Introduction of negative charge mimicking protein kinase C phosphorylation of cardiac troponin I. Effects on cardiac troponin C.

Protein kinase C phosphorylation of cardiac troponin, the Ca(2+)-sensing switch in muscle contraction, is capable of modulating the response of cardiac muscle to a Ca(2+) ion concentration. The N-domain of cardiac troponin I contains two protein kinase C phosphorylation sites. Although the physiological consequences of phosphorylation at Ser(43)/Ser(45) are known, the molecular mechanisms responsible for these functional changes have yet to be established. In this work, NMR was used to identify conformational and dynamic changes in cardiac troponin C upon binding a phosphomimetic troponin I, having Ser(43)/Ser(45) mutated to Asp. Chemical shift perturbation mapping indicated that residues in helix G were most affected. Smaller chemical shift changes were observed in residues located in the Ca(2+)/Mg(2+)-binding loops. Amide hydrogen/deuterium exchange rates in the C-lobe of troponin C were compared in complexes containing either the wild-type or phosphomimetic N-domain of troponin I. In the presence of a phosphomimetic domain, exchange rates in helix G increased, whereas a decrease in exchange rates for residues mapping to Ca(2+)/Mg(2+)-binding loops III and IV was observed. Increased exchange rates are consistent with destabilization of the Thr(129)-Asp(132) helix capping box previously characterized in helix G. The perturbation of helix G and metal binding loops III and IV suggests that phosphorylation alters metal ion affinity and inter-subunit interactions. Our studies support a novel mechanism for protein kinase C signal transduction, emphasizing the importance of C-lobe Ca(2+)/Mg(2+)-dependent troponin interactions.

Protein kinase C phosphorylation of cardiac troponin, the Ca 2؉ -sensing switch in muscle contraction, is capable of modulating the response of cardiac muscle to a Ca 2؉ ion concentration. The N-domain of cardiac troponin I contains two protein kinase C phosphorylation sites. Although the physiological consequences of phosphorylation at Ser 43 /Ser 45 are known, the molecular mechanisms responsible for these functional changes have yet to be established. In this work, NMR was used to identify conformational and dynamic changes in cardiac troponin C upon binding a phosphomimetic troponin I, having Ser 43 /Ser 45 mutated to Asp. Chemical shift perturbation mapping indicated that residues in helix G were most affected. Smaller chemical shift changes were observed in residues located in the Ca 2؉ /Mg 2؉ -binding loops. Amide hydrogen/deuterium exchange rates in the C-lobe of troponin C were compared in complexes containing either the wild-type or phosphomimetic N-domain of troponin I. In the presence of a phosphomimetic domain, exchange rates in helix G increased, whereas a decrease in exchange rates for residues mapping to Ca 2؉ /Mg 2؉ -binding loops III and IV was observed. Increased exchange rates are consistent with destabilization of the Thr 129 -Asp 132 helix capping box previously characterized in helix G. The perturbation of helix G and metal binding loops III and IV suggests that phosphorylation alters metal ion affinity and inter-subunit interactions. Our studies support a novel mechanism for protein kinase C signal transduction, emphasizing the importance of C-lobe Ca 2؉ /Mg 2؉ -dependent troponin interactions.
Troponin and tropomyosin form the Ca 2ϩ -sensitive switch that regulates striated muscle contraction. Troponin is a ternary assembly of proteins composed of the Ca 2ϩ -binding sub-unit troponin C (TnC), 1 the inhibitory subunit troponin I (TnI), and the tropomyosin-binding protein troponin T that anchors troponin to the thin filament. Troponin C, a member of the EF-hand family of Ca 2ϩ -binding proteins, contains two globular domains connected by a linker. Each domain of cTnC contains two EF-hand or Ca 2ϩ -binding motifs. The N-lobe contains two lower affinity Ca 2ϩ -binding motifs, sites I and II, that control muscle contraction. Site I is naturally inactive in the cardiac isoform because of several amino acid substitutions and an amino acid insertion (1). Thus, Ca 2ϩ binding at site II in cTnC regulates muscle contraction. The C-lobe contains two high affinity Ca 2ϩ -binding sites, III and IV, which also bind Mg 2ϩ with lower affinity. Interactions between the C-lobe of cTnC and the N-domain of cTnI form the Ca 2ϩ /Mg 2ϩ -dependent cTnC/cTnI interaction site. In addition, the C-lobe of cTnC also interacts tightly with the C terminus of cardiac troponin T (2). These interactions form the core of the troponin complex, tethering all three subunits throughout the contraction cycle.
