Characterization of smooth muscle caldesmon as a microtubule-associated protein

Cytoskeletal mechanisms regulating vascular endothelial barrier function in response to acute lung injury

Endothelial cells (EC) form a semi-permeable barrier between the interior space of blood vessels and the underlying tissues. In acute lung injury (ALI) the EC barrier is weakened leading to increased vascular permeability. It is widely accepted that EC barrier integrity is critically dependent upon intact cytoskeletal structure and cell junctions. Edemagenic agonists, like thrombin or endotoxin lipopolysaccharide (LPS), induced cytoskeletal rearrangement, and EC contractile responses leading to disruption of intercellular contacts and EC permeability increase. The highly clinically-relevant cytoskeletal mechanisms of EC barrier dysfunction are currently under intense investigation and will be described and discussed in the current review.

Amino acid mutations in the caldesmon COOH-terminal functional domain increase force generation in bladder smooth muscle

Caldesmon (CaD), a component of smooth muscle thin filaments, binds actin, tropomyosin, calmodulin, and myosin and inhibits actin-activated ATP hydrolysis by smooth muscle myosin. Internal deletions of the chicken CaD functional domain that spans from amino acids (aa) 718 to 731, which corresponds to aa 512-530 including the adjacent aa sequence in mouse CaD, lead to diminished CaD-induced inhibition of actin-activated ATP hydrolysis by myosin. Transgenic mice with mutations of five aa residues (Lys(523) to Gln, Val(524) to Leu, Ser(526) to Thr, Pro(527) to Cys, and Lys(529) to Ser), which encompass the ATPase inhibitory determinants located in exon 12, were generated by homologous recombination. Homozygous (-/-) animals did not develop, but heterozygous (+/-) mice carrying the expected mutations in the CaD ATPase inhibitory domain (CaD mutant) matured and reproduced normally. The peak force produced in response to KCl and electrical field stimulation by the detrusor smooth muscle from the CaD mutant was high compared with that of the wild type. CaD mutant mice revealed nonvoiding contractions during bladder filling on awake cystometry, suggesting that the CaD ATPase inhibitory domain suppresses force generation during the filling phase and this suppression is partially released by mutations in 50% of CaD in heterozygous. Our data show for the first time a functional phenotype, at the intact smooth muscle tissue and in vivo organ levels, following mutation of a functional domain at the COOH-terminal region of CaD.

Caldesmon is necessary for maintaining the actin and intermediate filaments in cultured bladder smooth muscle cells

Caldesmon (CaD), a component of microfilaments in all cells and thin filaments in smooth muscle cells, is known to bind to actin, tropomyosin, calmodulin, and myosin and to inhibit actin-activated ATP hydrolysis by smooth muscle myosin. Thus, it is believed to regulate smooth muscle contraction, cell motility and the cytoskeletal structure. Using bladder smooth muscle cell cultures and RNA interference (RNAi) technique, we show that the organization of actin into microfilaments in the cytoskeleton is diminished by siRNA-mediated CaD silencing. CaD silencing significantly decreased the amount of polymerized actin (F-actin), but the expression of actin was not altered. Additionally, we find that CaD is associated with 10 nm intermediate-sized filaments (IF) and in vitro binding assay reveals that it binds to vimentin and desmin proteins. Assembly of vimentin and desmin into IF is also affected by CaD silencing, although their expression is not significantly altered when CaD is silenced. Electronmicroscopic analyses of the siRNA-treated cells showed the presence of myosin filaments and a few surrounding actin filaments, but the distribution of microfilament bundles was sparse. Interestingly, the decrease in CaD expression had no effect on tubulin expression and distribution of microtubules in these cells. These results demonstrate that CaD is necessary for the maintenance of actin microfilaments and intermediate-sized filaments in the cytoskeletal structure. This finding raises the possibility that the cytoskeletal structure in smooth muscle is affected when CaD expression is altered, as in smooth muscle de-differentiation and hypertrophy seen in certain pathological conditions.

