|Home | About | Journals | Submit | Contact Us | Français|
Cardiac troponin I (cTnI) phosphorylation helps regulate myocardial contractility and relaxation during β-adrenergic stimulation. cTnI differs from the skeletal isoform in that it has a cardiac specific N′-extension of 32 residues (N′-extension). The role of the acidic-N′-region in modulating cardiac contractility has not been fully defined. To test the hypothesis that the acidic-N′-region of cTnI helps regulate myocardial function, we generated cardiac-specific transgenic mice in which residues 2–11 (cTnIΔ2–11) were deleted. The hearts displayed significantly decreased contraction and relaxation under basal and β-adrenergic stress compared to non-transgenic hearts, with a reduction in maximal Ca2+ dependent force and maximal Ca2+-activated Mg2+-ATPase activity. However, Ca2+ sensitivity of force development and cTnI-Ser23/24 phosphorylation were not affected. NMR amide proton/nitrogen chemical shift analysis shows that phosphorylation at Ser23/24 in cTnI and cTnIΔ2–11 decrease interactions with the N-lobe of cardiac troponin C. We hypothesized that phosphorylation at Ser23/24 induces a large conformational change positioning the conserved acidic-N-region to compete with actin for the inhibitory region of cTnI. Consistent with this hypothesis, deletion of the conserved acidic-N′-region results in a decrease in myocardial contractility in the cTnIΔ2–11 mice demonstrating the importance of acidic-N′-region in regulating myocardial contractility and mediating the heart’s response to β-AR stimulation.
β-adrenergic (β-AR) signaling plays a fundamental role in regulating cardiac performance (1). Physiological effects include increases in contractile force, heart rate and the rate of relaxation. The use of β-AR agonists and β-AR blockers to treat acute ventricular failure or chronic failure, respectively, likely represent a delicate balance of different cardioprotective mechanisms. During β-AR stimulation, multiple proteins in the sarcolemma, sarcoplasmic reticulum and myofilament are phosphorylated at multiple sites. One of the sarcomeric proteins, cTnI, is a key regulatory protein of the thin filament. There are three closely related troponin I (TnI) genes, each of which is selectively expressed in either the cardiac, fast skeletal or slow skeletal muscle fibers. The embryonic heart expresses mostly slow skeletal muscle TnI but expression gradually decreases during prenatal heart development as cTnI increases and becomes the only TnI isoform in the adult heart (2–4). Cardiac TnI interacts with the major proteins present in the sarcomeric thin filament, including actin, cTnC, α-tropomyosin (α-TM) and troponin T (cTnT). These interactions underlie its central role as a molecular switch, regulating muscle contraction in response to changes in intracellular Ca2+ concentrations.
Cardiac TnI differs from the slow skeletal isoform of TnI in that it contains a 32 amino acid (31 in the human) N′-extension. cTnI’s N′-extension is composed of three regions; an acidic-N′-region containing a single turn of helix, an extended rigid polyproline helix, and a C′-helix containing the bisphosphorylation motif. The β-AR signaling pathway controls phosphorylation of the two serine residues (Ser23/24) in the N′-extension by cAMP-dependent protein kinase (PKA) (5) and protein kinase D (PKD) (6). Phosphorylation of Ser23/24 results in a reduction in myofilament Ca2+ sensitivity (1) and an increase in cross-bridge cycling rate (7) by reducing the Ca2+ binding affinity of cTnC and allowing fine tuning of contractile function (8). This mechanism plays an important role in the functional adaptation of cardiac muscle to physiological or pathological stress.
Molecular modeling data have enabled predictions to be made and tested for cTnI’s functional domains. In the non-phosphorylated state, the N′-extension interacts with cTnC’s inactive Ca2+ binding site I and helix A, largely through a series of weak electrostatic and hydrophobic interactions such that the acidic-N′-region is does not strongly contact cTnC (9–11). However, an important unanswered question is the role of the conserved acidic-N′-region of cTnI in modulating cardiac function. Based on the available biochemical and physiological data, we hypothesized that modulation through phosphorylation might be partially mediated by electrostatic interactions between cTnI’s acidic-N′-region and available basic regions in cTnI, altering cross-bridge kinetics. Thus, the role of phosphorylation would be to stabilize a conformation able to facilitate these ionic interactions. Phosphorylation-induced loss of interaction with the N′-lobe of cTnC probably induces a hinge movement of the rigid PPII helix, positioning cTnI’s acidic-N′-region for electrostatic interaction with the conserved basic region of cTnI. The PPII helix is well suited for such conformational movement due to its restricted conformational rigidity, solvent exposure, and ability to present a hydrophobic surface as well as hydrogen bonding sites. Consistent with this hypothesis, NMR and modeling studies showed that bisphosphorylation of cTnI at Ser23/24 resulted in extension of its position on the N′-lobe of cTnC and interaction with the basic inhibitory region of cTnI (10).
