Role of Caldesmon in the Ca2+ Regulation of Smooth Muscle Thin Filaments

Smooth muscle thin filaments are made up of actin, tropomyosin, caldesmon, and a Ca2+-binding protein and their interaction with myosin is Ca2+-regulated. We suggested that Ca2+ regulation by caldesmon and Ca2+-calmodulin is achieved by controlling the state of thin filament through a cooperative-allosteric mechanism homologous to troponin-tropomyosin in striated muscles. In the present work, we have tested this hypothesis. We monitored directly the thin filament transition between the ON and OFF state using the excimer fluorescence of pyrene iodoacetamide (PIA)-labeled smooth muscle αα-tropomyosin homodimers. In steady state fluorescence measurements, myosin subfragment 1 (S1) cooperatively switches the thin filaments to the ON state, and this is exhibited as an increase in the excimer fluorescence. In contrast, caldesmon decreases the excimer fluorescence, indicating a switch of the thin filament to the OFF state. Addition of Ca2+-calmodulin increases the excimer fluorescence, indicating a switch of the thin filament to the ON state. The excimer fluorescence was also used to monitor the kinetics of the ON-OFF transition in a stopped-flow apparatus. When ATP induces S1 dissociation from actin-PIA-tropomyosin, the transition to the OFF state is delayed until all S1 molecules are dissociated actin. In contrast, caldesmon switches the thin filament to the OFF state in a cooperative way, and no lag is displayed in the time course of the caldesmon-induced fluorescence decrease. We have also studied caldesmon and Ca2+-calmodulin-caldesmon binding to actin-tropomyosin in the ON and OFF states. The results are used to discuss both caldesmon inhibition and Ca2+-calmodulin-caldesmon activation of actin-tropomyosin.

The contractile system of smooth muscles is based on actomyosin interaction like all muscles; however, it is adapted for the maintenance of sustained isometric force and slow contraction. Vertebrate smooth muscle is a dual regulated muscle: activation of myosin by phosphorylation of light chains by myosin light chain kinase and its dephosphorylation by a phosphatase is the prime regulator of smooth muscle contractility. However it is well established that the activity of the thin filaments toward myosin is independently regulated by Ca 2ϩ because native thin filaments isolated from smooth muscles confer a Ca 2ϩ -dependent regulation on unregulated myosin from skeletal or smooth muscle (1)(2)(3).
Smooth muscle thin filaments are made up of actin, tropomyosin, caldesmon, and a Ca 2ϩ -binding protein (CaBP) 2 with a stoichiometry of 1 caldesmon and CaBP per two tropomyosin and fourteen actin monomers in native thin filaments; synthetic Ca 2ϩ -regulated thin filaments may be reconstituted from these components at similar ratios (4,5). There is substantial evidence that caldesmon-based regulation is involved in modulating smooth muscle Ca 2ϩ sensitivity and relaxation (6 -8).
Like skeletal muscle thin filaments, the thin filaments of smooth muscles are negatively regulated: the function of caldesmon binding to actin-tropomyosin is to inhibit the activity of a constitutively active filament and the function of Ca 2ϩ binding to the CaBP is to reverse the inhibition. Caldesmon inhibition is cooperative with up to 14 actin monomers being inhibited by the binding of one caldesmon molecule to actin-tropomyosin; moreover, Ca 2ϩ and CaBP (or calmodulin) interacting with caldesmon activate thin filaments to up to 150% of the activity of actin-tropomyosin rather than simply neutralizing the inhibitory effect of caldesmon (5,9,10).
There has been considerable debate about the mechanism of smooth muscle thin filament regulation. We have consistently argued that the only model of regulation that can account for all the regulatory characteristics is a cooperative allosteric mechanism analogous to troponin-tropomyosin in striated muscle thin filaments (10 -15). Alternative models have been proposed that include a role for mutually exclusive competitive binding of caldesmon or myosin heads to actintropomyosin in determining thin filament interaction with myosin. The original model of Sobue et al. (16) proposed a purely competitive "flip-flop" mechanism, which was ruled out by measurements showing that caldesmon and S1⅐ADP⅐P i * This work was supported by a Research Career Development Fellowship could bind simultaneously to actin-tropomyosin (12). However, several studies appeared to show a relationship between caldesmon inhibition and S1⅐ADP⅐P i displacement (17)(18)(19)(20)(21)(22). We have criticized such experiments for using unphysiologically high concentrations of caldesmon relative to actin. Displacement models cannot account for the activating property of Ca 2ϩ -CaBP and Ca 2ϩ -calmodulin and are ruled out by the consistent observation that both caldesmon and calmodulin or CaBP remain bound to actin-tropomyosin under activating conditions (9,10,23). Nevertheless, it remains possible that smooth muscle thin filament regulation involves mixed cooperative and competitive regulation or even an entirely novel process (24).
