HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Dong, W.-J. Articles by Cheung, H. C. Search for Related Content PubMed PubMed Citation Articles by Dong, W.-J. Articles by Cheung, H. C. Social Bookmarking                      What's this?

J Biol Chem, Vol. 274, Issue 44, 31382-31390, October 29, 1999

Conformation of the Regulatory Domain of Cardiac Muscle Troponin C in Its Complex with Cardiac Troponin I^*

Wen-Ji Dong, Jun Xing, Matteo Villain^§, Matthew Hellinger, John M. Robinson, Murali Chandra^¶, R. John Solaro^¶, Patrick K. Umeda, and Herbert C. Cheung^**

From the  Department of Biochemistry and Molecular Genetics, ^§ Department of Physiology and and Biophysics, and  Department of Medicine, University of Alabama at Birmingham, Birmingham, Alabama 35294-2041 and the ^¶ Department of Physiology and Biophysics, College of Medicine, University of Illinois, Chicago, Illinois 60612-7342

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Calcium activation of fast striated muscle results from an opening of the regulatory N-terminal domain of fast skeletal troponin C (fsTnC), and a substantial exposure of a hydrophobic patch, essential for Ca²⁺-dependent interaction with fast skeletal troponin I (fsTnI). This interaction is obligatory to relieve the inhibition of strong, force-generating actin-myosin interactions. We have determined intersite distances in the N-terminal domain of cardiac TnC (cTnC) by fluorescence resonance energy transfer measurements and found negligible increases in these distances when the single regulatory site is saturated with Ca²⁺. However, in the presence of bound cardiac TnI (cTnI), activator Ca²⁺ induces significant increases in the distances and a substantial opening of the N-domain. This open conformation within the cTnC·cTnI complex has properties favorable for the Ca²⁺-induced interaction with an additional segment of cTnI. Thus, the binding of cTnI to cTnC is a prerequisite to achieve a Ca²⁺-induced open N-domain similar to that previously observed in fsTnC with no bound fsTnI. This role of cardiac TnI has not been previously recognized. Our results also indicate that structural information derived from a single protein may not be sufficient for inference of a structure/function relationship.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Contraction of striated muscle is regulated by a group of actin-binding proteins, the troponin-tropomyosin complex located on the actin filament (1). Troponin is a heterotrimer. The subunit TnC¹ binds Ca²⁺, TnI binds actin and inhibits actomyosin ATPase in relaxed muscle, and troponin T anchors the three-subunit complex to tropomyosin on the actin filament. These proteins form the thin filament. Strong force-generating interactions between myosin and actin are initiated by the binding of Ca²⁺ to the regulatory sites located in the N-terminal regulatory domain of TnC. The binding of activator Ca²⁺ to TnC weakens or breaks the interaction between TnI and actin, thus releasing the inhibition of actomyosin ATPase and initiating force generation.

The crystal structure of TnC from avian fast skeletal muscle TnC shows a dumbbell-shaped molecule with the N- and C-terminal segments folded into two globular domains connected by a long -helical central helix (2, 3). Each domain has two Ca²⁺-binding EF-hand (helix-loop-helix) motifs. The five helices in the N-domain are labeled the N-helix and helices A-D, starting from the N terminus (Fig. 1). The four helices in the C-domain are labeled helices E-H, starting from the C-terminal end of the central helix. The N-domain of fast skeletal TnC has two Ca²⁺-specific sites (sites I and II), which bind Ca²⁺ with a low affinity (K_a ~ 10⁵ M¹) and the C-domain has two high affinity Ca²⁺ sites (K_a ~ 10⁷ M¹), which also bind Mg²⁺ competitively (K_a ~ 10³ M¹). The two sites in the N-domain are the regulatory sites. Site I consists of the helix A-loop-helix B and site II the helix C-loop-helix D motif. Sites III and IV in the C-domain are believed to play a structural role and are occupied by Mg²⁺ under physiological conditions in relaxed muscle. Site I in cardiac TnC is inactive in chelating Ca²⁺ due to substitutions of two critical amino acids and an insertion in the binding loop; saturation of site II by Ca²⁺ is sufficient to trigger contraction in cardiac muscle (4). The crystal structure of fsTnC contains two bound Ca²⁺ ions in the C-domain and no bound cation in the N-domain. Based on the structure of the C-domain, an early computer model (5) (the HMJ model) suggests that Ca²⁺ binding to the regulatory sites induces reorientations of the B and C helices relative to the A and D helices, thus exposing a hydrophobic patch located in the B helix (see Fig. 1). The exposed hydrophobic patch in this open conformation becomes available for the Ca²⁺-dependent interaction with TnI. This model also has been used to interpret functional and drug binding properties of cardiac TnC.

Two types of spectroscopic studies have been reported to verify the HMJ model for fsTnC. A recent solution NMR study of the recombinant N-terminal fragment of fsTnC provides evidence for a Ca²⁺-induced open conformation in which the B and C helices move away from the A and D helices (6). We recently reported a fluorescence resonance energy transfer study of recombinant full-length fsTnC mutants in which the distributions of the distances between specific sites were determined (7). This study demonstrated large Ca²⁺-induced increases in the mean distances between residues 22 and 52 and between residues 90 and 52, indicating an open N-domain conformation in the holo-fsTnC state. The FRET data also showed a decrease in the half-width of the distributions of the distances in the presence of activator Ca²⁺, suggesting that the calcium-loaded open domain is highly constrained. The constrained conformation may be an important structural feature that allows a segment of fsTnI to interact with the exposed hydrophobic pocket in the B helix (7).

