C-Terminal Truncation of Cardiac Troponin I Causes Divergent Effects on ATPase and Force

Proteins and Peptides

cTnI, its C-terminal mutant (cTnI_1-192), and/or two synthetic peptides used in this study are numbered with respect to the primary amino acid sequence of human cTnI. cTnI_193-209 corresponds to the sequence of human cTnI equivalent to the portion cleaved in the stunned rat heart. A longer synthetic peptide, cTnI_188-209, was also synthesized to allow detection of any binding site overlapping residue 193 that might be lost or weakened in the shorter peptide. The peptides were prepared using standard F-moc synthesis. Human recombinant cTnI, cTnI_1-192, and cTnT were purified by ion exchange chromatography. Human recombinant cardiac TnC (cTnC) was supplied by Spectral Diagnostics (Toronto, Ontario, Canada). Purification of rabbit skeletal actin, rabbit back and/or chicken pectoral skeletal muscle myosin, bovine cardiac and rabbit skeletal αα-tropomyosin (Tm) is described in the online data supplement. Refolding of human recombinant cardiac troponin (cTn) from its constituent subunits was conducted as described previously.^17,18 Animals were used in compliance with the Animals for Research Act (Province of Ontario), the guidelines of the Canadian Council on Animal Care, and the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication No. 85-23, revised 1985).

Affinity Chromatography

Human recombinant cTnC (2 mg/mL) and cTnT (2 mg/mL) were coupled to CNBr-activated Sepharose 4B (Pharmacia). A mixture of cTnI and cTnI_1-192 (50 μg each), or each synthetic peptide individually, was applied to cTnC or cTnT affinity columns in equilibration buffer (10 mmol/L imidazole pH 7.5, 20 mmol/L NaCl, 1 mmol/L EGTA, 1 mmol/L DTT, and 0.01% NaN₃). The peptides and proteins were eluted using a gradient (linear or step) of increasing NaCl, followed by a urea gradient up to 8 mol/L urea. Visualization of cTnI or cTnI_1-192 was carried out by Coomassie Blue–stained 12% SDS-polyacrylamide gels, while the peptides cTnI_193-209 and cTnI_189-209 were analyzed by reverse-phase HPLC (Varian) as previously reported.²¹ Details are found in the online data supplement.

Actin-Binding Studies

F-actin (2 nmol) was mixed with 0.4 nmol of Tm and increasing quantities of cTnI or cTnI_1-192, in 200 μL of binding buffer. Protein mixtures were allowed to equilibrate for 30 minutes before centrifugation at 107 000g in a Beckman TL-100 ultracentrifuge at room temperature. This pelleted 90% to 99% of the F-actin. Pellets were rinsed, once gently, with binding buffer, and dissolved in 100 mL of reducing SDS sample buffer (New England Biolabs). Pellets and supernatants were analyzed by 12% SDS-PAGE. Gels were stained with Coomassie Blue. Densitometric analysis was conducted using Scanplot v5.06 (Cunningham Software).

Determination of ATPase Activity

Assays were conducted in a 96-well ELISA plate (200 μL of reaction volume). Inorganic phosphate was determined colorimetrically as described by Chifflet et al.¹⁹ All assays were performed in triplicate; means were plotted as standard deviations. Details are found in the online data supplement.

In Vitro Motility Assay

Thin filaments were reconstituted as previously described,²⁰ labeled with rhodamine-phalloidin at 1:1 actin/phalloidin ratio, and stored overnight at 4°C before use in the in vitro motility assay. The in vitro motility assay has been previously described in detail.^20–22 Statistical comparisons were performed, from the parameters of the fit, by use of the t test. All values are expressed as mean±SE. For further detail, see the online data supplement.

