Mutations of hydrophobic residues in the N-terminal domain of troponin C affect calcium binding and exchange with the troponin C-troponin I96-148 complex and muscle force production.

Interactions between troponin C and troponin I play a critical role in the regulation of skeletal muscle contraction and relaxation. We individually substituted 27 hydrophobic Phe, Ile, Leu, Val, and Met residues in the regulatory domain of the fluorescent troponin C(F29W) with polar Gln to examine the effects of these mutations on: (a) the calcium binding and dynamics of troponin C(F29W) complexed with the regulatory fragment of troponin I (troponin I(96-148)) and (b) the calcium sensitivity of force production. Troponin I(96-148) was an accurate mimic of intact troponin I for measuring the calcium dynamics of the troponin C(F29W)-troponin I complexes. The calcium affinities of the troponin C(F29W)-troponin I(96-148) complexes varied approximately 243-fold, whereas the calcium association and dissociation rates varied approximately 38- and approximately 33-fold, respectively. Interestingly, the effect of the mutations on the calcium sensitivity of force development could be better predicted from the calcium affinities of the troponin C(F29W)-troponin I(96-148) complexes than from that of the isolated troponin C(F29W) mutants. Most of the mutations did not dramatically affect the affinity of calcium-saturated troponin C(F29W) for troponin I(96-148). However, the Phe(26) to Gln and Ile(62) to Gln mutations led to >10-fold lower affinity of calcium-saturated troponin C(F29W) for troponin I(96-148), causing a drastic reduction in force recovery, even though these troponin C(F29W) mutants still bound to the thin filaments. In conclusion, elucidating the determinants of calcium binding and exchange with troponin C in the presence of troponin I provides a deeper understanding of how troponin C controls signal transduction.

Interactions between troponin C and troponin I play a critical role in the regulation of skeletal muscle contraction and relaxation. We individually substituted 27 hydrophobic Phe, Ile, Leu, Val, and Met residues in the regulatory domain of the fluorescent troponin C F29W with polar Gln to examine the effects of these mutations on: (a) the calcium binding and dynamics of troponin C F29W complexed with the regulatory fragment of troponin I (troponin I 96 -148 ) and (b) the calcium sensitivity of force production. Troponin I 96 -148 was an accurate mimic of intact troponin I for measuring the calcium dynamics of the troponin C F29W -troponin I complexes. The calcium affinities of the troponin C F29W -troponin I 96 -148 complexes varied ϳ243-fold, whereas the calcium association and dissociation rates varied ϳ38and ϳ33fold, respectively. Interestingly, the effect of the mutations on the calcium sensitivity of force development could be better predicted from the calcium affinities of the troponin C F29W -troponin I 96 -148 complexes than from that of the isolated troponin C F29W mutants. Most of the mutations did not dramatically affect the affinity of calcium-saturated troponin C F29W for troponin I 96 -148 . However, the Phe 26 to Gln and Ile 62 to Gln mutations led to >10-fold lower affinity of calcium-saturated troponin C F29W for troponin I 96 -148 , causing a drastic reduction in force recovery, even though these troponin C F29W mutants still bound to the thin filaments. In conclusion, elucidating the determinants of calcium binding and exchange with troponin C in the presence of troponin I provides a deeper understanding of how troponin C controls signal transduction.
Troponin C (TnC) 1 regulates striated muscle contraction and relaxation through the binding and release of Ca 2ϩ (for review see Refs. [1][2][3]. Skeletal muscle TnC (ϳ18 kDa) consists of globular N-and C-terminal domains connected by a 31-residue ␣-helix (for review see Refs. 4 and 5). Both domains bind two Ca 2ϩ ions through a pair of EF hand Ca 2ϩ -binding motifs. Each pair of EF hands interacts with one another through a short antiparallel ␤-sheet connecting the two Ca 2ϩ -binding loops (Ref. 6 and references within). The EF hands are numbered I-IV, and the helices flanking the loops are designated A-H, with an additional N-terminal 14-residue ␣-helix ( Fig. 1, Nhelix), which is absent in the closely related EF hand Ca 2ϩbinding protein calmodulin.
Much is known about the cation binding properties of TnC in solution. Each EF hand system binds Ca 2ϩ and Mg 2ϩ competitively, with the two C-terminal EF hands possessing higher Ca 2ϩ and Mg 2ϩ affinities (6 -8). In fact, the Ca 2ϩ -binding sites of the C-domain of TnC possess ϳ10-fold higher Ca 2ϩ affinity with a greater than 100-fold slower Ca 2ϩ dissociation rate compared with those in the N-domain (6,9). In part because of its high Ca 2ϩ and Mg 2ϩ affinities and slow Ca 2ϩ exchange rates (as compared with the kinetics of muscle contraction and relaxation), the C-domain is thought to play a structural role in muscle function by anchoring TnC into the Tn complex. In contrast, the Ca 2ϩ exchange rates of the N-domain of TnC are rapid enough to be involved in the dynamic Ca 2ϩ -dependent regulation of muscle mechanics (for review see Refs. 3 and 10).