The critical test for any regulatory scheme is that it produces experimentally testable predictions that can rule out other models. The concerted allosteric-cooperative transition model proposes two states of actin-tropomyosin termed ON and OFF. Myosin weak binding complex (M⅐ADP⅐P i ) affinity is the same for both states but strong binding complexes can only be formed with thin filaments in the ON state. Crossbridge cycling and force production is thus only possible when filaments are in the ON state, therefore the activity of the thin filament depends on the proportion of actin-tropomyosin in the ON state. Caldesmon is proposed to act as an allosteric inhibitor by binding preferentially to the OFF state while Ca 2ϩ -CaBP-caldesmon activates thin filaments by binding preferentially to the ON state. This mechanism is fundamentally the same as the mechanism for troponin regulation of actin-tropomyosin filaments in striated muscles (25)(26)(27).
We have previously demonstrated caldesmon inhibition with no change in the binding of myosin subfragment 1 (S1) weak binding complexes (S1⅐ADP⅐P i ) and the cooperative binding of strong-binding complexes (S1⅐ADP and S1⅐AMP⅐PNP) to caldesmon-inhibited actin-tropomyosin (12,13). These findings are compatible with the proposed mechanism. More recently we have investigated thoroughly the effect of caldesmon on the elementary steps of the actomyosin ATPase (28). We found that caldesmon had very little effect on the rate of S1 binding to actin-tropomyosin, acto-S1 dissociation by ATP and the rate of ADP release. The rate of phosphate release was drastically reduced. We suggested that caldesmon inhibition of phosphate release is caused by the thin filament being switched to an inactive (OFF) state.
In this report we have tested the hypothesis that smooth muscle thin filaments change state in both inhibitory and activating conditions. To do this we have used the excimer fluorescence of pyrene iodoacetamide-labeled smooth muscle tropomyosin as a probe of the thin filament ON-OFF transition and studied caldesmon and Ca 2ϩ -calmodulin-caldesmon binding to actin-tropomyosin and have related this to the control of thin filament activity as measured by actin activation of myosin S1 Mg 2ϩ -ATPase. The results support our hypothesis both for caldesmon inhibition of actin-tropomyosin and for Ca 2ϩ -calmodulin-caldesmon activation of actin-tropomyosin and also permit us to determine values of the size of the cooperative unit and the rates of transition and equilibrium constants between the ON and OFF states.
Steady State ATPase and Binding Measurements-Actintropomyosin activation of S1 Mg 2ϩ -ATPase was assayed as described previously (32). Phosphate liberated following ATP hydrolysis was measured by the method of Taussky and Schorr (38). [ 14 C]Caldesmon binding to actin-tropomyosin was measured by co-sedimentation as previously described (5).
Steady State Fluorescence Titrations-Steady state fluorescence measurements were obtained with a Fluoromax-2 photon counting fluorimeter in the ratio mode. Titrations were carried out with excitation at 343 nm and emission at 350 nm to monitor light scattering and 485 nm to monitor the excimer fluorescence of PIA-tropomyosin (37). Before titrations, all proteins were clarified by centrifugation in a refrigerated bench top centrifuge for 10 min at 25,000 ϫ g. All buffers used were filtered before use. The temperature was maintained at 25°C by a circulating water bath.
