Identification of the regulatory site in smooth muscle calponin that is phosphorylated by protein kinase C.

F-actin and tropomyosin inhibited the phosphorylation of calponin by protein kinase C, and the phosphorylation reduced the binding of calponin to F-actin and tropomyosin. Labeled phosphate from [gamma-32P]ATP was retained both on the chymotryptic NH2-terminal 22-kDa fragment, which contains the actin-, tropomyosin-, and calmodulin-binding regions, and on the COOH-terminal 12-kDa fragment. Fractionation of tryptic 32P-labeled peptides by high performance liquid chromatography allowed isolation of three phosphopeptides (designated T1, T2, and T3), each of which was located in three repeating amino acid motifs of calponin. Both the relative initial rates and extent of phosphorylation decreased in the order T2 > T3 > T1. Both serine and threonine residues were phosphorylated in T1 (GASQAGMTAPGTK), and only a threonine residue was phosphorylated in T2 (FASQQGMTAYGTR) and in T3 (GASQQGMTVYGLPR). As the 22-kDa fragment contained only T2, the phosphorylation site in T2 appeared to regulate the binding of calponin to F-actin and tropomyosin. The amino acid sequence of T2 indicates that protein kinase C phosphorylates Thr184. Thus Thr184 is the preferred site of phosphorylation and is functionally the most important of the sites phosphorylated by protein kinase C in smooth muscle calponin.

F-actin and tropomyosin inhibited the phosphorylation of calponin by protein kinase C, and the phosphorylation reduced the binding of calponin to F-actin and tropomyosin. Labeled phosphate from [y-32P]ATP was retained both on the chymotryptic NHz-terminal 22-kDa fragment, which contains the actin-, tropomyosin-, and calmodulin-binding regions, and on the COOH-terminal 12-kDa fragment. Fractionation of tryptic "P-labeled peptides by high performance liquid chromatography allowed isolation of three phosphopeptides (designated TI, Tz, and T3), each of which was located in three repeating amino acid motifs of calponin. Both the relative initial rates and extent of phosphorylation decreased in the order Tz > T3 > T1.

Both serine and threonine residues were phosphorylated in TI (GASQAGMTAPGTK), and only a threonine residue was phosphorylated in Tz (FASQQGM-TAYGTR) and in T3 (GASQQGMTVYGLPR). As the 22-kDa fragment contained only Tz, the phosphorylation site in Tz appeared to regulate the binding of calponin to F-actin and tropomyosin. The amino acid sequence of Tz indicates that protein kinase C phosphorylates T h P 4 . Thus T h P 4 is the preferred site of phosphorylation and is functionally the most important of the sites phosphorylated by protein kinase C in smooth muscle calponin.
The reversible phosphorylation of the 20-kDa regulatory light chain of smooth muscle myosin by a calcium-and calmodulin-dependent myosin light chain kinase has been established as the mechanism for activation of smooth muscle contraction (1)(2)(3). However, regulation by the phosphorylation of myosin light chain does not explain all aspects of the contractile functions of smooth muscle, in particular, the tonic contractile responses. Therefore, additional regulatory mechanisms have been postulated (4-6). Such mechanisms involve, for example, the calcium-and phospholipid-dependent protein kinase (protein kinase C (PKC)' ( 6 ) ) , the direct binding of calcium to myosin (7,8), and the actions of calciumsensitive factors that are associated with the thin filament. There are likely candidates for thin filament-linked regulatory proteins in smooth muscle, namely, caldesmon (9-16) and the recently discovered troponin-T-like protein, calponin (17,18).
Calponin, first purified from chicken gizzard (19) and later from bovine aorta (20), has been shown to interact with actin, tropomyosin, and . More recently, cDNA clones for chicken gizzard calponin have been sequenced, and the deduced amino acid sequences represent the two isoforms of chicken gizzard calponin (21). Smooth muscle calponin is known to be an excellent substrate for PKC in vitro (17,22), and this protein inhibits the actin-activated myosin Mg-ATPase of smooth muscle (17). Moreover, the phosphorylation of calponin by PKC results in loss of its activity via dissociation of calponin from actin (17). The binding of calponin to F-actin, tropomyosin, and calmodulin has been well documented (19), but the domains of interaction with these proteins have not been studied in detail. Chymotryptic proteolysis generates a 22-kDa fragment of turkey gizzard calponin, which has been shown to interact with tropomyosin (23). A cDNA sequence of chicken gizzard calponin (21) indicates that this 22-kDa fragment contains the region that corresponds to residues 7-182 of chicken gizzard calponin. Also, it is reported that NH2-terminal calponin region of residues 52-168 seems to contain the major determinants for F-actin and Ca2+-calmodulin binding (24).