Skeletal muscle contraction begins when cytoplasmic [Ca 2ϩ ] rises and binds to the N-terminal EF hands of TnC. The entire N-domain of TnC subsequently undergoes a large tertiary conformational change, in which helices B and C move away as a unit from helices N, A, and D, exposing a buried hydrophobic pocket to the solvent (Ref. 5 and references within). The newly formed hydrophobic pocket is thought to allow the N-domain of TnC to interact with the C-terminal of TnI transferring the inhibitory domain of TnI away from actin (11). Concurrently or subsequently, tropomyosin changes its position on the actin filament and myosin then binds cyclically to actin causing muscle contraction (for review see Refs. 1-3). As cytoplasmic [Ca 2ϩ ] lowers, the sequence of events above reverses (not necessarily in the same order), and the muscle relaxes. One of the steps that may influence the rate of muscle relaxation is Ca 2ϩ dissociation from the N-domain of TnC.
The influence that TnC has on the kinetics of muscle relaxation is controversial and incompletely understood (Ref. 12 and references within). The actual rate that Ca 2ϩ dissociates from TnC in muscle fibers has not been measured and thus must be inferred. Ca 2ϩ dissociates from the regulatory domain of isolated TnC ϳ20 -30 times faster than skeletal muscle relaxation and thus has been speculated to be too rapid to influence the rate of relaxation (for review see Ref. 3). However, the antiparallel binding of TnI to TnC increases the Ca 2ϩ sensitivity of the N-domain of TnC ϳ10-fold and slows the Ca 2ϩ dissociation  rate ϳ30-fold, with little additional change upon the formation  of the whole Tn complex (Refs. 8 and 13-16; for review see Ref. 10). Thus, the rate of Ca 2ϩ dissociation from the Tn complex and not TnC alone may be the more meaningful rate when considering factors that control muscle relaxation kinetics. Consistent with this idea, exchanging a TnC mutant with an ϳ2-fold slower N-terminal Ca 2ϩ dissociation rate into skeletal muscle fibers slowed the rate at which the fibers relaxed ϳ2fold (12). However, in the same study exchange of an ϳ1.5-fold faster TnC mutant into muscle did not statistically increase the rate of relaxation. Clearly, a broader range of Ca 2ϩ dissociation rates from TnC mutants is required to further probe the role of TnC in tuning the rate of striated muscle relaxation.
To better understand the regulation of muscle mechanics, it is important to elucidate the Ca 2ϩ -dependent interactions of TnC with TnI because TnC regulates muscle contraction as a part of the Tn complex and not in isolation. TnI residues 96 -116 (TnI 96 -116 ) bind actin and are primarily responsible for the ability of TnI to inhibit the ATPase activity of actomyosin, which can be reversed upon TnC-Ca 2ϩ binding to TnI 96 -116 (17)(18)(19). In conjunction with residues 96 -116, residues 117-148 of TnI are required for the complete inhibitory activity and regulatory interactions with actin and TnC (20 -23). Furthermore, the complete enhancement of the Ca 2ϩ sensitivity and the slowing of the Ca 2ϩ dissociation rate from the regulatory domain of TnC in the presence of intact TnI were mimicked by a peptide of TnI corresponding to residues 96 -148 (TnI 96 -148 ) (8). Thus, the Ca 2ϩ -dependent binding of the regulatory domain of TnC to TnI 96 -148 may be a good model system to study the Ca 2ϩ -dependent interactions between TnI and TnC that regulate muscle mechanics.
Recently, we investigated the effect of hydrophobic residue substitutions on the Ca 2ϩ binding properties of the regulatory domain of TnC with the Phe 29 3 Trp mutation (TnC F29W ). The global N-terminal Ca 2ϩ affinities of the TnC F29W mutants varied 2340-fold, whereas the Ca 2ϩ association and dissociation rates varied less than 70-fold and more than 45-fold, respectively (6). In the present study we have determined how these mutations affect the Ca 2ϩ binding properties and dynamics of the TnC F29W -TnI 96 -148 complex and located hydrophobic residues essential for high affinity binding of TnI 96 -148 . Furthermore, we have tested whether the TnC F29W -TnI 96 -148 complex is a better predictor than isolated TnC F29W for the Ca 2ϩ binding properties of the Tn complex in muscle and the potential for a particular TnC F29W mutant to support force production.

EXPERIMENTAL PROCEDURES
Materials-Phenyl-Sepharose CL-4B and EGTA were purchased from Sigma. Quin-2 was purchased from Calbiochem (La Jolla, CA). All other chemicals were of analytical grade. The TnI 96 -148 peptide was synthesized and purified by the Alberta Peptide Institute (Edmonton, Canada).
Protein Mutagenesis and Purification-The construction and expression of intact chicken skeletal TnC F29W in pET3a has been described (24). Chicken skeletal fast TnI was prepared as described for the rabbit protein (25). The TnC F29W mutants were constructed from the TnC F29W plasmid by primer based site-directed mutagenesis using a Stratagene QuikChange site-directed mutagenesis kit. The mutations were confirmed by DNA sequence analysis. The plasmids for TnC F29W and its mutants were transformed into Escherichia coli BL21(DE3)pLysS cells (Novagen) and purified as described previously (6). Aliquots of TnC F29W and I62QTnC F29W were labeled with the Cys-specific fluorescent probe 5-((((2-iodoacetyl)amino)ethyl)amino) naphthalene-1-sulfonic acid (IAEDANS) at position Cys 101 for the myofibril studies. Each TnC was reacted with 3-5-fold molar excess of IAEDANS for ϳ6 h at room temperature with constant shaking in 50 mM Tris, 90 mM KCl, 1 mM EGTA, 6 M urea, pH 7.5. The labeling reaction was stopped by the addition of 2 mM DTT, and the labeled proteins were exhaustively dialyzed against 10 mM MOPS, 90 mM KCl, pH 7.0, at 4°C to remove unreacted label.
