J Biol Chem, Vol. 274, Issue 42, 30115-30121, October 15, 1999

Mammal-specific, ERK-dependent, Caldesmon Phosphorylation in Smooth Muscle
QUANTITATION USING NOVEL ANTI-PHOSPHOPEPTIDE ANTIBODIES^*

Gerard D'Angelo^§¶, Philip Graceffa^§, C.-L. Albert Wang, John Wrangle, and Leonard P. Adam^**

From the  Boston Biomedical Research Institute and the  Harvard Medical School, Boston, Massachusetts 02114

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Extracellular signal-regulated kinases (ERKs) phosphorylate the high molecular mass isoform of the actin-binding protein caldesmon (h-CaD) at two sites (Ser⁷⁵⁹ and Ser⁷⁸⁹) during smooth muscle stimulation. To investigate the role of phosphorylation at these sites, antibodies were generated against phosphopeptides analogous to the sequences around Ser⁷⁵⁹ and Ser⁷⁸⁹. Affinity-purified antibodies were phosho- and sequence-specific. The major site of phosphorylation in h-CaD in porcine carotid arterial muscle strips was at Ser⁷⁸⁹; however, the amount of phosphate did not vary appreciably with either KCl or phorbol ester stimulation. Phosphorylation at Ser⁷⁵⁹ of h-CaD was almost undetectable (<0.005 mol of phosphate/mol of protein). Moreover, phosphorylation of the low molecular mass isoform of the protein (l-CaD) at the site analogous to Ser⁷⁸⁹ was greater in serum-stimulated cultured smooth muscle cells than in serum-starved cells. Serum-stimulated l-CaD phosphorylation was attenuated by the protein kinase inhibitor PD98059. These data 1) identify Ser⁷⁸⁹ of h-CaD as the major site of ERK-dependent phosphorylation in carotid arteries; 2) show that the level of phosphorylation at Ser⁷⁸⁹ is relatively constant following carotid arterial muscle stimulation, despite an increase in total protein phosphate content; and 3) suggest a functional role for ERK-dependent l-CaD phosphorylation in cell division.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Caldesmon is an actin-, tropomyosin-, myosin-, and calmodulin-binding protein that is alternatively spliced to yield either a 93-kDa protein (high molecular mass caldesmon (h-CaD)¹), specific to contractile smooth muscle, or a 60-kDa protein (low molecular mass caldesmon (l-CaD)), expressed ubiquitously (1-3). h-CaD is an inhibitor of actomyosin ATPase activity and can tether actin filaments to myosin filaments, suggesting that it may be a modulator of contractility in smooth muscle. In an attempt to identify a switch that may regulate caldesmon function in tissue, we previously showed that the protein is phosphorylated in resting smooth muscle and that the level of phosphorylation increases with muscle stimulation by many chemical agents, including phorbol esters, endothelin-1, histamine, and okadaic acid, and by depolarization with KCl (4). Sequencing of phosphopeptides generated from h-CaD purified from ³²P-labeled arterial smooth muscle led to the identification of at least two sites of phosphorylation, Ser⁷⁵⁹ and Ser⁷⁸⁹ (5), based on the numbering of the mammalian sequence (6), both in the carboxyl terminus of the protein. The site at Ser⁷⁵⁹ is conserved in gizzard h-CaD (equivalent to Ser⁷⁰²); however, the site at Ser⁷⁸⁹ is absent from the gizzard protein. Although not all of the phosphate in h-CaD could be accounted for by these two serine residues, it was later found that the both sites were capable of being phosphorylated in vitro by the 44- and 42-kDa isoforms of the extracellular signal-regulated kinase (ERK1 and ERK2, respectively) (7). Both sites are phosphorylated to the same extent by ERK1 in vitro; however, the relative amounts of phosphate in the two sites in intact tissue and the modulation of phosphate content in the two sites during muscle contraction are unknown.

l-CaD lacks a central helical domain present in h-CaD, but nevertheless possesses phosphorylation sites equivalent to Ser⁷⁵⁹ and Ser⁷⁸⁹ in its carboxyl terminus. l-CaD is known to be phosphorylated in cultured cells during mitosis (8, 9), where the responsible kinase is thought to be the 34-kDa protein p34^cdc2. Although phosphorylated during mitosis, l-CaD may be phosphorylated at other transition checkpoints in the cell cycle. There are no studies that have reported on this possibility. In addition, the specific sites of l-CaD phosphorylation in cells are unknown.

To better assess the relative phosphorylation of h-CaD at these two ERK sites in intact tissue and to address the question of l-CaD phosphorylation in cultured cells during cell cycle progression, antibodies were generated that specifically recognize the two ERK-dependent phosphopeptide sequences in mammalian h-CaD. The usefulness of these antibodies is demonstrated by monitoring phosphorylation (immunoreactivity) during muscle stimulation and cultured smooth muscle cell proliferation. We find that the major site of phosphorylation in resting smooth muscle is at Ser⁷⁸⁹, the site unique to mammalian CaD. Phosphorylation at this site is not significantly altered during muscle stimulation, suggesting that ERK-dependent h-CaD phosphorylation does not play a significant role in contractile regulation, in contrast to earlier suggestions by our laboratory (17). However, phosphorylation at the equivalent site in l-CaD is dynamically regulated by ERKs during proliferation of cultured smooth muscle cells, suggesting a role in cell division.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials

Most chemicals and reagents were purchased from Sigma. [-³²P]ATP and ¹²⁵I-protein A were purchased from NEN Life Science Products. Phosphocellulose (P-81) paper was from Whatman, and nitrocellulose was from Hoefer Scientific Instruments.