In contracting or resting arterial smooth muscle, calponin is not phosphorylated (25). However, in intact canine tracheal smooth muscle, the potentially important thin filament proteins calponin and caldesmon are phosphorylated and dephosphorylated at a rate sufficient to indicate a role in contraction and relaxation (26).
In the present study, we determined that F-actin and tropomyosin as well as calmodulin decrease the rate of phosphorylation of calponin by PKC, and that the phosphorylation of calponin reduces its ability to bind not only to F-actin but also to tropomyosin. We prepared fragments of calponin by digestion with a-chymotrypsin and analyzed their binding to F-actin, tropomyosin, and calmodulin. Furthermore, the phosphorylation of calponin a t multiple sites by PKC was analyzed, and the functionally most important and preferred site of phosphorylation was identified. namely, CM-Sephadex was used for the ion-exchange chromatography. PKC was purified from rat brain by the method previously described (27). Tropomyosin was purified from chicken gizzard by a modification of the procedure described by Brecher (28). The following proteins were purified by previously described methods: rabbit skeletal actin (29), bovine brain calmodulin (30), chicken gizzard myosin (31), and myosin light chain kinase (32).
Phosphorylation of Calponin-Phosphorylation of purified chicken gizzard calponin was carried out at 30 "C in a reaction mixture that contained 20 mM Tris-HC1 (pH 7.5), 5.8 p M calponin, 1 pg/ml PKC, 100 p~ [y3'P]ATP (25,000-30,000 cpm/nmol) or 100 p M ATP, 5 mM MgCI2, 0.8 mM CaC12, and 50 pg/ml phosphatidylserine. The reaction was initiated by the addition of [y-32P]ATP. Incorporation of radiolabeled phosphate was monitored by adding aliquots of the reaction mixture to 20% trichloroacetic acid at appropriate times. The precipitates, after centrifugation, were washed twice with 5% trichloroacetic acid, and then radioactivity was determined by scintillation counting. For the preparation of phosphorylated calponin, we took advantage of the heat stability of calponin (19). The reactions were stopped by the addition of 5 mM EGTA which was followed by boiling at 95 "C for 5 min. Then the mixtures were centrifuged at 100,000 X g at 30 min. This procedure effectively inactivates and removes the kinase without affecting the function of the calponin. The resultant supernatants were dialyzed exhaustively against buffer that contained 20 mM Tris-HC1 (pH 7.5), 50 mM NaC1, 0.1 mM EGTA, and 0. 5 mM DTT and used for subsequent binding or fluorescence assays.
Purification of Tryptic Phosphopeptides by HPLC-32P-Labeled calponin was purified by HPLC. In brief, 1 ml of calponin (0.3 mg/ ml) was phosphorylated by incubation with PKC at 30 "C for various times. The reactions were stopped by addition of 5 mM EGTA, boiling, and subsequent clarification, and 0.1% trifluoroacetic acid was added to the supernatants. Then each supernatant was applied to a Bondasphere 5 C-8 (Waters) reverse-phase column attached to a Shimadzu LC 6A HPLC system. HPLC-purified 32P-labeled calponin was lyophilized, and the residue was dissolved in a solution of 100 mM Tris-HC1 (pH 8.3), 1 mM DTT, and 20 mM KC1 which contained TPCK-treated trypsin (1:lOO (w/w)) and incubated at 25 'C. After 6 h, additional trypsin was added to bring the ratio of trypsin to calponin up to 1:50 (w/w), and the incubation was continued for another 12 h. The samples were adjusted to 0.1% trifluoroacetic acid, and then tryptic fragments were separated by HPLC on a C-8 reversephase column with a linear gradient from 0.1% trifluoroacetic acid in water to 0.1% trifluoroacetic acid in 50% acetonitrile in 100 min. The flow rate was 1 ml/min, and fractions were collected at 1-min intervals. The radioactivity of each fraction was determined by liquid scintillation counting. For further purification, fractions corresponding to the major peaks of radioactivity were pooled separately, lyophilized, and rechromatographed with more gradual gradient of acetonitrile.
Determination ofAmino Acid Sequence-The amino acid sequences of 32P-labeled phosphopeptides were determined with a gas-phase protein Sequencer (model 473A, Applied Biosystems).