To determine the role of the acidic-N′-region of cTnI on cardiac contractile function, we generated transgenic (TG) mice with cardiomyocyte-specific postnatal overexpression of a truncated cTnI that lacks the acidic-N′-region (cTnIΔ2–11). cTnIΔ2–11 cardiac myofibrils showed reductions in both maximal Mg2+-ATPase activity and absolute force, but no changes in Ca2+ affinity. cTnIΔ2–11 hearts had significantly reduced rates of contraction and relaxation under baseline and β-agonist treatment. These findings indicate that the acidic-N′-region of cTnI plays an important role in regulating cardiac function in non-stimulated hearts as well as during β-AR stimulation.
To investigate the role of the acidic-N′-region of cTnI on cardiac contractile function, we generated a cDNA that encodes mouse cTnI with a deletion of residues 2–11 (ADESSDAAGE). The full-length mouse wild-type (WT) cDNA cTnI was obtained by reverse transcription-PCR using total RNA isolated from mouse cardiac ventricles. The cDNA containing cTnIWT fragments were initially subcloned into pBluescript and sequenced as described earlier (14). The 10 amino acids of the acidic-N′-region (Fig. 1A) were deleted by standard PCR-based methods (cTnIΔ2–11). The cTnIΔ2–11 cDNA was subcloned into a site immediately downstream of the mouse cardiac α-myosin heavy chain promoter (α-MyHC) and the sequence verified by DNA sequencing. The TG cassette was then released from the vector backbone using NotI digestion, followed by gel purification. Multiple lines of FVB/N TG mice were generated using the purified NotI-digested DNAs. Founder mice were identified by PCR using tail clip DNA as template. Transgene copy number was determined by Southern blot analysis using an α-MyHC promoter probe. The founders were bred to non-transgenic (NTG) mice and lines showing Mendelian patterns of transmission selected for further analysis. Five or six mice per experiment, 12–15 weeks old of mixed gender were used for our studies after pilot experiments showed no gender differences. All TG mouse lines were viable and fertile. The mice had a normal lifespan and no gross cardiovascular pathology presented. All protocols complied with the Guide for the Use and Care of Laboratory Animals published by the National Institutes of Health.
Transcript levels were determined by RNA dot blot analysis with γ-32P-labeled cTnI and human growth hormone as well as probes for the cardiac hypertrophic markers atrial natriuretic factor and β-MyHC (12). Myofibrillar proteins were isolated from NTG and cTnIΔ2–11 mouse hearts using F60 buffer as previously described (8) and assayed for protein concentrations using the Bradford method (Bio-Rad, Hercules, CA, USA). The percentage of cTnI replacement was determined via SDS-PAGE (4–15% gradient Tris-HCl Ready Gel; Bio-Rad, Hercules, CA, USA) and Western blots using polyclonal antibodies against cTnI (Cell Signaling Technology, Danvers, MA, USA). Two-dimensional gel electrophoresis was carried out as described (8). Antibodies used for Western blot analysis are as follows: phospho-specific cTnI-Ser23/24 (Cell Signaling Technology, Danvers, MA, USA), total phospholamban (PLN) (Upstate, Lake Placid, NY), phosphorylated PLN-Ser16 (Upstate, Lake Placid, NY, USA), phosphorylated PLN-Thr17 (Cyclacel, Dundee, UK), calsequestrin (Research Diagnostics, Flanders, NJ), cardiac myosin binding protein-C (cMyBP-C) C0–C1 domain (12), phospho-specific cMyBP-CSer282 (generous gift of Lucie Carrier (13)), cardiac troponin T (Sigma, St. Louis, MO, USA) and cardiac α-tropomyosin (Chemicon, Temecula, CA, USA).
