Targeting of a Novel Ca+2/Calmodulin-Dependent Protein Kinase II Is Essential for Extracellular Signal-Regulated KinaseeCMediated Signaling
http://www.100md.com
循环研究杂志 2005年第9期
the Boston Biomedical Research Institute (W.A.M., S.S.G., H.-D.J., C.G., K.G.M.), Watertown
the Department of Medicine (K.G.M.), Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Mass.
Abstract
Subcellular targeting of kinases controls their activation and access to substrates. Although Ca2+/calmodulin-dependent protein kinase II (CaMKII) is known to regulate differentiated smooth muscle cell (dSMC) contractility, the importance of targeting in this regulation is not clear. The present study investigated the function in dSMCs of a novel variant of the isoform of CaMKII that contains a potential targeting sequence in its association domain (CaMKII G-2). Antisense knockdown of CaMKII G-2 inhibited extracellular signal-related kinase (ERK) activation, myosin phosphorylation, and contractile force in dSMCs. Confocal colocalization analysis revealed that in unstimulated dSMCs CaMKII G-2 is bound to a cytoskeletal scaffold consisting of interconnected vimentin intermediate filaments and cytosolic dense bodies. On activation with a depolarizing stimulus, CaMKII G-2 is released into the cytosol and subsequently targeted to cortical dense plaques. Comparison of phosphorylation and translocation time courses indicates that, after CaMKII G-2 activation, and before CaMKII G-2 translocation, vimentin is phosphorylated at a CaMKII-specific site. Differential centrifugation demonstrated that phosphorylation of vimentin in dSMCs is not sufficient to cause its disassembly, in contrast to results in cultured cells. Loading dSMCs with a decoy peptide containing the polyproline sequence within the association domain of CaMKII G-2 inhibited targeting. Furthermore, prevention of CaMKII G-2 targeting led to significant inhibition of ERK activation as well as contractility. Thus, for the first time, this study demonstrates the importance of CaMKII targeting in dSMC signaling and identifies a novel targeting function for the association domain in addition to its known role in oligomerization.
Key Words: CaMKII smooth muscle contractility targeting extracellular signal-regulated kinase
Introduction
Ca2+/calmodulin-dependent protein kinase II (CaMKII) is a Ser/Thr kinase that is expressed as 4 different isoforms (, , , and ). Each member of this kinase family possesses 3 major domains (Figure 1A): an N-terminal catalytic/regulatory domain containing a Ser/Thr kinase region and overlapping autoinhibitory and Ca2+/calmodulin binding regions, a central-linker domain containing variable regions, and a C-terminal association domain that promotes oligomerization.1 CaMKII is not active until Ca2+/calmodulin binds to its regulatory domain. This binding allows CaMKII to undergo autophosphorylation at Thr287 (numbering according to the isoform), which, in turn, increases its affinity for Ca2+/calmodulin and allows kinase activity even in the absence of Ca2+/calmodulin.
The and isoforms of CaMKII are predominantly expressed in neural tissue,2 where they have been demonstrated to be involved in synaptic plasticity, memory, and learning.3,4 In contrast, differentiated smooth muscle cells (dSMCs) express mainly the isoform of CaMKII,5 which has been linked to contractile activity by the use of antisense and pharmacological inhibitors.6,7
It is well known that subcellular localization and targeting of CaMKII are important in regulating signal transduction in many biological systems.8,9 For instance, in neuronal cells N-methyl-D-aspartate receptor stimulation causes the and isoforms of CaMKII to change from an F-actin bound state to a postsynaptic membrane bound state. This targeting is thought to increase synaptic strength because it brings CaMKII closer to its substrates, including ion channels and signaling proteins.3,4
The present study investigated the function of a novel variant of the isoform of CaMKII (CaMKII G-2; Figure 1A), recently cloned from an aorta tissue library,10 in dSMC contractility and signaling. The results obtained indicate that CaMKII G-2 is expressed as a protein in dSMCs and plays a significant role in dSMC signaling during contraction. In unstimulated dSMCs, CaMKII G-2 is bound to a cytoskeletal scaffold consisting of interconnected vimentin intermediate filaments and cytosolic dense bodies. On activation, CaMKII G-2 phosphorylates vimentin and is released into the cytosol, where it is subsequently targeted to cortical dense plaques. Targeting of CaMKII G-2 is dependent on a polyproline region confined within the last 99 residues of the association domain and is essential to dSMC signaling because it leads to ERK (extracellular signal-regulated kinase) 1 activation, LC20 phosphorylation, and contraction.
Materials and Methods
Fifty-three ferrets were euthanized with chloroform as approved by the Institutional Care and Use Committee. As previously described,6 the aorta was dissected, and then the tissue strips were attached to a force transducer, treated with a depolarizing stimulus (51 mmol/L KCl), and quick-frozen. Phosphorothioate oligonucleotides were synthesized to be complementary to residues 515 to 521 of the association domain of CaMKII G-2 and were loaded into the aorta tissue following a previously published protocol.6 A decoy peptide was designed against residues 571 to 586 of CaMKII G-2, which are predicted to contain 2 overlapping Src homology 3 (SH3)-binding domains,10 and was chemically loaded into the aorta tissue. Quick-frozen tissue samples were homogenized and analyzed via Western blot, as previously described.6 Overlay assays (also known as far Western blots) were performed as previously described.11 The overlay buffer contained 40 e/mL recombinant CaMKII G-2 association domain protein. The polymer-to-subunit ratio of vimentin in aorta tissue and COS-7 cells was measured using a previously published differential centrifugation method.12 dSMCs from aorta tissue were enzymatically isolated, paraformaldehyde fixed, and stained as previously described.13 Images were acquired with a Kr/Ar laser (Bio-Rad Radiance 2000) scanning confocal microscope equipped with a Nikon X-60 (NA1.4) oil immersion objective. The degree of colocalization between a pair of immunostained proteins was determined by computing the percentage of the total pixels that contain both fluorophores given predetermined thresholds.
An expanded Materials and Methods section, with additional details about the procedures, materials used, and data analysis methods, can be found in the online data supplement at http://circres.ahajournals.org.
Results
Tissue Expression of CaMKII G-2
Because the novel CaMKII G-2 variant has been previously described only as a cDNA isolated from an aorta library,10 the question arises whether it is expressed as a protein. To clarify this issue, an antibody that recognizes the association domain of CaMKII G-2 was generated and tested on aorta tissue whole-cell homogenate. The specificity of the antibody was validated by transiently transfecting COS-7 cells to express either CaMKII G-2 or a similar but distinct CaMKII variant, CaMKII G-1 (Figure 1A). 10 The antieCCaMKII G-2 antibody recognized a single band at the expected molecular mass of 62 kDa in lysates of CaMKII G-2eCtransfected COS-7 cells but recognized no bands in lysates of CaMKII G-1eCtransfected COS-7 cells (Figure 1B). Furthermore, the antieCCaMKII G-2 antibody recognized a single band at 62 kDa in aorta whole-cell homogenate, thereby demonstrating adequate specificity for imaging experiments. Interestingly, CaMKII G-2 expression was detectable in protein matched homogenates of several smooth muscles, as well as the heart and brain, but was essentially absent from skeletal muscle and the liver (Figure 1C).
CaMKII G-2 Specific Antisense Oligonucleotides Inhibit the Signaling Events That Regulate Contractility
To determine whether CaMKII G-2 is involved in dSMC signaling during contraction, aorta tissue was loaded with antisense oligonucleotides designed against part of the sequence that codes for the novel association domain. CaMKII G-2 protein expression in antisense-treated tissue was significantly decreased compared with both sham and random sequence-treated tissues (Figure 2A). The effect of decreased CaMKII G-2 protein expression on contractility was determined by stimulating the antisense-treated tissue with a depolarizing stimulus. The contractile response to a depolarizing stimulus was significantly less in antisense-treated tissue compared with both sham and random sequence-treated tissues, with the effect being relatively more pronounced at a lower level of depolarization (Figure 2B).
In some,6,14 but not all smooth muscles,15eC18 CaMKII-dependent activation of ERK1/2 has been shown to be essential in a pathway that leads to LC20 phosphorylation and dSMC contraction. In antisense-treated tissue quick-frozen after exposure to a depolarizing stimulus, ERK1 phosphorylation was significantly less than that detected in both depolarized sham and depolarized random sequence-treated tissues (Figure 2C). Interestingly, a significant decrease in the corresponding phosphorylated ERK2 signal was not obtained. LC20 phosphorylation in depolarized antisense-treated tissue was significantly less than that detected in both depolarized sham and depolarized random sequence-treated tissues (Figure 2D).