The sequences of cardiac TnC and fsTnC are about 70% identical. Most of the amino acid substitutions are in the N-domain including site I. Do these differences require a different mechanism for Ca²⁺ activation of cardiac muscle? In a preliminary FRET study (8), we showed that the N-domain of cTnC experienced a significant Ca²⁺-induced closed  open transition only in the presence of bound cTnI. This was the first demonstration that bound cTnI is required for Ca²⁺-induced opening of the N-domain conformation of cTnC. We report here details of this transition in cTnC observed with the complex cTnC·cTnI and with cTnC in the presence of fragments of cTnI. The results provide insights into the structural consequence of the binding of activator Ca²⁺ to cTnC and to cTnC in its complex with cTnI.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Protein and Peptide Preparations-- A cDNA encoding full-length TnC of chicken slow muscle was isolated from a gt10 cDNA library derived from embryonic chicken breast muscle cells (9). The sequence of this clone is identical to the sequence of the cDNA for cardiac TnC, and this clone was used for construction and expression of recombinant cTnC and cTnC mutants. The clone was first subcloned into a pAlter vector (Promega), and the pAlter mutagenesis kit was used to generate a cysteineless double mutant C35S/C84S. It was necessary to remove the two endogenous cysteines in order to obtain mutants with single cysteine in desired locations. This full-length double mutant was cleaved with NcoI and BamHI and inserted into the corresponding restriction sites of pET-3d vector (Novagen). This pET-3d plasmid containing the complete coding sequence of the TnC double mutant was used as a template to generate three single-tryptophan mutants containing the initial double mutations. These mutants were used to generate three final mutants each containing a single tryptophan and a single cysteine at desired locations: 1) F12W/N51C/C35S/C84S (12W/51C), 2) F20W/N51C/C35S/C84S (20W/51C), and 3) F20W/S89C/C35S/C84S (20W/89C). The locations of the two tryptophans and the two cysteines are indicated in the NMR structure shown in Fig. 1A. The sequences of the cDNA clones were verified by sequence analysis. To express cTnC and cTnC mutants, the appropriate pET-3d plasmid was transformed into Escherichia coli strain BL21(DE3) (Novagen) and expressed under isopropyl-1-thio--D-galactopyranoside induction. Subsequent steps and purification of expressed proteins were as described before for skeletal TnC (10). A tryptophanless cTnI mutant (W192C) and the N- and C-terminal fragments of cTnI were generated and characterized using the same method described previously (11). The identity of the expressed proteins was checked by matrix-assisted laser desorption ionization/time of flight mass spectrometry. The single cysteine in cTnC mutants was labeled with the fluorescence probe 1,5-IAEDANS under denatured conditions (11), and the labeled protein was subsequently renatured. The degree of labeling was found to be >95%.

View larger version (48K): [in this window] [in a new window]   Fig. 1.   A, a ribbon representation of the NMR structure of chicken cardiac holo-TnC in which all three sites are saturated with Ca2+ (Ref. 18 and Protein Data Bank code IAJ4). The structure is given here to show where the intersite distances reported in this paper are located. The five helices in the N-terminal regulatory domain are labeled N-helix and helices A-D, starting from the N terminus, and the four helices in the C-terminal domain are labeled as helices E-H. The helix A-loop-helix B motif is the first EF-hand and is the inactive Ca2+-binding site I. The helix C-loop-helix D is the second EF-hand and the Ca2+-binding site II. The helices B and C are linked by a short flexible linker (B-C linker). The N- and C-domain are linked by a 22-residue helix (central helix). The structure of the central helix is undefined in the NMR structure because it is highly flexible in solution. Shown in the figure are the four mutated residues (F12W, F20W, N51C, and S89C) and the three intersite distances determined in this work: 12W-51C, 20W-51C, and 20W-89C. Residue 12 is in the N-helix, residue 20 in the A helix, residue 51 in the B-C linker, and residue 89 in the N-terminal end of the central helix. Panel B is a representation indicating the three FRET distances determined in the holo-cTnC·cTnI complex, and showing an opening of the N-domain of cTnC in the Ca2+-loaded complex.

Peptides with sequences corresponding to segments of the C-terminal half of cTnI were synthesized as C-terminal amides using solid-state synthesis, by Fmoc (N-(9-fluorenyl)methoxycarbonyl) chemistry on a PerSeptive 9050 peptide synthesizer, with O-pentafluorophenyl ester amino acids activated with 1-hydroxyl-7-azabenzotriazole. Synthesized peptides were purified by reverse phase-HPLC, the purity of the products was checked by analytical reverse phase-HPLC, and the identity of the peptides was confirmed by matrix-assisted laser desorption ionization/time of flight mass spectrometry.

Myofibrillar ATPase assay was carried out with freshly prepared myofibrils from a rat heart. Endogenous cTnC was extracted at 4 °C from the myofibrils by a method slightly modified from that of Morimoto and Ohtsuki (12). The extraction solution containing CDTA included both pepstatin (1 µg/ml) and leupeptin (5 µg/ml), and the washing solution also included both pepstatin and leupeptin at 0.5 mg/ml. Reconstitution of cTnC-depleted myofibrils was done by incubation of the myofibrils (~1 mg/ml) with exogenous cTnC (~2-3 mg/ml) for 1 h at 25 °C, followed by at least three washings. The ATPase activity was determined from measurement of the release of inorganic phosphate (4) in a total volume of 1 ml. The assay mixture also included 0.2 mM phenylmethylsulfonyl fluoride, 0.5 mg/ml pepstatin, and 0.5 mg/ml leupeptin. The Ca²⁺-activated activity was determined at pCa 4.

Preparation of cTnC·cTnI Complex-- The binary complex cTnC·cTnI was prepared from cTnC mutants and the tryptophanless cTnI mutant, following our previous procedure (13). Briefly, the AEDANS-labeled cTnC was incubated with a large excess of cTnI in a buffer containing 50 mM MOPS at pH 7.2, 1 mM dithiothreitol, 5 mM Ca²⁺, and 6 M urea. The solution was then dialyzed against the same buffer containing 3 M urea and 1 M KCl. The urea and KCl concentrations were subsequently reduced stepwise by changing the dialysate to the final solution containing 50 mM MOPS at pH 7.2, 1 mM dithiothreitol, 1 mM EGTA, and 0.1 M KCl.