Role of protein phosphatase 2A in the regulation of endothelial cell cytoskeleton structure

Our recently published data suggested the involvement of protein phosphatase 2A (PP2A) in endothelial cell (EC) barrier regulation (Tar et al. [2004] J Cell Biochem 92:534-546). In order to further elucidate the role of PP2A in the regulation of EC cytoskeleton and permeability, PP2A catalytic (PP2Ac) and A regulatory (PP2Aa) subunits were cloned and human pulmonary arterial EC (HPAEC) were transfected with PP2A mammalian expression constructs or infected with PP2A recombinant adenoviruses. Immunostaining of PP2Ac or of PP2Aa + c overexpressing HPAEC indicated actin cytoskeleton rearrangement. PP2A overexpression hindered or at least dramatically reduced thrombin- or nocodazole-induced F-actin stress fiber formation and microtubule (MT) dissolution. Accordingly, it also attenuated thrombin- or nocodazole-induced decrease in transendothelial electrical resistance indicative of barrier protection. Inhibition of PP2A by okadaic acid abolished its effect on agonist-induced changes in EC cytoskeleton; this indicates a critical role of PP2A activity in EC cytoskeletal maintenance. The overexpression of PP2A significantly attenuated thrombin- or nocodazole-induced phosphorylation of HSP27 and tau, two cytoskeletal proteins, which potentially could be involved in agonist-induced cytoskeletal rearrangement and in the increase of permeability. PP2A-mediated dephosphorylation of HSP27 and tau correlated with PP2A-induced preservation of EC cytoskeleton and barrier maintenance. Collectively, our observations clearly demonstrate the crucial role of PP2A in EC barrier protection.

Role of the basic C-terminal half of caldesmon in its regulation of F-actin: comparison between caldesmon and calponin

We previously reported that caldesmon (CaD), together with tropomyosin (TM), effectively protects actin filaments from gelsolin, an actin-severing protein. To elucidate the structure/function relationship of CaD, we dissected the functional domain of CaD required for the protection. The basic C-terminal half of rat nonmuscle CaD (D3) inhibits gelsolin activity to the same degree as intact CaD, although a smaller C-terminal region of D3 does not. This smaller C-terminal region contains the minimum regulatory domain responsible for the inhibition of actomyosin ATPase, and for the binding to actin, calmodulin and TM. These results suggest that the domain responsible for the inhibition of gelsolin activity lies outside the minimum regulatory domain, and that the positive charge possessed by the C-terminal half of CaD is important for its interaction with actin. Moreover, while the D3 fragment promotes the aggregation of F-actin into bundles as reported previously, this bundle formation is inhibited by the acidic N-terminal half of CaD, as well as by poly-l-glutamate. It seems likely that the acidic N-terminal half of CaD neutralizes the superfluous basic feature of the C-terminal half. A comparison between D3 and calponin, another actin-binding protein that is also basic and has similar actin-regulatory activities, is also discussed.

Phosphatase 2A is involved in endothelial cell microtubule remodeling and barrier regulation

We have recently shown that microtubule (MT) inhibitor, nocodazole (2-5 microM) significantly increases endothelial cells (EC) actomyosin contraction and permeability indicating the importance of MT in maintaining the EC barrier (Verin et al. [2001]: Cell Mol Physiol 281:L565-L574). Okadaic acid (OA, 2-5 nM), a powerful inhibitor of protein phosphatase 2A (PP2A), significantly potentiates the effect of submaximal concentrations of nocodazole (50-200 nM) on transendothelial electrical resistance (TER) suggesting the involvement of PP2A activity in the MT-mediated EC barrier regulation. Immunofluorescent staining of EC revealed that in control cells PP2A distributes in a pattern similar to MT. Consistent with these results, we demonstrated that significant amounts of PP2A were present in MT-enriched EC fractions indicating tight association of PP2A with MT in endothelium. Treatment of EC with OA leads to disappearance of MT-like PP2A staining suggesting dissociation of PP2A from the MT network. Next, we examined the effect of PP2A inhibition on phosphorylation status of MT-associated protein tau, which in its unphosphorylated form promotes MT assembly. OA caused significant increases in tau phosphorylation confirming that tau is a substrate for PP2A in endothelium. Immunofluorescent experiments demonstrated that the OA-induced increases in tau phosphorylation strongly correlated with translocation of phospho-tau to cell periphery and disassembly of peripheral MT. These results suggest the involvement of PP2A-mediated tau dephosphorylation in alteration of EC MT structure and highlight the potential importance of PP2A in the regulation of EC the MT cytoskeleton and barrier function.

