Phosphorylation of the myosin phosphatase targeting subunit and CPI-17 during Ca²⁺ sensitization in rabbit smooth muscle

Abstract

Myosin phosphatase (MLCP) plays a critical regulatory role in the Ca²⁺ sensitivity of myosin phosphorylation and smooth muscle contraction. It has been suggested that phosphorylation at Thr⁶⁹⁵ of the MLCP regulatory subunit (MYPT1) and at Thr³⁸ of the MLCP inhibitor protein CPI-17 results in inhibition of MLCP activity. We have previously demonstrated that CPI-17 Thr³⁸ phosphorylation plays an important role in G-protein-mediated inhibition of MLCP in tonic arterial smooth muscle. Here, we attempted to evaluate the function of MYPT1 in phasic rabbit portal vein (PV) and vas deferens (VD) smooth muscles. Using site- and phospho-specific antibodies, phosphorylation of MYPT1 Thr⁶⁹⁵ and CPI-17 Thr³⁸ was examined along with MYPT1 Thr⁸⁵⁰, which is a non-inhibitory Rho-kinase site. We found that both CPI-17 Thr³⁸ and MYPT1 Thr⁸⁵⁰ were phosphorylated in response to agonists or GTPγS concurrently with contraction and myosin phosphorylation in α-toxin-permeabilized PV tissues. In contrast, phosphorylation of MYPT1 Thr⁶⁹⁵ did not increase. Comparable results were also obtained in both permeabilized and intact VD. The Rho-kinase inhibitor Y-27632 and the protein kinase C (PKC) inhibitor GF109203X suppressed phosphorylation of MYPT1 Thr⁸⁵⁰ and CPI-17 Thr³⁸, respectively, in intact VD while MYPT1 Thr⁶⁹⁵ phosphorylation was insensitive to both inhibitors. These results indicate that phosphorylation of MYPT1 Thr⁶⁹⁵ is independent of stimulation of G-proteins, Rho-kinase or PKC. In the phasic PV, phosphorylation of CPI-17 Thr³⁸ may contribute towards inhibition of MLCP while the phasic visceral VD, which has a low CPI-17 concentration, probably utilizes other Ca²⁺ sensitizing mechanisms for inhibiting MLCP besides phosphorylation of MYPT1 and CPI-17.

Smooth muscle contraction is mainly regulated by phosphorylation of myosin regulatory light chain (MLC), which is determined by the opposing actions of myosin light chain kinase (MLCK) and myosin light chain phosphatase (MLCP) (Hartshorne, 1987; Kamm & Stull, 1989). MLCK activity depends on Ca²⁺-calmodulin while MLCP functions independently of Ca²⁺ and is regulated by G-protein-coupled signalling pathways (Somlyo & Somlyo, 1994). MLCP inhibition leads to an increase in both MLC phosphorylation and contractile force of smooth muscle without any changes in Ca²⁺, hence causing an enhancement of the contractile Ca²⁺ sensitivity (Ca²⁺ sensitization; Kitazawa et al. 1991a,b; Kubota et al. 1992). This is an important physiological process in agonist-induced contraction of smooth muscle and also for cytoskeletal reorganization and movement of non-muscle cells (Somlyo & Somlyo, 2000).

MLCP is a holoenzyme composed of three subunits: a 38 kDa catalytic subunit of type 1 protein phosphatase (PP1c), a large 110–130 kDa regulatory subunit (MYPT1) and a small 20 kDa subunit of unknown function (Hartshorne et al. 1998). MYPT1 is the key subunit involved in binding to and activation of PP1c and in targeting myosin. RhoA, a Ras-related monomeric small GTPase, is thought to play a major role in the G-protein-coupled Ca²⁺ sensitization of smooth muscle contraction (Hirata et al. 1992; Somlyo & Somlyo, 2000). Several proteins have been identified as targets of GTP-bound RhoA, such as RhoA-activated kinase (Rho-kinase/ ROKα/ROCK-II) and MYPT1 (Kimura et al. 1996). Purified Rho-kinase phosphorylates the MYPT1 subunit mainly at Thr⁶⁹⁵ and Thr⁸⁵⁰ (in the chicken 133 kDa MYPT1 sequence; Feng et al. 1999a,b). The former phosphorylation suppresses the activity of the MLCP holoenzyme. Because agonist- and GTPγS-induced contractions at given submaximal Ca²⁺ concentrations are inhibited by treatment with Y-27632, a Rho-kinase specific inhibitor, it is thought that phosphorylation of MYPT1 at Thr⁶⁹⁵ by Rho-kinase evokes the contractile Ca²⁺ sensitization of smooth muscle by inhibiting MLCP activity (Uehata et al. 1997; Somlyo & Somlyo, 2000; Pfitzer, 2001; Fukata et al. 2001). In addition to Rho-kinase, purified kinases such as zipper-interacting protein kinase (ZIPK; MacDonald et al. 2001), integrin-linked kinase (ILK; Murányi et al. 2002) and myotonic dystrophy kinase (DMPK; Murányi et al. 2001) can phosphorylate MYPT1 Thr⁶⁹⁵. This has provoked controversy regarding the physiological MYPT1 kinase. The phosphorylation of Thr⁶⁹⁵ increases in response to lysophosphatidic acid in Swiss 3T3 cells (Feng et al. 1999a). In smooth muscle tissues, total phosphorylation of MYPT1 can be enhanced by stimulation with the G-protein activator GTPγS (Trinkle-Mulcahy et al. 1995; Swärd et al. 2000; Nagumo et al. 2000); however, the specific phosphorylation of MYPT1 Thr⁶⁹⁵ has not been investigated.