Fluorescence Measurements-- Steady-state fluorescence measurements were carried out at 20 ± 0.1 °C on an ISS PC1 photon-counting spectrofluorometer, using a band pass of 3 nm on both the excitation and emission sides. Emission spectra were corrected for variation of the detector system with wavelength. The quantum yield of tryptophan in proteins was determined by the comparative method as in previous work (14). For Ca²⁺ titration, the tryptophan was excited at 295 nm and the emission was detected at 331 nm. A standard Ca²⁺ solution (Orion) was used in Ca²⁺ titration experiments (13). EGTA was used to control the level of free Ca²⁺, which was calculated using known stability constants of the chelator for cations and proton (15).

Fluorescence intensity decay and anisotropy decay were measured at 20 ± 0.1 °C in the time domain with either a PRA 3000 single-photon lifetime spectrometer equipped with a rhodamine 6G dye laser cavity dumped and synchronously pumped by a mode-locked argon ion laser (14), or an IBH 5000 photon-counting lifetime system equipped with a very stable flash lamp operated at 40 kHz in 0.5 atm of hydrogen. The excitation wavelength was 295 nm and the emission wavelength was 333 nm for tryptophan and the corresponding wavelengths for the acceptor were 330 and 480 nm, respectively. For FRET studies, the donor was the single tryptophan and the acceptor was the extrinsic fluorophore AEDANS covalently linked to the cysteine. The intensity decay data of the donor collected from the donor-alone and donor-acceptor samples were used to calculate the distribution of the intersite distances as in our previous work (7, 16). All measurements, both steady-state and time-resolved, were made in 50 mM MOPS at pH 7.2, 100 mM KCl, 1 mM EGTA, and 5 mM Mg²⁺, unless stated otherwise. Since the main concern of this work is on the biochemical states of the regulatory N-domain of cTnC, cTnC, and cTnC·cTnI samples studied in these solution conditions are referred to as in the apo state. When Ca²⁺ was present, it was at pCa 4; cTnC and cTnC·cTnI at this pCa are referred to as in the holo state.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Characterization of cTnC Mutants-- The three single-tryptophan cTnC mutants were tested for their ability to confer Ca²⁺ activation in myofibrillar ATPase and to bind Ca²⁺ in the N-domain. These results are shown in Table I. All three mutants showed Ca²⁺-activated ATPase activities comparable to control. Ca²⁺ titration was carried out on two mutants (20W/51C and 20W/89C) that showed adequate changes in tryptophan fluorescence during the titration (titration curves not shown). Also carried out were the Ca²⁺ titration of the two cTnC mutants complexed with cTnI. A single binding constant was recovered in all four sets of titration data. These binding constants are very close to those which we previously reported (13) for isolated cTnC and the complex cTnC·cTnI and indicate that the mutations had little or no adverse effects on the functions of the cTnC mutants.

View this table: [in this window] [in a new window]   Table I Characterization of cTnC mutants The specific ATPase activity of control myofibrils was 58 and 161 nmol of Pi/min/mg of fibril protein in the presence of Mg2+ and Ca2+ (pCa 4), respectively. These values were obtained with cTnC-depleted myofibrils reconstituted with wild-type cTnC. The cTnC-extracted myofibrils showed essentially no Ca2+-activated activity. Ka is the binding constant of the cTnC mutants for Ca2+ determined by monitoring the change in the tryptophan fluorescence in the presence of Mg2+. The change in the fluorescence for cTnC(12W/51C) was not sufficiently large to allow a reliable estimate of Ka. A single binding constant was recovered from all four sets of data, using an equation as in previous work (reference 13). pCa0.5 = logKa and is given here for easy comparison with literature data.

Fluorescence Properties of cTnC Mutants-- The single-tryptophan cTnC mutants were studied by the steady-state and time-resolved methods. These results are summarized in Table II. The single tryptophan in one mutant was Trp¹² and the tryptophan in the other two mutants was Trp²⁰. Trp¹² showed a moderately high quantum, which was insensitive to the binding of Ca²⁺ to the regulatory site. Its intensity decay was biexponential, and the weighted mean lifetime increased slightly in the holo state. In the apo-cTnC·cTnI complex, its quantum yield increased 30% with a negligible effect on the mean lifetime. Ca²⁺ binding did not affect the quantum yield much, but increased the mean lifetime by 0.5 ns.

View this table: [in this window] [in a new window]   Table II Tryptophan fluorescence properties of cTnC mutants Protein concentration was 1-2 µM. The apo state refers to a solution containing Mg2+ such that the single regulatory site in the N-domain was unoccupied, and the holo state refers to a solution containing both Mg2+ and Ca2+ such that all three sites were saturated with Ca2+ (pCa 4). The numbers in parentheses are the fractional amplitudes (i) of the lifetime components (i). is the intensity-weighted mean fluorescence lifetime:  =  i i2/ ii.

As expected, the properties of Trp²⁰ were very similar in both mutants, indicating that mutations N51C and C89S had negligible effects on the tertiary structural environment of Trp²⁰. In the apo state of both isolated mutants and the cTnC·cTnI complexes, Trp²⁰ had a very similar and high quantum yield (0.34-0.35) and its intensity decay was single-exponential. The binding of activator Ca²⁺ to the regulatory site reduced the quantum yield 15-20% and resulted in a biexponential intensity decay with a mean lifetime slightly shorter than the single lifetime in the apo state. Formation of an apo complex with cTnI did not change the quantum yield or the intensity decay pattern, but in the holo complex the quantum yield increased about 15% with a mean lifetime of the two decay components increasing slightly. These results suggested that cTnI did not perturb the local environment of Trp²⁰ in the apo complex, but Ca²⁺ binding to the complex shifted the Trp²⁰ environment to a slightly less accessible environment. The high quantum yield, the single exponential intensity decay in the apo state, and the effect of Ca²⁺ on the spectral characteristics parallel those previously reported for the equivalent Trp²² in fast skeletal TnC (14). Molecular graphics and other results suggested that Trp²² is highly inaccessible to solvent in the apo state and becomes partially exposed in the holo state (10, 14). Trp²⁰ in cTnC appears to have structural environments similar to those of Trp²² in fsTnC in the two biochemical states.