Determination of Ca 2ϩ Affinities-All steady-state fluorescence measurements were performed using a Perkin-Elmer LS5 Spectrofluorimeter at 15°C. Trp fluorescence was excited at 275 nm and monitored at 345 nm as microliter amounts of CaCl 2 were added to 1 ml of each TnC F29W mutant (0.3 M) plus TnI 96 -148 (3 M) in 200 mM MOPS (to prevent pH changes upon the addition of metal), 90 mM KCl, 2 mM EGTA, 1 mM DTT, pH 7.0, at 15°C. The [Ca 2ϩ ] free was calculated using the computer program EGCA02 developed by Robertson and Potter (26). The Ca 2ϩ affinities are reported as dissociation constants (K d ).
Each K d represents the mean of 3-5 titrations fit with a logistic sigmoid function mathematically equivalent to the Hill equation, as previously described (6).
Determination of TnI 96 -148 Peptide Affinities-Trp fluorescence was monitored as described in the previous paragraph. Microliter amounts of TnI 96 -148 were added to 1 ml of each TnC F29W mutant (0.6 M) in 200 mM MOPS, 90 mM KCl, 2 mM EGTA, 1 mM [Ca 2ϩ ] free, 1 mM DTT, pH 7.0, at 15°C. Each peptide affinity, reported as a dissociation constant, represents the mean of three to five titrations fit to the root of a quadratic equation for binary complex formation as previously described (27).
Determination of Ca 2ϩ Dissociation Rates-Ca 2ϩ dissociation rates (k off ) were measured using an Applied Photophysics Ltd. (Leatherhead, UK) model SX.18 MV stopped flow instrument with a dead time of 1.4 ms at 15°C. The samples were excited using a 150-watt Xenon arc source. Ca 2ϩ dissociation from the N-terminal regulatory domain of TnC F29W and its mutants when complexed with TnI 96 -148 were obtained utilizing Trp fluorescence excited at 275 nm with emission monitored through a UV transmitting black glass filter (UG1 from Oriel, Stratford, CT). k off was also measured using the fluorescent Ca 2ϩ chelator Quin-2 (6,8). Quin-2 was excited at 330 nm with its emission monitored through a 510-nm broad band pass interference filter (Oriel, Stratford, CT). The buffer used in all stopped flow experiments was 10 mM MOPS, 90 mM KCl, 1 mM DTT, pH 7.0.
Calculation of Ca 2ϩ Association Rates-The Ca 2ϩ association rates (k on ) were calculated using the simple relationship k on ϭ k off /K d , where k off represents the concerted release of two Ca 2ϩ ions, and K d represents the binding event of two Ca 2ϩ ions to the N-domain of TnC in the presence of TnI 96 -148 , as previously described (8).
Muscle Fiber Experiments-Single fibers were isolated the day of use from bundles of rabbit psoas muscle that had been stored in a glycerinating solution at Ϫ20°C no longer than 1 month. Solutions and the mechanical setup utilized for force measurements were as previously described (28). Briefly, a single fiber was soaked in relaxing solution containing 1% (v/v) Triton X-100 for 5 min to remove any residual sarcolemma and sarcoplasmic reticulum. The fiber was then tied down in troughs attached to a servo-controlled DC torque motor (Cambridge Technologies, Watertown, MA) and an isometric force transducer (model 403A, Cambridge Technologies) as previously described (29). Fiber sarcomere length, width, and depth were measured with a video camera (Sony model XC-ST70) and an image analysis system (Simple PCI; Compix Inc., Cranberry Township, PA). Resting sarcomere length was set between 2.50 and 2.60 m. The fiber was then activated in a pCa 4.0 solution and rapidly slackened after isometric force reached plateau. The analogue output of the force transducer was digitized using a DaqBoard/2000 and Daqview software (Iotech Inc., Cleveland, OH). The total force was measured between the plateau and base-line levels. The same procedure was utilized to obtain the resting force level of the fiber in a pCa 9.0 solution. The active force generated by the fiber in the various pCa solutions was calculated as the total force minus the resting force. Three active force measurements were performed in pCa 4.0 with the final activation taken as the maximal force generated by the native fiber (i.e. prior to extraction of endogenous TnC), which led to an average force per cross-sectional area of 85 Ϯ 5 kN/m 2 . The fiber was then soaked for 2 min in a TnC extraction solution containing 5 mM EDTA, 10 mM HEPES, and 0.5 mM trifluoperazine dihydrochloride at pH 7.0 (30). The fiber was then washed three times in pCa 9.0 solution to remove any residual trifluoperazine dihydrochloride. If the residual force in pCa 4.0 solution was Ͼ10% of the maximal force, the extraction process was repeated. The fibers were then soaked for 2 min in a pCa 9.0 solution containing 16.7 M recombinant TnC F29W or its mutant. All of the reconstituted fibers were then exposed to a series of pCa solutions varying from pCa 9.0 to 4.0, and the active force versus pCa was measured. Every fourth activation was performed at pCa 4.0, to which each adjacent and randomized pCa was normalized.