In addition to phosphorylation of MYPT1 Thr⁶⁹⁵, a second MLCP inhibitory pathway involves a PKC-potentiated phosphatase inhibitor protein-17 kDa (CPI-17; Eto et al. 1995, 1997). Phosphorylation of CPI-17 at Thr³⁸ enhances its negative effect on purified PP1c and MLCP holoenzyme activity 1000-fold (Eto et al. 1997; Senba et al. 1999) and on in situ MLCP anchored to myofibrils (Li et al. 1998; Eto et al. 2000). Selective depletion of CPI-17 by skinning of smooth muscle cells eliminates PKC-induced Ca²⁺ sensitization of femoral artery strips and the response can be reconstituted by addition of PKC and CPI-17 together but not by PKC alone (Kitazawa et al. 1999). Stimulation of arterial smooth muscle with agonists, GTPγS, or the PKC activator phorbol ester induces phosphorylation of CPI-17 Thr³⁸ paralleling the contractile Ca²⁺ sensitization (Kitazawa et al. 2000). Phosphorylation of CPI-17 Thr³⁸ is partially suppressed by GF109203X (a PKC inhibitor) or Y-27632 (Kitazawa et al. 2000). The PKC delta isoform, isolated as a dominant CPI-17 kinase in pig aorta, is inhibited by both GF109203X and Y-27632, suggesting that CPI-17 functions in PKC-mediated regulation of MLCP (Eto et al. 2001). The expression level of CPI-17 is estimated to be 6.7 µm in rabbit femoral artery, 4.4 µm in portal vein, and 0.8 µm in vas deferens (Woodsome et al. 2001). In tonic arterial smooth muscle, which possesses a high CPI-17 expression, PKC is a dominant contributor towards Ca²⁺ sensitivity of contraction and MLC phosphorylation. These results imply that, in tonic vascular smooth muscle, the G-protein-mediated inhibition of MLCP is mainly through the PKC-CPI-17 signalling pathway. In contrast, PKC makes a minor contribution in phasic smooth muscles, such as vas deferens, that have low CPI-17 and high MLCP content. This indicates that a mechanism(s) other than CPI-17-MLCP exists for the G-protein-mediated inhibition of MLCP in phasic visceral smooth muscle (Woodsome et al. 2001).

The purpose of this study was to evaluate the physiological signalling pathways leading to MLCP inhibition. Using site- and phospho-specific MYPT1 Thr⁶⁹⁵ and CPI-17 Thr³⁸ antibodies we examined whether agonists and GTPγS induce phosphorylation of MYPT1 Thr⁶⁹⁵ or CPI-17 Thr³⁸ in parallel with MLC phosphorylation and contraction in intact and α-toxin-permeabilized phasic smooth muscles. We found that phosphorylation of both CPI-17 at Thr³⁸ and MYPT1 at Thr⁸⁵⁰ increased in response to stimulations by agonists and GTPγS while phosphorylation of MYPT1 at Thr⁶⁹⁵ was not significantly elevated. A portion of these findings has been presented at the Annual Biophysical Society Meeting (Kitazawa et al. 2002).

Methods

Tissue preparation and force measurement

All animal procedures were approved by the Animal Care and Use Committee of Georgetown University. White male albino rabbits (2–3 kg) were killed with an overdose of halothane and the desired smooth muscle tissues were dissected as described previously (Kitazawa et al. 1991a). Smooth muscles from the portal vein (PV) and femoral artery (FA) were then cut into strips (750 µm wide and 2.5–3 mm long with natural wall thickness) and freed of connective tissue. The vascular endothelium was removed by gentle rubbing with a razor blade. The vas deferens (VD) strips, without epithelium, lamina propria and connective tissues, were 250 µm in diameter and 3 mm long. The strips were mounted on a force transducer (AM801, SensoNor, Horten, Norway) and then immersed in a temperature-controlled bubble plate to allow rapid solution exchange and freezing for phosphorylation measurements (Masuo et al. 1994). Prior to experimentation, the strips were stimulated several times with a high K⁺ (154 mm) solution at 30 °C until a steady maximal response was obtained. Force levels were monitored throughout the experiments. Experiments using intact strips were carried out at 30 °C.

Solutions and permeabilization

The compositions of external and intracellular solutions have been described previously (Woodsome et al. 2001). For cell permeabilization, the strips were treated for 30 min at 30 ° C with 20 µg ml⁻¹ of Staphylococcus aureus α-toxin (List, Campbell, CA, USA) at pCa 6.3 (for PV and VD) or at pCa 6.7 (for FA) buffered with 10 mm EGTA in order to permeabilize the membrane while retaining cytosolic proteins (Masuo et al. 1994). The strips were further treated with 10 µm of the Ca²⁺ ionophore A23187 (Calbiochem) for 20 min at 20 °C in order to deplete the SR of Ca²⁺. To heavily permeabilize VD strips, we used 0.1 % Triton X-100 (Sigma) in the standard relaxing solution for 60 min at 5 °C. Compared with treatment at 20 °C, the cold temperature used during this protocol prevented a large reduction in Ca²⁺-activated force development in VD strips after treatment, indicating that demembranation at the lower temperature was milder. Experiments using all types of permeabilized tissues were carried out at 20 °C.

Measurement of MLC phosphorylation

Details of the measurement of MLC phosphorylation using two-dimensional gel electrophoresis have been described previously (Kitazawa et al. 1991a). For evaluation of phosphorylation levels, the equation used was:

where U = unphosphorylated MLC, P₁ = monophosphorylated MLC, and P₂ = diphosphorylated MLC.

