quantification of the calcium-induced secondary structural changes

Protein Science (1994), 3:1961-1974. Cambridge University Press. Printed in the USA. Copyright 0 1994 The Protein Society

Quantification of the calcium-induced secondary structural changes in the regulatory domain of troponin-C

STEPHANE M. GAGNE, SAKAE TSUDA, MONICA X. LI, MURAL1 CHANDRA, LAWRENCE B. SMILLIE, AND BRIAN D. SYKES Department of Biochemistry, Medical Research Council Group in Protein Structure and Function, University of Alberta, Edmonton T6G 2H7, Canada

(RECEIVED May 23, 1994; ACCEPTED August 30, 1994)

Abstract

The backbone resonance assignments have been completed for the apo ('H and "N) and calcium-loaded ( 'H, IsN, and 13C) regulatory N-domain of chicken skeletal troponin-C (1-90), using multidimensional homonuclear and heteronuclear NMR spectroscopy. The chemical-shift information, along with detailed NOE analysis and 3JHNHa coupling constants, permitted the determination and quantification of the Ca2+-induced secondary structural change in the N-domain of TnC. For both structures, 5 helices and 2 short 0-strands were found, as was observed in the apo N-domain of the crystal structure of whole TnC (Herzberg 0, James MNG, 1988, JMol Biol 203:761-779). The NMR solution structure of the apo form is indistinguishable from the crystal structure, whereas some structural differences are evident when comparing the 2Ca2+ state solution structure with the apo one. The major conformational change observed is the straightening of helix-B upon Ca2+ binding. The possible impor- tance and role of this conformational change is explored. Previous CD studies on the regulatory domain of TnC showed a significant Ca2+-induced increase in negative ellipticity, suggesting a significant increase in helical content upon Ca2+ binding. The present study shows that there is virtually no change in a-helical content associated with the transition from apo to the 2Ca2+ state of the N-domain of TnC. Therefore, the Ca2+-induced increase in ellipticity observed by CD does not relate to a change in helical content, but more likely to changes in spatial orientation of helices.

Keywords: calcium; CD; NMR; regulatory domain of troponin-C; secondary structural change

Troponin-C has a key role in muscle contraction of vertebrate striated muscle (skeletal and cardiac). The binding of Ca2+ to TnC induces a conformational change that affects the interaction between TnC, troponin-I (TnI), and troponin-T (TnT). This interaction blocks the inhibitory action of TnI, allowing forma- tion of the Mg2+-activated ATPase actomyosin complex, and ultimately leads to muscle contraction. The roles and interac- tions of proteins in the regulatory system of striated muscle have been studied extensively (Leavis & Gergely, 1984; Ohtsuki et al.,

Reprint requests to: Brian D. Sykes, Department of Biochemistry, 474 Medical Sciences Building, Edmonton, Alberta T6G 2H7, Canada; e-mail: [email protected].

Abbreviations: TnC, troponin-C; NTnC, N-terminal domain of troponin-C (1-90); RnC-apo, Ca2+-free troponin-c (1-90); NTnC . ~ a , Ca2+-saturated troponin-C (1-90); TRIG, tryptic fragment of troponin- C (12-87); NOESY, nuclear Overhauser effect spectroscopy; DQF- COSY, double quantum-filtered correlation spectroscopy; TOCSY, total correlation spectroscopy; 2D, 2-dimensional; 3D, 3-dimensional; SCUBA, stimulated crosspeaks under bleached alphas; CSI, chemical- shift index; HMQC, heteronuclear multiple-quantum coherence.

1986; Zot & Potter, 1987; Grabarek et al., 1992; Farah et al., 1 994).

TnC (M, = 18,000) has 4 Ca2+-binding sites: 2 high-affinity sites (I11 and IV), which are believed to be always occupied by either Ca2+ or Mg2+ under physiological conditions ( L a = =2 x lo7 M"; K M g = =2 x lo3 "I); and 2 low-affinity Ca2+- specific sites (I and 11, Kc, = 3 X 10' "I), which regulate muscle contraction. The crystal structure of turkey skeletal TnC (Herzberg & James, 1988) and of chicken skeletal TnC (Satyshur et al., 1988, 1994) revealed a 66% a-helical protein having 2 globular domains, each containing 2 calcium-binding sites, con- nected by an extended a-helix. In these structures, only sites I11 and IV in the C-terminal domain are occupied by Ca2+, whereas the regulatory sites in the N-domain are in the apo state. All 4 Ca2+-binding sites show the helix-loop-helix motif termed the EF-hand. The helix packing is, however, different in the Ca2+- free regulatory N-domain as compared to structures of homologous Ca2+-binding proteins, which have their Ca2+ sites filled (Strynadka & James, 1989). The NMR solution structure of the TRIC fragment of TnC in the apo-form (Findlay et al., 1994)

1961

1962

shows similar structural features to the apo N-domain of TnC. By comparing the crystal structure of the apo N-domain of TnC with its homologous C-domain, a model has been proposed for the conformational change that occurs in the N-domain of TnC upon Ca2+ binding (Herzberg et al., 1986). This model is the only structural representation of the Ca2+-saturated form of the regulatory domain of TnC. A detailed experimental characterization of this conformational change is required for a complete understanding of muscle contraction at the molecular level.

Although no structure for the Ca*+-saturated state of TnC is known, the Ca2+-induced conformational changes have been studied extensively using various spectroscopic techniques including NMR, CD, Raman, and fluorescence (Levine et al., 1977, 1978; Seamon et al., 1977; Hincke et al., 1978; Johnson & Pot- ter, 1978; Carew et al., 1980; Evans et al., 1980; Leavis et al., 1980; Tsuda et al., 1988, 1990; Krudy et al., 1992). In particular, a recent CD study has shown significant Ca2+-induced ellipticity changes in the isolated recombinant N-domain of TnC (Li et al., 1994). That these changes were a reflection of significant secondary structural alterations was considered a possibility.

2D-NMR has been used in recent years as a structural tool for small proteins (Wuthrich, 1986). Because spectral overlap is proportional to the size of the protein studied, analysis of larger proteins using 2D-NMR becomes tedious and structure determination is more difficult. With the development of multidimensional heteronuclear NMR (3D- and 4D-NMR) in the last few years (Clore & Gronenborn, 1991), and the feasible ''N/13C labeling of proteins via cloning and expression, NMR is now a powerful structural method for proteins in the 10-20-kDa range. These 3D- and 4D-NMR techniques allow removal of most of the overlap in NMR spectra, resulting in more complete analyses and high-resolution solution structure. Such techniques have been successfully applied to solve the solution structure of a highly homologous protein to TnC, calmodulin, complexed with a target peptide (Ikura et al., 1992).

This study reports the conformational changes that occur in the regulatory domain of chicken skeletal TnC. The N-domain (residues 1-90) of TnC (NTnC) has been successfully expressed and labeled with "N or lSN/I3C. Based on fluorescence and CD studies, the properties of NTnC are identical to those of the N-domain in intact TnC (Li et al., 1994). Thus, the structure of NTnC can be expected to accurately reflect that of the regulatory domain of whole TnC. Using 2D- and 3D-NMR, chemical- shift assignments have been completed for both NTnC-apo and NTnC .2Ca. The chemical-shift information, along with detailed NOE analysis and 3JHNHa coupling constants, have provided a detailed assessment of the secondary structural elements in the apo and 2Ca2+ states of NTnC. These analyses have demonstrated minimal Ca2+-induced changes in the a-helical content. Therefore, the significant increase in negative ellipticity previously observed by CD analyses (Li et al., 1994) must occur for other reasons, and an alternative explanation is provided. The present data on the Ca2+-induced conformational transition of NTnC provides important information necessary to our understanding of the mechanism of muscle regulation by TnC.

Results

2D and 3D multinuclear NMR provide high-resolution spectra with well-resolved signals from individual nuclei that can be used

S.M. Gagne et al.

to determine structure and probe conformational changes in de- tail. The occurrence of conformational changes in the regulatory domain of TnC is clearly indicated by the chemical-shift differences between the apo and 2Ca2+ states of NTnC. These changes are very well pictured when comparing the 2D( "N- 'H) HMQC spectrum of NTnC-apo and NTnC.2Ca (Fig. 1). With the exception of the asparagine and glutamine side-chain correlations, each crosspeak in the displayed region of the 2D( 15N-'H] HMQC represents a backbone amide NH. NTnC contains 1 proline, and thus a maximum of 89 "N-'HN backbone amide correlations may be observed; 86 and 84 backbone amides were observed and assigned for NTnC-apo and NTnC. 2Ca, respectively. The missing correlations correspond to amide hydrogens that are in fast exchange with the solvent and therefore are saturated during solvent presaturation. The very low overlap between the 15N and/or 'H chemical shifts of NTnC-apo and NTnC .2Ca reveals that the perturbations due to CaZ+ binding are propagated throughout the entire molecule. In order to properly attribute these changes to individual amino acids or regions in the protein, we first carefully assigned the backbone 'HN, I'N, and 'Ha resonances for NTnC-apo and the 'HN, I5N, 'Ha, I3Ca, and I3CO resonances for NTnC.2Ca.