Interaction of the betaIV-tubulin isotype with actin stress fibers in cultured rat kidney mesangial cells

Microtubules and actin filaments are two of the major components of the cytoskeleton. There is accumulating evidence for interaction between the two networks. Both the alpha- and beta-subunits of tubulin exist as numerous isotypes, some of which have been highly conserved in evolution. In an effort to better understand the functional significance of tubulin isotypes, we used a double immunofluorescence labeling technique to investigate the interactions between the tubulin beta-isotypes and the actin stress fiber network in cultured rat kidney mesangial cells, smooth-muscle-like cells from the renal glomerulus. Removal of the soluble cytoplasmic and nucleoplasmic proteins by detergent extraction caused the microtubule network to disappear while the stress fiber network was still present. In these extracted cells, the betaI- and betaII-tubulin isotypes were no longer present in the cytoplasm while the betaIV-isotype co-localized with actin stress fibers. Co-localization between betaIV-tubulin and actin stress fibers was also observed when the microtubule network was disrupted by the anti-tubulin drug colchicine and also by microinjection of the betaIV-tubulin antibody. Our results suggest that the betaIV isotype of tubulin may be involved in interactions between microtubules and actin.

Tyrosine phosphorylation of caldesmon is required for binding to the Shc.Grb2 complex

S3-v-erbB is a retroviral oncogene that encodes a ligand-independent, transforming mutant of the epidermal growth factor receptor. This oncogene has been shown to be sarcomagenic in vivo and to transform fibroblasts in vitro. Our previous studies (McManus, M. J., Lingle, W. L., Salisbury, J. L., and Maihle, N. J. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 11351-11356) showed that expression of S3-v-erbB in primary fibroblasts results in the tyrosine phosphorylation of caldesmon (CaD), an actin- and calmodulin-binding protein. This phosphorylation is transformation-associated, and the phosphorylated form of CaD is associated with a signaling complex consisting of Shc, Grb2, and Sos in transformed fibroblasts. To identify the tyrosine phosphorylation site(s) in the CaD molecule and to further elucidate the functional role of CaD tyrosine phosphorylation in S3-v-ErbB oncogenic signaling, we have generated a series of mutant CaDs in which one or more tyrosine residues have been replaced with phenylalanine. Using a CaD null cell line, DF1 cells (an immortalized chicken embryo fibroblast cell line), and transient transfection assays, we demonstrated that Tyr-27 and Tyr-393 are the major sites of tyrosine phosphorylation on CaD. Interestingly, Tyr-27 is located within the myosin binding domain of CaD, and Tyr-393 is adjacent to one of the major actin binding and actomyosin ATPase inhibitory domains. Our studies also show that the tyrosine phosphorylation of CaD enhances its binding to the Shc.Grb2 complex. Specifically, replacement of Tyr-27, but not of Tyr-165 or Tyr-393, significantly reduces the ability of CaD to interact with the Shc. Grb2 complex. Together, these studies demonstrate that the major sites of tyrosine phosphorylation on CaD are located in the myosin and actin binding domains of CaD and that Tyr-27 is the major tyrosine phosphorylation site through which CaD interacts with the Shc.Grb2 complex.

