Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Mar 15.
Published in final edited form as: J Mol Biol. 2024 Feb 20;436(6):168498. doi: 10.1016/j.jmb.2024.168498

Troponin structural dynamics in the native cardiac thin filament revealed by cryo electron microscopy

Cristina M Risi 1, Betty Belknap 1, Jennifer Atherton 1, Isabella Leite Coscarella 2, Howard D White 1, P Bryant Chase 3, Jose R Pinto 2, Vitold E Galkin 1,*
PMCID: PMC11007730  NIHMSID: NIHMS1973614  PMID: 38387550

Abstract

Cardiac muscle contraction occurs due to repetitive interactions between myosin thick and actin thin filaments (TF) regulated by Ca2+ levels, active cross-bridges, and cardiac myosin-binding protein C (cMyBP-C). The cardiac TF (cTF) has two nonequivalent strands, each comprised of actin, tropomyosin (Tm), and troponin (Tn). Tn shifts Tm away from myosin-binding sites on actin at elevated Ca2+ levels to allow formation of force-producing actomyosin cross-bridges. The Tn complex is comprised of three distinct polypeptides – Ca2+-binding TnC, inhibitory TnI, and Tm-binding TnT. The molecular mechanism of their collective action is unresolved due to lack of comprehensive structural information on Tn region of cTF. C1 domain of cMyBP-C activates cTF in the absence of Ca2+ to the same extent as rigor myosin. Here we used cryo-EM of native cTFs to show that cTF Tn core adopts multiple structural conformations at high and low Ca2+ levels and that the two strands are structurally distinct. At high Ca2+ levels, cTF is not entirely activated by Ca2+ but exists in either partially or fully activated state. Complete dissociation of TnI C-terminus is required for full activation. In presence of cMyBP-C C1 domain, Tn core adopts a fully activated conformation, even in absence of Ca2+. Our data provide a structural description for the requirement of myosin to fully activate cTFs and explain increased affinity of TnC to Ca2+ in presence of active cross-bridges. We suggest that allosteric coupling between Tn subunits and Tm is required to control actomyosin interactions.

Keywords: thin filament, troponin, myosin binding protein C, cryo-EM, muscle regulation

Graphical Abstract

graphic file with name nihms-1973614-f0001.jpg

Introduction

Every single heartbeat depends on the interaction between the two sarcomeric filamentous systems – thick (myosin) and thin (actin) filaments, a process that is regulated by Ca2+ [1-4]. Recent seminal advances in cryo electron microscopy (cryo-EM) yielded previously missing structural information about the organization of the thick [5, 6] and thin [7, 8] filaments. The innovative work by Yamada et al [7] defined the structural transition between the inhibited Ca2+-free and activated Ca2+-bound states of the cardiac thin filament (cTF). These findings provided a solid structural foundation for the steric blocking model of muscle regulation, which holds that at low Ca2+ (relaxing conditions) Tm blocks myosin from binding to actin, while at high Ca2+ (activating conditions) Tm allows cross-bridge formation [9, 10]. Subsequent cryo-EM studies with native cTFs at systolic Ca2+ levels confirmed the overall structure of the cTF proposed by Yamada et al [7].

The backbone of the cTF is formed by filamentous actin (F-actin) (Figure 1A, tan ribbons) that holds tropomyosin (Tm) – the steric blocker of actomyosin interactions (Figure 1A, yellow ribbons), and the troponin (Tn) complex (Figure 1A, green, red and blue ribbons) that regulates the position of Tm on F-actin in a Ca2+-dependent manner [4, 7, 8, 10]. Tn complex itself is a lengthy structure comprised of the Tn core (Figure 1A, large black bracket), TnT1 region (Figure 1A, blue arrows) that stabilizes the junction between the adjacent Tm molecules (Figure 1A, small black bracket), and TnI C-terminus region (Tn-CT) (Figure 1A, red arrows) that stabilizes the inhibited state of the cTF via its interactions with actin and Tm. The major features of the Tn core are TnC (Figure 1A, green ribbons), a Ca2+-binding subunit, and an IT-arm (Figure 1A, arm) that is comprised of a central section of TnI (Figure 1A, red ribbons) and a C-terminal section of TnT (Figure 1A, blue ribbons). Seven actin molecules along with one Tn complex and one Tm comprise cTF regulatory unit (RU). Additionally, due to the two-stranded, non-helical nature of cTF, residing opposite each other are two such RUs that are not equivalent, one upper strand and one lower strand RU due to their relative positions on the F-actin structure [7, 8].

Figure 1. Structure of the Tn core at low Ca2+ (pCa=8).

Figure 1.

Open in a new tab

(A) The original model of the cTF in Ca2+-free state [7, 8] (colored ribbons) is docked into the overall (e.g., consensus) 3D reconstruction of the cTF at pCa=8 (grey transparent surface) portrayed in two views related by 90° azimuthal rotation. The model consists of actin (tan ribbons), Tn core (large square bracket) comprised of TnC, TnI, and TnT (green, red and blue ribbons, respectively), and Tm molecules (yellow ribbons). TnT is a part of both Tn core (large bracket) and junction region (small bracket) that is annotated as TnT1 (blue arrows). TnI C-terminus spans up from the Tn core over three actin subunits (red arrows). The tip of the IT arm is missing in the overall 3D reconstruction (black arrow). The featureless TnC-NTD is marked by a green arrow. The area of the cTF used to refine the Tn core structure is marked with a dotted rectangle. (B-D) 3D reconstructions of the Tn core (transparent grey surfaces) are shown for Ca2+-free or CF (B), Ca2+-free rotated or CF-R (C), and Ca2+-free tilted or CF-T (D) structural states. The color codes for the Tn core components are the same as in (A). (E) Comparison of the CF state reported here (blue ribbons) with previous data show small clockwise rotation of the CF IT arm by 5° and dramatic counter clock rotation of the CF TnC N-terminal domain by ~25°. (F) Structure of the TnC N-terminal domain reported here (blue ribbons) matches the NMR structure of this domain (PDB 1SPY, red ribbons). (G) Transition from CF (blue) to CF-R (green) structural state involves rotation of the Tn core by ~30° from Ca2+-free to the Ca2+-bound state (red arrow). (H) Transition from CF (blue) to CF-T (cyan) structural state implicates the tilt of the Tn core upwards by ~20° (red arrow).

The multifaceted structural arrangement of the cTF implies that every component of the Tn complex has its specific role in balancing the inhibited and activated states of the cTF. For example, TnT1 (Figure 1A, blue arrows) role in cTF regulation is to stabilize the Tm head-to-tail interaction by spanning across the Tm overlap region [7, 8, 11, 12]. A high resolution cryo-EM structure of the junction region revealed molecular details of stabilization of the Tm-Tm overlap by TnT1 of this intricate interaction, i.e., Tm overlap region and TnT1 [13]. The position of TnI-CT, which is crucial for stabilization of the inhibited state of the cTF[14-17], has been shown by cryo-EM to span across three actin molecules above the Tn core (Figure 1A, red arrows) [7, 8]. In silico modeling confirmed its interactions with both actin and Tm [18]. Unfortunately, the secondary structure of one of the most crucial parts of the Tn complex – the TnC N-terminal domain (TnC-NTD) that harbors the low affinity Ca2+-binding site that is responsible for Ca2+-dependent regulation [19, 20] was not resolved in the cTF context [7, 8]. In overall maps, its density appeared as a featureless blob that did not show any structural details (Figure 1A, green arrow), suggesting structural heterogeneity within that region of the Tn core. Furthermore, cTF Ca2+ sensitivity relies on the structural dynamics of the TnC-NTD [21] where both Ca2+ and TnI-CTD (i.e., where the switch helix or switch peptide is located) promote opening of the TnC-NTD cleft [22, 23] that is required for cTF activation [7, 8]. Therefore, the lack of a reliable structure of the TnC-NTD region poses a tremendous drawback on our understanding of cTF regulation.

