Does Calponin Interact with Caldesmon?*

The roles of calponin and caldesmon and their interaction in regulation of smooth muscle contraction are controversial. Recently, strong binding between these two proteins has been reported (Graceffa, P., Adam, L. P., and Morgan, K. G. (1996) J. Biol. Chem. 271, 30336–30339). Results in this paper fail to confirm their data and are consistent with the concept of independent functions for calponin and caldesmon. To examine the ability of duck gizzard caldesmon to interact with calponin, three caldesmon derivatives, each containing a different sulfhydryl-specific reporter probe (6-acryloyl-2-dimethylaminonaphtalene,N-(1-pyrenyl)iodoacetamide, andN-iodoacetyl-N′-(5-sulfo-1-naphtylo)ethylenediamine) attached to a single cysteine located in the C-terminal domain, were synthesized. Addition of calponin to labeled caldesmon at both low and physiological salt concentrations did not induce any changes in fluorescence intensity or maximum shift. Under the same conditions, calmodulin and tropomyosin (known to bind to the C terminus of caldesmon) produced substantial changes in these spectral parameters. Gel filtration of an equimolar caldesmon-calponin mixture on a fast protein liquid chromatography Superose-12 column revealed two base-line-separated peaks, the first containing only caldesmon and the second only calponin, thus confirming the lack of any interaction between these two proteins. Also, the addition of calponin did not change the fluorescence parameters of labeled caldesmon in complexes with F-actin and F-actin-tropomyosin.

The primary mechanism for smooth muscle regulation, necessary to initiate contraction, involves phosphorylation of the 20-kDa myosin light chains by a specific Ca 2ϩ /calmodulin-dependent myosin light chain kinase (for reviews see Refs. [1][2][3]. It is postulated that the specific thin filament-associated proteins caldesmon and calponin take part in the secondary mechanism of regulation, probably during the relaxation phase. This view is supported by in vitro experiments showing that both of these proteins inhibit actin-activated ATPase activity of myosin, mobility of actin filaments over immobilized myosin in motility assay, and skinned muscle contraction. This inhibition can be reversed by Ca 2ϩ /calmodulin, or other Ca 2ϩ -binding proteins (like S100 or caltropin), and by phosphorylation with protein kinase C or casein kinase II (Refs. 3-6 and references therein).
Studies by Makuch et al. (7) revealed that calponin and caldesmon compete for the binding sites on actin filament, calponin being more effective at displacing caldesmon than vice versa. Similar results were obtained in the Chalovich laboratory (8). This suggested that calponin and caldesmon do not form a mutual complex on actin and reside on different populations of thin filaments in vivo. Immunofluorescence and immunogold electron microscopy confirmed this suggestion showing that, whereas caldesmon is present exclusively in the contractile domain, calponin is primarily, although not exclusively, located in the cytoskeletal domain of chicken gizzard muscle (9,10). This result implied that calponin and caldesmon are segregated in different thin filaments, as postulated earlier (11) on the basis of immunoprecipitation by anti-caldesmon and anti-filamin antibodies of the two different subsets of thin filaments.
In view of these works, the results showing a strong, specific interaction of calponin and caldesmon (12) were unexpected. Moreover, a very weak association of these two proteins was earlier reported and was considered to be too weak to affect the function of each of the two proteins (13). These controversial results (12) sparked our interest in resolving the problem.