Direct observation of microtubule-f-actin interaction in cell free lysates

Coordinated interplay of the microtubule and actin cytoskeletons has long been known to be crucial for many cellular processes including cell migration and cytokinesis. However, interactions between these two systems have been difficult to document by conventional approaches, for a variety of technical reasons. Here the distribution of f-actin and microtubules were analyzed in the absence of fixation using Xenopus egg extracts as an in vitro source of microtubules and f-actin, demembranated Xenopus sperm to nucleate microtubule asters, fluorescent phalloidin as a probe for f-actin, and fluorescent tubulin as a probe for microtubules. F-actin consistently colocalized in a lengthwise manner with microtubules of asters subjected to extensive washing in flow chambers. F-actin-microtubule association was heterogenous within a given aster, such that f-actin is most abundant toward the distal (plus) ends of microtubules, and microtubules heavily labeled with f-actin are found in close proximity to microtubules devoid of f-actin. However, this distribution changed over time, in that 5 minute asters had more f-actin in their interiors than did 15 minute asters. Microtubule association with f-actin was correlated with microtubule bending and kinking, while elimination of f-actin resulted in straighter microtubules, indicating that the in vitro interaction between f-actin and microtubules is functionally significant. F-actin was also found to associate in a lengthwise fashion with microtubules in asters centrifuged through 30% sucrose, and microtubules alone (i.e. microtubules not seeded from demembranated sperm) centrifuged through sucrose, indicating that the association cannot be explained by flow-induced trapping and alignment of f-actin by aster microtubules. Further, cosedimentation analysis revealed that microtubule-f-actin association could be reconstituted from microtubules assembled from purified brain tubulin and f-actin assembled from purified muscle actin in the presence, but not the absence, of Xenopus oocyte microtubule binding proteins. The results provide direct evidence for an association between microtubules and f-actin in vitro, indicate that this interaction is mediated by one or more microtubule binding proteins, and suggest that this interaction may be responsible for the mutual regulation of the microtubule and actomyosin cytoskeletons observed in vivo.

The interaction of Mip-90 with microtubules and actin filaments in human fibroblasts

The novel microtubule-interacting protein Mip-90 was originally isolated from HeLa cells by using affinity columns of agarose derivatized with peptides from the C-terminal regulatory domain on beta-tubulin. Biochemical and immunocytochemical data have suggested that the association of Mip-90 with the microtubule system contributes to its cellular organization. Here we report the interaction patterns of Mip-90 with microtubules and actin filaments in interphase human fibroblasts. A polyclonal monospecific antibody against Mip-90 was used for immunofluorescence microscopy analysis to compare the distribution patterns of this protein with tubulin and actin. A detailed observation of fibroblasts revealed the colocalization of Mip-90 with microtubules and actin filaments. These studies were complemented with experiments using cytoskeleton-disrupting drugs which showed that colocalization patterns of Mip-90 with microtubules and actin filaments requires the integrity of these cytoskeletal components. Interestingly, a colocalization of Mip-90 with actin at the leading edge of fibroblasts grown under subconfluency was observed, suggesting that Mip-90 could play a role in actin organization, particularly at this cellular domain. Mip-90 interaction with actin polymers was further supported in vitro by cosedimentation and immunoprecipitation experiments. The cosedimentation analysis indicated that Mip-90 bound to actin filaments with an association constant Ka = 1 x 10(6) M-1, while an stoichiometry Mip-90/actin of 1:12 mol/mol was calculated. Western blots of the immunoprecipitates revealed that Mip-90 associated to both actin and tubulin in fibroblasts extracts. These studies indicate that Mip-90, described as a microtubule-interacting protein, also bears the capacity to interact with the microfilament network, suggesting that it may play a role in modulating the interactions between these cytoskeletal filaments in nonneuronal cells.