Myofibril Preparation and Experiments-Rabbit skeletal psoas myo-fibrils were prepared and stored as previously described (31). Endogenous TnC was extracted from a sample of the stock myofibrils by first washing the myofibrils three times in a myofibril TnC extraction solution (10 mM MOPS, 90 mM KCl, 5 mM EDTA, 2 mM DTT, 0.02% Tween 20, pH 8.0) to remove any residual glycerol. The myofibrils were then soaked in the TnC extraction solution for approximately 10 min at room temperature, pelleted, and resuspended in fresh TnC extraction solution an additional three times. The TnC extracted myofibrils were then washed three times in a pCa 9 solution after which the concentration of the myofibrils was determined (31). 0.1 mg/ml aliquots of the TnC extracted myofibrils in the pCa 9.0 solution were then exposed to 1 M TnC F29W or I62QTnC F29W labeled with IAEDANS for approximately 5 min. The myofibrils were then diluted with a pCa 3.0 solution to bring the final pCa of the solution to 4.0 according to a mixing table. The reconstituted myofibrils were then plated on a glass slide with a coverslip and imaged as previously described (31,32). Briefly, the images were collected using a Zeiss Axiovert TV (Thornwood, NJ) epifluorescence microscope equipped with a 100ϫ oil immersion phase contrast lens and a Chroma filter set #11000UV (360 nm broad excitation, 400-nm-long pass dichroic and 420-nm-long pass emission). The digital images were obtained using a 12-bit intensity resolution CCD camera (Kodak KAF 1300 chip; Photometrics, Tucson, AZ) controlled by a Matrox board and IPLab Spectrum software (version 3.0, Signal Analytics, Vienna, VA) run by a Macintosh 840AV.

Effect of Ca 2ϩ on the Fluorescence Spectra of TnC F29W and Mutants in the Presence of TnI 96 -148 at 15°C-
The influence of individual hydrophobic to Gln mutations in the regulatory domain of TnC F29W on Ca 2ϩ binding and exchange was previously characterized (6). However, TnC does not function in muscle in isolation but as part of the Tn complex, primarily interacting with TnI. Thus, we determined the influence of 27 TnC F29W mutants in which all N-terminal hydrophobic residues were individually substituted with Gln ( Fig. 1) on Ca 2ϩ binding and exchange in the presence of the regulatory peptide of TnI, TnI 96 -148 . The N-terminal domain of TnC F29W undergoes a large increase in Trp fluorescence upon forming the Ca 2ϩ -TnC F29W -TnI 96 -148 complex (8,20). In the absence of Ca 2ϩ , the Trp fluorescence of TnC F29W upon the addition of TnI 96 -148 marginally decreased Ͻ1.1-fold at 345 nm ( Fig. 2A Table I summarizes the Ca 2ϩ binding data for these and the remaining TnC F29W mutants. In the presence of TnI 96 -148 , TnC F29W exhibited a half-maximal increase in its Trp fluorescence upon the addition of Ca 2ϩ at 267 Ϯ 3 nM. The Ca 2ϩ affinities for the mutants ranged from 70 Ϯ 1 nM for L49QTnC F29W to 17 Ϯ 3 M for F26QTnC F29W . Therefore, substitution of hydrophobic residues with polar Gln produced N-domain TnC F29W mutants that exhibited ϳ243-fold variation in their Ca 2ϩ affinities in the presence of TnI 96 -148 . The Hill coefficients for all but two of the TnC F29W -TnI 96 -148 mutant complexes (F26QTnC F29W and I37QTnC F29W ) were between 1.6 and 2.8 (see Table I), implying cooperative binding of Ca 2ϩ and TnI 96 -148 to TnC F29W and its mutants.
Similar to the binding of TnI to TnC, the binding of TnI 96 -148 to TnC F29W increases the Ca 2ϩ sensitivity of the regulatory domain of TnC F29W ϳ12-fold (Table I and Refs. 8 and 13-15). On average, the Ca 2ϩ sensitivity of the TnC F29W mutants increased ϳ11-fold (Table I) Table I summarizes the TnI 96 -148 binding data for these and the remaining TnC F29W mutants. In the presence of a saturating concentration of Ca 2ϩ , TnC F29W bound to TnI 96 -148 with an affinity of 146 Ϯ 19 nM, which is in agreement with a previously reported value (20). Hydrophobic interactions are important for the Ca 2ϩ -dependent binding of the N-domain of TnC to the C-domain of TnI. Thus, it seemed logical that substitution of hydrophobic residues with polar Gln in the N-domain of TnC would likely decrease its affinity for TnI 96 -148 . However, Fig. 4 (Table I). Therefore, substitution of hydrophobic residues with polar Gln in the regulatory domain of TnC F29W increased (ϳ16-fold) and decreased (ϳ2-fold) the Ca 2ϩ dissociation rate from the TnC F29W -TnI 96 -148 complex, creating an ϳ33-fold variation.