Stopped-flow Experiments-All transient kinetic measurements were performed on a Hi-Tech Scientific SF-61 double mixing stopped-flow system using a 100 watt Xe/Hg lamp and a monochromator for excitation wavelength selection as previously described (28). Pyrene iodoacetamide fluorescence was excited at 364 nm and emission of excimer fluorescence was monitored through a 455 nm cut-off filter. Light scattering was observed at 90 o to the incident beam using a UG-5 filter (light over 400 nm was cut off). The measurements were carried out in ATPase buffer at 20°C unless otherwise stated (28). The temperature of the stopped-flow machine was maintained within 0.1°C during the course of the experiment. All buffers were filtered and all proteins were clarified by centrifugation in a refrigerated bench top centrifuge for 10 min at 25,000 ϫ g just before use. Usually four to nine transients were collected and averaged. The data were then fitted to one or two exponentials by a non-linear least square curve fit using the software provided by Hi-Tech. The stated concentrations of reactants are those after mixing in the stopped-flow observation cell. Fig. 1 illustrates the regulatory properties of reconstituted smooth muscle thin filaments. Addition of caldesmon to actin-smooth muscle tropomyosin resulted in inhibition to 6.6 Ϯ 3.6% of the actin-tropomyosin-activated ATPase activity at low stoichiometry ( Fig.  1A). 80% inhibition of ATPase was obtained with 0.074 Ϯ 0.16 caldesmon added per actin (six preparations) in agreement with previous measurements (5,11,12). Chicken gizzard and sheep aorta caldesmon were indistinguishable in this assay, and the inhibition was independent of temperature (20 -37°C) and KCl concentration. In suitable conditions (120 mM KCl, 37°C), 10 -20 M Ca 2ϩ -calmodulin (Ca 2ϩ -CaM)-activated actin-tropomyosin, which had been inhibited by caldesmon. ATPase activity with saturating quantities of Ca 2ϩ -calmodulin was activated to 115 Ϯ 2% of the uninhibited actin-tropomyosin ATPase (six preparations) (Fig. 1B). This level of activation is similar to that previously observed for Ca 2ϩ -calmodulin (9,39). Under comparable conditions NEM-S1, which switches thin filaments to the ON state, activated actin-tropomyosin to 200% of actin-tropomyosin ATPase at a ratio of 0.15 NEM-S1:1 actintropomyosin (Fig. 1C).

Inhibition of Actin-Tropomyosin by Caldesmon and Activation by Caldesmon-Ca 2ϩ -Calmodulin-
Detection of the ON/OFF Equilibrium by PIA-Tropomyosin Excimer Fluorescence Changes-We monitored the equilibrium between the ON and OFF states of actin-tropomyosin by the excimer fluorescence technique of Ischii and Lehrer (36).
Smooth muscle ␣␣-tropomyosin dimers were covalently labeled with pyrene iodoacetamide at 1.9 Ϯ 0.08 mol/mol (number of experiments ϭ 10). Fig. 2A shows fluorescence emission spectra of reconstituted actin-tropomyosin. Monomer fluorescence peaks are at 385 and 410 nm and excimer fluorescence is at 480 nm. When S1 is added to actin to turn the filaments ON, the excimer fluorescence is increased, and when S1 is displaced from actin by adding Mg 2ϩ -ATP, the excimer fluorescence returns to the original level. The average increase in excimer fluorescence induced by S1 using smooth muscle PIA-␣␣-tropomyosin was 55.6 Ϯ 1.5% (number of experiments ϭ 10).
Titration of actin-tropomyosin with increasing concentrations of S1 shows that the switch between states is cooperative: pyrene excimer fluorescence increased to a plateau at much lower S1 concentrations than S1 binding monitored by light scatter (Fig. 2B). The fraction of the thin filaments in the ON state, ƒ on determined from the excimer fluorescence, was plotted against F b , the fraction of S1 bound to the thin filament calculated from the light scattering signal. Fitting the data to the equation ƒ on ϭ 1 Ϫ (1 Ϫ F b ) n generates the cooperative unit size, n (Fig. 2C) as described by Geeves and Lehrer (68). A mean value for n of 10 Ϯ 1 was obtained from three separate experiments.
Cooperativity of the OFF/ON transition of actin-tropomyosin can also be determined from the transient kinetics of S1 binding to actin-tropomyosin (Fig. 3). When S1 is in excess over actin, the excimer fluorescence transient is faster than the light scattering supporting the assumption that the 2 signals monitor 2 different processes. Increasing S1 concentration lead to a hyperbolic increase of the 2 observed rate constants (k obs fluorescence and k obs light scattering ) indicating that each of the 2 signals is monitoring a 2-step reaction. The first step correspond to S1 binding to the thin filaments, while the second step correspond to different processes. In the case of light scattering (monitoring S1 binding), the second step corresponds to an isomerization to a strongly bound acto-S1 com- NEM S1, µM A B C ATPase, % actin-tropomyosin FIGURE 1. Control of actin-tropomyosin activation of S1 ATPase by caldesmon, Ca 2؉ -calmodulin, and NEM-S1. A represents the effect of caldesmon (0 -12 M) on actin-tropomyosin activation of skeletal muscle myosin S1 Mg 2ϩ -ATPase activity. The ATPase was measured at 37°C, using 2 M skeletal muscle S1, 12 M skeletal muscle actin, 4 M smooth muscle tropomyosin, in 120 mM KCl, 2.5 mM MgCl 2 , 5 mM PIPES, pH 7.1, 1 mM NaN 3 , 1 mM DTT. The reaction was initiated with 5 mM Mg 2ϩ -ATP. B represents reactivation of the Mg 2ϩ -ATPase activity preinhibited by 0.8 M caldesmon, by increasing concentration of Ca 2ϩ -calmodulin (0 -50 M). C shows the activation of actin-tropomyosin-S1 Mg 2ϩ -ATPase activity by increasing concentration of NEM-S1 (0 -4 M). The data are expressed as percent of uninhibited actin-tropomyosin activation of skeletal muscle myosin S1 Mg 2ϩ -ATPase activity (mean value of uninhibited ATPase is 0.35 s Ϫ1 ).
plex. In the absence of any added nucleotide, this usually plateaus around 250 s Ϫ1 (28). In the case of the excimer fluorescence (monitoring the thin filament switching between the OFF and ON states), the second step correspond to the transition from the OFF to the ON state. The hyperbolic fit gave a maximum rate of 833 s Ϫ1 for the OFF to ON transition albeit, it is anticipated that this value contains a large error because the experimental data cover less than half of the entire curve. Nevertheless we can put a lower limit of 350 s Ϫ1 for the maximum rate of the OFF to ON transition (The highest experimentally determined value, Fig. 3B). The slopes in the initial part of the 2 curves depend on the rate of S1 binding. It was previously modeled that the ratio of the 2 slopes (k obs fluorescence /k obs light scattering ) cor-responds to the size of the cooperative unit n (68). Using this kinetic technique n was found to be 10, which is the same as determined from the steady state titrations. Caldesmon and Ca 2ϩ -Calmodulin Switch Actin-Tropomyosin between the OFF and ON States-We used the change in pyrene excimer fluorescence to determine whether caldesmon would switch actin-tropomyosin to the OFF state. Caldesmon reduced pyrene excimer fluorescence by 19% and the quantity needed to fully switch the thin filament OFF, corresponded to the quantity needed to inhibit actin-tropomyosin activation of Mg 2ϩ -ATPase activity (Fig. 4A). The excimer fluorescence reached 80 Ϯ 5% of the minimum level with 1:7 caldesmon added per actin. We confirmed that caldesmon switches actintropomyosin completely to the OFF state by comparison with skeletal muscle troponin in the absence of Ca 2ϩ : this induced the same decrease of excimer fluorescence of smooth muscle tropomyosin and subsequent addition of caldesmon had no effect on the excimer fluorescence (data not shown). αα-smTm αα-smTm + actin αα-smTm + actin + S1 αα-smTm + actin + S1 + MgATP Wavelength, nm   We determined the effect of Ca 2ϩ -calmodulin upon actin-PIA-tropomyosin excimer fluorescence at 10 M caldesmon where the inhibition of ATPase and decrease of excimer fluorescence was at a maximum (Fig. 4B). Calmodulin in the presence of activating Ca 2ϩ concentrations switched the filaments toward the ON state. The maximum fluorescence was 30% greater than that of actin-tropomyosin (K T ϭ 0.9 compared with 0.6 for actin-tropomyosin) and the concentration of Ca 2ϩcalmodulin needed to give maximal increase in excimer fluorescence was in the same range as reversal of inhibition (Fig. 1).
Addition of S1 to saturating actin-tropomyosin-caldesmon increased the fluorescence back up to the level of the fully ON filament (Fig. 5A). The transition was cooperative with an estimated cooperative unit size in the range 3-5. The further addition of caldesmon to actin-tropomyosin-S1 gave a decrease in fluorescence, which was only partial in the concentration range studied, where caldesmon affinity for actin is less than the S1 affinity (Fig. 5B).
Because we have measured the fluorescence of the fully OFF state (the fluorescence plateau in the presence of caldesmon) and the fully ON state (the fluorescence plateau in the presence of S1) we can calculate the equilibrium constant K T for actintropomyosin. For smooth muscle ␣␣-tropomyosin under our conditions K T was 0.60 Ϯ 0.05 (number of experiments ϭ 3). K T could be increased by increasing temperature or KCl concentrations.