Analysis of Phosphoamino Acid-Phosphoamino acids from 32Plabeled tryptic peptides were analyzed by one-dimensional thin layer electrophoresis as described previously (33). Purified 32P-labeled peptides were dried in a centrifugal concentrator (model CC-181; Tomy, Tokyo), and each sample was dissolved in 6 N HCI and hydrolyzed for 3 h at 110 "C. The samples were dried and then dissolved in buffer of pH 3.5 that contained 1 pg each of phenol red, phosphoserine, phosphothreonine, and phosphotyrosine. The samples were spotted on a 20 X 20-cm2 cellulose plate and subjected to electrophoresis (origin closer to the negative electrode) at 1000 V for 60 min. Under these conditions, all of the applied radioactivity remained on the plate.
Assay for Binding of Proteins to F-actin-Assessment of the binding of unphosphorylated or phosphorylated chymotryptic fragments of calponin to F-actin was achieved by the co-sedimentation method. The samples were incubated for 30 min at 25 "C in a solution that contained 20 mM Tris-HC1 (pH 7.5), 30 mM KCl, 2 mM MgCI,, 1 mM ATP, 1 mM DTT, and 0.2 mM CaCIZ. The samples were then centrifuged for 1 h at 100,000 X g (TL 100; Beckman) in order to sediment actin and the fragments of calponin that had bound to actin. After centrifugation, the pellets and supernatants were analyzed by SDSpolyacrylamide gel electrophoresis (PAGE). Gels were stained with Coomassie Brilliant Blue R-250 and the fragments distributed in the pellets and supernatants were quantified by densitometric scanning. Determination of stoichiometry and dissociation constant ( K d ) were also carried out by the co-sedimentation method, in the same solution that contained 10 p~ actin, 0.75-4 p M calponin and in the presence and absence of 1.7 p M tropomyosin.
Affinity Chromatography-Tropomyosin or calmodulin was coupled to cyanogen bromide-activated Sepharose 4B according to the instructions supplied by Pharmacia LKB Biotechnology Inc. The columns were equilibrated at 4 "C with starting buffer before application of the chymotryptic fragments of calponin. Samples of 300 pg of digested protein were applied to the column, and the column was washed with equilibration buffer A (20 mM Tris-HC1 (pH 7.5), 1 mM DTT, 1 mM CaClZ, 2 mM MgCI2, and 1 p M p-amidinophenyl methanesulfonyl fluoride hydrochloride) for tropomyosin affinity column, or with buffer A that contained 100 mM NaCl for calmodulin-sepharose 4B affinity. After elution of the unbound protein or fragments and extensive washing of the column, the elution of the bound protein or fragments was achieved with buffer A that contained 20,100, and 500 mM NaCl in a stepwise manner, for the tropomyosin affinity column. For the calmodulin affinity column, the elution of the bound protein or fragments was achieved with buffer A that contained 100 mM NaCl and 1 mM EGTA instead of 1 mM CaCl2. The eluted fractions were analyzed by SDS-PAGE.
Fluorescence Measurements-Dansyl-calmodulin was prepared by the method described by Kincaid et al. (34). Typically, 30 pl of acetone containing 6 mM dansyl chloride (1.5 mol/mol of calmodulin) were added slowly with stirring at 0 "C to 2 ml of a solution of calmodulin (1 mg/ml) that had been dialyzed against 10 mM NaHC03 (pH 10.0). After incubation for 90 min at 30 "C, the mixture was dialyzed exhaustively against 20 mM Tris-HCI (pH 8.0), 250 mM NaC1, 5 mM MgCl2, 0.1 mM EGTA, and 1 mM DTT at 4 "C. Dansyl-tropomyosin was prepared by the method described by Burtnick and Bhangu (35). Tropomyosin was labeled with dansyl chloride in a reaction mixture composed of 40 mM sodium borate (pH 8.0), 0.5 M KCl, and 1 mM DTT for 3 h at 30 "C. After labeling, the protein was dialyzed exhaustively against the same buffer as used for dialysis of dansylcalmodulin. Fluorescence measurements were made at 25 "C in a spectrofluorometer equipped with a thermostatted cell holder (RF-5000; Shimadzu). Emission spectra were recorded with an excitation wavelength of 340 nm for dansyl-calmodulin and 365 nm for dansyltropomyosin, with slit widths of 5 nm. Binding data were analyzed by double-reciprocal plots and lines were fitted by least-squares analysis.
Other Techniques-Measurement of chemical phosphate in protein was determined by the method described by Buss and Stull (36). SDS-PAGE was carried out in the buffer system established by Laemmli (37). Gels were autoradiographed, if necessary, on Kodak X-Omat RP film. Protein concentrations were determined by the method of Bradford (38). Statistical significance was determined using paired or unpaired Student's t test.