Both N-terminal myosin-binding and C-terminal actin-binding sites on smooth muscle caldesmon are required for caldesmon-mediated inhibition of actin filament velocity

It has been suggested that the tethering caused by binding of the N-terminal region of smooth muscle caldesmon (CaD) to myosin and its C-terminal region to actin contributes to the inhibition of actin-filament movement over myosin heads in an in vitro motility assay. However, direct evidence for this assumption has been lacking. In this study, analysis of baculovirus-generated N-terminal and C-terminal deletion mutants of chicken-gizzard CaD revealed that the major myosin-binding site on the CaD molecule resides in a 30-amino acid stretch between residues 24 and 53, based on the very low level of binding of CaDDelta24-53 lacking the residues 24-53 to myosin compared with the level of binding of CaDDelta54-85 missing the adjacent residues 54-85 or of the full-length CaD. As expected, deletion of the region between residues 24 and 53 or between residues 54 and 85 had no effect on either actin-binding or inhibition of actomyosin ATPase activity. Deletion of residues 24-53 nearly abolished the ability of CaD to inhibit actin filament velocity in the in vitro motility experiments, whereas CaDDelta54-85 strongly inhibited actin filament velocity in a manner similar to that of full-length CaD. Moreover, CaD1-597, which lacks the major actin-binding site(s), did not inhibit actin-filament velocity despite the presence of the major myosin-binding site. These data provide direct evidence for the inhibition of actin filament velocity in the in vitro motility assay caused by the tethering of myosin to actin through binding of both the CaD N-terminal region to myosin and the C-terminal region to actin.

Microtubule-microfilament synergy in the cytoskeleton

This review describes examples of structural and functional synergy of the microtubule and actin filament cytoskeleton. An analysis of basal body (centriole)-associated fibrillar networks includes studies of ciliated epithelium, neurosensory epithelium, centrosomes, and ciliated protozoa. Microtubule and actin filament interactions in cell division and development are illustrated by centrosome motility, cleavage furrow positioning, centriole migration, nuclear migration, dynamics in the phragmoplast, growth cone motility, syncytial organization, and ring canals. Model systems currently used for studies on organelle transport are described in relation to mitochondrial transport in axons and vesicular transport in polarized epithelium. Evidence that both anterograde and retrograde motors are associated with one organelle is also discussed. The final section reviews proteins that bind both microtubules and actin filaments and are possible regulators of microtubule-microfilament interactions. Regulatory roles for posttranslational modifications, microtubule and microfilament dynamics, and multisubunit complexes are considered.

Mutagenesis analysis of functionally important domains within the C-terminal end of smooth muscle caldesmon

The ability of chicken gizzard smooth muscle caldesmon (CaD) to inhibit actomyosin ATPase activity is due mainly to an inhibitory domain that resides within the C-terminal 67 amino acid residues of the CaD molecule. In the present study, a series of C-terminal truncation and internal deletion mutants of chicken gizzard smooth muscle CaD were systematically designed using a site-directed mutagenesis approach, and these mutant proteins were overexpressed in a baculovirus expression system. Analysis of actin binding and inhibition of actomyosin ATPase activity using these mutants identified a strong actin-binding motif of 6 amino acid residues (from Lys718 to Glu723), which also form the core sequence for CaD-induced inhibition of actomyosin ATPase. However, maximal inhibition by CaD requires the presence of residues 728-731, which are not associated with actin binding. Our data provide direct evidence for the requirement of actin binding to a specific region in CaD for CaD-induced inhibition of actin activation of smooth muscle myosin ATPase. Furthermore, our findings also show that the region between residues 690 and 717 is responsible for the weak inhibition of actomyosin ATPase and reveal that the inhibitory determinants located in the regions between residues 690 and 717 and residues 718 and 756 can function independently.