Methods

Physiological Preparation-- Porcine carotid arteries were transported from the slaughterhouse in an ice-cold physiological saline solution (PSS) that consisted of 140 mM NaCl, 4.7 mM KCl, 1.2 mM Na₂HPO₄, 1.2 mM MgSO₄, 1.6 mM CaCl₂, 0.02 mM EDTA, 5.6 mM glucose, and 2 mM MOPS, pH 7.4. Arteries were dissected free of fat and connective tissue and either used immediately or stored overnight at 4 °C in fresh PSS before use. Strips of artery 5 mm in width were dissected; the endothelium was gently rubbed off; and the muscle strip was attached to a force transducer (Grass Instrument Co.) for the measurement of tension at 37 °C in PSS. Experimental solutions consisted of either PSS with (i) the addition of either phorbol 12,13-dibutyrate (PDBu; 1 µM) or okadaic acid (50 µM) or (ii) the replacement of KCl for NaCl to give a final KCl concentration of 110 mM (KPSS). Contractile responses were terminated by quick-freezing the tissue with liquid N₂-cooled tongs after either 60 min (PDBu and okadaic acid) or 30 min (KPSS) of stimulation.

Porcine carotid arterial smooth muscle cells were cultured from explants taken from the medial sections of the arteries. The cells were maintained in culture for <10 passages in Dulbecco's modified essential medium containing 10% fetal bovine serum and the antibiotics penicillin G, streptomycin, and amphotericin B (Life Technologies, Inc.).

Antibody Preparation-- Three types of polyclonal antibodies were used in the study. First, rabbit polyclonal antiserum was generated against full-length porcine stomach h-CaD (-CaD). Second, affinity-purified antibodies generated against the phosphopeptide PDGNKS(PO₄)PAPKPGC, a sequence based on Ser⁷⁵⁹ of mammalian h-CaD (equivalent to Ser⁷⁰² of gizzard h-CaD), were prepared as described previously (10). Third, antibodies were generated against the phosphopeptide CQSVDKVTS(PO₄)PTKV, a sequence that is analogous to Ser⁷⁸⁹ of mammalian h-CaD (6) and absent from avian h-CaD. Briefly, the phosphopeptide was synthesized, coupled via the cysteine residue to keyhole limpet hemacyanin using m-maleimidobenzoyl-N-hydroxysulfosuccinimide ester (Pierce), and injected into rabbits, and at appropriate times, serum was collected. Antibodies specific for the phosphorylated form of the peptide/protein were purified by a combination of (i) ammonium sulfate precipitation, (ii) passage over a column of the corresponding unphosphorylated peptide, and (iii) passage of the flow-through from Step ii over a column of the phosphopeptide.

Immunoblotting-- Frozen arterial muscle strips were ground to a fine powder under liquid N₂, and proteins were extracted into sample buffer containing 3% SDS and heated for 10 min in a boiling water bath. The extract was clarified by centrifugation at 14,000 × g for 15 min, and proteins were separated by SDS-polyacrylamide gel electrophoresis and electrophoretically transferred to nitrocellulose. Western blots were incubated overnight with an appropriate dilution (empirically derived) of antibody, washed, incubated with ¹²⁵I-protein A, and then washed again. Antibody binding was detected by autoradiography and quantitated using a -counter. For all experiments, tissue samples were run in triplicate under identical loading conditions for each replicate to permit a comparison of anti-phosphopeptide antibody binding, either at Ser⁷⁵⁹ or Ser⁷⁸⁹, with -CaD antibody binding.

Because of the possibility of caldesmon dephosphorylation during protein extraction and gel electrophoresis, additional confirmatory studies were performed as follows. Proteins in the finely ground tissue powder were denatured by the addition of dry ice-cooled acetone containing 10% (w/v) trichloroacetic acid and 10 mM dithiothreitol. The tissue homogenate was then allowed to come to room temperature, washed with acetone/dithiothreitol to remove the acid, and lyophilized. SDS-containing sample buffer was added to the lyophilized tissue, and immunoblotting procedures were continued as described above. Results obtained using this alternative procedure were not different from those using the procedure described above.

Proteins-- Recombinant ERK1 was prepared using a baculovirus expression system (10) and was activated by incubation with a recombinant, constitutively active mutant of GST-MEK1 (an isoform of MAP kinase and ERK kinase). h-CaD was purified from either porcine stomach or chicken gizzard smooth muscle according to Bretscher (11). The recombinant catalytic subunit of PP1 was generously donated by Dr. Anna DePaoli-Roach (Indiana University). Purified myosin regulatory light chains (LC₂₀), used in determining ATP-specific activity, were generously supplied by Dr. Renne Lu (Boston Biomedical Research Institute) and were phosphorylated by gizzard myosin light chain kinase in the presence of calcium and calmodulin (purified from bovine testicles). Calf intestinal alkaline phosphatase was purchased from Roche Molecular Biochemicals. Actin and tropomyosin were purified from rabbit skeletal muscle and chicken gizzards, respectively, according to methods we have used previously (12).