The heart weight (HW) and the ratio of heart weight:body weight (HW/BW) were measured to determine if cardiac hypertrophy had occurred. Gross examination and histopathological analysis were carried out as described (8). The paraffin-embedded longitudinal sections of whole mouse hearts stained with hematoxylin-eosin or Masson’s trichrome were examined for overall morphology, presence of necrosis, fibrosis, myocyte disarray and calcification using an Olympus B-60 microscope and SPOT software (Diagnostic Instruments, Sterling Heights, MI, USA). Localization and integration of cTnIΔ2–11 into the sarcomere was determined by confocal microscopy (12). Five-μm cryostat sections were probed with cTnI antibodies (Cell Signaling Technology, Danvers, MA, USA) followed by incubation with Alexa-488 conjugated secondary antibody (Invitrogen, Carlsbad, CA, USA).
For two-dimensional M-mode echocardiography, mice with the implanted osmotic pumps were anesthetized with 2% isoflurane. Hearts were visualized with a Hewlett Packard Sonos 5500 instrument and a 15 MHz transducer (12). Measurements were taken three times per mouse from different areas and then averaged for left ventricular (LV) diastolic and systolic dimensions and septal and posterior wall thickness, from which fractional shortening (FS) and LV mass was derived. Invasive hemodynamic studies was performed in the intact animals as previously described (14). Data were analyzed using a PowerLab system (ADInstruments, Colorado Springs, CO, USA).
To determine the stress tolerance of cTnIΔ2–11 hearts, NTG and cTnIΔ2–11 animals underwent two weeks of continuous infusion of the β-agonist ISO (Sigma, St. Louis, MO). Alzet miniosmotic pumps (Durect Corporation, Cupertino, CA, USA) containing either ISO (60 mg/kg/day) in 0.02% ascorbic acid (Sigma) or vehicle only (sham) were surgically implanted between the scapulae in 12-week-old NTG and cTnIΔ2–11 mice for 14 days as described (15). Cardiac function was measured by M-mode echocardiography before, 7, and 14 days after implantation.
To phosphorylate myofibrillar proteins, total myofibrils were incubated with the catalytic subunit of PKA as described earlier (8). Ca2+-activated Mg2+-ATPase activity was measured by titrating Ca2+Δsensitivity of the NTG and cTnIΔ2–11 mouse hearts and measuring Pi release (8). Data were analyzed by fitting the data obtained for each individual and then averaging the derived Hill parameters as described (8).
Procedures for mechanical analysis of murine papillary fibers have been described (16, 17). In brief, mice were injected with heparin (500 IU/kg intraperitonally) 5 min before being killed. To prepare skinned fibers, the heart was removed and placed in relaxing solution (5.37 mM ATP, 30 mM phosphocreatine, 5.0 mM EGTA, 20 mM BES, 7.33 mM MgCl2, 0.12 mM CaCl2, 10 mM DTE, 10μg/mL leupeptin, and 32 mM potassium methansulfonate, pH 7.0) at 4 °C. The solution also contained 30 mM 2,3-butanedione monoxime designed to protect myocardial tissue from mechanical injury. Muscle fibers of approximately 0.5-mm diameter and 2–3-mm length were isolated from left ventricular papillary muscles. The fiber strips were skinned by incubation in 5.5 mM ATP, 5.0 mM EGTA, 20 mM BES, 6.13 mM MgCl2, 0.11 mM CaCl2, 10 mM DTE, 10 μg/mL leupeptin, 121.8 mM potassium methansulfonate, pH 7.0, and 50% glycerol with 0.5% (wt/vol) Triton X-100 for 12 h at 4 °C. The fiber strips were then transferred to fresh solution without Triton X-100, and stored at −20 °C until used. Dissected fibers were mounted isometrically between a force transducer and a length-step generator in relaxing solution (Scientific Instruments, Heidelberg, Germany). Sarcomere length (determined by laser diffraction analysis) at resting tension was always 2.0–2.1 μm. We determined that the cross-sectional area at the base of the muscle was between 0.05 to 0.1 mm2. Contraction solution had the same composition as the relaxing solution, except that EGTA was substituted with 5 mM [Ca2+]EGTA. Initial maximum isometric force was measured in activating solution (pCa 5.0). Force was determined while the fibers were bathed in sequentially increasing Ca2+ concentrations ranging from pCa 8 to 5.0 pCa values and recorded on a chart recorder. Strip tension (mN/mm2) was calculated by dividing force by fiber cross-sectional area, calculated from widths measured at the major axis. To examine the effects of PKA phosphorylation on the pCa-force relationship in vitro, skinned fibers were treated with PKA (Novagen, San Diego, CA, USA). After measurements of the pCa-force relationship (before PKA treated), the fiber was incubated in relaxing solution (pCa 8.0) plus 0.5 μM PKA for 10 min. The fibers were relaxed for 15 min, and the pCa-force relationship was measured after PKA treatment.