A variety of effectors can modulate the frequency and intensity of myocardial contraction by charge modification upon the phosphorylation of cTnI and cardiac troponin T. In the heart, phosphorylation of cTnI appears to be of particular importance in cardiac hypertrophy and failure (3)(4)(5). Cardiac TnI can be phosphorylated by protein kinase A at Ser 23 and Ser 24 (6). ␤-Adrenergic stimulation leads to protein kinase A phosphorylation at Ser 23 and Ser 24 of cTnI, enhancing relaxation by decreasing the Ca 2ϩ affinity at site II (7,8). Cardiac TnI can also be phosphorylated by PKC at Ser 43 , Ser 45 , and Thr 144 (9). Phosphorylation at Ser 43 and Ser 45 is known to decrease maximal actomyosin MgATPase activity, Ca 2ϩ sensitivity, and cross-bridge binding to the thin filament (4, 9 -11). These biochemical changes lead to maladaptive growth and diminished contractility, culminating in end-stage heart failure (5). Although it is clear that PKC phosphorylation of cardiac troponin modulates contraction in response to hemodynamic stressors such as hypertension and myocardial infarction, the molecular mechanisms of this modulation remain unknown.
The effects of PKC phosphorylation at Ser 43 and Ser 45 of cTnI can be mimicked, both in vitro and in vivo, by the introduction of negative charge at positions 43/45 (3,4,11). Serine residues 43/45 of cTnI are at the N terminus of an amphiphilic ␣-helix that binds in a hydrophobic cleft in the C-lobe of cTnC (2). Stability of the hydrophobic core in CcTnC is largely governed by metal binding at site III (12). The exchange of Ca 2ϩ for Mg 2ϩ in Ca 2ϩ /Mg 2ϩ -binding sites III and IV resulted in a partial closure of the hydrophobic binding cleft around site IV, allowing the possibility that Ca 2ϩ /Mg 2ϩ exchange can modulate contraction via Ca 2ϩ /Mg 2ϩ -dependent cTnC/cTnI interactions (13,14). However, even absent excess Ca 2ϩ , Mg 2ϩ did not readily displace Ca 2ϩ in the C-lobe of cTnC bound to NcTnI (14).
To define the structural consequences resulting from cTnI phosphorylation at Ser 43 /Ser 45 by PKC, we have utilized solution NMR to identify residues in cTnC that are important in transmission of the phosphorylation signal. Backbone resonances in Ca 2ϩ -loaded [ 13 C, 15 N]CcTnC bound to NcTnI(S43D/ S45D) were assigned and used for secondary structure determination. The phosphorylation mimetics did not significantly alter the secondary structural elements of the paired Ca 2ϩbinding motifs in CcTnC. Residues in CcTnC important for transmission of the phosphorylation signal were identified using chemical shift perturbation mapping and amide proton exchange. The phosphorylation mimetics induced localized conformational/dynamic perturbations in the N terminus of helix G and Ca 2ϩ /Mg 2ϩ -binding loops III and IV. Specifically, the presence of a negative charge at Ser 43 /Ser 45 of cTnI was found to destabilize the Thr 129 -Asp 132 N-cap in helix G of CcTnC. Perturbation of N-cap interactions in helix G have been shown previously to decrease metal ion affinities at site III (15). These findings support a mechanism for PKC modulation of cardiac contractility wherein the introduction of negative charge at Ser 43 /Ser 45 of cTnI results in altered C-lobe metal ion affinities and perturbation of Ca 2ϩ /Mg 2ϩ -dependent protein-protein interactions. Presumably, these changes are transmitted to other regulatory and switch regions within cardiac troponin. These changes establish a role for the Ca 2ϩ /Mg 2ϩ -dependent cTnC/ cTnI interaction in transmitting the phosphorylation signal and modulating Ca 2ϩ sensitivity.