RESULTS
Phosphorylation of Calponin by PKC-In a previous study we reported that smooth muscle calponin is a good substrate for PKC and that the phosphorylation of calponin by PKC is inhibited by calmodulin in a noncompetitive fashion with respect to calponin (22). We investigated whether or not actin and tropomyosin can also inhibit the phosphorylation of calponin, and we found that indeed they can. Kinetic analysis of actin-induced inhibition of the phosphorylation of calponin by PKC revealed that actin inhibited the phosphorylation in a competitive fashion with respect to calponin (Fig. U) and the Ki value was 2.2 WM. In contrast to actin, tropomyosin inhibited the phosphorylation in an uncompetitive fashion ( Fig. 1B). When using histone type 111-S from calf thymus (Sigma) as the substrate for PKC, no inhibition of PKC was observed in the presence of actin, tropomyosin, or calmodulin (data not shown). These results indicate that the sites of binding of actin, tropomyosin, and calmodulin to the calponin molecule may be different from one another. Furthermore, we examined the phosphorylation of calponin by PKC in the presence of both actin and tropomyosin. In the presence of excess concentrations of both actin and tropomyosin, the phosphorylation was markedly reduced, but not completely.

Measurement of Chemical Phosphate in Purified Calponin-
We determined whether the purified calponin is already partially phosphorylated or not, by the chemical phosphate determinations according to the method described by Buss and Stull (36), before determining the effect of phosphorylation of calponin on its binding abilities to F-actin, tropomyosin, and calmodulin. We found that endogenous phosphate in purified calponin was not detected.
Effect of Phosphorylation on the Ability of Calponin to Bind F-actin, Tropomyosin, and Calmodulin-The binding of phosphorylated and unphosphorylated calponin to F-actin in the presence and absence of tropomyosin were compared by the ultracentrifugation method, as described under "Experimental Procedures.'' As shown in Table I, the amount of phosphorylated calponin remaining unbound was 59.2 +. 2.5%, as compared with 12.2 f 3.4% in control reactions with unphosphorylated calponin. The apparent dissociation constant (Kd) value of calponin for F-actin in the absence of tropomyosin, which was obtained from a double-reciprocal plot (Fig. M ) , increased from 0.19 +-0.06 p~ for the unphosphorylated protein to 1.6 +-0.54 p~ for the phosphorylated protein. These differences are statistically significance ( p < 0.05). The corresponding values of calponin for F-actin in the presence of tropomyosin also increased from 0.18 f 0.04 pM to 1.2 f 0.16 p~ ( p < 0.01) (Fig. 2B). Tropomyosin did not affect the Kd value of unphosphorylated and phosphorylated calponin for F-actin statistically. These results suggest that phosphorylation of calponin by PKC can reduce the binding affinity for F-actin both in the presence and absence of tropomyosin. These results were in good agreement with the findings, reported previously (17), that phosphorylated calponin has a much reduced affinity for F-actin.
To determine whether the phosphorylation of calponin affects its binding to tropomyosin and calmodulin, we examined the effect of the phosphorylation of calponin on its ability to bind dansyl-tropomyosin and dansyl-calmodulin. The effects of tropomyosin-binding proteins on the fluorescence of dansyl-tropomyosin have been reported previously (35). In the presence and absence of Caz+, the fluorescence spectrum of dansyl-tropomyosin exhibited a maximum at 505 nm when it was excited at 365 nm. Addition of calponin caused an Calponin by PKC  (3 p~) and its chymotryptic fragments were incubated with F-actin (10 p~) , and then the reaction mixtures were sedimented at high speed, as described under "Experimental Procedures." Supernatants and pellets were analyzed by SDS-PAGE. Results were determined from densitometric analysis of gels after SDS-PAGE of supernatants and pellets. Cap, calponin; P-Cap, phosphorylated calponin; 22K, 22-kDa fragment; P-22K, phosphorylated 22-kDa fragment; 12-K, 12-kDa fragment; P-12K, phosphorylated 12-kDa fragment. The values given are mean f S.E. of four to six observations. Distribution of calponin, 22-kDa and 12-kDa fragments Conditions Supernatant Pellet increase in fluorescence intensity of dansyl-tropomyosin but did not cause any shift in the emission maximum of the spectrum (data not shown). The effects of calcium and calmodulin-binding proteins on the fluorescence of dansyl-calmodulin have been also reported previously (34). With excitation at 340 nm, the fluorescence spectrum of dansyl-calmodulin in the presence of 0.1 mM EGTA exhibited a maximum at 520 nm; in the presence of 0.3 mM Ca2', the maximum shifted to 490 nm and the fluorescence intensity at 490 nm increased. In the presence of Ca2+, the addition of calponin caused a blue shift in the fluorescence spectrum and the fluorescence intensity at 490 nm of dansyl-calmodulin increased (data not shown). In the absence of Ca2+, the addition of calponin did not alter the spectrum of dansylcalmodulin (data not shown). These results correspond closely to earlier observation that calponin can bind to a tropomyosin affinity column in a Ca2+-independent manner and to a calmodulin affinity column in a Ca2+-dependent manner (19).