The cTnIWT and cTnIΔ2–11 cDNAs were subcloned into the pET23a+ expression vector (Novagen, San Diego, CA, USA). [15N,2H]cTnC, cTnIWT and cTnIΔ2–11 proteins were expressed in bacteria, purified and complex formation was carried out as described (23, 24). NMR experiments were performed at 40 °C on Varian 600- or 800-MHz Inova spectrometers (25) and spectral widths in the t1 and t2 dimensions were 3.3 and 12 kHz, respectively. Composite amide chemical shift differences were determined from the square root of the weighted sum of the squares of proton and nitrogen chemical shift differences. 15N chemical shift differences were weighted by a factor of one-seventh to scale their contributions to a magnitude similar to 1H chemical shift differences and the data processed as described (24). Spectra were processed with Felix 2000 and analyzed with Sparky (T. D. Goddard and D. G. Kneller, Sparky3, University of California, San Francisco, CA, USA) software packages.
All values are expressed as means ± s.e. The statistical significance of differences between two groups and multiple groups was determined by Student’s t test and two-way ANOVA (SigmaStat 3.1, Systat Software, San Jose, CA, USA), respectively. For all tests, P<0.05 was considered significant. The Hill coefficient was calculated using Origin 7.5 NLSF tool (OriginLab Corporation, Northampton, MA, USA) as described (8). The theoretical molecular weight and isoelectric point of cTnI were calculated at http://prometheus.brc.mcw.edu/promost/.
To investigate the functional effects of cTnI’s acidic N′-region (Fig. 1A) in intact mouse hearts we generated TG mice in which this region was deleted (cTnIΔ2–11). The cDNA was linked to the mouse α-MyHC promoter to drive cardiac specific expression and used to generate multiple TG mouse lines. Nineteen TG lines were obtained with varying copy numbers as analyzed by Southern blot analysis. For all of the lines, normal Mendelian ratios were observed, indicating that expression did not result in any detectable embryonic lethality. The amount of TG transcript being expressed was determined for each line and three lines (lines 93, 81 and 78) were selected in which cTnIΔ2–11 expression was roughly equivalent to a TG line that expresses high levels of normal cTnI (cTnIWT) (Fig. 1B), which served as a control to rule out any phenotype that might occur merely as a result of high TG expression levels of cTnI. (18). Lines 93, 81 and 78 showed 5.5-, 4.0- and 2.8-fold increases over NTG controls, respectively (Fig. 1C). The sarcomere has the capacity to maintain protein stoichiometry even when the transcript levels are increased via TG manipulation. Thus the degree of replacement is a function of the level of overexpression (8). The degree of cTnIΔ2–11 protein replacement was confirmed by SDS-PAGE (Fig. 2A) as well as by Western blots using cTnI-specific antibodies that recognize both the endogenous cTnI and TG cTnIΔ2–11 proteins (Fig. 2B). Densitometry showed that we obtained essentially complete replacement of the endogenous protein with the transgenically-encoded cTnIΔ2–11 in line 93, which was chosen for subsequent functional analyses. Replacement of endogenous cTnI with cTnIΔ2–11 had no effect on expression levels of the other major contractile proteins, including MyHC, actin, myosin light chains, cTnT and α-TM (Fig. 2A). The deletion of the acidic-N′-region of cTnI was benign with no signs of increased morbidity, mortality or cardiac hypertrophy, implying that cTnIΔ2–11 is functionally active. Heart weight/body weight (HW/BW) ratios did not differ significantly between NTG (0.0051 ± 0.006, n = 8) and cTnIΔ2–11 (0.0050 ± 0.005, n = 6) littermates. The cTnIΔ2–11 hearts were unremarkable and showed no evidence of abnormal morphology, myofibrillar disarray, necrosis, or ventricular fibrosis (Fig. 3A and B). Normal integration of cTnIΔ2–11 into the sarcomere was confirmed by immunofluorescent detection using cTnI-specific antibodies (Fig. 3C).
Although the TG animals appeared overtly normal and showed no obvious pathology, we hypothesized that the cTnIΔ2–11 hearts would have reduced basal cardiac function and this deficit would remain or 9 become more pronounced during β-AR stimulation. To test this, NTG and cTnIΔ2–11 animals underwent two weeks of continuous β-agonist infusion with isoproterenol (ISO). We utilized four, 12-week old groups; NTG (sham), NTG (ISO), cTnIΔ2–11 (sham) and cTnIΔ2–11 (ISO). At the end of the two-week infusion, cardiac function was assessed by M-mode echocardiography. Both the NTG and cTnIΔ2–11 sham animals showed normal fractional shortening and the other functional parameters did not differ between the two groups (Fig. 4A and C, Table 1). In contrast, both the NTG and cTnIΔ2–11 ISO groups had significantly increased left ventricular inner diameters in diastole and systole and decreased fractional shortening. Importantly, significant differences presented between the NTG and cTnIΔ2–11 groups, with the cTnIΔ2–11 ISO hearts displaying both decreased heart rates and fractional shortening compared to the NTG ISO hearts. Continuous ISO infusion increased the HW/BW ratio significantly in both NTG and cTnIΔ2–11 ISO groups compared to the respective sham groups but no differences presented between the NTG and cTnIΔ2–11 ISO groups, ruling out a differential hypertrophic response as being responsible for the functional differences that were observed (Fig. 4B and D).
After the two-week infusion, in vivo hemodynamics were determined using the closed chest model and invasive catheterization (EXPERIMENTAL PROCEDURES). Ventricular contraction and relaxation were measured at baseline and during dobutamine infusion. As expected, dobutamine infusion increased dP/dtmax and dP/dtmin in both the NTG and cTnIΔ2–11 sham hearts (Fig. 4E and F) but basal and stimulated values for both these parameters were decreased in the cTnIΔ2–11 hearts as compared to the NTG hearts. That is, dP/dtmax increased in NTG sham hearts by 50% at the highest dobutamine dose, but only by 30% in the cTnIΔ2–11 sham hearts (Fig. 4E and F, Table 2). After two weeks of continuous ISO infusion, the cTnIΔ2–11 ISO hearts displayed a relatively depressed heart rate at baseline and upon dobutamine infusion (Table 2). Hemodynamic parameters, including LV pressure, dP/dtmax and dP/dtmin were also significantly compromised in the cTnIΔ2–11 ISO hearts, compared to the NTG ISO group (Table 2). The NTG ISO hearts showed conserved cardiac function with values for both dP/dtmax and dP/dtmin at baseline almost equivalent to the rates that obtained with a maximum dose of dobutamine in the NTG sham mice. As expected, NTG ISO hearts were unable to respond to increasing concentrations of dobutamine as they were already fully stimulated. These results indicate that, despite appearing overtly normal and healthy under standard cage conditions, the acidic-N′-region deletion does compromise cardiac function and this is exacerbated under conditions of chronic and acute β-AR stimulation.
We used skinned papillary muscle fibers to assess the effect of cTnIΔ2–11 on the kinetics and Ca2+ sensitivity of force generating cross-bridge formation. Fiber preparations were also subjected to PKA treatment in order to compare fully phosphorylated cTnI with the truncated TG protein (Fig. 5A). Average maximally developed isometric tension (Fmax) was significantly decreased in cTnI 2–11–derived fibers relative to NTG fibers (8.9 ± 0.4 versus 6.9 ± 0.6 kN/m2, P<0.002, n = 5), although Ca2+ sensitivity was unaffected. The same fibers were then treated with PKA and the effects on Ca2+ sensitivity and maximum force determined for both the NTG and cTnIΔ2–11 fibers. Exogenous treatment of fibers with PKA did not affect Fmax in either group (Table 3). Ca2+ sensitivity was significantly decreased by PKA treatment compared to the untreated fibers for both groups. In this isolated system, cTnIΔ2–11’s effect on maximal force development is consistent with decreased myocardial contractility in vivo and a concomitant decrease in the number of force-generating cross-bridges. There was no difference in the basal and PKA-activated Ca2+-sensitivity, indicating that the deletion did not affect phosphorylation of Ser23/24 or myofilament Ca2+ sensitivity.
To understand the consequences of cTnIΔ2–11 on the Ca2+ activated actomyosin Mg2+-ATPase activity, myofibril Mg2+-ATPase activity and Ca2+ sensitivity were measured with or without PKA treatment (Fig. 5B and Table 4). The cTnIΔ2–11 myofibrils showed a significant decrease in the maximal Mg2+-ATPase activity (163.07 ± 4.85 nmol Pi/min/mg, P<0.01) compared to the NTG (188.96 ± 4.67 nmol Pi/min/mg) but Ca2+ sensitivity was unaffected. As expected, PKA treatment of either cTnIΔ2–11 or NTG myofibrils resulted in decreased Ca2+ sensitivity. Hill coefficients and EC50 values fall within the range of those previously reported (8). The Hill coefficients for the NTG and cTnIΔ2–11 proteins did not differ, reflecting the similarities in the shape of the Ca2+ binding curves. Phosphorylation of cTnI residues was confirmed by Western blots using phospho-specific cTnI antibodies (Fig. 5B, inset). Taken together, the data show that the acidic-N′-region plays a major role in regulating the actomyosin Ca2+-activated maximal Mg2+-ATPase activity.