A modified in vitro motility assay was used to measure the effects of cTnI truncation on an index of isometric force. VanBuren et al²³ and others²⁴ have characterized this technique, in detail, with consistent results. Briefly, α-actinin was attached to a myosin S1-decorated nitrocellulose coverslip surface (15 to 100 μg/mL) followed by a bovine serum albumin wash (0.5 mg/mL). Reconstituted thin filaments (10 to 20 nmol/L) were then added to the motility surface. Motility was initiated with the addition of low-salt motility buffer (see online data supplement). Relative isometric force was defined as the minimum amount of α-actinin needed to completely arrest thin-filament motility. Force:pCa relations were then determined. Force:pCa data were fit and statistical significance determined as described for the velocity data. For more details, see the online data supplement.

An expanded Materials and Methods section can be found in the online data supplement available at http://www.circresaha.org.

Affinity Chromatography

The aim of these experiments was to assess any large-scale change in affinity between the Tn subunits and whether correct stoichiometry would exist under the conditions of the A-Tm-S1 ATPase and in vitro motility assays, and to test whether C-terminal peptides harbored binding determinants for cTnT or cTnC. The Table summarizes the results obtained from binding of cTnI, cTnI_1-192, and the synthetic peptides to the cTnT and cTnC affinity columns. As shown in Figure 1, cTnI and cTnI_1-192 did not elute from the cTnT-Sepharose column in the presence of 1 mol/L NaCl, but rather, elution required denaturation with 8 mol/L urea. In fact, cTnI and cTnI_1-192 eluted together in a urea gradient; this suggests that the affinity of cTnI_1-192 for cTnT is equal or very similar to that of cTnI. The interaction between cTnI and cTnC, in the presence of Ca²⁺, was sufficiently strong that cTnI failed to elute from the cTnC-Sepharose column in 1 mol/L NaCl or 8 mol/L urea. Elution of either cTnI or cTnI_1-192 required the addition of EDTA to chelate the Ca²⁺ (Table), indicating the Ca²⁺-sensitive interaction between cTnI and cTnC was not overtly affected by deletion of the C-terminus of cTnI. Furthermore, neither synthetic peptide that comprised the amino acid sequence from the C-terminus of cTnI bound the cTnT- or cTnC-Sepharose affinity columns (Table). Taken together, these data demonstrate that there is no direct interaction between the C-terminus of cTnI and cTnC or cTnT, and that the C-terminal deletion does not grossly disrupt the global interaction between cTnI and the other Tn subunits under these experimental conditions. Nor does deletion of the C-terminus of TnI cause any impediment to the formation of binary cTnI-cTnC complexes or refolded ternary troponin complexes that were used in subsequent functional assays.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Affinity chromatography. C-terminus of cTnI does not affect interaction with cTnT. Approximately equimolar amounts of cTnI and cTnI_1-192 were applied to a cTnT-Sepharose equilibrated as described in Materials and Methods. The column was washed with equilibration buffer before high-salt buffer was applied. A urea gradient followed the high-salt wash. Fractions (1.5 mL) were analyzed by 12% SDS-PAGE. Gels were stained with Coomassie Blue. The bulk (>90%) of the cTnI and cTnI_1-192 remained bound to the column in the presence of 1 mol/L NaCl and could only be eluted under denaturing conditions.

Actin Binding and Effects on A-Tm-S1 ATPase Activity

cTnI and cTnI_1-192 displayed equal affinity for actin (F-actin), as assessed by co-sedimentation on ultracentrifugation (Figure 2). The affinity of both cTnI and cTnI_1-192 for actin increased in the presence of Tm, consistent with a previous report.¹⁶ Under the conditions used in this assay (100 mmol/L NaCl), both cTnI and cTnI_1-192 caused actin to form thick bundles (determined by electron microscopy; not shown), when added in excess of F-actin or F-actin-Tm, which precluded proper determination of the maximum binding stoichiometry. TnI-induced aggregation of actin is a complication reported by others,¹⁶ but has little impact on the primary observations that the Tm-dependent enhancement of actin binding is preserved in TnI_1-192, and that C-terminal truncation does not grossly mitigate binding of cTnI to actin.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Ultracentrifugation. C-terminus of cTnI does not alter binding to filamentous actin or actin-Tm. Increasing concentrations of cTnI and cTnI_1-192 were incubated with F-actin or actin-Tm (5:1 mole ratio) in buffer containing 100 mmol/L NaCl and 5 mmol/L MgCl₂ and were ultracentrifuged to pellet both F-actin and any bound protein. Both pellet (P) and supernatant (S, unbound protein) were analyzed. A, Representative Coomassie Blue–stained gels showing protein composition of the supernatants and pellets. B, Gels from panel A, and a duplicate set of gels, were analyzed by densitometry. cTnI (•) and cTnI_1-192 (▿) values are shown. Each data point is an average (and range) of the values obtained from the two sets of gels.