Recombinant Protein Expression and
Purification-Isotopically enriched and unlabeled recombinant proteins were expressed, purified, and quantified as described (16,17).
NMR Methodology-Experiments were collected at 40°C using 400, 600, and 800 MHz Varian Inova spectrometers equipped with pulsefield gradient units and triple resonance probes. Acquisition parameters for heteronuclear multidimensional NMR experiments and chemical shift referencing details can be found in Gasmi-Seabrook et al. (17). Inter-residue NOEs obtained from 15 N-editied nuclear Overhauser effect spectroscopy-HSQC experiments at mixing times of 70 and 150 ms were used to confirm consecutive assignments. 1 H-15 N transverse relaxation-optimized spectroscopy spectra of the intact binary complex [ 2 H, 15 N]cTnC-cTnI(S43D/S45D) were acquired with 2048 points in the direct dimension, 64 points in the indirect dimension, and 320 scans per increment. Felix 2000 was employed to process and analyze NMR data.
Amide 1 H and 15 N chemical shift differences were obtained by subtracting the respective chemical shifts for each residue in CcTnC-NcTnI from the amide 1 H and 15 N chemical shifts for the same residue in CcTnC-NcTnI(S43D/S45D). The combined 1 H and 15 15 N HSQC spectra were acquired with 1280 and 64 points in the direct and indirect dimensions, respectively, with eight scans per increment. Spectra were processed and analyzed using Felix. The volume of each amide cross-peak was normalized, and the ratio was fitted to a single exponential function, . Differences in CcTnC amide proton exchange rates, ⌬k ex , were obtained by subtracting exchange rates for each residue in CcTnC-NcTnI from the same residue in CcTnC-NcTnI(S43D/S45D).

RESULTS AND DISCUSSION
NMR Signal Assignment and Secondary Structure-Changing a phosphorylation site to a negatively charged residue such as Asp or Glu can often be used to mimic phosphorylation and facilitate biophysical studies. Incorporation of a negative charge at Ser 43 /Ser 45 by mutation was shown to provide functional mimetics for PKC phosphorylation of cTnI (4,18). A comparison between cTnI and cTnI(S43D/S45D) bound to Ca 2ϩ -loaded [ 2 H, 15 N]cTnC was made. Chemical shift perturbation mapping was used to monitor conformational changes in cTnC induced by the introduction of a negative charge at Ser 43 / Ser 45 of cTnI in the intact binary complex. Amide proton chemical shift differences between Ca 2ϩ -loaded [ 2 H, 15 N]cTnC-cTnI(S43D/S45D) and [ 2 H, 15 N]cTnC-cTnI are shown in Fig. 1A. The majority of chemical shift perturbations induced by the mutation of Ser 43 /Ser 45 to Asp were localized to the C-lobe of cTnC, with residues in helix G experiencing the largest chemical shift changes (Fig. 1A). Small amide proton chemical shift perturbations were observed for Glu 66 , Gly 68 , and Ser 69 in the N-lobe of cTnC (Fig. 1A). These residues are located in the regulatory Ca 2ϩ -binding site (site II). Titration of cTnC with a cTnI peptide phosphorylated at Ser 43 /Ser 45 was also found to induce small amide proton chemical shift perturbations in Nlobe residues corresponding to Gly 42 , Val 72 , and Val 79 (19). Chemical shift perturbations may reflect charge-induced changes in local electrostatic interactions or changes in protein structure. It is unlikely that N-lobe chemical shift perturbations result from local electrostatic interactions, because these residues are Ͼ15 Å from the sites of cTnI phosphorylation in the core cardiac troponin structure (2).