CaMKII G-2 Colocalizes With Vimentin and -Actinin in Unstimulated dSMCs
To elucidate whether CaMKII G-2 is targeted to specific proteins, unstimulated dSMCs were fixed and immunostained for CaMKII G-2. Confocal microscopy indicated that CaMKII G-2 was localized primarily to filamentous structures throughout the dSMC (Figure 3A). Because some isoforms of CaMKII are known to bind to F-actin,3,4,9 dSMCs were costained for CaMKII G-2 and actin (Figure 3B). Although there were a few regions where CaMKII G-2 and actin overlapped, in general, the 2 proteins did not appear to colocalize (Figure 3C).
Unstimulated dSMCs were also costained for CaMKII G-2 and vimentin (Figure 3D through 3F), the primary intermediate filament protein expressed in aortic vascular smooth muscle.19Merged images (Figure 3F) illustrated that CaMKII G-2 (Figure 3D) and vimentin (Figure 3E) highly colocalized. However, a number of green spots where CaMKII G-2 and vimentin did not colocalize were consistently observed (white arrows in Figure 3F). These spots were 0.5 e and were closely associated with the vimentin filament network, which are both characteristic properties of cytosolic dense bodies.20eC22 Therefore, unstimulated dSMCs were also costained for CaMKII G-2 and -actinin (Figure 3G through 3I), a known dense body protein.22 Merged images (Figure 3I) illustrated that CaMKII G-2 (Figure 3G) and -actinin (Figure 3H) highly colocalized, except for a few regions (white arrows) where presumably CaMKII G-2 colocalized with vimentin.
The degree of colocalization between CaMKII G-2 and the cytoskeletal components was quantified. The relatively low degree of colocalization between CaMKII G-2 and actin in unstimulated dSMCs was confirmed because CG-2eCactin=0.30±0.02 and CactineCG-2=0.31±0.02 for 16 cells. In contrast, the colocalization coefficients measured for CaMKII G-2 and either vimentin or -actinin in unstimulated dSMCs were both significantly higher (P<0.001; CG-2eCvimentin=0.63±0.02 and CvimentineCG-2=0.67±0.01 for 12 cells; CG-2eC-actinin=0.65±0.02 and C-actinineCG-2=0.64±0.02 for 10 cells).
The Subcellular Location of CaMKII G-2 Is Dynamically Regulated by Depolarization
To see whether the localization of CaMKII G-2 to vimentin and -actinin was constitutive or regulated, dissociated dSMCs were depolarized for up to 10 minutes to activate CaMKII G-2 and then immunostained for CaMKII G-2. After 10 minutes of depolarization, CaMKII G-2 was relatively depleted from the cytosol and was concentrated at regions near the cell cortex (Figure 4A and 4B). The changes in targeting were quantified by measuring the cortex/cytosol fluorescence ratio of CaMKII G-2 at different time points during depolarization. The ratio increased significantly after 2 minutes of depolarization, reaching a value of 4 at 10 minutes (Figure 4C).
The Subcellular Locations of ERK1/2 and Mitogen-Activated Protein Kinase Kinase 1/2 Do Not Change During Depolarization
Previously, ERK1/2 has been shown to be downstream of CaMKII in the depolarization-mediated activation pathway.6,14 ERK1/2 has also been shown to translocate to a cortical signaling complex in dSMCs stimulated with an -adrenergic agonist.23 Therefore, to determine whether CaMKII G-2 translocates to increase its proximity to a cortical signaling complex during depolarization-mediated activation, dSMCs were immunostained for mitogen-activated protein kinase kinase (MEK) 1/2 and ERK1/2. Line scan analysis revealed that MEK1/2 and ERK1/2 were homogeneously distributed throughout the unstimulated dSMC, and, furthermore, these distributions remain unchanged during depolarization (Figure 4D). This demonstrates that although MEK1/2 and ERK1/2 are involved in both the depolarization- and -adrenergiceCactivation pathways, they are differentially targeted by different stimuli.
CaMKII G-2 Is Targeted to Dense Plaques in Depolarized dSMCs
CaMKII G-2 is targeted to cytoplasmic dense bodies in unstimulated dSMCs. To determine whether CaMKII G-2 is targeted to cortical dense plaques (which also contain -actinin) after stimulation, depolarized dSMCs were costained for CaMKII G-2 and the dense plaque protein vinculin.24 Merged images (Figure 4G and 4J) illustrated that CaMKII G-2 (Figure 4E and 4H) and vinculin (Figure 4F and 4I) highly colocalized. Furthermore, the cortical dense plaques formed a rib-like pattern (white arrows in Figure 4J) on the surface of the dSMC, as has been previously reported.25,26 The high degree to which CaMKII G-2 colocalized with vinculin in depolarized dSMCs was confirmed by quantifying the colocalization coefficients (CG-2eCvinculin=0.62±0.03 and CvinculineCG-2=0.62±0.03 for 12 cells).
CaMKII G-2 Binds Directly to Vimentin and -Actinin
Overlay assays were performed to determine whether CaMKII G-2 can bind directly to vimentin, -actinin, or vinculin. Significant binding of recombinant CaMKII G-2eCassociation domain protein to -actinin and vimentin in the whole-cell homogenate was found by immunoblotting (Figure 5A). In contrast, there was no significant binding between CaMKII G-2 and vinculin (Figure 5A). The degree of binding was quantified by measuring the densitometry of the overlay band normalized to the densitometry of the Napthol Blue Black protein stain on the same membrane (Figure 5B). Binding between CaMKII G-2 and either vimentin or -actinin was significantly higher than binding between CaMKII G-2 and vinculin. As a negative control, the binding between CaMKII G-2 and BSA (protein matched to the amount of vimentin in the homogenate) was analyzed and found to be negligible.
CaMKII Autophosphorylation and Vimentin Phosphorylation Precede CaMKII G-2 Translocation
To begin to address the mechanism of CaMKII G-2 translocation, the time course of CaMKII Thr287 autophosphorylation during depolarization was monitored and compared with the time course of CaMKII G-2 translocation. CaMKII autophosphorylation increased significantly above baseline at 0.5 minute of depolarization and reached 250% baseline by 10 minutes (Figure 6A). Therefore, increases in CaMKII autophosphorylation occurred several minutes before the onset of CaMKII G-2 translocation (Figure 4C).
Because activated CaMKII has been reported to phosphorylate intermediate filaments, vimentin phosphorylation was measured at a CaMKII-specific site.27 Vimentin phosphorylation increased significantly above baseline at 2 minutes of depolarization and reached 200% baseline by 10 minutes (Figure 6B). Therefore, increases in vimentin phosphorylation occurred after CaMKII autophosphorylation (Figure 6A) and were either coincident with or occurred slightly before CaMKII G-2 translocation (Figure 4C).
Because phosphorylation of vimentin at the CaMKII-specific site could be caused by any of the 6 variants of CaMKII, the level of vimentin phosphorylation was measured in depolarized CaMKII G-2 antisense-treated tissue (Figure 6C). Vimentin phosphorylation in depolarized antisense-treated tissue was significantly less than that measured in both depolarized sham and depolarized random sequence-treated tissues. Thus, the increase in vimentin phosphorylation during depolarization was at least partially attributable to the kinase activity of CaMKII G-2.
Depolarization Does Not Cause Vimentin Disassembly in dSMCs
The assembly/disassembly dynamics of vimentin intermediate filaments in cultured cells are known to be regulated by several kinases including CaMKII, RhoA-binding kinase , protein kinase A, and protein kinase C.12,28eC30 To test whether the stability of the vimentin filament network is regulated by CaMKII-mediated phosphorylation in dSMCs, the polymer-to-subunit ratio of vimentin was measured in aorta tissue and for comparison in COS-7 cells. As a positive control, cells were treated with calyculin-A, which is a type-1 and type-2A protein phosphatase inhibitor that has been shown to depolymerize at least 50% of the total vimentin pool in cultured BHK-21 cells.12 Unstimulated COS-7 cells contained an abundance of polymerized vimentin compared with the amount of soluble subunits (Figure 6D). Treatment of the COS-7 cells with calyculin-A caused disassembly of vimentin, as indicated by a relative increase in the soluble fraction. In unstimulated dSMCs, the majority of vimentin was also found in the polymerized fraction (Figure 6E). However, depolarization of dSMCs did not change the fractional distribution of vimentin. Interestingly, calyculin-A also did not affect the distribution of vimentin in dSMCs. Thus, the vimentin filament network in dSMCs appears to be more structurally stable than that of cultured COS-7 or BHK-21 cells.