The effects on the fluorescence intensity of the addition of increasing amounts of calponin to dansyl-tropomyosin and dansyl-calmodulin are illustrated in Fig. 3, A and 3 B. In the case of binding to dansyl-tropomyosin (Fig. 3A), the apparent dissociation constant (Kd) value for calponin, which was obtained from a double-reciprocal plot, increased from 1.8 f 0.19 p M for the unphosphorylated protein to 4.1 k 0.94 p M for the phosphorylated protein significantly ( p < 0.05), and the maximum change in the fluorescence intensity (Urn-/ Fo) decreased from 0.79 k 0.074 for the unphosphorylated protein to 0.25 f 0.034 for the phosphorylated protein significantly ( p < 0.01). In the case of binding to dansyl-calmodulin (Fig. 3B), Urn,/Fo decreased from 0.64 -+ 0.026 for the unphosphorylated protein to 0.30 f 0.061 for the phosphorylated protein significantly ( p < 0.05), but similar apparent Kd values were obtained for both unphosphorylated (0.64 k 0.026 p~) and phosphorylated calponin (0.55 f 0.032 pM). These results suggest that the phosphorylation of calponin by PKC reduces the binding affinity of calponin not only for F-actin but also for tropomyosin. In the case of calmodulin, phosphorylation of calponin does not reduce the binding affinity for calmodulin. However, phosphorylation of calponin probably reduced the conformational change of calmodulin-calponin complex, because the increase in fluorescence intensity is directly related to complex formation (39). Location of Phosphorylation Sites and the F-actin-, Tropomyosin-, and Calmodulin-binding Regions in Calponin-Preliminary experiments with a number of proteases revealed that digestion with a-chymotrypsin of chicken gizzard calponin gave the most clearly differentiated set of digestion products. In an effort to obtain the functional fragments of calponin, samples of unphosphorylated and phosphorylated calponin were digested with a-chymotrypsin. Calponin was digested more slowly after the protein had been phosphorylated by PKC (data not shown). However, limited chymotryptic digestion of both unphosphorylated and phosphorylated calponin gave rise to the same fragments. The pattern of digestion of phosphorylated calponin after SDS-PAGE is shown in Fig. 4A. During the course of a 60-min reaction, the first relatively stable products are the 31-, 22-, and 12-kDa fragments. Amino-terminal sequence analysis of the 22-and 12-kDa fragments revealed that the 22-and 12-kDa fragments begin at Asn7 and Glyla3, respectively. However, in the absence of confirmation of carboxyl-terminal sequences of these fragments, our data on the amino-terminal sequences of these fragments, their estimated molecular weights, and the sequence of calponin (21) lead us to tentatively conclude the following. The 22-kDa fragment begins at Asn7 and ends at, or very near, Tyr"' or Trylag; the 12-kDa fragment begins at Glyla3 and ends at, or very near, the carboxyl terminus. Fig.  4B shows an autoradiogram of the same SDS-polyacrylamide gel as in Fig. 4A. Radioactivity was associated with the 22and 12-kDa fragments, as well as with the intact calponin. These results indicate that the sites of phosphorylation are present in both the 22-and 12-kDa fragments generated by the digestion with a-chymotrypsin.
The binding of phosphorylated and unphosphorylated 22and 12-kDa fragments to F-actin was compared by the ultracentrifugation method described in the "Experimental Procedures." As shown in Table I, although the 22-kDa fragment has weaker affinity for F-actin than intact calponin, this fragment co-sediments with F-actin in a Ca2+-independent manner. In contrast to the 22-kDa fragment, the 12-kDa fragment does not co-sediment with F-actin. Moreover, as shown in Table I These results suggest that the actin-binding region(s) is located on the 22-kDa fragment, and that the phosphorylation of the 22-kDa fragment, as well as that of intact calponin, reduces its ability to bind to F-actin.
To identify the fragments that bind tropomyosin and calmodulin, the digestion mixture containing unphosphorylated calponin was subjected to affinity chromatography on immobilized tropomyosin and immobilized calmodulin. The 22-kDa fragment was retained on the immobilized tropomyosin in a Ca2+-independent manner, and on the immobilized calmodulin in a Ca2+-dependent manner, but the 12-kDa fragment flowed through the two affinity columns in both the presence and the absence of Ca2+ (Fig. 5 , A and B). The changes in fluorescence intensity of dansyl-tropomyosin and dansyl-calmodulin, as a function of the concentration of calponin, of the 22-and 12-kDa fragments, are shown in Fig. 5, C and D.
The 12-kDa fragment had essentially no effect on the fluorescence of either dansyl-tropomyosin or dansyl-calmodulin. By contrast, calponin and the 22-kDa fragment caused substantial increases in the intensity of fluorescence. These findings are in agreement with the results of binding assays using affinity chromatography. These results indicate that the tropomyosin-and calmodulin-binding regions are located in the 22-kDa fragment.
Purification and Analysis of Tryptic Phosphopeptides-To identify the sites in calponin that are phosphorylated by PKC, calponin was phosphorylated (1.2 mol of phosphate/mol of calponin) and subjected to the complete tryptic hydrolysis (see "Experimental Procedures"). The tryptic peptides were separated by C-8 reverse-phase HPLC. Fractionation of tryptic phosphopeptide peaks of calponin that had been phosphorylated by PKC are shown in Fig. 6, A and B. It is clear from Fig. 6B that three major radioactive peaks were found. Each radioactive peak fraction was rechromatographed using more gradual CHJCN gradient to achieve further purification by reverse-phase HPLC. As a result, each of the three peaks contains only one phosphopeptide. The three purified phosphopeptides were designated TI, Tz, and TS on the basis of the order of their elution from the first reverse-phase column.
The recovery of 32P from the column was 80-90% of the applied radioactivity. To determine the relative initial rates , (pM) [Total added protein] , (pM) FIG. 5. Binding of a-chymotryptic fragments to immobilized tropomyosin and calmodulin and effects of these fragmentn on the fluorescence of dansyl-calmodulin and dansyl-tropomyosin. Calponin was digested with a-chymotrypsin as descrlhed under "Experimental Procedures." Digestion was stopped by the addition of 2 mM diisopropyl fluorophosphate, and the digests were dialvzetl against buffer A (see "Experimental Procedures") for the binding assay with a tropomyosin affinity column and against huffer A that contained 100 mM NaCl for the binding assay with a calmodulin affinity column. The dialyzed digests were loaded on columns of tropomyosin-Sepharose 4B and calmodulin-Sepharose 4B. After washing of unbound proteins from the columns, bound proteins were eluted as described under "Experimental Procedures." A, 12.5% gel after SDS-PAGE of fractions from tropomyosin affinity column; lune I . total fragments; lone 2, of phosphorylation and the extent of phosphorylation of these phopeptides generated by tryptic digestion of chicken gizzard phosphopeptides, we prepared calponin that was phosphoryl-calponin that had been phosphorylated by PKC are shown in ated by PKC to different extents and then we repeated the Table 11. T o identify the location of the phosphorylated amino analysis after tryptic hydrolysis. From Fig. 7, it is clear that both the relative initial rates and the extent of phosphorylation of these three phosphopeptides decreased in the order T P > T3 > TI. Among these three phosphopeptides, the preferred site of phosphorylation is located in Tz, which accounts for about 50% of the total phosphorylation by PKC after 60 min and about 67% of the total phosphorylation within the first 5 min.
Amino Acid Sequencing and Determination of Phosphoamino Acids-The actual amino acid sequences of the phos-acids, phosphoamino acid analysis of these phosphopeptides was carried out. Both serine and threonine residues were found to be the phosphorylated amino acids in TI, and only threonine was phosphorylated in TP and Ta (Fig. 8). This finding is in good agreement with our previous report that PKC preferentially phosphorylated threonine in calponin from chicken gizzard (22). Moreover, similar results were obtained in the presence of both 50 p~ CaCI2 and 1 p~ phorbol 12,13-dibutyrate or 30 pg/ml 1-oleoyl-2-acetyl-racglycerol in the reaction mixture for the phosphorylation of calponin by PKC (data not shown). These results indicate that the Tz of calponin, which contains the preferred sites of phosphorylation, was determined to be calponin (Phe'73-Arg'=), with Thr" and/or Thr" being phosphorylated. T3, which contains less preferred sites of phosphorylation, was determined to be calponin (Gly251-Arg265), with ThrZ5' being the site of phosphorylation. T, was determined to be calponin ( G l~~~~-L y s~~~) , with Ser215 and ThrZz0 and/or ThrZz4 being the sites of phosphorylation. Thus multiple sites of phosphorylation were identified and those contained within the 22-kDa fragment, which had the actin-, tropomyosin-, and calmodulin-binding regions, appeared to be functionally important. However, two threonine residues, Thr'@' and T h P , were present within the phosphopeptide T,. Therefore, it was important to establish the site phosphorylated by PKC in tryptic Tz peptide. The most probable sites of phosphorylation were identified by failure to identify a Thr when peptide sequences were compared with the known sequence (40). To eliminate the possibility that either of these residues (Thr'"'' and Thr") was phosphorylated, we used the synthetic peptide corresponding to Tz. As shown in Table 111, Thr'@' in the tryptic phosphopeptide Tz and both Thr" and Thr" in the corresponding synthetic peptide were detected. But the yield of ThrIa in the tryptic phosphopeptide TZ was found to be reduced markedly. These data suggest that Thr" is the phosphorylated amino acid in the Tz, which accounts for about 50% of the total phosphorylation (about 0.6 mol of phosphate/mol of calponin) (Fig. 7).

DISCUSSION
PKC is now recognized as a major regulatory enzyme and it has been implicated in the control of a wide variety of physiological processes (41). There are numerous examples of endogenous substrates, including cytoskeletal proteins (41), that can be phosphorylated by PKC. Since this kinase may be implicated in smooth muscle functions, such as contraction and relaxation (6,42-44), substrates of PKC in smooth muscle are of particular interest. We reported previously that calponin from chicken gizzard smooth muscle is an excellent substrate for PKC (22).
The experiments reported herein demonstrate that three tryptic phosphopeptides can be generated from chicken gizzard calponin, (TI, (Gly213-Lys225); Tz, (Phe173-Arg'85); and T3, (Gl~'~'-Arg2~)), when calponin has been phosphorylated by PKC. This result is in good agreement with the results of two-dimensional phosphopeptide mapping by Winder and Walsh (li'), in which three major tryptic phosphopeptides were detected. From our results of amino acid sequence and phosphoamino acid analysis of these phosphopeptides, we have determined that PKC phosphorylates Thrla in T,, S e P 5 and T h P O and/or ThrZz4 in TI, and ThrZ5' in T3. From these findings, together with the observations that (i) the F-actin-, tropomyosin-, and calmodulin-binding domains are all located in the 22-kDa fragment; (ii) the binding of F-actin and tropomyosin to calponin inhibits the phosphorylation of calponin by PKC; (iii) the phosphorylation of calponin by PKC reduces its ability to bind F-actin and tropomyosin; and (iv) the phosphorylation of the 22-kDa fragment also reduces its ability to bind F-actin, we argue that the site of phosphorylation located in the 22-kDa fragment is functionally the most important of all the sites of phosphorylation in calponin by

I1
Sequence analysis of phosphopeptides generated by tryptic proteolysis of calponin that had been phosphorylated by PKC Calponin was exhaustively phosphorylated by PKC and was digested with trypsin, and phosphopeptides were separated by reverse-phase HPLC, as described under "Experimental Procedures." Purified phosphopeptides were subjected to amino acid sequencing.

G l y -A l a -S e r -G l n -G l n -G l y -M e t -T h r -V a l -T~r -G l~-L e u -P r~-A r~
"The number is based on the published sequence (21). Aliquots that contained approximately equimolar amounts of the tryptic phosphopeptides separated by reverse-phase HPLC were hydrolyzed at 110 "C for 120 min prior to electrophoresis at pH 3.5, as descrihed under "Experimental Procedures." ori, origin; p-.srr, phosphoserine; p-thr, phosphothreonine; p-tyr, phosphotyrosine.  (21), and protease specificity, it is suggested that only ThrlR4 is included in the 22-kDa fragment, and the other phosphorylated amino acids are included in the 12-kDa fragment. Moreover, it is of interest to note that the three phosphopeptides that were preferentially phosphorylated by PKC were located within the repeating motifs of amino acid sequence within the calponin molecule.
