Article Information

• PDF (186 kb)
• Full Text with Large Figures
• Cited by in Scopus (6)

PubMed

• Articles by Tharin M.A. Blumenschein
• Articles by Brian D. Sykes

Related Articles

• Energetics of the binding of calcium and troponin I to tropo...

  Energetics of the binding of calcium and troponin I to troponin C from rabbit skeletal muscle Biophysical Journal, Volume 48, Issue 5, 1 November 1985, Pages 727-739 C.K. Wang and H.C. Cheung Abstract We determined the free energy of interaction between rabbit skeletal troponin I (TNI) and troponin C (TNC) at 10 degrees and 20 degrees C with fluorescently labeled proteins. The sulfhydryl probe 5-iodoacetamidoeosin (IAE) was attached to cysteine (Cys)-98 of TNC and to Cys-133 of TNI, and each of the labeled proteins was titrated with the other unlabeled protein. The association constant for formation of the complex between labeled TNC (TNC*) and TNI was 6.67 X 10(5) M-1 in 0.3 M KCl, and pH 7.5 at 20 degrees C. In the presence of bound Mg2+, the binding constant increased to 4.58 X 10(7) M-1 and in the presence of excess of Ca2+, the association constant was 5.58 X 10(9) M-1. Very similar association constants were obtained when labeled TNI was titrated with unlabeled TNC. The energetics of Ca2+ binding to TNC* and the complex TNI X TNC* were also determined at 20 degrees C. The two sets of results were used to separately determine the coupling free energy for binding TNI and Mg2+, or Ca2+ to TNC. The results yielded a total coupling free energy of -5.4 kcal. This free energy appeared evenly partitioned into the two species: TNI X TNC(Mg)2 or TNI X TNC(Ca)2, and TNI X TNC(Ca)4. The first two species were each stabilized by -2.6 kcal, with respect to the Ca2+ free TNI X TNC complex, and TNI X TNC(Ca)4 was stabilized by -2.8 kcal, respect to TNI X TNC(Ca)2 or TNI X TNC(Mg)2. The coupling free energy was shown to produce cooperatively complexes formed between TNI and TNC in which the high affinity sites were initially saturated as a function of free Ca2+ to yield TNI X TNC(Ca)4. This saturation occurred in the free Ca2+ concentration range 10(-7) to 10(-5) M. The cooperative strengthening of the linkage between TNI and TNC induced by Ca2+ binding to the Ca2+-specific sites of TNC may have a direct relationship to activation of actomyosin ATPase. The nature of the forces involved in the Ca2+-induced strengthening of the complex is discussed. Abstract | PDF (1221 kb)

• The Structure of Lethocerus Troponin C: Insights into the Me...

  The Structure of Lethocerus Troponin C: Insights into the Mechanism of Stretch Activation in Muscles Structure, Volume 15, Issue 7, 18 July 2007, Pages 813-824 Gianfelice De Nicola, Christoph Burkart, Feng Qiu, Bogos Agianian, Siegfried Labeit, Stephen Martin, Belinda Bullard and Annalisa Pastore Summary To gain a molecular description of how muscles can be activated by mechanical stretch, we have solved the structure of the calcium-loaded F1 isoform of troponin C (TnC) from and characterized its interactions with troponin I (TnI). We show that the presence of only one calcium cation in the fourth EF hand motif is sufficient to induce an open conformation in the C-terminal lobe of F1 TnC, in contrast with what is observed in vertebrate muscle. This lobe interacts in a calcium-independent way both with the N terminus of TnI and, with lower affinity, with a region of TnI equivalent to the switch and inhibitory peptides of vertebrate muscles. Using both synthetic peptides and recombinant proteins, we show that the N lobe of F1 TnC is not engaged in interactions with TnI, excluding a regulatory role of this domain. These findings provide insights into mechanically stimulated muscle contraction. Summary | Full Text | PDF (1654 kb)

• Toward Protein Structure In Situ: Comparison of Two Bifuncti...