CaMKII G-2 Targeting Is Controlled by a Polyproline Sequence Confined Within the Association Domain
Because vimentin depolymerization does not explain the redistribution of CaMKII G-2 during depolarization, it is possible that phosphorylation of vimentin simply allows CaMKII G-2 to be released and targeted elsewhere. One potential targeting sequence within CaMKII G-2 is a polyproline region within the association domain that is predicted to contain 2 overlapping SH3-binding domains (Figure 1A). This hypothesis was tested by synthesizing a decoy peptide of this region and loading the peptide into dSMCs to determine its effect on targeting and CaMKII G-2eCdependent signaling. Imaging of a chemically loaded dSMC demonstrated that the decoy peptide was distributed throughout the cell (Figure 7A). Furthermore, this type of distribution was observed in both unstimulated and depolarized dSMCs.
To determine whether the decoy peptide had any effect on CaMKII G-2 targeting, the distribution of CaMKII G-2 was measured in depolarized dSMCs that lacked a peptide, were loaded with the decoy peptide, or were loaded with a control peptide. The decoy peptide prevented CaMKII G-2 translocation during depolarization, whereas the control peptide had no effect (Figure 7B). The cortex/cytosol ratio of CaMKII G-2 fluorescence was significantly smaller in depolarized decoy peptide-loaded dSMCs compared with both depolarized sham and depolarized control peptide-loaded dSMCs (Figure 7C).
Inhibiting CaMKII G-2 translocation also negatively affected downstream signaling because phosphorylated ERK1 levels in depolarized decoy peptide-loaded tissue were significantly less than that measured in both depolarized sham and depolarized control peptide-loaded tissues (Figure 7D). Interestingly, the decoy peptide had no significant effect on phosphorylated ERK2 levels. Furthermore, there was a small but significant decrease in the contractile response (computed by normalizing the KCl-induced contraction measured after peptide loading with the KCl-induced contraction measured before peptide loading) to a depolarizing stimulus (51 mmol/L KCl for 5 minutes) in decoy peptide-loaded tissue (121±5%; n=5) compared with both sham (143±3%; P<0.01; n=5) and control peptide-loaded (140±4%; P<0.05; n=4) tissues.
Discussion
CaMKII targeting is known to be an essential aspect of signal transduction in many biological systems,8,9 particularly in neuronal cells, where it is involved in the signaling events that regulate cell division, differentiation, and synaptic plasticity.3,4 This study is the first to demonstrate the importance of CaMKII targeting in vascular smooth muscle signaling. In unstimulated dSMCs, a novel variant of the isoform of CaMKII, CaMKII G-2, is bound to a cytoskeletal scaffold consisting of intermediate filaments and cytosolic dense bodies. Interactions between CaMKII and vimentin/-actinin have been observed in nonmuscle cell types28,29,31 but have not been previously demonstrated in dSMCs. Anchorage of CaMKII G-2 to the cytoskeletal scaffold in unstimulated dSMCs might be necessary for optimal activation of CaMKII G-2 or to ensure that CaMKII G-2 is separated from its substrates in the quiescent dSMC. Tissue differential centrifugation assays ruled out the possibility that the observed binding of CaMKII G-2 to intermediate filaments, and the subsequent phosphorylation of those filaments, serves as a mechanism of regulation of filament disassembly in dSMCs, even though such a mechanism has been reported for cultured cells.28,29 Interestingly, it seems that the network of intermediate filaments within the dSMC is more stable than that of cultured cells. This inherent stability may be important in providing structural support to the dSMC during force generation.
Depolarization causes CaMKII G-2 to become activated, released into the cytosol, and subsequently targeted to cortical dense plaques. It does not appear that CaMKII G-2 translocates to the cortex to form a signaling complex with MEK1/2 and ERK1/2, which are known binding partners of CaMKII in smooth muscle,6,14 because they remain homogeneously distributed. Therefore, there must be an unknown downstream substrate within the cortex to which CaMKII G-2 translocates and binds, which then leads to ERK1 activation, LC20 phosphorylation, and dSMC contraction. Because CaMKII G-2 colocalizes but does not bind with vinculin in depolarized dSMCs, there could be some other protein within the cortical dense plaques, such as -actinin, to which CaMKII G-2 binds. Evidence in the literature also suggests that this unknown substrate could be a nonreceptor tyrosine kinase such as proline-rich tyrosine kinase 2 or a src family member.32
Previously, CaMKII has been shown to be targeted by its catalytic and variable domains and even by the N-terminal association domain sequence adjacent to its variable domain.8,9 This study is the first to demonstrate that CaMKII can be targeted by a sequence that is confined to its association domain. This part of CaMKII has generally been thought to function only in oligomerization. However, CaMKII G-2 is novel among CaMKII variants in that it has a polyproline sequence confined within its association domain that is predicted to contain 2 overlapping SH3-binding domains. Loading dSMCs with a decoy peptide of this sequence not only prevents CaMKII G-2 targeting but also results in significant inhibition of ERK1 phosphorylation as well as contractility. Thus, targeting of CaMKII G-2 to the cortex is an essential aspect of dSMC signaling during contraction.
Previously, a significant decrease in contractility in response to 51mmol/L KCl was measured in aorta tissue treated with noneCvariant-specific CaMKII antisense oligonucleotides.6 Surprisingly, this decrease in contraction is almost identical to that measured in aorta tissue treated with only CaMKII G-2 specific antisense oligonucleotides. Furthermore, the levels of phosphorylated ERK1 and LC20 in CaMKII G-2 antisense-treated tissue are also decreased a similar magnitude to that previously measured in the noneCvariant-specific CaMKII antisense-treated tissue.6 Therefore, the data suggest that either CaMKII G-2 is the primary variant of CaMKII responsible for regulating contraction or that CaMKII G-2 is a component of the majority of heterooligomers formed between the variants of CaMKII. However, 1 significant difference between the current study and the study, which used noneCvariant-specific CaMKII antisense oligonucleotides,6 is that the latter did not observe a greater inhibition of contraction at a lower concentration of KCl. In contrast, in the present study, greater inhibition of contraction was observed at a lower concentration of KCl in CaMKII G-2 antisense-treated tissue. Thus, these combined results suggest that the CaMKII variants may differ in their sensitivity to KCl in the extracellular environment and, hence, intracellular Ca2+ levels, with CaMKII G-2 being relatively more sensitive.
In both antisense-treated and decoy peptide-loaded aorta tissue, the decrease in contractility was significant but relatively small in magnitude. The relatively small inhibition of contractility could be attributable to the fact that antisense treatment only decreased the protein expression of CaMKII G-2 by 50%. It is quite possible that better knockdown of the CaMKII message might provide a larger inhibition of contractility. Furthermore, there are known to be other parallel, redundant pathways that become activated during depolarization and that also regulate dSMC contractility. For example, LC20 can be phosphorylated directly by Ca2+/calmodulin and MLCK in the absence of activated CaMKII.33
Another result found in both the antisense and decoy peptide studies was that phosphorylated ERK1 levels were significantly decreased, whereas no significant changes in phosphorylated ERK2 were detected. This suggests that ERK1 and ERK2 might play different roles in vascular smooth muscle function, with ERK1 playing a role in the depolarization-mediated regulation of contractility, as shown here. In contrast, it is possible that ERK2 plays a relatively greater role in the -adrenergiceCmediated contraction pathway, previously shown to involve ERK1/2 in this same tissue.34,35 Interestingly, the present study determined that ERK1/2 does not translocate during depolarization, but a previous study, using the same cell type, found ERK1/2 to transiently translocate to the cortex during -adrenergic stimulation.23 Additionally, it is important to point out that ERK1/2 is known to be involved in many other dSMC functions besides contraction, including adhesion-dependent signaling,36,37 reactive oxygen species generation, proliferation, and migration,38eC41 and that there is considerable tissue and species diversity in the degree to which ERK1/2 activation regulates smooth muscle contractility. In some cases, a lack of any role for ERK1/2 in the regulation of smooth muscle contractility has been reported.15eC18
In summary, the results presented here demonstrate that a novel variant of the isoform of CaMKII plays a significant role in dSMC signaling during contraction. Furthermore, a polyproline region within the association domain of this novel CaMKII is identified as a targeting domain important to this signaling.
Acknowledgments
Funding provided by NIH grants HL31704, HL42293, and HD43054 (to K.G.M.) and HL074470 (to S.S.G).
References
Hudmon A, Schulman H. Structure-function of the multifunctional Ca2+/calmodulin-dependent protein kinase II. Biochem J. 2002; 364: 593eC611.
Yamauchi T, Fujisawa H. Evidence for three distinct forms of calmodulin-dependent protein kinases from rat brain. FEBS Lett. 1980; 116: 141eC144.
Shen K, Meyer T. Dynamic control of CaMKII translocation and localization in hippocampal neurons by NMDA receptor stimulation. Science. 1999; 284: 162eC166.