To verify that the time course of the EGTA-induced Trp fluorescence decreases for the TnC F29W -TnI 96 -148 complexes followed Ca 2ϩ dissociation and not a slower structural change, Ca 2ϩ dissociation was also measured using the fluorescent Ca 2ϩ chelator Quin-2. Whereas the fluorescence of Trp was selective for the events of N-terminal Ca 2ϩ dissociation, Quin-2 fluorescence reported Ca 2ϩ dissociation from both the N-and C-domains of TnC F29W and its mutants complexed with TnI 96 -148 . However, the Ca 2ϩ dissociation rates from the Nterminal domain of TnC F29W and its mutants were easily distinguished from the rates of Ca 2ϩ dissociation from the Cterminal domain (on average 0.159 Ϯ 0.007 s Ϫ1 ) because the latter rates were Ͼ30-fold slower in the presence of TnI 96 -148 or intact TnI. Fig. 5B demonstrates that for all of the mutants, the Ca 2ϩ dissociation rate reported by Trp was in excellent agreement with the N-terminal rate determined by Quin-2. Therefore, the fluorescent Trp signal accurately reports Ca 2ϩ binding and dissociation from the TnC F29W -TnI 96 -148 mutant complexes.
To verify that TnI 96 -148 is a satisfactory model system for the regulatory domain binding of TnC to TnI, stopped flow studies were also conducted with intact chicken skeletal TnI. Fig. 5C shows the time course of the increases in Quin-2 fluorescence as Ca 2ϩ was dissociated from the N-terminal domains of M81QTnC F29W (4 s Ϫ1 ), L49QTnC F29W (5 s Ϫ1 ), TnC F29W (9 s Ϫ1 ), I73QTnC F29W (22 s Ϫ1 ), and F26QTnC F29W (135 s Ϫ1 ) complexed with intact TnI. The Ca 2ϩ dissociation rates measured from the regulatory domain of TnC F29W and its mutants in the presence of TnI 96 -148 were similar to that measured in the presence of intact TnI. Therefore, the TnC F29W -TnI 96 -148 complex is a good model system to study the regulatory mechanisms of the Ca 2ϩdependent binding of TnC to TnI.
Calculation of Ca 2ϩ Association Rates-The Ca 2ϩ association rates to TnC F29W and its mutants in the presence of TnI 96 -148 were calculated using the Ca 2ϩ K d and k off values determined by Trp (k on ϭ k off /K d ; Table I). The calculated k on for Tnc F29W in the presence of TnI 96 -148 was ϳ4.5 ϫ 10 7 M Ϫ1 s Ϫ1 , which was ϳ2.4-fold slower than the k on calculated or measured in the absence of the peptide (6,8 Fig. 6A shows that V45QTnC F29W and M46QTnC F29W possess ϳ19and 3.6-fold higher Ca 2ϩ affinity than TnC F29W in the absence of TnI 96 -148 , respectively (Table  II and Ref. 6). On the other hand, M81QTnC F29W and F78QTnC F29W display ϳ5.9and 8.4-fold lower Ca 2ϩ affinity than TnC F29W in the absence of TnI 96 -148 , respectively (Table  II and Ref. 6). However, Fig. 6B shows that only the Ca 2ϩ sensitivities of V45QTnC F29W and F78QTnC F29W in the presence of TnI 96 -148 remain higher (ϳ3-fold) and lower (ϳ19-fold) than TnC F29W , respectively (see also Table II). Thus, the qualitative and quantitative changes in N-terminal Ca 2ϩ sensitivities for several of the TnC F29W mutants compared with TnC F29W in the presence or absence of TnI 96 -148 were not the same.
To test which TnC F29W system (with or without TnI 96 -148 ) better represents the Ca 2ϩ sensitivity of force production in muscle, the endogenous TnC in psoas muscle fibers was extracted and then replaced with TnC F29W or its mutants, and force versus pCa was measured. After TnC extraction, the average force generated by the single skinned muscle fibers was 2.3 Ϯ 0.5% of the maximal force (data not shown). Subsequent reconstitution of the muscle fibers with V45QTnC F29W , M46QTnC F29W , TnC F29W , M81QTnC F29W , or F78QTnC F29W recovered 82 Ϯ 5, 73 Ϯ 4, 90 Ϯ 3, 65 Ϯ 8, and 80 Ϯ 2% of the maximal force at pCa 4, respectively. Fig. 6C demonstrates that the Ca 2ϩ dependence of force generation with TnC F29W or its mutants followed qualitatively more closely to the Ca 2ϩ sensitivities of the mutant TnC F29W -TnI 96 -148 complexes and not to that of the isolated TnC F29W proteins (see also Table II). Thus, the Ca 2ϩ -dependent behavior of the TnC F29W -TnI 96 -148 complex is a better indicator for how the TnC mutant will control the Ca 2ϩ dependence of force generation when incorporated into muscle fibers.

FIG. 4. TnI 96 -148 binding to Ca 2؉saturated TnC F29W and its mutants.