To examine in more detail conformational perturbations induced in the C-lobe of cTnC by PKC phosphorylation at Ser 43 / Ser 45 of cTnI, we have studied a model phosphomimetic complex, CcTnC-NcTnI(S43D/S45D). NMR resonance assignments and solution structures for both Mg 2ϩ -and Ca 2ϩ -loaded CcTnC in the CcTnC-NcTnI complex are available (13). Comparison of C-lobe chemical shift perturbations in cTnC-cTnI(S43D/S45D) and CcTnC-NcTnI(S43D/S45D) demonstrate that CcTnC-NcTnI(S43D/S45D) provides a suitable model for PKC phosphorylation-induced structural changes in the Ca 2ϩ /Mg 2ϩ -dependent cTnC-cTnI interaction site (Fig. 1). This finding is consistent with the core cardiac troponin x-ray structure showing that NcTnI primarily makes contacts with the C-lobe of cTnC (2). Phosphorylation sites Ser 43 /Ser 45 in cTnI are located at the N terminus of helix H1 in cTnI, corresponding to residues 43-79 (2). Residues 43-65 of cTnI bind to the hydrophobic cleft in CcTnC (2).
Two dimensional 1 H- 15  Backbone chemical shift assignments for Ca 2ϩ -loaded [ 13 C, 15 N]CcTnC bound to NcTnI(S43D/S45D) were obtained using standard triple resonance assignment strategies (13,17). The resonance assignment strategy relied primarily on (H ␤ )C ␤ C ␣ (CO)NNH, HNC ␣ C ␤ , HNC ␣ , and HNCO triple resonance experiments. Assignments were obtained for 75 of the 81 CcTnC amino acid residues. Backbone chemical shifts for the C ␣ , C ␤ , CЈ, and H ␣ resonances were used to determine chemical shift index values for each residue in Ca 2ϩ -loaded CcTnC bound to NcTnI(S43D/S45D) (21). The chemical shift indexdetermined secondary structure showed a characteristic paired EF-hand motif with four helices spanning residues 93-103 (E), 114 -123 (F), 130 -139 (G), and 150 -158 (H), as well as two short ␤-strands extending from residues 111-113 and 147-149 (Fig. 1B). The overall secondary structure is analogous to that previously determined for Ca 2ϩ -and Mg 2ϩ -loaded CcTnC bound to NcTnI (13). The similarity in secondary structure for CcTnC bound to NcTnI and NcTnI(S43D/S45D) justified the use of available C-lobe cTnC structures for interpreting changes in chemical shifts and H/D exchange rates upon the introduction of a negative charge at the two PKC phosphorylation sites (2,13).
Chemical Shift Mapping-Chemical shifts are sensitive to the local environment and can be used to monitor subtle structural changes. We have shown previously that amide chemical shifts for cTnC are extremely sensitive to the small conformational changes that occur upon cAMP-dependent protein kinase A phosphorylation of cTnI (22,23). Chemical shifts for the 1 H N , 15  used for a residue-by-residue comparison with the chemical shifts obtained in CcTnC-NcTnI(S43D/S45D). Perturbations in the combined 1 H N / 15 N H and 13 C ␣ / 13 C ␤ chemical shifts are plotted on the structure of Ca 2ϩ -loaded CcTnC bound to NcTnI (Fig. 2).
The largest 1 H N / 15 N H chemical shift perturbations were observed in the N terminus of helix G, Ile 128 and Thr 129 , and in helix E, Glu 95 (Fig. 1B). These residues cluster around the N-cTnI-binding site (2). Threonine 129 is the N-cap residue in the N-terminal helix capping box of helix G in both Ca 2ϩ -and Mg 2ϩ -loaded CcTnC-NcTnI complexes (13,17). In both complexes, the side-chain hydroxyl of Thr 129 hydrogen bonds to the amide of Asp 132 , and the Asp side-chain, in turn, forms a hydrogen bond with the amide of Thr 129 (13, 17). The upfield FIG. 2. Residue-specific chemical shift differences between CcTnC bound to NcTnI and NcTnI(S43D/S45D) mapped onto the Ca 2؉ -loaded CcTnC structure. A, changes in combined amide 1 H/ 15 N chemical shifts mapped onto