Hudmon A, Schulman H. Neuronal Ca2+/calmodulin-dependent protein kinase II: the role of structure and autoregulation in cellular function. Annu Rev Biochem. 2002; 71: 473eC510.
Zhou ZL, Ikebe M. New isoforms of Ca2+/calmodulin-dependent protein kinase II in smooth muscle. Biochem J. 1994; 299: 489eC495.
Kim I, Je HD, Gallant C, Zhan Q, Riper DV, Badwey JA, Singer HA, Morgan KG. Ca2+-calmodulin-dependent protein kinase II-dependent activation of contractility in ferret aorta. J Physiol. 2000; 526 (Pt 2): 367eC74.
Rokolya A, Singer HA. Inhibition of CaM kinase II activation and force maintenance by KN-93 in arterial smooth muscle. Am J Physiol Cell Physiol. 2000; 278: C537eCC545.
Bayer KU, Schulman H. Regulation of signal transduction by protein targeting: the case for CaMKII. Biochem Biophys Res Commun. 2001; 289: 917eC923.
Colbran RJ. Targeting of calcium/calmodulin-dependent protein kinase II. Biochem J. 2004; 378: 1eC16.
Gangopadhyay SS, Barber AL, Gallant C, Grabarek Z, Smith JL, Morgan KG. Differential functional properties of calmodulin-dependent protein kinase IIgamma variants isolated from smooth muscle. Biochem J. 2003; 372: 347eC357.
Gangopadhyay SS, Takizawa N, Gallant C, Barber AL, Je HD, Smith TC, Luna EJ, Morgan KG. Smooth muscle archvillin: a novel regulator of signaling and contractility in vascular smooth muscle. J Cell Sci. 2004; 117: 5043eC5057.
Eriksson JE, He T, Trejo-Skalli AV, Harmala-Brasken AS, Hellman J, Chou YH, Goldman RD. Specific in vivo phosphorylation sites determine the assembly dynamics of vimentin intermediate filaments. J Cell Sci. 2004; 117: 919eC932.
Je HD, Gallant C, Leavis PC, Morgan KG. Caveolin-1 regulates contractility in differentiated vascular smooth muscle. Am J Physiol Heart Circ Physiol. 2004; 286: H91eCH98.
Abraham ST, Benscoter H, Schworer CM, Singer HA. A role for Ca2+/calmoduin-dependent protein kinase II in the mitogen-activated protein kinase signaling cascade of cultured rat aortic vascular smooth muscle cells. Circ Res. 1997; 81: 575eC584.
D’Angelo G, Graceffa P, Wang C-LA, Wrangle J, Adam LP. Mammal-specific, ERK-dependent, caldesmon phosphorylation in smooth muscle. J Biol Chem. 1999; 274: 30115eC30121.
Nixon GF, Iizuka K, Haystead CMM, Haystead TAJ, Somlyo AP, Somlyo AV. Phosphorylation of caldesmon by mitogen-activated protein kinase with no effect on Ca2+ sensitivity in rabbit smooth muscle. J Physiol (Lond). 1995; 487: 283eC289.
Gorenne I, Su X, Moreland RS. Inhibition of p42 and p44 MAP kinase does not alter smooth muscle contraction in swine carotid artery. Am J Physiol. 1998; 44: H131eCH138.
Hedges JC, Oxhorn BC, Carty M, Adam LP, Yamboliev IA, Gerthoffer WT. Phosphorylation of caldesmon by ERK MAP kinases in smooth muscle. Am J Physiol Cell Physiol. 2000; 278: C718eCC726.
Frank ED, Warren L. Aortic smooth muscle cells contain vimentin instead of desmin. Proc Natl Acad Sci U S A. 1981; 78: 3020eC3024.
Bond M, Somlyo AV. Dense bodies and actin polarity in vertebrate smooth muscle. J Cell Biol. 1982; 95: 403eC413.
Kuo KH, Herrera AM, Seow CY. Ultrastructure of airway smooth muscle. Respir Physiol Neurobiol. 2003; 137: 197eC208.
Bagby R. Organization of contractile/cytoskeletal elements. In: Stephens NL, ed. Biochemistry of Smooth Muscle. Boca Raton, FL: CRC Press Inc; 1983: 1eC84.
Khalil RA, Morgan KG. PKC-mediated redistribution of mitogen-activated protein kinase during smooth muscle activation. Am J Physiol. 1993; 265: C406eCC411.
Small JV, Gimona M. The cytoskeleton of the vertebrate smooth muscle cell. Acta Physiol Scand. 1998; 164: 341eC348.
North AJ, Galazkiewicz B, Byers TJ, Glenney JR Jr, Small JV. Complementary distributions of vinculin and dystrophin define two distinct sarcolemma domains in smooth muscle. J Cell Biol. 1993; 120: 1159eC1167.
Small JV. Geometry of actin-membrane attachments in the smooth muscle cell: the localisations of vinculin and alpha-actinin. EMBO J. 1985; 4: 45eC49.
Ando S, Tokui T, Yamauchi T, Sugiura H, Tanabe K, Inagaki M. Evidence that Ser-82 is a unique phosphorylation site on vimentin for Ca2(+)-calmodulin-dependent protein kinase II. Biochem Biophys Res Commun. 1991; 175: 955eC962.
Ogawara M, Inagaki N, Tsujimura K, Takai Y, Sekimata M, Ha MH, Imajoh-Ohmi S, Hirai S, Ohno S, Sugiura H, Yamauchi T, Inagaki M. Differential targeting of protein kinase C and CaM kinase II signalings to vimentin. J Cell Biol. 1995; 131: 1055eC1066.
Inagaki N, Goto H, Ogawara M, Nishi Y, Ando S, Inagaki M. Spatial patterns of Ca2+ signals define intracellular distribution of a signaling by Ca2+/Calmodulin-dependent protein kinase II. J Biol Chem. 1997; 272: 25195eC25199.
Sin WC, Chen XQ, Leung T, Lim L. RhoA-binding kinase alpha translocation is facilitated by the collapse of the vimentin intermediate filament network. Mol Cell Biol. 1998; 18: 6325eC6339.
Dhavan R, Greer PL, Morabito MA, Orlando LR, Tsai LH. The cyclin-dependent kinase 5 activators p35 and p39 interact with the alpha-subunit of Ca2+/calmodulin-dependent protein kinase II and alpha-actinin-1 in a calcium-dependent manner. J Neurosci. 2002; 22: 7879eC7891.
Ginnan R, Singer HA. CaM kinase II-dependent activation of tyrosine kinases and ERK1/2 in vascular smooth muscle. Am J Physiol Cell Physiol. 2002; 282: C754eCC761.
Horowitz A, Menice CB, Laporte R, Morgan KG. Mechanisms of smooth muscle contraction. Physiol Rev. 1996; 76: 967eC1003.
Dessy C, Kim I, Sougnez CL, Laporte R, Morgan KG. A role for MAP kinase in differentiated smooth muscle contraction evoked by a-adrenoceptor stimulation. Am J Physiol Cell Physiol. 1998; 275: C1081eCC1086.
Khalil RA, Menice CB, Wang C-LA, Morgan KG. Phosphotyrosine-dependent targeting of mitogen-activated protein kinase in differentiated contractile vascular cells. Circ Res. 1995; 76: 1101eC1108.
Lu KK, Armstrong SE, Ginnan R, Singer HA. Adhesion-dependent activation of CaM Kinase II and regulation of ERK activation in vascular smooth muscle. Am J Physiol Cell Physiol. 2005:June 15;Epub.
Sundberg-Smith LJ, Doherty JT, Mack CP, Taylor JM. Adhesion stimulates direct PAK1/ERK2 association and leads to ERK-dependent PAK1 Thr212 phosphorylation. J Biol Chem. 2005; 280: 2055eC2064.
Lo IC, Shih JM, Jiang MJ. Reactive oxygen species and ERK 1/2 mediate monocyte chemotactic protein-1-stimulated smooth muscle cell migration. J Biomed Sci. 2005; 12: 377eC388.
Blaschke F, Stawowy P, Kappert K, Goetze S, Kintscher U, Wollert-Wulf B, Fleck E, Graf K. Angiotensin II-augmented migration of VSMCs towards PDGF-BB involves Pyk2 and ERK 1/2 activation. Basic Res Cardiol. 2002; 97: 334eC342.
Lee JH, Johnson PR, Roth M, Hunt NH, Black JL. ERK activation and mitogenesis in human airway smooth muscle cells. Am J Physiol Lung Cell Mol Physiol. 2001; 280: L1019eCL1029.