Assignment

The assignment strategy for NTnC-apo was based on conven- tional methods using 2D-COSY, 2D-TOCSY, and 3D-I'N-edited TOCSY experiments to identify spin systems, and 2D-NOESY and 3D-I'N-edited NOESY experiments to identify interresidue connectivities along the polypeptide chain (Wuthrich, 1986). The assignment of the NTnC-apo was based in part on the assignment of the TR,C fragment (Findlay & Sykes, 1993) and followed a

r 106.0

108 0

1100

112.0

114.0

1160

118.0

120.0 1 5 N

122.0

124.0

126.0

, , , , , , , , : :;, ,o;* *..;" , , , , ,I 128.0

130.0

6' I320

I 0 8 10.4 100 9.6 9 2 8.8 8.4 8 0 7.6 1.2 6.8

1 H

Fig. 1. Superposition of the 2D ( "N-IH] HMQC spectra of NTnC- apo (filled crosspeaks) and NTnC.2Ca (opened crosspeaks). The dotted lines illustrate the Caz+-induced chemical-shift change for a few backbone amides.

Structural change in regulatory domain of troponin-C 1963

similar procedure except for the fact that the "N-edited NOESY and TOCSY experiments were used to resolve some ambiguities. The quality of the NTnC-apo spectra was equivalent to the TRlC spectra (Findlay & Sykes, 1993). The complete 'HN, "N, and 'Ha assignments for NTnC-apo are listed in Table 1.

For NTnC .2Ca, a different approach, which relies on heteronuclear scalar coupling, was used to obtain sequential connectivities along the protein backbone (Ikura et al., 1990). Similar approaches that also rely on various combinations of correlations between 'HN, "N, 'Ha, I3Ca, and I3CO nuclei have been successfully applied for various proteins, including the homologous protein calmodulin (Ikura et al., 1990, 1991). Two types of sequential assignment were used: one centered on the I3Ca and another centered on the I3CO. The first type of sequential assignment was accomplished by combining the strong intraresidue 1HN(i)-'5N(i)-'3Ca(i) and the weaker sequential 'HN(i)- '5N(i)- '3Ca(i - 1) connectivities from the HNCA experiment, the sequential 'HN(i)-I5N(i)-l3Ca(i - 1) connectivity from the HNCOCA experiment, and the dNN(i, i k 1) NOE connectivities from the "N-edited NOESY experiment. The role of the 4 chemical-shift coordinates of each dNN ['HN(i); 15N(i); I5N(i k 1); 'HN(i k l)], which were auto- matically found in the peak list using the in-house program CHAINS (R. Boyko & S. Gagnt, MRC group in Protein Structure and Function, University of Alberta, unpubl.), was to confirm

the Ca-based sequential assignment and to resolve some ambiguities resulting from overlapping Ca. An example of HNCA, HNCOCA, and dNN connectivities is shown in Figure 2A, B, and E, respectively. A total of 153 connectivities were found in the HNCA (80 intraresidue and 73 sequential) and 77 sequential connectivities in the HNCOCA. The second strategy for sequential assignment, now relying on the l3CO, was performed by combining the intraresidue ' H a ( i)-I3Ca( i)-I3C0( i) connectivity from the HCACO experiment, the sequential 'HN( i + 1)- "N(i + l)-I3CO(i) connectivity from the HNCO experiment, and the intraresidue 'HN(i)-"N(i)-'Ha(i) from the I5N- edited TOCSY experiment. An example of these connectivities is shown in Figure 2C, D, and F. Overall, 89 intraresidue connectivities were found in the HCACO and 82 sequential ones in the HNCO. In both procedures, the "N-edited TOCSY was used to identify the spin system related to each 'HN/"N pair. Not all spin systems could be identified in the TOCSY experiment, mainly because of the small 'HN-'Ha coupling constants that are characteristic of a-helices. However, residue identification could often be helped by comparing the observed 'Ha, I3Ca, and I3CO chemical shifts to their corresponding expected value (Wishart & Sykes, 1994a). Using this procedure, the near-complete 'HN, "N, 'Ha, I3Ca, and l3CO resonance assignments of NTnC.2Ca have been achieved and listed in Ta- ble 2.

Table 1. Polypeptide backbone 'Hand I5N chemical shifts for the apo N-domain of chicken TnC a t p H 6.6 and 30 "C ~ ___

Residue ''N 'HN IHa Residue I5N

AI - s 2 -

M3 - T4 115.7 D5 121.6 4 6 121.5 Q7 122.3 A8 124.3 E9 120.6 A10 124.8 R11 116.4 A12 120.8 F13 119.4 L14 119.8 SI5 117.3 E16 122.9 E17 120.8 MI8 122.7 I19 120.7 A20 121.1 E2 1 122.8 F22 122.7 K23 122.7 A24 121.7 A25 120.8 F26 120.3 D27 116.8 M28 119.3 F29 117.3 D30 123.7

- -

8.12 8.34 8.29 8.18 8.05 8.12 8.10 7.93 7.61 7.69 7.42 7.86 9.00 8.87 7.75 8.09 7.83 7.85 9.04 9.17 7.84 7.38 8.35 9.02 7.45 7.64 7.24

-

4.12 A3 1 4.45 D32 4.60 G33 4.22 G34 4.50 G35 4.14 D36 4.00 I37 4.13 S38 4.07 T39 4.13 K40 3.90 E4 1 4.25 L42 4.35 G43 4.38 T44 4.50 v45 3.91 M46 4.05 R47 3.99 M48 3.37 L49 4.17 G50 4.17 Q5 1 4.90 N52 3.75 P53 4.17 T54 4.32 K55 3.66 E56 4.33 E57 3.94 L58 4.34 D59 4.87 A60

128.4 115.8 112.6 111.6 114.0 119.1 111.9 115.8 119.6 121.4 117.3 121.4 107.1 119.8 123.2 118.7 120.2 123.2 119.0 108.2 119.6 118.5

115.1 123.7 118.3 122.4 122.6 120.6 122.7

-

'HN ~

8.15 8.33 8.13 8.78

10.28 7.81 8.31 8.58 8.58 8.10 7.53 7.19 8.84 7.45 7.63 8.35 8.12 8.01 7.33 7.74 8.03 8.69 -

8.79 8.78 8.75 7.79 8.50 8.63 7.42

' H a Residue "N ' HN 'Ha

4.10 4.72 3.79, 4.05 3.92, 4.16 3.80, 4.34 5.58 4.65 4.86 3.76 4.14 4.70 3.77 3.49, 3.76 3.76 3.54 3.81 4.61 4.23 4.57 3.81, 4.30 4.46 5.15 4.75 4.35 3.91 4.05 4.01 4.02 4.32 4.19

I6 1 I62 E63 E64 V65 D66 E67 D68 G69 S70 G7 1 T72 173 D74 F75 E76 E77 F78 L79 V80 M8 1 M82 V83 R84 Q85 M86 K87 E88 D89 A90

122.3 121.8 119.1 118.1 115.0 123.1 124.3 117.2 111.3 120.1 115.1 113.3 118.3 126.0 120.1 117.8 120.5 122.7 120.2 118.8 119.9 120.3 121.5 120.2 117.7 118.8 120.9 122.3 122.8 130.8

7.94 8.45 7.76 7.46 8.11 8.34 8.07 8.45 7.93 9.05

10.28 7.88 8.87 8.57 8.36 7.96 8.27 8.46 8.11 7.25 7.75 8.18 8.19 7.68 7.59 7.93 7.79 8.10 8.20 7.70

3.75 3.45 3.99 4.13 4.21 5.15 4.12 4.66 3.85, 3.91 4.26 3.86, 4.16 5.43 4.79 5.61 3.50 3.75 3.97 3.98 3.42 3.28 3.91 4.04 3.55 3.87 3.98 4.20 4.16 4.26 4.63 4.12

1964 S.M. Gagne' et al.

Secondary structure determination strategy

There are 3 reliable secondary structure determination methods that are commonly used in NMR: one is based on the NOE connectivities characteristic of different secondary structures (Wiith- rich, 1986), another is based on the 3&NHa coupling constants (Pardi et al., 1984), and the last one is based on the characteristic backbone chemical-shift values of the a-helix, @-sheet, and random coil residues (Wishart & Sykes, 1994a, 199413). We have applied the 3 approaches to NTnC-apo and NTnC.2Ca, with special attention to the quantification aspects of the NOE approach.

NOE connectivities found in NOESY spectra usually contain, among others, the daN ( i - 1, i) and the dNrr (i, i) . The dNa (i, i ) NOE is noninformative by itself, the intraresidue HN-Ha distance being covalently restricted to between 2.2 and 3.1 A, and being realistically found in the 2.7-3.1-A range for a-helices and @-sheets. ForthedaN(i- 1, i) NOE, theallowedHN(j,-Ha(j-l, distance is found in a wider range (2.2 to 3.6 A), and the magnitude of the dmN ( i - 1, i) NOE could be used to differentiate between a residue found in the right-handed a-helix region (3.4 A < HN(i,-Ha(i-l, < 3.6 A) and one in the @-sheet region (2.2 A < HN(j,-Ha(j-l, < 3.1 A). However, the inaccuracy of this measurement (mainly due to the variation of amide exchange rate along the sequence) makes it a poor criterion for

b..., I MN(l+l); N(I+l); Cq l ) ~

Fl ('*CO) 181.0 180.0 1790 178.0 If7.0 176.0 175.0 1,h.o 1730 ' 7.2

F3 @N)

7.4

7.2

n c w 7.4

FI Paca) 62.0 60.0 58.0 56.0 Y.O 52.0 m.0 48.0 6.0

Fig. 2. Connectivities related to 1 residue of NTnC.2Ca, M28, in various 3D experiments. The HNCA (A) and HNCOCA (B) experiments provide sequential information via the I3Ca chemical shift. Similar sequential information can also be obtained via the I3CO chemical shift by combining the HNCO (C), the HCACO (D), and the "N-edited TOCSY (F) experiments. The "N-edited NOESY (E) experiment also provides sequential information (dNN, d a N ( i - 3, i ) , and d u ~ ( i -1, i ) ) .