EXPERIMENTAL PROCEDURES
Preparation of Proteins-Caldesmon was prepared from duck gizzards according to the method of Bretscher (14) with some modifications (15). Calponin was purified from chicken gizzard by the procedure of Takahashi (16). Chicken gizzard tropomyosin was obtained as described earlier by Dą browska et al. (17). Rabbit skeletal muscle actin was isolated from acetone-dried muscle powder and purified as described by Spudich and Watt (18). Calmodulin was isolated from bovine brain by the method of Gopalakrishna and Anderson (19).
Caldesmon Labeling-Labeling of caldesmon cysteine with acrylodan 1 (Molecular Probes) was carried out as described by Graceffa et al. (12) at a probe to caldesmon molar ratio of 5. The degree of labeling determined using the acrylodan molar extinction coefficient of 1.29 ϫ 10 4 M Ϫ1 cm Ϫ1 at 360 nm (20) was 0.70 -0.90 mol of acrylodan/mol of caldesmon.
Labeling of caldesmon with pyrenyl (Molecular Probes) was carried out according to Kouyama and Mihashi (21) at a probe to caldesmon molar ratio of 4.5. The degree of labeling determined using the pyrenyl molar extinction coefficient of 2.2 ϫ 10 4 M Ϫ1 cm Ϫ1 at 344 nm was 0.57-0.63 mol of pyrenyl/mol of caldesmon.
Labeling of caldesmon with AEDANS (Sigma) was performed as described in Tawada et al. (22) at a probe to caldesmon molar ratio of 7.5. The degree of labeling determined using the AEDANS molar extinction coefficient of 0.61 ϫ 10 4 M Ϫ1 cm Ϫ1 at 337 nm (23) was 0.50 -0.65 mol of AEDANS/mol of caldesmon.
Each labeling procedure was repeated twice for different caldesmon preparations. Labeling reactions were stopped by the addition of an excess of DTT and followed by an exhaustive prolonged dialysis with fluorescence control of the dialysate (80%) mixture with ethanol (20%) for the presence of the DTT derivative of each dye. The actomyosin ATPase activity assay (7) was used to control whether the described * This research was supported by a grant from the State Committee for Scientific Research to the Nencki Institute. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Fluorescence Measurements-Fluorescence spectra were recorded on a SPEX Fluorolog-2 photon-counting fluorimeter at a temperature of 20°C and caldesmon concentration of 1 mM in 20 mM Tris-HCl buffer, pH 7.5, containing 1 mM MgCl 2 , 1 mM CaCl 2 (buffer T) and either 50 or 150 mM NaCl. The spectra were corrected for solvent and, if needed, for dilution.
FPLC Gel Filtration-FPLC experiments were performed on a Superose-12 column (Pharmacia Biotech, Inc.). A caldesmon/calponin 1:1 (mol/mol) mixture was applied on the column in buffer T containing either 50 or 150 mM NaCl and eluted isocratically with the loading buffer. Peak fractions were precipitated with 20% trichloroacetic acid and centrifuged. The resulting pellets were dissolved in 50 ml of 6 M urea and analyzed by SDS-PAGE.
Electrophoresis-The quality of the protein samples at all steps of the various procedures was monitored electrophoretically. SDS-PAGE was carried out on polyacrylamide minislab gels using the discontinuous Tris-glycine buffer system of Laemmli (24