The TnI 96 -148 -dependent decreases in Trp fluorescence are shown for Ca 2ϩ -saturated L49QTnC F29W (q), TnC F29W (OE), M81QTnC F29W (Ⅺ), I73QTnC F29W (‚), and F26QTnC F29W (E) as a function of TnI 96 -148 concentration. Microliter amounts of TnI 96 -148 were added to each TnC (0.6 M) in the same buffer and temperature as in Fig. 2 (6). Because these data were collected under identical experimental conditions as were used here, we have for comparative purposes reproduced these data in this figure and Table II. Microliter amounts of Ca 2ϩ were added to 1 ml of each protein (0.3 M) in the same buffer and temperature as described in the legend to Fig. 2. Trp fluorescence was monitored as described in the legend to Fig. 3. Another effect observed in the reconstituted muscle, which could not be predicted by the Ca 2ϩ binding properties of the isolated TnC, was the maximal amount of force recovered by a particular mutant. Fig. 7A shows the Ca 2ϩ -dependent increases in force recovered by TnC F29W (OE), L42QTnC F29W (ૺ), I73QTnC F29W (‚), and I62QTnC F29W (f). At pCa 4.0, TnC F29W , L42QTnC F29W , I73QTnC F29W , and I62QTnC F29W recovered 90 Ϯ 3, 79 Ϯ 3, 45 Ϯ 6, and 12 Ϯ 4% of the force generated by the endogenous TnC prior to extraction, respectively. Similar to I62QTnC F29W , I37QTnC F29W and F26QTnC F29W also recovered force poorly at 15 Ϯ 1 and 13 Ϯ 1% of the maximal amount of force generated by the fiber prior to endogenous TnC extraction, respectively (data not shown). Thus, the amount of maximal force sustained by the TnC F29W mutants was variable, with three of the mutants only marginally allowing any force production. As will be discussed, the binding of TnI 96 -148 to the Ca 2ϩ -saturated TnC F29W mutants may offer clues as to why some of the mutants support little force.
To test whether I62QTnC F29W was actually binding to the thin filaments in the TnC extracted muscle fibers, additional TnC exchange experiments were performed on the muscle. Fig.  7B (Ⅺ) at time 0 shows the maximal force recovered by TnC F29W at pCa 4.0 in a reconstituted muscle fiber. The fiber was then transferred to a relaxing solution containing 16.7 M I62QTnC F29W , and the force generated at pCa 4.0 was measured at several time intervals. The amount of force production decreased with time, eventually reaching a value similar to that generated by fibers solely reconstituted with I62QTnC F29W . The data indicate that I62QTnC F29W was able to bind to the thin filaments and competitively displace TnC F29W from the Tn complex. Furthermore, when a muscle fiber was initially reconstituted with I62QTnC F29W (Fig. 7B, time 0, f) and then competitively displaced with 16.7 M TnC in relaxing solution, maximal force increased with a time course similar to that at which I62QTnC F29W inhibited the force generated with TnC F29W . When the TnC-binding sites in the fiber were vacant (i.e. after endogenous TnC extraction), the addition of TnC at the concentration used for the competitive binding studies caused force to be maximal within 2 min (data not shown). Results similar to that of I62QTnC F29W were obtained in the displacement studies when I37QTnC F29W or F26QTnC F29W were tested (data not shown). Thus, the data supports the hypothesis that the TnC F29W mutants that minimally support force (Ͻ15%) bind to the thin filament and form the Tn complex.
To directly visualize whether I62QTnC F29W was able to incorporate into the TnC-depleted muscle fiber, both TnC F29W and I62QTnC F29W were labeled with the extrinsic fluorescent probe IAEDANS and reconstituted into psoas myofibrils. Fig.  7C shows representative phase contrast images of TnC F29W -IAEDANS (top left panel) and I62QTnC F29W -IAEDANS (top right panel) reconstituted myofibrils. As can be seen from the fluorescent images (Fig. 7C, bottom left panel for TnC F29W -IAEDANS and bottom right panel for I62QTnC F29W -IAE-DANS), both IAEDANS-labeled TnC proteins incorporate into the myofibril at the myosin-actin filament overlap and nonoverlap space. Thus, as predicted from the physiological competition experiments, I62QTnC F29W binds to the thin filament at a similar location, as does TnC F29W , and forms the Tn complex, albeit in an inactive state. DISCUSSION The goal of the present study was to examine the effect of the hydrophobic mutations on the Ca 2ϩ binding properties of the TnC F29W -TnI 96 -148 complex and on the affinity of TnC F29W for TnI 96 -148 . Furthermore, we wanted to examine whether the effect of hydrophobic mutations on the Ca 2ϩ sensitivity of force development could be better predicted by the Ca 2ϩ and TnI 96 -148 binding properties of the TnC F29W mutants than by that of the isolated TnC F29W . Because the regulatory domain of chicken TnC is spectroscopically silent, the Phe 29 3 Trp mutation was utilized to follow the structural changes in the N-domain of TnC induced by changes in Ca 2ϩ concentration (6,8,9,12,24,(33)(34)(35). In our previous study, all 27 Phe, Ile, Leu, Val, and Met residues were individually mutated to polar Gln to examine the role of hydrophobic residues in Ca 2ϩ binding and exchange with the regulatory domain of intact TnC F29W in isolation (6). The hydrophobic TnC F29W mutants exhibited ϳ2340-fold variation in their Ca 2ϩ binding affinities. Indicative of the Ca 2ϩ affinity changes, the hydrophobic TnC F29W mutants also exhibited less than 70-fold and more than 45-fold variation in their Ca 2ϩ association rates and dissociation rates, respectively (6). The data indicated that the local side chain interactions of the hydrophobic residues within the tertiary structures of the apo and Ca 2ϩ -bound regulatory domain of TnC F29W played an important role in dictating the Ca 2ϩ binding properties of the protein.