FRET Distance Determination-- Fig. 2 shows the steady-state emission spectra of the cTnC mutants (panels A-C) and these mutants in the presence of cTnI (panels A'-C'). The major band in the 330-nm region is the donor emission, and the other band in the 480-nm region is the sensitized acceptor emission. In the absence of bound cTnI, Ca²⁺ induced small to negligible increases in the donor intensity and small reciprocal decreases in the acceptor intensity (changes from curve 1 to curve 2) for all three samples. These changes suggested small to negligible decreases in energy transfer for the three donor-acceptor pairs. By contrast, Ca²⁺ induced a considerably larger enhancement of the donor emission and decrease of the sensitized acceptor fluorescence for the two mutants 12W/51C and 20W/51C bound to cTnI (panels A' and B'). These changes suggested significant decreases in energy transfer and, therefore, increases in the two distances induced by Ca²⁺ within the cTnC·cTnI complex. The Ca²⁺ effect on the third mutant 20W/89C complexed with cTnI (panel C') was less pronounced than on the other samples. These steady-state spectra were used to calculate the intersite separations (r) for the three distances (Table III).

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Steady-state fluorescence emission spectra of single-tryptophan cTnC mutants labeled with IAEDANS, and the spectra of the mutants complexed with cTnI (cTnC·cTnI). The mutants are 12W/51C, 20W/51C, and 20W/89C. The spectra were determined with 1 µM protein in either 5 mM Mg2+ (curve 1) or 5 mM Mg2+ and pCa 4 (curve 2). Panels A-C, the spectra of unbound cTnC mutants; panels A'-C', the spectra of the corresponding cTnC·cTnI complexes.

View this table: [in this window] [in a new window]   Table III Intersite distances in cTnC mutants The three indicated distances were determined with both isolated cTnC mutants and the complex (cTnC mutant)-cTnI in both the apo and holo states. r is the inter-site distance determined by the steady-state intensity method, is the mean inter-site distance recovered from the distribution of the distances, and hw is the half-width of the distribution. The value of given in parentheses indicates that the mean distance was held fixed at the indicated value during the least squares analysis, and the corresponding R2 value is also given in parentheses. The value of distance change refers to the observed change in the distance between the holo and apo states. The NMR distance of holo cTnC is the separation between the C carbon atoms of the two residues determined with cTnC (residues 1-89) (Ref. 18 and PDB IAJ4). The NMR distance of holo complex is the distance between the two C atoms of cTnC (residues 1-89) in its complex with cTnI peptide (residues 147-163), taken from the best representative conformer in an ensemble of 40 conformers (Ref. 21 and RCSB protein data bank, accession code 1MXL).

Time-resolved data are needed to recover the distribution of the distances for each donor-acceptor pair. The intensity decay curves of the tryptophan in the isolated mutants are shown in Fig. 3 (panels A-C); the corresponding decay curves of the mutants complexed with cTnI are shown in the adjacent panels (A'-C'). These data were quantified and displayed in Fig. 4 as a distribution of the distances for each donor-acceptor pair. The transition of cTnC from the apo state to the holo state (Fig. 4, curves 1 to 2) was accompanied by a very small shift of the distribution toward longer distances for the distance Trp²⁰-Cys⁵¹. The distributions for apo-cTnC (curve 1) and the apo-cTnC·cTnI complex (curve 3) were very similar. However, the transition of the complex from the apo state (curve 3) to the holo state (curve 4) was accompanied by a large shift of the distribution with the mean distance (^-r) increasing by 6.5 Å. For the distance Trp¹²-Cys⁵¹, Ca²⁺ induced a small increase in ^-r (2.1 Å) in unbound cTnC and a larger increase (6.5 Å) in the complex. These distributions were relatively narrow with values of the half-width in the range of 2.1-3.5 Å for the distance Trp²⁰-Cys⁵¹ and 3-5 Å for the distance Trp¹²-Cys⁵¹. The observed increases in the mean values of the two distances are comparable to those detected with isolated apo skeletal TnC. These results provide direct evidence for a Ca²⁺-induced opening of the regulatory N-domain of cTnC only when it is bound to cTnI. The distance parameters and other relevant information are listed in Table III. The distance derived from the steady state measurements (r) are in excellent agreement with the mean distance (^-r) derived from the distributions. Also listed in the table are the separations of the C carbon atoms between the donor and acceptor residues from the NMR structure.

View larger version (36K): [in this window] [in a new window]   Fig. 3.   Representative fluorescence intensity decay of donor tryptophan in three single-tryptophan cTnC mutants, 12W/51C, 20W/51C, and 20W/89C, and their complexes with cTnI (cTnC·cTnI). Panels A-C are for unbound cTnC mutants, and panels A'-C' are for the corresponding cTnC·cTnI complexes. Curves 1 (in Mg 2+) and 2 (Mg2+ plus pCa 4) in each panel are from the donor-only samples. Curves 3 (in Mg2+) and 4 (in Mg2+ plus pCa 4) are from the donor-acceptor samples. Experimental conditions were the same as given for Fig. 2.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Area-normalized distribution of intersite distances determined from cTnC mutants and their complexes with cTnI. The energy donor was the tryptophan residue, and the acceptor was AEDANS attached to the cysteine residue. A, distribution for distances Trp20-Cys51 (20W-51C); B, distribution for the distances Trp12 -Cys51 (12W-51C); C, distribution for the distances Trp20-Cys89 (20W-89C). Curves 1 (apo) and 2 (holo) are for unbound cTnC; curves 3 (apo) and 4 (holo) are for cTnC bound to cTnI. The distance parameters recovered from the distributions are summarized in Table III, along with other relevant structural information. Experimental conditions were the same as given in Fig. 3.

The mean distances for Trp²⁰-Cys⁸⁹ in the apo and holo states of cTnC were within 1 Å of each other, and the corresponding values in the cTnC·cTnI complex were only 2.4 Å apart. The latter increase was much smaller than the changes observed for the other two distances in the complex. The distance parameters for this distance are also listed in Table III. The three sets of distances obtained in the holo-cTnC·cTnI complex are indicated in Fig. 1B for visual comparison with the distances obtained in the absence of bound cTnI (Fig. 1A).