RESULTS AND DISCUSSION
Duck gizzard caldesmon is a single thiol protein; its Cys residue is located in the C-terminal part of the molecule at a position corresponding to 580 in the amino acid sequence of chicken gizzard caldesmon (30). Since the C-terminal sequences of avian isoforms of caldesmon are similar (31), and the site of the interaction between caldesmon and calponin was ascribed to just its C-terminal domain (12), duck gizzard caldesmon appeared to be convenient for fluorescent studies after modification with sulfhydyl-specific reporter probes.
As shown in Fig. 1, the fluorescence spectrum of acrylodanduck gizzard caldesmon adduct has a very broad peak with a maximum at 517 nm. This means that the acrylodan label is located in a polar environment (20). The fluorescence maximum does not alter upon a change of excitation wavelength from 360 to 390 nm. Our spectra were blue shifted by 5 nm as compared with those reported for acrylodan-labeled porcine stomach caldesmon (12). According to our observations, the DTT derivative of acrylodan could only be removed after an extensive and prolonged dialysis possibly because it fits into the molten globules located at both ends of the caldesmon molecule (32). An insufficient volume of dialysis buffer may be an explanation why Graceffa et al. (12) obtained a labeling ratio of up to 1.1 for porcine stomach caldesmon which is unexpected for a single thiol protein, as discussed in (20).
Addition of calponin to acrylodan-labeled caldesmon (at a 1:1 molar ratio) did not affect its fluorescence spectrum at 50 and 150 mM NaCl (Fig. 1). Under the same conditions, calmodulin, a protein with a similar hydrophobicity as calponin, induced marked changes in fluorescence spectra parameters (intensity and maximum blue shift) of acrylodan-labeled caldesmon. It is also noteworthy that the calponin to caldesmon binding constant of 9.5 ϫ 10 7 M Ϫ1 reported by Graceffa et al. (12) is of the same range as we observed earlier for the calmodulin-caldesmon complex (15). Significant changes in acrylodan-labeled caldesmon were also evoked by tropomyosin (data not shown) though its binding constant to caldesmon is 2 orders of magnitude lower (15) than that reported for the calponin-caldesmon complex (12). The shape of the fluorescence spectra of AE-DANS-labeled caldesmon is very similar to that presented in Fig. 1. In this case, we also did not observe any effect of calponin on the caldesmon spectra (data not shown).
The shape of the fluorescence spectrum of pyrenyl-labeled duck gizzard caldesmon (Fig. 2) differs from the fluorescence spectrum of pyrenyl-protein adducts often presented in the literature (e.g. chicken gizzard isoform of caldesmon labeled at both thiol groups (33) or pyrenyl-labeled F-actin (21)). The extremely polar (charged) surface of the C-terminal domain of the caldesmon molecule results in spectra looking like that for pyrenyl-G-actin (21). Pyrenyl-labeled caldesmon spectra also do not demonstrate relevant changes in the presence of calponin, while upon addition of calmodulin or tropomyosin (not shown), the fluorescence intensity in both low and physiological salt undergoes substantial changes. Partially dialysed sample of pyrenyl-labeled caldesmon resulted in fluorescence changes similar to those observed by Graceffa et al. (12) upon addition of calponin.
The fluorescence studies indicate that calponin does not induce any conformational changes in the spatial vicinity of the probe attachment to the caldesmon molecule and points to a lack of an interaction between calponin and this part of caldesmon. To prove our findings, we have performed additional experiments with molecular sieves. An equimolar mixture of duck gizzard caldesmon and calponin in various salt solutions was loaded onto an FPLC Superose-12 column and eluted with the loading buffer. The results are presented in Fig. 3. In both low and physiological salt concentrations, we have obtained two separated protein peaks, the first containing caldesmon and the second calponin. We did not observe the formation of a caldesmon complex with calponin. This confirms our fluorescence data and extends the conclusion about the lack of an interaction with the entire caldesmon molecule.
In earlier work on calponin, Vancompernolle et al. (13) reported that caldesmon bound to a calponin affinity column at 5 mM KCl was eluted at 70 mM KCl. However, calponin is known as a protein extremely sensitive to ionic strength and even able to aggregate at low ionic strength. In our FPLC experiments performed at 20 mM NaCl, calponin was retarded on the column and eluted only after an extensive wash with buffer containing 200 mM NaCl (data not shown). Therefore, one may infer that in the experiment described by Vancompernolle et al. (13), caldesmon was possibly bound to a partially unfolded immobilized calponin and was released when the ionic strength became high enough to restore the native conformation of the immobilized protein.
The fluorescence spectra of acrylodan-labeled caldesmon complexed with F-actin show a 27-nm blue shift and an almost 150% increase in the quantum yield in 50 mM NaCl as well as a 17-nm blue shift and almost double the quantum yield in 150 mM NaCl when compared with the spectra of acrylodan-labeled caldesmon alone under the same conditions (Table I). These dramatic changes did not depend practically on the molar ratio of the proteins in the range studied. They might be caused by the acrylodan label transfer from a strongly polar (the surface of caldesmon molten globules) to a strongly hydrophobic (Factin interior) environment. Binding of tropomyosin to this complex results in a further small (1-3 nm) blue shift and a very small increase in the quantum yield (see Table I). These small changes of the fluorescence parameters may be due to the structural changes in F-actin induced by the binding of tropomyosin and/or direct interaction of tropomyosin with caldesmon when bound to F-actin. Calponin binding to F-actincaldesmon or F-actin-caldesmon-tropomyosin complexes causes a small (1-2 nm) red shift and a negligible decrease in the quantum yield. The largest decrease (10%) in the quantum yield was observed in the case of calponin binding to the Factin-caldesmon complex (7:1 molar ratio) in 50 mM NaCl (see Table I). Conceivably this decrease might reflect the structural changes induced in F-actin by calponin binding; higher ionic strength or binding of tropomyosin reduces it by half. The structural changes in F-actin induced by calponin binding are rather local because they are ineffective when the quantity of F-actin in solution is doubled.
Taken together, the results presented above empower us to give a negative answer to the question "Does calponin interact with caldesmon?" Note Added in Proof-In agreement with our findings, Drs. Graceffa, Adam, and Morgan have communicated to us that they have further verified a lack of strong interaction between caldesmon and calponin by analytical ultracentrifugation in collaboration with Dr. Walter Stafford.

TABLE I
The fluorescence parameters of acrylodan-labeled duck gizzard caldesmon in complex with F-actin, tropomyosin, and calponin Fluorescence spectra were measured at 20°C in buffer T (see "Experimental Procedures") containing 50 or 150 mM NaCl, respectively. The molar ratios of proteins are shown in parentheses. Both slits were adjusted to 5 nm, and samples were excited at a wavelength of 380 nm. Evaluation of spectra maximum and integration were done using commercial software supplied by SPEX. The relative quantum yield was defined as Q rel ϭ (Q sample /Q CaD ϩAc ). CaD, calmodulin; Ac, actin; CaP, calponin; TM, tropomyosin.