Vol. 49 No. 2/2002 471–479 QUARTERLY

Kinetics of the smooth muscle calponin-F-actin interaction was studied by stopped-flow measurements of light scattering and fluorescence intensity of pyrene-labelled F-actin. The intensity and character of the changes in light scattering, and thus the mode of calponin binding to actin filaments leading to changes in their shape and bundling, depend on the molar ratio of the two proteins. Parallel measurements of pyrene-fluorescence quenching upon calponin binding revealed that intrinsic conformational changes in actin filaments are delayed relative to the binding process and are not markedly influenced by the mode of calponin binding. Bundling of actin filaments by calponin was not correlated with fluorescence changes and thus with alterations in the structure of actin filaments.

Calponin, originally discovered in chicken gizzard smooth muscle by Takahashi et al. (1986), is an actin-binging protein of molecular mass 32.3 kDa and dimensions of 16.2 2.6 nm, sufficient to span three monomers along the long pitch helix of actin filament (Stafford et al., 1995).Three dimensional electron microscopic reconstructions suggest that calponin is located peripherally along the long pitch actin helix forming a shield-like density over subdomain 2 of actin and connects the axially adjacent actin monomers by binding to the upper and lower edges of subdomain 1 of each actin monomer (Hodgkinson et al., 1997).It was initially suggested by El- Mezgueldi et al. (1992) that the region of calponin encompassing 145-182 amino-acid residues is implicated in actin-binding.Later on, it appeared that this region is actually composed of two distinct actin-binding sites on the calponin molecule: the first within amino acids 145-163, responsible for the inhibition of the actin-activated Mg-ATPase activity of myosin (El-Mezgueldi et al., 1995;El-Mezgueldi et al., 1996;Mino et al., 1998) and the second, within amino acids 172-187, essential and sufficient for suppression of actin binding (Mino et al., 1998) due to its phosphorylation by protein kinase C (Itoh et al., 1995).The mechanism of the inhibition of actin-activated ATPase of myosin by calponin is a matter of controversy: some authors claim that calponin affects strong binding of actin to myosin (El-Mezgueldi & Marston, 1996) while others do not confirm this idea (Ko³akowski et al., 1997;Szymanski et al., 1997).
It has been shown that calponin forms two types of complexes with actin: one at a molar ratio to actin up to 1:3-4, sufficient for almost full inhibition of the actomyosin ATPase activity (Ko³akowski et al., 1997;Szymanski et al., 1997) and the other at a higher ratio leading to the formation of actin filament bundles (Ko³akowski et al., 1995;Ko³akowski et al., 1997;Lu et al., 1995;Tang et al., 1997).The ability of calponin to crosslink actin filaments implicates its structural role, i.e. formation, together with other actin-binding proteins (a-actinin, filamin), of the actin cytoskeleton (Leinweber et al., 1999).Such dual role of calponin is supported by its distribution both in the contractile apparatus and cytoskeletal domain of smooth muscle cell (North et al., 1994;Mabuchi et al., 1996).It colocalizes with actin filaments (Walsh et al., 1993) composed of smooth muscle a and g isoforms as well as the cytoskeletal b isoform of actin (North et al., 1994).
In order to gain further insight into the calponin-actin interaction we have studied its kinetics by monitoring changes in the light scattering and fluorescence intensity of pyrene-labelled F-actin using a stopped flow apparatus.The results show that the binding rate of calponin to actin filaments depends on the ratio of both proteins and always precedes the conformational changes in actin monomers.
Stopped-flow fluorescence and light scattering measurements.The measurement of the time-course of calponin binding to actin was performed on a Spex Fluorolog spectrofluorometer equipped with Hi-Tech stoppedflow accessory (SFA-20) or on an SF-51 Hi-Tech stopped flow apparatus.The solutions containing various concentrations of calponin and 4 mM pyrene-labelled actin in 30 mM KCl, 2 mM MgCl 2 , 0.5 mM EGTA, 0.5 mM 2-mercaptoethanol, 0.02% NaN 3 and 10 mM imidazole-HCl, pH 7, were loaded into the drive syringes.The samples were then pushed into the sample loop where equilibration at the experimental temperature lasted for 5-10 s before the reaction was initiated.Light scattering was monitored at 90°to the incident light at a 365 nm.Pyrenyl fluorescence, excited at 365 nm, was monitored at 397 nm.All concentrations refer to the reactants after they have been mixed.
Curve fitting routines.The curves were fitted using the nonlinear Levenberg-Marquardt algorithm.Changes in the light scattering or fluorescence clearly deviated from a single exponential fit.Therefore, fitting to a double ex-ponential plot was performed according to equation: where k 1 and k 2 are the observed rate constants and A 1 and A 2 are the corresponding amplitudes of the changes.The observed rate constants (k 1obs and k 2obs ) were calculated from at least three traces.Typical traces are shown in the Figures.Simulation of the binding process was made using the KINSIM program (Barshop et al., 1983).