Here we used an unsupervised classification approach to show that the Tn core of the cTF adopts multiple structural conformations at both high and low Ca2+ levels and that the frequencies of those conformations differ between the two strands (i.e., upper and lower strands). Visualization of the Tn core secondary structure allowed us to fill in this important portion of existing models of the cTF, models that are widely used for in silico experiments [24-26]. Firstly, we show that at high Ca2+ levels (pCa=4) the cTF is not entirely activated by Ca2+ and RUs exist in two structural modes – a fully activated state and a novel partially activated state. These two cTF states differ in the conformation of the TnC-NTD, interactions of the TnI-CTD with TnC and actin-Tm, and the position of Tm on actin. These novel data provide a structural explanation for the requirement of cross-bridges to fully activate the cTF [27, 28]. Secondly, we show that at low Ca2+ levels (pCa=8) the Tn core displays three different conformations while Tm position remains the same in the inhibited state. Finally, we show that C1 domain of cardiac myosin binding protein C (cMyBP-C), which has been shown to activate cTF to the same extent as myosin in the absence of Ca2+ [29], converts the structure of the Tn core into the Ca2+-bound fully activated state in the absence of Ca2+ which corroborates with the increased affinity of TnC to Ca2+ during contraction [30, 31]. Overall, our work shows how TnC, TnI-CTD and Tm work in concert to control actomyosin interactions.

Results

Structure of the Tn core of the Ca2+-free cTF.

First, we aimed to analyze the structure of the Tn core at Ca2+-free conditions (pCa=8, which is a sub-diastolic Ca2+ concentration). We used the same set of manually selected segments of native cTFs previously used for reconstruction of the junction region of the cTF [13]. The initial steps of image analysis were analogous to those used for classification of cTFs at systolic Ca2+ levels [8] (summarized in Supplementary data Figure S1). In agreement with the previous observations [7, 8] the Tn core region of the cTF yielded the lowest resolution of ~11 Å when compared to the rest of the map (Figure S1I) and appear as a featureless blob (Figure 1A, large black bracket). The density at the tip of each IT arm was partially missing as well (Figure 1A, black arrow). This indicated that the Tn core density in the map was an average of multiple conformations.

We used the particle subtraction routine in RELION [32] to extract fragments of the cTF containing either the upper or lower Tn complexes (Figure S1J) and those two sets were processed separately (Figures S2 and 3) to: i) reveal if the two Tn complexes were structurally different; and ii) increase robustness of obtained structures. First, we analyzed the upper Tn core (Figure S2). Of note, all attempts to analyze both Tn cores together failed, which suggests that the conformations of upper and lower Tn complexes are not tightly correlated. Due to a lack of precise structural information regarding the structural state and/or positioning of the Tn core relative to the actin backbone of the cTF we used unsupervised 3D sorting (Figure S2B) that revealed three distinct structural classes that were equally populated (Figure S2B, black numbers). We termed those Ca2+-free (CF), Ca2+-free rotated (CF-R), and Ca2+-free tilted (CF-T) (Figures 1B, C, and D, respectively). Local resolution measurements showed that the highest resolution was achieved in the actin region (~4.5 Å) followed by the IT arm region (~6 Å) and the TnC N-terminal regulatory domain (~7.5 Å) (Figures S2D, G, and K). The resultant maps were filtered to 7.5 Å for CF and CF-R, while CF-T was filtered to 8 Å resolution.

All three maps displayed very well defined helices across the Tn complex and were suitable for building reliable atomic models (see Supplementary Information for details and Table S2 for the validation of the models). Of the three resultant atomic models, the CF model was closest to previously adopted models [7, 8, 33] that share similar IT arm position with only 5° rotational difference (Figure 1E, small black arrow). However, there was a significant difference in the position of TnC’s regulatory (N-terminal) domain (NTD) – in our structure the regulatory domain was rotated counter-clockwise around the linker region between the two TnC lobes by ~25° from its position in earlier models [7, 8, 33] (Figure 1E large black arrow, Movie S1). Nonetheless, the folding of the TnC-NTD was the same as in NMR structure PDB 1SPY [34] (Figure 1F) with RMSD of 2.5 Å2 when the loop between helices A and B (Figure 1F, red circle) was excluded. Comparison of the entire CF TnC-NTD with the ensemble of the 40 NMR models provided in 1SPY yielded an average RMSD of 3.0 Å2. The main difference between the CF and CF-R states was the swing of the entire Tn core in the Ca2+-free conformation by ~30° from its Ca2+-free to the Ca2+-bound azimuthal position (Figure 1G, red arrow and Movie S2). Finally, the CF-T state has its Tn core tilted up by ~20° without any azimuthal rotation (Figure 1H, red arrow and Movie S3). The position of Tm in all three of these states (i.e., CF, CF-R, and CF-T) was the same and consistent with previous reports [7, 8, 33]. Flexible fitting of Tm into the actin-Tm part of the consensus map filtered to 4.2 Å resolution yielded a model consistent with the recently published alternative model of the Ca2+-free TF (PDB 7UTL) [33] (Figure S4) except the Tm junction region did not match [13].

Image analysis of the lower Tn core set was done using the same routine (Figure S3) and yielded three very similar structures to the ones obtained for the upper strand. Comparison between the upper and the lower resultant models shows that the CF and CF-R structures are very similar between the two strands (Figure S5A-D). The RMSD between the upper and the lower CF models across all 367 pairs (e.g. TnC, TnI and TnT subunits) was 2.6 Å2, while the statistics for comparison of the upper and the lower CF-R models yielded RMSD of 2.4 Å2. The upper and the lower CF and CF-R structures were obtained independently via unsupervised sorting, therefore, their semblance proves the robustness of the CF and CF-R Tn core structures reported here. CF-T classes from the upper and the lower strands had similar positions of the IT arm (Figures S5E-G, red arrows), while the TnC-NTD in the lower CF-T structure was shifted by ~12 Å from its position in the upper counterpart (Figure S5G, black arrow and Movie S4). Both the upper and lower CF-T classes possessed disordering of the TnC-NTD which required additional unsupervised sorting of those classes (Figures S2I and S3I). This suggests that when Tn core is tilted up, its TnC-NTD interactions with the rest of the filament are weakened and that this weakening results in different patterns of interaction of the TnC-NTD with the actin-Tm backbone on the upper and lower strands. Notably, the position of Tm was indistinguishable in all the Ca2+-free (CF) states on the upper and lower strands.

To evaluate how different conformations of the Tn core (e.g., CF, CF-R, and CF-T) affect the structure of the regulatory unit (RU) we used the same approach as for the Tn core, but this time we used the particle subtraction routine to extract the entire RU for both upper and lower strands (Figures S6 and S7, respectively). The actual maps and associated models for the upper RUs in CF, CF-R and CF-T states are shown in Figures 2A, B and C, respectively. We did not find any differences in either the TnT1 interactions with actin-Tm (Figure 2, black arrows) or the TnI C-terminus interactions with actin and Tm (Figure 2, red arrows). The same results were obtained for the lower strand.

Figure 2. 3D reconstructions of the regulatory units (RUs) in CF, CF-R and CF-T structural states.

Figure 2.

Open in a new tab

(A-C) The actual maps are shown as grey transparent surface, actin molecules are tan, and Tm is yellow, while Tn core in CF state is shown in blue (A), in CF-R state in green (B), and in CF-T state in cyan (C). TnI-CT is shown in red and its trajectory over the three consecutive actins is marked with red arrows. TnT1 is marked with black arrows.

Finally, to evaluate whether the CF, CF-R or CF-T states are coupled across the strands, we sorted the entire cTF segments (i.e., comprised of both strands) into nine classes based on the structure of the Tn core (Figure S8). Our data showed that the frequencies of resultant classes were comparable (Figure S8, black numbers). Hence, the Tn core conformations in the Ca2+-free cTF were uncoupled suggesting low (if any) cooperativity between the two strands in the inhibited cTF.

Structure of the Tn core of the Ca2+-bound cTF.