Protein Phosphorylation-- Purified h-CaD was phosphorylated in vitro using empirically derived amounts of recombinant ERK1 and GST-MEK1. Two phosphorylation reactions were run in parallel: one contained [-³²P]ATP, and the other contained nonradioactive ATP. The reaction mixture containing radioactivity was used to determine the level of h-CaD phosphorylation by dividing the specific activity of the protein (cpm/mol h-CaD) by the ATP-specific activity (determined as described in Ref. 4). All reactions were terminated either by boiling or by the addition of gel loading buffer containing 3% SDS. Heat-denatured proteins were sedimented by centrifugation, and CaD was separated from reaction constituents and concentrated by a combination of P-11 or DEAE column chromatography and dialysis against a storage buffer consisting of 20 mM MOPS, pH 7.0, 100 mM NaCl, 0.5 mM EGTA, 0.5 mM EDTA, 5 mM dithiothreitol, 1 mM benzamidine, and 0.1 mM phenylmethylsulfonyl fluoride.

For the "back-phosphorylation" experiments, proteins were extracted from frozen ground porcine carotid arteries into a buffer consisting of 20 mM MOPS, pH 7.0, 250 mM NaCl, 0.5 mM EDTA, 0.5 mM EGTA, 1 mM dithiothreitol, 1 mM benzamidine, and 0.1 mM phenylmethylsulfonyl fluoride. The mixture was rapidly heated to denature most proteins and clarified by centrifugation. ERK1, GST-MEK1, and ATP (0.25 mM) were added to half of the extract, and the reaction was terminated by the addition of 10% (w/v) trichloroacetic acid. As a control, the other half of the tissue extract was incubated under the same conditions in the presence of added ATP, but without the addition of ERK1 and GST-MEK1. Acid-denatured proteins were collected by centrifugation, washed with acetone, and lyophilized. Gel sample buffer containing 3% SDS was added, and Western blots were generated, in triplicate, for the subsequent assessment of antibody binding to caldesmon.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The carboxyl-terminal h-CaD sequences containing the two ERK-dependent phosphorylation sites at Ser⁷⁵⁹ and Ser⁷⁸⁹ are shown in Fig. 1. Whereas the peptide sequence surrounding Ser⁷⁵⁹ (Ser(P)⁷⁵⁹) is common to both avian and mammalian CaD, the sequence surrounding Ser⁷⁸⁹ (Ser(P)⁷⁸⁹) is unique to mammalian CaD. In mammals, Ser(P)⁷⁵⁹ and Ser(P)⁷⁸⁹ are identical in human, porcine, canine, and rat CaD. Antibodies against Ser(P)⁷⁵⁹ and Ser(P)⁷⁸⁹ were generated in rabbits and affinity-purified as described under "Experimental Procedures." As shown in Fig. 2A, anti-phospho-Ser⁷⁵⁹ (-Ser(P)⁷⁵⁹) antibodies did not bind to purified porcine stomach h-CaD on Western blots, but did bind to h-CaD that was phosphorylated with ERK1. Conversely, anti-phospho-Ser⁷⁸⁹ (-Ser(P)⁷⁸⁹) antibodies exhibited some binding to purified h-CaD, and this binding was markedly enhanced with ERK1-dependent phosphorylation (Fig. 2B). The addition of alkaline phosphatase to purified h-CaD resulted in a loss of basal -Ser(P)⁷⁸⁹ antibody binding (Fig. 2B), suggesting that some phosphate is contained in Ser⁷⁸⁹. These data are consistent with previously reported observations that purified h-CaD contains covalently bound phosphate (7). Treatment with alkaline phosphatase was not found to artifactually destroy the -Ser(P)⁷⁸⁹ antibody-binding epitope since antibody binding was recovered after subsequent phosphorylation with ERK1. Collectively, these data show that the two sets of polyclonal antibodies are specific for the phosphorylated form of CaD.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Phosphopeptide sequences used for the generation of phospho- and site-specific antibodies. Phosphopeptides analogous to the two ERK1 phosphorylation sequences in h-CaD were synthesized as described under "Experimental Procedures," and antibodies against these peptides were generated in rabbits. The two phosphopeptides are named Ser(P)759 (PSer759) and Ser(P)789 (PSer789), named for the two serine residues phosphorylated by ERK1 in vitro. Ser(P)759 is present in both mammalian and avian CaD (designated Ser(P)702 (PSer702)), whereas Ser(P)789 is specific for mammalian caldesmon. Cysteine residues were added for methodological reasons for the covalent attachment of the peptides to affinity columns or keyhole limpet hemacyanin. The mammalian sequence depicted is the human sequence and varies slightly in dog (Ser774 versus Gly), rat (Thr753 versus Ser and Ser774 versus Gly), and rabbit (Asn757 versus Ser and Ser774 versus Gly). Although not all of the porcine sequence is known, the amino acids indicated by asterisks are identical in the two species.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Anti-phosphopeptide antibody binding to purified caldesmon in vitro. In A, immunoblots are shown depicting the binding of affinity-purified antibodies generated against Ser(P)759 (-PSer759) to either purified porcine stomach h-CaD (CaD) or h-CaD that was phosphorylated with recombinant ERK1 (P-CaD). B shows immunoblots of -Ser(P)789 antibody binding to purified porcine stomach h-CaD (CaD), h-CaD treated with alkaline phosphatase (dephos-CaD), or dephosphorylated h-Cad subsequently phosphorylated with ERK1 (P-CaD).