To evaluate the extent to which the difference between the apo and holo distributions of the distances can be considered distinct, we examined the uncertainties in the mean distance and the half-width of each distribution by calculating the variance ratio surface (results not shown) as in our previous work (7, 16). The calculated surfaces of both the half-width and for the apo and holo states of the distances Trp²⁰-Cys⁵¹ and Trp¹²-Cys⁵¹ in isolated cTnC are not steep and have significant overlaps, making it difficult to distinguish between the distributions of the two biochemical states determined with cTnC. In contrast, the variance ratio surfaces of the two parameters for the two distances in cTnC·cTnI are quite steep and well separated between the apo and holo states. These surfaces do not intersect at any reasonable values for _R², which indicates that the distributions of these two distances in the binary complex under the apo and holo conditions are different from each other. The surfaces of the half-width for the two biochemical states of the third distance Trp²⁰-Cys⁸⁹ in the complex are also steep, but the surfaces for the mean distance are shallow and significantly overlap. This analysis indicates that the apo and holo distributions for this third distance in the binary complex are not statistically distinguishable from each other.

We refitted the same set of data for the distance Trp²⁰-Cys⁵¹ for the apo state and the holo state by holding the mean distance at 18 Å. The _R² value was elevated 10-fold from 1.17 to 12.3 for the apo state and 9-fold from 1.05 to 9.89 for the holo state (Table III), which indicates that the uncertainty of the best fitted value was not likely more than 2-3 Å. A similar analysis was applied to the distance Trp¹²-Cys⁵¹ with the same conclusion of the uncertainty in the best fitted mean distance. These results together with those from the variance ratio surfaces indicate that the data are easily adequate to demonstrate significantly longer intersite distances of Trp²⁰-Cys⁵¹ and Trp¹²-Cys⁵¹ in the holo state than in the apo state in the binary complex cTnC·cTnI.

The half-width recovered from the apo distributions of Trp²⁰-Cys⁵¹ was very narrow (3-4 Å) when compared with the half-width (10-11 Å) of the distribution for the corresponding apo Trp²²-Cys⁵² distance in fsTnC. The difference may reflect a difference in the structural dynamics of the N-domain between the two isoforms. It is necessary, however, to examine the extent to which other factors could contribute to the apparent half-width. Although these factors are not well understood, the angular motion of the donor and acceptor could contribute to the observed width if their rotational correlation times are 10 ps (7, 17). A narrower distribution for Trp²⁰-Cys⁵¹ compared with Trp²²-Cys⁵² in fsTnC could result from reduced angular motions of the fluorophores in the cardiac isoform. The amplitude of this motion can be assessed by the model-independent order parameter S². A value of S² < 1 indicates motion with respect to the protein. The S² values were calculated from the anisotropy data and are shown in Table IV. For the same donor-acceptor pair in equivalent positions in fsTnC, S² was 0.83 for the donor and 0.50 for the acceptor (7). The corresponding S² values were 0.82 and 0.62, respectively, for cTnC. The donor tryptophan in both isoforms had the same order parameter and hence very similar angular motions, but the acceptor in cTnC had a larger S² value and, therefore, a smaller amplitude in its angular motion than in fsTnC. This difference in the acceptor motion alone could contribute to the observed narrower distribution of the apo Trp²⁰-Cys⁵¹ distance in cTnC. The observed axial depolarization factors of both donor (Trp²⁰) and acceptor (Cys⁵¹-AEDANS) (Table IV) are quite comparable to those for Trp²² and Cys⁵²-AEDANS in fsTnC, suggesting similar upper and lower limits of orientational contributions to both sets of distribution curves. While these considerations are qualitative, they suggest that the observed narrow distribution for apo Trp²⁰-Cys⁵¹ in cTnC is a characteristic of the protein's structural dynamics rather than fluorophore motions.

View this table: [in this window] [in a new window]   Table IV Anisotropy parameters of tryptophan (donor) and AEDANS (acceptor) in cTnC mutants The limiting anisotropy values of the donor (ro,D) and acceptor (ro,A) were determined from the respective anisotropy decay curves. The axial depolarization factor of donor (dxD) and acceptor (dxA) were calculated as previously described (27). The order parameter S2 was calculated according to Lipari and Szabo (28), and the lower and upper limits of the orientation factors (2min and 2max) were calculated from the limiting anisotropies of donor and acceptor (27).

Anisotropy Decays-- In the calculation of the mean distances from the distributions of the distances, a value of 2/3 was used for the orientation factor ² based on the assumption of isotropic and rapid tumbling of the fluorophores. If this value was inappropriate or if the ² was different for the two biochemical states, the calculated distance parameters would be subjected to an uncertainty that cannot be quantitatively assessed. We measured the anisotropy decay of both the donor and acceptor in each mutant and its complex with cTnI (data not shown). The limiting anisotropy values of both fluorophores were used to calculate their respective axial depolarization factors and other related parameters. These results are listed in Table IV. The range of ² (²_min  ²_max) calculated from the depolarization factors was 0.20-2.76 for the three distances under most conditions. The expected error resulting from the use of the value 2/3 for the orientation factor to calculate the mean distance is less than 25%. The error was slightly higher for two cases in which the range of ² was slightly larger.

Effect of cTnI Peptides on cTnC Conformation-- Since the presence of cTnI appeared essential to elicit an open N-domain conformation in the presence of activator Ca²⁺, we next examined the effect of several cTnI peptides to induce a partially open or open cTnC conformation. Fig. 5 shows a set of steady-state emission spectra of the cTnC mutant 12W/51C in the presence of different cTnI peptides. The spectra in Fig. 5A show the maximal increase in the donor intensity and the decrease in the sensitized acceptor intensity (curve 2) (corresponding to maximal opening of the N-domain) recorded with the binary complex cTnC·cTnI induced by the addition of activator Ca²⁺. This loss in energy transfer serves as a control for the effect of cTnI peptides on domain opening. In the presence of the C-terminal half fragment of cTnI (P(128-211)), the donor enhancement induced by Ca²⁺ (Fig. 5B) was almost as large as that observed with the cTnC·cTnI complex, indicating that this long fragment was almost as effective as full-length cTnI in modulating the Ca²⁺-induced opening. The N-terminal half fragment (P(1-128)) was very ineffective in enhancing the donor peak (Fig. 5C) and had a negligible effect on the donor-acceptor distance. Within the sequence of the C-terminal fragment, peptide P(129-149) was not effective in causing domain opening (Fig. 5D). Peptide P(150-149) was more effective and the longer peptide P(129-166) induced a donor enhancement approaching that of cTnI. These peptide results are normalized to that obtained with full-length cTnI and are listed in Table V.