  Toward Protein Structure In Situ: Comparison of Two Bifunctional Rhodamine Adducts of Troponin C Biophysical Journal, Volume 93, Issue 3, 1 August 2007, Pages 1008-1020 Olivier Julien, Yin-Biao Sun, Andrea C. Knowles, Birgit D. Brandmeier, Robert E. Dale, David R. Trentham, John E.T. Corrie, Brian D. Sykes and Malcolm Irving Abstract As part of a program to develop methods for determining protein structure in situ, sTnC was labeled with a bifunctional rhodamine (BR or BSR), cross-linking residues 56 and 63 of its C-helix. NMR spectroscopy of the N-terminal domain of BSR-labeled sTnC in complex with Ca and the troponin I switch peptide (residues 115–131) showed that BSR labeling does not significantly affect the secondary structure of the protein or its dynamics in solution. BR-labeling was previously shown to have no effect on the solution structure of this complex. Isometric force generation in isolated demembranated fibers from rabbit psoas muscle into which BR- or BSR-labeled sTnC had been exchanged showed reduced Ca-sensitivity, and this effect was larger with the BSR label. The orientation of rhodamine dipoles with respect to the fiber axis was determined by polarized fluorescence. The mean orientations of the BR and BSR dipoles were almost identical in relaxed muscle, suggesting that both probes accurately report the orientation of the C-helix to which they are attached. The BSR dipole had smaller orientational dispersion, consistent with less flexible linkers between the rhodamine dipole and cysteine-reactive groups. Abstract | Full Text | PDF (360 kb)

• …more

Copyright © 2006 The Biophysical Society. All rights reserved.
Biophysical Journal, Volume 90, Issue 7, 2436-2444, 1 April 2006

doi:10.1529/biophysj.105.076216

Muscle and Contractility

Dynamics of the C-Terminal Region of TnI in the Troponin Complex in Solution

Tharin M.A. Blumenschein^*, Deborah B. Stone^†, ‡, Robert J. Fletterick^†, Robert A. Mendelson^‡, 1 and Brian D. Sykes^*, ,  

^* CIHR Group in Structure and Function and Department of Biochemistry, University of Alberta, Edmonton, Alberta T6G 2H7, Canada
^† Department of Biochemistry and Biophysics, University of California, San Francisco, California 94143-2240
^‡ Cardiovascular Research Institute, University of California, San Francisco, California 94143-2240

Address reprint requests to Brian D. Sykes, Tel: 780-492-5460; Fax: 780-492-0886.

¹ This article is dedicated to deceased colleague Robert A. Mendelson.

In skeletal and cardiac muscle, contraction is regulated by changes in the Ca²⁺ concentration inside the cell. These changes are sensed by the troponin (Tn)¹ complex and transmitted to the other components of the contractile unit. The troponin complex is formed by three subunits, troponin C (TnC), troponin I (TnI), and troponin T (TnT) (for reviews, see Sykes 1, Gordon et al. 2, Tobacman 3, and Farah and Reinach 4). TnC is an EF-hand protein that can bind four Ca²⁺ ions and is responsible for sensing the increase in calcium concentration. Two of its EF-hand sites, located in the C-domain, are high-affinity sites, capable of binding both Ca²⁺ and Mg²⁺. These sites are occupied both during contraction and in the relaxed state. The other two EF-hand sites, located in the N-domain, are Ca²⁺-specific and have lower affinity, being occupied when the intracellular Ca²⁺ levels rise during contraction, and empty in the relaxed state. This change in occupancy changes the interactions between TnC and TnI, and these conformational changes are transmitted to the rest of the thin filament, regulating muscle contraction.

TnI interacts with actin, TnC, and TnT. The interaction between TnI and actin inhibits muscle contraction and is independent of Ca²⁺ in the absence of TnC 4. Two regions of TnI are capable of interacting with actin: the inhibitory region, comprising residues 104–115, and the C-terminal region, residues 148 and 182 5. TnC and TnI interact in a Ca²⁺-dependent manner. In the relaxed state, Ca²⁺ concentrations inside the cell are low, and the only interaction between TnC and TnI is the structural interaction between the C-domain of TnC and the N-terminal region of TnI. When Ca²⁺ levels inside the cell rise, the N-domain of TnC interacts with the “switch” region of TnI (residues 116–131 in skeletal TnI), removing the interactions between TnI and actin and, consequently, the inhibition. The interaction between TnI and TnT is a structural coiled-coil, involving residues 58–102 from TnI and 200–245 from TnT, in chicken fast skeletal muscle 6.

Besides TnI, TnT also interacts with tropomyosin and actin, keeping the troponin complex anchored to the thin filament in the absence of Ca²⁺ and transmitting the conformational changes from the troponin complex to the rest of the regulatory unit (tropomyosin and seven monomers of actin).

We have previously reported dynamics measurements of TnC in the troponin complex 7. Here we study the dynamics of TnI in the troponin complex. The skeletal muscle isoform of TnI has 182 amino acids, but only the first 143 residues can be seen in the crystal structure of the troponin complex 6. The “missing” C-terminal region of TnI interacts with actin in muscle and is necessary to achieve wild-type levels of inhibition in the presence of TnC but absence of Ca²⁺5. A nascent structure of this region has recently been determined 8.

A reconstituted troponin complex containing whole TnC, whole TnI, and the T2 domain of TnT was used to study the dynamics of TnI in the troponin complex. The T1 domain of TnT was not included because it causes aggregation of the complex 9 and it does not interact with TnC and TnI 2. A group of peaks was much more intense than the others, making it difficult to analyze both groups of peaks at the same time. For that reason, we focused on the intense peaks, which were assigned to the key actin binding C-terminal region of TnI, comprising residues 136–182.

A uniformly deuterated deletion mutant of chicken skeletal muscle TnT (isoform TnT-3), corresponding to the T2 domain, was expressed and purified as described 9. Uniformly deuterated chicken skeletal TnC was prepared as described previously 10. Cysteine-free chicken fast skeletal muscle TnI was expressed from a plasmid kindly provided by Dr. Larry Smillie 11. [²H, ¹³C, ¹⁵N]-labeled cysteine-free TnI was expressed as previously described for TnC 12. Isolated inclusion bodies were resuspended in 8M urea, 50mM TrisCl pH 7.0, 1mM EDTA. After clarification, the solution was loaded on a DEAE-Sephadex A-25 column equilibrated with the same buffer. The flow-through was collected, applied to a CM-Sepharose column equilibrated with the same buffer, and eluted with a 0–300mM NaCl gradient. For further purification, the fractions containing TnI were dialyzed against the same buffer, with no salt and at pH 8.0, and applied to a CM-Sepharose column equilibrated with the same buffer at pH 8.0. Fractions containing TnI were dialyzed against 1% formic acid and lyophilized. The troponin complex containing these subunits was assembled and purified as described 9. The resulting complex was concentrated, dialyzed against 100mM NH₄HCO₃, 0.1mM dithiothreitol (DTT), and lyophilized.

For the Ca²⁺ NMR sample, 12.2mg of freeze-dried complex were dissolved in 200μL of buffer containing 10mM Hepes, 250mM KCl, 5mM CaCl₂, 2mM DTT, 0.03% sodium azide, and 0.2mM 2,2-dimethyl-2-silapentane-5-sulfonic acid, in 90% H₂O, 10% D₂O. The pH was adjusted to 6.8. The solution was filtered using a microcentrifuge filter with 0.22μm pore size, and transferred to an NMR tube (3mm diameter). For the EGTA sample, 12.2mg of freeze-dried complex were dissolved in 540μL of buffer containing 10mM Hepes, 250mM KCl, 10mM EGTA, 5mM MgCl₂, 2mM DTT, 0.03% sodium azide, and 0.2mM 2,2-dimethyl-2-silapentane-5-sulfonic acid, in 90% H₂O, 10% D₂O. The pH was adjusted to 7.1. The solution was filtered using a microcentrifuge filter with 0.22μm pore size and transferred to an NMR tube (5mm diameter). Expected free and bound concentrations of each ion and binding site were calculated with MaxChelator 2.50 13, as described 7. This program can be obtained from http://www.stanford.edu/∼cpatton/maxc.html. In the conditions used, <10% of the N-domain sites are expected to be occupied in the presence of EGTA/Mg²⁺.

For chemical shift assignments of TnI in the troponin complex, ¹H, ¹⁵N-TROSY-HSQC, three-dimensional (3D)-TROSY-HNCACB, and 3D-TROSY-HNCOCA spectra were acquired in a Varian INOVA 800MHz spectrometer using BioPack (update version 2005-02-21, Varian, Palo Alto, CA) pulse sequences. For the backbone amide ¹⁵N relaxation measurements, spectra were acquired on Varian INOVA 600MHz and 800MHz spectrometers. For relaxation time measurements, we used the T₁ or T₂ options in the gNhsqc BioPack (Varian) pulse sequence, based on pulse sequences developed by Kay and co-workers 14,15. These pulse sequences use ¹H 180° pulses during the relaxation period to eliminate the effects of cross correlation between dipolar and chemical shift anisotropy (CSA) relaxation mechanisms and were chosen to facilitate comparison with previously published data. For ¹H-¹⁵N nuclear Overhsauser effect (NOE) heteronuclear relaxation measurements, the pulse sequence used was gNnoe (BioPack, Varian), based on a pulse sequence developed by Kay and co-workers 15, with a delay inserted before each transient to allow for recovery of magnetization without unblanking of the amplifier. Relaxation delays were 10, 30, 50, 70, 90, 100, and 130ms for T₂ and 0.01, 0.21, 0.41, 0.61, 0.81, 1.01, and 1.21s for T₁. The delay for nuclear Overhauser effect (NOE) build-up in the gNnoe experiments was 4s, with a recovery delay of 6s. In the no-nOe experiments, both delays were replaced by a recovery delay of 10s. Two-dimensional (2D) NMR spectra were collected with 2048 (t₂)×96 (t₁) complex points in TROSY mode and from 768 (t₂)×96 (t₁) to 512 (t₂)×58 (t₁) complex points when decoupling the indirectly detected dimension during acquisition. 3D NMR spectra were collected with 2048 (t₃)×82 (t₁)×32 (t₂) (for HNCACB) or 2048 (t₃)×64 (t₁)×32 (t₂) (for HNCOCA) complex points. All spectra were acquired at 30°C. All experimental free induced decays (FIDs) were processed with NMRPipe 16 and analyzed with NMRView 17. Linear prediction for up to half the number of experimental points was used in the indirectly detected dimension of experiments used for assigning. No linear prediction was used in the relaxation experiments. Data were zero-filled to the next power of 2 and multiplied by a sine-bell apodization function shifted by 90° before Fourier transformation. Data used for chemical shift assignments were also multiplied by a Gaussian function before Fourier transformation. Backbone amide ¹⁵N T₁ and T₂ relaxation data were fitted to a two-parameter exponential decay using the “Rate Analysis” function of NMRView. (The error bars in the center and bottom panels of Fig. 3 represent the confidence intervals calculated by NMRView, using the noise level as an estimate for the standard deviation in intensity.) The intensity of the peaks in the nOe and no-nOe spectra was measured in NMRView, and the ratio between them was calculated using an in-house tk/tcl script. (The error bars in the top panel of Fig. 3 were calculated from the noise level in each spectrum.)

Reduced spectral density mapping analysis was performed according to the equations 18

Skeletal troponin complex, reconstituted with ²H-TnC, [²H, ¹³C, ¹⁵N]-TnI, and ²H-TnT-T2, was used to acquire NMR spectra of TnI in the troponin complex in the presence of Ca²⁺ and EGTA/Mg²⁺ (Fig. 1). Peaks corresponding to the C-terminal region of TnI were much sharper and stronger than any other region of the protein, offering a unique opportunity to study the dynamics of that region. Peak widths are inversely proportional to the transverse relaxation time (Δν=1/πT₂), and sharper peaks indicate that this region is more flexible than the remainder of the protein. Since line widths cannot be accurately measured in 2D spectra, direct relaxation measurements are a more accurate way to study protein dynamics.

Display large version of this figure

Figure 1

Comparison between 800MHz ¹H, ¹⁵N-HSQC spectra of TnI in the troponin complex in the presence of Ca²⁺ (in black) and EGTA/Mg²⁺ (in red). Most of the visible peaks are not significantly affected by the presence or absence of Ca²⁺. Some assignments are indicated in the figure. The residue most affected by the Ca²⁺ state is 137, which moves 0.2ppm in the ¹H dimension as indicated by the dashed line.

Fig. 1 shows 2D ¹H, ¹⁵N-HSQC NMR spectra of TnI in the troponin complex, with the vertical scale adjusted to show only the intense peaks. The high intensity of the peaks corresponding to the C-terminal region of TnI, relative to the size of the troponin complex and the intensity of peaks from other regions of the protein, suggests that this is a very flexible region which moves independently from the rest of the complex. Most peaks are not affected by the presence or absence of Ca²⁺, showing no change in chemical shift values between the two conditions used. The main exception to that is residue 137, which moves almost 0.2ppm in the ¹H dimension, as indicated by the dashed line in Fig. 1. Residue 136 could not be assigned in the presence of EGTA/Mg²⁺.

Chemical shift assignments for N, H_N, C_α, and C_β atoms of residues 135–182 of TnI were obtained by sequential assignment, from 3D HNCACB and HNCOCA NMR experiments in the presence of Ca²⁺ and 3D HNCACB in the presence of EGTA/Mg²⁺.

The chemical shift index (CSI) is a measurement of how much the chemical shift value for each atom differs from the standard value for that atom in a random coil environment and can therefore be used to determine the secondary structure of the protein 19. The ¹³C_α CSI is not significantly affected by local sequence variation and is therefore more reliable than ¹H_N, ¹³CO, or ¹⁵N 20. CSI values were calculated by subtracting the reference value for random coil amino acids 21 from the assigned chemical shift for the residues in the C-terminal region of TnI in the presence of Ca²⁺ or EGTA/Mg²⁺. To correct for deuterium isotope effects, 0.5ppm were added to the result 22. The CSIs for the C-terminal region of TnI are presented in the top row of Fig. 2 in parts per million. The dashed lines represent the thresholds for any residue to be considered in an α-helical (positive) or β-strand (negative) conformation. An α-helical or β-strand segment is formed by a continuous stretch of at least three (for β-strands) or four (for α-helices) residues in the protein 21. No stable secondary structure can be seen in this region of TnI.

Display large version of this figure

Figure 2

¹³C_α CSI plots. (Top row) The visible region of TnI in the troponin complex, in the presence of Ca²⁺ or EGTA/Mg²⁺. The dashed lines represent the threshold values for α-helices (0.7ppm) and β-strands (–0.7ppm). (Bottom row) Comparison between TnI and the N-terminal domain of TnC in the troponin complex. Residues 135, 141, and 167–169 of TnI could not be assigned in the presence of EGTA/Mg²⁺. Residues 1–3, 18, 47, and 86 of TnC could not be assigned in the presence of EGTA/Mg²⁺, and residues 1, 16, 17, 25–28, 46, 47, 57, 58, 64, 65, 76, 79, 81, 86, 88–91, 93, and 94 of TnC could not be assigned in the presence of Ca²⁺.

A comparison between the CSIs for the N-terminal domain of TnC in the troponin complex and the C-terminal region of TnI, presented on the same scale, is shown in the bottom row of Fig. 2. TnC has well-defined secondary structure, which is reflected on chemical shift values of up to six times the threshold value and on average of three times the threshold value. The incipient quality of the secondary structure in the C-terminal region of TnI is reflected in the low indexes that most often do not achieve the threshold value for secondary structures.

Backbone amide ¹⁵N NMR relaxation measurements were performed for the C-terminal region of TnI in the troponin complex and are presented in Fig. 3. The presence of Ca²⁺ or EGTA/Mg²⁺, in general, did not affect the relaxation parameters. ¹H-¹⁵N heteronuclear NOE values were calculated from the ratio of the intensity of ¹H-¹⁵N-HSQC peaks in the presence and absence of proton saturation, as described in Materials and Methods, and presented in the top panel. The field at which the experiments were acquired strongly influences the NOE values due to effects on the CSA, and there was a significant difference between NOEs measured at 600 and 800MHz, as expected. At 600MHz, NOE ratios for many of the residues were too close to zero to be measured accurately. The measured values ranged from −0.53 to 0.28 for most of this region of TnI, reaching −2.0 for the most C-terminal residue. At 800MHz, the NOE ratios could be measured for most residues and were between 0.24 and 0.63. Residue 182 had the lowest NOE, at −1.1.

Display large version of this figure

Figure 3

Backbone amide ¹⁵N NMR relaxation measurements for the visible region of TnI in the troponin complex, in the presence of Ca²⁺ and EGTA/Mg²⁺, at two different fields: 600MHz and 800MHz. (●) In the presence of Ca²⁺ at 600MHz; (▴) in the presence of EGTA/Mg²⁺ at 600MHz; (♢) in the presence of Ca²⁺ at 800MHz; and (□) in the presence of EGTA/Mg²⁺ at 800MHz. (Top panel) NOE measurements; (center panel) T₁ relaxation times; and (bottom panel) T₂ relaxation times. The inset shows the lower region of the T₂ relaxation times plot expanded.

Backbone amide ¹⁵N T₁ relaxation times are plotted in the central panel of Fig. 3. Residues 143–179 had relatively constant T₁ times in both conditions and fields, with values between 500 and 800ms. Only a few residues had T₁ times above 800ms. Between residues 137 and 141, inclusive, there was a larger variation of T₁ times for each residue in the different conditions. T₁ values were consistently higher in the presence of Ca²⁺ than in the presence of EGTA/Mg²⁺. For residues 180 and 182 (residues 181, 155, and 142 are superimposed in the ¹H, ¹⁵N-HSQC spectrum and their relaxation parameters could not be measured), T₁ times are longer, ranging from 828ms to 1.29s in both conditions.

Backbone amide ¹⁵N T₂ relaxation times were also measured and are plotted in the bottom panel of Fig. 3. The inset shows an expanded view of the lower region of the plot, with T₂ times between 40 and 350ms. All the residues, with exception of residue 182, fall within that region. It can be seen from the plot that the T₂ relaxation times gradually increase for the first 10 residues, from residue 137 (the first residue for which T₂ could be measured) to 146. For the first five residues, T₂ values are consistently larger in the presence of EGTA/Mg²⁺ than in the presence of Ca²⁺. Between residues 147 and 177, the values of T₂ remain relatively constant, between 100 and 300ms. The five most C-terminal residues of the protein show another increase in T₂ values, reaching up to 1.3s for residue 182. For most of this region, T₂ values are larger at 600MHz than at 800MHz, as expected from the effect of field intensity on transverse relaxation.

¹⁵N relaxation measurements were analyzed by reduced spectral density mapping 18,23. The intensity of the ¹H, ¹⁵N-HSQC peaks for the C-terminal region of TnI, together with the backbone amide ¹⁵N relaxation measurements for this region, indicate that this is a flexible region, which does not conform to the assumptions required for the commonly used model-free formalism 24,25.

The spectral density function was evaluated at five different frequencies, as described in Materials and Methods. The results are presented in Fig. 4. Fewer residues could be evaluated at the frequencies observed at 600MHz, due to the lack of NOE data for these residues. J(0) shows clear difference between values in the presence of Ca²⁺ or EGTA/Mg²⁺ for the first five residues, whereas the remaining residues have very similar values for both fields and conditions. These initial residues have significantly higher values of J(0) in the presence of Ca²⁺ than in the presence of EGTA/Mg²⁺, and the values decrease along the sequence. From residue 144 on, J(0) becomes constant until the last five residues when the value of J(0) drops again.

Display large version of this figure

Figure 4

Spectral densities of backbone motions of the C-terminal region of TnI in the troponin complex, at five different frequencies, in nanoseconds/radians. (Top panel) J(0); (center panel) J(ω_N); and (bottom panel) J(0.87ω_H). Spectral densities calculated from data acquired at 600MHz are represented in black: J(0) in the top panel, J(61MHz) in the center panel, and J(522MHz) in the bottom panel. (●) In the presence of Ca²⁺, and (▴) in the presence of EGTA/Mg²⁺. Spectral densities calculated from data acquired at 800MHz are represented in white: J(0) in the top panel, J(81MHz) in the center panel, and J(696MHz) in the bottom panel. (♢) In the presence of Ca²⁺, and (□) in the presence of EGTA/Mg²⁺.

J(ω_N) shows field dependence for the first five residues and between residues 155–165. Its values are approximately constant along most of the C-terminal region of TnI, decreasing sharply along the last five residues. J(0.87ω_H) increases gradually along the sequence of TnI, and J(522MHz) is always larger than J(696MHz), as expected.

Chemical shifts for the C-terminal region of TnI in the troponin complex were not significantly different in the presence of Ca²⁺ or EGTA/Mg²⁺. This suggests that there are no conformational changes in this region in the presence or absence of Ca²⁺, unlike TnC in the troponin complex 7. Although the interactions between TnC and TnI are known to depend on the presence of Ca²⁺2, our results confirm that those interactions do not affect the last 40 residues of TnI.

CSI analysis of this region of TnI shows that it has no stable secondary structure. NMR data reflect the behavior of a population of molecules, and the CSI reflects what fraction of the population has well-defined secondary structure. Using 2.8ppm as the CSI for completely folded α helices 26, the helical residues in the C-terminal region of TnI are, on average, 15% folded. Recently, a structure for this region was determined using NOE-based techniques 8. Whereas CSI is equally sensitive to folded and unfolded fractions of the population, NOE is more sensitive to the smaller distances between nuclei in the folded population. This can be seen from the following equations:

Since random coil chemical shifts are associated with mobility 19, the CSI analysis indicates that the C-terminal of TnI is very flexible. This had already been demonstrated by the higher intensity for the peaks from this region when compared to the rest of the protein. Measurements of backbone amide ¹⁵N NMR relaxation parameters offer a more direct look at the dynamics of the C-terminal region of TnI. NOE ratios are relatively constant for most of the protein, with a pronounced decrease for the last four residues, indicating a very flexible C-terminal tail.

T₁ measurements show constant values for most of the residues measured, with slightly higher values for the first five residues, especially in the presence of calcium, and markedly higher values for the last four residues. Whereas the increase in T₁ for the first five residues indicates less flexibility in that region due to the proximity of TnC, the longer T₁ times at the C-terminus of TnI are a result of the higher flexibility of the C-terminal residues. The apparent correlation time for this region to TnI and the fields used place our T₁ measurements around the minimum of the T₁×τ_m plot (as represented in Spyracopoulos et al. 27), rendering T₁ measurements especially insensitive to changes in mobility in the C-terminal region of TnI.

T₂ measurements, in contrast, start at relatively low values and increase gradually for the first 10 measured residues. This is followed by a region of constant T₂, ending with a sharp increase in T₂ times for the last five residues of TnI. The C-terminal increase in T₂ times is characteristic of flexible termini and widespread among proteins. The slow increase in T₂ times between residues 137 and 146 is less common and indicates a gradual increase in mobility along the backbone of this region of TnI. T₂ times for the C-terminal region of TnI suggest it is a tethered domain, analogous to a ball at the end of a chain. The domain is formed by a collapsed region of constant T₂ times between residues 147 and 178, preceded by a “chain” of residues 137–146, which attach this region to the rest of the troponin complex.

A reduced spectral density mapping approach 18,23 was used to further analyze the relaxation data. The model-free approach 24,25 is the most commonly used method to analyze relaxation data. However, it assumes that the molecule is relatively rigid and its internal motions happen in a much faster timescale than the tumbling of the molecule (τ_m≫τ_e). Peak intensities, CSI, and the values of the relaxation parameters indicate that this is not the case for the C-terminal region of TnI.

J(0.87ω_H) values calculated from data acquired at 600MHz were consistently higher than J(0.87ω_H) values calculated from data acquired at 800MHz. This shows that the spectral density function for the C-terminal region of TnI can be described by a sum of exponentials and can be accurately analyzed by the reduced spectral density mapping approach 23.

The J(0) term in the reduced spectral density mapping analysis contains contributions of the exchange term R_ex, which affects the transverse relaxation when present and for that reason is sometimes referred to as J_e(0) 18. The contribution of R_ex is highly dependent on magnetic field and can be explicitly calculated when relaxation data are acquired at more than one field 23. In the case of the C-terminal region of TnI, however, both transverse relaxation times T₂s and J(0) show almost identical values at the two different fields, indicating that any contribution of an exchange term would be minimal and does not need to be calculated explicitly.

High J(0) values are indicative of slow motions (millisecond-microsecond), whereas J(0.87ω_H) is sensitive only to fast internal motions in the picosecond-nanosecond timescale 28. Taken together, the spectral density plots (Fig. 4) indicate that the first five measured residues are dominated by slow motions, which are more prominent in the presence of Ca²⁺ than in the presence of EGTA/Mg²⁺, suggesting a different anchoring for the tethered domain in the presence of Ca²⁺ or EGTA/Mg²⁺. A central region containing ∼30 residues (roughly residues 145–175) has a smaller amount of slow motions, with the J(0) constant around 1; and the C-terminal tail is very flexible, dominated by fast motions. Underlying that pattern, fast motions increase constantly from residue 137 to 182.

An apparent correlation time can be calculated for the central region encompassing residues 145–175, from the spectral density functions 18, according to the formula

The resulting value is ∼4ns, which corresponds roughly to a molecular mass of 8kDa 27. This is the same apparent molecular weight found from a simple relationship between transverse relaxation rate and molecular weight, derived from a number of proteins 7,29, and corresponds to ∼70 residues. This is more than the 47 residues contained in the C-terminal region of TnI and likely reflects the fact that this region is attached to the much larger troponin complex and is not rotating freely in solution. It also indicates that the C-terminal region of TnI is collapsed in the context of the troponin complex, as suggested by the NOE, T₂, and J(0) plots. At the same time, NOE values indicate that this is not a rigid 8-kDa domain. Expected NOE values for a rigid domain that size are ∼0.8.

Although the dynamics data show that the C-terminal of TnI is collapsed, the CSI shows no sign of stable secondary structure. Although this may seem contradictory at first, different models in polymer theory predict at least some degree of collapse for an unstructured protein independently of secondary structure 30. NMR spectroscopy describes the average ensemble of molecules in the sample; and therefore the possibility of transient secondary structure being formed in part of the ensemble, at levels too low to affect the average, cannot be excluded.

The picture that emerges for the C-terminal region, encompassing residues 137–182 of TnI, is that of a tethered collapsed domain, very flexible and with at most nascent secondary structure. This kind of domain had already been suggested by King et al. 9 as consistent with small angle neutron scattering data.

The domain is attached to the rest of the troponin complex only by the interaction between TnC and the “switch” region of TnI, which is known to be a Ca²⁺-dependent interaction 31. Nonetheless, the switch region does not become completely flexible in the presence of EGTA/Mg²⁺, since it cannot be visualized as sharper NMR peaks as observed for the C-terminal region. Rather, it stays close to TnC, as suggested by the crystal structure 6. The initial residues of the C-terminal region (137–141) become significantly more flexible in the presence of EGTA/Mg²⁺, indicating a change in the interactions between TnC and the preceding regions of TnI.

Despite being collapsed, fast motions (in the picosecond-nanosecond timescale) in the C-terminal region of TnI increase proportionally to the distance from the attachment point to the rest of the troponin complex. This suggests that along the whole sequence of this region, the point of attachment is a main factor limiting fast motions.

In the muscle, the C-terminal region of TnI interacts with actin in the relaxed state and is necessary for the inhibition of muscle contraction in the presence of TnC and absence of Ca²⁺5. This suggests that although this region of TnI is mostly unstructured in the presence of Ca²⁺, it becomes structured upon interaction with actin, as many examples of unstructured proteins and domains that fold upon binding to their targets 32. The coupling of folding and binding can offer high specificity at a modest affinity, since there is a balance between the enthalpic energy in the bound state and the entropic energy in the free, unfolded state. This is especially important for complexes that need to dissociate easily. To respond to the Ca²⁺ signal and start muscle contraction, TnI and actin have to be able to dissociate promptly.

Once dissociated from actin in the presence of Ca²⁺, the region of the troponin complex that contains the C-terminal region of TnI is at a larger distance from actin than in the absence of Ca²⁺, when TnI is interacting with actin. The “fly-casting mechanism” proposed for unstructured proteins 33 would allow TnI to bind to actin from that larger distance, folding and “reeling in” the whole troponin complex to the position that it occupies once TnI is bound to actin. According to this model, the C-terminal region of TnI remains unfolded when not interacting with actin and has a larger radius of gyration than it would when folded. When TnI approaches actin, weak interactions between unfolded TnI and actin can therefore begin to occur at a much larger distance than if TnI had to be folded. As these weak interactions occur, they increase the likelihood of other residues in TnI also interacting with actin, the enthalpic energy of the interactions compensates for the increased entropy of the unfolded state, and the C-terminal region of TnI assumes its folded conformation. The interaction between this region of TnI and actin, which holds the troponin complex in place in the absence of Ca²⁺, is likely part of the extra mass located in the absence of Ca²⁺ on the outer domain of actin by electron microscopy 8,34,35.

The homologous C-terminal region in cardiac TnI contains a few mutations associated with familial hypertrophic cardiomyopathy, K205S, G202S, and ΔK182. However, there is very limited functional information available about mutations in that region 36. The C-terminal region is very similar in the skeletal and cardiac isoforms of TnI, with >75% identity. It is therefore expected that both have the same structural properties, and the C-terminal region of cardiac TnI should also be unstructured when not bound to actin. Degradation of this region is associated with ischemia and myocardial stunning 36. Being unstructured when not bound to actin, this region of TnI can be particularly susceptible to proteolysis.

Although the crystal structures of the troponin complex did not show the C-terminal region of TnI, NMR dynamics reveals that it is intrinsically unstructured when not bound to actin. The coupling between folding and binding to actin allows the C-terminal of TnI to bind actin effectively from the Ca²⁺ state and to dissociate easily to allow muscle contraction.