Although the data of Fig. 2 show specificity for phosphorylated versus unphosphorylated h-CaD, the utility of the antibodies requires that they also be site-specific. To determine site specificity of antibody binding, two different approaches were chosen. First, we took advantage of the differences in the sequences at the carboxyl terminus of mammalian and avian caldesmon. -Ser(P)⁷⁸⁹ antibodies did not bind to ERK1-phosphorylated avian h-CaD since it does not contain the sequence analogous to the Ser(P)⁷⁸⁹ peptide (Fig. 3), thus demonstrating that the -Ser(P)⁷⁸⁹ antibodies do not recognize the alternate ERK1 phosphorylation site or any other site in avian h-CaD that may be phosphorylated by ERK1. Because -Ser(P)⁷⁵⁹ antibodies did not bind to purified gizzard h-CaD, the data also demonstrate that the purified protein does not contain phosphate in Ser⁷⁰², the site analogous to mammalian Ser⁷⁵⁹ after purification. Second, we used the two phosphopeptides against which the antibodies were generated to competitively antagonize antibody binding to ERK1-phosphorylated h-CaD on Western blots. Antibody binding to phosphorylated h-CaD was completely inhibited only with the phosphopeptide (0.5 mg/ml) against which the antibody was generated (Fig. 4). Although not shown, -Ser(P)⁷⁵⁹ and -Ser(P)⁷⁸⁹ antibody binding was not inhibited by the unphosphorylated peptides.

View larger version (67K): [in this window] [in a new window]   Fig. 3.   Anti-phosphopeptide antibody binding to unphosphorylated and ERK1-phosphorylated gizzard caldesmon. Purified unphosphorylated gizzard h-CaD (upper panel) and ERK1-phosphorylated gizzard h-CaD (lower panel) were subjected to SDS-polyacrylamide gel electrophoresis and electrophoretically transferred to nitrocellulose, and antibody binding was assessed by immunoblot analysis using antibodies generated against either full-length caldesmon (-CaD) or the phosphopeptide sequences Ser(P)759 (-PSer759) and Ser(P)789 (-PSer789). -Ser(P)759 antibodies bound only to ERK1-phosphorylated gizzard h-CaD. Antibodies generated against Ser(P)789, the mammal-specific sequence, did not bind to either unphosphorylated or ERK1-phosphorylated gizzard h-CaD.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Phosphopeptide competition of antibody binding to ERK1-phosphorylated caldesmon. Purified porcine stomach h-CaD was phosphorylated with ERK1, subjected to SDS-polyacrylamide gel electrophoresis, and electrophoretically transferred to nitrocellulose. Antibody binding using either -Ser(P)759 (-PSer759) or -Ser(P)789 (-PSer789) was then performed in the presence or absence of a 0.5 mg/ml concentration of the phosphopeptide Ser(P)759 or Ser(P)789. Only the phosphopeptide against which a given antibody was generated completely inhibited the binding of that antibody to recombinant ERK1-phosphorylated h-CaD (P-CaD).

Previously reported results from sequencing experiments showed that both Ser⁷⁵⁹ and Ser⁷⁸⁹ of h-CaD contained phosphate when the protein was extracted from arterial muscle (5). These experiments did not seek to determine the amount of phosphate incorporated into each of the two sites and at which site, if either, there was a modulation of phosphate content during muscle stimulation. We addressed these questions using the phosphopeptide antibodies characterized above. The binding of -Ser(P)⁷⁸⁹ antibodies to h-CaD on Western blots generated from whole tissue extracts was easily detectable, whereas the binding of -Ser(P)⁷⁵⁹ antibodies was very difficult to detect (Fig. 5). By comparing the ratio of anti-phosphopeptide antibody binding in one set of Western blots to caldesmon content (-CaD antibody binding) in a duplicate set of Western blots, we did not detect a change in the amount of phosphate in Ser⁷⁸⁹ of h-CaD with KCl stimulation of the muscle for 30 min. To maximize the possibility of detecting phosphate in Ser⁷⁵⁹, we incubated the arteries in a bathing solution containing the phosphatase inhibitor okadaic acid (50 µM, 60 min). Inexplicably, there were reductions in -Ser(P)⁷⁸⁹ antibody binding to h-CaD of 42 and 61% in two muscles stimulated for 60 min with okadaic acid. However, it was only with muscle stimulation using okadaic acid that we were able to detect -Ser(P)⁷⁵⁹ antibody binding to h-CaD in whole tissue extracts (Fig. 5).

View larger version (42K): [in this window] [in a new window]   Fig. 5.   Immunoblots of h-CaD in extracts generated from control and stimulated porcine carotid arterial smooth muscle. Strips of porcine carotid arteries were attached to force transducers and either remained unstimulated (Ctl) or were incubated with KCl (110 mM, 60 min) or okadaic acid (OA; 50 µM, 60 min). Proteins were extracted from the tissue with 3% SDS, separated (in triplicate) by SDS-polyacrylamide gel electrophoresis, and transferred to nitrocellulose, and antibody binding was performed as described under "Experimental Procedures." Western blots were probed with antibodies generated against full-length caldesmon (-CaD), Ser(P)789 (-PSer789), and Ser(P)759 (-PSer759) in A-C, respectively. Arrows refer to the position of h-CaD. The identities of the proteins to which the -Ser(P)759 antibodies bind (migrating at 116 and 97 kDa) are unknown.