secondary structure identification. This inadequacy can be over- come by using the ratio of these 2 NOEs, dNa/duN (for simplic- ity, dNa( i , i) NOE/daN( i - 1, i) NOE is abbreviated to dNu/ daN). This ratio is smaller than I for @-strands, and larger than 1 for right-handed a-helices (see Fig. 3). Because these 2 NOEs are related to the same amide HN, the dNa/dmN ratio has the advantage of being independent of amide exchange and can be quantitated more safely. Note that the magnitude of this ratio relates mainly to the $ torsion angle of residue (i - l ) , if only the right-handed a-helix and the @-sheet regions are considered. We therefore used this ratio as one of the criteria for secondary structure characterization, by measuring the ratio of the intensity of the d m N ( i - 1, i ) and dNm( i , i ) NOE found on the I5N plane of residue (i) in the "N-edited NOESY. The second NOE criterion used is the daN( i - 3, i ) connectivity, which is representative of a-helices (Wiithrich, 1986). Examples of dNm, d a N , and daN ( i - 3, i) connectivities can be found in Figure 2E. The third criterion included in the determination of the secondary structure is the CSI, which is applied using the methodology described by Wishart and Sykes (1994a, 1994b). The CSI was determined using the 'Ha chemical shifts for NTnC-apo, and the consensus of the 'Ha , I3Coc, and 13C0 indexes without any smoothing for NTnC .2Ca. Structural information about torsion angles can also be obtained from scalar coupling constants based on Karplus equations (Karplus, 1963). The relation between the backbone 4 angle and the ' & N & coupling constant is available (Pardi et al., 1984) and is used as the fourth criterion to defined secondary structure. 3 & N H a < 6 Hz was taken to indicate helices, whereas 35HNHm > 8 Hz was taken to indicate @ secondary structure. Unambiguous absence of splitting at high-resolution enhancement (LB = -28) was interpreted

Definition of helices can be a delicate issue because different methods can be used (Richards & Kundrot, 1988). Therefore, it is vital to stipulate clearly the approach used before stating that a helix starts or ends at a certain residue. The information obtained from the data presented in this paper is the following: (1) the occurrence of a dNa/daN > 1 indicates that residue i - 1 possesses a $ angle in the a-helical region; (2) d,N(i - 3, i ) NOE suggests that residue i - 3, i - 2, and i - 1 are in helical conformation but does not give any information about the 4/$ angle of residue i, although it may suggest that HNi is hydro- gen bonded to either COj-4 or COiW3; (3) the CSI reports pre- dominantly information about the 4 and $ angle of its related residue because backbone chemical shifts are closely related to the main-chain $/$ angle (Wishart et al., 1991); (4) the ' J H N H ~ reports estimation of the 4 angle. Ideally all 4 criteria should agree, but due to factors that are not related to secondary structure (overlap, spin diffusion, chemical shift affected by aromatic ring, . . .), a consensus is used instead. Therefore, helices were defined using the following rules: (1) residues were defined as helical when more than half of the available criteria (consensus of available dNm/daN, daN(i - 3, i), CSI, and ' & H a ) were characteristic of an a-helix; (2) at least 4 consecutive helical residues were necessary to start a helix; (3) the N-terminal residue of a helix must possess a consensus of the following: a dNa/ daN > 1 with the next residue, a darN( i - 3, i), an a-helix CSI, and a JHNHm < 6 Hz; (4) the C-terminus of a helix must have a consensus of the following: a dNa/daN > 1 with the next residue, a daN ( i - 3, i) between residue i - 2 and i + 1, an a-helix CSI, and a 3JHNH, < 6 Hz. The methodology used in this study

as 3 J ~ ~ ~ , < 6 HZ.

Structural change in regulatory domain of troponin-C

Table 2. Polypeptide backbone 'H, 13C, and "N chemical shifts for the Ca2+-saturated

- - state of the N-domain of chicken TnC at pH 6.7 and 30 "C

Residue I5N I3Ca 13C0 'HN IHQ!

A1 s 2 M3 T4 D5 Q6 4 7 A8 E9 A10 R11 A12 F13 L14 S15 E16 E17 M18 I19 A20 E2 1 F22 K23 A24 A25 F26 D27 M28 F29 D30 A3 1 D32 G33 G34 G35 D36 137 S38 T39 K40 E41 L42 G43 T44 v45

- -

116.1 123 .O 120.9 123.0 124.4 120.5 124.5 116.5 120.4 119.4 119.4 117.2 122.7 120.9 121.6 121.2 122.2 122.2 121.2 125.1 122.2 122.0 121.2 118.6 120.3 119.4 119.7 130.4 114.3 111.8 109.6 113.8 116.2 125.9 124.0 115.1 122.1 122.3 122.4 107.1 119.7 122.0

-

51.8 58.3 56.0 61.1 57.3 58.8 58.3 54.9 58.8 55.5 59.4 53.7 59.2 53.6 56.3 60.1 60.2 58.3 66.8 55.1 59.1 58.8 58.3 54.9 55.0 62.2 56.9 57.9 58.2 52.1 55.0 52.7 46.7 46.3 45.3 52.6 60.1 55.7 66.9 59.3 58.8 57.8 47.8 67.2 66.3

174.2 173.9 176.6 175.0 178.7 179.1 177.8 181.1 178.8 179.2 178.0 178.6 175.7 176.0 175.0 179.6 179.1 178.1 177.6 181.3 180.3 178.5 177.2 180.7 177.9 176.6 177.8 177.9 177.7 176.7 179.1 177.8 175.7 175.9 173.2 173.1 175.7 175.9 177.0 179.7 179.7 178.7 175.5 176.3 177.5

- -

8.01 8.69 8.87 7.83 8.50 8.13 7.92 7.96 7.69 7.68 7.37 7.83 9.05 8.89 7.73 8.24 7.84 7.77 8.72 9.24 7.57 7.57 8.88 8.57 7.30 7.78 8.08 7.68 8.19 8.02 8.14

10.69 7.75 9.61 8.70 9.22 7.84 7.70 8.52 8.58 7.94 7.30

-

4.14 4.52 4.64 4.44 4.44 4.07 3.93 4.14 4.04 4.11 3.85 4.23 4.38 4.50 4.59 3.97 4.07 4.03 3.72 4.19 4.14 5.03 3.95 4.19 3.98 3.17 4.19 4.05 4.30 4.50 4.1 I 4.59 3.85, 3.85 3.98, 4.09 3.67, 4.47 5.17 4.94 4.85 3.82 4.12 4.13 3.99 3.59, 3.98 3.94 3.69

for secondary structure determination is therefore closely related to a secondary structure definition based on +I$ angle (Rich- ards & Kundrot, 1988). The compilation of the dNa/daN, daN ( i - 3, i ) , CSI, and 3 & N H a , along with the secondary structure consensus for NTnC-apo and NTnC.2Ca are represented in Figures 4 and 5, respectively.

NTnC-apo; secondary structure

Five helices were found by NMR for NTnC-apo: helices N, A, B. C. and D. This is the same as is found in the crvstal struc-

Residue

M46 R47 M48 L49 G50 Q5 1 N52 P53 T54 K55 E56 E57 L58 D59 A60 I61 I62 E63 E64 V65 D66 E67 D68 G69 S70 G7 1 T72 I73 D74 F75 E76 E77 F78 L79 V80 M81 M82 V83 R84 Q85 M86 K87 E88 D89 A90

1965

117.7 120.5 122.8 -

106.9 119.4 118.0 -

114.3 123.3 118.3 122.6 121.6 120.1 123.0 121.5 119.3 119.1 116.2 108.3

128.7 116.2 110.4 118.2 118.0 109.3 126.5 131.9 119.7 117.3 120.7 123.4 119.8 120.3 121.5 118.6 122.3 120.8 117.2 120.1 120.4 121.9 122.3 130.7

-

59.4 58.9 58.8 53.6 45.7 53.9 51.0 62.6 60.2 60.0 60.6 60.4 58.0 57.2 54.8 65.0 65.0 59.3 58.9 60.8 53.5 59.2 52.4 47.1 60.1 45.6 58.1 60.4 53.1 61.5 58.8 57.9 61.2 58.0 67.2 59.3 56.4 66.9 59.4 57.3 57.9 57.5 57.0 54.0 53.7

178.8 181.0 175.6 176.9 174.6 174.3 172.1 177.8 175.2 178.3 180.1 179.5 178.6 179.0 180.3 178.3 177.7 178.2 179.4 175.4 177.3 176.9 177.7 175.3 176.2 172.8 173.5 176.2 175.9 176.4 180.3 178.8 177.3 179.1 178.2 178.2 179.4 178.1 179.0 177.9 177.6 177.1 176.5 174.9 170.6

I H N

8.01 8.22 7.94 7.19 7.75 7.95 8.66

8.60 8.79 8.80 7.77 7.96 8.54 7.63 7.57 7.46 8.35 7.27 7.26

~

-

8.53 8.01 7.75 8.54

10.97 7.68 9.19 9.53 8.79 7.77 8.19 8.83 8.02 7.29 7.85 8.07 8.38 8.06 7.40 7.92 7.86 7.91 8.08 7.65

-

IHff

3.91 4.64 4.22 4.34 3.70, 4.12 4.43 5.14 4.74 4.43 3.88 4.08 4.00 4.03 4.30 4.18 3.71 3.42 4.00 4.04 4.59 4.70 4.23 4.73 3.81, 3.88 4.22 3.44, 4.13 4.92 5.07 5.32 3.59 3.80 4.17 4.04 3.47 3.27 3.69 4.05 3.64 3.86 4.15 4.25 4.17 4.30 4.63 4.12

tures of TnC. According to the criteria described above, the N-helix is well defined from T4 to F13. The tight turn linking the N-helix and the A-helix is characterized by a dNu/daN < 1 observed for S15 and E16, indicating positive $ angles for both L14 and S15. The A-helix unambiguously spans from E16 to M28. The negative CSI of F22 in the center of the A-helix might occur due to a deshielding orientation of the aromatic ring. No regular secondary structure is observed for the F29-G35 segment, whereas the next 3 residues are well defined as B-strand by the CSI. The (dNa/daN < 1) between 137 and T39 indicate

I , ~ ~~~ ~ ~d~ ~ ~ positive $ angles between D36 and S38, and the 3 J ~ , . , ~ u > 8 Hz

1966 S.M. GagnP et al.

R O R O . .

+ I 8 0

+ I 2 0

-I20 4 -180 -120 -50 0 +el +I20 +I80

Fig. 3. $I$ Map showing the relation between the dNu/duN ratio and the major secondary structure regions. The shaded areas represent the energy-favored + i / $ i - l regions for 2 residues in a protein. @A, Pp, (YR, and aL represent ideal +I$ angles for antiparallel &sheet, parallel 0- sheet, right-handed a-helix, and left-handed a-helix, respectively. The thick line is the contour for dNu/dddddddddddddddbN = 1 from a contour map gener- ated by varying the 62 and $, in a dipeptide. dNu/duN represents the expected ratio between the dNu(i , i ) and duN(i - 1 , i ) NOES obtained from the measure of [r(HNi-Hai-l)/r(HNi-H~i)]1'6, where r is the distance between 2 protons. A representation of the various parameters used in the +/$ map for 2 residues is shown above the map. The &sheet region is clearly characterized by a dNu/duN e I , whereas the right- handed a-helix is localize in the d ~ ~ / d ~ N > 1 region, making the dNu/ dbN ratio an accurate criterion for secondary structure determination.

of D36 and I37 indicate a @-strand characteristic 4 angle for these residues.

The B-helix of NTnC-apo is less well defined by the NOES due to some 'Ha resonances appearing very close to the water resonance (4.70 and 4.61 for E41 and R47, respectively) and 'Ha overlap (in particular in the L42-T44 segment). Even if some of the dN,/dmN and dmN( i - 3, i ) are ambiguous in the B-helix, the absence of an NOE between K40-Ha and G43-HN is unambiguous (Fig. 6). This is in perfect agreement with the crystal structure, where these 2 protons are separated by 5.19 dr. The 3 J H N H u > 8 Hz of E41 is also in agreement with the X-ray coordinates, where this residue had a $I angle of -96". The T39- M48 stretch shows 2 @-sheet-like CSI; one for D l , and the other for R47. These 2 nonhelical chemical shifts are, however, consistent with the geometry of the B-helix in the crystal structure, as demonstrated by chemical-shift calculations made using the crystal coordinates and the SHIFTS program (Jellard et al.,

PENCE, University of Alberta, unpubl.; Osapay & Case, 1991). These calculations showed an a-helix-like upfield shift relative to the random coil value for all 'Ha in the T39-M48 segment, except for E41, due to irregular c#d$ angles (-96'/-7O), and for K47 due to the particular orientation of the Q5l C=O bond. The rationalization related to K47Ha is based on chemical-shift calculations with and without QSlCO, which demonstrated that Q5lCO induces a downfield shift on K47Ha. Helix-B therefore spans from L42 to M48 but can also be viewed as spanning from T39 to M48 with a kink at residue E41.

No particular secondary structure is associated with L49-T54. The last 2 helices, C and D, are very well defined, spanning from K55 to E64 and F75 to K87, respectively. The second Ca*+- binding loop, like the first one, possesses no distinct NOE or CSI features for the first 6 residues, and a @-strand for the last 3 residues, T72-D74. The 3 C-terminal residues are relatively flexible/unstructured.

NTnCaZCa; secondary structure

Five helices are also found in the calcium-saturated form of NTnC, with slightly different lengths for some of them in comparison to the apo form. Considering that the amide protons of the first 3 residues are not observed due to fast exchange, and that the CSI is variable for the Al-T4 segment, the first 4 residues are probably disordered. The N-helix is very well defined from D5 to F13. The A-helix is also well delineated from E16 to F29 using our criteria, having only 1 ambiguous duN(i - 3, i ) . The first calcium-binding loop of NTnC does not show any regular secondary structure features from D30 to G35. The binding-loop ends with a short @-strand (D36-S38) defined by negative CSIs for D36.137, and S38, dN,/daN < 1 for 137 and

least well defined using the dN,/daN criterion, due to overlap or very weak NOES. However, the daN(i - 3, i ) , CSI, and 3JHNHol unambiguously characterize the B-helix from T39 to R47. Unfortunately, because the "N of L49 is unassigned, the only secondary structure information available for M48 is its c s l and it 3JHNHol , which are both characteristic of helices. Helix-B is therefore assumed to extend from T39 to M48. The L49-T54 section of NTnC-2Ca shows variations for both the dNu/duN and the CSI. and is categorized as a 6-residue linker between helices B and C. The C-helix is also well characterized from K55 to E63, with some ambiguity about its C-terminus. As can be seen in the 2D( "N-'H) HMQC NMR spectrum (Fig. 1). V65 is one of the weakest observable amides, leading to virtually no observable NOES in the 3D-"N-edited NOESY. In addition, the amide of D66 is unassigned or nonobservable, so that only the CSI and the '&Ha are available to characterize E64. The C-helix therefore ended at E64. The V65-G71 segment does not have any a-helix or @-sheet character. Finally, the second @-strand (T72-D74) and helix-D (F75-K87) are well defined by the NMR data. The C-terminal residues E88-A90 are flexible, as for the apo structure.

S38, and 3JHNHp > 8 HZ for I37 and S38. The B-helix is the

Ca'+-induced conformational change

In order to fully characterize the secondary structural changes that occur in NTnC upon calcium binding, we compared the secondary structures obtained for NTnC-apo and NTnC - 2Ca from


5 10 1 5

A S M T D Q Q A E A R A F L S

Sec. struct. - 2 0 2 5 3 0 3 5 4 0 4 5 5 0

E E M I A E F K A A F D M F D A D G G G D I S T K E L G T V M R M L G Q

3JHN-Ha a a a a a a a a a - a a a p p a p - - - p P - a a p a a a a a ? a p - p

da(i-3,i) " " - "

_""". """" """" - """" """" """"

dNddclN

CSI

5 5 6 0 6 5 7 0 1 5 80 8 5 9 0 N P T K E E L D A I I E E V D E D G S G T I D F E E F L V M M V R Q M K E D A

Sec. struct. r Fig. 4. Schematic representation of the secondary structure determination of NTnC-apo. daN(i - 3, i ) represents the NOE connectivity between Ha,-3 and HNi; d ~ , / d , ~ represents the ratio between the dNa(i , i ) and d u N ( i - 1, i ) NOE intensities (upward and downward relate to d N , / d , ~ > 1 and d N a / d a ~ < I , respectively); CSI represents the chemical-shift index (Wishart & Sykes, 1994a, 1994b) for the a-protons only (upward and downward relate to a-helix and 6-strand characteristics, respectively); 3JHN&, reports values of J < 6 Hz (a ) , J > 8 Hz (6). Intermediate J s are represented by a dash (-), and ambiguous or unavailable Js by a question mark (?). Dotted lines and open squares indicate ambiguities. The compilation of the secondary structure as determined by these criteria is found on the last line of each segment; rectangles and arrows indicate helices and &strands, respectively. The kink in the B-helix (see text) is represented by a break at residue 41.

this study, the NMR secondary structure of the TR,C fragment (Findlay & Sykes, 1993), and the crystal structure of TnC (Herz- berg & James, 1988). To obtain a comparison as consistent as possible for the helices, all structures were compared in terms of "NMR helices." For the X-ray structure, we extracted the dNa/daN NOE ratio, the d a N ( i - 3, i ) NOE, and the 3JHNHa expected from the crystal coordinates based on distance alone or on C#I angle and used them to define the helices with the approach described above for NTnC. The TR,C helices were defined based on the NOE data presented by Findlay and Sykes (1993). The comparison of the helices from these 4 structures is summarized in Table 3. The N-helix can be considered iden-

tical for the X-ray, NTnC-apo, and NTnC.2Ca, the only possible variance being different flexibility for the first few residues. The A-helix is invariant between the X-ray and NTnC-apo, and is 1 residue longer in NTnC.2Ca. Although the NOES presented previously for TR,C indicated D26 as being the C-terminus of helix-A, the chemical-shift identities between NTnC-apo and TRIC for the 'Ha residues E16-A31 (A6 s 0.02 ppm) suggests that both have identical C#I/$ angles. A revision of the NOE data related to helix-A of TRIC (data not shown) showed NOE connectivities identical to the ones presented in this paper for NTnC- apo, and therefore helix-A is believed to end at M28 in all 3 apo structures and at F29 in the Ca2+ structure.

1968 S.M. Gagne et ai.

5 10 15 A S M T D Q Q A E A R A P L S

Sec. struct.

2 0 2 5 3 0 3 5 4 0 4 5 5 0 E E M I A E P K A A F D M P D A D G G G D I S T K E L G T V M R M L G Q

3JHN-Ha a a ? a a a ? a a a a a a a ? ? p a a ? ? p p a a a a a a a a a a ? a p

5 5 60 65 70 7 5 80 85 9 0

N P T K E E L D A I I E E V D E D G S G T I D P E E P L V M M V R Q M K E D A

Sec. rtmct. - Fig. 5. Schematic representation of the secondary structure determination of NTnC.2Ca. The representation used is the same as in Figure 4, except for the CSI, which is a consensus of the 'Ha, I3Ca, and 13C0 index.

The B-helix in the crystal structure possesses a kink due to a deviation from ideal helical +/+ angles at E41 (4 = -96", + = -7"). This 4 angle is reflected in the NTnC-apo solution structure by a 35HNH, > 8 Hz (Fig. 7). In terms of expected daN(i - 3, i) in the crystal structure, there should be a strong NOE for L42 (3.0 A), a nonobservable NOE for G43 (5.2 A), and medium NOES for the T44-G50 stretch (3.3-3.8 A). The NTnC-apo NOESY spectrum shows an ambiguous daN( i - 3, i ) for L42 and none for G43, whereas the NTnC.2Ca spectrum reveals daN ( i - 3, i ) for both L42 and G43 (Fig. 6). Consequently, the kink at the N-terminus of helix-B is observed in the X-ray, TR,C, and NTnC-apo structures. This kink is, however, absent in the structure of NTnC-2Ca, as indicated by the NOE connectivities (Figs. 5 , 6) and 3 5 H N H a < 6 Hz (Figs. 5 , 7).

Site I1 of NTnC-apo is similar to the crystal conformation, as indicated by the G69 daN( i - 3, i ) NOE (Fig. 4). This NOE is the only daN(i - 3, i) expected for a nonhelical region in the

crystal structure (3.69 A). The NOE spectrum of NTnC.2Ca does not exhibit this d a N ( i - 3, i ) , indicating some Ca2+- induced conformational change in site 11. Finally, the conformation of helix-C and helix-D are not affected by Ca2+ binding.

Because the chemical shift of a nucleus is affected by its environment, the chemical-shift changes that occur in NTnC upon Ca2+ binding can be interpreted in terms of structural changes. Figure 8 compiles those changes for each residue, using a relative chemical-shift change that includes 'Ha, 'HN, and I5N. As expected, sites I and I1 are the most affected by the binding of Ca2+ due to conformational changes and the proximate presence of 2 calcium ions. The most affected residue is I37 (Figs. 1, 8), located in the middle of the first 0-strand. The second @-strand chemical shifts are also strongly affected. Inter- estingly, the largest variation observed outside of the binding loops occurs at E41 and L42; this observation is consistent with the NOE variation observed at the beginning of the B-helix


NTnC-apo 663: 15N=107.1

F3 ('HN) -- ~ ~ n C . 2 c . a 663 : 15N=107.1

- 0

1 I K4Oa

G43a2 I

8 L42a T39a

I -

4.2 4.0 3.8 36 3.4

F1 (lH)

Fig. 6 . Comparison of the NOE connectivities of the G43 amide proton for NTnC-apo and NTnC.2Ca in their "N-edited NOESY. The 3D NOESY spectrum of NTnC .2Ca unambiguously indicates the presence of a doN(i - 3, i ) between K40 and (343. This NOE is absent in the spectrum of NTnC-apo, as would be expected from the crystal structure (see text); the expected position of this d a N ( i - 3, i ) is indicated by a box. Note that the "N-edited NOESY spectra of NTnC-apo is plotted at the noise level. This difference in NOE connectivity between NTnC-apo and NTnC .2Ca is one of the indications that the kink of helix-B at E41 is straightened upon Ca2+-binding. The observation of a very weak dolN (i - 4, i) between T39 and G43 in the spectra of NTnC .2Ca also supports this idea (those 2 protons are separated by 5.39 A in the crystal structure of TnC).

(Fig. 6). On average, the B/A helix pair is strongly affected by the calcium-induced structural change, whereas the C-D pair is less perturbed. The N-helix is only weakly perturbed, as is the B-C linker.

Helical content

In order to assess the helical content of NTnC-apo and NTnC.2Ca, we used 2 approaches, one based upon number of residues in helical segments, and a second based upon all residues susceptible of having helical Cp/$ angles. We first directly count the number of residues found in helices according to the data presented in Figures 4 and 5 for the 2 forms of NTnC. The NMR secondary structure of NTnC-apo shows 55 residues in well-defined helices, and a similar number, 56, is found for NTnC.2Ca. Even if this helical residue count is not perfectly accurate, these results do not explain the Ca2+-induced ellipticity increase observed in the CD spectra of NTnC (Li et al., 1994).

The second approach attempts to estimate the proportion of time that each residue spends with helical 4/$ angles. Due to the relationships between a-proton chemical shifts and backbone dihedral angles (Wishart et al., 1991), the helical content of a protein (in terms of Cp, $) can also be approximated by using the chemical-shift differences ( A s ) from the random-coil value (Wishart & Sykes, 1YY4a). In this case, every residue with an upfield shift relative to the random-coil value is accounted for, whether it is in a helix or not. By its nature, this method is quite comparable to the quantification of secondary structure made by CD. We calculated the average A6s from the random-coil values for the 'Has of NTnC-apo and NTnC.2Ca. Sixty-two residues had an upfield shift characteristic of a-helical structure in

Table 3 . Location of a-helices in the N-domain of TnC

Residue range ~~

a-Helix X-ray (apo)" TRIC (apo)b NTnC-apoC NTnC.2Ca'

N M3-Fl3 - T4-Fl3 D5-FI3 A E16-M28 E16-D26d E16-M28 E16-F29

C K55-E64 K55-E64 K55-E64 K55-E64 D F75-A90 F75-M86 F75-R87 F75-R87

a Herzberg and James (1988). Helix limits based on expected '&NHa

Findlay and Sykes (1993). NMR secondary structure of turkey skel-

'Present study. Helix range based on '&NHa, CSI, daN( i - 3, i )

Revision of the NOE data of TRIC suggests that the C-termini of

and NOES from crystal coordinates.

etal TnC fragment (residues 12-87).

NOE, and d ~ ~ / d ~ ~ NOE ratio.

helix-A is M28 (see text).

NTnC-apo, with an average A6 of -0.37 ppm. The NTnC.2Ca assignment revealed 63 upfield-shifted 'Ha , with an average A6 of -0.36 ppm. The regular shift from random-coil chemical shift in helical residues is -0.38 ppm (Wishart & Sykes, 1994a). The crystal structure of TnC has 65 residues with $ between - 10" and -70". The number of upfield-shifted a-protons for NTnC-apo and NTnC.2Ca is in agreement with the crystal structure, especially when considering that truncation at residue YO in NTnC is likely to reduce helix-D by 2-3 residues compared to the whole TnC. As for the results from the residue count, the ASS do not show any increase of helical content in NTnC upon Ca2+-binding.

Discussion

It is well recognized that muscle contraction is triggered by a Ca2+-induced structural change in the regulatory domain of TnC (see review by Crabarek et al., 1992). The results presented in this paper indicate that the secondary structural features of NTnC are not strongly affected by the binding of calcium. For both NTnC-apo and NTnC.2Ca, an N-terminal helix and 2 helix-loop-helix motifs joined by a short &sheet are observed in solution. When compared to the crystal structure of turkey TnC, where the N-domain is found in the apo form (Herzberg & James, 1988), the NMR solution data of NTnC-apo shows only similarities. Even the kink observed by X-ray crystallography at residue 41 of helix-B is observed in solution, as sup- ported by the absence of the G43 daN ( i - 3, i ) , the nonhelical ' H a chemical shift of E41, and the 3JHNHol > 8 Hz.

The secondary structure of NTnC.2Ca, as determined by NMR, is very similar to NTnC-apo (Figs. 4, 5 ) . The binding of calcium to NTnC does not induce any noticeable differences in terms of secondary structure for 3 of the 5 helices. Helix-A ap- pears to be extended by 1 residue at its C-termini in NTnC .2Ca. The length and relative position of helix-B and helix-C in NTnC-2Ca are exactly the same as observed in the NMR secondary structure of the homologous Ca2+-binding protein calmodulin (Ikura et al., 1991). Helices A and D were also found

1970 S.M. Gagne et a/.

-1 NTnC-apo E41 J = 9.2 HZ

J = 9.5 HZ F29

1 NTnC*2Ca

7.61

\ -

B . O j

8.1 4 J = 6.3 Hz u

Fig. 7. Comparison of the 35HNH, coupling constant of E41 for NTnC-apo and NTnC.2Ca in their HMQC-J spectrum. A coupling constant of 9.2 Hz was measured for NTnC-apo using the HMQCJFIT program (see Materials and methods), indicating that E41 has a /3 characteristic 4 angle in the apo form of NTnC. In the spectrum of NTnC.2Ca, no splitting is observed, indicating that E41 has a '&NHn < 6 Hz and that this residue has an cy characteristic 4 angle in the Ca2+ form of NTnC. Note the splitting observed for E88 (3JHNHa = 6.3 Hz), which clearly indicates that E41 has a 3JHNHm < 6 Hz in NTnC.2Ca. Both spectra were processed in the same way (VNMR parameters GF = 0.05, LB = -28).

at the same position in calmodulin as in NTnC .2Ca, even TnC (Satyshur et al., 1988), is removed upon calcium binding, though calmodulin lacks the N-helix. leading to a regular helix from T39 to M48. These changes re-

The geometry of the B-helix is affected. The kink at residue 41, sult in an increase of symmetry between the 2 binding sites upon which was well characterized in the crystal structure of chicken addition of calcium. This is in agreement with the model pro-

posed by Herzberg et al. (1986) and many helix-loop-helix protein structures, which were solved in the Ca2+-saturated state either by X-ray crystallography (see reviews by Strynadka & James, 1989; McPhalen et al., 1991) or NMR (Akke et al., 1992;

ring in the 2 Ca2+-binding loops are not presented here in de- tail and must await the determination of the tertiary structure of NTnC .2Ca, the straightening of helix-B is the major Ca2+- induced structural change we can assess at this time. This change is localized at E41, a position that corresponds to the third most conserved residue in calcium-binding sites of proteins having the helix-loop-helix motif (Marsden et al., 1990). This invariant glu-

nating the Ca2+ ion in a bidentate manner (Strynadka & James,

TnC are -96"/-7", whereas the average angles for that posi-

6 137 Ikura et al., 1992). Because the conformational changes occur-

5

4

3

2 tamate has both oxygen atoms of its carboxylate group coordi-

1 1989). The E41 +/$ dihedral angles in the crystal structure of

0 tion in 11 loops that have Ca2+ bound (from 5 different Ca2+- 6 io is io i5 $0 35 40 45 50 55 60 65 70 75 80 85 90 binding proteins) are -660/-410 (Strynadka & James, 1989).

Sequence The E41 b/$ transition from -96"/-7" to -66"/-41" can, by . .

Fig. 8. Chemical-shift change along the sequence of NTnC induced by itself, account for most of the Ca2+-induced reorientation of the binding of calcium. The chemical-shift change for each residue is the B- and C-helix Proposed by Herzberg at al. (1986) in their obtained by averaging the normalized change of the backbone 'Ha, model. The consequence of this dihedral change is pictured in 'HN, and ''N chemical shifts. The normalized chemical-shift change Figure 9, where the removal of the kink in the &helix of the cry+ for a particular nucleus of each residue is obtained by dividing the ob- tal structure not only puts this helix in an orientation similar to served shift change by the average shift change for all residues. There- fore, a chemical-shift change of is equivalent to the average change the model but also reorients the C-helix, giving a helical arrange- observed for all residues. The horizontal bars indicated the average ment, which is consistent with the model and other CaZ+- change for various segments of NTnC: N, A, B, C, and D represent the binding protein structures solved in the Ca2+ form. 5 respective helices; I and I1 represent the binding loop in site I and 11; important outcome of the present study is the conclu- and L represents the B-C linker. As expected, the largest changes are sion that there is virtually no change in a-helical content in observed in the 2 Ca2+-binding loops. The third largest change is found in helix-B, mainly due to ~ 4 1 and ~ 4 2 ; this observation is one piece of the transition from aP0 to the 2Ca2+ state of the N-domain of evidence for the straightening of that helix upon Ca2+ binding. TnC. Although previous estimates of the magnitude of the


Fig. 9. Schematic representation of the Ca2+-induced straightening of helix-B and its possible impact on the tertiary fold of NTnC. I: The Ca*+-free crystal structure (Herzberg & James, 1988). 11: The predicted Ca2+-saturated structure (Herzberg et at., 1986). 111: Same as I , with helix-B straightened by changing the $I/$ backbone angle of E41 from -961-7 to -661-41. As can be seen by comparing 11 and 111, the straightening of helix-B, which is the predominant Ca2+-induced secondary structure change observed by NMR in NTnC, could account for most of the opening of NTnC upon Ca2+ binding.

Ca2+-induced far-UV CD ellipticity changes attributable to the N-domain transition have been somewhat contradictory (Hincke et al., 1978; Johnson &Potter, 1978; Leavis et al., 1978; Nagy & Gergely, 1979), a recent study (Li et at., 1994) on isolated NTnC has shown clearly a very significant increase (-23%) in the negative values of [B]221nm. This Ca2+-induced ellipticity change must now be attributed to features other than an increase in helical content. Among several possible structural factors af- fecting the relationship between helical content and ellipticity (critically reviewed by Manning, 1989), that attributable to helix-helix interaction seems the most plausible for NTnC. As cited in this review, both experimental and theoretical studies suggest a decrease in the CD intensity of - 10% for 2 antiparallel helices separated by 7-10 A compared to a single a-helix. In the proposed Ca2+-induced structural transition of NTnC (Herzberg et at., 1986), the B- and C-helices move as a unit relative to the N-, A-, and D-helices. As a result, the orientations of helix-B and helix-C to those of A and D change from nearly antiparallel to roughly perpendicular. In other words, the helix packing in the apo form of NTnC would produce a smaller ellipticity signal than expected assuming only contributions from the secondary structure. Other contributions could include the removal of the kink at residue 41 of helix-B (Chen et al., 1974), Ca2+-induced aggregation, and the effects of alterations in the environment of clustered Phe side chains (Manning, 1989). However, these are likely to be of lesser significance than contributions arising from reorientation of helices. In any case the present study shows that it is the Ca2+-induced tertiary structural changes, not the secondary structural ones, that are respon- sible for the CD ellipticity change in the N-domain of TnC.

Materials and methods

NTnC preparation and ”C and ‘’N labeling

For the preparation of NTnC, PCR was performed by using 2 30-mer oligonucleotides and pTZ18.TnC (91 stop) as a template. The construction of pTZ18.TnC (91 stop) was described in Li et at. (1994). The 5‘-end oligonucleotide for PCR was designed

to include codons corresponding to amino acid residues 1-6 of TnC (underlined) and flanked on the 5’ direction by an initiation codon (bold letters) and nucleotides of the pET3a vector (Studier et al., 1990) including the Nde I restriction enzyme site:

5”GAGATATACATATGGCGTCAATGACGGACC-3’.

The 3’-end oligonucleotide includes a sequence (underlined) corresponding to amino acid residues 85-90 of the noncoding strand of the TnC gene. This is preceded in the 5’ direction by a com- plimentary sequence for a stop codon (bold letters) and a BamH I restriction site:

5”AATATGGATCCTAGGCGTCCTCTTTCATCT-3’

PCR was performed in a Perkin Elmer thermal cycler (model 480) using Taq polymerase and conditions essentially as described by H o et al. (1989). The amplified DNA fragment was digested with restriction enzymes Nde I and BamH I and ligated into the corresponding sites of expression vector pET3a plasmid DNA cleaved with the same enzymes (Studier et al., 1990). The ligation mixture was transformed into competent Escherichia coli BL21(DE3) pLys S cells and the entire region of the TnC gene sequenced to ascertain the correctness of the amplification by the Taq polymerase enzyme. The expression of NTnC was carried out in minimal medium utilizing the isopropyl P-D-thiogalacto- pyranoside induction protocol of Studier et at. (1990). The minimal medium consisted of M9 salts as described by Maniatis et al. (1982) with (NH4)2S04 replacing NH4CI. Each 1 L of medium at pH 7.5 contained 6 g Na2HP04, 3 g KH2P04, 0.5 g NaCI, and 1 g (NH&S04 to which was added 2 mL of mineral mixture (1 M MgS04, 0.1 mM FeCI3, and 12.5 mM ZnS04), 1 mL of 100 mM CaCI2, and 1 mL of vitamin mixture (0.1 g/100 mL each of biotin, choline chloride, folic acid, niacinamide, D-panto- thenic acid, and pyridoxal chloride, 0.5 g/100 mL thiamine, 0.01 g/100 mL riboflavin, all in H20). Glucose (3 g) dissolved in 20 m L H 2 0 was added with supplements of ampicillin and chloramphenicol to final concentrations of 100 pg/mL and

1972 S.M. Gagne et al.

25 pg/mL, respectively. The above additives were sterilized sep- arately by filtration. For expression of uniformly enriched [I5N]- and/or [I3C]NTnC, the (NH4),S04 and/or glucose of the medium were replaced with [I5N] (NH4),S04 (99.9 atom 070)

and/or [13C] glucose (99.0 atom Vo), respectively, both pur- chased from Isotec Inc. Purification of NTnC followed the previously published procedure for fusion TnC (Golosinska et al., 1991). During expression in E. coli, the N-terminal Met, corresponding to the initiation codon, is cleaved off, leaving Ala-1 as the N-terminal residue. Overall recovery of NTnC was - 100 mg/L of growth culture. To make the calcium-free sample (NTnC-apo), decalcification was accomplished by using (3-25 gel filtration: 10 mg of the sample dissolved in 1 mL of 0.5 M EDTA, pH 8.0, was applied to a 160-mL (1.5 x 90 cm) G-25 gel filtration column and eluted with 25 mM NH4HC03. Frac- tions containing protein were pooled, lyophilized, dissolved in Ca2+-free water, and lyophilized again to volatilize all of the NH4HC03. For 2D- and 3D-NMR experiments, 5 mg or 10 mg of the NTnC-apo sample were dissolved in 0.5 mL of 50 mM KCI, pH 6.6, in either H20 (H20:D20, 9: 1) or D20 (99.95%), to give final concentrations of 1 mM or 2 mM, respectively. The Ca2+-saturated NMR sample was prepared as follows: 2 mM NTnC, 4.2 mM CaCI,, 100 mM KCI, pH 6.7, in either a 90: 10 H20:D20 mixture or a D20 solution (for HCACO experiment only).

NMR spectroscopy

All experiments were carried out on a Varian Unity-600 NMR spectrometer operating at a temperature of 30 "C. Assignments of 'H- and "N-resonances of NTnC-apo were carried out by acquiring the following sets of 2D- and 3D-NMR data: (1) DQF- COSY in H 2 0 (Rance et al., 1983); (2) TOCSY (50 ms) in H 2 0 (Braunschweiler & Ernst, 1983; Davis & Bax, 1985); (3) NOESY (40, 80, 120, 160, 200, 250 ms) in H 2 0 (Jeener et al., 1979; Macura & Ernst, 1980); (4) DQF-COSY in DzO; ( 5 ) TOCSY (50 ms) in DzO; (6) NOESY (150 ms) in DzO; (7) 2D( 15N-'HJ HMQC (Bax et al., 1983); (8) 3D-I5N-edited NOESY (150ms) in H,O (Kay et al., 1989); and (9) 3D-I'N-edited TOCSY (75 ms) in H 2 0 (Marion et al., 1989a). The following experiments were used for the study of NTnC.2Ca: (10) 2D(I5N-IH] HMQC; (1 1) 3D-HNCA (Grzesiek & Bax, 1992); (12) 3D-HNCO (Grze- siek & Bax, 1992); (13) 3D-HNCOCA (Grzesiek & Bax, 1992); (14) 3D-HCACO (Powers et al., 1991); (15) 3D-15N-edited NOESY (150 ms); and (16) 3D-I5N-edited TOCSY (70 ms). The

J H N - H a coupling constants were obtained from HMQC-J experiments (Kay & Bax, 1990) with the following numbers of complex points and spectral widths: ( 'H) 1,024 and 8,000 Hz; (15N) 384 and 1,800 Hz. For the spectra recorded in H20, water suppression was achieved by presaturation (1 .O-1.2 s). The water suppression for (3) was achieved using the SCUBA pulse sequence (Brown et al., 1988).

In all experiments, the ' H acquisition dimension was centered at the water frequency (4.67 ppm) with a spectral width of 13.33 ppm (5.00 ppm in the HCACO case). The indirectly de- tected "N carrier position and spectral width were either 119.22 ppm and 28.01 ppm (HNCA, HNCOCA, and HNCO), or 117.44 ppm and 23.03 ppm (15N-edited NOESY and I5N- edited TOCSY). The corresponding parameters for the I3Ca were either 55.83 ppm and 30.25 ppm (HNCA, HCACO), or

3

56.1 1 and 24.18 (HNCOCA). Finally, the I3CO carrier position and spectral width were either 177.11 ppm and 11.93 ppm (HNCO), or 176.34 pprn and 18.05 ppm (HCACO). For the HCACO experiment, the number of complex points acquired was 512 in F3 ('H); for all other 3D experiments, 1,024 complex points were acquired in F3 ('H). For the HNCA and HNCOCA experiments, there were 48 and 32 complex points acquired in F2 (I3Ca) and F1 (15N), respectively; the time domain was increased by 32 complex points in both F2 and F1 di- mensions by means of linear prediction and, after zero-filling, the final spectra consisted of 1,024 x 256 x 64 points. The number of complex points acquired for the HNCO was 64 in F2 (I3CO) and 32 in F1 (I5N); 32 additional complex points were predicted in the F1 dimension to finally generate, after zero- filling, a 1,024 X 128 X 64-point spectrum. For the HCACO, we acquired 35 complex points in F2 (13CO) and 28 complex points in F1 (I3Ca); both F2 and F1 time domains were sub- sequently extended using linear prediction to give a 512 x 128 x 64-point spectrum. In the case of the I5N-edited NOESY and "N-edited TOCSY, 32 and 128 complex points were acquired in F2 ("N) and F1 ('H), respectively; 32 complex points were added to the F2 dimension by linear prediction to obtain a final 1,024 X 64 X 256-point spectrum. For all experiments except the HCACO, a postacquisition solvent suppression by convolution of the time-domain data was applied prior to Fourier transform (Marion et al., 1989b). In most of thecases, a 60"-shifted squared sinebell was applied in F1 and F2, and a 90"-shifted sinebell function in F3. After Fourier transform of the F3 dimension, parts of the spectra without resonances were discarded, when possible, prior to the processing of F2 and F1, thus reducing the size of the final spectra by 50% or 75%. The HMQC-Js were processed to a final size of 2,048 x 2,048 points with a 90"- shifted sinebell in F2 ('H) and the following filtering VNMR parameters in F1 (I5N): GF = 0.06 (NTnC-apo), GF = 0.05 (NTnC.2Ca), LB = -8, -10, -12, -14, -16, -18, -20, -22, -24, -26, -28. The splittings at various LBs were fitted with the HMQCJFIT software (Goodgame & Geer, 1993) to obtain the 3 J ~ ~ ~ , coupling constants. The HMQCJFIT program was first modified to properly handle the VNMR weighting functions.

Processing of the 2D and 3D data sets was accomplished using either the VNMR software (VNMR 4. IA, Varian, Palo Alto, California) or NMRPIPE (F. Delaglio, NIDDK, NIH, Maryland, unpubl.). When used within the VNMR software, extension of the time domain was achieved using the linear prediction algorithm lpfft. Automatic peak picking of the transformed 3D spectra was achieved using the CAPP program (Gar- rett et al., 1991). Because CAPP is run at the noise level and every spectrum contains imperfections, a number of false peaks are usually picked. Most of those peaks were removed automat- ically by using the in-house program PPFILT (L. Willard & S. GagnC, PENCE, University of Alberta, unpubl). PPFILT fil- ters the peak list of a 3D spectrum through a high signal-to-noise 2D spectrum (or another 3D spectrum). For example, the 3D- 15N-edited NOESY has been filtered through a 2D-I5N-HMQC. In the last step of the peak-picking, we used the interactive graphic-based program PIPP (Garrett et al., 1991) to visualize the output of CAPP/PPFILT and edit the peak-pick table. Both processing and peak-picking were accomplished on a Sun Sparc2 workstation. The chemical-shift reference used for 'H and 13C is 2,2-dimethyl-2-silapentane-5-sulfonic acid. I5N chemical shifts are reported relative to external acidic NH4Cl (24.93 pprn).


Acknowledgments

We are grateful to Leigh Willard and Robert Boyko for providing various programs essential to this study. We thank Dan Garrett and Frank Delaglio (NIH) for supplying the excellent programs CAPP, PIPP, and NMRPipe. We also thank Dr. Frank D. Sonnichsen and Carolyn M. Slupsky for stimulating discussions and critical reading of the present paper, Carolyn M. Slupsky for providing the protocol for I3C and 15N labeling, Dr. Michael Bloemendal for bringing a useful reference (Man- ning, 1989) to our attention, and Dr. Michael N.G. James for providing us with the coordinates of the model of Ca2+-loaded TnC. Financial support from the MRC and Fonds FCAR is gratefully acknowledged.

References

Akke M, Drakenberg T, Chazin WJ. 1992. Three-dimensional solution structure of Ca2+-loaded porcine calbindin Dgk determined by nuclear magnetic resonance spectroscopy. Biochemistry 31:1011-1020.

Bax A, Griffey RH, Hawkins BL. 1983. Correlation of proton and nitrogen- 15 chemical shifts by multiple quantum NMR. JMagn Reson 55:301-315.

Braunschweiler L, Ernst RR. 1983. Coherence transfer by isotropic mixing: Application to proton correlation spectroscopy. JMagn Reson53521-528.

Brown SC, Weber PL, Mueller L. 1988. Toward complete 'H NMR spectra in proteins. JMagn Reson 77:166-169.

Carew EB, Leavis PC, Stanley HE, Gergely J. 1980. A laser raman spectroscopic study of Ca2+ binding to troponin C. Biophys J30:351-358.

Chen YH, Yang TY, Chan KM. 1974. Determination of the secondary structures of proteins by circular dichroism and optical rotary dispersion. Bio- chemistry 13:3350-3359.

Clore GM, Gronenborn AM. 1991. Structures of larger proteins in solution:

252:1390-1399. Three- and four-dimensional heteronuclear NMR spectroscopy. Science

Davis DG, Bax A. 1985. Assignment of complex 'H NMR spectra via two- dimensional homonuclear Hartmann-Hahn spectroscopy. J A m Chem Soc 107:2820-2821,

Evans JS, Levine BA, Leavis PC, Gergely J, Grabarek Z, Drabikowski W. 1980. Proton magnetic resonance studies on proteolytic fragments of troponin-C. Structural homology with the native molecule. Biochim Bio- phys Acta 623:lO-20.

Farah CS, Miyamoto CA, Ramos CHI, Silva ACR, Quaggio KF, Smillie LB, Reinach FC. 1994. Structural and regulatory functions of the N- and C-terminal regions of skeletal muscle troponin 1. J Biol Chem 269: 5230-5240.

Findlay WA, Sonnichsen FD, Sykes BD. 1994. Solution structure of the TRIC fragment of skeletal muscle troponin-C. J Biol Chem 269:6773-6778.

Findlay WA. Sykes BD. 1993. 'H-NMR resonance assignments, secondary structure, and global fold of the TR,C fragment of turkey skeletal troponin C in the calcium-free state. Biochemistry 32:3461-3467.

Garrett DS, Powers R, Gronenborn AM, Clore GM. 1991. A common sense approach to peak-picking in two-, three-, and four-dimensional spectra using automatic computer analysis of contour diagrams. J Magn Reson 95:214-220.

Golosinska K, Pearlstone JR, Borgford T, Oikawa K, Kay CM, Carpenter MR, Smillie LB. 1991. Determination of and corrections to sequences of turkey and chicken troponins-C. J Biol Chem 266:15797-15809.

Goodgame MM, Geer SM. 1993. HMQCJFIT-software for analyzing HMQC-J spectra. J Magn Reson Ser A 102:246-248.

Grabarek 2, Tao T, Gergely J. 1992. Molecular mechanism of troponin-C function. /Muscle Res Cell Motil 13:383-393.

Grzesiek S, Bax A. 1992. Improved 3D triple-resonance NMR techniques applied to a 31 kDa protein. J Magn Reson 96:432-440.

Herzberg 0, James MNG. 1988. Refined crystal structure of troponin C from turkey skeletal muscle at 2.0 A resolution. J Mol Biol203:761-779.

Herzberg 0, Moult J, James MNG. 1986. A model for the Ca2+-induced

Hincke MT, McCubbin WD, Kay CM. 1978. Calcium-binding properties of conformational transition of troponin C. JBiol Chem 261:2638-2644.

cardiac and skeletal troponin C as determined by circular dichroism and ultraviolet difference spectroscopy. Can J Biochem 56:384-395

Ho SN, Hunt HD, Horton RM, Pullen JK, Pease LR. 1989. Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77:51-59.

Ikura M, Clore GM, Gronenborn AM, Zhu G, Klee CB, Bax A. 1992. So-

sional NMR. Science 256532-638. lution structure of a calmodulin-target peptide complex by multidimen-

Ikura M, Kay LE, Bax A. 1990. A novel approach for sequential assignment of 'H, I3C, and "N spectra of larger proteins: Heteronuclear triple- resonance three-dimensional NMR spectroscopy. Application to calmed- d in . Biochemistry 29:4659-4667.

Ikura M, Spera S, Barbato G , Kay LE, Krinks M, Bax A. 1991. Secondary structure and side-chain 'H and I3C resonance assignments of calmodulin in solution by heteronuclear multidimensional NMR spectroscopy. Biochemistry 30:9216-9228.

Jeener J, Meier BH, Bachmann P, Ernst RR. 1979. Investigation of exchange processes by two dimensional NMR spectroscopy. J Chem Phys 71: 4546-4553.

Johnson JD, Potter JD. 1978. Detection of two classes of CaZ+ binding sites in troponin C with circular dichroism and tyrosine fluorescence. J Biol Chem 253:3775-3777.

Karplus M. 1963. Vicinal proton coupling in nuclear magnetic resonance. J Am Chem Soc 85 :2870-287 1.

Kay LE, Bax A. 1990. New methods for the measurement of NH-CaH coupling constants in 15N-labeled proteins. J Magn Reson 86:llO-126.

Kay LE, Marion D, Bax A. 1989. Practical aspects of 3D heteronuclear NMR of proteins. J Magn Reson 84:72-84.

Krudy GA, Brito RMM, Putkey JA, Rosevear PR. 1992. Conformational changes in the metal-binding sites of cardiac troponin C induced by calcium binding. Biochemistry 31:1595-1602.

Leavis PC, Gergely J. 1984. Thin filament proteins and thin filament-linked regulation of vertebrate muscle contraction. CRC Crif Rev Biochem

Leavis PC, Nagy B, Lehrer SS, Bialkowska H, Gergely J. 1980. Terbium binding to troponin C: Binding stoichiometry and structural changes induced in the protein. Arch Biochem Biophys 200:17-21.

Leavis PC, Rosenfeld SS, Gergely J, Grabarek Z, Drabikowski W. 1978. Pro- teolytic fragments of troponin C. J Biol Chem 2535452-5459.

Levine BA, Mercola D, Coffman D, Thornton JM. 1977. Calcium binding by troponin-C. A proton magnetic resonance study. J Mol Biol 115: 743-760.

Levine BA, Thornton JM, Fernandes R, Kelly CM, Mercola D. 1978. Com- parison of the calcium- and magnesium-induced structural changes of

535: 11-24. troponin-C. A proton magnetic resonance study. Biochim Biophys Acta

Li MX, Chandra M, Pearlstone JR, Racher KI, Trigo-Gonzalez G, Borg- ford T, Kay CM, Smillie LB. 1994. Properties of isolated recombinant N and C domains of chicken troponin C. Biochemistry 33:917-925.

Macura S , Ernst RR. 1980. Elucidation of cross relaxation in liquids by two- dimensional NMR spectroscopy. Mol Phys 4 1 9 - 1 17.

Maniatis T, Fritsch EF, Sambrook J. 1982. Molecular cloning: A laboratory manual. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory.

Manning MC. 1989. Underlying assumptions in the estimation of secondary structure content in proteins by circular dichroism spectroscopy- A critical review. J Pharm Biomed Analysis 7:1103-1119.

Marion D, Driscoll PC, Kay LE, Wingfield PT, Bax A, Gronenborn AM, Clore GM. 1989a. Overcoming the overlap problem in the assignment of ' H NMR spectra of larger proteins by use of three-dimensional heteronuclear 'H-I5N Hartmann-Hahn-multiple quantum coherence and

tion to interleukin 18. Biochemistry 28:6150-6156. nuclear overhauser-multiple quantum coherence spectroscopy: Applica-

Marion D, Ikura M, Bax A. 1989b. Improved solvent suppression in one- and two-dimensional NMR spectra by convolution of time-domain data. J Magn Reson 84:425-430.

Marsden BJ, Shaw GS, Sykes BD. 1990. Calcium binding proteins. Eluci- dating the contributions to calcium affinity from an analysis of species variants and peptide fragments. Biochem Cell Biol68:587-601.

McPhalen CA, Strynadka NCJ, James MNG. 1991. Calcium-binding sites in proteins: A structural perspective. Adv Protein Chem 42:77-144.

Nagy B, Gergely J. 1979. Extent and localization of conformational changes in troponin C caused by calcium binding. JBiol Chem 254:12732-12737.

Ohtsuki I , Maruyama K. Ebashi S . 1986. Regulatory and cytoskeletal proteins of vertebrate skeletal muscle. Adv Protein Chem 38:l-67.

Osapay K, Case DA. 1991. A new analysis of proton chemical shifts in proteins. J A m Chem Soc 113:9436-9444.

Pardi A, Billeter M, Wiithrich K. 1984. Calibration of the angular depen- dence of the amide proton-Ca coupling constants, J H N ~ , in a globular protein. Use of 3 J ~ ~ , for identification of helical secondary structure. JMoI Biol 180:741-751.

Powers R, Gronenborn AM, Clore GM, Bax A. 1991. Three-dimensional triple-resonance NMR of I3C/l5N-enriched proteins using constant-time evolution. J Magn Reson 94:209-213.

Rance M, S~crensen OW, Bodenhausen G, Wagner G, Ernst RR, Wiithrich K . 1983. Improved spectral resolution in COSY 'H NMR spectra of proteins via double quantum filtering. Biochem Biophys Res Commun 117:479-485.

Richards FM, Kundrot CE. 1988. Identification of structural motifs from protein coordinate data: Secondary structure and first-level supersecond-

Satyshur KA, Pyzalska D, Greaser M, Rao ST, Sundaralingam M. 1994. ary structure. Proteins Struct Funct Genet 3:71-84.

16~235-305.

1974 S.M. Gagne et ai.

Structure of chicken skeletal muscle troponin C at 1.78 A resolution. Acta Cryslallogr D 50:40-49.

Satyshur K A , Rao ST, Pyzalska D, Drendel W, Greaser M, Sundaralingam M. 1988. Refined structure of chicken skeletal muscle troponin C in the two-

Seamon KB, Hartshorne DJ, Bothner-By AA. 1977. Ca2+ and Mg2+ depen- calcium state at 2-A resolution. J B i d Chem 263:1628-1647.

dent conformations of troponin C as determined by 'H and I9F nuclear magnetic resonance. Biochemistry 16:4039-4046.

Strynadka NCJ, James MNG. 1989. Crystal structures of the helix-loop-helix calcium-binding proteins. Annu Rev Biochem 58:951-998.

Studier FW, Rosenberg AH, Dunn JJ, Dubendorff JW. 1990. Use of T7 RNA polymerase to direct expression of cloned genes. Methods Enzymol

Tsuda S, Hasegawa Y, Yoshida M, Yagi K, Hikichi K. 1988. Nuclear magnetic resonance study on rabbit skeletal troponin C: Calcium-induced conformational change. Biochemislry 27:4120-4126.

185:60-89.

Tsuda S, Ogura K, Hasegawa YM, Yagi K, Hikichi K. 1990. 'H NMR study of rabbit skeletal muscle troponin C: Mg*+-induced conformational change. Biochemistry 29:4951-4958.

Wishart DS, Sykes BD. 1994a. Chemical shifts as a tool for structure de-

Wishart DS, Sykes BD. 1994b. The 13C chemical shift index: A simple termination. Methods Enzymo/239:363-392.

method for the identification of protein secondary structure using 13C

Wishart DS, Sykes BD, Richards FM. 1991. Relationship between nuclear chemical shift data. J Biomol NMR 4:171-180.

Biol 222:311-333. magnetic resonance chemical shift and protein secondary structure. JMol

Wiithrich K. 1986. NMR of proteins and nucleic acids. New York: John Wiley & Sons.

Zot AS, Potter JD. 1987. Structural aspects of troponin-tropomyosin regulation of skeletal muscle contraction. Annu Rev Biophys Biophys Chem 16:535-559.

quantification of the calcium-induced secondary structural changes

Documents