The Ca 2ϩ affinities of the mutant TnC F29W -TnI 96 -148 complexes varied ϳ243-fold. However, the variation in the Ca 2ϩ sensitivity of the mutants in the absence of TnI 96 -148 was an order of magnitude larger (6). It would appear that Ca 2ϩ binding is optimized when the regulatory domain of TnC is in the open state (helices B and C swing away from helices N, A, and D) (6). The binding of TnI or C-terminal peptides of TnI to the regulatory domain of TnC help to lock TnC into the open state and thus enhance the Ca 2ϩ binding affinity of TnC ϳ10 -12fold (8,(13)(14)(15). The high Ca 2ϩ affinity mutants of TnC F29W (F22QTnC F29W , V45QTnC F29W , M46QTnC F29W , L49QTnC F29W , and M82QTnC F29W ) may mimic TnI binding to TnC by shifting the equilibrium of the regulatory domain of TnC into the open state and away from the closed state (6). Thus, binding of TnI 96 -148 to the high Ca 2ϩ affinity mutants of TnC F29W increases their Ca 2ϩ sensitivities less than 5-fold as compared with an ϳ12-fold increase in Ca 2ϩ sensitivity for TnC F29W . The reduced Ca 2ϩ sensitivity enhancement to the high affinity Ca 2ϩ -binding mutants by TnI 96 -148 is the primary reason for the decreased variation in the Ca 2ϩ binding affinities of the TnC F29W -TnI 96 -148 mutant complexes.
On the other hand, hydrophobic mutations to polar Gln in the regulatory domain of TnC F29W that may impede the forma-  side chain interaction. Consistent with this idea, analysis of the NMR structure of the Ca 2ϩ -TnC-TnI 115-131 complex (36) or a modeled structure of TnC-TnI (37) indicates that there are nine different hydrophobic residue side chains within the N-domain of TnC that come within 4 Å of six different hydrophobic side chains within TnI 115-131 . All of the high Ca 2ϩ affinity mutant hydrophobic residue side chains (Phe 22 , Val 45 , Met 46 , Leu 49 , and Met 82 ) come in close contact to TnI 115-131 and modestly decrease TnI 96 -148 binding ϳ1.7-3-fold. However, as mentioned above, neither Phe 26 nor Ile 62 come in close contact with TnI but interact with the Ca 2ϩ -binding loop ␤-sheet residues Ile 37 and Ile 62 (Fig. 8). Thus, Phe 26 and Ile 62 may help maintain the open state in such a way as to allow high affinity TnI 96 -148 binding to the regulatory domain of TnC F29W . Consistent with this idea, inhibiting the Ca 2ϩ -dependent opening of the regulatory domain of TnC by the introduction of a disulfide bond between the NAD and BC units decreased the affinity of TnI binding ϳ15-fold (38).
The ϳ12-fold increase in TnC F29W Ca 2ϩ affinity upon binding of TnI or TnI 96 -148 is primarily reflected by an ϳ30-fold slower rate of Ca 2ϩ dissociation from the TnC F29W -TnI 96 -148 complex (8). Hydrophobic residue substitutions to polar Gln in the regulatory domain of TnC F29W varied the Ca 2ϩ dissociation rates from the TnC F29W -TnI 96 -148 complex ϳ33-fold. This broad range in Ca 2ϩ dissociation rates is primarily reflected by the ability of some of the hydrophobic mutations to speed the rate of Ca 2ϩ dissociation, up to ϳ15-fold compared with theTnC F29W -TnI 96 -148 complex. Consistent with the inability of TnI 96 -148 to enhance the Ca 2ϩ sensitivity of the TnC F29W mutants with increased Ca 2ϩ affinity, the Ca 2ϩ dissociation rate from the mutant TnC F29W -TnI 96 -148 complexes could only be slowed ϳ2-fold.
The TnC F29W -TnI 96 -148 complex may be a good model system to study how different Tn complexes respond to changes in Ca 2ϩ concentration in muscle. Interestingly, the rate of Ca 2ϩ dissociation from the TnC F29W -TnI 96 -148 complex, and not isolated TnC F29W , is similar to the rate of fast twitch skeletal muscle relaxation (Fig. 9). Previous experiments with TnC mutants in which the Ca 2ϩ sensitivity of the regulatory domain in solution was increased or decreased demonstrated similar  (43) and was rendered using Rasmol (44). qualitative shifts in the force-pCa relationship upon reconstitution in skeletal muscle fibers (8,12,34,35,39,40). This may be the case only in those circumstances when the Ca 2ϩ sensitivity of the isolated TnC and the TnC-TnI complex are shifted in a similar direction. To test this hypothesis we reconstituted skinned rabbit psoas fibers with TnC F29W mutants that displayed the same qualitative changes in Ca 2ϩ sensitivities in the absence or presence of TnI 96 -148 (V45QTnC F29W and F78QTnC F29W ) and two that did not (M46QTnC F29W and M81QTnC F29W ). Comparison of the effects of these mutations on the force-pCa relationship suggests that the TnC F29W -TnI 96 -148 complex is a better predictor than isolated TnC F29W for how changes in Ca 2ϩ binding to TnC modulate the Ca 2ϩ sensitivity of force production. For instance, even though both Val 45 3 Gln and Met 46 3 Gln mutations increase the Ca 2ϩ affinity of isolated TnC F29W , only the Val 45 3 Gln mutation increased the Ca 2ϩ sensitivity of force development. These results are consistent with the fact that only Val 45 3 Gln, but not Met 46 3 Gln increases the Ca 2ϩ affinity of the TnC F29W -TnI 96 -148 complex.