Yamboliev IA, Gerthoffer WT. Modulatory role of ERK MAPK-caldesmon pathway in PDGF-stimulated migration of cultured pulmonary artery SMCs. Am J Physiol Cell Physiol. 2001; 280: C1680eCC1688.(William A. Marganski, Sam)
the Department of Medicine (K.G.M.), Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Mass.
Abstract
Subcellular targeting of kinases controls their activation and access to substrates. Although Ca2+/calmodulin-dependent protein kinase II (CaMKII) is known to regulate differentiated smooth muscle cell (dSMC) contractility, the importance of targeting in this regulation is not clear. The present study investigated the function in dSMCs of a novel variant of the isoform of CaMKII that contains a potential targeting sequence in its association domain (CaMKII G-2). Antisense knockdown of CaMKII G-2 inhibited extracellular signal-related kinase (ERK) activation, myosin phosphorylation, and contractile force in dSMCs. Confocal colocalization analysis revealed that in unstimulated dSMCs CaMKII G-2 is bound to a cytoskeletal scaffold consisting of interconnected vimentin intermediate filaments and cytosolic dense bodies. On activation with a depolarizing stimulus, CaMKII G-2 is released into the cytosol and subsequently targeted to cortical dense plaques. Comparison of phosphorylation and translocation time courses indicates that, after CaMKII G-2 activation, and before CaMKII G-2 translocation, vimentin is phosphorylated at a CaMKII-specific site. Differential centrifugation demonstrated that phosphorylation of vimentin in dSMCs is not sufficient to cause its disassembly, in contrast to results in cultured cells. Loading dSMCs with a decoy peptide containing the polyproline sequence within the association domain of CaMKII G-2 inhibited targeting. Furthermore, prevention of CaMKII G-2 targeting led to significant inhibition of ERK activation as well as contractility. Thus, for the first time, this study demonstrates the importance of CaMKII targeting in dSMC signaling and identifies a novel targeting function for the association domain in addition to its known role in oligomerization.
Key Words: CaMKII smooth muscle contractility targeting extracellular signal-regulated kinase
Introduction
Ca2+/calmodulin-dependent protein kinase II (CaMKII) is a Ser/Thr kinase that is expressed as 4 different isoforms (, , , and ). Each member of this kinase family possesses 3 major domains (Figure 1A): an N-terminal catalytic/regulatory domain containing a Ser/Thr kinase region and overlapping autoinhibitory and Ca2+/calmodulin binding regions, a central-linker domain containing variable regions, and a C-terminal association domain that promotes oligomerization.1 CaMKII is not active until Ca2+/calmodulin binds to its regulatory domain. This binding allows CaMKII to undergo autophosphorylation at Thr287 (numbering according to the isoform), which, in turn, increases its affinity for Ca2+/calmodulin and allows kinase activity even in the absence of Ca2+/calmodulin.
The and isoforms of CaMKII are predominantly expressed in neural tissue,2 where they have been demonstrated to be involved in synaptic plasticity, memory, and learning.3,4 In contrast, differentiated smooth muscle cells (dSMCs) express mainly the isoform of CaMKII,5 which has been linked to contractile activity by the use of antisense and pharmacological inhibitors.6,7
It is well known that subcellular localization and targeting of CaMKII are important in regulating signal transduction in many biological systems.8,9 For instance, in neuronal cells N-methyl-D-aspartate receptor stimulation causes the and isoforms of CaMKII to change from an F-actin bound state to a postsynaptic membrane bound state. This targeting is thought to increase synaptic strength because it brings CaMKII closer to its substrates, including ion channels and signaling proteins.3,4
The present study investigated the function of a novel variant of the isoform of CaMKII (CaMKII G-2; Figure 1A), recently cloned from an aorta tissue library,10 in dSMC contractility and signaling. The results obtained indicate that CaMKII G-2 is expressed as a protein in dSMCs and plays a significant role in dSMC signaling during contraction. In unstimulated dSMCs, CaMKII G-2 is bound to a cytoskeletal scaffold consisting of interconnected vimentin intermediate filaments and cytosolic dense bodies. On activation, CaMKII G-2 phosphorylates vimentin and is released into the cytosol, where it is subsequently targeted to cortical dense plaques. Targeting of CaMKII G-2 is dependent on a polyproline region confined within the last 99 residues of the association domain and is essential to dSMC signaling because it leads to ERK (extracellular signal-regulated kinase) 1 activation, LC20 phosphorylation, and contraction.
Materials and Methods
Fifty-three ferrets were euthanized with chloroform as approved by the Institutional Care and Use Committee. As previously described,6 the aorta was dissected, and then the tissue strips were attached to a force transducer, treated with a depolarizing stimulus (51 mmol/L KCl), and quick-frozen. Phosphorothioate oligonucleotides were synthesized to be complementary to residues 515 to 521 of the association domain of CaMKII G-2 and were loaded into the aorta tissue following a previously published protocol.6 A decoy peptide was designed against residues 571 to 586 of CaMKII G-2, which are predicted to contain 2 overlapping Src homology 3 (SH3)-binding domains,10 and was chemically loaded into the aorta tissue. Quick-frozen tissue samples were homogenized and analyzed via Western blot, as previously described.6 Overlay assays (also known as far Western blots) were performed as previously described.11 The overlay buffer contained 40 e/mL recombinant CaMKII G-2 association domain protein. The polymer-to-subunit ratio of vimentin in aorta tissue and COS-7 cells was measured using a previously published differential centrifugation method.12 dSMCs from aorta tissue were enzymatically isolated, paraformaldehyde fixed, and stained as previously described.13 Images were acquired with a Kr/Ar laser (Bio-Rad Radiance 2000) scanning confocal microscope equipped with a Nikon X-60 (NA1.4) oil immersion objective. The degree of colocalization between a pair of immunostained proteins was determined by computing the percentage of the total pixels that contain both fluorophores given predetermined thresholds.
An expanded Materials and Methods section, with additional details about the procedures, materials used, and data analysis methods, can be found in the online data supplement at http://circres.ahajournals.org.
Results
Tissue Expression of CaMKII G-2
Because the novel CaMKII G-2 variant has been previously described only as a cDNA isolated from an aorta library,10 the question arises whether it is expressed as a protein. To clarify this issue, an antibody that recognizes the association domain of CaMKII G-2 was generated and tested on aorta tissue whole-cell homogenate. The specificity of the antibody was validated by transiently transfecting COS-7 cells to express either CaMKII G-2 or a similar but distinct CaMKII variant, CaMKII G-1 (Figure 1A). 10 The antieCCaMKII G-2 antibody recognized a single band at the expected molecular mass of 62 kDa in lysates of CaMKII G-2eCtransfected COS-7 cells but recognized no bands in lysates of CaMKII G-1eCtransfected COS-7 cells (Figure 1B). Furthermore, the antieCCaMKII G-2 antibody recognized a single band at 62 kDa in aorta whole-cell homogenate, thereby demonstrating adequate specificity for imaging experiments. Interestingly, CaMKII G-2 expression was detectable in protein matched homogenates of several smooth muscles, as well as the heart and brain, but was essentially absent from skeletal muscle and the liver (Figure 1C).
CaMKII G-2 Specific Antisense Oligonucleotides Inhibit the Signaling Events That Regulate Contractility
To determine whether CaMKII G-2 is involved in dSMC signaling during contraction, aorta tissue was loaded with antisense oligonucleotides designed against part of the sequence that codes for the novel association domain. CaMKII G-2 protein expression in antisense-treated tissue was significantly decreased compared with both sham and random sequence-treated tissues (Figure 2A). The effect of decreased CaMKII G-2 protein expression on contractility was determined by stimulating the antisense-treated tissue with a depolarizing stimulus. The contractile response to a depolarizing stimulus was significantly less in antisense-treated tissue compared with both sham and random sequence-treated tissues, with the effect being relatively more pronounced at a lower level of depolarization (Figure 2B).
In some,6,14 but not all smooth muscles,15eC18 CaMKII-dependent activation of ERK1/2 has been shown to be essential in a pathway that leads to LC20 phosphorylation and dSMC contraction. In antisense-treated tissue quick-frozen after exposure to a depolarizing stimulus, ERK1 phosphorylation was significantly less than that detected in both depolarized sham and depolarized random sequence-treated tissues (Figure 2C). Interestingly, a significant decrease in the corresponding phosphorylated ERK2 signal was not obtained. LC20 phosphorylation in depolarized antisense-treated tissue was significantly less than that detected in both depolarized sham and depolarized random sequence-treated tissues (Figure 2D).
