THE JOURNAL. OF BIOL~CICAL 0 1994 by The American Society for

CHEMISTRY ' Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 38, Issue of September 23, pp. 23731-23735, 1994 Printed in U.S.A.

N"R Studies Delineating Spatial Relationships within the Cardiac Troponin I-Troponin C Complex*

(Received for publication, May 11, 1994, and in revised form, July 22, 1994)

George A. KrudyS, Quinn Kleerekoperl, Xiaodu GuoO, Jack W. HowarthS, R. John SolaroPn, and Paul R. RosevearSll From the Wepartment of Biochemistry and Molecular Biology, University of Texas Medical School, Houston, Texas 77225 and the $Department of Physiology and Biophysics, College of Medicine, University of Illinois, Chicago, Illinois 60612-7342

NMR spectroscopy and selective isotope labeling of both recombinant cardiac troponin C (cTnC3) and a trun- cated cardiac troponin I (cTnI/NH,) lacking the N-termi- nal 32-amino acid cardiac-specific sequence have been used to probe protein-protein interactions central to muscle contraction. Using [methyl-'3C]Met-labeled cTnC3, all 10 cTnC Met residues of Ca2+-saturated cTnC3 could be resolved in the two-dimensional heteronuclear single- and multiple-quantum coherence spectrum of the cTnI-cTnC complex. Based on the known Met assign- ments in cTnC3, the largest chemical shift changes were observed for Mets1, MePo, Met'37, and Met"?. Methionines 120,137, and 157 are all located in the C-terminal domain of cTnC. Methionine 81 is located at the N terminus of the central helix. Minimal chemical shift changes were ob- served for Met4s, Met4?, and Metlo3 of cTnC3 in the cTnI.cTnC complex. All 6 Met residues in [methyl- 'SC]Met-labeled cTnUNH, could be resolved in the cTnI-cTnC complex, suggesting that both cTnI and cTnC form a stable homogeneous binary complex under the conditions of the NMR experiment. In the absence of added protease inhibitors in the cTnI-cTnC complex, cTnI/NH, was found to undergo selective proteolysis to yield a 5.5-kDa N-terminal fragment corresponding to residues 33-80. Judging from the NMR spectra of [methyl- "CIMet-labeled cTnC3, cTnI433-80) was sufficient for in- teraction with the C-terminal domain of cTnC in a man- ner identical to that observed for native cTnI/NH,. However, in the presence of the proteolytic fragment cTnI-(33-80), the chemical shift of Mets1 was not per- turbed from its position in free cTnC3. Thus, residues located C-terminal to Arc in cTnI appear to be respon- sible for interaction with the N-terminal half of cTnC. Taken together, these results provide strong evidence for an antiparallel arrangement for the two proteins in the troponin complex such that the N-terminal portion of cTnI interacts with the C-terminal domain of cTnC. This interaction likely plays a role in maintaining the stability of the TnI-TnC complex.

Contraction in striated muscle is regulated through Ca2+- dependent protein-protein interactions that modulate the ac- tivity of actomyosin ATPase (1, 2). Key to transmission of the Ca2+ signal that controls actomyosin ATPase is the interaction

* This work was supported by National Institutes of Health Grants HL22231 and HL49934 (to R. J. S.), by American Heart Association Grant AHA92015340 (to P. R. R.), by a fellowship from the American Heart Association of Metropolitan Chicago (to X. G.), and by National Institutes of Health Grant HL45724 (to Dr. John Putkey). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

ll To whom correspondence should be addressed.

between troponin C (TnC)' and troponin I (TnI). TnC and TnI, along with TnT, are members of the troponin complex. The troponin complex and tropomyosin are located on the thin fila- ment. Each Tn subunit has been assigned or implicated in a distinct function. TnC is the Ca2+-binding subunit of the Tn complex; TnT makes primary protein-protein contacts with tro- pomyosin; and TnI, the so-called inhibitory subunit, partici- pates in the major protein-protein interaction with TnC. Upon binding Ca2+ to TnC, protein-protein interactions in the com- plex are modified.

Troponin C is a member of a family of proteins having at least one helix-loop-helix or EF-hand Ca2+-binding site (3). The two isoforms of TnC found in vertebrate muscle are fast skeletal (sTnC) and cardiac or slow skeletal (cTnC). Both proteins con- tain four EF-hand Ca2+-binding sites (I-IV). The cardiac iso- form differs significantly from the skeletal isoform in that the first Ca2+-binding site, site I, is inactivated by several critical amino acid substitutions (4, 5).

From the crystallographic structures of sTnC and calmodu- lin, cTnC is predicted to be a dumbbell-shaped molecule with Ca2+-binding sites I/II and III/N separated on opposite ends of a long central helix (6-8). Site 11, located in the N-terminal domain, has been demonstrated to be the Ca2+-binding site responsible for regulating cardiac muscle contraction (9). The high affinity Ca2+/Mg2"dependent sites I11 and IV, in the C- terminal domain of cTnC, are thought to be largely responsible for maintaining the stability of the interaction between cTnI and cTnC within the troponin complex (10).

Troponin I, the inhibitory component of the troponin system, demonstrates Ca2+-dependent protein-protein interactions. The relief of inhibition of actomyosin ATPase upon Ca2+ binding a t site I1 of cTnC and the subsequent interactions with TnI are the key steps in the initiation of muscle contraction. Evidence to date strongly suggests multiple interactions between TnI and TnC involving segments located some distance from each other in the primary sequences (11, 12).

At least three interaction sites between TnI and TnC have been postulated: 1) a metal-independent structural site, 2) a Ca2+/M$+-dependent site, and 3) a Ca2+-specific binding site (1, 12-14). In the absence of metals, an association constant of 6.67 x lo5 M - ~ was measured for the TnI-TnC interaction (15). This is in agreement with biochemical studies demonstrating that sTnI binds to a sTnC affinity column even in the presence of 2 mM EDTA and 1 M NaCl(12) and with the observation that the TnI.TnC complex is stable in 8 M urea (16, 17). The binding of

I The abbreviations used are: Tn, troponin; sTn, skeletal troponin; cTn, cardiac troponin; cTnC3, recombinant cardiac des-Met'- [Ala21troponin C; cTnI/NH,, recombinant cardiac des-(Met1-Thr3*)-tro- ponin I; PMSF, phenylmethylsulfonyl fluoride; HSMQC, two-dimen- sional heteronuclear single- and multiple-quantum coherence.

23731

23732 Spatial Relationships in the TnI-TnC Complex

Mg2' ion to the Ca2+/Mg2+-dependent sites has been shown to increase the affinity of TnI for TnC by -100-fold (15). Using an N-terminal deletion mutant, sTnI,,,, which lacks residues 1-57, it could be demonstrated that the Ca2+/Mg2"dependent site interaction was lost (12). The C-terminal portion of sTnI contains the well known inhibitory binding region and is thought to contain the Ca2+-specific site-dependent interaction. Binding studies utilizing truncated sTnI proteins have also been used to propose that the N-terminal region of sTnI inter- acts with the C-terminal domain of sTnC and that sTnI-(103- 182) binds in a Ca2+-dependent manner to the N-terminal do- main of sTnC (16). Thus, deletion mutants, proteolytic fragments, cross-linking, and fluorescence approaches suggest that the N-terminal domain of sTnI is responsible for a Ca2+/ Me-dependent site interaction and that the C-terminal region of sTnI is responsible for the Ca2+-specific site-dependent interaction.

To facilitate structure/function studies of the cTnI.cTnC com- plex, a rat cardiac troponin I has recently been cloned and sequenced (17-19). Sequence comparisons between species in- dicate that cTnI is a highly conserved protein, with the cardiac isoforms having a variable length cardiac-specific N-terminal sequence (19). A truncated form of rat cTnI (cTnI/NH,) lacking the N-terminal 32-amino acid cardiac-specific sequence has been successfully expressed in bacteria, and quantities suffi- cient for biophysical studies have been readily obtained from 1 liter of bacterial growth (18-20). Purified cTnUNH, has been exchanged into myofibrils and shown to restore regulation fully

To initiate structural studies on the cTnI.cTnC complex, we have utilized the 10 Met residues in [rnethyZ-*3ClMet-labeled cTnC3 as selective markers aimed a t monitoring conforma- tional changes in cTnC upon cTnI binding. This study not only demonstrates the feasibility of obtaining biologically relevant structural data using this approach, but also provides struc- tural data demonstrating that the C terminus of cTnC interacts with the N terminus of cTnI.

(18-20).

EXPERIMENTAL PROCEDURES Materials-~-[rnethyl-'~C]Methionine, deuterium oxide, and Tris-dl,

were obtained from Cambridge Isotope Laboratories (Andover, MA). All other chemicals were of the highest purity available commercially.

Labeled Recombinant Proteins"cTnC3 was overexpressed in Escherichia coli JM109 (21). [methyl-'3ClMet-labeled cTnC3 was ob- tained by growing the bacteria in an enriched defined medium contain- ing 0.5 g of alanine, 0.4 g of arginine, 0.4 g of aspartic acid, 0.4 g of asparagine, 0.05 g of cystine, 0.4 g of glutamine, 0.65 g of glutamic acid, 0.55 g of glycine, 0.1 g of histidine, 0.23 g of isoleucine, 0.23 g of leucine, 0.42 g of lysine HCl, 25 mg of [methyl-'3Clmethionine, 0.13 g of phenyl- alanine, 0.1 g of proline, 0.21 g of serine, 0.23 g of threonine, 0.17 g of tyrosine, 0.23 g of valine, 0.5 g of adenine, 0.65 g of guanosine, 0.2 g of thymine, 0.5 g of uracil, 0.2 g of cytosine, 1.5 g of sodium acetate, 1.5 g of succinic acid, 0.5 g of ammonium chloride, 0.85 g of sodium hydroxide, 10.5 g of GHPO,, 50 ml of 20% glucose, 4 ml of 1 M MgSO,, 1 ml of 0.01 M FeCI,, 125 p1 of 1 M CaCl,, 50 rng of L-tryptophan, 50 mg of thiamine, 1 mg of biotin, 50 mg of niacin, 30 mg of kanamycin, and 50 mgfliter ampicillin. Mixed metals were added as previously described (22). Re- combinant cTnC was purified to apparent homogeneity as described previously by Putkey et al. (9).

Recombinant cTnC samples for NMR experiments were extensively dialyzed against 50 mM (NH,),C03; lyophilized; and dissolved in NMR buffer containing 20 ELM Tris-d,,, 200 mM KCI, and 0.1 mM PMSF in 100% ,H,O a t pH 7.5. For preparation of apo-cTnC3, samples were dialyzed three times against 50 mM (NH,),CO, containing 5 mM EGTA and 10 mM EDTA. desalted on a Sephadex G-25 column, and lyophilized.

Selective labeline of cTnI/NH, was achieved by sowing BL21(DE3) cells, harboring thepET-9a vector containing rat &&NH,~in the above enriched defined medium supplemented with [methyl-'3C]Met and con- taining 25 mgfliter kanamycin and 25 mgfliter chloramphenicol. Rou- tinely, 2 liters of enriched defined medium were inoculated with cells

from a 200-ml overnight growth in L broth, and 0.4 mM isopropyl-l- thio-p-D-galactopyranoside was added at an absorbance of 0.7. Cells were harvested 4 h after induction, collected by centrifugation, and stored at -70 "C until use. Recombinant cTnI/NH, was purified from inclusion bodies by resuspending in 20 mMiter Buffer A (8 M urea, 500 mM Tris-HC1, pH 7.9, 500 mM NaCl, 30 mM /3-mercaptoethanol, 1 mM EDTA, and 1 mM PMSF). Cells were sonicated at 2-min intervals for 20 s for a total of 6 min. Cell lysate was centrifuged a t 36,000 x g for 10 min in an SS-34 rotor. The supernatant was dialyzed overnight against 4 liters of Buffer B (6 M urea, 25 rnM Tris-HC1, pH 8.0, 1 mM EDTA, 1 mM PMSF, and 1 mM 1,4-dithiothreitol) and loaded onto a Pharmacia Bio- tech CM-Sepharose column equilibrated in Buffer B. Recombinant cTnVNH, was purified to apparent homogeneity by elution from the CM-Sepharose column using a 0-1 M NaCl gradient. cTnI samples were stored in 6 M urea until use in cTnI.cTnC complex formation.

Boponin I-Doponin C Complex Formation-The binary cTnC3.cTnY NH, complex was prepared by slowly adding aliquots of cTnI/NH, in 6 M urea to a concentrated solution of cTnC3. Routinely, 0.81 pmol of cTnC3 in 3 ml of NMR buffer was placed in a 5-ml Amicon cell contain- ing a PM-10 membrane. To this solution was added 0.2 pmol of cTnY NH,, and the resulting solution was concentrated to -1 ml and diluted with 5 ml of NMR buffer to dilute the residual urea. Additional 0.2-pmol aliquots of cTnI/NH, were added until an equimolar concentration of cTnC3 and cTnI/NH, was obtained. The resulting complex was repeat- edly washed by pressure filtration to remove all traces of urea. The total protein concentration of the complex was used to determine that both cTnC3 and cTnI/NH, were present in approximately equimolar amounts, making the assumption that there was minimal loss of cTnC during complex formation. Protein concentrations were determined us- ing the Bradford assay (24) and the bicinchoninic acid assay (25). The cTnC3.cTnI/NH2 complex was visualized on 15% SDS-polyacrylamide or 10% native polyacrylamide gels by staining with Coomassie Brilliant Blue.

NMR Methods-Heteronuclear single- and multiple-quantum coher- ence (HSMQC) (26) spectra of the [methyl-13C1Met-labeled proteins were typically collected with 1024 complex data points in the t, domain and 200 increments in the t, domain. The 'H and I3C spectral widths were 5556 and 2000 Hz, respectively. The water resonance was sup- pressed by continuous irradiation during the relaxation delay. HSMQC spectra were processed with a 60"-shifted sine-bell squared function and zero-filled to 1024 points in both t, and t,. 'H and I3C chemical shifts were reported relative to the HDO signal at 4.563 ppm and to the methyl carbon of [methyl-'3ClMet a t 14.86 ppm, respectively. All spectra were processed using the FELIX 2.0 software package (Hare Research, Inc.). The 'H and I3C assignments for the methionine methyl groups in cTnC3 were previously assigned by comparison of HSMQC spectra of a set of [methyl-'3C]Met-labeled recombinant proteins having known Met residues mutated to Leu? This was accomplished using six sets of triple Met mutants having different combinations of Met residues altered to Leu and six sets of single Met to Leu mutations?

Identification ofcTn1-(33-80/ Fragment"cTnI/NH, was found to un- dergo selective proteolysis in samples of the cTnC3.cTnLNH, complex maintained at 40 "C in the absence of protease inhibitors. SDS-polyac- rylamide gel electrophoresis showed that cTnI/NH, had been selectively proteolyzed to yield a fragment of -6 kDa. The cTnI/NH, fragment was isolated from native cTnC3 by chromatography on a Pharmacia Biotech Mono Q column equilibrated in 20 m~ Tris-HC1, pH 7.5, containing 200 mM KCI. N-terminal amino acid sequencing of the purified cTnI/NH, fragment was performed by Edman degradation on an Applied Biosys- tems model 4738 protein sequencer. The N-terminal sequence was found to be Met-Glu-Pro-His-Ala-Lys-Lys-Lys-Ser-Lys-Ile-Ser, which corresponds to the N terminus of recombinant cTnI/NH,. The molecular mass of the N-terminal cTnI fragment was determined by matrix-as- sisted UV laser desorption mass spectroscopy to be 5520 Da, which is consistent with Arrg8' being the C-terminal amino acid (18). Thus, the cTnUNH, fragment represents amino acids 1-80 of recombinant cTnY NH,. This fragment (cTnI-(33-80)) lacks the 32-amino acid cardiac- specific N terminus.

RESULTS AND DISCUSSION

As part of a general program to understand Ca2+-dependent protein-protein interactions that modulate the activity of acto- myosin ATPase in cardiac muscle, we have initiated NMR stud- ies on the cardiac troponin I-troponin C complex. The large

X. Lin, G. A. Krudy, J . W. Howarth, P. R. Rosevear, and J. A. Putkey, submitted for publication.

Spatial Relationships in the TnI-TnC Complex 23733

molecular weight of this system (40,000), the predicted high per- centage of a-helical structure in TnC, three active EF-hand Ca2+-binding motifs in cTnC, the lack of structural information on cTnI, and the known insolubility of cTnI have impeded the acquisition of atomic resolution structural information on this system. We have previously shown using NMR that individual Met methyl groups can be detected and used to monitor ligand- :nduced conformational changes in proteins having an approx- mate molecular mass of 66 kDa (28). Methionine methyl groups

were chosen as spectral probes based on our previous experience with isotopic labeling of Met residues in large molecular weight systems, the even spatial distribution of the 10 Met residues in cTnC, and suggestions that several of the Met residues were involved in the proposed hydrophobic pocket responsible for Ca2+-specific interactions with TnI. Previously, the 10 Met methyl groups in cTnC3 were assigned in our laboratory by com- parison of HSMQC spectra of wild-type protein, [SePs]cTnC, and a series of mutant proteins in which Met residues were systematically changed to Leu.,

To study the interaction of cTnC with cTnI, a recombinant truncated form of cTnI (cTnVNH,) lacking the N-terminal 32- amino acid cardiac-specific sequence was used to form a 1:l complex with [n~ethyl-'~C]Met-labeled cTnC3. Under conditions of complex formation, any unreacted cTnVNH, was insoluble and could be removed by centrifugation. The total protein con- centration of the complex was used to determine that both cTnC3 and cTnVNH, were present in approximately equimolar amounts. Complex formation was also monitored by SDS-poly- acrylamide and native acrylamide gel electrophoresis (Fig. 1). Native acrylamide gel electrophoresis demonstrated that in the absence of Ca2+, nearly all of the [methyZ-'3ClMet-labeled cTnC was complexed with cTnVNH,. SDS-polyacrylamide gel elec- trophoresis was used to visualize the purity of the final complex and to monitor proteolysis of cTnI (Fig. 1B ).

Fig. 2A shows the HSMQC spectrum of Ca2+-saturated [n~ethyZ-~~C]Met-labeled cTnC3. Each of the 10 Met methyl groups could be resolved and a sequence specifically assigned permitting their use as structural probes? The HSMQC spec- trum of the Ca2+-saturated cTnC3-cTnVNH2 complex is shown in Fig. 223. Although the molecular mass of the complex is 39.5 kDa, all 10 Met methyl groups could be individually resolved. Complex formation resulted in large chemical shifts in the 'H-13C correlations for methionines 81, 120, 137, and 157 (Fig. 2, A and B; and Table I). Methionines 120,137, and 157 are all located in the C-terminal domain of cTnC. Methionine 81 is lo- cated in the N-terminal portion of the central helix. These chem- ical shift changes result from alterations in the unique local electronic environment around each of the individual Met resi- dues. Changes in the electronic environment could result either from direct interactions between the individual Met residues and a region of cTnI or from conformational changes in cTnC induced by the binding of cTnI.

In the presence of cTnI/NH,, broadening of the 'H-13C corre- lations for Met6" and Metss, compared with the other Met 'HJ3C correlations in the complex, was observed, suggesting confor- mational heterogeneity at these locations in the complex. This conformational heterogeneity could reflect multiple orienta- tions, each having a slightly different magnetic environment, for these methyl groups (Fig. 2B). Methionine 85 is located near the center of the central helix. Chemical shifts ofmethionines 45,47, 80, and 103 were not significantly perturbed upon binding cTnI/ NH, to Ca2+-saturated cTnC3 (Fig. 2 and Table I). Thus, the major interaction between Ca2+-saturated cTnC and cTnVNH,, as monitored by chemical shift changes in the ' H W correla- tions for Met methyl groups of cTnC, occurs within the C-ter- minal domain and in the N-terminal portion of the central helix.

A v

TnC/Tnl

- - TnC

1 2

B kDa 66 - 45 - 36 - 24 -

cTnVNH2

cTnC 14- Tnl(33-80)

1 2 3 4 5 6

vine trypsinogen (24 kDa) c T n m . A, 15% native acrylamide gel FIG. 1. Complex formation between recombinant cTnC and bo-

electrophoresis of cTnC3 and cTnC3dI'nVNH2 used to visualize the cTnI.cTnC complex before NMR structural studies. B, 15% SDS-poly- acrylamide gel electrophoresis of cTnC3, cTnI/NH,, cTnC3.cTnI/NH2, proteolyzed cTnCB.cTnI/NH,, and purified cTnI433-80). Lune 1, molec- ular mass standards (bovine serum albumin (66.2 m a ) , hen egg white albumin (45 kDa), glyceraldehyde-3-phosphate dehydrogenase (36 kDa), and hen egg white lysozyme (14.4 m a ) ; lune 2, purified recom- binant cTnI/NH,; lune 3, cTnC3.cTnl/NH, complex; lune 4, proteolyzed cTnC3dhI/NH, complex; lune 5, purified recombinant cTnC3; lune 6, purified cTnI-(33-80) fragment from the proteolyzed cTnCB.cTnVNH, complex. Gels were stained with Coomassie Brilliant Blue.

Using recombinant TnI deletion mutants, multiple interac- tion sites between sTnC and sTnI have previously been proposed (12,16): specifically, a metal-independent structural site, a Ca2+/ Me-dependent site, and a Ca2+-specific binding site (12). We propose that the conformational response observed in the C- terminal domain of cTnC upon binding cTnI/NH, is largely re- sponsible for maintaining the stability of the cTnI.cTnC com- plex. Interactions observed between the central helix of cTnC and cTnI/NH, may be involved in the regulation of muscle con- traction by Ca2+. In the model of Ca2+-saturated cTnC, based on the crystal structure of sTnC (7,27), both Mets1 and Met" lie on the same side of helix D and are partially exposed to solvent. Both Mets' and MetRS form a portion of the hydrophobic patch that is implicated in regulating Ca2+-dependent interactions with cTnI (7, 23). In contrast, the side chain of Met'" is less exposed, is located more toward the interior of the protein, and appears not to be perturbed by TnI binding (Fig. 2B and Table I). In support of this model, the measured solvent exposure of Metsn is significantly less than that determined for Met" and MetR5 in Ca2+-saturated cTnC3.3 The absence of a significant change in the 'H-13C correlation for Met4' was somewhat sur- prising since Met4s is one of several hydrophobic residues (in-

J. W. Howarth, G. A. Krudy, X. Lin, J. A. Putkey, and P. R. Rosevear, submitted for publication.

23734 Spatial Relationships in the TnIeTnC Complex

A

157 0

85 43 e 103

6 81

0 8 0 8

45

B

0 103

47 * (tg 6o

137

81 8 Q 80

120 @ 45 0

157

C 157 @

85 a 0 103

47 @ 60

0 81 0 137

@ 80

120 Q * 45 2.4 2.2 2.0 1.8 1.6 1.4 1.2

proton (ppm)

beled cTnC3 (A), Ca--saturated [methyZ-lSCIMet-labeled FIG. 2. HSMQC spectra of Ca"-saturated [methyZ-lSCIMet-la-

cTnC3dIhl/NH, complex (B) , and [methyZ-lSC]Met-labeled cTnC3.cTn1-(33-80) complex (C). Samples were -1.5 mM cTnC3 in 20 mM Tris-dl,, 200 mM KC1,O.l mM PMSF, and 1 PM leupeptin in 100% 2H,0 at pH 7.5 (A); 1.2 mM cTnC3,cTnI/NH, complex in 20 mM Tris-dl,, 200 mM KCl, 0.1 mM PMSF, and 1 p~ leupeptin in 100% 'H,O at pH 7.5 ( B ) ; and 1.2 mM cTnC3.cTnI-(33-80) complex in 20 mM Tris-d,, and 200 mM KC1 in 100% ,H,O at pH 7.5 ( C ) . Spectra were obtained at 500 MHz and 40 "C as described under "Experimental Procedures."

cluding Phe27, Phe77, Val64, Met", and Mets5) that, based on the model of Ca2+-saturated cTnC, line the hydrophobic patch (7,23, 27).

TABLE I

Ca2+-saturated [methyl-'3CIMet-labeled cardiac troponin C Proton and carbon methionine methyl chemical shifts in

free and bound to cTnIlNH, or cTnl-(33-80) Proton chemical shifts are reported relative to the 2H'H0 signal at

4.563 ppm and are accurate to kO.01 ppm. Carbon chemical shifts are reported relative to [rnethyZ-'3Clmethionine at 14.86 ppm and are accu-

"Experimental Procedures." rate to k O . 1 ppm. The spectra were acquired at 40 "C as described under

cTnC cTnC.cTnI/NH, cTnC~TnI-(33-80) Residue

'H 13C 'H 1% 'H 13c

PPm PPm Met45 1.83 16.4 1.82 16.5 1.83 16.6 Met47 2.18 15.5 2.17 15.5 2.19 15.5 Met6" 1.92 15.6 1.92 15.5 1.92 15.5 Metso 1.77 16.2 1.77 16.2 1.79 16.3 Mets' 1.39 15.8 1.84 16.2 1.38 15.7 Mets5 2.10 15.2 2.06 15.2 2.09 15.1 M e P 3 1.90 15.0 1.87 15.0 1.87 15.0 MetI2" 1.86 15.6 1.98" 16.7" 1.99" 16.7" Met'37 1.90 15.7 1.89" 15.9" 1.90" 15.9" Met'57 1.80 14.4 1.06 13.6 1.06 13.6

ppm

The resonance assignments for Met'," and Met'37 in the cTnI/NH, or cTnI-(33-80) complexes could be reversed.

The cTnI.cTnC complex was also prepared using [methyl- 13C]Met-labeled cTnI/NH,. Purified cTnI/NH, contains 6 Met residues, including the initiator Met. In the absence of cTnC, cTnI/NH, is only soluble in high concentrations of urea or salt. In the presence of 6 M urea, onIy a single 'H-13C correlation, corresponding to that observed for Met in a random-coil struc- ture, was observed for the six Met methyl groups in cTnI/NH,. However, in the presence of cTnC, all six of the Met methyl groups in cTnI/NH, could be resolved in the HSMQC spectrum (data not shown). Although we do not yet have sequence-spe- cific assignments for the methyl group 'H-13C correlations in cTnI/NH,, the observation of all six methyl group correlations in [rnethyl-'3C]Met-labeled cTnI/NH, in combination with the observation of all 10 Met residues in [rnethyl-'3ClMet-labeled cTnC3 demonstrates the homogeneity of the cTnI.cTnC com- plex under the conditions of the NMR experiment.

We observed that incubation of the [methyZ-13ClMet-labeled cTnC-cTnI/NH, complex at 40 "C, in the absence of protease inhibitors, resulted in chemical shift changes in the HSMQC spectrum of several cTnC3 Met methyl groups (Fig. 2C) . SDS- polyacrylamide gel electrophoresis of this sample showed that cTnI/NH, had undergone proteolytic degradation to yield a fragment of -6 kDa (Fig. 1B). No apparent proteolytic degra- dation had occurred in [rnethyZ-13ClMet-labeled cTnC3. The proteolytic fragment of cTnI/NH, was purified by anion ex- change chromatography in the absence of metals (Fig. 1B). Edman degradation defined the N-terminal sequence as Met- Glu-Pro-His-Ala-Lys-Lys-Lys-Ser-Lys-Ile-Ser, corresponding to the N terminus of recombinant cTnI/NH,. A molecular mass of 5.5 kDa for the cTnI/NH, fragment was determined by matrix- assisted W laser desorption mass spectroscopy. This corre- sponds to the first 49 amino acids in recombinant cTnUNH,. However, with respect to the full-length rat cardiac troponin I sequence, this fragment corresponds to amino acids 33-80, where the first 32 amino acids compose the cardiac-specific N terminus. Thus, the nomenclature adopted for this fragment was cTnI-(33-80).

In the [methyZ-'3C]Met-labeled cTnC3.cTnI-(33-80) complex, resulting from proteolysis of cTnI/NH,, large changes in the 'H-W correlations were observed only for methionines 120, 137, and 157 (Fig. 2C and Table I). These changes are similar to those observed previously in the native cTnC.cTnI/NH, com- plex (Fig. 2B and Table I). Methionines 120, 137, and 150 are located in the C-terminal domain of cTnC, suggesting that the

Spatial Relationships in the TnI.TnC Complex 23735

N-terminal region of cTnI interacts with the C-terminal do- main of cTnC such that the two proteins are oriented in a head-to-tail fashion or are antiparallel. An antiparallel ar- rangement for TnI and TnC in the TnI.TnC complex has pre- viously been suggested by experiments using truncated TnI pro- teins (12,161. In support of this, the 'H-13C correlation for Mets1 in helix D was found not to be significantly perturbed relative to free cTnC3 in the cTnC3.cTnI-(33-80) complex (Fig. 2 and Table I). In addition, the broadening of the Met" and Mets5 methyl 'H-13C correlations observed in the intact cTnC3.cTnV NH, complex (Fig. 2 B ) was not observed in the cTnC3.cTnI- (33-80) complex (Fig. 2C) since the C-terminal region of cTnI is no longer available for interaction with the N-terminal domain of cTnC3. This suggests that C-terminal regions of cTnI-(81- 211) are likely responsible for interactions with the N-terminal domain and central helix of cTnC. Taken together, these studies provide strong structural evidence for an antiparallel arrange- ment for these two proteins in the troponin complex.

Based on biochemical studies (12, 16, 29) and on our struc- tural studies on the cTnI.cTnC complex, it is now clear that the C-terminal domain of cTnC interacts with the N-terminal do- main of cTnI to help maintain the integrity of the TnI.TnC complex. Moreover, the largest cTnI-induced conformational changes in cTnC, based on changes in the magnetic environ- ments of the 'H-W correlations for the methionine methyl groups, occur in the C-terminal domain of cTnC. This may be a consequence of the tight association between TnI and TnC. Calcium-specific regulatory interactions involve the N-termi- nal domain of cTnC and the C-terminal domain of cTnI. Spe- cifically, our studies suggest that helix D of cTnC makes strong contacts with the C-terminal portion of cTnI-(81-211). Addi- tional structural studies using a variety of truncated TnI pro- teins in the presence and absence of Ca2+ will permit us to further delineate regions of interaction between these two pro- teins and the mechanism by which Ca2+ binding to cTnC is transmitted through TnI to the other proteins in the thin filament.

1. 2. 3. 4.

5. 6.

7. 8.

9.

10.

11.

12.

13.

14. 15.

16.

17.

18.

20. 19.

21.

22.

23.

25. 24.

26. 27.

28. 29.

REFERENCES Leavis, P. C., and Gergely, J. (1984) CRC Crit. Rev. Biochem. 16,235305 Adelstein, R. S. (1980)Annu. Reu. Biochem. 49,921-956 Kretsinger, R. H., and Nockolds, L. E. (1973) J. Biol. Chem. 248,33133326 Collins, J. H., Greaser, M. L., Potter, J. D., and Horn, M. J. (1977) J. Biol.

VanEerd, J. P., and Takahashi, K. (1976) Biochemistry 15, 1171-1180 Babu, Y. S., Sack, J. S., Greenough, T. L., Bugg, C. E., Means, A. R., and Cook,

Herzberg, O., and James, M. N. (1985) Nature 313,653459 Satyshur, K. A., Sambhoroa, T. R., Pyzalska, D., Drendael, W., Greaser, M., and

Sundaralingam, M. (1988) J. Biol. Chem. 263, 162S1647 Putkey, J. A,, Sweeney, H. L., and Campbell, S. T. (1989) J. Biol. Chem. 264,

12370-12378 Negele, J. C., Dotson, D. G., Liu, W., Sweeney, H. L., and Putkey, J. A. (1992)

J. Biol. Chem. 267,825-831 Syska, H., Wilkinson, J. M., Grand, R. J. A,, and Perry, S. V. (1976) Biochem.

J . 153.375487

Chem. 252,635M362

W. J. (1985) Nature 315, 3 7 4 2

Sheng, Z., Pan, B.-S., Miller, T. E., and Potter, J. D. (1992) J. B i d . Chem. 267, 25407-25413

Zot, A. S., and Potter, J. D. (1987) Annu. Rev. Biophys. Biophys. Chem. 16,

Zot, A. S., and Potter, J. D. (1982) J. Biol. Chem. 257, 7678-7683 Grabarek, Z., Drabikowski, W., Leavis, P. C., Rosenfeld, S. S., and Gergely, J.

Farah, C. S., Miyamoto, C. A., Ramos, C. H. I., desilva, A. C. R., Quaggio, R. (1981) J. Biol. Chem. 256, 13121-13127

B., Fujimori, IC, Smillie, L. B., and Reinach, F. C. (1994) J. Biol. Chem. 269, 5230-5240

Murphy, A. M., Jones, L., 11, Sims, H. F., and Strauss, A. W. (1991) Biochem- istry 30, 707-712

Guo, X., Wattanapermpool, J., Palmiter, K. A,, Murphy, A. M., and Solaro, R. J. (1994) J. B i d . Chem. 269, 15210-15216

Guo, X., Murphy, A. M., and Solaro, R. J. (1992) Biophys. J. 61, A158 (abstr.) Wattanapermpool, J. (1994) Biophys. J . 66, W338 (abstr.) Krudy, G. A., Brito, R. M. M., Putkey, J. A., and Rosevear, P. R. (1992) Bio-

Xu, B., Krudy, G. A., and Rosevear, P. R. (1993) J. Biol. Chem. 268, 16259-

Herzberg, O., Moult, J., and James, M. N. G. (1986) J. Biol. Chem. 261,

Bradford, M. M. (1976) Anal. Biochem. 72, 248-254 Smith, P. E., Krohn, R. I., Hermanson, G. T., Mallia, A. K., Gartner, F. H.,

Provenzano, M. D., Fujimoto, E. K., Goeke, N. M., Olson, B. J., and Klenk, D. C. (1985) Anal. Biochem. 150, 76-85

535-559

chemistry 31,1595-1602

16264

2638-2644

hiderwig, E. R. P. (1990) J. Magn. Reson. 86,346-354 Brito, R. M. M., Putkey, J. A., Strynadka, N. C. J., James, M. N. G., and

Rosevear, P. R. (1988) Biochemistry 27, 7931-7939 Rosevear, P. R. (1991) Biochemistry 30, 10236-10245

Kobayashi, T., Tao, T., Gergely, J., and Collins, J. H. (1994) J. Biol. Chem. 269, 5725-5729