Although the -Ser(P)⁷⁵⁹ antibodies are site- and phosphorylation-specific, it is apparent from the data of Fig. 5 that they bind to proteins other than CaD migrating near the 116- and 97-kDa molecular mass markers. Because the polyclonal antibodies generated against full-length CaD do not bind in this molecular mass region, these proteins are not proteolytic fragments of CaD. The identities of the proteins are unknown; however, -Ser(P)⁷⁵⁹ antibody binding increased upon muscle stimulation with okadaic acid (Fig. 5C), suggesting the possibility that they may be phosphoproteins containing phosphopeptide sequences similar to those present in h-CaD.

To determine if the phosphorylation levels in Ser⁷⁵⁹ or Ser⁷⁸⁹ are of a magnitude suggestive of a physiological role, we used the phospho-specific antibodies to quantitate the absolute level of phosphate in the two sites. This assessment was made by first determining the antibody binding ratios of -Ser(P)⁷⁸⁹ or -Ser(P)⁷⁵⁹ to -CaD for h-CaD from control or PDBu-stimulated muscles. These ratios were compared with the antibody binding ratios for h-CaD phosphorylated by ERK1 in vitro, for which the level of phosphate in the two sites could be explicitly determined. In the experiment of Fig. 6, purified h-CaD was phosphorylated to a level of 1.8 mol of phosphate/mol of protein in vitro. Approximately 50% of the phosphate incorporated into h-CaD was in each of the two sites, based on phosphopeptide map analyses of ERK1-phosphorylated h-CaD (data not shown). Therefore, ~0.9 mol of phosphate/mol of protein was incorporated into each of the two sites (Ser⁷⁵⁹ and Ser⁷⁸⁹) in vitro. Using these values, for the specific experiment depicted in Fig. 6, the amount of phosphate in Ser⁷⁸⁹ was calculated to increase with PDBu stimulation from 0.27 to 0.32 mol of phosphate/mol of protein. In four separate experiments analyzed in this manner, the phosphate content in Ser⁷⁸⁹ was, on average, 0.24 ± 0.03 and 0.28 ± 0.04 mol of phosphate/mol of protein for h-CaD from control and PDBu-stimulated muscles, respectively. The phosphate content in Ser⁷⁵⁹ was always calculated to be <0.005 mol of phosphate/mol of protein. Collectively, these data show that significant levels of phosphate are present in Ser⁷⁸⁹ of h-CaD and that these levels do not vary much with muscle stimulation by KCl or PDBu.

View larger version (34K): [in this window] [in a new window]   Fig. 6.   Estimation of the amount of phosphate in the two ERK1-phosphorylated caldesmon sites in porcine carotid arteries. Western blots (in triplicate) were generated from proteins extracted with 3% SDS from either unstimulated or PDBu-stimulated (1 µM for 60 min) porcine carotid arteries (left and center panels, respectively) and purified porcine stomach h-CaD that was phosphorylated in vitro to a level of 1.8 mol of phosphate/mol of protein with ERK1 (right panel). The binding of antibodies generated against full-length caldesmon (CaD), Ser(P)759 (PS759), or Ser(P)789 (PS789) was then assessed as described under "Experimental Procedures." *, these immunoblots were exposed approximately nine times as long as the other blots in the figure.

Two possibilities could explain a lack of -Ser(P)⁷⁵⁹ antibody binding to caldesmon from tissue. The first is that the binding epitope for the antibody is destroyed during some phase in the manipulation of the tissue and the subsequent handling of the protein extracts. The second, more likely explanation is that there is a genuine differential in phosphorylation at the two sites in tissue. To address the first possibility, we performed a back-phosphorylation experiment in which caldesmon and other proteins were extracted from carotid arterial muscle tissue and subsequently phosphorylated with ERK1 in vitro (Fig. 7). As before, little to no binding was observed with -Ser(P)⁷⁵⁹ antibodies in Western blots generated from control tissue extracts; however, binding was easily detectable in Western blots generated from the tissue extracts incubated with exogenous ERK1 and GST-MEK1. Phosphorylation at Ser⁷⁸⁹ also increased with ERK1 phosphorylation, demonstrating that caldesmon was not stoichiometrically phosphorylated in the original protein extract. These data provide more convincing evidence that the second possibility stated above is true, i.e. that phosphate levels in Ser⁷⁵⁹ are very low in tissue.

View larger version (54K): [in this window] [in a new window]   Fig. 7.   Back-phosphorylation of porcine carotid arterial caldesmon. Proteins from ground porcine carotid arterial smooth muscle were extracted into a buffer as described under "Experimental Procedures." Western blots were generated (in triplicate) from this extract and from the same protein extract after the addition of MEK and ERK1. Antibody binding to caldesmon was then performed. The upper panel is a naphthol blue/black-stained nitrocellulose membrane before immunoblotting; the lower panels are autoradiograms depicting antibody binding to h-CaD and l-CaD in the protein extracts. (The band at 66 kDa in the upper panel is bovine serum albumin from the ERK1 preparation used for enzyme stability.) -PS789, anti-Ser(P)789 antibody; -PS759, anti-Ser(P)759 antibody.