View larger version (30K): [in this window] [in a new window]   Fig. 5.   Effect of cTnI peptides on the distance between Trp12 and Cys51 in cTnC mutant 12W/51C. The energy donor was Trp12, and the acceptor was AEDANS linked to Cys51. A, steady-state emission spectra of the cTnC mutant complexed with full-length cTnI in the presence of Mg2+ (curve 1) and in pCa 4 (curve 2). B-F, spectra of the cTnC mutant in the presence of different cTnI peptides in the presence of Mg2+ (curve 1) and in pCa 4 (curve 2). In each panel, the holo spectrum (curve 2) was normalized to the peak intensity of the apo donor band (curve 1). The Ca2+-induced enhancement of the donor band shown in A for cTnC·cTnI was the maximum, and the enhancement observed with cTnC-peptide samples was always less than this maximum value. The samples in B-F all contained 1 µM cTnC and 10-20 µM of cTnI peptides. Other conditions: 1 mM EGTA, 0.3 M KCl, 50 mM MOPS at pH 7.1, and either 5 mM Mg2+ or 5 mM Mg2+ plus pCa 4.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Biochemical studies have established that in the absence of activator Ca²⁺ TnC from striated muscle interacts with TnI with a low affinity, and in the presence of activator Ca²⁺ a second mode of interaction occurs resulting in a large stabilization of the TnC-TnI complex. This Ca²⁺-dependent interaction triggers the contractile cycle by removing the inhibition of actomyosin ATPase and initiating strong interaction between myosin and the thin filament. The Ca²⁺ switch is the TnC-TnI linkage. The structural aspect of the switch mechanism cannot be fully understood without detailed knowledge of the component proteins of the switch, particularly the conformation of the Ca²⁺-loaded regulatory domain of TnC. Toward this goal, numerous studies have been reported to elucidate the Ca²⁺-induced structural changes in the N-terminal regulatory domain of fast skeletal TnC. From these studies of fsTnC, which include the crystal structures, solution NMR structures, and FRET, a detailed description has emerged for the Ca²⁺-induced transition of the N-domain of fsTnC from a closed to an open conformation. We have shown in the present FRET study that this domain opening is negligible in cardiac TnC and becomes substantial only in the presence of bound cardiac TnI.

For comparison with the previous FRET results of fsTnC, we used a cTnC mutant that contained the same donor-acceptor pair located in residues (Trp²⁰-Cys⁵¹) equivalent to those that were studied in fsTnC mutant (Trp²²-Cys⁵²). We also determined two additional intersite distances between Trp¹² and Cys⁵¹ and between Trp²⁰ and Cys⁸⁹. Of the four residues used in FRET measurements, three are within the N-domain. Residue 89 is at the N-terminal end of the central helix, and the results for the distance Trp²⁰-Cys⁸⁹ may reflect, in part, the property of the central helix. For this reason, we first discuss the energy transfer results from the first two distances.

A distinct feature of the distributions of the two distances Trp²⁰-Cys⁵¹ and Trp¹²-Cys⁵¹ is a narrow half-width in both the apo and holo states of isolated cTnC. The half-width values are a factor of 2-3 smaller than those observed with fsTnC. Trp²⁰ behaves very similarly to Trp²² in fsTnC, and it seems unlikely that the narrow half-width of the Trp²⁰-Cys⁵¹ distribution is due to a different angular mobility of the acceptor probe AEDANS linked to Cys⁵¹ as we have pointed out. The half-width of the distribution for the second distance Trp¹²-Cys⁵¹ is slightly larger, but still small when compared with the half-width values previously determined for the distances in fsTnC. Thus, the N-domain of cTnC in the apo state is considerably more constrained than that of apo-fsTnC. The mean distance of 15.7 Å for apo Trp²⁰-Cys⁵¹ is significantly longer than the values 9-10 Å for the corresponding distance in fsTnC, suggesting a partially open apo conformation in cTnC. This interpretation is consistent with the mean distance of 18 Å determined with holo-fsTnC. In the holo-cTnC state of Trp²⁰-Cys⁵¹, the half-width decreases by <1 Å and the mean distance increases by <1 Å. We have shown that these small changes are not statistically significant. Similarly, Ca²⁺ does not induce any change in the half-width and only a small increase (2 Å) in the mean distance for the Trp¹²-Cys⁵¹ distance. The small increase in the mean distance is not considered significant. These results are strong evidence that the apo N-domain of cTnC is already constrained and partially open and has a different average conformation from the apo N-domain of fsTnC. These surprising results are consistent with recent NMR studies which showed that the holo N-domain of cTnC has a more compact conformation compared with that of holo-fsTnC and is very similar to that of the apo-cTnC (18-20). The inability of activator Ca²⁺ to open up the N-domain of cTnC is demonstrated by two very different approaches and raises the possibility of different mechanisms for activation of skeletal and cardiac muscles.

The distance parameters recovered from the distribution of the Trp²⁰-Cys⁸⁹ distances require special considerations. The distribution of these distances in cTnC is insensitive to Ca²⁺, just as the distributions of the other two sets of distances are. However, the half-width is large (>8 Å) and approaching that observed with fsTnC. Since the conformation of the Trp²⁰ segment does not seem to be flexible, a contribution to the apparent dynamics of the polypeptide segment containing the acceptor would be a large angular motion of the acceptor probe. This is not a likely explanation because the fluorescence anisotropy properties of the acceptor linked to Cys⁸⁹ are not consistent with this possibility. The central helix of cTnC in solution is very flexible, and the structure of this segment of cTnC is undefined in the NMR structure (18). This segmental flexibility likely contributes to the observed apparent half-width. It is of interest that the half-width of this distribution decreases by about 2 Å in the apo-cTnC·cTnI complex, and by another 2 Å to a final value of 4.3 Å in the holo-cTnC·cTnI complex. Regardless of how cTnI interacts with cTnC, this interaction appears to reduce the segmental flexibility of the central helix, and the binding of Ca²⁺ to the single regulatory site in the N-domain further reduces the flexibility. The small increase (2.4 Å) in the mean distance in the holo complex is consistent with the notion that the spatial coordinates of Trp²⁰ (helix A) are not affected by Ca²⁺ binding to the N-domain. This apparent increase may reflect a constrained segment of the central helix imposed by bound cTnI in the holo complex.