RESULTS
The rate of complex formation between calponin and filamentous actin at various molar ratio and low ionic strength was measured by light scattering.Since the interaction of the elongated protein molecules with protein polymers affects not only the mass but also the shape (diameter) of the polymers (Cantor & Schimmel, 1980), the intensity of light scattering which is sensitive to the changes in those parameters appeared to be convenient in studying the mode of calponin binding to actin filaments.As shown in Fig. 1, the character of changes in light scattering depended on the calponin to actin molar ratio.At the ratio of 0.4 : 1 light scattering was gradually increased (Fig. 1A).The obtained curve was best described by a double exponential fit (k 1obs = 1.1 s -1 and k 2obs = 0.13 s -1 ) since a single exponential plot did not fit the data adequately.In this case, calponin molecules were likely to be initially bound randomly through one or another actin-binding site along the actin filament and instantly were rearranged by utilising both actin-binding sites to the longitudinal position, which is thermodynamically favoured.A small increase in the actin filament diameter caused, in consequence, a small (about 7%) increase in the light scattering signal over that produced by actin filaments alone.Measurements of the rate of associa-tion of calponin with actin was complicated by bundling of actin filaments occurring at higher calponin concentrations (Fig. 1C) (Ko³akowski et al., 1995;1997;Lu et al., 1995;Tang et al., 1997).Therefore, we estimated the association rate constants using the KINSIM program, on the assumption that binding was irreversible.The estimated rate of binding describes the lower limit of the velocity for this reaction.The simulation was made according to a simple binding scheme (1) or according to a two step binding Scheme (2).
In Scheme 2, we assume that binding is followed by a very fast reorganization connected with additional binding of calponin to two actin molecules.This assumption leads to a better fit of the experimental data.The calculated association rate constant k 1* = 2 ´10 5 M -1 is in agreement with that estimated by Lu et al. (1995) by measurement of the rate of caldesmon dissociation from actin filaments after addition of calponin.
On increasing ratio of calponin to actin to 0.6:1 (still below the saturation level) the light scattering changes accompanying the binding of both proteins were much faster (< 0.5 s) and more pronounced (Fig. 1B).The character of the changes was typical for a multiphasic reaction (Cantor & Schimmel, 1980) and can be interpreted in terms of the two mode binding model proposed previously (Ko³akowski et al., 1997).The fast initial binding induced a marked (about 60%) enhancement of the light scattering, probably due to the mostly transverse orientation of calponin molecules with one site of attachment to actin filaments and a  1.The units on the ordinate axes are arbitrary, however, they are the same for all three panels.Dashed lines (A) represent a simulation of binding of calponin to actin according to scheme 1 with association rate constant k 1 = 2 ´10 5 M -1 s -1 and bold dashed lines according to scheme 2 with association rate constant k 1* = 2 10 5 M -1 s -1 .Measurements were performed at 20°C.
significant increase in their diameter.The calculation of the binding rate constant was, in this case, impossible due to the superimposition of the binding phase with a subsequent two-step decrease in the light scattering, which probably reflects the rearrangement of calponin molecules on the actin filament to the longitudinal position.First, an about 10% decrease in light scattering signal occured relatively quickly because there were still some unoccupied, free binding sites on the actin filaments.The following steps were very slow (k 1obs = 0.1 s -1 , k 2obs = 0.011 s -1 ) (Table 1) due to steric hindrance and a probable requirement for the dissociation of calponin molecules from actin filaments, which seems to be rate limiting since the observed rate constant of this step was in the range of the dissociation rate constant (Lu et al., 1995).Moreover, double exponential fitting suggests that calponin dissociated from the two different actin-binding sites.A small number of calponin molecules which preserved a transverse orientation caused, after a long lag phase (200 s), cooperative bundling of actin filaments manifested by a large increase in the light scattering (Fig. 1C, inset).At the calponin to actin ratio of 0.8 : 1, when almost all actin monomers in the filament were saturated by calponin and its longitudinal arrangement on the filament was impossible, the number of calponin molecules in the transverse, bundling orientation was much higher.Therefore, as demonstrated by an almost 10-fold increase in light scattering signal, bundling of actin filaments appeared immediately (k 1obs = 1.9 s -1 ) upon mixing of both proteins (Fig. 1C).This explains why, in this case, we could not distinguish between the binding and bundling phases.The amplitudes of changes in light scattering signal accompanying bundling were much higher than observed during binding of calponin to actin.Thus, the complicated double exponential character of the registered curve was probably caused by crosslinking and subsequent reorganisation of the actin filament network.The bundling process itself is likely to show complicated kinetics as the observed signal is a weighted sum from a heterogeneous popula-Table 1. Kinetic parameters of calponin binding to F-actin.
The relative changes, amplitudes and observed rate constants were calculated from the light scattering and the fluorescence curves using the procedure described in the Material and Methods section.tion, with larger bundles having a stronger scattering characteristics.Thus, the higher rate constants correspond to the rate of bundling.