Proteins and antibodies

Recombinant MYPT1 was kindly provided by Drs K. Mabuchi and T. Tao (Boston Biomedical Research Institute). N-terminal, His-tagged, fusion protein corresponding to amino acids 1–543 of rat Rho-kinase (ROKα/ROCK-II) was from Upstate Biotechnology (Waltham, MA, USA). Phosphorylation of the MYPT1 fusion protein was carried out at 30 °C for different times in 50 mmNaCl, 20 mm Tris-HCl, 5 mm magnesium methanesulphonate, 1 mm EDTA, 0.1 mm dithiothreitol, 100 µg ml⁻¹ MYPT1, 0.5 units Rho-kinase, and 0.1 mm ATP. Anti-CPI-17 IgY antibody was raised in chicken eggs using His6-tagged porcine recombinant CPI-17 and purified using Affigel-10 resin conjugated with untagged CPI-17. Anti-phospho[Thr³⁸]-CPI-17 (anti-p[Thr³⁸]) antibody was raised in rabbit using the CPI-17 phosphopeptide ARV(phospho-T)VKYDRREL (corresponding to the sequence of amino acids 35–46 of porcine CPI-17; Eto et al. 1997) and purified using Affigel-10 resin conjugated with thiophosphorylated CPI-17 (Kitazawa et al. 2000). Monoclonal MYPT1 antibody was from BabCO (Richmond, CA, USA). Anti-phospho[Thr⁶⁹⁵]-MYPT1 antibody (anti-p[Thr⁶⁹⁵]) was raised in rabbit using the MYPT1 Thr⁶⁹⁵ phosphopeptide CQSRRS(phospho-T)QGVTL (corresponding to the sequence of amino acids 689–699 of human MYPT1 with an additional cysteine at the N-terminus; Chen et al. 1994; Feng et al. 1999a). Antibodies were absorbed from the antiserum onto unphosphorylated peptide conjugated to resin, prior to affinity purification of the phospho-specific antibody using a phospho-peptide resin. Polyclonal anti-phospho[Thr⁸⁵⁰]-MYPT1 antibody (anti-p[Thr⁸⁵⁰]) against a MYPT1 Thr⁸⁵⁰ phosphopeptide EKRRS(phospho-T)GVSFW (corresponding to amino acid residues 847–857 of human MYPT1) was obtained from Upstate Biotechnology. Secondary antibodies against chicken were from Promega (Madison, WI, USA). Anti-mouse and anti-rabbit secondary antibodies were from Chemicon (Temecula, CA, USA).

Immunoblotting

Non-treated intact and permeabilized smooth muscle strips were rapidly frozen and treated as previously described (Woodsome et al. 2001). Briefly, frozen strips were transferred onto frozen acetone containing 10 % trichloroacetic acid, incubated at −80 °C overnight and then gradually warmed up to room temperature. The acid-fixed strips were washed with acetone several times and then air-dried. The dried strips were homogenized with hand glass-glass homogenizer in Laemmli sample buffer (Woodsome et al. 2001). To examine the phosphorylation level of both CPI-17 and MYPT1 in the same sample, Western blotting experiments were always carried out in duplicate. Equal amounts of each extract were loaded onto two identical polyacrylamide gels composed of 15 % acrylamide at the bottom (for CPI-17) and 8 % in the middle (for MYPT1), with a stacking gel on top. Separated proteins were transferred to the same nitrocellulose membranes. The membranes were blocked in a Tris-buffered saline solution containing 0.05 % Tween-20, 5 % non-fat milk, and 1 % bovine serum albumin. The membranes were then incubated with a primary antibody followed by an alkaline phosphatase-conjugated secondary antibody. The immunoblots were developed with an alkaline phosphatase substrate solution (Sigma) to visualize immunoreactive proteins. The bands of alkaline phosphatase products were digitized and analysed as previously described (Woodsome et al. 2001). We compared the ratios of phosphorylated CPI-17 and MYPT1 to the total amount of CPI-17 and MYPT1, respectively, in the paired set of Western blots as described previously (Kitazawa et al. 2000).

Antibody specificity

The specificity of the rabbit antibody for phosphorylated Thr³⁸ of CPI-17 has been demonstrated previously (Kitazawa et al. 2000). To determine the specificity of the MYPT1 anti-p[Thr⁶⁹⁵] antibody to phosphorylated Thr⁶⁹⁵ of MYPT1, dot blot analyses were first performed using various MYPT1 peptides. This antibody reacted with MYPT1 Thr⁶⁹⁵ phosphopeptide (QSRRSpTQGVTL) but not unphosphorylated peptide with the same amino acid sequence or MYPT1 Thr⁸⁵⁰ phosphopeptide with a different sequence (EKRRSpTGVSFW). The monoclonal anti-MYPT1 antibody was previously used for quantitative immunoassays for MYPT1 content in various smooth muscle tissues (Woodsome et al. 2001). Western blotting with the anti-MYPT1 antibody showed equal reactivity toward the unphosphorylated and Rho-kinase-phosphorylated forms of recombinant MYPT1. In contrast, the anti-p[Thr⁶⁹⁵] antibody hardly reacted with unphosphorylated recombinant MYPT1 while anti-p[Thr⁸⁵⁰] weakly responded to unphosphorylated MYPT1; however, the interaction was 5.7 ± 0.1 % (n = 9) of the response seen with the phosphorylated protein. Both phospho-specific antibodies offered a near-linear response to Rho-kinase-phosphorylated recombinant MYPT1. Using an extract from PV smooth muscle tissues stimulated with 30 µm phenylephrine and 1 µm endothelin-1, the linearity of the antibody response was examined in the range of protein concentrations used in this study. There were no significant differences in Western blot linearity between the three antibodies. The MYPT1 blots of either the Rho-kinase-phosphorylated recombinant protein or agonist-stimulated PV extracts by the anti-p[Thr⁶⁹⁵] antibody were erased by the presence of 1 µm phosphorylated Thr⁶⁹⁵ peptide. However, 3 µm of phosphorylated threonine did not affect either Western blots of PV extracts by the anti-p[Thr⁶⁹⁵] or the anti-p[Thr⁸⁵⁰] antibody.