Furthermore, the Ca 2ϩ sensitivity of force development generated with F78QTnC F29W was dramatically lower than that generated with M81QTnC F29W or TnC F29W , even though both Phe 78 3 Gln and Met 81 3 Gln mutations in isolated TnC F29W decreased the Ca 2ϩ sensitivity of the regulatory domain to a similar extent (6). These results are consistent with the fact that the Phe 78 3 Gln but not Met 81 3 Gln mutation leads to a dramatic decrease in Ca 2ϩ affinity of the TnC F29W -TnI 96 -148 complex. However, the Phe 78 3 Gln mutation appears to have a larger effect on the Ca 2ϩ affinity of the TnC F29W -TnI 96 -148 complex than on the Ca 2ϩ sensitivity of force development. Thus, the Ca 2ϩ binding properties of the TnC-TnI complex are not the only determinants of Ca 2ϩ sensitivity of force development. There is evidence that skeletal troponin T, tropomyosin, and actomyosin can modulate the Ca 2ϩ sensitivity of muscle mechanics either directly through TnC or through mechanisms yet to be explained (for review see Ref. 3).
A striking result observed with I62QTnC F29W was the dramatic reduction of force production generated by muscle fibers reconstituted with this mutant, even though the data shows it is able to bind to the thin filament. The near lack of force production cannot be explained by the low Ca 2ϩ sensitivity of the I62QTnC F29W -TnI 96 -148 complex because the F78QTnC F29W -TnI 96 -148 complex has ϳ 2-fold lower Ca 2ϩ sensitivity but is able to produce ϳ80% of maximal force. However, Ca 2ϩ -saturated I62QTnC F29W has an ϳ14-fold decreased affinity for TnI 96 -148 as compared with TnC F29W . The large decrease in TnI 96 -148 binding affinity for I62QTnC F29W is the likely reason why this mutant is unable to support force. It appears that Ca 2ϩ -I62QTnC F29W in the fiber might not effectively compete with actin binding to the regulatory domain of TnI, thus keeping the muscle fiber in a state of inactivation. Consistent with this interpretation, F26QTnC F29W , which had an ϳ10-fold lower affinity for TnI 96 -148 , also produced only ϳ13% of the maximal force upon reconstitution in the muscle fibers (data not shown). The exact opposite effect occurred when the regulatory regions of TnC and TnI were cross-linked, causing a regulated thin filament system to be permanently activated even in the absence of Ca 2ϩ (41). However, Ca 2ϩ -saturated I37QTnC F29W (a ␤-sheet mutant) bound TnI 96 -148 with only an ϳ2-fold lower affinity than TnC F29W but still only produced ϳ15% maximal force upon reconstitution in the muscle fibers (data not shown). Furthermore, Ca 2ϩ -saturated I73QTnC F29W , another ␤-sheet mutant, bound TnI 96 -148 with an affinity nearly identical to that of TnC F29W but only produced ϳ45% maximal force. Again, this points out that additional events besides Ca 2ϩ binding and subsequent TnI binding are involved in the signal pathway of force production. Consistent with this idea, a mutant TnC with a decreased skeletal troponin T affinity (but similar affinity for TnI) has been implicated in a loss of reconstituted thin filament ATPase activity (27). However, another mutant TnC with apparently normal Ca 2ϩ , TnI, and skeletal troponin T binding also displayed a diminished reconstituted thin filament ATPase activity through an unidentified mechanism apparently important for the Ca 2ϩ -dependent regulation of signal transduction (42).
In summary, we utilized TnC F29W to study Ca 2ϩ binding and exchange with a series of hydrophobic N-domain TnC mutants in the presence of TnI 96 -148 and intact TnI. The TnC F29W -TnI 96 -148 mutant complexes exhibited ϳ243-fold variation in their Ca 2ϩ binding affinities, ϳ38-fold variation in their Ca 2ϩ association rates, and ϳ33-fold variation in their Ca 2ϩ dissociation rates. The regulatory peptide of TnI, TnI 96 -148 , was an accurate mimic of intact TnI for measuring Ca 2ϩ dissociation rates from the TnC-TnI complexes. Furthermore, the effect of hydrophobic mutations on the Ca 2ϩ sensitivity of force development could be better predicted from the Ca 2ϩ affinities of the TnC F29W -TnI 96 -148 mutant complexes than from that of the isolated TnC F29W mutants. Interestingly, TnC F29W mutants with Ͼ10-fold lower TnI 96 -148 affinities in the presence of saturating Ca 2ϩ , compared with that of TnC F29W , were able to bind to the thin filaments but led to dramatic reduction of force recovery in reconstituted muscle fibers. Thus, not just Ca 2ϩ binding to TnC but the changes in the interactions with other regulatory proteins are critical in the pathway of signal transduction of force development. In conclusion, elucidating the determinants of Ca 2ϩ binding and exchange with TnC in the presence of its target protein TnI may provide a deeper understanding of how TnC and other closely related EF hand proteins respond to Ca 2ϩ and control signal transduction.