Contraction due to microtubule disruption is associated with increased phosphorylation of myosin regulatory light chain

Microtubules have been proposed to function as rigid struts which oppose cellular contraction. Consistent with this hypothesis, microtubule disruption strengthens the contractile force exerted by many cell types. We have investigated alternative explanation for the mechanical effects of microtubule disruption: that microtubules modulate the mechanochemical activity of myosin by influencing phosphorylation of the myosin regulatory light chain (LC20). We measured the force produced by a population of fibroblasts within a collagen lattice attached to an isometric force transducer. Treatment of cells with nocodazole, an inhibitor of microtubule polymerization, stimulated an isometric contraction that reached its peak level within 30 min and was typically 30-45% of the force increase following maximal stimulation with 30% fetal bovine serum. The contraction following nocodazole treatment was associated with a 2- to 4-fold increase in LC20 phosphorylation. The increases in both force and LC20 phosphorylation, after addition of nocodazole, could be blocked or reversed by stabilizing the microtubules with paclitaxel (former generic name, taxol). Increasing force and LC20 phosphorylation by pretreatment with fetal bovine serum decreased the subsequent additional contraction upon microtubule disruption, a finding that appears inconsistent with a load-shifting mechanism. Our results suggest that phosphorylation of LC20 is a common mechanism for the contractions stimulated both by microtubule poisons and receptor-mediated agonists. The modulation of myosin activity by alterations in microtubule assembly may coordinate the physiological functions of these cytoskeletal components.

Overexpression, purification, and characterization of full-length and mutant caldesmons using a baculovirus expression system

Three recombinant chicken gizzard caldesmon (CaD) baculovirus vectors that contained the full-length CaD codon sequence (Pv1CaD), the full-length CaD codon sequence and a six-histidine tag at the 5'-end (pBlueBacHisCaD), or the full-length CaD codon sequence and an extra six-histidine codon sequence at the 3'-end (PvlHisCaD) were constructed. Spodoptera frugiperda (Sf9) cells transfected with these constructs overexpressed full-length CaD, yielding 2, 20, and 50 micrograms per 10(6) cells for pBlueBacHisCaD, PvlHisCaD, and PvlCaD, respectively. Time course assays for the expressed proteins demonstrated that the optimum harvest time was 36 h postinfection. Immunofluorescence microscopy revealed PvlCaD localized on the plasma membrane of Sf9 cells at 24 h postinfection and distributed throughout the cytoplasm at 36-48 h postinfection. Analysis of the purified recombinant full-length CaD revealed most of the characteristics of the authentic CaD, including (a) an electrophoretic mobility corresponding to 125 kDa, (b) heat stability, (c) binding to actin, tropomyosin-actin, myosin, and calmodulin, (d) ability to inhibit actin-activated ATP hydrolysis by smooth muscle myosin, and (e) ability of Ca(2+)-calmodulin to reverse the inhibition. A CaD mutant with a deletion of 159 amino acids from the carboxyl terminus of the full-length CaD was also expressed at high levels in Sf9 cells. However, this mutant showed a decreased ability to bind to actin, tropomyosin-actin, and calmodulin, whereas the myosin binding was unaffected; actin-activated ATP hydrolysis by smooth muscle myosin was not inhibited by this mutant.

Smooth-muscle mitogen-activated protein (MAP) kinase: purification and characterization, and the phosphorylation of caldesmon

A single 42 kDa isoform of mitogen-activated protein (MAP) kinase is expressed in both embryonic and adult chicken gizzard. The gizzard MAP kinase, which cross-reacts with anti-p44mpk antibody, has been purified from adult chicken gizzard and partially characterized. The purification protocol employs phenyl-Sepharose, polylysine-agarose, hydroxyapatite, Mono-Q and phenyl-Superose column chromatography. The purified enzyme phosphorylates myelin basic protein and gizzard high-molecular-mass (h-)caldesmon. Sea-star p44mpk and gizzard MAP kinase phosphorylate h-caldesmon at identical sites at the C-terminal domain, as revealed by tryptic-peptide mapping of the phosphorylated protein. Phosphorylation of h-caldesmon by gizzard MAP kinase abolishes its interaction with polymerized tubulin. The specific activity of the purified gizzard kinase toward myelin basic protein is similar to that of brain tau kinase, but is only a fraction of that of activated sea-star p44mpk. This suggests that, although a large amount of MAP kinase is present in the gizzard, only a small percentage of the enzyme is activated normally. Autophosphorylation of the gizzard kinase, at least in part on tyrosine residues, activates its kinase activity.