It is well documented that binding of cTnI to actin-Tm inhibits A-Tm-S1 ATPase activity by inhibiting binding of myosin S1 to actin.^16,25 Figure 3A depicts the concentration-dependent inhibition of A-Tm-S1 ATPase activity by cTnI and cTnI_1-192. Both cTnI and cTnI_1-192 inhibited the A-Tm-S1 ATPase activity to the same extent at a 1:1 mole ratio of cTnI to actin. At intermediate concentrations, cTnI_1-192 was a marginally less effective inhibitor than cTnI, with IC₅₀ values of 0.32 nmol and 0.27 nmol, respectively, and implies reduced affinity of cTnI_1-192 for actin-Tm. The reduced affinity was not detected in the ultracentrifugation studies and thus may reveal sensitivity differences between the two types of experiments. Importantly, however, once cTnI_1-192 is bound to actin, it inhibits A-Tm-S1 ATPase activity as effectively as full-length cTnI.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Inhibition of A-Tm-S1 ATPase activity by cTnI and cTnI_1-192 and its reversal by cTnC. A, A-Tm-S1 ATPase activity was measured as a function of increasing concentration of cTnI (•) or cTnI_1-192 (○). B, cTnI (•) and cTnI_1-192 (○) were added to actin-Tm such that the mole ratio of cTnI (or cTnI_1-192):actin:Tm was 5:1:1. Reversal of this inhibition was determined by increasing the concentrations of cTnC in the presence of Ca²⁺. A-Tm-S1 ATPase activity is taken to be 100%. Averages and their standard deviations were calculated from triplicate data points. Where error bars are not visible, the standard deviation is small and lies under the symbol.

In the presence of Ca²⁺, cTnC releases the cTnI- and cTnI_1-192-mediated inhibition of the A-Tm-S1 ATPase activity (Figure 3B). Specifically, cTnI or cTnI_1-192 was added to actin-Tm in the ratio 5 A/1 Tm/1 TnI (or TnI_1-192) in the presence of S1. Addition of cTnC, in the presence of Ca²⁺, released both cTnI and cTnI_1-192 inhibition (Figure 3B); however, the concentration of cTnC required to achieve 50% reversal of A-Tm-S1 ATPase inhibition by cTnI_1-192 was less than that required to reverse inhibition by intact cTnI (IC₅₀ values of 0.2 nmol and 0.4 nmol, respectively). It may be interpreted that cTnI_1-192 either binds more tightly to Ca²⁺-cTnC than does full-length cTnI, or that the weakened/altered interaction between cTnI_1-192 and actin-Tm (observed in the inhibition studies) enhances the switch of cTnI_1-192 to cTnC.

Analysis of Reconstituted cTn

To assess the Ca²⁺-dependent regulation of ATPase activity, cTn complexes were formed using either cTnI or cTnI_1-192, mixed with cTnC and cTnT. Functional Tn complexes composed of either intact cTnI or cTnI_1-192 (cTn and cTn_1-192, respectively) were prepared and mixed at a 5:1:1 mole ratio of actin:Tm:cTn (or cTn_1-192). At low Ca²⁺, cTn_1-192 was not as fully inhibitory as wild-type cTn (45.3±1.9 versus 29.2±0.3 nmol · min⁻¹ · mg⁻¹). At pCa 4.6, cTn_1-192 conferred increased maximum ATPase activity, reaching 218±10 nmol · min⁻¹ · mg⁻¹ relative to 142±3 for cTn (Figure 4A), an increase of 54±11%. From the plots normalized for the maximum Ca²⁺-dependent change in A-Tm-S1 ATPase activity for each troponin (Figure 4A, inset), cTnI_1-192 has increased Ca²⁺ sensitivity compared with intact cTn by approximately 0.1 pCa units. Importantly, under these conditions, cTn_1-192 binds to actin-Tm with the same stoichiometry as cTn, when assessed by ultracentrifugation and SDS-PAGE (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Ca²⁺ sensitivity of cTn and cTn_1-192. A, Ca²⁺ dependence of the A-Tm-cTn-S1 activity was determined for reconstituted cTn (•) and cTn_1-192 (○). Inset, Results from panel A, normalized with respect to the maximum change in ATPase activity in Ca²⁺-free and Ca²⁺-replete conditions, to allow comparison of Ca²⁺ sensitivity. B, Maximum Ca²⁺-activated ATPase activity (1 mmol/L CaCl₂) was determined for reactions containing mixtures of cTn and cTn_1-192. Activity is expressed relative to control A-Tm-Tn-S1 ATPase activity (100%). Averages and their standard deviations were calculated from triplicate data points.

To determine whether cTnI_1-192 acts as a “poison peptide” that disproportionately affects the function of regular troponin, as has been observed with FHC mutants of cTnI²⁹ and cTnT,³⁰ Ca²⁺-activated A-Tm-S1 ATPase activity was determined for various mole fractions of cTn_1-192 with respect to Tn (ie, fixed total Tn, Figure 4B). Results are expressed relative to the A-Tm-Tn-S1 ATPase activity (A-Tm-Tn-S1 ATPase=100%). Figure 4B shows that ATPase increased linearly with the increased percentage of cTn_1-192. At a cTn_1-192:cTn ratio of 0.2, a level of cTn_1-192 comparable with levels observed in stunned rat hearts and transgenic mouse hearts that express cTnI_1-193 activity was 9.4±3.0% greater than that obtained observed with wild-type cTn alone.

In Vitro Motility and Force Measurements

To determine whether cTnI proteolysis alone is sufficient to cause the decreased force production in previous studies of muscle fibers, we examined both the sliding velocity and in vitro force index of cTn- and cTn_1-192-bearing reconstituted thin filaments, using the in vitro motility assay. Figure 5A shows that thin filaments bearing cTn_1-192 moved with the same mean unloaded velocity as those bearing cTn (compare 6.2±0.3 versus 6.0±0.3 μm/s, P=NS, respectively). Consistent with results from the A-Tm-S1 ATPase assay, there was an increased Ca²⁺ sensitization of the velocity-pCa relation when the thin filaments contained cTn_1-192 rather than cTn (compare 6.42±0.03 versus 5.99±0.05, respectively. P<0.001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5. Measurement of unloaded filament velocity and relative force using the in vitro motility assay. A, Velocity of actin-Tm filaments with bound cTn (•) or bound cTn_1-192 (□) was measured as a function of Ca²⁺. B, Relative isometric force was determined for actin-Tm filaments with bound cTn (▪) or bound cTn_1-192 (○) as a function of Ca²⁺, as described in Materials and Methods.

Thin filaments bearing cTn or cTn_1-192 were then compared in a modified in vitro motility assay that provides a reliable proxy for isometric force. In this assay, the load on a thin filament is a function of the relative concentration of the force generator (myosin) and a motion inhibitor (α-actinin) on a nitrocellulose coverslip (see Materials and Methods). The results of these experiments are depicted in Figure 5B. Interaction of these anchored thin filaments, decorated with cTn_1-192, with myosin generated only 76±4% (P<0.001) of the force produced by the interaction of myosin with actin-Tm filaments comprised with cTn. The Ca²⁺ sensitivity of the force-pCa relation was not statistically different between the two types of filaments studied (actin-Tm-cTn–induced pCa₅₀=6.27±0.03, actin-Tm-Tn_1-192–induced pCa₅₀=6.25±0.06).