To provide an explanation for the low level of phosphorylation at Ser⁷⁵⁹ in comparison with Ser⁷⁸⁹, we investigated the possibility that these sites may be subject to differential phosphatase or kinase regulation. The purified catalytic subunit of protein phosphatase type 1 catalyzed the dephosphorylation of Ser⁷⁵⁹, but not Ser⁷⁸⁹ (Fig. 8A). These data are consistent with the results of Fig. 5 since okadaic acid is an inhibitor of protein phosphatases 1 and 2a. In addition, because CaD is bound to thin filaments in tissue, the possibility exists that one of the ERK phosphorylation sites is inaccessible to the kinase. In support of this possibility, actin and tropomyosin were found to inhibit ERK1-dependent phosphorylation of h-CaD at Ser⁷⁵⁹ more so than at Ser⁷⁸⁹ (Fig. 8B). There was no effect of tropomyosin alone on h-CaD phosphorylation, and there was no apparent difference between F- and G-actin in blocking ERK1-dependent phosphorylation at Ser⁷⁸⁹ (data not shown); however, tropomyosin did enhance the inhibition of ERK1-dependent h-CaD phosphorylation by either F- or G-actin. These results suggest the possibility of differential phosphatase and/or kinase regulation as a means for differential phosphate content in the two ERK1 phosphorylation sites of CaD in intact tissue.

View larger version (32K): [in this window] [in a new window]   Fig. 8.   Effects of phosphatase or actin/tropomyosin on caldesmon dephosphorylation and phosphorylation by ERK1. In A, immunoblots were generated from ERK1-phosphorylated porcine stomach h-CaD before (upper panel) and after (lower panel) treatment with the catalytic subunit of rabbit skeletal muscle protein phosphatase type 1 (csPP1). In B, site-specific h-CaD phosphorylation was assessed by antibody binding to purified porcine stomach h-CaD phosphorylated with ERK1 in the absence (Cont) or presence (+A/Tm) of actin and tropomyosin at a ratio with caldesmon of 9:1:1. -PSer789, anti-Ser(P)789 antibody; -PSer759, anti-Ser(P)759 antibody.

Although phosphorylation of l-CaD in cultured cells is known to occur in the carboxyl terminus of the protein (8, 9), the specific sites phosphorylated, the extent of phosphorylation, and the precise timing of phosphorylation in the cell cycle are unknown. To begin to address these questions, the phospho-specific antibodies were used to assess l-CaD phosphorylation at the sites analogous to Ser⁷⁵⁹ and Ser⁷⁸⁹ of h-CaD. Similar to results obtained with carotid arteries, there was little to no detectable -Ser(P)⁷⁵⁹ antibody binding to l-CaD from cultured porcine carotid arterial smooth muscle or A10 cells (data not shown); however, -Ser(P)⁷⁸⁹ antibody binding was relatively easy to detect. The level of phosphate in the position analogous to Ser⁷⁸⁹ was much lower in l-CaD from serum-starved cells compared with l-CaD from cells in the exponential growth phase (Fig. 9). This effect was also observed with cultured A10 cells (data not shown). Although l-CaD was first described as a phosphoprotein during mitosis, the level of phosphate in the Ser⁷⁸⁹-analogous sequence was not elevated in nocodazole-treated cells, known to arrest cells in mitosis, in comparison with exponentially growing cells (in which the majority of cells are not in mitosis) (Fig. 9A). Furthermore, if cells were serum-starved for 24 h and then serum was added back for 1 h, there was an increase in -Ser(P)⁷⁸⁹ antibody binding to a level greater than that with exponentially growing cells (Fig. 9B). The relatively rapid increase in phosphorylation at this site upon serum stimulation was markedly inhibited (59%) by preincubation with the MEK inhibitor PD98059 (50 µM; n = 3). Although not shown, the addition of PD98059 (50 µM for 2 h) to exponentially growing cells inhibited phosphorylation by 30 and 41% in two additional experiments. There was no affect of the p38 MAP kinase inhibitor SB203580 (25 µM) on l-CaD phosphorylation levels using the -Ser(P)⁷⁸⁹ antibody. These results show that l-CaD is rapidly phosphorylated as cultured smooth muscle cells progress from serum-starved conditions into the cell cycle and not only during mitosis.

View larger version (41K): [in this window] [in a new window]   Fig. 9.   l-CaD phosphorylation in cultured porcine carotid arterial smooth muscle cells. A, smooth muscle cells were harvested after serum starvation for 48 h, under conditions of exponential growth, or after nocodazole treatment (25 µg/ml) for 20 h. Proteins from the cells were extracted with 3% SDS, separated by SDS-polyacrylamide gel electrophoresis, and transferred to nitrocellulose, and immunoblotting was performed using either -CaD or -Ser(P)789 (-PSer789). B, smooth muscle cells were subjected to serum starvation (SS) and then stimulated by the readdition of serum for 1 h in the absence (10%; PD) or presence (10%; +PD) of 50 µM PD98059 for 3 h. Immunoblots were generated from these cells as described above using -CaD and -Ser(P)789 and compared with immunoblots generated from exponentially growing cells (EXP). Numbers at the bottom refer to the level of phosphorylation at Ser789 relative to serum stimulation for 1 h. FBS, fetal bovine serum.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