Kinetics of the ON-OFF Transition-We induced the transition from the ON to the OFF state using either ATP to dissociate pre-mixed actin-tropomyosin-S1 complex or caldesmon added to actin-tropomyosin. Fig. 6A shows light scattering and excimer fluorescence changes following acto-S1 dissociation by 40 M ATP. The decrease in light scattering monitoring S1 dissociation from actin-tropomyosin is fast (130 s Ϫ1 ) and without delay. In contrast the excimer fluorescence transient representing the ON to OFF transition showed a lag (displayed as an upward curvature) before an exponential decrease (rate 67 s Ϫ1 ). The transition to the OFF state is delayed until all S1 molecules are dissociated from any single cooperative unit. If caldesmon is rapidly mixed with actin-tropomyosin, the excimer fluorescence decreased exponentially (Fig. 6B). The transient was best fit by a sum of 2 exponentials. The faster component represented about 75% of the total signal. Because the fluorescence change is very rapid (Ͼ400 s Ϫ1 ) and because of the relatively poor signal to noise ratio, there is a large error in the fitted rate constants; however, this rate is comparable with the value of Ͼ350 s Ϫ1 calculated for the transition from the OFF to the ON state (see section above for Fig. 3B). In contrast to the ATP induced acto-S1 dissociation, no lag is displayed in the time course of the caldesmon induced fluorescence decrease (Fig.  6B). These results suggest that each caldesmon binding to one actin switches the whole cooperative unit to the OFF state.

Caldesmon Binding to the ON and OFF States of Actin-Tropomyosin-
If actin-tropomyosin is fixed in either the ON or OFF state then predictions follow from the cooperative allosteric model that can uniquely distinguish it from other models. If actin-tropomyosin is in the ON state then caldesmon binding should be strongly inhibited and would become cooperative, conversely fixing actin-tropomyosin in the OFF state should enhance caldesmon binding. We used NEM-S1 to switch actin-tropomyosin ON and troponin I to switch actin-tropomyosin OFF. Fig. 7A shows [ 14 C]iodoacetamide-labeled caldesmon binding to 2.5 M actin-tropomyosin in the absence and presence of NEM-S1. Caldesmon binding to actin-tropo- myosin (open circle) is clearly biphasic, as has been previously noted (5). The affinity of the strong binding component is greater than 10 6 M Ϫ1 under these conditions. 50 ϫ 10 6 M Ϫ1 was determined by Smith et al. (5). The mean affinity of the weakly binding component was 0.32 ϫ 10 6 Ϯ 0.19 M Ϫ1 (number of experiments is 4). In the presence of NEM-S1 caldesmon affinity was greatly reduced, and the binding curve became sigmoid, indicating positive cooperativity (closed circles). The decreased caldesmon affinity at low caldesmon:actin ratios was 0.044 Ϯ 0.003 times the affinity in the absence of NEM-S1 based on the ratio of the initial slopes (four preparations). When actin-tropomyosin was switched OFF with troponin I, caldesmon binding was essentially the same as in the absence of troponin I (Fig.  7B). Because we have estimated K T to be 0.6 we would expect a maximum of 40% increase in affinity of the tight binding component, which would not be detectable in these measurements.
We then measured caldesmon binding to actin-tropomyosin in activating conditions (Ͼ0.5 M Ca 2ϩ and sufficient calmodulin to fully saturate the caldesmon). The calmodulin concentration had to be enough to saturate binding in all the samples; therefore, the concentration of calmodulin used was the concentration at which the plateau of reversal of inhibition occurred in the ATPase assay, which was 20 M plus the total caldesmon concentration used for the binding assays. The results are shown in Fig. 8A. Both curves are very similar and follow the same trend. There is high affinity binding at the low concentrations followed by low affinity binding. The presence of Ca 2ϩ -calmodulin did not appear to affect either phase of the binding. The effects of NEM-S1 and troponin I on CaD-Ca 2ϩ -CaM binding to actin tropomyosin was opposite to that of inhibitory caldesmon. NEM-S1 (ON state) produced no detectable effect upon the strong binding of CaD-Ca 2ϩ -CaM (Fig.  8B), while CaD-Ca 2ϩ -CaM binding to actin-tropomyosin in the presence of inhibitory concentration of troponin I (OFF state) is weakened at least 20ϫ and is now cooperative (Fig. 8C), indicating that the activated complex binds preferentially to the ON state as predicted by the cooperative-allosteric model.