CaMKII G-2 Colocalizes With Vimentin and -Actinin in Unstimulated dSMCs
To elucidate whether CaMKII G-2 is targeted to specific proteins, unstimulated dSMCs were fixed and immunostained for CaMKII G-2. Confocal microscopy indicated that CaMKII G-2 was localized primarily to filamentous structures throughout the dSMC (Figure 3A). Because some isoforms of CaMKII are known to bind to F-actin,3,4,9 dSMCs were costained for CaMKII G-2 and actin (Figure 3B). Although there were a few regions where CaMKII G-2 and actin overlapped, in general, the 2 proteins did not appear to colocalize (Figure 3C).
Unstimulated dSMCs were also costained for CaMKII G-2 and vimentin (Figure 3D through 3F), the primary intermediate filament protein expressed in aortic vascular smooth muscle.19Merged images (Figure 3F) illustrated that CaMKII G-2 (Figure 3D) and vimentin (Figure 3E) highly colocalized. However, a number of green spots where CaMKII G-2 and vimentin did not colocalize were consistently observed (white arrows in Figure 3F). These spots were 0.5 e and were closely associated with the vimentin filament network, which are both characteristic properties of cytosolic dense bodies.20eC22 Therefore, unstimulated dSMCs were also costained for CaMKII G-2 and -actinin (Figure 3G through 3I), a known dense body protein.22 Merged images (Figure 3I) illustrated that CaMKII G-2 (Figure 3G) and -actinin (Figure 3H) highly colocalized, except for a few regions (white arrows) where presumably CaMKII G-2 colocalized with vimentin.
The degree of colocalization between CaMKII G-2 and the cytoskeletal components was quantified. The relatively low degree of colocalization between CaMKII G-2 and actin in unstimulated dSMCs was confirmed because CG-2eCactin=0.30±0.02 and CactineCG-2=0.31±0.02 for 16 cells. In contrast, the colocalization coefficients measured for CaMKII G-2 and either vimentin or -actinin in unstimulated dSMCs were both significantly higher (P<0.001; CG-2eCvimentin=0.63±0.02 and CvimentineCG-2=0.67±0.01 for 12 cells; CG-2eC-actinin=0.65±0.02 and C-actinineCG-2=0.64±0.02 for 10 cells).
The Subcellular Location of CaMKII G-2 Is Dynamically Regulated by Depolarization
To see whether the localization of CaMKII G-2 to vimentin and -actinin was constitutive or regulated, dissociated dSMCs were depolarized for up to 10 minutes to activate CaMKII G-2 and then immunostained for CaMKII G-2. After 10 minutes of depolarization, CaMKII G-2 was relatively depleted from the cytosol and was concentrated at regions near the cell cortex (Figure 4A and 4B). The changes in targeting were quantified by measuring the cortex/cytosol fluorescence ratio of CaMKII G-2 at different time points during depolarization. The ratio increased significantly after 2 minutes of depolarization, reaching a value of 4 at 10 minutes (Figure 4C).
The Subcellular Locations of ERK1/2 and Mitogen-Activated Protein Kinase Kinase 1/2 Do Not Change During Depolarization
Previously, ERK1/2 has been shown to be downstream of CaMKII in the depolarization-mediated activation pathway.6,14 ERK1/2 has also been shown to translocate to a cortical signaling complex in dSMCs stimulated with an -adrenergic agonist.23 Therefore, to determine whether CaMKII G-2 translocates to increase its proximity to a cortical signaling complex during depolarization-mediated activation, dSMCs were immunostained for mitogen-activated protein kinase kinase (MEK) 1/2 and ERK1/2. Line scan analysis revealed that MEK1/2 and ERK1/2 were homogeneously distributed throughout the unstimulated dSMC, and, furthermore, these distributions remain unchanged during depolarization (Figure 4D). This demonstrates that although MEK1/2 and ERK1/2 are involved in both the depolarization- and -adrenergiceCactivation pathways, they are differentially targeted by different stimuli.
CaMKII G-2 Is Targeted to Dense Plaques in Depolarized dSMCs
CaMKII G-2 is targeted to cytoplasmic dense bodies in unstimulated dSMCs. To determine whether CaMKII G-2 is targeted to cortical dense plaques (which also contain -actinin) after stimulation, depolarized dSMCs were costained for CaMKII G-2 and the dense plaque protein vinculin.24 Merged images (Figure 4G and 4J) illustrated that CaMKII G-2 (Figure 4E and 4H) and vinculin (Figure 4F and 4I) highly colocalized. Furthermore, the cortical dense plaques formed a rib-like pattern (white arrows in Figure 4J) on the surface of the dSMC, as has been previously reported.25,26 The high degree to which CaMKII G-2 colocalized with vinculin in depolarized dSMCs was confirmed by quantifying the colocalization coefficients (CG-2eCvinculin=0.62±0.03 and CvinculineCG-2=0.62±0.03 for 12 cells).
CaMKII G-2 Binds Directly to Vimentin and -Actinin
Overlay assays were performed to determine whether CaMKII G-2 can bind directly to vimentin, -actinin, or vinculin. Significant binding of recombinant CaMKII G-2eCassociation domain protein to -actinin and vimentin in the whole-cell homogenate was found by immunoblotting (Figure 5A). In contrast, there was no significant binding between CaMKII G-2 and vinculin (Figure 5A). The degree of binding was quantified by measuring the densitometry of the overlay band normalized to the densitometry of the Napthol Blue Black protein stain on the same membrane (Figure 5B). Binding between CaMKII G-2 and either vimentin or -actinin was significantly higher than binding between CaMKII G-2 and vinculin. As a negative control, the binding between CaMKII G-2 and BSA (protein matched to the amount of vimentin in the homogenate) was analyzed and found to be negligible.
CaMKII Autophosphorylation and Vimentin Phosphorylation Precede CaMKII G-2 Translocation
To begin to address the mechanism of CaMKII G-2 translocation, the time course of CaMKII Thr287 autophosphorylation during depolarization was monitored and compared with the time course of CaMKII G-2 translocation. CaMKII autophosphorylation increased significantly above baseline at 0.5 minute of depolarization and reached 250% baseline by 10 minutes (Figure 6A). Therefore, increases in CaMKII autophosphorylation occurred several minutes before the onset of CaMKII G-2 translocation (Figure 4C).
Because activated CaMKII has been reported to phosphorylate intermediate filaments, vimentin phosphorylation was measured at a CaMKII-specific site.27 Vimentin phosphorylation increased significantly above baseline at 2 minutes of depolarization and reached 200% baseline by 10 minutes (Figure 6B). Therefore, increases in vimentin phosphorylation occurred after CaMKII autophosphorylation (Figure 6A) and were either coincident with or occurred slightly before CaMKII G-2 translocation (Figure 4C).
Because phosphorylation of vimentin at the CaMKII-specific site could be caused by any of the 6 variants of CaMKII, the level of vimentin phosphorylation was measured in depolarized CaMKII G-2 antisense-treated tissue (Figure 6C). Vimentin phosphorylation in depolarized antisense-treated tissue was significantly less than that measured in both depolarized sham and depolarized random sequence-treated tissues. Thus, the increase in vimentin phosphorylation during depolarization was at least partially attributable to the kinase activity of CaMKII G-2.
Depolarization Does Not Cause Vimentin Disassembly in dSMCs
The assembly/disassembly dynamics of vimentin intermediate filaments in cultured cells are known to be regulated by several kinases including CaMKII, RhoA-binding kinase , protein kinase A, and protein kinase C.12,28eC30 To test whether the stability of the vimentin filament network is regulated by CaMKII-mediated phosphorylation in dSMCs, the polymer-to-subunit ratio of vimentin was measured in aorta tissue and for comparison in COS-7 cells. As a positive control, cells were treated with calyculin-A, which is a type-1 and type-2A protein phosphatase inhibitor that has been shown to depolymerize at least 50% of the total vimentin pool in cultured BHK-21 cells.12 Unstimulated COS-7 cells contained an abundance of polymerized vimentin compared with the amount of soluble subunits (Figure 6D). Treatment of the COS-7 cells with calyculin-A caused disassembly of vimentin, as indicated by a relative increase in the soluble fraction. In unstimulated dSMCs, the majority of vimentin was also found in the polymerized fraction (Figure 6E). However, depolarization of dSMCs did not change the fractional distribution of vimentin. Interestingly, calyculin-A also did not affect the distribution of vimentin in dSMCs. Thus, the vimentin filament network in dSMCs appears to be more structurally stable than that of cultured COS-7 or BHK-21 cells.