Next, we analyzed the structure of the Tn core at high Ca2+ conditions (pCa=4, which is a saturating concentration of Ca2+). We used a set of automatically selected segments of native cTFs. The initial steps of image analysis were analogous to those used for the classification of cTFs at systolic Ca2+ levels [8] and are summarized in Figure S9. In agreement with the previous observations [7, 8] the Tn core region of the cTF yielded the worst resolution (~11 Å) when compared to the rest of the map (Figure S9I) and resembled a featureless blob (Figure 3A, large black bracket). Similar to the Ca2+-free overall map (Figure 1A), the density at the tip of each IT arm was partially missing as well (Figure 3A, black arrow). This indicated that at high Ca2+ levels the Tn core does not exist in a single structural conformation. We extracted fragments of the cTF containing either the upper or lower Tn complexes (Figure S9J) and those two sets were processed separately using an unsupervised classification routine (Figures S10 and S11). Analysis of the upper strand particles revealed two distinct structural states (Figures 3B and C) that were present at comparable frequencies (Figure S10B, black numbers). Local resolution calculations indicated that the highest resolution was achieved in the actin region (~4.8 Å) followed by the IT arm region and the TnC N-terminal regulatory domain (~7 Å) (Figure S10E and H). Both density maps filtered to 7 Å resolution displayed well defined helices all across the Tn core and were suitable for building reliable atomic models (see Supplementary Information for details and Table S4 for the validation of the models). The atomic model of the first class (Figure 3B) was close to the earlier model of the Ca2+-bound Tn core [7, 8, 33] (Figure 3D) with RMSD of 2.9 Å2. Comparison of the TnC-NTD alone from our model with the same region from the crystal structure of the Ca2+-bound Tn (PDB 4Y99) (Figure 3E) yielded RMSD of 3.0 Å2. Importantly, the Tm density in the first structure matched Tm position in the preceding models of the fully activated cTF [7, 8, 33]. We termed the first structure as Ca2+-bound fully activated (CB-FA) state. The second class was prominently different from the first class (Figure 3F) – its IT arm and Tm (Figure 3F, purple surface) resided between the CF and the CB-FA states (Figure 3F, blue and pink surfaces, respectively), while its TnC-NTD cleft was only partially open (Figure 3G, cyan arrow and Movie S5) despite TnI’s switch helix bound in the cleft (Figure 3C, red arrow). Due to its partial activation state we called that second class Ca2+-bound partially activated (CB-PA). To summarize, the unsupervised classification of Tn core complexes at high Ca2+ levels revealed that not all the Tn cores and associated Tm cables were fully activated by Ca2+.

Figure 3. Structure of the Tn core at high Ca2+ (pCa=4).

Figure 3.

Open in a new tab

(A) The original model of the cTF in Ca2+-bound state [7, 8] (colored ribbons) is docked into the overall (e.g., consensus) 3D reconstruction of the cTF at pCa=4 (grey transparent surface) depicted in two views related by 90° azimuthal rotation. The model consists of actin (tan ribbons), Tn core (large square bracket) comprised of TnC, TnI, and TnT (green, red and blue ribbons, respectively), and Tm molecules (magenta ribbons). TnT is a part of both the Tn core and junction region (marked as TnT1, blue arrows). TnI switch helix or S-helix (human residues 149-160, red ribbons) is marked with red arrow. The tip of the IT arm is missing in the overall 3D reconstruction (black arrow). The area of the cTF used to refine the Tn core structure is marked as dashed rectangle. (B-C) 3D reconstructions of the Tn core (transparent grey surfaces) are shown for Ca2+-bound fully activated or CB-FA (B), and Ca2+-bound partially activated or CB-PA (C) structural states. The color codes for the Tn core components are the same as in (A). (D) CB-FA state reported here (pink ribbons) is consistent with the previous models (PDB 6KN8, 7KO5 and 7UTI; grey ribbons). (E) The conformation of the TnC-NTD in the CB-FA state matches that in the crystal structure of the Tn complex in the Ca2+-bound state (PDB 4Y99, grey ribbons). (F) Superimposition of the CF (blue), CB-PA (medium purple) and CB-FA (pink) 3D reconstructions shows that CB-PA resides in-between the CF and CB-FA states. The swing of the IT arm and azimuthal rotation of the Tm cable upon cTF activation are shown as yellow and red arrows, respectively. (G) Opening of the TnC-NTD cleft upon transition from CB-PA to CB-FA state is marked with a cyan arrow. (H - I) Superimposition of the CF-R (green) and CB-FA (pink) structures shows similar swing of their IT arms (black arrow), but different Tm positions (green and pink arrows) and different state of the TnC NTD cleft (I).

Comparison of the CF-R (Figure 3H, green) and CB-FA (Figure 3H, pink) states revealed that even though the positions of their IT arms were very similar (Figure 3H, black arrow), the Tm cable in the CF-R state was in the blocked state (Figure 3H, green arrow) while its TnC-NTD cleft was tightly closed (Figure 3I). Therefore, the azimuthal swing of the Tn core itself was not sufficient for the activation of the cTF and could occur in the absence of Ca2+.

Image analysis of the lower Tn core set (Figure S11) was done using the same routine used for the upper Tn core set and yielded fully (e.g. CB-FA) and partially activated (e.g. CB-PA) classes. In contrast to the upper strand, the frequency of the CB-FA class was 2-fold higher than that for the CB-PA class (Figure S11B, black numbers). The CB-FA class was not entirely homogeneous and was divided into two sub-classes termed CB-FA1 and CB-FA2 using unsupervised 3D sorting (see Supplementary Information methods and Figure S11F and G). Comparison between the lower CB-FA1 and CB-FA2 models (Figure S12A-E) revealed two differences. First, Tm in the CB-FA1 state was slightly shifted towards the inhibited state by ~3° from its position in either upper CB-FA or lower CB-FA2 (Figure S12B, cyan arrow). Second, the B-helix of TnC-NTD in CB-FA2 was tilted down by ~25° from its position in either upper CB-FA or lower CB-FA1 (Figure S12C, black arrow, Movie S6). Therefore, the conformation of the Tn core in upper CB-FA was more similar to that in lower CB-FA1 (RMSD 3.0 Å2) (Figure S12D and E), while Tm position was identical in the upper CB-FA and lower CB-FA2 (Figure S12B). Comparison of the partially activated classes (e.g. CB-PA) between the two strands revealed a small difference in IT arm orientation (Figure S12F and G, black arrow) between the two otherwise very similar structures (RMSD 3.4 Å2). Overall, unsupervised sorting revealed minor structural differences in the conformations of the Tn cores between the upper and the lower strands, but revealed that the lower strand has 2-fold more fully activated Tn complexes when compared with the upper strand, which provides further experimental evidence for differences in the regulation of the two strands by Ca2+ reported previously [8].

To evaluate how the two conformations of the Tn core at high Ca2+ levels (i.e., CB-FA and CB-PA) affect RU structure, we used particle subtraction to extract the entire RU for both upper and lower strands (Figures S13 and S14, respectively). The actual maps and associated models for the upper RUs in CB-FA and CF-R are shown in Figure 4B and C, respectively. We compared those RU maps with the CF state and found that the extent of RU activation was proportional to the interaction of the TnI C-terminus (TnI-CT) with actin-Tm. In the absence of Ca2+, TnI-CT had extensive interactions with actin-Tm (Figure 4A, red arrows) while at high Ca2+ levels, only partially activated RUs (CB-PA state) had traces of TnI-CT bound to actin (Figure 4B, red arrows) or Tm (Figure 4B, red spheres). The fully activated RU map (CB-FA) did not show any signs of TnI-CT in association with actin-Tm (Figure 4C). Of note, the TnI switch helix was bound to TnC-NTD in both partially or fully activated RUs (Figure 4B and C, red brackets). Next, we compared the positions of Tm in the aforementioned states of the RU (Figure 4D, Movie S7) – the trajectory of the Tm swing (Figure 4D, black arrow) from the CF (Figure 4D, yellow ribbons) to the CB-FA (Figure 4D, magenta ribbons) state was equivalent to that reported previously [7, 8], while the CB-PA state (Figure 4D, orange ribbons) was located between those two states. To summarize, our data fully confirmed previous findings regarding the movement of the Tm cable on the surface of actin upon cTF activation [7, 8], whereas the Tm movement can be linked to the interaction of TnI-CT with actin-Tm.

Figure 4. Comparison of the 3D reconstructions of regulatory units (RUs) in CF, CB-PA and CB-FA structural states.

Figure 4.

Open in a new tab

(A-C) The actual maps are shown as grey transparent surfaces, actin molecules are tan and Tm is yellow in CF state (A), orange in CB-PA state (B) or magenta in CB-FA state (C), while Tn core is depicted in blue in CF state (A), purple in CB-PA state (B) or pink in CB-FA state (C). TnI-CT is shown in red and its trajectory over the three consecutive actins is marked with red arrows. The position of the TnI switch helix (S-helix) is marked with red brackets. (D) The full trajectory of the Tm swing upon cTF activation by Ca2+. Color codes for Tm are the same as in A-C.

Finally, to evaluate whether the CB-FA and CB-PA states are correlated across the two strands, we sorted the entire cTF segments (e.g. comprised of both strands) into six classes based on the structure of the Tn core (Figure S15). Our data showed that the frequencies of resultant classes were comparable (Figure S15, black numbers), except segments that had both RUs in the CB-PA state were more frequent. Hence, the Tn core conformations at high Ca2+ levels confirmed the existence of low cooperativity across the two strands reported previously [8].

Structure of Tn core in the Ca2+-free cTF activated by the C1 domain of cMyBP-C.

Lastly, we aimed to elucidate how activation of the cTF by other sarcomeric proteins rather than Ca2+ affects structure of the Tn core. We loaded cTFs at low Ca2+ (pCa=8) with sub-stoichiometric amounts of cMyBP-C C1 domain previously reported to activate the cTF to the same extent as myosin [29, 35]. The initial steps of image analysis were analogous to those used for the classification of cTFs at systolic Ca2+ levels [8] and are summarized in Figure S16A-C. To evaluate how small quantities of C1 interact with the cTFs, we used an unsupervised sorting approach (Figure S16D) to determine: i)what are preferred sites of interaction of C1 with the cTF; and ii) what is the Tn core structure near sites of C1 binding. The cTF segments were sorted into eight classes and their frequencies are depicted in black in Figure S16D, with the best matching structural states of the upper and the lower RUs marked using abbreviations from Figures 1 and 3. The first two classes (~24%) were in the CF state. Despite the absence of Ca2+, classes 3 – 5 (~40%) had either the upper or the lower strand partially activated (i.e., it gave the best match with the CB-PA model) and class 6 (~13%) had both strands matching the CB-PA state. Finally, classes 7 and 8 (~23%) showed C1 bound to the cTF; however, C1 did not bind over the entire filament, but strictly at sites predicted earlier [36]. In either class, C1 density was prominent only when C1 was cross-linking actin of the lower strand with the Tm of the upper strand (detailed in Figure 5A) so that the upper strand Tm was in the myosin state (M-state) as previously demonstrated, despite the absence of Ca2+ [29, 35]. Overall, the unsupervised sorting provided strong evidence that C1 cooperatively activates the cTF, since ~53% of cTF segments (classes 3 – 6) had one or both strands partially activated. However, full activation of RUs (e.g. to the C1/M state) required prominent C1 occupancy (e.g., upper RUs in classes 7 and 8) (Figure S16D, dark green arrows). Further work will be required to characterize this cooperative activation of the cTF by C1.

Figure 5. Structure of the Tn core at low Ca2+ (pCa=8) in presence of cMyBP-C C1 domain peptide.

Figure 5.

Open in a new tab

(A) Four views of the overall (consensus) map are related by 90° azimuthal rotations. The map is grey transparent surface while the model is shown as ribbons. The upper RU activated by C1 has Tn core in steel gray and Tm in cyan. The lower RU has Tn core in purple and Tm in orange. Location of the missing tip of the lower Tn is marked with a red circle. Traces of TnI-CT on the lower strand are marked with red numbers and arrows. Three C1 peptides bound to three sites between the lower-actin and upper-Tm are in dark green, while the closest to the TnT linker actin is denoted with a red asterisk. The area of the cTF used to refine the Tn core structure is marked with a dashed rectangle. (B) Refined map (transparent grey surface) and the resultant model with actin in tan, Tn core in steel gray, and Tm in cyan are shown for the upper strand RU. TnI switch helix (S-helix) is marked in red. (C) Tn core structure activated by C1 (steel gray ribbons) is consistent with the CB-FA state (pink ribbons) despite the absence of Ca2+ in the C1 experiments. (D) Trajectory of Tm swing upon cTF activation by Ca2+ or cMyBP-C C1 domain.

Since class 8 upper strand had the highest occupancy by C1 (Figure S16D, dark green arrows) it was chosen for further processing (Figure S17). Local resolution measurements in the overall (e.g., consensus) map (Figure S17C) showed that actin was resolved to ~4.3 Å resolution and Tm to ~6 Å resolution, while Tn core and C1 were resolved to ~8 Å resolution. Detailed views of the consensus map filtered to 8 Å resolution are shown in Figure 5A. The upper strand Tm was in the C1/M-state (Figure 5A cyan ribbons), while the Tm cable on the lower strand was in the CB-PA state (Figure 5A, orange ribbons). The lower strand Tn core matched the CB-PA model (Figure 5A, purple ribbons) but the map was missing the IT arm tip (Figure 5A, red circle), suggesting heterogeneity. Due to the small size of the class, classification of the lower Tn core was not conceivable. Consistent with the CB-PA state (Figure 4B, red arrows), the lower strand also possessed traces of TnI-CT (Figure 5A, red arrows) which were completely missing on the C1-activated upper strand. C1 (Figure 5A, green arrows) was preferably bound to three adjacent actins starting from the actin subunit adjacent to the TnT linker (Figure 5A, red asterisk).

The Tn core on the upper activated strand (Figure 5A, steel grey ribbons) matched well the CB-FA state. To further elucidate its structure, we extracted the Tn core region (Figure 5A, dotted rectangle) particles and sorted those into two classes using unsupervised classification routine. The set was found to be quite homogeneous with the majority of segments (~72%) binned into the first class (Figure S17E) which in turn yielded a 7 Å resolution (Figure S17G) map of the core region (Figure 5B). Comparison of the resultant atomic model showed that the conformation of the Tn core was very similar to the CB-FA structure obtained from the Ca2+-activated cTF (Figure 5C) with RMSD of 3.1 Å2. However, the position of the Tm cable obtained from the consensus map (Figure 5A, cyan ribbons) was distinct from CB-FA since Tm is in the M-state (Figure 5D). The full trajectory of Tm from CF towards its C1/M-state (Movie S7) is consistent with the original mechanism proposed by Yamada et al. [7]. We conclude that C1 preferentially binds between the lower-actin and the upper-Tm site on the cTF presumably due to its interactions with the more compressed TnT linker. In the absence of Ca2+, C1 activates the upper strand so that the Tn core resembles the CB-FA state (i.e., Ca2+-bound state) with its TnC-NTD cleft open and the TnI switch helix bound in the cleft (Figure 5B, red arrow). At the same time, in the C1-activated RUs, TnI-CT is completely dissociated from its actin-Tm site while Tm is in the M-state.

Effect of C1 domain of cMyBP-C binding to fluorescently-labeled reconstituted thin filaments (RTFs).

To evaluate our structural observations we used fluorescently-labeled RTFs that report on changes upon Ca2+ binding to the regulatory site II of cTnC. We performed two different approaches to test the binding of C1 to RTF and its modulatory function to cTnC site II. In one set of assays, we titrated first with C1 followed by Ca2+ (Group 1), while in separate assays, we titrated first with Ca2+ and followed by C1 (Group 2) (Figure S18A). In Group 1, titration of C1 increased fluorescence intensity, qualitatively mimicking the effect of Ca2+ binding to cTnC site II in RTF. Ca2+ addition after C1 did not show any further increases in RTF fluorescence (Figure S18B, C and D). In Group 2, as expected, Ca2+ titration increased RTF fluorescence intensity. Surprisingly, C1 titration (following pCa 5.0, which is a Ca2+ saturating condition for cTnC Ca2+ binding site II, measured at 30°C) increases RTF fluorescence intensity even further (Figure S18E, F, G). Of note, ~2.5 μM of C1 was used for cryo-EM experiments, while up to 15 μM of C1 was used in fluorescent work to achieve maximal effect of C1 on steady-state fluorescence of RTFs.

Our results demonstrate that: i) in agreement with the structural data (Figure 5), C1 produces similar effects of Ca2+ binding to cTnC regulatory site II; and ii) further fluorescence changes in cTnC occur by the binding of C1 to RTF even when saturated with Ca2+. Additional rise in the fluorescence presumably reflects the absence of CB-PA RUs since all the Tn cores are in the CB-FA states.

Discussion

Our main goal was to resolve the secondary structure the TnC-NTD that harbors the low affinity Ca2+-binding site responsible for Ca2+-dependent regulation of the cTF [19, 20] and to correlate its conformation(s) to structural states of other regulatory components of the cTF (e.g. TnI C-terminus and Tm). The global resolution for the 3D reconstructions of the cTF reported previously was from 6.6 Å [7] to 8 Å [8] but that metric was to some extent misleading because the resolution was not uniform in the density maps with the lowest resolution observed in the Tn core region. Therefore, TnC NTD secondary structure was not determined in previous cryo-EM studies, hence, an approximate position of TnC-NTD was used in the original cryo-EM models of the cTF [7, 8]. A focused approach to the cTF reconstruction allowed us to achieve near atomic resolution for the Tm junction region [13]. Here we used similar strategy and focused on the structure of the Tn core region – the most mobile part of the cTF after the TnT and TnI N-termini. Due to its intrinsic heterogeneity we could not achieve near atomic resolution for this part of the cTF. However, we were able to visualize its secondary structure in multiple structural states and link those structural modes to the TnI C-terminus mobile region interactions with actin-Tm and Tm position on the actin surface. Our data shows that the TnC-NTD, TnI-CTD and Tm are allosterically coupled.

We used unsupervised sorting to separately evaluate Tn core complexes from the upper and the lower strands to exclude any bias in the TnC-NTD structure determination. Of note, all the attempts to process the Tn cores from the two strands together were unsuccessful due to very weak correlation between the structural states of the upper and the lower strands (Figures S8 and S15). Importantly, our approach allowed us to compare independently obtained structures of the upper and the lower Tn cores to structurally characterize differences between the two strands. The Tn structural modes that we report here may represent the ones with the lowest free energy and therefore the most populated states, but transient states should also exist and their presence presumably lowers the resolution reported here for the Tn core structures. Further efforts are required to obtain atomic resolution for this cTF region.

At low Ca2+ levels we found three major conformations of the Tn core that differed by the orientation of the entire Tn core with respect to the actin backbone of the cTF (Figure 1B-D). Indeed, averaging over these three states significantly reduced resolution of the resultant map (compare Figure 1A with Figure 1B-C). Our data indicate that in the relaxed cTF all three Tn core structures exist on both strands with very similar frequencies (Figures S2B and S3B, black numbers). We found that CF and CF-R structures are very similar between the strands (Figure S5A-D), while CF-T states are somewhat different due to a shift in the TnC-NTD position (Figure S5G, Movie S4). How this difference may affect regulation of the cTF remains to be determined. We achieved a 7.5 – 8 Å resolution of the Tn core which was sufficient to visualize the secondary structure of the core components and build more reliable atomic models of the cTF. The position of TnC-NTD in the original Ca2+-free cTF model was off by a 25° counter-clockwise rotation of TnC-NTD with respect to the IT arm (Figure 1E) and this positioning of the TnC-NTD was preserved in the CF-R and CF-T structures on both strands. Importantly, the resultant atomic model of the TnC-NTD matched the corresponding NMR structure (Figure 1F). TnI-CTD in all three structures is bound at the actin-Tm interface as proposed previously [7, 8, 18] (Figure 2). It is important to note that our data on the structure of the entire cTF (Figure S8) explicitly show that the cTF in the relaxed state may exist in at least nine structural states which are equally probable. Therefore, a single model for the relaxed cTF does not exist and this should be carefully considered when designing in silico studies.

Unsupervised sorting of the Tn core from the cTF at high Ca2+ levels revealed two structural states (Figure 3B and C). The first one (Figure 3B) is consistent with the original model of the activated cTF [7, 8] (Figure 3D) and, therefore, is termed a fully activated state (e.g. CB-FA), whereas the second one (Figure 3C) is a novel structure that has not been reported previously. This novel state was found to have its IT arm position, Tm position and the TnC-NTD cleft opening between the CF and CB-FA states (Figure 3F and G), and for that reason, was termed partially activated (e.g. CB-PA). It was exhilarating to visualize the partially activated state, since it has been demonstrated that Ca2+ alone cannot fully activate the TF and cross-bridges are required [27, 28]. It seems quite possible that the PA state may represent the kinetically predicted closed state of the thin filament [9]. In silico experiments also proposed an intermediate state between the relaxed and activated conformations of the cTF and pointed to the role of TnI-CTD interactions with both actin-Tm and the TnC-NTD upon cTF activation [26].

The TnC-NTD studies indicate that both Ca2+ and the TnI switch helix (porcine residues 150-161, human residues 149-160) that is located in the TnI-CTD are required to open the TnC-NTD cleft [22, 23] which leads to the transition of the cTF from its relaxed to the activated state. However, the C-terminal 17 residues of TnI (porcine residues 194-211, human residues 193-210) are involved in the opposite activity – stabilization of the cTF relaxed state [14-17]. Therefore, cTF Ca2+-dependent regulation relies on the structural dynamics of the TnC-NTD, where Ca2+ and TnI-CTD govern transitions between the energetically different structural states of the TnC [21]. Here we directly show the interplay between the structural state of the TnC-NTD and the interaction of the TnI-CTD with actin-Tm and TnC (Figure 4). In any of the CF states (e.g. CF, CF-R, and CF-T) TnI-CTD is tightly bound to its actin-Tm interface (Figure 2, red arrows), while the TnC-NTD cleft is closed (Figure 1) irrespective to the orientation of the Tn core relative to the cTF backbone. In the partially activated state (e.g. CB-PA) the switch helix of TnI-CTD is already bound to TnC-NTD (Figure 4B, red bracket), whereas partial interaction of TnI-CTD residues 175-188 with actin and 198-201 with Tm is still evident (Figure 4B, red ribbons and spheres, respectively). Both 175-188 [16] and 198-201 [14, 15, 17] segments of TnI are essential for the inhibitory activity of TnI. In the fully activated CB-FA state no traces are detected of TnI-CTD bound to actin-Tm (Figure 4C). Hence, complete dissociation of the TnI-CTD from actin-Tm is required for the full activation of the cTF. Presumably, in the presence of Ca2+ TnI-CTD comes on and off from its actin-Tm interface, which in turn modulates the activation state of the cTF [37]. This conclusion aligns well with the previously proposed “fly casting” model of TnC-CTD regulation of the TF [38].

In contrast to the Ca2+-free cTF, we found significant differences between the structural states of the upper and the lower strands. We identified two structural conformations of the CB-FA state on the lower strand (Figure S12) and, importantly, the frequency of the CB-PA state was reduced on the lower relative to the upper stand (Figures S10B and S11B, black numbers), suggesting that TnI-CTD dynamics differs between the two strands. Additional experiments will be required to elucidate the mechanism(s) of these differences.

Despite the lack of atomic resolution for the TnC-NTD in our structures, comparison of the CF, CF-PA, and CB-FA states leads to a hypothesis regarding the structural transitions within the TnC-NTD upon cTF activation (Figure 6A-E). Ca2+ on its own cannot open the TnC-NTD to expose its hydrophobic core (Figure 6A, cyan circle), hence binding of TnI switch helix in the hydrophobic cleft is also required to stabilize the open, activated configuration [22]. TnC-NTD in the CB-PA state shows partial opening of the cleft (Figure 6B, green arrow) where hydrophobic interactions between helices A and B are still preserved (Figure 6B, cyan circle). We believe that the switch helix binding to this intermediate state (e.g. CB-PA) (Figure 6C) is the first step of TnI interaction with the TnC-NTD which still allows the C-terminal region of TnC to interact with the actin-Tm (Figure 4B). Complete dissociation of TnI-CTD from actin-Tm results in a more extensive interaction of the switch helix with TnC-NTD that prompts the full opening of the cleft (Figure 6D and E) and completes the activation of the cTF (Figure 4C).

Figure 6. Concerted action of TnC-NTD, Tni-CTD and Tm during cTF activation.

Figure 6.

Open in a new tab

(A-E) A possible mechanism of transition between CF, CB-PA, and CB-FA states. Hydrophobic residues that comprise the TnC-NTD core are highlighted in yellow, and TnI S-helix in orange (C) or red (E) helix with atoms shown only for the hydrophobic residues. The TnC-NTD helices are marked with black arrows. In (B) and (D) TnC-NTD is shown without the S-helix only for illustrative purposes, since in both CB states the S-helix is bound to TnC-NTD (Fig. 3B and C). (A) In CF state, TnC-NTD cleft is stabilized by hydrophobic interactions (yellow) that form a hydrophobic core (cyan circle). (B and C) Upon transition from CF to CB-PA the TnC-NTD cleft partially opens up (green arrow) to maintain the hydrophobic interactions between helices B and C (cyan circle). TnI-CTD S-helix binds to that interface (orange ribbons) via its hydrophobic patch (orange atoms). (D and E) The bound S-helix promotes wider opening of the cleft (green arrow) that completes TnC-NTD domain transition into CB-FA state. (F-I) Equilibriums between TnC-NTD cleft opening, TnI-CTD actin-Tm interaction and Tm position determine the activation state of the cTF. (F) In the absence of Ca2+, TnC-NTD cleft is closed which prevents the TnI-CTD S-helix from binding to the TnC-NTD and stabilizes TnI-CTD interaction with actin-Tm. (G and H) Ca2+ binding to the TnC-NTD shifts TnC cleft equilibrium towards open conformation and allows interactions with the TnI-CTD S-helix which weakens TnI-CTD interactions with actin-Tm to yield CB-PA state (G) or completely dissociates TnI-CTF from actin-Tm to yield CB-FA state (H). (I) Myosin and/or cMyBP-C interactions with the cTF moves Tm to the open/myosin state which completely shifts TnI-CTD equilibrium to the unbound state. cMyBP-C/myosin can completely activate cTF without Ca2+ by shifting Tm into open state and dissociating TnI-CTD from its actin-Tm interface which allows TnI-CTD S-helix to bind to the TnC-NTD cleft and turn it into the CB-FA conformation.

Activation of the cTF is ultimately determined by the position of the Tm cable on the actin’s surface which either blocks or permits the formation of active cross-bridges. To understand how the dynamic interaction of TnC-NTD and TnI-CTD are coupled to the Tm swing we decorated the cTF by C1 domain of cMyBP-C without Ca2+. The interactions of C1 with the cTF do not overlap with either TnI-CTD or TnC-NTD, but azimuthally shift Tm into the M-state [29]. Remarkably, we found that Tm shift into the M-state by C1 in the absence of Ca2+ causes complete dissociation of the TnI-CTD from its actin-Tm interface and converts the conformation of the Tn core from the CF into the CB-FA state (Figure 5). This directly shows that the swing of the Tm distant from the TnI-CTD interface has a dramatic effect on the conformation of TnC-NTD. This further implies that interaction of the distal C-terminal region of the TnI-CTD (e.g. h193-210) with Tm is critical for preventing Ca2+ activation of the cTF in agreement with previous results [14, 15, 17, 39]. Finally, conversion of the Tn core into the CB-FA state in the absence of Ca2+ corroborates with the observation of increased affinity of TnC for Ca2+ during contraction [30, 31]. Activation of the cTF by cMyBP-C or active cross-bridges can allosterically activate adjacent RUs by turning those into the CB-FA states, which presumably may cooperatively recruit more cross-bridges and simultaneously increase Ca2+ binding to TnC [40] promoting the completion of the twitch even at falling Ca2+ levels.

Overall, our cryo-EM work reveals that there are multiple structural states of Tn core domain in both the presence and absence of Ca2+ and shows the concerted action of TnC-NTD, TnI-CTD and Tm to the process of cTF activation. We conclude that activation of the cTF relies on two tightly coupled dynamic equilibria – opening/closing of the TnC-NTD cleft and binding/dissociation of the TnI-CTD from its actin-Tm interface (Figure 6F-I). In other words, while bound to its actin-Tm interface, TnI-CTD dynamically probes the state of the TnC-NTD by checking if its switch helix can bind into the TnC-NTD cleft in a “fly casting” fashion [38]. In the absence of Ca2+ the TnC-NTD cleft equilibrium is shifted to the closed state (Figure 6A), which prevents the TnI-CTD switch helix from binding to the TnC-NTD and, therefore, provides a stabilizing effect on the interaction of TnI-CTD with actin-Tm (Figure 6F). Ca2+ binding to TnC-NTD shifts its equilibrium towards a more open conformation and allows interactions with the TnI-CTD switch helix (Figure 6B and C). The ability of the switch helix to interact with TnC-NTD shifts TnI-CTD interactions with actin-Tm from the strongly bound (Figure 6F) to weakly bound (Figure 6G) or unbound (Figure 6H). Myosin cross-bridges and/or cMyBP-C interaction with the cTF moves the Tm cable further away from its closed position to the open or myosin state [8, 29] which presumably precludes the interaction of the TnI-CTD with Tm and completely shifts TnI-CTD equilibrium to unbound state (Figure 6I) that in turn eliminates any partially activated (e.g. CB-PA) RUs completing cTF activation. Importantly, cMyBP-C and presumably myosin cross-bridges can completely activate cTF even without Ca2+ via altering TnI-CTD interaction with its actin-Tm interface by securing Tm in the open state (Figure 6, bottom arrow). Dissociation of the TnI-CTD from actin-Tm makes its switch helix more available for binding to the TnC-NTD cleft and shifts its equilibrium towards the open state.

Materials and Methods

Proteins and Buffers:

Native porcine cardiac TFs were prepared as described in [8]. C1 was prepared as reported in [35]. A-buffer was used for cryo-EM experiments: 50 mM potassium acetate, 10 mM 3-(N-morpholino) propane sulfonic acid (MOPS), 3 mM MgCl2 and either 2.0 mM (ethyleneglycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA) for pCa 8 or 0.1 mM CaCl2 for pCa=4. The pH was adjusted to 7.0 after all of the components were mixed.

Cryo EM.

TFs:

A total of 1.8 μL of 1 μM native TFs in A-buffer in Ca2+-free (pCa 8) or high Ca2+ (pCa=4) conditions was applied to the center of the glow discharged lacey carbon grid so that they form a small droplet in the center of the grid. The small droplet was immediately blotted with Whatman Grade 1 Filter Paper for 3 to 5 s and vitrified in a Vitrobot Mark IV (FEI, Inc.). Summary for imaging conditions and image reconstruction is provided in SI Appendix Tables S1 (pCa=8) and S3 (pCa=4). Image analysis was performed using RELION [32], while modeling was done using UCSF Chimera [41] and PHENIX software suites [42]. Models validation statistics are provided in SI Appendix Tables S2 (pCa=8) and S4 (pCa=4). Experimental details are provided in the SI Materials and Methods.

TFs+C1:

A total of 1.0 μL of 1.5 μM native TFs and 0.8 μL of 5 μM C1 in A-buffer in Ca2+-free (pCa 8) conditions were mixed on the the glow discharged lacey carbon, blotted with Whatman Grade 1 Filter Paper for 4 to 5.5 s and vitrified in a Vitrobot Mark IV (FEI, Inc.). Summary for imaging conditions and image reconstruction is provided in SI Appendix Tables S5. Image analysis and modeling was performed as described for the TFs. Models validation statistics are provided in SI Appendix Table S6. Experimental details are provided in the SI Materials and Methods.

Binding of C1 and Ca2+ to fluorescently labeled reconstituted TFs.

Reconstituted wild-type fluorescent regulated thin filaments (RTFs) were prepared by standard methods published by our laboratory. Briefly, bacterially expressed, recombinant human cTnC had Cys-35 mutated to Ser and Cys-84 labeled with the fluorophore IAANS. IAANS was obtained from Molecular Probes, Plano, TX. Purification of cTnC and IAANS labeling were performed according to established methods [43]. Recombinant human troponin complex containing the IAANS-labeled cTnC was assembled according to Pinto [44]. Proper stoichiometry of the troponin complex was verified by SDS-PAGE and Coomassie blue staining. The RTFs were prepared by mixing fluorescent human recombinant cardiac troponin complex, pig cardiac tropomyosin, and F-actin from rabbit skeletal muscle acetone powder. A detailed protocol for the preparation of IAANS-labeled RTFs is described in [43]. Binding of C1 domain of MyBP-C and Ca2+ to fluorescently labeled RTF was measured by changes in steady-state fluorescence intensity using a spectrofluorometer FP-8300 PCT-818 (Jasco Analytical Instruments). IAANS fluorescence was excited at 330 nm and emission was detected at 440 nm. Measurements were performed using a 2 mL Quartz cuvette containing an initial volume of 500 μL with fluorescently-labeled RTFs at an initial concentration of 0.25 μM (measured by Coomassie Plus kit (Pierce), using bovine serum albumin as standard). Experiments were performed in a buffer consisting of 2.0 mM EGTA, 5.0 mM NTA, 1.25 mM MgCl2, 90 mM KCl, 120 mM MOPS, 1 mM freshly prepared DTT, pH 7.0 at 30°C. Addition of each aliquot of titrant was followed by 1 min pause before fluorescence intensity was recorded. Steady state fluorescence was plotted as ΔF/ΔFmax, where ΔF is the change in fluorescence after subtraction of the initial baseline value (RTF without C1 and Ca2+), and ΔFmax is the maximum fluorescence intensity recorded.

Supplementary Material

1
2
3
4
5
6
7

Highlights.

  • Thin filament troponin adopts multiple conformations at high and low Ca2+ levels

  • The two strands of thin filament are structurally diverse

  • At high Ca2+ levels thin filament exists in partially or fully activated state

  • cMyBP-C C1 domain fully activates thin filament in absence of Ca2+

  • Troponin and tropomyosin are allosterically coupled upon thin filament regulation

Acknowledgments

This work was supported by NIH grants R01 HL160966 (to V.E.G., J.R.P and P.B.C) and S10-RR025067.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Accession numbers

The atomic models have been deposited to the Protein Data Bank (www.rcsb.org) and the Electron Microscopy Data Bank (www.ebi.ac.uk/pdbe/emdb) with accession codes: 8UWW and 42681 for CF upper, 8UWY and 42683 for CF-R upper, 8UWX and 42682 for CF-T upper, 8UYD and 42800 for CF lower, 8UZ5 and 42833 for CF-R lower, 8UZ6 and 42835 for CF-T lower, 8UZX and 42846 for CB-FA upper, 8UZY and 42847 for CB-PA upper, 8V01 and 42849 for CB-FA1 lower, 8V0I and 42856 for CB-FA2 lower, 8V0K and 42858 for CB-PA lower.

Competing Interest Statement: J.R.P. provides consulting to Kate Therapeutics, but such work is unrelated to the content of this article. Other authors do not have any competing interests.

References

  • 1.Ebashi S & Endo M (1968). Calcium ion and muscle contraction. Prog Biophys Mol Biol. 18, 123–183. 10.1016/0079-6107(68)90023-0. [DOI] [PubMed] [Google Scholar]
  • 2.Ebashi S. (1972). Calcium ions and muscle contraction. Nature. 240, 217–218. 10.1038/240217a0. [DOI] [PubMed] [Google Scholar]
  • 3.Huxley HE (1972). Structural changes in the actin- and myosin-containing filaments during contraction. Cold Spring Harb. Symp. Quant. Biol 37, 361–376. [Google Scholar]
  • 4.Spudich JA, Huxley HE & Finch JT (1972). Regulation of skeletal muscle contraction. II. Structural studies of the interaction of the tropomyosin-troponin complex with actin. J. Mol. Biol 72, 619–632. [DOI] [PubMed] [Google Scholar]
  • 5.Dutta D, Nguyen V, Campbell KS, Padron R & Craig R (2023). Cryo-EM structure of the human cardiac myosin filament. Nature. 623, 853–862. 10.1038/s41586-023-06691-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Tamborrini D, Wang Z, Wagner T, Tacke S, Stabrin M, Grange M, Kho AL, Rees M, Bennett P, Gautel M & Raunser S (2023). Structure of the native myosin filament in the relaxed cardiac sarcomere. Nature. 623, 863–871. 10.1038/s41586-023-06690-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Yamada Y, Namba K & Fujii T (2020). Cardiac muscle thin filament structures reveal calcium regulatory mechanism. Nat Commun. 11, 153. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Risi CM, Pepper I, Belknap B, Landim-Vieira M, White HD, Dryden K, Pinto JR, Chase PB & Galkin VE (2021). The structure of the native cardiac thin filament at systolic Ca(2+) levels. Proc Natl Acad Sci U S A. 118. 10.1073/pnas.2024288118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.McKillop DFA & Geeves MA (1993). Regulation of the interaction between actin and myosin subfragment-1: evidence for three states of the thin filament. Biophys. J 65, 693–701. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Vibert P, Craig R & Lehman W (1997). Steric-model for activation of muscle thin filaments. Journal of Molecular Biology. 266, 8–14. [DOI] [PubMed] [Google Scholar]
  • 11.Jackson P, Amphlett GW & Perry SV (1975). The primary structure of troponin T and the interaction with tropomyosin. Biochem J. 151, 85–97. 10.1042/bj1510085. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Pearlstone JR & Smillie LB (1977). The binding site of skeletal alpha-tropomyosin on troponin-T. Can J Biochem. 55, 1032–1038. 10.1139/o77-154. [DOI] [PubMed] [Google Scholar]
  • 13.Risi CM, Belknap B, White HD, Dryden K, Pinto JR, Chase PB & Galkin VE (2023). High-resolution cryo-EM structure of the junction region of the native cardiac thin filament in relaxed state. PNAS Nexus. 2, pgac298. 10.1093/pnasnexus/pgac298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Narolska NA, Piroddi N, Belus A, Boontje NM, Scellini B, Deppermann S, Zaremba R, Musters RJ, dos Remedios C, Jaquet K, Foster DB, Murphy AM, van Eyk JE, Tesi C, Poggesi C, van der Velden J & Stienen GJ (2006). Impaired diastolic function after exchange of endogenous troponin I with C-terminal truncated troponin I in human cardiac muscle. Circ Res. 99, 1012–1020. 10.1161/01.RES.0000248753.30340.af. [DOI] [PubMed] [Google Scholar]
  • 15.Foster DB, Noguchi T, VanBuren P, Murphy AM & Van Eyk JE (2003). C-terminal truncation of cardiac troponin I causes divergent effects on ATPase and force: implications for the pathophysiology of myocardial stunning. Circ Res. 93, 917–924. 10.1161/01.RES.0000099889.35340.6F. [DOI] [PubMed] [Google Scholar]
  • 16.Rarick HM, Tu XH, Solaro RJ & Martin AF (1997). The C terminus of cardiac troponin I is essential for full inhibitory activity and Ca2+ sensitivity of rat myofibrils. J Biol Chem. 272, 26887–26892. 10.1074/jbc.272.43.26887. [DOI] [PubMed] [Google Scholar]
  • 17.Galinska A, Hatch V, Craig R, Murphy AM, Van Eyk JE, Wang CL, Lehman W & Foster DB (2010). The C terminus of cardiac troponin I stabilizes the Ca2+-activated state of tropomyosin on actin filaments. Circ Res. 106, 705–711. 10.1161/CIRCRESAHA.109.210047. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Lehman W, Pavadai E & Rynkiewicz MJ (2021). C-terminal troponin-I residue strap tropomyosin in the muscle thin filament blocked-state. Biochem Biophys Res Commun. 551, 27–32. 10.1016/j.bbrc.2021.03.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Li MX, Wang X & Sykes BD (2004). Structural based insights into the role of troponin in cardiac muscle pathophysiology. J Muscle Res Cell Motil. 25, 559–579. 10.1007/s10974-004-5879-2. [DOI] [PubMed] [Google Scholar]
  • 20.Kobayashi T & Solaro RJ (2005). Calcium, thin filaments, and the integrative biology of cardiac contractility. Annu Rev Physiol. 67, 39–67. 10.1146/annurev.physiol.67.040403.114025. [DOI] [PubMed] [Google Scholar]
  • 21.Stevens CM, Rayani K, Singh G, Lotfalisalmasi B, Tieleman DP & Tibbits GF (2017). Changes in the dynamics of the cardiac troponin C molecule explain the effects of Ca(2+)-sensitizing mutations. J Biol Chem. 292, 11915–11926. 10.1074/jbc.M116.770776. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Li MX, Spyracopoulos L & Sykes BD (1999). Binding of cardiac troponin-I147-163 induces a structural opening in human cardiac troponin-C. Biochemistry. 38, 8289–8298. 10.1021/bi9901679. [DOI] [PubMed] [Google Scholar]
  • 23.Dong WJ, Xing J, Villain M, Hellinger M, Robinson JM, Chandra M, Solaro RJ, Umeda PK & Cheung HC (1999). Conformation of the regulatory domain of cardiac muscle troponin C in its complex with cardiac troponin I. J Biol Chem. 274, 31382–31390. 10.1074/jbc.274.44.31382. [DOI] [PubMed] [Google Scholar]
  • 24.Rynkiewicz MJ, Pavadai E & Lehman W (2022). Modeling Human Cardiac Thin Filament Structures. Front Physiol. 13, 932333. 10.3389/fphys.2022.932333. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Pavadai E, Rynkiewicz MJ, Yang Z, Gould IR, Marston SB & Lehman W (2022). Modulation of cardiac thin filament structure by phosphorylated troponin-I analyzed by protein-protein docking and molecular dynamics simulation. Arch Biochem Biophys. 725, 109282. 10.1016/j.abb.2022.109282. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Mason AB, Tardiff JC & Schwartz SD (2022). Free-Energy Surfaces of Two Cardiac Thin Filament Conformational Changes during Muscle Contraction. J Phys Chem B. 126, 3844–3851. 10.1021/acs.jpcb.2c01337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Risi C, Eisner J, Belknap B, Heeley DH, White HD, Schroder GF & Galkin VE (2017). Ca2+-induced movement of tropomyosin on native cardiac thin filaments revealed by cryoelectron microscopy. Proc. Natl. Acad. Sci. U. S. A 114, 6782–6787. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Swartz DR, Moss RL & Greaser ML (1996). Calcium alone does not fully activate the thin filament for S1 binding to rigor myofibrils. Biophys J. 71, 1891–1904. 10.1016/S0006-3495(96)79388-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Risi C, Belknap B, Forgacs-Lonart E, Harris SP, Schroder GF, White HD & Galkin VE (2018). N-Terminal Domains of Cardiac Myosin Binding Protein C Cooperatively Activate the Thin Filament. Structure. 26, 1604–1611. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Bremel RD & Weber A (1972). Cooperation within actin filament in vertebrate skeletal muscle. Nat. New Biol 238, 97–101. [DOI] [PubMed] [Google Scholar]
  • 31.Zot AS & Potter JD (1989). Reciprocal coupling between troponin C and myosin crossbridge attachment. Biochemistry. 28, 6751–6756. [DOI] [PubMed] [Google Scholar]
  • 32.Scheres SH (2012). RELION: implementation of a Bayesian approach to cryo-EM structure determination. J Struct Biol. 180, 519–530. 10.1016/j.jsb.2012.09.006j.bpj.2020.05.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Pavadai E, Lehman W & Rynkiewicz MJ (2020). Protein-Protein Docking Reveals Dynamic Interactions of Tropomyosin on Actin Filaments. Biophys J. 119, 75–86. 10.1016/j.bpj.2020.05.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Spyracopoulos L, Li MX, Sia SK, Gagne SM, Chandra M, Solaro RJ & Sykes BD (1997). Calcium-induced structural transition in the regulatory domain of human cardiac troponin C. Biochemistry. 36, 12138–12146. 10.1021/bi971223d. [DOI] [PubMed] [Google Scholar]
  • 35.Harris SP, Belknap B, Van Sciver RE, White HD & Galkin VE (2016). C0 and C1 N-terminal Ig domains of myosin binding protein C exert different effects on thin filament activation. Proc. Natl. Acad. Sci. U. S. A 113, 1558–1563. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Risi CM, Villanueva E, Belknap B, Sadler RL, Harris SP, White HD & Galkin VE (2022). Cryo-Electron Microscopy Reveals Cardiac Myosin Binding Protein-C M-Domain Interactions with the Thin Filament. J Mol Biol. 434, 167879. 10.1016/j.jmb.2022.167879. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Siddiqui JK, Tikunova SB, Walton SD, Liu B, Meyer M, de Tombe PP, Neilson N, Kekenes-Huskey PM, Salhi HE, Janssen PM, Biesiadecki BJ & Davis JP (2016). Myofilament Calcium Sensitivity: Consequences of the Effective Concentration of Troponin I. Front Physiol. 7, 632. 10.3389/fphys.2016.00632. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Hoffman RM, Blumenschein TM & Sykes BD (2006). An interplay between protein disorder and structure confers the Ca2+ regulation of striated muscle. J Mol Biol. 361, 625–633. 10.1016/j.jmb.2006.06.031. [DOI] [PubMed] [Google Scholar]
  • 39.Meyer NL & Chase PB (2016). Role of cardiac troponin I carboxy terminal mobile domain and linker sequence in regulating cardiac contraction. Arch Biochem Biophys. 601, 80–87. 10.1016/j.abb.2016.03.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Moss RL, Razumova M & Fitzsimons DP (2004). Myosin crossbridge activation of cardiac thin filaments: implications for myocardial function in health and disease. Circ Res. 94, 1290–1300. 10.1161/01.RES.0000127125.61647.4F. [DOI] [PubMed] [Google Scholar]
  • 41.Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC & Ferrin TE (2004). UCSF Chimera--a visualization system for exploratory research and analysis. J. Comput. Chem 25, 1605–1612. [DOI] [PubMed] [Google Scholar]
  • 42.Liebschner D, Afonine PV, Baker ML, Bunkoczi G, Chen VB, Croll TI, Hintze B, Hung LW, Jain S, McCoy AJ, Moriarty NW, Oeffner RD, Poon BK, Prisant MG, Read RJ, Richardson JS, Richardson DC, Sammito MD, Sobolev OV, Stockwell DH, Terwilliger TC, Urzhumtsev AG, Videau LL, Williams CJ & Adams PD (2019). Macromolecular structure determination using X-rays, neutrons and electrons: recent developments in Phenix. Acta Crystallogr D Struct Biol. 75, 861–877. 10.1107/S2059798319011471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Dweck D, Hus N & Potter JD (2008). Challenging current paradigms related to cardiomyopathies. Are changes in the Ca2+ sensitivity of myofilaments containing cardiac troponin C mutations (G159D and L29Q) good predictors of the phenotypic outcomes? J Biol Chem. 283, 33119–33128. 10.1074/jbc.M804070200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Pinto JR, Parvatiyar MS, Jones MA, Liang J, Ackerman MJ & Potter JD (2009). A functional and structural study of troponin C mutations related to hypertrophic cardiomyopathy. J Biol Chem. 284, 19090–19100. 10.1074/jbc.M109.007021. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1973614-supplement-1.gif (131.4KB, gif)
2
NIHMS1973614-supplement-2.gif (173.2KB, gif)
3
NIHMS1973614-supplement-3.gif (183KB, gif)
4
NIHMS1973614-supplement-4.gif (243.4KB, gif)
5
NIHMS1973614-supplement-5.gif (328.6KB, gif)
6
NIHMS1973614-supplement-6.gif (1.6MB, gif)
7
NIHMS1973614-supplement-7.docx (12.8MB, docx)

RESOURCES