h-CaD is hypothesized to be a contractile regulatory protein because of its abilities to bind to actin, tropomyosin, myosin, and calmodulin in vitro and to inhibit actin-activated myosin ATPase activity and its localization to thin filaments (2, 3). However, a mechanism for the potential regulation of contraction by h-CaD is unknown, and a definitive role for the protein in this functional process has been a point of debate. l-CaD is present in most cells, where it is hypothesized to regulate the dynamics of actin filament organization (8); however, the possibility remains that the function of both l-CaD and h-CaD is similar and may involve the regulation of some process other than cell shortening. Because both isoforms of CaD are phosphorylated in cells (4, 5, 8, 9, 13-16), it is likely that the covalent addition and removal of phosphate serve to regulate its function in a reversible manner. A better understanding of the phosphorylation process should help to determine the physiological role of the protein in tissue.

h-CaD is phosphorylated in arterial smooth muscle, and total phosphate content in the protein increases with muscle stimulation by a variety of agents. Although the identities of all phosphorylation sites in h-CaD have not been elucidated, phosphopeptide sequencing of the protein isolated from ³²P-loaded muscle shows that phosphate is contained in Ser⁷⁵⁹ and Ser⁷⁸⁹, based on the numbering of the human CaD sequence (6). Approximately two-thirds of the total amount of phosphate in h-CaD is incorporated into these two ERK-dependent sites, whereas the site(s) of the remaining one-third are unknown (5). Using the newly developed phosphopeptide-specific antibodies designed to monitor the phosphorylation levels in Ser⁷⁵⁹ and Ser⁷⁸⁹, the major site of ERK-dependent phosphorylation in h-CaD was determined to be at Ser⁷⁸⁹. Binding of -Ser(P)⁷⁵⁹ antibodies was detectable only with a significant purification of h-CaD from muscle tissue. In a previous study, we found that binding of the -Ser(P)⁷⁵⁹ antibodies to h-CaD increased with various manipulations of muscle tissue (10). However, in this study, we show that the level of phosphate in Ser⁷⁵⁹ is <0.005 mol of phosphate/mol of protein. Because of the low level of phosphorylation at this site, it is unlikely that reversible addition of phosphate at Ser⁷⁵⁹ is a means through which h-CaD function is altered in smooth muscle tissue. Ser⁷⁵⁹ phosphorylation may be relevant only in circumstances where phosphatase activity is dramatically altered, as we observed with okadaic acid stimulation. On the other hand, we estimate that the level of phosphate in Ser⁷⁸⁹ of h-CaD is significant in muscle, but that the level of phosphorylation is not rapidly modified with muscle stimulation. These data suggest that ERK phosphorylation of h-CaD does not play a role in the series of rapidly occurring events that culminate in the generation of force. This conclusion is in agreement with the results of Nixon et al. (18), who showed that the phosphorylation of h-CaD following the addition of ERK to permeabilized smooth muscle does not induce the generation of force.

The observation that Ser⁷⁸⁹ phosphorylation is relatively constant during muscle stimulation contrasts with our previously published observations that total h-CaD phosphate content increases with muscle stimulation (4). One explanation for this apparent discrepancy is that the site of phosphorylation in h-CaD modulated during contraction is in a sequence different from those monitored with our site-specific antibodies, suggesting that ERK-dependent phosphorylation at Ser⁷⁸⁹ contributes only to the basal level of h-CaD phosphorylation. Although all of the phosphate incorporated into h-CaD by the ERKs is in either Ser⁷⁵⁹ or Ser⁷⁸⁹, not all of the phosphate in h-CaD purified from tissue is accounted for by these two sites (5). The location of the additional phosphorylation site(s) and the responsible kinase(s) are unknown.

The observation that Ser⁷⁸⁹ phosphorylation levels are not significantly altered with muscle stimulation by phorbol esters and KCl is surprising because of the known increase in ERK activity (15, 17). An explanation for this apparent discrepancy is provided by the results of ERK immunolocalization studies in smooth muscle (19). ERK can be found at the cell membrane or more centrally within the cytosol, in some circumstances appearing to have a filamentous distribution. The population of ERK exhibiting the more dynamic change in activity after tissue stimulation may not necessarily be the one responsible for modifying h-CaD. Relatively static, low levels of active ERK may be contained in filamentous structures such as those to which h-CaD binds. Of course, a second possibility for this apparent discrepancy is that a kinase other than ERK is responsible for h-CaD modification at Ser⁷⁸⁹. Clearly, however, our data show that the major site of h-CaD phosphorylation in tissue is at the mammal-specific site (Ser⁷⁸⁹) and that the level of phosphate contained in this site is not rapidly altered following muscle stimulation with either KCl or phorbol esters.