the ribbon structure of Ca 2ϩ -loaded CcTnC bound to NcTnI (Protein Data Bank accession code 1SCV). B, changes in 13 C␣ and 13 C␤ chemical shifts mapped onto the ribbon structure of Ca 2ϩ -loaded CcTnC bound to NcTnI. C, surface representation of the Ca 2ϩ -loaded CcTnC structure with consensus residues highlighted in red. Highlighted residues show significant chemical shift perturbations in two of the three ( 1 H/ 15 N, 13 C␣, and 13 C␤) difference maps or have neighboring residues showing significant chemical shift differences in at least one of the three difference maps. amide proton chemical shift for Thr 129 in the phosphomimetic complex is consistent with destabilization of the hydrogen bond between the amide of Thr 129 (N-cap residue) and the carboxylate side-chain of Asp 132 (N3 residue). The 3 J NH-C␣ coupling constant for Thr 129 decreased slightly from 9 Hz (17) to 8.3 Hz in the phosphomimetic complex, consistent with a decrease in the average backbone torsion angle. Changes in the amide proton chemical shift and the 3 J NH-C␣ coupling constant for Thr 129 are consistent with destabilization of the N-terminal helix G capping box. In addition, upfield amide nitrogen chemical shifts are also observed for Ile 128 and Thr 129 in the phosphomimetic complex (data not shown). Hydrophobic interactions between the NЈ residue (Ile 128 ) and the N4 residue (Ile 133 ) stabilize the N-cap box (24). Such stabilizing interactions would be expected to deshield the amide nitrogen resonances of Ile 128 and Thr 129 . However, destabilization of the hydrogenbonding network in the Thr 129 -Asp 132 N-cap box would reverse this effect, shifting the amide nitrogen resonances upfield as observed in the phosphomimetic complex.
Chemical shift perturbations were also observed in 13 (Fig. 2B). Hydrophobic and polar residues identified by chemical shift indexing form contiguous surfaces that define the binding interface between CcTnC and NcTnI within the Ca 2ϩ /Mg 2ϩ -dependent cTnC-NcTnI interaction site (Fig. 2C). Residues 43-65 of cTnI form an amphiphilic ␣-helix that binds to the C-lobe hydrophobic crevice via multiple polar and Van der Waals interactions (2), similar to the interactions observed in the skeletal sTnC-sTnI (1-47) complex (25).
Hydrogen/Deuterium Exchange-Changes in local dynamic behavior have also been utilized to identify residues that undergo conformational change accompanying protein binding. Hydrogen/deuterium exchange allows characterization of changes in global thermodynamic stability and local conformational motion with exchange time constants on the order of minutes to days. Whereas chemical shifts report on the magnetic environment of nuclei, H/D exchange kinetics provides information about backbone dynamics, conformation, and proton-solvent interactions. Comparison of H/D exchange rates provides a mechanism to assess the effects of the PKC phosphorylation mimetics on conformational fluctuations in CcTnC. To this end, amide proton H/D exchange rates in [ 15 N]CcTnC bound to either NcTnI or NcTnI(S43D/S45D) were measured at 25°C. At temperatures Ͼ25°C, many of the exchange rates were too fast to reliably measure, whereas temperatures Ͻ25°C resulted in the broadening of 1 H-15 N correlations. Hydrogen/deuterium exchange rates could be monitored for 59 of the 75 assigned 1 H-15 N correlations in the HSQC spectrum of CcTnC. The remaining resides were excluded from analysis because of weak signal intensity and/or resonance overlap. The amide proton exchange kinetics could be classified into five categories, namely rapidly exchanging (within the first 20 min), fast (k ex Ͼ 8.5 ϫ 10 Ϫ2 min Ϫ1 ), moderately fast (8.5 ϫ 