DISCUSSION
The Ca 2ϩ -sensitive regulation of smooth muscle thin filaments and the requirement of actin, tropomyosin, caldesmon and a Ca 2ϩ -binding protein in this regulation is well established (1, 2, 40) but the mechanism of Ca 2ϩ regulation has been a controversial topic. We and others have produced indirect evidence for  a troponin-like cooperative allosteric mechanism in smooth muscle thin filaments (1,(11)(12)(13)(14)(15)41), however this article reports for the first time direct and unambiguous evidence for this mechanism. We demonstrate that smooth muscle thin filaments are a cooperative allosteric regulated system in which actintropomyosin exist in two activity states, ON and OFF that are linked by a concerted cooperative equilibrium. Caldesmon acts as an allosteric inhibitor by preferentially binding to actin-tropomyosin in the OFF state while at activating Ca 2ϩ concentration the Ca 2ϩ -CaBP-caldesmon complex acts as an activator of actin-tropomyosin activity by preferentially binding to the ON state (42). This mechanism is very similar to that established for troponin-tropomyosin as recently described by Lehrer and Geeves (43): caldesmon is equivalent to troponin I or troponin in the absence of Ca 2ϩ and Ca 2ϩ -CaBP-caldesmon is equivalent to troponin in the presence of activating Ca 2ϩ .
We have verified the hypothesis by direct measurement of the effects of caldesmon and Ca 2ϩ -CaBP-caldesmon on the actin-tropomyosin ON-OFF transition. Changes in the excimer fluorescence of pyrene iodoacetamide conjugated to skeletal muscle tropomyosin are sensitive to the states of actin-tropomyosin (36). In this article, we report that change in the excimer fluorescence of smooth muscle tropomyosin is also a direct probe of the state of the thin filament. We found that addition of S-1 to smooth muscle tropomyosin-actin increased excimer fluorescence. We believe that an increase in the excimer fluorescence is a probe of the cooperative switching of the smooth muscle thin filament to the ON state for the following reasons: 1) the fluorescence signal reaches a plateau before the actin filament is saturated with S1 as monitored by light scattering. 2) The excimer fluorescence does not change if the actin-PIAtropomyosin filament is already switched to the ON state (by preincubation with sub-saturating concentrations of NEM-S1 or S1). 3) The fluorescence increase is paralleled by activation of actin-tropomyosin ATPase activity by NEM-S1. Using the change in the excimer fluorescence (monitoring the ON-OFF transition) and light scattering (monitoring S1 binding to actin) we measured a cooperative unit size of 10 compared with 7 for skeletal muscle tropomyosin. A similar value for smooth muscle ␣␣-tropomyosin was previously reported by Lehrer et al. (44).
Addition of caldesmon decreased the excimer fluorescence, indicating that actin-tropomyosin was being switched to the OFF state (i.e. the opposite effect to adding S-1) and that the decrease was cooperative since it required less than 0.15 caldesmon/actin-tropomyosin for 80% of the fluorescence decrease. The decrease in excimer fluorescence depends on the amount of thin filament in the ON state. In our measurements it is detectable because under our conditions smooth muscle ␣␣-tropomyosin is initially about 60% in the OFF state and 40% in the ON state (K T ϭ 0.6 in contrast to less than 0.2 for skeletal muscle tropomyosin) (37). Addition of skeletal muscle troponin in the absence of Ca 2ϩ (at 1 troponin/7 actin) to the actinsmooth muscle tropomyosin filament induced the same decrease of excimer fluorescence and subsequent addition of caldesmon had no effect on the fluorescence. This is further evidence that a decrease in the excimer fluorescence by caldesmon is a probe of the filament being switched to the OFF state. It is possible to switch actin-tropomyosin-caldesmon back to the ON state by adding S1 or Ca 2ϩ -calmodulin.
The excimer fluorescence can be used to monitor the rate of transition between the ON and the OFF states. Addition of ATP to actin-smooth muscle tropomyosin preloaded with S1 (fully switched ON) led to a decrease in the excimer fluorescence and the transient showed a lag. The lag is due to the ability of S1 to maintain the thin filament in the ON state until the last S1 dissociates from a cooperative unit. If caldesmon is used to switch the thin filament to the OFF state, the transient showed no lag because each caldesmon is able to switch the whole cooperative unit to the OFF state. The rate of switching of the smooth muscle thin filament to the OFF state by caldesmon is Ͼ400 s Ϫ1 . Similarly, values of 250 s Ϫ1 (45) and Ͼ500 s Ϫ1 (68) have been reported for skeletal muscle thin filaments.
Adding calmodulin in the presence of activating Ca 2ϩ concentrations switches the filaments toward the ON state. The maximum fluorescence was 30% greater than that of actin-tropomyosin (K T ϭ 0.9 compared with 0.6 for actin-tropomyosin), indicating that Ca 2ϩ -calmodulin did not simply neutralize the inhibitory effect of caldesmon by dissociating it from actin as has been suggested (16,46,47). Control experiments also clearly show that caldesmon binding to actin-tropomyosin is unaffected by Ca 2ϩ -calmodulin (Fig. 8A). Thus Ca 2ϩ -calmodulin activated caldesmon-inhibited actin-tropomyosin by switching the thin filament to the ON state.