Purified ERK phosphorylates Ser⁷⁵⁹ and Ser⁷⁸⁹ of mammalian h-CaD at roughly the same rate in vitro.² To explain how it is that h-CaD is more extensively phosphorylated in tissue at Ser⁷⁸⁹ compared with Ser⁷⁵⁹, several experiments were performed. In agreement with the observation by Childs et al. (20), who showed that actin and tropomyosin inhibit gizzard h-CaD phosphorylation by ERK1, we found that actin and tropomyosin inhibit h-CaD phosphorylation and that this inhibition is greater at the site common among all caldesmons (Ser⁷⁵⁹) than at Ser⁷⁸⁹. Therefore, one explanation for the differential level of phosphate in the two sites is that Ser⁷⁵⁹ is sterically inhibited by thin filament proteins in the cell, resulting in the low level of phosphorylation at this site. The possibility also remains that there are mechanisms in place to remove any phosphate in Ser⁷⁵⁹ by the enzymatic actions of a specific intracellular protein phosphatase. Consistent with this concept, an increase in Ser⁷⁵⁹ phosphorylation was observed with okadaic acid stimulation of the muscle strips. In addition, the catalytic subunit of protein phosphatase type 1, similar in class to the catalytic subunit of smooth muscle myosin phosphatase (21-23), already known to be present in the actomyosin domain along with h-CaD, removes phosphate from Ser⁷⁵⁹ selectively versus Ser⁷⁸⁹. Therefore, two mechanisms may account for differential phosphorylation at the two ERK sites: steric inhibition of the kinase by actin/tropomyosin and differential susceptibility to phosphatases.

l-CaD is phosphorylated during mitosis at sites modified by the proline-directed protein kinase p34^cdc2 in vitro (8, 9). Because the ERK phosphorylation sites in CaD are a subset of the p34^cdc2 sites (24), the possibility exists that phosphorylation in cells may be by either or both of these classes of protein kinases (25). An evaluation of our data demonstrates that l-CaD phosphorylation levels are low in cells withdrawn from the cell cycle, but that the addition of serum to initiate cell proliferation results in the relatively rapid addition of phosphate to l-CaD. l-CaD phosphorylation levels 1 h after serum stimulation are not significantly different from the levels in exponential growing cells or in nocodazole-treated mitotic cells. Because cells are in a cycle that is not associated with elevated levels of p34^cdc2 after 1 h of serum stimulation (i.e. they are not in G₂ or mitosis) (25) and because much of the increase in phosphate is prevented by the MEK inhibitor PD98059, it appears that l-CaD is modified early during the initiation of cell proliferation by MAP kinases. MAP kinases are required for certain transitions of the cell cycle, including G₀/G₁, G₁/S, and G₂/M (25-27) and, once activated, may therefore target l-CaD in addition to other substrates involved in cell proliferation.

Interestingly, the ERK-dependent phosphorylation site in mammalian caldesmons is at a point of sequence diversity in comparison with avian CaD. It is striking that such a prominent phosphorylation event is species-specific. Until the precise physiological function of CaD and what role phosphorylation has in altering this function are known, it will not be possible to determine if 1) the sequence difference in avian CaD acts to confer the functional equivalent of phosphorylation of mammalian CaD, or 2) the function of ERK-dependent CaD phosphorylation in mammalian smooth muscle is unnecessary or redundant to some other process in avian smooth muscle.

In sum, we have generated unique phospho- and site-specific antibodies to analyze CaD phosphorylation in tissue. The major ERK-dependent phosphorylation site in h-CaD is at Ser⁷⁸⁹, a site unique to mammalian versus avian CaD. Based on the observation that the level of h-CaD phosphorylation at Ser⁷⁸⁹ is not significantly elevated after prolonged muscle contraction, it is unlikely that phosphorylation at this site is involved in regulating cell shortening, as previously hypothesized by our laboratory (17). Nevertheless, our results do not rule out a role for phosphorylation of h-CaD at unknown sites (other than Ser⁷⁵⁹ and Ser⁷⁸⁹) in this regulatory process. The equivalent position in l-CaD of cultured smooth muscle cells is phosphorylated rapidly during the initiation of cell proliferation, suggesting a role in cell division. The methods we have developed to directly assess the level of phosphate incorporated into the two specific ERK-dependent CaD phosphorylation sites will help to uncover the role of CaD phosphorylation by MAP kinases in the regulation of smooth muscle function.

	ACKNOWLEDGEMENT

We thank Samantha Matson for technical expertise in performing many of the experiments.

	FOOTNOTES

^* This work was supported in part by National Institutes of Health Grants HL56035 (to L. P. A.), AR41637 (to C.-L. A. W., L. P. A., and P. G.), AR30917 (to P. G.), and HL09457 (to G. D.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ These authors contributed equally to this work.

^¶ Present address: Cardiovascular Drug Discovery, F14-07, Bristol-Myers Squibb Co., P. O. Box 4000, Princeton, NJ 08543.

^** To whom correspondence should be addressed: Cardiovascular Drug Discovery, F14-07, Bristol-Myers Squibb Co., P. O. Box 4000, Princeton, NJ 08543. Tel.: 609-252-4066; Fax: 609-252-3315; E-mail: adaml@ bms.com.

² G. D'Angelo, P. Graceffa, C.-L. A. Wang, J. Wrangle, and L. P. Adam, unpublished observations.

	ABBREVIATIONS

The abbreviations used are: h-CaD and l-CaD, high and low molecular mass isoforms of caldesmon, respectively; ERK, extracellular signal-regulated kinase; PSS, physiological saline solution; MOPS, 4-morpholinepropanesulfonic acid; PDBu, phorbol 12,13-dibutyrate; GST, glutathione S-transferase; MAP, mitogen-activated protein; MEK, MAP kinase/ERK kinase.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES