Smooth Muscle Calponin INHIBITION OF ACTOMYOSIN MgATPase AND REGULATION BY PHOSPHORYLATION*

Calponin isolated from chicken smooth cle inhibits the actin-activated MgATPase activity of smooth muscle myosin in a reconstituted system com- posed of contractile and regulatory proteins.

Primary regulation of contraction in smooth muscle involves phosphorylation of the 20,000-dalton light chains of myosin by Ca'+/calmodulin-dependent myosin light chain kinase (1,2). It is becoming increasingly clear, however, that other regulatory systems, having both direct and indirect calcium dependence, may have a role to play in the regulation of smooth muscle contraction (3). These mechanisms of regulation include caldesmon/calmodulin (4,5), the calcium-and phospholipid-dependent protein kinase (protein kinase C) (6), and perhaps the direct binding of Ca2+ to myosin (7,8). Recently, another smooth muscle protein has been described which may function to regulate the contractile state of the muscle (9,10 To whom all correspondence should be addressed. and calmodulin-binding troponin T-like protein. Calponin, which has been purified from chicken gizzard (9) and bovine aorta (lo), is a heat-stable, basic, 34-kDa protein which interacts with F-actin and tropomyosin in a Ca'+independent manner and with calmodulin in a Ca'+-dependent manner. It is present in smooth muscle at the same molar concentration as tropomyosin (9). Electron microscopy supports the idea that calponin is a bona fide thin filament protein: electron microscopy of smooth muscle tropomyosin paracrystals indicated that calponin binds to a site 16-17 nm from the C terminus of tropomyosin with 40 nm periodicity, i.e. identical to the binding pattern of skeletal muscle troponin T (11). The thin filament-bound form of calponin is degraded 500 times more slowly by calpain than is the free form of calponin, suggesting a very close association between calponin and the thin filament similar to the association of troponin T with the skeletal muscle thin filament (12). Calponin is clearly a distinct protein from caldesmon and myosin light chain kinase (lo), but it is antigenically related to the C-terminal half of rabbit skeletal and bovine cardiac troponin T (13). Recently, however, Lehman (14) has suggested that calponin may be a cytoskeletal or nuclear matrix protein rather than a thin filament component (see "Discussion").
During characterization of native thin filaments prepared from chicken gizzard smooth muscle, we observed a 32-kDa protein in addition to actin, tropomyosin, and caldesmon (15). This protein was identified as calponin (13), a conclusion which we have confirmed using specific polyclonal antibodies to isolated gizzard calponin.
The properties, including the binding to actin and tropomyosin, suggested that calponin may function in the regulation of smooth muscle contraction. To investigate this possibility, we carried out preliminary studies of the effects of purified calponin on the actin-activated myosin MgATPase in vitro using purified smooth muscle proteins: actin, myosin, tropomyosin, calmodulin and myosin light chain kinase (16 1 shows that calponin produced a concentration-dependent inhibition of actomyosin ATPase with maximal (79%) inhibition being reached at concentrations of calponin >2 pM. Inhibition was shown to be due to calponin since inhibitory activity could be removed by immunoprecipitation with specific polyclonal antibodies to calponin (Table I). The inhibitory effect of calponin is not a nonspecific effect due to its basic nature (p1 = 8.4-9.1 (13)). Two other basic proteins, ribonuclease A (p1 = 9.6) and chymotrypsinogen (p1 = 9.5), exhibited no inhibitory effects on the actomyosin MgATPase, even at concentrations as high as 10 PM: ATPase rates of Actomyosin ATPase rates were measured as described under "Experimental Procedures" at the indicated concentrations of calponin in the presence (0) and absence (W) of Ca*+. Myosin phosphorylation levels were quantified on the same samples as described under "Experimental Procedures" in the presence (0) and absence (0) of Ca*+. Myosin phosphorylation data are presented as mean + SE. of 5-10 observations at each calponin concentration, and in the case of MgATPase rates data are the mean k S.E. of four to six observations, except at 7.5 and 10 pM calponin where these were three and two observations, respectively. Actomyosin" -Ca*+ 6.6 Actomyosin + Caz+ 141.7 Actomyosin + Ca*+ + supernatantb 125.0 Actomyosin + Ca2+ + buffer control' 133.5 Actomyosin + Ca*+ + antibody controk' 37.6 ' Actomyosin refers to the standard actomyosin ATPase reaction system.
b Supernatant refers to the supernatant remaining after immunoprecipitation of calponin.
'Buffer control refers to the supernatant from a control immunoprecipitation reaction from which the antigen (calponin) was omitted.
dAntibody control refers to the supernetant from a control immunoprecipitation reaction from which the antibody (anti-calponin) was omitted.
122.2 and 120.0 nmol of Pi/mg myosin.min were observed in the presence of 2 and 10 JLM ribonuclease A, respectively, and of 118.8 and 114.9 nmol of P,/mg myosin. min in the presence of 2 and 10 PM chymotrypsinogen, respectively. The control ATPase rate in these experiments was 120.7 nmol of PJmg myosin f min, which was reduced to 26.9 nmol Pi/mg myosin ' min in the presence of 2 PM calponin. Calponin also inhibited superprecipitation of actomyosin in the reconstituted system (data not shown). On the other hand, myosin light chain phosphorylation was unaffected (Fig. l), even at calponin concentrations as high as 17.5 pM, in which case the phosphorylation level was determined to be 1.82 2 0.05 mol of Pi/ mol myosin (n = 9). SDS-PAGE followed by autoradiography verified that specific phosphorylation of the 20,000-dalton myosin light chain occurred in all cases (data not shown).
These results indicate that inhibition of the actomyosin ATPase is not due to inhibition of myosin phosphorylation as a consequence, for example, of removal of calmodulin by binding to calponin. The mechanism of inhibition was further investigated by examining the effect of calponin on the actinactivated MgATPase activity of prephosphorylated myosin (Fig. 2). Myosin was first. phosphorylated in the presence of myosin light chain kinase and Ca'+-calmodulin to a level of 1.7 mol of Pi/m01 myosin. Actin was then added in order to activate the myosin MgATPase and the incubation continued in the presence or absence of calponin. In a third experiment, calponin was added 5 min after the addition of actin rather than with actin. As expected, the MgATPase of phosphorylated myosin was strongly activated by actin, to 181.7 nmol of Pi/mg myosin + min (Fig. 2, 0); in a separate experiment the ATPase rate of unphosphorylated myosin in the presence of actin was determined to be only 1.1 nmol of Pi/mg myosin. min. Calponin added with actin caused a 78% decrease in the rate of ATP hydrolysis, to 39.7 nmol of Pi/mg myosin.min (Fig. 2, A), i.e. comparable to the inhibition seen in Fig. 1 where ATP hydrolysis and myosin phosphorylation occurred concomitantly.
Calponin is capable not only of inhibiting actin activation of the MgATPase of prephosphorylated myosin, but also inhibits the actin-activated MgATPase when added to the fully activated system: addition of calponin 5 min after actin resulted in 96% inhibition of the maximal rate of ATP hydrolysis, i.e. from 173 to 7 nmol of Pi/mg myosin. min (Fig. 2, A).
Inhibition of the actomyosin MgATPase by calponin is reversible by increasing the concentration of actin and tropomyosin. Under standard conditions, 2 PM calponin resulted in 74% inhibition (Fig. 3). This inhibition was progressively lost as the actin and tropomyosin concentrations were raised (maintaining a 3:l molar ratio of actin:tropomyosin) until, at 60 pM actin and 20 PM tropomyosin, inhibition almost completely disappeared.
On the other hand, if the calponin concentration was also increased (e.g. 20  Myosin was maximally phosphorylated by incubation for 8 min under the ATPase assay conditions described under "Experimental Procedures" with the exception that actin was omitted (tropomyosin was included). The following additions were then made simultaneously: 6 pM actin (0, A) and 6 pM actin + 5 pM calponin (A). Samples of reaction mixtures were withdrawn at the indicated times following actin addition for quantification of ATP hydrolysis. In one case (A) calponin (5 pM) was added 5 min after the addition of actin. Actomyosin MgATPase rates were measured as described under "Experimental Procedures" in the presence and absence of 2 pM calponin and at the indicated concentrations of tropomyosin. The actin concentration was also varied so as to maintain a 3:l molar ratio of actin:tropomyosin. Control ATPase rates were 114 nmol of PJmg myosin.min in the presence of Ca2+ and 7 nmol of P,/mg myosin. min in the absence of Ca*+. Myosin was prephosphorylated as described in the legend to Fig. 3 prior to the addition of 6 pM actin with or without 5 pM calponin in the presence of 0.1 mM CaCl* or 1 mM EGTA. ATPases and myosin phosphorylation levels were quantified as described under "Experimental Procedures." However, a clear signal was observed after transblotting 10 kg of purified calponin. Myosin was prephosphorylated as before prior to addition of actin with or without calponin in the presence or absence of Ca*+ (Table II). Inhibition by calponin was observed both in the presence of Ca*+ (65% inhibition) and in the absence of Ca*+ (78% inhibition).
