consensus MAPK phosphorylation site rather than a PKG site. With the present findings, we conclude that MAPK, which is activated by PKG signaling, and PKG phosphorylate telokin, thus triggering the translocation to near plasma membrane in smooth muscle cells.

cGMP signaling induces relaxation of smooth muscle (10, 27, 39,42). This is at least partly due to a change in [Ca2+]i. Nitroprusside decreases inositol 1,4,5-trisphosphate (IP3) production (33) and L-type Ca2+ channel conductance (4, 25). It has been also suggested that cGMP enhances Ca2+ extrusion by activating the sarcoplasmic reticulum (SR) Ca2+ pump by phosphorylating phospholamban (35, 53), increasing the activity of the plasma membrane Ca2+-ATPase (12,43) and Na+/Ca2+ exchanger (11). Recently, it was shown that cGMP signaling also directly modulates the smooth muscle contractile apparatus. With clamped [Ca2+] using permeabilized strips, it was shown that cGMP decreases force with constant [Ca2+] and this is due to the increase in MLCP activity (34). The mechanism by which cGMP decreases force by activation of MLCP is not clear as yet. On the other hand, telokin is a candidate that is involved in cGMP-dependent relaxation of smooth muscle. It was reported that telokin accelerates the relaxation of Triton X-100-skinned smooth muscle strips and that cGMP potentiates this effect (55). While this manuscript was in preparation, it was reported that telokin phosphorylation at Ser13/Ser19 potentiates smooth muscle relaxation (36), although the mechanism by which telokin and telokin phosphorylation induce relaxation is unclear. The present study clarifies that the telokin phosphorylation at Ser13/Ser19 triggers translocation of telokin to near plasma membrane. The translocation kinetics of telokin-GFP was relatively slow at 25°C. Because the phosphorylation at Ser13 and/or Ser19 would be responsible for the translocation, it is plausible that the phosphorylation of telokin after activation of cGMP signaling is a slow process. The telokin translocation kinetics found in the present study is consistent with the time course of the telokin-induced potentiation of the relaxation of smooth muscle strips as well as PKG-induced phosphorylation of telokin (55).

Although the present results clearly showed that the activation of cGMP signaling results in the translocation of telokin-GFP, not all of the cells showed significant translocation. We think that this is related to the expression level of PKG in each cell because 1) the translocation was not observed with cells over passage 5, and the expression level of PKG diminishes with increased passage of the cells; 2) even the cells at less than three passages showed an expression level of PKG significantly lower than that of fresh isolates (mean value of ∼60% of the fresh isolates);3) the PKG expression level of the cultured cells varies (93–38% of the level of the fresh isolates) and the cells expressing PKG above 70% of the fresh isolates were ∼30%.

Although the functional role of telokin translocation is unclear, it is plausible that cGMP/PKG-induced translocation of telokin is related to the relaxation of smooth muscle because the same phosphorylation sites are responsible for both events. There are several possibilities to account for telokin-induced relaxation. First, telokin may directly inhibit myosin motor activity. However, this is less likely because addition of telokin hardly inhibited in vitro motility activity of smooth muscle myosin (not shown). Second, telokin may inhibit MLCK, and it was reported that telokin could inhibit myosin phosphorylation by MLCK in vitro (40, 49). Third, telokin may activate MLCP activity, thus decreasing myosin phosphorylation. However, there has been no evidence for the activation of MLCP by telokin. Fourth, telokin may interact with the thin filament binding proteins and may change actomyosin contractile activity. Interestingly, it was found that calponin in smooth muscle translocates to near plasma membrane after stimulation (38). The translocation of these two proteins to near membrane might be related to each other.

How the phosphorylation of telokin translocates telokin is unclear, but it is plausible that the “target” proteins are present at near membrane. One such candidate is myosin, because telokin binds to myosin in vitro (47). However, this is less likely because the phosphorylation of telokin does not change the binding activity to phosphorylated myosin or dephosphorylated myosin (Fig. 7). Quite recently, it was reported (54) that the truncation of the COOH-terminal domain of telokin that is responsible for the myosin binding of telokin failed to eliminate the relaxation effect of telokin in skinned smooth muscle strips, suggesting that the myosin binding activity of telokin does not affect its relaxation activity. Further studies are required to clarify the translocation mechanism of telokin.

Acknowledgments

This work is supported by National Institutes of Health Grants HL-60831, HL-61426, and AR-41653.

Footnotes

  • Address for reprint requests and other correspondence: M. Ikebe, Dept. of Physiology, Univ. of Massachusetts Medical School, Worcester, MA 01655 (E-mail:mitsuo.ikebe{at}umassmed.edu).

  • 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.

  • May 8, 2002;10.1152/ajpcell.00501.2001

REFERENCES

View Abstract