10 Ϫ3 min Ϫ1 Ͻ k ex Ͻ 8.5 ϫ 10 Ϫ2 min Ϫ1 ), slow (2.5 ϫ 10 Ϫ3 min Ϫ1 Ͻ k ex Ͻ 8.5 ϫ 10 Ϫ3 min Ϫ1 ), and very slow (k ex Ͻ 2.5 ϫ 10 Ϫ3 min Ϫ1 ) (Table I). Rapidly exchanging amide correlations, disappearing within the first 20 min of exchange in both complexes, were assigned to residues Asp 87 , Asp 88 , Ser 89 , Lys 90 , Gly 91 , Lys 92 , Thr 93 , Glu 94 , Ser 98 , Asp 99 , Arg 102 , Met 120 , Thr 124 , Gly 125 , Glu 126 , Thr 127 , Glu 130 , Asn 144 , Glu 152 , and Gly 159 . Most of these correspond to residues that are located in unstructured or mobile loop regions (13). The time course of H/D exchange for 28 residues could be followed in both complexes and used to calculate exchange rates (Table I). The introduction of a negative charge mimicking PKC phosphorylation in NcTnI was found to both increase and decrease the amide proton H/D exchange rates of selective residues in CcTnC (Table I and Fig.  3). A comparison of 1 H-15 N correlation spectra of CcTnC bound to NcTnI(S43D/S45D) and NcTnI after 330 min of exchange into 2 H 2 O shows that the amide proton of Ile 133 exchanges faster in the phosphomimetic complex (Fig. 3). Isoleucine 133 is located in helix G immediately following the Thr 129 -Asp 132 N-cap box. The H/D exchange rates for Thr 129 , Asp 132 , Ile 133 , and Glu 135 , located in helix G of CcTnC, were also increased in the phosphomimetic complex (Table I and Fig. 4). These residues are involved both in stabilizing N-terminal helix capping interactions and in the (i) to (i ϩ 3) hydrogen bonds typical of ␣-helices. In addition, Asp 132 and Glu 135 participate in interresidue polar interactions with Arg 45 and Lys 46 of cTnI (2). Thus, the introduction of proximal negative charge in cTnI weakens the substantial hydrogen-bonding lattice in helix G, altering Ca 2ϩ /Mg 2ϩ -dependent interactions between cTnI and the C-lobe of cTnC.
In contrast, comparison of the Asn 107 and Asn 143 1 H-15 N cross-peaks in correlation spectra after 330 min of exchange in 2 H 2 O show that these amide protons exchange slower in the PKC phosphomimetic complex (Fig. 3). Both Asn 107 and Asn 143  (Table I and Fig. 4). The side chain of Asp 113 , located at the ninth position in Ca 2ϩ /Mg 2ϩ binding loop III, directly coordinates the bound metal ion. Amide protons of Gly 110 and Gly 146 , located at the sixth position in Ca 2ϩ /Mg 2ϩ binding loop III and IV, respectively, hydrogen bond to the carboxylate side chain of the conserved Asp residues at position 1 (12). Decreased H/D exchange for residues within Ca 2ϩ /Mg 2ϩ binding loops III and IV is consistent with the stabilization of intra-loop hydrogen bonding interactions and the compaction of the metal binding loops in the phosphomimetic complex. These changes suggest that phosphorylation of NcTnI alters the conformation of Ca 2ϩ /Mg 2ϩ -binding sites III and IV, possibly resulting in altered metal ion affinity and Ca 2ϩ /Mg 2ϩ exchange. Summary-The structural consequences of PKC phosphorylation at Ser 43 /Ser 45 of cTnI on cTnC have been examined using chemical shift mapping and H/D exchange. To facilitate struc-tural studies, PKC phosphorylation mimetics of cTnI having Ser 43 /Ser 45 mutated to Asp were utilized. The overall picture obtained from chemical shift mapping shows structural perturbations predominately localized to the C-lobe of cTnC, with smaller N-lobe perturbations around Ca 2ϩ -binding site II (Fig. 1). Distances between N-lobe Ca 2ϩ -binding sites and Ser 43 /Ser 45 of cTnI in the core cardiac troponin structure (2) suggest that chemical shift perturbations in Ca 2ϩ -binding site II are the result of long range effects as opposed to direct binding interactions.