CaMKII G-2 Targeting Is Controlled by a Polyproline Sequence Confined Within the Association Domain
Because vimentin depolymerization does not explain the redistribution of CaMKII G-2 during depolarization, it is possible that phosphorylation of vimentin simply allows CaMKII G-2 to be released and targeted elsewhere. One potential targeting sequence within CaMKII G-2 is a polyproline region within the association domain that is predicted to contain 2 overlapping SH3-binding domains (Figure 1A). This hypothesis was tested by synthesizing a decoy peptide of this region and loading the peptide into dSMCs to determine its effect on targeting and CaMKII G-2eCdependent signaling. Imaging of a chemically loaded dSMC demonstrated that the decoy peptide was distributed throughout the cell (Figure 7A). Furthermore, this type of distribution was observed in both unstimulated and depolarized dSMCs.
To determine whether the decoy peptide had any effect on CaMKII G-2 targeting, the distribution of CaMKII G-2 was measured in depolarized dSMCs that lacked a peptide, were loaded with the decoy peptide, or were loaded with a control peptide. The decoy peptide prevented CaMKII G-2 translocation during depolarization, whereas the control peptide had no effect (Figure 7B). The cortex/cytosol ratio of CaMKII G-2 fluorescence was significantly smaller in depolarized decoy peptide-loaded dSMCs compared with both depolarized sham and depolarized control peptide-loaded dSMCs (Figure 7C).
Inhibiting CaMKII G-2 translocation also negatively affected downstream signaling because phosphorylated ERK1 levels in depolarized decoy peptide-loaded tissue were significantly less than that measured in both depolarized sham and depolarized control peptide-loaded tissues (Figure 7D). Interestingly, the decoy peptide had no significant effect on phosphorylated ERK2 levels. Furthermore, there was a small but significant decrease in the contractile response (computed by normalizing the KCl-induced contraction measured after peptide loading with the KCl-induced contraction measured before peptide loading) to a depolarizing stimulus (51 mmol/L KCl for 5 minutes) in decoy peptide-loaded tissue (121±5%; n=5) compared with both sham (143±3%; P<0.01; n=5) and control peptide-loaded (140±4%; P<0.05; n=4) tissues.
Discussion
CaMKII targeting is known to be an essential aspect of signal transduction in many biological systems,8,9 particularly in neuronal cells, where it is involved in the signaling events that regulate cell division, differentiation, and synaptic plasticity.3,4 This study is the first to demonstrate the importance of CaMKII targeting in vascular smooth muscle signaling. In unstimulated dSMCs, a novel variant of the isoform of CaMKII, CaMKII G-2, is bound to a cytoskeletal scaffold consisting of intermediate filaments and cytosolic dense bodies. Interactions between CaMKII and vimentin/-actinin have been observed in nonmuscle cell types28,29,31 but have not been previously demonstrated in dSMCs. Anchorage of CaMKII G-2 to the cytoskeletal scaffold in unstimulated dSMCs might be necessary for optimal activation of CaMKII G-2 or to ensure that CaMKII G-2 is separated from its substrates in the quiescent dSMC. Tissue differential centrifugation assays ruled out the possibility that the observed binding of CaMKII G-2 to intermediate filaments, and the subsequent phosphorylation of those filaments, serves as a mechanism of regulation of filament disassembly in dSMCs, even though such a mechanism has been reported for cultured cells.28,29 Interestingly, it seems that the network of intermediate filaments within the dSMC is more stable than that of cultured cells. This inherent stability may be important in providing structural support to the dSMC during force generation.
Depolarization causes CaMKII G-2 to become activated, released into the cytosol, and subsequently targeted to cortical dense plaques. It does not appear that CaMKII G-2 translocates to the cortex to form a signaling complex with MEK1/2 and ERK1/2, which are known binding partners of CaMKII in smooth muscle,6,14 because they remain homogeneously distributed. Therefore, there must be an unknown downstream substrate within the cortex to which CaMKII G-2 translocates and binds, which then leads to ERK1 activation, LC20 phosphorylation, and dSMC contraction. Because CaMKII G-2 colocalizes but does not bind with vinculin in depolarized dSMCs, there could be some other protein within the cortical dense plaques, such as -actinin, to which CaMKII G-2 binds. Evidence in the literature also suggests that this unknown substrate could be a nonreceptor tyrosine kinase such as proline-rich tyrosine kinase 2 or a src family member.32
Previously, CaMKII has been shown to be targeted by its catalytic and variable domains and even by the N-terminal association domain sequence adjacent to its variable domain.8,9 This study is the first to demonstrate that CaMKII can be targeted by a sequence that is confined to its association domain. This part of CaMKII has generally been thought to function only in oligomerization. However, CaMKII G-2 is novel among CaMKII variants in that it has a polyproline sequence confined within its association domain that is predicted to contain 2 overlapping SH3-binding domains. Loading dSMCs with a decoy peptide of this sequence not only prevents CaMKII G-2 targeting but also results in significant inhibition of ERK1 phosphorylation as well as contractility. Thus, targeting of CaMKII G-2 to the cortex is an essential aspect of dSMC signaling during contraction.
Previously, a significant decrease in contractility in response to 51mmol/L KCl was measured in aorta tissue treated with noneCvariant-specific CaMKII antisense oligonucleotides.6 Surprisingly, this decrease in contraction is almost identical to that measured in aorta tissue treated with only CaMKII G-2 specific antisense oligonucleotides. Furthermore, the levels of phosphorylated ERK1 and LC20 in CaMKII G-2 antisense-treated tissue are also decreased a similar magnitude to that previously measured in the noneCvariant-specific CaMKII antisense-treated tissue.6 Therefore, the data suggest that either CaMKII G-2 is the primary variant of CaMKII responsible for regulating contraction or that CaMKII G-2 is a component of the majority of heterooligomers formed between the variants of CaMKII. However, 1 significant difference between the current study and the study, which used noneCvariant-specific CaMKII antisense oligonucleotides,6 is that the latter did not observe a greater inhibition of contraction at a lower concentration of KCl. In contrast, in the present study, greater inhibition of contraction was observed at a lower concentration of KCl in CaMKII G-2 antisense-treated tissue. Thus, these combined results suggest that the CaMKII variants may differ in their sensitivity to KCl in the extracellular environment and, hence, intracellular Ca2+ levels, with CaMKII G-2 being relatively more sensitive.
In both antisense-treated and decoy peptide-loaded aorta tissue, the decrease in contractility was significant but relatively small in magnitude. The relatively small inhibition of contractility could be attributable to the fact that antisense treatment only decreased the protein expression of CaMKII G-2 by 50%. It is quite possible that better knockdown of the CaMKII message might provide a larger inhibition of contractility. Furthermore, there are known to be other parallel, redundant pathways that become activated during depolarization and that also regulate dSMC contractility. For example, LC20 can be phosphorylated directly by Ca2+/calmodulin and MLCK in the absence of activated CaMKII.33
Another result found in both the antisense and decoy peptide studies was that phosphorylated ERK1 levels were significantly decreased, whereas no significant changes in phosphorylated ERK2 were detected. This suggests that ERK1 and ERK2 might play different roles in vascular smooth muscle function, with ERK1 playing a role in the depolarization-mediated regulation of contractility, as shown here. In contrast, it is possible that ERK2 plays a relatively greater role in the -adrenergiceCmediated contraction pathway, previously shown to involve ERK1/2 in this same tissue.34,35 Interestingly, the present study determined that ERK1/2 does not translocate during depolarization, but a previous study, using the same cell type, found ERK1/2 to transiently translocate to the cortex during -adrenergic stimulation.23 Additionally, it is important to point out that ERK1/2 is known to be involved in many other dSMC functions besides contraction, including adhesion-dependent signaling,36,37 reactive oxygen species generation, proliferation, and migration,38eC41 and that there is considerable tissue and species diversity in the degree to which ERK1/2 activation regulates smooth muscle contractility. In some cases, a lack of any role for ERK1/2 in the regulation of smooth muscle contractility has been reported.15eC18
In summary, the results presented here demonstrate that a novel variant of the isoform of CaMKII plays a significant role in dSMC signaling during contraction. Furthermore, a polyproline region within the association domain of this novel CaMKII is identified as a targeting domain important to this signaling.
Acknowledgments
Funding provided by NIH grants HL31704, HL42293, and HD43054 (to K.G.M.) and HL074470 (to S.S.G).
References
Hudmon A, Schulman H. Structure-function of the multifunctional Ca2+/calmodulin-dependent protein kinase II. Biochem J. 2002; 364: 593eC611.
Yamauchi T, Fujisawa H. Evidence for three distinct forms of calmodulin-dependent protein kinases from rat brain. FEBS Lett. 1980; 116: 141eC144.
Shen K, Meyer T. Dynamic control of CaMKII translocation and localization in hippocampal neurons by NMDA receptor stimulation. Science. 1999; 284: 162eC166.
Hudmon A, Schulman H. Neuronal Ca2+/calmodulin-dependent protein kinase II: the role of structure and autoregulation in cellular function. Annu Rev Biochem. 2002; 71: 473eC510.