³²P-incorporation into glutathione S-transferase (GST)-MYPT1 (654–880) fusion protein by Rho-kinase was assayed by immunoblotting using anti-p[Thr⁶⁹⁵] and anti-p[Thr⁸⁵⁰]-MYPT1 antibodies. Conditions were 0.16 mg ml⁻¹ GST-MYPT1 (654- 880) (Upstate), 0.01 unit of constitutively active Rho-kinase (ROK), and 0.1 mm [γ-³²P]ATP (3 µCi), in 25 mm Mops-NaOH pH 7.0, with 10 mm magnesium acetate, 1 mm dithiothreitol, 1 µm microcystin-LR, 4 mm Pefabloc (Roche Diagnostics), for 30 min at 30 °C. The reaction was terminated by addition of an equal volume of 2 × Laemmli sample buffer, and a 5 µl aliquot was subjected to immunoblotting. Duplicate membranes were prepared for anti-p[Thr⁶⁹⁵] and anti-p[Thr⁸⁵⁰]-MYPT1 staining. The blot was exposed against an imaging plate, and ³²P-labelled GST-MYPT1(654–880) on the blot was analysed by PhosphorImager (Amersham Pharmacia Biotech) using ImageQuant software.

Statistics

Results are expressed as the mean ± s.e.m. of n experiments. Statistical significance was evaluated with the Student’ s two-tailed t test, with P < 0.05 being considered significant.

Results

Contraction and phosphorylation of MLC, MYPT1, and CPI-17 in α-toxin-permeabilized PV and VD

To investigate the effect of G-protein activation on phosphorylation of signalling molecules at given levels of Ca²⁺, we first used α-toxin-permeabilized phasic portal vein (PV) smooth muscle, as it has a relatively high CPI-17 and MLCP content (Woodsome et al. 2001). Permeabilization with α-toxin allowed the control of intracellular Ca²⁺ concentrations (buffered with 10 mm EGTA) and other small soluble molecules, such as GTP and ATP without loss of cellular proteins (Kitazawa et al. 1991a). At the submaximal level of pCa 6.3, both the α₁-agonist phenylephrine (PE) and the unhydrolysable GTP analogue GTPγS significantly increased both contraction and MLC phosphorylation several-fold (Fig. 1A and B).

Figure 1. G-protein-mediated Ca²⁺ sensitization of contraction and myosin phosphorylation in rabbit portal vein (PV) smooth muscle.

Effects of rigor, pCa > 8, pCa 6.3 alone, pCa 6.3 + phenylephrine (PE) and pCa 6.3 + GTPγS on contraction (A) and MLC phosphorylation (B) in α-toxin-permeabilized rabbit portal vein (PV) smooth muscle. Rigor: after permeabilized with α-toxin at pCa 6.3 and treated with A23187 in the Ca²⁺-free relaxing solution, the PV strips were incubated in the Ca²⁺-free, MgATP-free, creatine phosphate-free solution for 30 min at 20 °C and then frozen. pCa > 8: the permeabilized strips were soaked in the 10 mm EGTA-containing relaxing solution for 30 min. pCa 6.3: after incubation in the relaxing solution for 30 min, the strips were stimulated with the pCa 6.3-containing solution for 13 min. pCa 6.3 + PE: the strips were first incubated in the pCa 6.3 solution alone for 10 min and then stimulated with the same solution except containing 30 µm PE for 3 min. pCa 6.3 + GTPγS: the strips were treated with the same procedure as in the pCa 6.3 + PE except that GTPγS was added instead of PE during stimulation. Force levels were normalized with the maximal levels developed at pCa 4.5 plus 30 µm GTPγS in individual strips. Phosphorylation levels of MLC were measured as described in Methods. The number of experiments was 8–13.

The specificities of phosphorylation-dependent antibodies for CPI-17 Thr³⁸, MYPT1 Thr⁸⁵⁰ and MYPT1 Thr⁶⁹⁵ have been described in Methods in detail. Phosphorylation-specific recognition of MYPT1 Thr⁶⁹⁵ and Thr⁸⁵⁰ by the antibodies was confirmed using ³²P-labelled recombinant MYPT1 fragment. Figure 2A shows a reasonable correlation between ³²P incorporation and immunoblots of the antibodies. We then used these antibodies to monitor the phosphorylation levels of the MYPT1 at Thr⁶⁹⁵ and Thr⁸⁵⁰ and CPI-17 at Thr³⁸ extracted from smooth muscle tissues. In parallel to contraction and MLC phosphorylation, the two Ca²⁺ sensitizing compounds significantly increased phosphorylation of CPI-17 at Thr³⁸ and of MYPT1 at Thr⁸⁵⁰ several-fold (Fig. 2B and C). However, phosphorylation of MYPT1 at Thr⁶⁹⁵ was not significantly increased even by GTPγS (Fig. 2D).

Figure 2. Phosphorylation of CPI-17 and MYPT1 in rabbit portal vein.

The figure shows in vitro phosphorylation of recombinant MYPT1 by Rho-kinase (A) and in vivo phosphorylation of CPI-17 Thr³⁸ (B), MYPT1 Thr⁸⁵⁰ (C), and MYPT1 Thr⁶⁹⁵ (D) at rigor, pCa 6.3 alone, pCa 6.3 + PE and pCa 6.3 + GTPγS in α-toxin-permeabilized rabbit PV smooth muscle. A, phosphorylation of GST-MYPT1 (654–880) with [γ-³²P]ATP was performed as described in Methods. Upper panels show immunoblots using anti-p[Thr⁶⁹⁵] (left) and anti-p[Thr⁸⁵⁰] (right). Lower panels show their autoradiograms. In B, C and D, upper panel shows a representative paired set of Western blots of total and phosphorylated CPI-17 Thr³⁸, MYPT1 Thr⁸⁵⁰ and MYPT1 Thr⁶⁹⁵, respectively. Average phosphorylation levels (lower panel) of CPI-17 Thr³⁸ (B), MYPT1 Thr⁸⁵⁰ (C) and MYPT1 Thr⁶⁹⁵ (D) are expressed as a percentage of a value induced by 30 µm GTPγS at pCa 6.3 for each site. The number of experiments was 3–4 for B, C and D.