Molecular Mechanistic Insights

The data are considered in the context of prominent three-state models of thin-filament regulation. In the biochemical three-state model, advanced by McKillop and Geeves¹⁵ and refined by Geeves and Lehrer,³³ tropomyosin can exist in equilibrium between three states. In the absence of Ca²⁺, troponin stabilizes tropomyosin in a state that blocks myosin binding to actin (blocked state). In the presence of Ca²⁺, the equilibrium shifts and tropomyosin adopts a state that allows myosin to bind strongly to actin (closed state). In the presence of myosin, equilibrium shifts yet again and tropomyosin adopts a state that permits myosin to bind actin, hydrolyze ATP, and cooperatively propagate further binding of myosin to actin (open state). Vibert and colleagues³² have proposed a structural three-state model in which the Ca²⁺-free troponin constrains tropomyosin in a position on the inner edge of the outer domain of actin that sterically occludes myosin-binding sites (EGTA state). In the presence of Ca²⁺, troponin allows tropomyosin to move to the outer aspect of the inner domain of actin, which exposes some amino acids on actin that interact strongly with myosin (Ca²⁺-induced state). Subsequent binding of myosin further nudges tropomyosin onto the inner domain of actin, which exposes all strong myosin-binding determinants on actin (myosin-induced state). The biochemical and structural three-state models are not necessarily equivalent, but provide complementary views of regulation nonetheless. Hereafter, the blocked/EGTA states are designated the B-state, the closed/Ca²⁺-induced states are designated the C-state, and the open/myosin-induced state is designated M-state.

cTnI, as part of the troponin complex, is posited to act in concert with cTnT^27,28 to stabilize/constrain Tm in the B-state. Therefore, the observation that the cTn_1-192 does not fully inhibit thin-filament ATPase in the absence of Ca²⁺ (Figure 4A) could result if (1) the nature of Tn-Tm in the B-state is fundamentally altered, (2) cTnI_1-192 alters the B-state–C-state equilibrium transition, or (3) the B-state is in equilibrium with an altered C-state. We believe that (1) is unlikely since cTnI_1-192 itself is capable of full ATPase inhibition under saturating conditions (Figure 3A), and preliminary reports of 3D-helical reconstruction from electron micrographs of cTn_1-192-bearing filaments show that the B-state position of Tm is unaltered.³¹ The present study cannot discriminate between (2) and (3), yet the helical reconstruction study, mentioned previously, is consistent with option (3), as it describes an altered C-state in which Tm is shifted azimuthally toward the inner aspect of the inner domain of actin.³¹

Our observation that cTn_1-192 causes Ca²⁺ sensitization of both the S1 ATPase (Figure 4A) and sliding velocity (Figure 5A) is consistent with previous work by Rarick et al,²⁶ who, by using myofibrils reconstituted with different truncation mutants of cTnI, demonstrated Ca²⁺ sensitization of myofibrillar ATPase with progressive truncation from the C-terminus. In the context of TnI interactions with its binding partners, the affinity chromatography and ultracentrifugation studies establish that the interactions between cTnI_1-192 and cTnT, cTnC, or actin-Tm are not grossly altered when compared with intact cTnI. Rather, it appears that subtle changes have occurred that alter the ability of cTnI_1-192 to inhibit the A-Tm-S1 ATPase activity (Figure 3A) and the ability of cTnC to release this inhibition (Figure 3B). The latter experiment suggests that cTnI_1-192 might be less effective at maintaining the blocked B-state. Therefore, the Ca²⁺ sensitization of ATPase and thin-filament sliding velocity may result from a combination of scenarios (2) and (3) above.