The lower ATPase rates observed in the absence of Ca*+ in each case were due to a low level of dephosphorylation of myosin by Ca'+-independent phosphatase activity once Ca2+ was removed from the assay system (Table II). However, to overcome this problem of dephospho-rylation during reactions in the absence of Ca2+, we repeated the experiments of Table II using thiophosphorylated myosin which is resistant to the action of myosin phosphatase (42). Experimental conditions were exactly as described in Table  II except that ATPyS replaced ATP in the prephosphorylation step. ATPase rates were then measured after addition of ATP and actin in the presence and absence of calponin and in the presence and absence of Ca'+. In the presence of Ca*', calponin inhibited the ATPase rate by 63.1% and in the absence of Ca2+ by 63.6%. We conclude, therefore, that the inhibitory effect of calponin on the actin-activated myosin MgATPase is Ca2+-independent.
Phosphorylation of Culponin-Calponin inhibition of the actomyosin ATPase is not, therefore, regulated directly by Ca*+ or by Ca"-calmodulin (calponin binds calmodulin in a Ca*+-dependent manner (9)). We considered the possibility, however, that calponin may be regulated by phosphorylation. Calponin was found to be phosphorylated by protein kinase C (to 1.0 mol of PJmol), by Ca*+/calmodulin-dependent protein kinase II (to 1.0 mol of Pi/mol) and by a Ca*+/calmodulindependent protein kinase copurifying with smooth muscle caldesmon (to 1.9 mol of PJmol), but not by CAMP-or cGMPdependent protein kinases or myosin light chain kinase (~0.01 mol of PJmol).
In each case, specific phosphorylation of calponin was verified by SDS-PAGE and autoradiography (data not shown). We previously published evidence suggesting that the Ca'+/calmodulin-dependent protein kinase activity copurifying with caldesmon actually resides within the caldesmon molecule itself (24). However, this conclusion is not consistent with the recently deduced amono acid sequence of caldesmon (43). The properties of this kinase most closely resemble those of Ca'+/calmodulin-dependent protein kinase II (44). We, therefore, examined the site specificity of phosphorylation of calponin by this kinase activity and by purified bovine brain Ca*+/calmodulin-dependent protein kinase II using two-dimensional phosphopeptide mapping of limit tryptic peptides (Fig. 4). Identical phosphopeptide maps were obtained (compare Fig. 4, A and B), so we conclude that the kinase activity copurifying with caldesmon is a smooth muscle form of Ca2'/calmodulin-dependent protein kinase II. Phosphopeptide mapping of calponin phosphorylated by protein kinase C revealed a simpler pattern of phosphopeptides ( Fig.  4C), only three major phosphopeptides being detected. If tryptic digests of calponin phosphorylated by protein kinase C and Ca"/calmodulin-dependent protein kinase II were combined prior to two-dimensional phosphopeptide mapping, the three phosphopeptides (labeled 1-3) comigrated (data not shown) suggesting that the two major sites of phosphorylation by Ca*+/calmodulin-dependent protein kinase II (sites 1 and 2) are also phosphorylated by protein kinase C. Site 3 which is strongly phosphorylated by protein kinase C is poorly phosphorylated by Ca2+/calmodulin-dependent protein kinase II. Additional sites of phosphorylation by Ca2+/calmodulindependent protein kinase II are apparent (Fig. 4, A and B), but are minor relative to sites 1 and 2.
Thin layer electrophoresis of acid hydrolysates of phosphorylated calponin followed by autoradiography indicated that Ca*+/calmodulin-dependent protein kinase II and protein kinase C phosphorylated both serine and threonine residues. Neither kinase incorporated radiolabeled phosphate onto tyrosine residues.
The effect of phosphorylation of calponin on its ability to inhibit the actomyosin MgATPase was then investigated. Calponin was phosphorylated to 1.11 mol of Pi/m01 calponin by Ca'+/calmodulin-dependent protein kinase II and to 1. Approximately 1 rg of each limit tryptic digest of phosphorylated calponin was subjected to two-dimensional phosphopeptide mapping as described by Colburn et al. (29). Thin layer electrophoresis (TLE) was performed from anode @ to cathode 8 followed by ascending thin layer chromatography (TLC). Autoradiographs presented are: A, calponin phosphorylated by Ca'+/calmodulin-dependent protein kinase II (1.0 mol of P,/mol calponin); B, calponin phosphorylated by caldesmon with copurifying kinase activity (1.9 mol of P,/mol calponin); C, calponin phosphorylated by protein kinase C (1.0 mol of P,/mol calponin). 0, represents the origin and 1, 2, and 3 represent the major phosphopeptides referred to in the text. ATPase were examined (Fig. 5). At concentrations as high as 5 FM, calponin phosphorylated by Ca'+/calmodulin-dependent protein kinase II or protein kinase C had little or no effect on the actinactivated myosin MgATPase when compared with unphosphorylated calponin. Partial inhibition (from 117 to 89 nmol of Pi/mg myosin. min) was observed in other experiments in the presence of 8 j.tM phosphorylated calponin. The same concentration of unphosphorylated calponin reduced the ATPase rate to 6 nmol of Pi/mg myosin.min. Therefore, calponin loses its ability to inhibit the actomyosin ATPase upon phosphorylation.
Although regulation of actomyosin MgATPase by calponin is not mediated directly by calcium, but via calcium-dependent kinases, the possibility exists that calponin could be acting in a troponin-like manner by immobilizing tropomyosin on the actin filament, i.e. not permitting tropomyosin to move relative to the actin filament in such a way as to permit normal actin-myosin interaction. In the presence of actin (7 PM) and presence or absence of 1 pM tropomyosin, calponin produced a concentration-dependent inhibition of the actin-activated myosin MgATPase (Fig. 6A) reaching maximum (70-80%) inhibition in both cases at or above '2 pM calponin. This effect can be seen more clearly in Fig. 6B where the data have been normalized in order to compensate for the lower ATPase rates seen in the absence of tropomyosin. Calponin inhibition of the actin-activated MgATPase of myosin is not, therefore, dependent on the presence of tropomyosin.
It is known from the work of Takahashi et al. (9,13) that calponin binds to calmodulin, tropomyosin, and actin. Examination of the effects of phosphorylation on these binding properties of calponin may help to shed light on the mechanism whereby calponin inhibits the actin-activated myosin MgATPase. We found that phosphorylation of calponin did not affect its interaction with either calmodulin-Sepharose or tropomyosin-Sepharose (data not shown) but, as shown in Fig. 7A, did affect its binding to actin. Unphosphorylated calponin bound to actin and to actin-tropomyosin as determined in a sedimentation assay (Fig. 7A, lanes 1 and 2, 5 and 6). Upon phosphorylation by either protein kinase C or Ca*+/ calmodulin-dependent protein kinase II, most of the calponin did not bind to actin or actin-tropomyosin ( Actomyosin ATPase rates were measured as described under "Experimental Procedures" at the indicated concentrations of calponin in the presence of actin (7 FM) and in the presence (0, n ) or absence (0) of 1 pM tropomyosin and in the presence (0,O) or absence (m) of Ca'+. A, data presented as ATPase rate in nanomoles of Pi/mg myosin. min. B, data normalized by expressing ATPase rate in the presence or absence of tropomyosin as a percentage of the maximum. A, actin and phosphorylated or unphosphorylated calponin were sedimented at high speed in the presence and absence of tropomyosin as described under "Experimental Procedures." Supernatants (S) and pellets (P) were analyzed by SDS-PAGE. Lanes 1 and 2, actin + unphosphorylated calponin; lanes 3 and 4, actin + phosphorylated calponin; lanes 5 and 6, actin + tropomyosin + unphosphorylated calponin; lanes 7 and 8, actin + tropomyosin + phosphorylated calponin. In this experiment, calponin was phosphorylated to 1.0 mol of Pi/mol calponin by Ca'+/calmodulin-dependent protein kinase II. Identical results were obtained if calponin was phosphorylated to 1.0 mol of Pi/mol calponin by protein kinase C. In separate experiments, calponin alone (phosphorylated or unphosphorylated) did not sediment under these conditions. B, phosphorylated or unphosphorylated calponin (2 pM) was sedimented at high speed in the presence of actin (6 j.~cM), myosin (1 PM), tropomyosin (2 PM), myosin light chain kinase (74 nti) and calmodulin (1 pM) as described under "Experimental Procedures." Supernatants (S) and pellets (P) were analyzed by SDS-PAGE. Lane 1, molecular weight markers (a, phosphorylase b,-97,400, b, bovine serum albumin, 66,200; c, ovalbumin, 45,000; d, carbonic anhydrase, 29,000; e, soybean trypsin inhibitor, 20,100, /, lysozyme, 14,400); lanes 2 and 3, myosin, actin, tropomyosin, myosin light chain kinase, and calmodulin alone; lanes 4 and 5, as lanes 2 and 3 but plus unphosphorylated calponin; lanes 6 and 7, as lanes 2 and 3 but plus phosphorylated calponin. In this experiment calponin was phosphorylated to 2.3 mol of Pi/m01 calponin by Ca*+/calmodulin-dependent protein kinase II. M, myosin; A, actin; Tm, tropomyosin; Cap, calponin; L&, 20-kDa light chain of myosin; LC17, 17-kDa light chain of myosin. tropomyosin (Fig. 7B, lanes 4 and 5) and was recovered predominantly in the pellet along with myosin. Phosphorylated calponin clearly does not bind to actin-tropomyosin and was recovered in the supernatant along with some unsedimented actin (Fig. 7B, lanes 6 and 7). Quantitative data (Table  III) were obtained by laser densitometry of the gels shown in Fig. 7.

DISCUSSION
In spite of being a major protein component of smooth muscle, calponin was only recently reported (9). Its properties, as determined by , suggest that it is a bona fide thin filament protein, although this has recently been challenged by Lehman (14) who suggested that a 35-kDa protein identified as calponin may be part of the insoluble cytoskeleton or possibly a component of the nuclear matrix. This suggestion was based on two observations: first, thin filaments immunoprecipitated with anti-caldesmon did ATPase system + P-Cap 100.0 0.0 ' Results were obtained by laser densitometry of the gels shown in Fig. 7, A and B. not contain calponin and, secondly, gizzard "ghost" cells from which actin and myosin were extracted were enriched in the 35-kDa protein. However, the results of the immunoprecipitation experiments suggest that calponin may have been completely proteolyzed, a reasonable possibility given the prolonged incubation at 25 "C and its known susceptibility to proteolysis. Furthermore, if comparable amounts of the histones in washed muscle and ghost cells were loaded on the gels in Fig. 6, a and b, of Ref. 14, it would be apparent that the amount of calponin retained in the ghost cells is small relative to that extracted with actin and myosin. The weight of evidence, therefore, favors calponin as a thin filamentassociated protein which, as shown by Takahashi et al. (lo), is specific to smooth muscle and is, therefore, of interest as a potential regulator of smooth muscle contraction. This possibility is supported by the in vitro binding properties of calponin: it interacts with F-actin and tropomyosin in the presence or absence of Ca*+, and with calmodulin only in the presence of Ca*+ (9, 13). We initially encountered this protein in preparations of native thin filaments from chicken gizzard (15); it is clear that the 32-kDa protein observed in such preparations is identical to calponin (13).* Consequently, we have investigated the effects of purified calponin on the actinactivated myosin MgATPase in an in vitro system reconstructed from purified smooth muscle contractile and regulatory proteins: actin, myosin, tropomyosin, calmodulin, and myosin light chain kinase.
Calponin inhibits the actomyosin ATPase maximally (-80%) when present in the assay system at a concentration equimolar to tropomyosin. Takahashi et al. (9) have shown that the molar concentrations of calponin and tropomyosin in chicken gizzard smooth muscle are equal. The inhibitory effect of calponin on the actomyosin ATPase is apparently due to its interaction with actin and is clearly not due to inhibition of myosin phosphorylation. Furthermore, although calponin can bind Ca*+ directly (40), its ability to inhibit the actomyosin ATPase is not Ca*+-dependent. This observation prompted us to search for an alternative mechanism of regulating calponin function. Phosphorylation was an obvious one to investigate. We found that calponin could be phosphorylated by protein kinase C and Ca*+/calmodulin-dependent protein kinase II, but not by CAMP-or cGMP-dependent protein kinases or by myosin light chain kinase. Most interestingly, phosphorylation by either protein kinase C or Ca"+/ calmodulin-dependent protein kinase II abolished inhibition of actomyosin ATPase activity by calponin. Phosphorylation also blocked the actin-calponin interaction. However, phosphorylation did not affect the interaction of calponin with ' S. J. Winder and M. P. Walsh, unpublished observations.