It is important to note that the changes in actin-tropomyosin state as monitored by excimer fluorescence correspond precisely to changes in thin filament activity as determined from measurements of activation of S1 ATPase activity. The level of excimer fluorescence and the level of actin activation of S1 ATPase is determined by the proportion of thin filaments in the ON state. Fluorescence and ATPase are reduced by low concentrations of caldesmon relative to actin-tropomyosin similar to the physiological ratio found in native thin filaments (1:16) (4) and activated by similar concentrations of Ca 2ϩ -calmodulin ( Figs. 1 and 3). Ca 2ϩ -calmodulin increased ATPase and fluorescence to 18 -30% above the levels for actin-tropomyosin alone in accord with our earlier measurements (5,9,10).
The direct evidence for caldesmon regulation by a troponin-like mechanism is supported by measuring how S1 and caldesmon binding to smooth muscle thin filaments depends on the ON-OFF equilibrium. The weak binding of S1⅐ADP⅐P i to actin-tropomyosin is not affected by the binding of inhibitory concentrations of caldesmon, indicating that the ON and OFF states have equal affinities for S1⅐ADP⅐P i (12,48) while the binding of strong binding complexes (S1⅐ADP, S1⅐AMP⅐PNP) is inhibited by caldesmon and becomes cooperative (13): this effect parallels troponin-inhibited actintropomyosin (49) and is due to strong binding complexes having a much higher affinity for the ON state than the OFF state leading to cooperative switching of the thin filament to the ON state (25,43).
Caldesmon binding to actin-tropomyosin depends critically on the activity state of actin-tropomyosin in the proposed regulatory mechanism. The basis of cooperative inhibition is that the inhibitory ligand has a higher affinity for the OFF state than the ON state. As Fig. 4 shows, initial caldesmon binding to actin-tropomyosin in the ON state is weakened at least 20-fold and becomes cooperative, while caldesmon binding to actintropomyosin in the OFF state is similar to binding to actintropomyosin alone, indicating that caldesmon preferentially binds to the OFF state as expected.
The activating effect of Ca 2ϩ -CaBP-caldesmon produces the opposite effect on affinity for actin-tropomyosin (Fig. 5): Ca 2ϩ -CaBP-caldesmon binding to actin-tropomyosin in the ON state is similar to binding to actin-tropomyosin alone, but Ca 2ϩ -CaBP-caldesmon binding to actin-tropomyosin in the OFF state is weakened about 20ϫ and is now cooperative, indicating that the activated complex binds preferentially to the ON state as predicted by the cooperative-allosteric model.
In addition, we have recently measured the effects of caldesmon on the elementary steps of the actin-tropomyosin-S1 ATPase pathway using transient kinetics. Using a fluorescent phosphate sensor, MDCC-labeled phosphate-binding protein, in double mixing stopped-flow experiments, we observed that caldesmon (1:7 actin) reduced the rate of phosphate release from the actin-S1⅐ADP⅐P i complex from 58 s Ϫ1 to 2.5 s Ϫ1 and introduced a lag phase of up to 200 ms, which could be eliminated by preincubation with S1 (1/5actin). These findings suggest that at physiological ratios of caldesmon to actin-tropomyosin; inhibition of the actomyosin ATPase is primarily due to an effect on the rate of phosphate release as was demonstrated for troponin-tropomyosin (28,50). The cooperative mechanism described here with caldesmon acting as an allosteric inhibitor is the only mechanism that can account for both the reduced P i release rate and the initial lag.
It is necessary to comment on the shape of the caldesmon binding curve, because different interpretations of this binding curve have been used as evidence for different caldesmon inhibitory mechanisms. We have consistently observed biphasic binding curves consisting of a high affinity (Ͼ10 7 M Ϫ1 affinity) low stoichiometry component and a low affinity component (3 ϫ 10 5 M Ϫ1 ), which saturates at a total amount bound of 1 caldesmon per actin (5,(51)(52). It is the tight binding component that is associated with the inhibition of actin-tropomyosin and is sensitive to the ON-OFF equilibrium (Figs. 4 and 5) (11). This component of binding is absent in pure actin filaments. This biphasic binding has been reported for caldesmon binding to actin-tropomyosin by others (53); it has also been found for troponin IC binding to actin-tropomyosin. Zhou et al. (54) reported 0.14 mol troponin IϩC per mol actin binding at 6 ϫ 10 6 M Ϫ1 and 0.86 troponin IϩC/actin binding at 3 ϫ 10 5 M Ϫ1 . It is therefore possible that this biphasic pattern of binding is characteristic of all cooperative regulatory actinbinding proteins. It should, however, be noted that at physiological ratios of actin to caldesmon or troponin IϩC only the tight binding sites can be occupied.
The Chalovich group have published binding curves in which the transition between the two phases is much less distinct than those shown here, and these curves were fitted to the McGee and Von Hippel algorithm and used as evidence for an inhibitory mechanism that primarily involved dissociation of S1⅐ADP⅐P i from actin by caldesmon (18,20,21,55,56). It should be noted that the McGee and Von Hippel algorithm cannot fit the curves shown in Figs. 4 and 5 with a sharp transition between strong and weak binding, 3 whereas the two-state model can account for both types of observed data. We have reported previously that the affinity of the tight binding component is weakened by increasing ionic strength while the weak binding component is not (57,58). Furthermore, the ionic strength effect is much more pronounced with small C-terminal fragments of caldesmon; and as a consequence the biphasic nature of the curve becomes hard to distinguish at moderate KCl concentrations (Ͼ20 mM) because the difference between the two dissociation constants is reduced. This property may account for reports of variable relationship between caldesmon bound and inhibition for different-sized fragments of caldesmon based on fitting to binding curves. In our experience, the bound/inhibited relationship is the same for full-length caldesmon, domain 3 ϩ 4, domain 4 and domain 4b alone (12, 58 -59). This is compatible with the allosteric inhibition of actin-tropomyosin, demonstrated in this article, in which the cooperative parameters are defined by actin and tropomyosin rather than by the nature of the inhibitor species (11,32).
Investigation of the structural changes associated with smooth muscle thin filament regulation supports the cooperative model. There is evidence for concerted conformational changes in tropomyosin (60), actin (61), and caldesmon (62) during the ON-OFF transition. Electron microscopy of actin-tropomyosin-caldesmon shows tropomyosin to be in the OFF (or closed) position relative to actin. However, from electron microscopy and transient kinetic measurements, there is no evidence that caldesmon induces a blocked state in smooth muscle thin filaments in contrast to troponin (28,63). The caldesmon-actin interaction is critical for determining the thin filament activity state. Our work using caldesmon truncation mutants and NMR spectroscopy has shown that caldesmon binds to actin through three short sequences, each of about 9 amino acids, separated by spacers 30 amino acids long (15,58). Cooperative switching of actin to the OFF state is obtained when at least 2 contact sequences are bound (64,65) and the main consequence of Ca 2ϩ -calmodulin binding to actin is to release all except one of the contact sites, thus permitting actin to change to the ON state (66,67).
The estimated parameters of thin filament switching by caldesmon and Ca 2ϩ -calmodulin are shown in Fig. 9. The values determined here are compatible with our previous measurements of cooperative S1⅐ADP binding and cooperative inhibition (13). Quantitatively they differ from troponin-tropomyosin and the differences may explain why caldesmon is not the primary regulator of smooth muscle thin filament activity. Compared with troponin, caldesmon has a lower affinity for actin-tropomyosin than troponin in all activity states, and the calculated cooperative unit size is reduced by the binding of caldesmon, whereas it is increased by binding troponin (68). As a consequence the physiological function of caldesmon seems to be modulation of Ca 2ϩ sensitivity and maintaining relaxation at low Ca 2ϩ (6 -8) while the main Ca 2ϩdependent switch is due to myosin light chain phosphorylation and dephosphorylation. . Cooperative allosteric mechanism of smooth muscle thin filament Ca 2؉ regulation. Calculated parameters for the cooperative ON-OFF transition are given for actin-tropomyosin (ATm), actin-tropomyosin-caldesmon (ATm.CaD) and actin-tropomyosin-caldesmon-Ca 2ϩ -calmodulin (ATm. CaD.Ca.CaM). Caldesmon decreases K T while caldesmon-Ca 2ϩ -calmodulin increases it, thus the system is switched OFF by caldesmon and switched ON by caldesmon-Ca 2ϩ -calmodulin. This is because caldesmon has a higher affinity for the OFF state than for the ON state while caldesmon-Ca 2ϩ -calmodulin has a higher affinity for the ON state than the OFF state.