To determine whether the phosphorylation level of MYPT1 Thr⁶⁹⁵ was regulated in situ, MgATP and creatine phosphate were removed from the relaxing solution to block protein kinase activity. Under these rigor conditions for 30 min, phosphorylation at Thr⁶⁹⁵, along with phosphorylation at MYPT1 Thr⁸⁵⁰ and CPI-17 Thr³⁸, was reduced to a negligible level (Fig. 2). To determine whether the phosphorylation level of MYPT1 Thr⁶⁹⁵ at pCa 6.3 alone was already maximal, the potent PP1 and PP2A phosphatase inhibitor calyculin A was added for 20 min in the presence of GTPγS at pCa 6.3 to block in situ phospho-Thr⁶⁹⁵ phosphatase activity. MYPT1 Thr⁶⁹⁵ phosphorylation was significantly increased to 156 ± 11 % (n = 4) of the level at pCa 6.3 + GTPγS, and phosphorylation of MYPT1 Thr⁸⁵⁰ and CPI-17 Thr³⁸ was also enhanced to 173 ± 10 % and 121 ± 8 %, respectively.

To compare the results obtained with the PV, we used permeabilized vas deferens (VD), which has low (5 times lower than in PV) CPI-17 but high MLCP expressions. Therefore, in this tissue, CPI-17 content is thought to be too low to significantly contribute to Ca²⁺ sensitization (Woodsome et al. 2001). Application of GTPγS to VD in a control pCa 6.3 solution resulted in a several-fold increase in the phosphorylation of both CPI-17 Thr³⁸ and MYPT1 Thr⁸⁵⁰ (Fig. 3B and C), which mirrored the contractile response (Fig. 3A), while MYPT1 phosphorylation at Thr⁶⁹⁵ was only increased by about 30 % (Fig. 3D). We observed that an increase in Ca²⁺ from pCa > 8 to 6.3 had no effect on MYPT1 phosphorylation at either Thr⁶⁹⁵ or Thr⁸⁵⁰, while phosphorylation of CPI-17 at Thr³⁸ had a slight tendency to increase. To examine the effect of skinning on MYPT1 phosphorylation, we treated VD muscle strips with 0.1 % Triton X-100 for 60 min at 4 °C in the pCa > 8 solution. After the Triton X-100 treatment, the MYPT1 content was slightly decreased to 75 ± 11 % (n = 3) of that of untreated tissues. No significant reduction was found in Thr⁸⁵⁰ phosphorylation in the remaining MYPT1 (94 ± 11 % of that at pCa > 8 in α-toxin-permeabilized tissues). In contrast, Thr⁶⁹⁵ phosphorylation after the skinning treatment significantly decreased to 47 ± 8 %.

Figure 3. Uncoupling of GTPγS-induced force generation and phosphorylation of MYPT1 at Thr⁶⁹⁵ in rabbit vas deferens (VD).

The figure shows effects of rigor, pCa > 8, pCa 6.3 alone and pCa 6.3 + GTPγS on contraction (A), CPI-17 Thr³⁸ phosphorylation (B), MYPT1 Thr⁸⁵⁰ phosphorylation (C), and MYPT1 Thr⁶⁹⁵ phosphorylation (D) in α-toxin-permeabilized rabbit vas deferens (VD) smooth muscle. Conditions used were the same as those in Figs 1 and 2, except that different smooth muscle tissues were used (n = 3–6).

Effects of depolarization and Ca²⁺ sensitizing agonists on contraction and phosphorylation of MYPT1 and CPI-17 in intact VD smooth muscle

To confirm the results obtained in permeabilized smooth muscle tissues, we used untreated intact VD smooth muscle. Membrane depolarization with high (154 mm) K⁺ solution caused a biphasic contraction (see Fig. 5A); a rapid transient component of force development (peak was reached 15 s after stimulation) immediately followed by a rapid partial relaxation and a secondary, slow rise in force of variable size. Following the so-called ‘after-contraction’, force gradually declined to a lower steady state level (Fig. 4A and Fig. 5A). The Ca²⁺ sensitizing agonists phenylephrine (PE) and endothelin-1 (ET) evoked a comparatively slow development followed by a sustained contraction (Fig. 4A).

Figure 5. Rho-kinase independent phosphorylation of MYPT1 Thr⁶⁹⁵ in intact VD.

Effect of 1 µm prazosin, 10 µm Y-27632 or 3 µm GF109203X on contraction (A), CPI-17 Thr³⁸ phosphorylation (B), MYPT1 Thr⁸⁵⁰ phosphorylation (C), and MYPT1 Thr⁶⁹⁵ phosphorylation (D) is shown either at rest, at the peak (high K⁺ peak) or during the plateau phase of high K⁺-induced contraction (high K⁺ plateau) in intact rabbit VD smooth muscle. The inhibitors were applied from 10 min before stimulations throughout experiments. The strips were stimulated and frozen as described in Fig. 3 (n = 4–7).

Figure 4. Uncoupling of agonist-induced force generation and phosphorylation of MYPT1 Thr⁶⁹⁵ in intact rabbit VD.