Our observation of increased maximum A-Tm-Tn_1-192-S1 ATPase activity at pCa 4.6 (Figure 4A) also suggests that the C- to M-state transition is affected by the TnI truncation. Again, the data might be explained by fundamental structural differences in the C- and M-states or a shift of the C- to M-state equilibrium that favors the M-state. As previously stated, preliminary work by Foster et al,³¹ studying cTn_1-192-bearing reconstituted filaments, shows that the C-state position of Tm is “nudged” toward the M-state, which exposes more of the myosin-binding interface on actin. Thus, C-terminal truncation of TnI, through its influence on Tn-Tm may promote binding of myosin to the actin filament. Interestingly, the effect of Tn_1-192/Tn ratio (1 mmol/L CaCl₂) on ATPase activity was linear (Figure 4B). This suggests that the effect of cTn_1-192 is confined locally to a single regulatory unit (7:1:1 mole ratio of actin:Tm:cTn) and is not propagated along the actin filament. Restated, cTnI does not appear to act as a poisoned polypeptide or dominant-negative protein that disproportionately affects the function of the wild-type protein, as has been observed for some Tn mutations that cause familial hypertrophic cardiomyopathy (FHC).^29,30

The hypercontractility of filaments regulated by cTn_1-192, observed in in vitro ATPase studies, is clearly not reflected under isometric conditions, however. The diminished force index, measured with the modified motility assay, indicates that cTn_1-192 causes one of the strongly bound actomyosin intermediates of the crossbridge cycle to be destabilized. The fact that cTn_1-192 engenders decreased force production (Figure 5B), yet causes supramaximal ATPase (Figure 4A) and has no effect on maximum thin-filament sliding velocity (Figure 5A), would seem to indicate that the hypocontractility is strain-dependent. The strain-dependent step of the crossbridge cycle, which becomes rate-limiting under isometric conditions, is an isomerism of strongly bound A-M-ADP.³⁵ How might truncation of TnI truncation lead to crossbridge destabilization? Under isometric conditions, actin-bound myosin heads push Tn-controlled Tm to the M-state. It is conceivable that cTn_1-192 alters the M-state position of Tm such that thin-filament contacts, with strained and strongly bound myosin, are not fully supported. Ongoing structural studies of cTn_1-192-bearing filaments may yet provide an explanation.

Contribution of cTnI Proteolysis in Heart Disease

This study provides an indication of how the proteolysis of cTnI may contribute to the cardiac pathophysiology observed with myocardial stunning and/or pressure overload. First, cTnI_1-192 confers decreased maximum force in a defined reconstituted assay (Figure 5B). This confirms that cTnI proteolysis is sufficient to account for the decreased force (steady-state tension) observed in detergent skinned muscle fibers and trabeculae obtained from either rat hearts that have undergone global ischemia⁵ or from transgenic mouse hearts that express cTnI_1-193.⁸

Second, our observations that cTn_1-192 causes divergent effects on ATPase and force in vitro suggest, yet do not prove per se, that proteolysis of cTnI may cause these parameters to be uncoupled in stunned myocardium. It will be important to test this prospect directly, since chemomechanical uncoupling would reduce the efficiency of contraction and have important consequences for the ailing myocardium. Specifically, given that the intracellular concentration of ATP is already reduced during myocardial stunning, inefficient contraction due to poor chemomechanical coupling would further tax low energy reserves. Thus, exertion would lead to increased energy demands within the heart that could not be met.

In conclusion, it is now clear that cTnI modifications occur in diseased human myocardium.^11,12 It has become particularly important to determine how C-terminal truncations of cTnI alter crossbridge cycling, since recent studies have shown that cTnI undergoes proteolysis, at its N- and C-termini, in the myocardium of heart-bypass patients.¹¹ Although the fragments in that study may or may not correspond to cTnI_1-192 exactly, our present study shows that removal of only 17 amino acids from the C-terminus substantively alters molecular function.

Note Added in Proof

Since submission of this manuscript, the structure of the core domain of Ca²⁺-saturated cardiac Tn has been determined by X-ray crystallography.³⁴ The C-terminal residues 192 to 209 are not visible in the structure, and the region is believed to be highly flexible.