Since the TnC-TnI linkage is generally considered the Ca²⁺ switch, the relevant structural changes in cTnC that may play a key role in the activation mechanism needs to be investigated also with the cTnC·cTnI complex. The distributions of the two distances Trp²⁰-Cys⁵¹ and Trp¹²-Cys⁵¹ clearly show a substantial Ca²⁺-induced opening when cTnC is complexed with cTnI. The formation of an apo-cTnC·cTnI complex has little or no effect on the distance parameters. In the presence of bound Ca²⁺ in the N-domain, the mean values of the two distances increase by more than 6 Å without much effect on the half-widths. The open conformation that is reported by both intersite distances has characteristics comparable to those previously observed for intersite distances in the N-domain of fsTnC. Thus, the binding of cTnI to cTnC is a prerequisite to achieve a Ca²⁺-induced open N-domain conformation. This open conformation must contain a segment of cTnI bound to the hydrophobic patch and represents the conformation of the regulatory domain resulting from the Ca²⁺-dependent interaction with cTnI. We suggest that a similar structural basis provides the mechanism by which the Ca²⁺-dependent interaction between TnC and TnI occurs in both skeletal and cardiac muscles. The difference between the two systems is in the mechanism by which the optimal open conformation of the regulatory domain is achieved. In skeletal TnC, the optimal conformation is modulated by the binding of two Ca²⁺ ions and is detectable with isolated fsTnC; in cardiac TnC, the open conformation is modulated by the binding of one Ca²⁺ ion coupled to formation of an apo complex with cTnI. The domain opening in cardiac TnC cannot be directly demonstrated with isolated cTnC, but is demonstrated in the apo-cTnC·cTnI complex. Once the optimal open conformation is achieved in either isoform, the Ca²⁺ modulated interaction with TnI can proceed in a similar manner. A recent NMR study of the structure of a complex formed between the N-domain of cTnC (residues 1-89) and a cTnI peptide (residues 147-163) showed a Ca²⁺-induced open conformation (21), similar to that previously observed with fsTnC in the absence of peptide. These NMR results are consistent with our FRET studies. Listed in Table III are relevant NMR distances corresponding to the distances determined by FRET. There is a general agreement between the FRET results obtained from cTnC·cTnI and the NMR results derived from the structure of (cTnC N-domain)-peptide complex.

Previous FRET results of fsTnC show a 10% overlap of the distributions between the apo and holo states, suggesting a fraction of the fsTnC molecules to be open in the apo state and an equilibrium between the closed and open conformations in both biochemical states. It has not been established whether Ca²⁺ binds to the fraction of fsTnC molecules with an open conformation, thus driving the equilibrium toward the open state, or Ca²⁺ binds to the closed conformation and forces the domain to open up. In contrast, within the cTnC·cTnI complex there is no overlap between the apo and holo distributions of the Trp²⁰-Cys⁵¹ distances (Fig. 4), and the corresponding overlap of the Trp¹²-Cys⁵¹ distributions is very small. These results suggest two dominant N-domain conformations of cTnC in cTnC·cTnI, one conformation in the apo complex and the other in the holo complex. This apo-cTnC conformation may be characterized as constrained and "partially open." Activator Ca²⁺ binds to this initial, partially open apo conformation and induces further opening to the optimal conformation.

We previously showed that the single regulatory site of cTnC in the cTnC·cTnI complex and cardiac troponin have essentially the same equilibrium binding constants for Ca²⁺ and very similar association and dissociation kinetic parameters for Ca²⁺ (13). What is structurally relevant in Ca²⁺ activation is the cTnC conformation in the presence of the second subunit, cTnI. Our results show that if structural information is to be used to gain insight into its relationship with function, it is necessary to investigate a minimal assembly that can support the function in question. In the present case, the cTnC·cTnI complex meets this requirement and yields information not available from cTnC alone.

The B helix has a kink at Glu⁴¹ in fsTnC and at Glu⁴⁰ in cTnC. These residues are at the beginning of their respective helix B in the two isoforms. In fsTnC, the side chain of Glu⁴¹ is the seventh Ca²⁺ ligand within the 12-residue binding loop of site I. This kink is straightened out upon binding Ca²⁺ to both sites I and II, thus allowing helix B to reorient and move away. The removal of the kink has been attributed to Ca²⁺-induced changes in the backbone dihedral angle at the base of helix B, which allows the two carboxylate oxygens of Glu⁴¹ to reach the bound Ca²⁺ ion (22). In cTnC, this kink remains in the holo state because of the absence of bound Ca²⁺ at the inactive site I (which includes Glu⁴⁰) (18). If a full opening of the cTnC N-domain is structurally dependent upon removal of the kink at Glu⁴⁰, the present FRET results would suggest that the Ca²⁺-independent formation of the cTnC·cTnI complex may result in a conformation in the cTnC N-domain favorable for subsequent removal of the kink.