The figure shows contraction (A), CPI-17 Thr³⁸ phosphorylation (B), MYPT1 Thr⁸⁵⁰ phosphorylation (C), and MYPT1 Thr⁶⁹⁵ phosphorylation (D) at rest or during contraction stimulated with either high K⁺, PE, or endothelin-1 (ET) in intact rabbit VD smooth muscle. Force and phosphorylation levels were measured in the normal external solution either at rest, 15 s after high K⁺ was applied (K-peak), 10 min after high K⁺ (K-plateau), 1.5 min after 30 µm PE or 3 min after 1 µm ET. Force levels were normalized to the peak of high K⁺-induced contraction. The relative phosphorylation was determined as described in Methods. The mean phosphorylation at rest was expressed as 1 (n = 4–7).

Neither K⁺ nor agonists enhanced the resting phosphorylation level of MYPT1 at Thr⁶⁹⁵ (Fig. 4D), whereas both treatments significantly increased the phosphorylation levels of CPI-17 Thr³⁸ and MYPT1 Thr⁸⁵⁰ (Fig. 4B and C), similar to the results seen in permeabilized tissues (Fig. 2 and Fig. 3). ET increased contraction and CPI-17 Thr³⁸ phosphorylation significantly but less than PE, while MYPT1 phosphorylation at Thr⁸⁵⁰ was more markedly enhanced by ET than by PE. The lack of change in MYPT1 Thr⁶⁹⁵ phosphorylation by K⁺ and agonists was confirmed using an additional rabbit anti-phospho-Thr⁶⁹⁵ antibody (data not shown). In addition, no significant change (122 ± 13 % of rest; n = 3) in the phosphorylation of MYPT1 at Thr⁶⁹⁵ was found in intact tonic FA smooth muscle (with high CPI-17 and low MYPT1 expression) stimulated by histamine, while phosphorylation of CPI-17 at Thr³⁸ was markedly elevated to 25-fold (n = 3) of the resting level as previously demonstrated (Kitazawa et al. 2000).

Pharmacological aspects of contraction and phosphorylation of MYPT1 and CPI-17 in intact VD

We examined the effects of treatment with either the α₁-antagonist prazosin, the Rho-kinase inhibitor Y-27632, or the PKC inhibitor GF109203X. Prazosin (1 µm) completely blocked the development of the 30 µm PE-induced contraction of intact VD. Y-27632 (10 µm) and GF109203X (3 µm) reduced PE-induced contractions to 6 ± 3 and 40 ± 5 % (n = 3), respectively, of the control values caused by PE alone. As shown in Fig. 5A, prazosin slightly affected the peak of the high K⁺-induced contraction (to 95 ± 1 % of control; n = 3) but markedly reduced the plateau phase (to 39 ± 4 %). In contrast, prazosin had no observable effect on the high-K⁺-induced contractions in PV and FA strips (n = 3). Y-27632 and GF109203X both potently reduced the plateau phase of the high-K⁺-induced contraction in the VD to 42 ± 8 and 28 ± 7 % (n = 3), respectively, with a slight inhibition of the phasic component of contraction similar to the effect of prazosin. In tonic FA strips, Y-27632 also reduced the tonic phase of high-K⁺-induced contraction to 26 ± 5 % (n = 5), with little effect (94 ± 5 % of control) on the peak of contraction (also see Mita et al. 2002), while GF109203X decreased the tonic phase to 58 ± 7 % (n = 3).

Figure 5 shows effects of the above inhibitors on CPI-17 phosphorylation at Thr³⁸ (B) and on MYPT1 phosphorylation at Thr⁸⁵⁰ (C) and Thr⁶⁹⁵ (D). None of inhibitors significantly affected the phosphorylation level of the Thr⁶⁹⁵. Prazosin abolished the enhanced phosphorylation of both MYPT1 Thr⁸⁵⁰ and CPI-17 Thr³⁸ during the tonic component of high K⁺-induced contraction. Y-27632 more potently inhibited MYPT1 phosphorylation at Thr⁸⁵⁰ than GF109203X (Fig. 5C). In contrast, CPI-17 Thr³⁸ phosphorylation during PE-induced contraction was reduced more by GF109203X (21 ± 2 % of that induced by PE alone; n = 8) than Y-27632 (61 ± 3 %), similar to the effects of these agents on phosphorylation of CPI-17 Thr³⁸ during the prazosin-sensitive, high K⁺-induced contraction (Fig. 5B).

Discussion

Phosphorylation of MYPT1

Our study has disclosed that phosphorylation of MYPT1 at Thr⁶⁹⁵, a crucial inhibitory site for MLCP, is hardly elevated in response to G-protein activation in intact and α-toxin-permeabilized PV and VD smooth muscles. Previous reports have shown that agonists and GTPγS can increase phosphorylation of MYPT1 along with MLC phosphorylation and contraction in smooth muscle. Total ³²P radioactivity incorporated into MYPT1 is enhanced by activation of G-proteins (Trinkle-Mulcahy et al. 1995; Swärd et al. 2000; Nagumo et al. 2000). In this study, however, Ca²⁺ sensitizing agonists and even the non-hydrolysable GTP analogue GTPγS, which activates all G-proteins linked to the Ca²⁺ sensitizing receptors, did not significantly increase MYPT1 Thr⁶⁹⁵ phosphorylation in smooth muscle tissues. The results were confirmed by using two specific antibodies against the phospho-Thr⁶⁹⁵ peptide. This implies that there are no G-protein-coupled signalling pathways leading to Thr⁶⁹⁵ phosphorylation in smooth muscle. In contrast to Thr⁶⁹⁵, MYPT1 Thr⁸⁵⁰ is significantly phosphorylated in response to activation of G-proteins. Therefore, Thr⁸⁵⁰ phosphorylation might explain the previous reports of elevation of total MYPT1 phosphorylation. Two papers have documented in situ Thr⁶⁹⁵ phosphorylation using non-muscle cells. Feng et al. (1999a) provided a pioneer work showing that, in Swiss 3T3 cells, Thr⁶⁹⁵ was phosphorylated up to 4-fold after lysophosphatidic acid stimulation and this phosphorylation was inhibited by Y-27632 at the same concentration used in this study. Watanabe et al. (2001) reported that in human platelets, phosphorylation of both MYPT1 Thr⁶⁹⁵ and CPI-17 Thr³⁸ was increased in response to a thromboxane A₂ analogue and inhibited by, respectively, Y-27632 and GF109203X. These conflicting results regarding Thr⁶⁹⁵ phosphorylation may be due to variations in intracellular signalling cascades downstream of G-protein activation in different cell types.