When calponin was phosphorylated by PKC, we observed that the phosphorylation was inhibited by the presence of either F-actin or tropomyosin. \Ye evaluated the kinetics of the actin-and tropomyosin-induced inhibition of the phosphorylation of calponin under the condition that most of the nnP was incorporated into the T, (Fig. 7 ) . The results revealed that actin inhibits the phosphorylation in a competitive fashion while tropomyosin inhihits it in an uncompetitive fashion with respect to calponin.
N'e reported previously (22) that calmodulin inhibits phosphor>tlation in a noncompetitive fashion with respect to calponin.
From these results, we speculate that the site of phosphorylation by P K C located within the T2 of calponin are affected hy the binding of Factin, tropomyosin, and calmodulin to calponin and that the T2 may be adjacent to the actin-binding domain, but not to the calmodulin-and tropomyosin-binding domains, because of the t-ypes of their inhihitory effects on the phosphorylation of calponin by PKC.
We also found that the phosphorylation of calponin by PKC alters the capacity of this protein for binding tropomyosin, as well as actin. Although phosphorylated calponin retained its tropomyosin-binding capacity. phosphorylation caused a decrease in the affinity of calponin for tropomynsin, as indicated by results obtained using fluorescence techniques. In contrast to our result, it was shown pre\riously (17) that the phosphorylation of calponin by PKC' does nnt affect the interaction of this protein with immobilized tropnmyosin. The discrepancy between our result and the other may t w due to difference in the binding assay. namely. the u s e o f flwrescence technique in our study and affinity chromatoEraphy in the other. It has been reported that calponin inhibits the actin-activated myosin Mg-ATPase of smooth muscle and that the phosphorylation of this protein by J'KC results in loss of this ability via the dissociation of calponin from act in (17). Our findings suggest that the inhibitory effect a f calponin on the actin-activated mvosin Mg-ATl'ase in smooth muscle is due to its ability to bind not only actin hut also tropomyosin, which is reduced when calponin is phosphorylated by PKC.
Calponin interacts with F-actin and tropomyosin in a Ca2+independent manner and with calmodulin in a Ca'+-dependent manner (19). In the present study, using binding assays, we found that F-actin-, tropomyosin-. and calmodulin-binding domains were all located in the chymotryptic 22-kDa fragment. These findings are in good agreement with other recent findings (21, 23, 24). Mezgueldi et al. (24) reported that NH2terminal ragion of residues 52-lfi8 in chymotryptic NH,terminal 22-kDa fragment contains the major determinats for F-actin and Ca2+-calmodulin binding. Takahashi and Nadal-Ginard (21) published the sequence of a cDNA clone for calponin from chicken gizzard and suggested the locations of the putative actin-, tropomyosin-, and calmodulin-binding domains of this protein. They suggested that amino acid residues 129-149 of calponin may he responsible for the interaction with actin filaments and calmodulin, and that the tropomyosin-binding structure of calponin may be composed of several regions, which extend to the amino-terminal half of the molecule. In an other report, \.'ancompernolle ct nl. (29) demonstrated that the chymotryptic NH2-terminal 22-kDa fragment of turkey gizzard calponin hinds t n tropomyosin-Sepharose. In addition, Takahashi et al. reported that amino acids residues 185-193 of calponin, which are homologous to the calmodulin-binding domain of neuromodulin, are possible candidates for the site of binding of calmodulin in a Ca2+independent manner. However, our results indicated that the chymotryptic 12-kDa fragment, which contained amino acid residues 185-193 of calponin, did not interact with calmodulin either in the presence or in the absence of Ca2+, as evidenced both by the inability to affect the fluorescence spectra of dansyl-calmodulin and by the inability to bind to immobilized calmodulin on an affinity column.
Finally, it is reported that in uiuo calponin is not phosphorylated in contracting arterial smooth muscle (25). By contrast, in tracheal smooth muscle, calponin phosphorylation increased rapidly in response to the contraction of smooth muscle by carbachol, and remained elevated at steady state (26). In our present study, it is suggested that calponin is phosphorylated by PKC in vitro and that the phosphorylation is inhibited partially in the presence of both actin and tropomyosin. These data suggest that calponin can be phosphorylated by PKC under some physiological conditions. Further investigations are needed to establish the physiological relevance of calponin phosphorylation.
The ability of PKC to phosphorylate calponin and to affect the functional properties of smooth muscle calponin suggests that this enzyme may also regulate certain calponin-mediated functions of smooth muscle. It appears that the site of phosphorylation in the 22-kDa fragment, which contains the Factin-, tropomyosin-, and calmodulin-binding regions, is functionally the most important of all the sites phosphorylated by PKC.