Calponin to actin ratio
To obtain information on the effect of calponin binding on the actin filament structure, in parallel with the rates of light scattering, the rates of fluorescence intensity changes of pyrene-labelled F-actin upon its interaction with calponin were measured.At all ratios of calponin to actin, the fluorescence intensity of pyrene-labelled actin dropped gradually, at a rate proportional to the ratio of both proteins (Fig. 2, Table 1).Fluorescence quenching for the ratio of calponin 0.4, 0.6 and 0.8 moles per actin monomer reached 32, 37 and 43%, respectively.The fluorescence quenching curves were best fitted by a double exponential equation.As shown by the fluorescence parameters (Table 1) the quenching is induced by two processes of comparable amplitude but with rates differing by a factor of 10.This probably reflects the binding of calponin molecules to actin filaments through distinct actin-binding sites.
Comparison of the time-course of the light scattering and fluorescence intensity changes showed that, at the lowest calponin to actin ratio, the fluorescence quenching rate was lower (k 1obs = 0.19 s -1 ) than the rate of binding (k 1obs = 1.1 s -1 ).Acceleration of the binding processes at a higher calponin ratio (0.6 : 1) and immediate rearrangement of calponin molecules on the actin filament prevented calculation of the binding rate constant.However, comparison of the curves presented in Fig. 1B and Fig. 2 (curve b) showed that, during the initial binding phase, the fluorescence intensity decreased by only 5%.Thus, the delay in the decrease of fluorescence intensity indicates that it has been caused not by calponin binding itself but by slow changes in the intrinsic structure of actin filaments induced by this binding.The bundling of actin filaments by calponin was not correlated with fluorescence quenching: at an intermediate ratios of calponin to actin monomer (0.6 : 1), the bundling followed, but at a higher ratio (0.8 : 1) it preceded the fluorescence quenching (Table 1).Nevertheless, the observation of  Table 1.Measurements were performed at 20°C a significant change in light scattering signal (Fig. 1C) without an appreciable change in fluorescence argues that the quenching in fluorescence observed in Fig. 2 (curve c) is not an artefact arising from the increasing turbidity producing an inner filter effect but reflects an environmental change in the pyrene labelled actin residue.