Y-27632 inhibits agonist- and GTPγS-induced, but not Ca²⁺-activated, contraction in smooth muscle, suggesting that Rho-kinase signalling is involved in smooth muscle Ca²⁺ sensitization (Uehata et al. 1998; Somlyo & Somlyo, 2000). However, Y-27632 had no effect on phosphorylation of MYPT1 Thr⁶⁹⁵ in either the PV or VD, suggesting that Thr⁶⁹⁵ is not an endogenous target of Rho-kinase in smooth muscle. Other MYPT1 kinases, such as ZIPK (MacDonald et al. 2001), ILK (Kiss et al. 2002) and DMK (Murányi et al. 2001), are insensitive to Y-27632 and can phosphorylate Thr⁶⁹⁵ and other sites but not Thr⁸⁵⁰, and therefore are possible candidates responsible for Thr⁶⁹⁵ phosphorylation. Inhibition of protein kinases by removal of MgATP and inhibition of phosphatases by calyculin A, respectively, decreased and increased the phosphorylation level of in situ MYPT1 Thr⁶⁹⁵, indicating that the phosphorylation levels are submaximal. Nonetheless, the phosphorylation of Thr⁶⁹⁵ is not required for G-protein-mediated Ca²⁺ sensitization of smooth muscle.

In contrast, phosphorylation of MYPT1 at Thr⁸⁵⁰ was increased during G-protein activation. Y-27632 effectively inhibited Thr⁸⁵⁰ phosphorylation and phenylephrine-induced contraction in VD. These results suggest that Rho-kinase is activated in response to G-protein activation and is responsible for the phosphorylation of Thr⁸⁵⁰, which does not have clear functional consequences (Feng et al. 1999a). Interestingly, however, Velasco et al. (2002) have recently reported that phosphorylation of Thr⁸⁵⁰ reduces the affinity of N-terminal fragments of MYPT1 for myosin. The involvement of Thr⁸⁵⁰ phosphorylation in smooth muscle Ca²⁺ sensitization requires further investigation. The effects of the Rho-kinase phosphorylation at other minor sites on MLCP activity have not been studied. The findings in this study, however, do not eliminate the possibility that Rho-kinase has a role in Ca²⁺ sensitization through MLCP inhibition. The Rho-kinase pathway may be linked to phosphorylation of other target(s) that contribute to MLCP inhibition, such as the Thr⁸⁵⁰ of MYPT1 and the cytosolic inhibitor protein CPI-17.

Dominant role of CPI-17 Thr³⁸ phosphorylation in Ca²⁺ sensitization of vascular smooth muscle

CPI-17 is a strong candidate for physiological involvement in the G-protein- and PKC-induced inhibition of MLCP and Ca²⁺ sensitization. In human platelets, both MYPT1 Thr⁶⁹⁵ and CPI-17 Thr³⁸ are phosphorylated in response to stimulation with a thromboxane A₂ analogue (Watanabe et al. 2001). Inhibition of Rho-kinase and PKC reduces phosphorylation of MYPT1 Thr⁶⁹⁵ and CPI-17 Thr³⁸, respectively, which results in a decrease in MLC phosphorylation. These results suggest that both the Rho-kinase-MYPT1 and PKC-CPI-17 pathways contribute to regulation of MLC phosphorylation in thromboxane-induced secretion of ATP from the platelets. In contrast, we have showed that CPI-17 Thr³⁸, but not MYPT1 Thr⁶⁹⁵, is phosphorylated in response to G-protein activation in phasic PV smooth muscle where both CPI-17 and MYPT1 are highly expressed. Stimulation of PV with PKC activator induces a contraction that is 80 % of the GTPγS-induced contraction (Woodsome et al. 2001). Thus, phosphorylation of CPI-17 Thr³⁸ seems to play a primary role in the regulation of MLCP in agonist-induced signalling in phasic PV smooth muscle. Consistent with this, in tonic arterial smooth muscle such as the FA, which has high CPI-17 and low MLCP expression, both PKC and G-protein activators equally increase phosphorylation of CPI-17 at Thr³⁸ to a high level with a large increase in Ca²⁺ sensitization of both MLC phosphorylation and contraction through a 50 % reduction in MLCP activity (Masuo et al. 1994; Woodsome et al. 2001). Inhibition of histamine-induced Thr³⁸ phosphorylation by inhibitors such as GDPβS, Y-27632 and GF109203X is accompanied by reductions in contractile force in FA smooth muscle (Kitazawa et al. 2000). These results imply that the G-protein-mediated inhibition of MLCP is predominantly through the PKC-CPI-17-MLCP signalling pathway in vascular smooth muscles, such as FA and PV.