Zhou ZL, Ikebe M. New isoforms of Ca2+/calmodulin-dependent protein kinase II in smooth muscle. Biochem J. 1994; 299: 489eC495.
Kim I, Je HD, Gallant C, Zhan Q, Riper DV, Badwey JA, Singer HA, Morgan KG. Ca2+-calmodulin-dependent protein kinase II-dependent activation of contractility in ferret aorta. J Physiol. 2000; 526 (Pt 2): 367eC74.
Rokolya A, Singer HA. Inhibition of CaM kinase II activation and force maintenance by KN-93 in arterial smooth muscle. Am J Physiol Cell Physiol. 2000; 278: C537eCC545.
Bayer KU, Schulman H. Regulation of signal transduction by protein targeting: the case for CaMKII. Biochem Biophys Res Commun. 2001; 289: 917eC923.
Colbran RJ. Targeting of calcium/calmodulin-dependent protein kinase II. Biochem J. 2004; 378: 1eC16.
Gangopadhyay SS, Barber AL, Gallant C, Grabarek Z, Smith JL, Morgan KG. Differential functional properties of calmodulin-dependent protein kinase IIgamma variants isolated from smooth muscle. Biochem J. 2003; 372: 347eC357.
Gangopadhyay SS, Takizawa N, Gallant C, Barber AL, Je HD, Smith TC, Luna EJ, Morgan KG. Smooth muscle archvillin: a novel regulator of signaling and contractility in vascular smooth muscle. J Cell Sci. 2004; 117: 5043eC5057.
Eriksson JE, He T, Trejo-Skalli AV, Harmala-Brasken AS, Hellman J, Chou YH, Goldman RD. Specific in vivo phosphorylation sites determine the assembly dynamics of vimentin intermediate filaments. J Cell Sci. 2004; 117: 919eC932.
Je HD, Gallant C, Leavis PC, Morgan KG. Caveolin-1 regulates contractility in differentiated vascular smooth muscle. Am J Physiol Heart Circ Physiol. 2004; 286: H91eCH98.
Abraham ST, Benscoter H, Schworer CM, Singer HA. A role for Ca2+/calmoduin-dependent protein kinase II in the mitogen-activated protein kinase signaling cascade of cultured rat aortic vascular smooth muscle cells. Circ Res. 1997; 81: 575eC584.
D’Angelo G, Graceffa P, Wang C-LA, Wrangle J, Adam LP. Mammal-specific, ERK-dependent, caldesmon phosphorylation in smooth muscle. J Biol Chem. 1999; 274: 30115eC30121.
Nixon GF, Iizuka K, Haystead CMM, Haystead TAJ, Somlyo AP, Somlyo AV. Phosphorylation of caldesmon by mitogen-activated protein kinase with no effect on Ca2+ sensitivity in rabbit smooth muscle. J Physiol (Lond). 1995; 487: 283eC289.
Gorenne I, Su X, Moreland RS. Inhibition of p42 and p44 MAP kinase does not alter smooth muscle contraction in swine carotid artery. Am J Physiol. 1998; 44: H131eCH138.
Hedges JC, Oxhorn BC, Carty M, Adam LP, Yamboliev IA, Gerthoffer WT. Phosphorylation of caldesmon by ERK MAP kinases in smooth muscle. Am J Physiol Cell Physiol. 2000; 278: C718eCC726.
Frank ED, Warren L. Aortic smooth muscle cells contain vimentin instead of desmin. Proc Natl Acad Sci U S A. 1981; 78: 3020eC3024.
Bond M, Somlyo AV. Dense bodies and actin polarity in vertebrate smooth muscle. J Cell Biol. 1982; 95: 403eC413.
Kuo KH, Herrera AM, Seow CY. Ultrastructure of airway smooth muscle. Respir Physiol Neurobiol. 2003; 137: 197eC208.
Bagby R. Organization of contractile/cytoskeletal elements. In: Stephens NL, ed. Biochemistry of Smooth Muscle. Boca Raton, FL: CRC Press Inc; 1983: 1eC84.
Khalil RA, Morgan KG. PKC-mediated redistribution of mitogen-activated protein kinase during smooth muscle activation. Am J Physiol. 1993; 265: C406eCC411.
Small JV, Gimona M. The cytoskeleton of the vertebrate smooth muscle cell. Acta Physiol Scand. 1998; 164: 341eC348.
North AJ, Galazkiewicz B, Byers TJ, Glenney JR Jr, Small JV. Complementary distributions of vinculin and dystrophin define two distinct sarcolemma domains in smooth muscle. J Cell Biol. 1993; 120: 1159eC1167.
Small JV. Geometry of actin-membrane attachments in the smooth muscle cell: the localisations of vinculin and alpha-actinin. EMBO J. 1985; 4: 45eC49.
Ando S, Tokui T, Yamauchi T, Sugiura H, Tanabe K, Inagaki M. Evidence that Ser-82 is a unique phosphorylation site on vimentin for Ca2(+)-calmodulin-dependent protein kinase II. Biochem Biophys Res Commun. 1991; 175: 955eC962.
Ogawara M, Inagaki N, Tsujimura K, Takai Y, Sekimata M, Ha MH, Imajoh-Ohmi S, Hirai S, Ohno S, Sugiura H, Yamauchi T, Inagaki M. Differential targeting of protein kinase C and CaM kinase II signalings to vimentin. J Cell Biol. 1995; 131: 1055eC1066.
Inagaki N, Goto H, Ogawara M, Nishi Y, Ando S, Inagaki M. Spatial patterns of Ca2+ signals define intracellular distribution of a signaling by Ca2+/Calmodulin-dependent protein kinase II. J Biol Chem. 1997; 272: 25195eC25199.
Sin WC, Chen XQ, Leung T, Lim L. RhoA-binding kinase alpha translocation is facilitated by the collapse of the vimentin intermediate filament network. Mol Cell Biol. 1998; 18: 6325eC6339.
Dhavan R, Greer PL, Morabito MA, Orlando LR, Tsai LH. The cyclin-dependent kinase 5 activators p35 and p39 interact with the alpha-subunit of Ca2+/calmodulin-dependent protein kinase II and alpha-actinin-1 in a calcium-dependent manner. J Neurosci. 2002; 22: 7879eC7891.
Ginnan R, Singer HA. CaM kinase II-dependent activation of tyrosine kinases and ERK1/2 in vascular smooth muscle. Am J Physiol Cell Physiol. 2002; 282: C754eCC761.
Horowitz A, Menice CB, Laporte R, Morgan KG. Mechanisms of smooth muscle contraction. Physiol Rev. 1996; 76: 967eC1003.
Dessy C, Kim I, Sougnez CL, Laporte R, Morgan KG. A role for MAP kinase in differentiated smooth muscle contraction evoked by a-adrenoceptor stimulation. Am J Physiol Cell Physiol. 1998; 275: C1081eCC1086.
Khalil RA, Menice CB, Wang C-LA, Morgan KG. Phosphotyrosine-dependent targeting of mitogen-activated protein kinase in differentiated contractile vascular cells. Circ Res. 1995; 76: 1101eC1108.
Lu KK, Armstrong SE, Ginnan R, Singer HA. Adhesion-dependent activation of CaM Kinase II and regulation of ERK activation in vascular smooth muscle. Am J Physiol Cell Physiol. 2005:June 15;Epub.
Sundberg-Smith LJ, Doherty JT, Mack CP, Taylor JM. Adhesion stimulates direct PAK1/ERK2 association and leads to ERK-dependent PAK1 Thr212 phosphorylation. J Biol Chem. 2005; 280: 2055eC2064.
Lo IC, Shih JM, Jiang MJ. Reactive oxygen species and ERK 1/2 mediate monocyte chemotactic protein-1-stimulated smooth muscle cell migration. J Biomed Sci. 2005; 12: 377eC388.
Blaschke F, Stawowy P, Kappert K, Goetze S, Kintscher U, Wollert-Wulf B, Fleck E, Graf K. Angiotensin II-augmented migration of VSMCs towards PDGF-BB involves Pyk2 and ERK 1/2 activation. Basic Res Cardiol. 2002; 97: 334eC342.
Lee JH, Johnson PR, Roth M, Hunt NH, Black JL. ERK activation and mitogenesis in human airway smooth muscle cells. Am J Physiol Lung Cell Mol Physiol. 2001; 280: L1019eCL1029.
Yamboliev IA, Gerthoffer WT. Modulatory role of ERK MAPK-caldesmon pathway in PDGF-stimulated migration of cultured pulmonary artery SMCs. Am J Physiol Cell Physiol. 2001; 280: C1680eCC1688.(William A. Marganski, Sam)