DISCUSSION
Combination of the light scattering and fluorescence methods to analyse the kinetics of the calponin-actin interaction allowed us to investigate both the binding of calponin to actin filaments and structural changes in the actin filaments accompanying this process.The results revealed that the mode of calponin binding to actin filaments depends on the ratio of both proteins: at a low ratio, elongated calponin molecules bind longitudinally along the filaments, at high a ratio transverse binding predominates facilitating bundling of actin filaments.At intermediate ratios, calponin molecules initially bind in a random fashion and subsequently rearrange into the longitudinal position.The molecules, which preserve the transverse orientation, form bundles of actin filaments.These results confirm our earlier proposition (Ko³akowski et al., 1997) of a two-site model of calponin-Factin interaction.
The profiles of the fluorescence intensity changes were similar and independent of the calponin to actin ratio, thus, the mode of calponin binding to actin filaments.The detailed kinetic analysis showed that the fluorescence quenching process consists of two components similar in amplitude and different in rate.This is in agreement with the observation of Mino et al. (1998) that the peptides covering the two distinct actin-binding sites of calponin reduce the fluorescence intensity of pyrene-labelled actin to the same extent.The rate of fluorescence decrease was lower than the rate of binding of calponin to F-actin, which indicates that fluorescence quenching of pyrene-conjugated to Cys-374 of actin is produced by an intrinsic conformational change in the actin filaments induced by calponin binding and not by calponin binding per se.This is consistent with earlier observations of Noda et al. (1992) that, upon treatment with calponin, the spectra of filamentous actin became similar to those of monomeric actin pointing to a reversal of actin monomer conformation in filaments towards that of a free state.Lack of correlation between the rate of fluorescence changes and the rate of the light scattering increase resulting from bundling of actin filaments by calponin excluded the participation of actin filament bundling in inducing those structural alterations.A similar result was previously reported by Tang et al. (1997).
The different mode of calponin binding to actin, proposed above, seems to be in agreement with the data of Bartegi et al. (1999) on the interaction of the C-terminal domain of calponin with F-actin.The authors observed, in the presence of an excess of labelled F-actin (at Cys-374) over acrylodan-labelled calponin (at Cys-273), significant changes in the intensity and maximum emission wavelength of fluorescence, which indicates a close proximity between the labelled amino-acid residues of both proteins.At the molar ratio of calponin to actin of 1 : 1.2, the distance measured between the labels increased to 2.9 nm.
The physicochemical methods applied so far in studies of the calponin-F-actin interaction, i.e., fluorescence intensity of pyrene-labelled actin (Mino et al., 1998;Noda et al., 1992;and this work), polarized fluorescence (Borovikov et al., 1996) and light scattering (this work) indicate that calponin induces changes in the shape and flexibility of actin filaments due to conformational alterations in actin monomers.Direct evidence for such alterations was obtained by subtracting the densities associated with pure F-actin from those for the F-actin-calponin complexes obtained from the electron micrographs of neg-atively stained filaments (Hodgkinson et al., 1997).
In conclusion, we suggest that the physiological significance of calponin to actin binding consist in that: u the changes induced by calponin in actin filament structure, slows down the crossbridge cycle and thus inhibit actomyosin ATPase activity, u independent of the above changes, fast bundling of actin filaments by a larger amount of calponin takes part in the formation of the actin cytoskeleton and its remodelling during the contraction-relaxation cycle of smooth muscle.
We are grateful to Dr. A.A. Kasprzak and Professor M. Kochman for critical reading of the manuscript and valuable suggestions.

Figure 1 .
Figure 1.Changes in light scattering during the interaction of calponin with F-actin at various ratios of both proteins.Samples contained various amounts of calponin and 2 mM F-actin in 30 mM KCl, 2 mM MgCl 2 , 0.5 mM EGTA, 0.5 mM 2-mercaptoethanol, 0.02% NaN 3 and 10 mM imidazole-HCl, pH, 7.0.A. Binding of 0.8 mM calponin to actin.Inset shows the whole binding process.B. Binding of 1.2 mM calponin to actin and sequential rearrangement of calponin on actin filaments.Inset shows the slow rearrangement phase.C. Bundling of actin filaments by 1.6 mM calponin.Inset shows sequential long time interval view of the light scattering curve presented in B, corresponding to the bundling process.Bold lines represent double exponential fits.Corresponding rate constants and amplitudes are given in Table1.The units on the ordinate axes are arbitrary, however, they are the same for all three panels.Dashed lines (A) represent a simulation of binding of calponin to actin according to scheme 1 with association rate constant k 1 = 2 ´10 5 M -1 s -1 and bold dashed lines according to scheme 2 with association rate constant k 1* = 2 10 5 M -1 s -1 .Measurements were performed at 20°C.

Figure 2 .
Figure 2. Changes in fluorescence intensity of pyrene-labelled F-actin during its interaction with calponin at various ratios of both proteinsCurves a, b and c correspond to 0.8, 1.2 and 1.6 mM calponin, respectively, mixed with 2 mM F-actin in solutions described in Fig.1.Bold lines are double exponential fits.Corresponding rate constants and amplitudes are given in Table1.Measurements were performed at 20°C