Ca²⁺ sensitization of phasic visceral smooth muscle

Compared to tonic vascular smooth muscle, CPI-17 expression in the VD is minimal. Even though PKC stimulation causes a large relative increase in Thr³⁸ phosphorylation to a level similar to that induced by the direct G-protein activator GTPγS, increases in MLC phosphorylation and contraction caused by a PKC activator are minimal (Woodsome et al. 2001). This clearly suggests that CPI-17 phosphorylation alone cannot cause a significant Ca²⁺ sensitization in VD. Compared with FA (Kitazawa et al. 2000), Y-27632 had a greater inhibitory effect on agonist-induced contractions in VD. These results suggest that Rho-kinase is involved in the MLCP inhibition and Ca²⁺ sensitization in phasic visceral smooth muscle, though not through phosphorylation of MYPT1 Thr⁶⁹⁵ and/or CPI-17 Thr³⁸. Here a question is raised: how does Rho-kinase induce myosin phosphorylation and contraction without any contribution from phosphorylation of MYPT1 Thr⁶⁹⁵ and/or CPI-17 Thr³⁸? One possible mechanism for Y-27632-sensitive Ca²⁺ sensitization is the direct phosphorylation of MLC at Ser¹⁹ by Rho-kinase. This phosphorylation induces Ca²⁺ sensitization of contraction without inhibiting MLCP (Amano et al. 1996). In fact, a constitutively active fragment of Rho-kinase produced a contraction and MLC phosphorylation in skinned smooth muscle at a given level of Ca²⁺ (Kureishi et al. 1997). However, the direct phosphorylation hypothesis is inconsistent with the fact that agonists and GTPγS significantly decrease the dephosphorylation rate but do not increase the phosphorylation rate of in situ MLC during Ca²⁺ sensitization in the PV and FA (Kitazawa et al. 1991b; Masuo et al. 1994). Therefore, the direct phosphorylation is very unlikely to occur in smooth muscle Ca²⁺ sensitization. It has been reported that Rho-kinase phosphorylates multiple cytoskeletal proteins, such as ezrin/radixin/moesin and adducin (Fukata et al. 2001). As the primary target(s) of Rho-kinase in smooth muscle Ca²⁺ sensitization is unclear, the VD might offer a suitable experimental model for studying novel signalling mediated by Rho-kinase.

Membrane depolarization with high K⁺ markedly enhanced phosphorylation of CPI-17 at Thr³⁸ in parallel to force development in the VD. In contrast, in the tonic FA, high K⁺ does not significantly increase CPI-17 Thr³⁸ phosphorylation (Kitazawa et al. 2000). Prazosin, a specific α₁-antagonist, inhibited both the K⁺-induced contraction and CPI-17 phosphorylation in the VD but not in the FA, suggesting that the high K⁺-induced increase in CPI-17 Thr³⁸ phosphorylation is due to the depolarization-induced neurotransmitter (noradrenaline) release from sympathetic nerves in the VD tissue (Burnstock, 1995).

Significance of phosphorylation of MYPT1 Thr⁶⁹⁵

The steady, considerable level of Thr⁶⁹⁵ phosphorylation found in all intact and α-toxin-permeabilized smooth muscle tissues, regardless of contractile state, is suggestive of a partial but substantial inhibition of the in situ MLCP activity. Treatment with Triton X-100 induced a significant reduction in the Thr⁶⁹⁵ phosphorylation. This suggests a lesser inhibition of MLCP via Thr⁶⁹⁵ phosphorylation in the demembranated compared to the intact and α-toxin-permeabilized smooth muscle. This also indicates that the endogenous Thr⁶⁹⁵ kinase(s) is not as tightly associated with the myofibrils as the in situ Thr⁶⁹⁵ phosphatase, or that the latter enzyme is activated by Triton X-100 treatment. The rate constant of dephosphorylation of MLC in α-toxin-permeabilized PV smooth muscle in a solution containing no MgATP, pCa > 8, and a MLCK inhibitor (ML-9; Kitazawa et al. 1991b) was several-fold lower than that found in the Triton X-100-skinned tissue in a solution containing MgATP at pCa 4.5 (Butler et al. 1994). Although the conditions used were different, this differentiation in the in situ MLCP activity may be due to a decrease in the phosphorylation level of MYPT1 Thr⁶⁹⁵ caused by Triton X-100 treatment. Borman et al. (2002) recently found that constitutively active recombinant ZIPK provoked a Ca²⁺-independent contraction and increased Thr⁶⁹⁵ phosphorylation of MYPT1 several-fold in β-escin-permeabilized smooth muscle tissues. This apparently large increase in phosphorylation compared to the basal level might be due to a lowering of Thr⁶⁹⁵ phosphorylation by the skinning procedure. Although it remains to be further investigated, the finding of a significant level of Thr⁶⁹⁵ phosphorylation raises the interesting possibility that the in situ MLCP activity, regardless of contractile states, is always partially but significantly inhibited. This might offer a reasonable explanation for the previously observed activation of in situ MLCP by cGMP signalling pathway (Wu et al. 1996; Lee et al. 1997). Indeed, dephosphorylation of MYPT1 Thr⁶⁹⁵ occurs in cultured aortic smooth muscle cells on addition of a cGMP analogue (Sandu et al. 2001), although the effect of cGMP on MYPT1 Thr⁶⁹⁵ phosphorylation in intact smooth muscle tissues needs to be investigated.

In conclusion, neither PV nor VD requires phosphorylation of MYPT1 Thr⁶⁹⁵ downstream of RhoA/Rho-kinase activation during smooth muscle Ca²⁺ sensitization. The Thr⁶⁹⁵ phosphorylation is not regulated during a contraction-relaxation cycle but may have an important role in maintaining MLCP activity at a submaximal level that may vary among smooth muscle types. On the other hand, phosphorylation of CPI-17 Thr³⁸ is responsible for G-protein-mediated Ca²⁺ sensitization in vascular smooth muscles, whereas in phasic visceral muscles a novel mechanism might regulate the activity of MLCP in response to stimuli.