Recently, measurement of fluorescence resonance energy transfer distance distributions from a single cTnC donor/acceptor pair in cTnC-cTnI and cTnC-cTnI(S43E/S45E, T144E) showed that the introduction of a negative charge in cTnI alters N-lobe conformational equilibria (3). Previously we showed that amide chemical shifts could be utilized to monitor conformation equilibria between open and closed N-lobe substates (16). We see no evidence that the introduction of a negative charge at Ser 43 /Ser 45 of cTnI significantly alters conformational equilibria in the N-lobe of cTnC (Fig. 1). It is likely that the N-lobe conformational change detected by fluorescence resonance energy transfer analysis (3) results from the additional negative charge at Thr 144 of cTnI. This probability is consistent with N-lobe conformational changes detected by NMR chemical shift analysis of cTnC binding to a cTnI regulatory peptide phosphorylated at Thr 144 (19). Taken together, these results help clarify the molecular consequences of PKC phosphorylation at Ser 43 , Ser 45 , and Thr 144 in cTnI on Ca 2ϩ -loaded cTnC. Phosphorylation of Ser 43 /Ser 45 in cTnI induces conformational perturbations in the C-lobe of cTnC containing Ca 2ϩ /Mg 2ϩbinding sites III and IV, whereas phosphorylation at Thr 144 directly alters conformational equilibria in the N-lobe of cTnC containing Ca 2ϩ -binding site II.
A combination of chemical shift mapping and H/D exchange was used to examine conformational perturbations in the Ca 2ϩ / Mg 2ϩ -dependent cTnC-cTnI interaction site induced by the introduction of a negative charge at Ser 43 /Ser 45 of cTnI. Chemical shift mapping identified structural perturbations in CcTnC residues lining the NcTnI hydrophobic binding cleft (Fig. 2). Amide hydrogen/deuterium exchange results are consistent with the destabilization of N-terminal helix G-capping interactions and the conformational perturbation of Ca 2ϩ /Mg 2ϩ -binding sites III and IV (Fig. 4). Whereas metal binding at site III is primarily responsible for stabilizing the hydrophobic core (12), helix-capping interactions are known to increase domain stability and accelerate folding (28). Destabilization of N-cap interactions in helices C and G by mutation have been shown to decrease Ca 2ϩ -binding affinity in TnC (15,27). Thus, weakening of the hydrogen-bonding lattice in helix G would be expected to alter metal ion affinity and Ca 2ϩ /Mg 2ϩ exchange at site IV. Microcalorimetry data suggest that Ca 2ϩ /Mg 2ϩ -dependent protein-protein interactions are 8-fold stronger in the presence of Ca 2ϩ than in the presence of Mg 2ϩ (29). Substitution of Mg 2ϩ for Ca 2ϩ in CcTnC bound to NcTnI is characterized by condensation of the C-terminal portion of the metal binding loops and partial closure of the cTnI hydrophobic binding cleft around site IV (13). The close association between helix stability and metal ion affinity provides an attractive model for modulating Ca 2ϩ /Mg 2ϩ exchange and Ca 2ϩ /Mg 2ϩ -dependent protein-protein interactions by PKC phosphorylation.
A negative charge introduced at Ser 43 /Ser 45 of cTnI, either by mutation or phosphorylation, is expected to stabilize the NcTnI helix through favorable electrostatic interactions between neighboring polar side chains and the helix backbone (30). Residues 43-65 in cTnI form a ␣-helix that binds to the C-lobe of cTnC through multiple polar and van der Waals interactions (2). Computational analyses of cTnI having Glu substituted at Ser 43 /Ser 45 suggest that the incorporation of a negative charge extends the N terminus of the cTnI helix to residue 40 (4). Solution analysis of the secondary structure of NcTnI bound to CcTnC, based on experimentally determined chemical shift index values, suggests that the NcTnI helix begins at residue 46. The mutation of Ser 43 /Ser 45 to Asp stabilizes and extends the N terminus by three residues, from residue 46 in NcTnI to residue 43 in NcTnI(S43D/S45D). 2 Structural changes in CcTnC and NcTnI induced by the phosphorylation of Ser 43 / Ser 45 support a mechanism for PKC modulation of cardiac contractility wherein the introduction of a negative charge results in altered C-lobe metal ion affinities and perturbation of Ca 2ϩ /Mg 2ϩ -dependent protein-protein interactions. Structural changes in NcTnI and CcTnC, forming the Ca 2ϩ /Mg 2ϩdependent protein-protein interaction site, can then be transmitted to other regulatory or switch regions in troponin. These studies emphasize the importance of Ca 2ϩ /Mg 2ϩ -dependent cTnC-cTnI protein-protein interactions in the modulation of cardiac contractility.