The quantum yield of an extrinsic probe attached to Cys35 within the inactive binding loop of site I (residues 29-40) in apo-cTnC is 3-fold enhanced upon binding with cTnI (23). This result indicates that formation of an apo-cTnC·cTnI complex induces considerable structural perturbation on the inactive binding loop I and suggests a potential perturbation of the side chain of Glu⁴⁰. The nature of this perturbation is not known, but it is unlikely that the kink at Glu⁴⁰ is straightened out because apo complex formation does not shift the distributions of the two distances Trp²⁰-Cys⁵¹ and Trp¹²-Cys⁵¹. The perturbation, however, may cause the Glu⁴⁰ side chain to move to an intermediate position. The reorientation and movement of the helix C resulting from Ca²⁺ binding to site II then could provide the final perturbation resulting in movement of the Glu⁴⁰ side chain to the optimal position similar to that found for Glu⁴¹ in fsTnC in the presence of bound Ca²⁺ at site I. A central feature of this model is that the domain opening of cTnC is coupled to a perturbation of the inactive Ca²⁺ binding loop I. This perturbation is induced by cTnI in the apo state. This mechanism is not intrinsically different from that for fsTnC, although both Ca²⁺ site II and cTnI share the coupling in cardiac muscle. Besides structural consideration, the energetics for domain opening and exposure of a hydrophobic pocket should be balanced by the stabilization of the N-domain upon binding activator Ca²⁺. Presumably, this balance is available to fsTnC, but not to cTnC. The holo N-domain of cardiac TnC is expected to be less stable than that of fast skeletal TnC by about 7 kcal/mol because of the absence of a second bound Ca²⁺ (24).This destabilization relative to Ca²⁺-loaded fsTnC can be compensated by the formation of an apo-cTnC·cTnI, which releases about 10 kcal/mol (25). Thus, holo-cTnC·cTnI would be energetically favorable for domain opening to occur. In this model, a structural perturbation is coupled with a decrease of the free energy of the Ca²⁺-loaded complex to bring about the N-domain opening.

A question arises as to which region of cTnI is responsible for conferring the putative structural perturbation in cTnC. The N-terminal half of cTnI is ineffective in opening the domain conformation with E_r = 0.17, indicating a very small Ca²⁺-induced increase in the distance between the two sites Trp¹² and Cys⁵¹ and an essentially closed domain conformation. The effectiveness of the C-terminal half fragment (residues 128-211) approaches that of full-length cTnI in causing a domain opening (E_r = 0.87). Within the C-terminal half sequence, the sequence P(129-149) is substantially less effective (E_r = 0.17) than the sequence P(150-166) (E_r = 0.41). The sequence of the latter peptide corresponds to a segment in fsTnI (residues 115-131) that has been shown to bind to the hydrophobic patch of the B helix in fsTnC in the presence of Ca²⁺ (26). When P(150-166) is extended in the N-terminal end to include the sequence 129-149, the resulting peptide P(129-166) is more effective (E_r = 0.63) than the sequence 150-166 by itself. This enhanced effect was not observed when the two individual peptides P(129-149) and P(150-166) were added together to cTnC (results not shown). These results suggest that the cTnI sequence 150-166 plays a role in the opening of the N-domain of cTnC upon Ca²⁺ binding to site II. This conclusion is corroborated by the NMR study of the structure of cTnC N-domain complexed with the cTnI peptide (residues 147-163) (21). Results from the two complimentary methods have provided evidence of a role of a cTnI segment in modulating the Ca²⁺-dependent opening of the regulatory domain of cTnC. Since P(129-149) is known to bind to the C-domain, the central helix, and the N-domain, the longer peptide P(129-166) used in this study may simply provide a better anchor of the 150-166 segment to cTnC and enable the segment to interact with the hydrophobic patch more effectively.

If the cTnI segment between residues 150 and 166 interacts with the hydrophobic patch in helix B, it is possible that reorientation of helix C induced by Ca²⁺ binding at site II enables the 150-166 segment to become juxtaposed to its hydrophobic patch target and to interact with it. In this alternate mode of domain opening, the mechanism does not necessarily depend upon removal of the kink at Glu⁴⁰. Domain opening, however, would be facilitated by "partial opening" of the regulatory domain prior to the binding of Ca²⁺ to site II. The FRET results show that this conformation exists because the mean values of the two distances in either isolated apo-cTnC or apo-cTnC·cTnI are several angstroms longer than the corresponding separations in the apo regulatory domain of fsTnC. This partially open conformation of the apo N-domain in the complex is highly constrained, and these structural features could facilitate insertion of a cTnI segment toward the hydrophobic patch. This model would explain the change in energy transfer evident with P(150-166). The role of full-length cTnI would be to provide an anchor and optimal proximity of the critical segment of residues 150-166 to the B helix rather than driving a conformational coupling between the inactive binding loop I and the open conformation. Further work is required to understand what roles other regions of cTnI may play in modulating the Ca²⁺-induced conformational transition of the N-domain of cTnC in cardiac troponin.

In summary, we have provided direct evidence to show for the first time that bound cTnI is required for the Ca²⁺-dependent transition of the regulatory N-domain of intact cardiac TnC to a substantial open conformation. The binding of cTnI is a prerequisite for this transition to occur. The open regulatory domain is sufficiently large for interaction of a second region of cTnI with a hydrophobic pocket. The requirement of cTnI for domain opening in cTnC has not been previously recognized and suggests an additional role of cTnI in Ca²⁺ activation of cardiac muscle.

	FOOTNOTES

^* This work was supported, in part, by National Institutes of Health Grants HL52508 (to H. C. C.) and HL63377 (to R. J. S.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^** To whom correspondence should be addressed: 520 CH-19, Dept. of Biochemistry and Molecular Genetics, University of Alabama at Birmingham, Birmingham, AL 35294-2041. Tel.: 205-934-2485; Fax: 205-975-4621; E-mail: hccheung@uab.edu.

	ABBREVIATIONS

The abbreviations used are: TnC, troponin C; TnI, troponin I; fsTnC, fast skeletal muscle TnC; fsTnI, fast skeletal muscle TnI; cTnC, cardiac muscle TnC; cTnI, cardiac muscle TnI; FRET, fluorescence resonance energy transfer; CDTA, trans-1,2-diaminocyclohexane-N, N,N',N'- tetraacetic acid; HPLC, high pressure liquid chromatography; IAEDANS/1,5-IAEDANS, 5-(iodoacetamidoethyl)aminonaphthalene-1-sulfonic acid; MOPS, 3-(N-morpholino)propanesulfonic acid.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea