2.1. Evolutionary Significance of TnC N-Helix
TnCs, along with myosin light chains, parvalbumins and calmodulin (CaM) itself, are part of the calmodulin family of EF-hand, Ca
2+-binding proteins [
81]. Interestingly, of the large number of proteins in the calmodulin family, only vertebrate TnCs have sequences that correspond to an N-helix, the one α-helix of TnC that is not directly involved in the formation of an EF-hand [
81,
82,
83]. This can be seen in
Figure 2, which illustrates the presence of an N-helix (violet)—adjacent to the A-helix (dark red) and B-helix (red) that surround EF-hand site I—in sTnC (
Figure 2a) and cTnC (
Figure 2b), while there is no sequence or structure comparable to the N-helix preceding the first helix-loop-helix domain in CaM (
Figure 2c).
While the evolutionary origin of the N-helix of TnC is not known, it is unique to vertebrate striated muscles [
81,
83]. There is diversity among invertebrate TnC’s in regard to which of the four EF-hands is functionally capable of binding divalent cations or presumed to be functional based upon sequence analysis [
84,
85,
86,
87]. In stark contrast to vertebrate TnC discussed above, the C-lobe of TnC appears to be more significant than the N-lobe for Ca
2+-activation of most invertebrate striated muscles [
11,
84,
85,
87]. Thus, it seems likely that the N-helix could have a significant role in Ca
2+-activation of vertebrate striated muscle, although the exact nature of that role could differ between cardiac and skeletal TnC. The α-helical structures at the N-termini of sTnC and cTnC appear similar (
Figure 2a,b), but the sequences differ among species, particularly for sTnC, and isoforms within a species [
47,
48,
49,
81,
83]. The latter point is illustrated in
Figure 3 for the human genetic isoforms. Among the first 12 amino acids, only four are identical (
Figure 3) and one of those four is the N-terminal Met that is typically removed post-translationally from sTnC but not cTnC of vertebrate species [
48,
49,
76,
81,
88,
89,
90,
91,
92]. Protein sequencing indicates that the N-terminus of both cTnC and sTnC is further modified post-translationally by acetylation of the remaining N-terminal amino group [
48,
49,
76,
88,
89,
90,
91] which neutralizes what would otherwise be the positive charge of the N-terminus of the mature polypeptide. These differences in primary sequence between the N-helices of sTnC and cTnC are substantial enough to suggest that they may not be functionally identical. Thus, the exact nature of the role of the N-helix in Ca
2+-activation of vertebrate striated muscle could differ between cardiac and skeletal TnC.
2.2. Structural Evidence for the Functional Significance of the TnC N-Helix
High resolution structural studies have provided important insights into Ca2+-activation of the thin filament by applying an evolving variety of structural determination methods to elucidate the structure of TnC. The structures include isolated TnC, TnC associated with portions of or all of the troponin complex, and TnC within the thin filament, in the absence and presence of divalent cations in the N-lobe site (cTnC) or sites (sTnC).
X-ray crystallography provided the earliest structures of native sTnC purified from chicken or turkey skeletal muscles [
55,
56,
57]. These studies supplied structural details not available in an earlier report of diffraction from crystals of rabbit and chicken sTnC [
93]. All of the earliest crystal structures had two metal ions bound to the C-lobe, with 2 Ca
2+ in the turkey sTnC structures [
55,
56] and either 2 Mn
2+, 2 Nd
3+ or 2 Au
3+ in the chicken sTnC structures [
57,
93]. In these structures, the apo N-lobe and metal ion-bound C-lobe take on similar but distinct conformations that allowed for predictions about the structural changes that occur upon regulatory Ca
2+ binding in the N-lobe. Almost a decade later, Houdusse et al. [
60] filled in the gap by obtaining crystal structures of recombinantly expressed chicken sTnC with 4 Ca
2+ bound, thereby demonstrating experimentally what could only be speculated upon in earlier studies. None of these studies provided a structure with physiological Mg
2+ bound at the C-lobe.
The D/E-linker is of particular note among all of these sTnC structures because it forms a single, extended α-helix separating the N- and C-lobes. When analogizing sTnC structure to a dumbbell, the central D/E-linker forms the handlebar between the N- and C-lobe masses at either end (
Figure 2a) [
55,
56,
57,
60]. The linear D/E-linker separating the N- and C-lobes in all of these sTnC structures could have resulted from the conditions of crystallization (e.g., acidic pH ~ 5) and structural constraints within the crystal lattices formed. In addition, the structures may have been affected by the absence of the other troponin complex proteins, especially the portions of TnI that bind the N- and C-lobes when divalent cations are present [
9,
94,
95,
96,
97,
98]. An X-ray crystallography study by Saijo et al. [
99] demonstrated flexibility of the D/E-linker of recombinant rabbit sTnC in the 2 Ca
2+ state (i.e., apo N-lobe) crystallized at basic pH (pH ~ 8) in the presence of sTnI
1-47 that binds the C-lobe. Solution NMR spectroscopy further demonstrated flexibility of the D/E-linker of Ca
2+-saturated recombinant chicken sTnC (4 Ca
2+) [
59] and recombinant Cys-less chicken cTnC (3 Ca
2+) [
61]. In solution, i.e., without the constraints of a crystal lattice, the N- and C-lobe motions are essentially independent—recognizing that they are not completely independent because they are covalently linked to each other—and thus for many purposes the N- and C-lobes can be studied as separate constructs [
94,
95,
96,
100,
101,
102]. While the structures discussed above included the N-helix, a major focus was on divalent cation-induced structural changes within the N- and C-lobes, and little attention was paid to the N-helix. It was even concluded from solution NMR studies that the absence of the N-helix had little or no effect on the proteolytically isolated, apo N-lobe of native turkey sTnC [
100,
103]. A similar conclusion was obtained from lower resolution circular dichroism studies on recombinant chicken sTnC [
54], although the thermal stability of chicken and rabbit sTnC structure was reduced in the absence of the N-helix regardless of divalent cation binding [
51,
54].
Building upon the work of Saijo et al. [
99], Takeda et al. [
62] determined the X-ray crystal structure of Ca
2+-saturated (3 Ca
2+) cardiac troponin core domain. They showed a collapsed—albeit not fully resolved—DE-linker. In the Takeda et al. [
62] structure, the N-helix of cTnC’s N-lobe is in close proximity to the C-lobe and, more specifically, the portion of cTnI that is bound to the C-lobe. Within two years of the Takeda et al. publication [
62], structures became available for the skeletal troponin core domain in both the Ca
2+-free, Mg
2+-saturated state (apo N-lobe and 2 Mg
2+ C-lobe) as well as the Ca
2+-saturated (4 Ca
2+) state [
79]. In contrast to the Ca
2+-saturated cardiac troponin core domain of Takeda et al. [
62], the Ca
2+-saturated (4 Ca
2+) skeletal troponin core domain exhibited an elongated, α-helical D/E-linker [
79] comparable to that observed in the crystal structures of sTnC with only 2 Ca
2+ bound to the C-lobe [
55,
56,
57,
60]. However, the D/E-linker of the skeletal troponin core domain becomes disordered in the Ca
2+-free (2 Mg
2+) state. Ca
2+-saturated TnC from these crystal structures of the skeletal and cardiac troponin complexes are shown in
Figure 2a,b, respectively, illustrating what was inferred from sequence differences between the N-helices of sTnC and cTnC—that the role of the N-helix in Ca
2+-activation of vertebrate striated muscle could differ between cTnC and sTnC.
With the advancement of cryo-EM—now the ‘gold-standard’ technology of macromolecular structure analysis—we have been able to visualize structural changes within the thin filament upon Ca
2+ binding to cTnC, and to more deeply analyze its regulatory mechanism [
19,
20,
22,
104,
105]. While an individual structural model determined by cryo-EM represents a single structural state, capture of multiple structural states allows cryo-EM to be an excellent technique for the study of thin filament structural dynamics.
Current cryo-EM-based models of the thin filament place the N-helix of cTnC close to three structural components of cTn: cTnC D-helix associated with regulatory Ca
2+-binding at EF-hand site II; the D/E-linker connecting the N- and C-lobes of cTnC; and cTnI helix H1 (residues 42-80), located in the N-terminal portion of cTnI that binds to the C-lobe of cTnC, which anchors cTnC within cTn (
Figure 1,
Figure 4 &
Supplemental movies) [
19,
20,
22]. This central location and the altered interactions with other portions of the troponin complex upon Ca
2+ binding strongly imply that cTnC N-helix could play a critical role in Ca
2+-regulation of cardiac contraction by widely transmitting information about regulatory Ca
2+-binding at the cTnC N-lobe to other parts of the cardiac thin filament.
2.3. Biophysical Studies Demonstrate a Critical Role for the N-Helix of sTnC in Normal Ca2+-Regulation of Skeletal Muscle, But Evidence for That of cTnC Is Lacking
Functional assays have been employed to address the question of the functional role(s) of TnC’s N-helix in light of the evidence presented above that (i) evolution has uniquely favored this structural element in vertebrate TnC’s among calmodulin family proteins [
60,
76,
81,
82,
83,
88] and (ii) that the N-helix is situated between TnC’s N-lobe and other parts of troponin on the thin filament [
19,
20,
22,
104], while in contrast (iii) that removal of the N-helix has little effect on the secondary structure of TnC [
54,
100,
103]. This research involving sTnC from multiple species has shown that the N-helix does play an essential role in both structure and function of sTnC [
51,
52,
53,
54,
106,
107].
Functionally, the absence of sTnC’s N-helix markedly decreases the Ca
2+-affinity of TnC’s regulatory N-lobe. Solution studies have shown that the absence of the N-helix caused a 3-fold change in K
d for Ca
2+-binding at chicken sTnC N-lobe—with or without an F29W substitution that provides a fluorescence readout of Ca
2+-specific changes in structure [
108]—with little or no change in higher-affinity binding of Ca
2+ or Mg
2+ at the C-lobe [
53,
54]. The absence of sTnC’s N-helix in reconstituted thin filaments not only reduced the apparent Ca
2+-sensitivity by more than 2-fold for actomyosin solution MgATPase activity, but also caused a marked reduction in the maximum activity [
53,
54]. Interestingly, removal of only part (~½) of the N-helix had little effect on Ca
2+-binding to sTnC in solution, but was still associated with a reduction in Ca
2+-sensitivity of actomyosin MgATPase activity [
52]. Reversing the scenario to examine a non-muscle system that normally relies on CaM, either sTnC or sTnC that is missing the N-helix can replace CaM to activate erythrocyte Ca
2+-ATPase, albeit with a requirement for substantially higher Ca
2+ levels [
109]; in this non-muscle system, less Ca
2+ is required with sTnC that is missing the N-helix compared to sTnC—the opposite of what is observed in muscle—and markedly higher concentrations of protein are required for either TnC relative to CaM.
A study followed up on the observation that CaM can functionally replace sTnC to regulate isometric force generation by skeletal muscle [
110] even though it does not integrate fully into the troponin complex. The fact that CaM does not serve well as a structural replacement of TnC in the troponin complex suggests that one possible role of the N-helix of TnC is to maintain TnC bound to the thin filament. The same group also utilized a recombinant rabbit sTnC construct with the N-helix missing [
51,
107] to test its impact on skeletal muscle force and found that the highest Ca
2+-activated force of skeletal muscle was reduced when the N-helix was missing [
51,
107]. This effect was similar to CaM, which also reduced Ca
2+-sensitivity of steady-state isometric force [
110]. However, the authors did not attribute these CaM-associated differences exclusively to the missing N-helix (
Figure 2), but instead to incomplete occupancy of thin filament regulatory units by CaM. Interestingly, invertebrate striated muscle where the N-helix of TnC is absent also has reduced Ca
2+-sensitivity compared to vertebrate muscle [
84].
Removal of chicken sTnC N-helix was associated with a marked reduction in Ca
2+-sensitivity of steady-state isometric force generation using recombinantly expressed proteins reconstituted into rabbit skeletal fibers [
53,
106]. The deletion of the N-helix in these studies was associated with a small reduction in maximum Ca
2+-activated force; the difference with studies described above may be due to higher Ca
2+ concentrations when studying the N-helix deletion [
53,
106]. N-helix removal had little or no effect on cooperativity of steady-state Ca
2+-binding, MgATPase activity, or isometric force generation, as assessed by the Hill coefficient [
53,
54,
106,
111]. Beyond steady-state phenomena, the Ca
2+-dependence of the kinetics of isometric tension redevelopment (
kTR) were shifted rightward to a similar extent as steady-state isometric force such that there was little effect on the relationship between force and
kTR [
106].
Based on the discussion above, the N-helix plays a significant role in maintaining the structural stability of TnC and the Ca
2+-regulated contraction in skeletal muscle. However, comparable studies have not yet been performed with cTnC. Although the N-helix is evolutionarily favored by vertebrate striated muscle, it is notable (per sequence analysis described in Sec. 2.1 above) that the sequence of cTnC N-helix is not the same as that for sTnC; and also (per structure analysis described in Sec. 2.2 above, esp.
Figure 4) that the sTnC N-helix is not tucked into a central location like that of cTnC. Additionally, evidence also supports the structural studies indicating that the N-helix is close to the C-lobe at Ca
2+-saturated state in the tertiary structure of cTnC (
Supplemental movies) [
19,
20,
61,
62], which is not true for sTnC [
79].
Secondly, previous research also shows that the second half of N-helix is sufficient to maintain normal sTnC function [
52], but this is not true for the cardiac isoform. Multiple amino acid variants spanning the whole helix region are reported to be related to all three major types of cardiomyopathy [
112,
113,
114,
115], which will be further discussed in the following section. Thus, it is unreasonable to directly presume that the N-helix of the two TnC isoforms are functionally identical.
2.4. Pathophysiological Evidence for the Significance of the TnC N-Helix: Variants in the N-Helix of Troponin C Are Associated with Human Cardiomyopathies
Although the cTnC N-helix only accounts for 12 amino acids out of 161 residues in total, multiple mutations in this region have been linked to three major types of human cardiomyopathy: dilated cardiomyopathy (DCM), hypertrophic cardiomyopathy (HCM) and restrictive cardiomyopathy (RCM) (
Figure 5), indicating the importance of the N-helix in regulation of cardiac contraction [
112,
116,
117,
118,
119,
120,
121,
122,
123,
124,
125,
126,
127,
128,
129,
130,
131,
132,
133,
134,
135,
136,
137,
138]. For these reasons, attention should be directed towards understanding how the cTnC N-helix participates in the physiological function of the cardiac thin filament and the genesis of cardiomyopathy.
Among the seven mutations identified in the N-helix of cTnC shown in
Figure 5, four are published in peer-reviewed journals, and the A8V variant is the most studied. Clinical evidence has shown that the variant relates to both HCM and RCM in human, with different inheritance patterns: autosomal dominant for HCM, while autosomal recessive for RCM [
116,
117,
139,
140]. The two-copy variant usually results in an earlier disease onset and poorer prognosis compared with the single-copy variant [
140], which is consistent with the data obtained from the murine model [
121]. Mechanistic studies examining the A8V variant of cTnC N-helix reveal that this mutation may strengthen the binding of the cTnI switch helix to cTnC, which enhances the Ca
2+ affinity of myofibrils without affecting that of isolated cTnC [
121,
123]. A subsequent study that focused on the same variant showed that the N-helix also appears to communicate with the D-helix of cTnC (part of activating site II), increasing the Ca
2+ sensitivity of isometric force generation and modulating both the number of and rate of cycling crossbridges [
128]. These mechanical effects are consistent with the A8V variant’s association with HCM, and the structural interpretation is supported by cryo-EM structures shown in
Figure 1. Interestingly, another study reported the presence of a sexually dimorphic transcriptome of the A8V variant suggesting a possible novel role of the N-helix in promoting the disease phenotype [
130].
Additional evidence suggests that the N-terminal helix of cTnC is close in tertiary structure to the intrinsically disordered C-terminus of cTnT—a structural element that remains unresolved in existing Ca
2+-bound cryo-EM structures. This interaction is strengthened in the presence of the I4M pathogenic variant of the cTnC N helix, a variant that is associated with pediatric DCM, probably by negatively influencing the allosteric regulation of Ca
2+ activation through the cTnC-cTnI switching mechanism [
129].
2.5. Cardiomyopathy Variants May Alter Communication Between the cTnC N-Helix and Other Parts of Troponin
In the cardiac isoform of vertebrate TnC, there are only three divalent cation binding sites, with two sites (site III and site IV) in the C-lobe always occupied and participating in the critical structural role of the C-lobe—binding TnC to TnI—under physiological conditions. In contrast, the remaining site (site II) in the N-lobe responsive to variations in cytoplasmic Ca
2+ concentration (
Figure 1). The D/E linker is an obvious pathway for communication between the N- and C-lobes of cTnC. Measurements of protein dynamics in solution have shown that the HCM-related cTnC variant (D145E) could enhance the Ca
2+ affinity of site II in N-lobe, probably by abnormal hydrophobic surface exposure, although the Ca
2+-binding sites in the C-lobe are destabilized [
78,
141,
142].
The cryo-EM structural evidence described above points to the N-helix as an additional pathway. As shown in
Figure 4 and
Supplemental movies, the N-helix is located at the center of the cTnC D-helix, D/E-linker and cTnI helix H1 in both Ca
2+-free and Ca
2+-bound states, indicating the possibility of the N-helix as a ‘communicator’ not only between lobes of cTnC, but also subunits of troponin. Thus, this location of the N-helix is likely important for the mechanism by which Ca
2+-initiated activation is transmitted from site II in the N-lobe of cTnC to other parts of the thin filament. This presumption is further supported by the pathophysiological research on the A8V variant, which is associated with both HCM and RCM (see Sec. 2.4). Furthermore, the N-helix of cTnC is also presumed to be close to the intrinsically disordered C-terminus of cTnT [
129].