To confirm that these functional, hemodynamic and biochemical changes were not due to compensatory changes in contractile protein phosphorylation states, PLN (Fig. 6A), the myosin light chains (Fig. 6B) and cMyBP-C (Fig. 6C) phosphorylation was examined by Western blot analysis using phospho-specific antibodies and two-dimensional electrophoresis. The data showed that phosphorylation levels of these proteins were unchanged between the cTnIΔ2–11 and NTG sham and ISO hearts, respectively.
Backbone amide resonances provide excellent probes for monitoring protein-protein interactions and identifying interaction surfaces. The chemical shift differences can be used to map the cTnC binding site of the N′-extension (2–33) of cTnI on Ca2+-loaded cTnC. Interactions of the N′-extension (2–33) and, in particular, the conserved acidic-N′-region (2–11) were analyzed by comparing combined amide 1H and 15N chemical shift differences for Ca2+-loaded cTnC bound to cTnI and cTnIΔ2–33 (Fig. 7A) and for cTnC bound to cTnI and cTnIΔ2–11 (Fig. 7B). Residues showing significant chemical shift perturbations induced by the N′-extension (2–33) were located in the N′-lobe of cTnC with the largest perturbations centered in defunct Ca2+-binding site I (28–38) and Ca2+ binding site II (65–76) (Fig. 1A). Bis-phosphorylation or introduction of negative charge at Ser23/24 in the N′-extension results in a loss of these interactions (10, 11, 19). No significant chemical shift perturbations were observed in the linker region or in the C′-lobe of cTnC. Composite amide 1H and 15N chemical shift differences for Ca2+-loaded cTnC bound to cTnI and cTnIΔ2–11 were considerably smaller, generally <0.05 ppm (Fig. 7B). The magnitude of the observed chemical shift perturbations suggest that residues 2–11, comprising the conserved acidic N′-region, do not interact strongly with Ca2+-loaded cTnC. This is consistent with previously published chemical shift perturbation analysis (19, 20) and binding studies (21) supporting interactions between residues 19–30 in the N′-extension and the N′-lobe of Ca2+-loaded cTnC.
Amide proton and amide nitrogen chemical shifts of residues in the N′-lobe of cTnC can be used to monitor opening of the regulatory domain upon binding cTnI (19, 20). The conformational transitions for two residues, Glu66 and Ile128, located in the N′- and C′-lobes of cTnC respectively, were examined in detail. Glu66 can exhibit conformational dependent chemical shift changes that are correlated with open and closed N′-lobe conformations (20). The binding of full-length cTnI to Ca2+-loaded cTnC results in a downfield shift for the amide cross-peak of Glu66 (Fig. 7C). This shift is consistent with an opening of the N′-lobe of cTnC and exposure of the hydrophobic cleft for binding the switch region of cTnI (20). A similar downfield shift for the amide cross-peak of Glu66 is observed in the presence of cTnIΔ2–11 (Fig. 7C). However, the magnitude of the shift is intermediate between that observed for free Ca2+-loaded cTnC and the Ca2+-loaded cTnC/cTnI complex (Fig. 7C). The magnetically different amide 1H-15N environments presumably result from alterations in the N′-lobe open/closed equilibrium (20), indicating that the conserved acidic-N′-region (2–11) shows some propensity for stabilizing a more open N′-lobe conformation. Loss of the conserved acidic-N′-region may alter the interaction of residues 19–30 of cTnI with the N′-lobe of cTnC. It is known that phosphorylation of Ser23/24 of cTnI results in weakening of the interactions between the N′-extension and the N′-lobe of cTnC, providing a molecular basis by which the N′-lobe conformational equilibrium is modified in response to physiological stimuli (20). No significant chemical shift changes in Ile128 were observed between Ca2+-loaded cTnC and Ca2+-loaded cTnC bound to either cTnI or cTnIΔ2–11 (Fig. 7C).
Human cTnI has a 31-residue N′-extension that is not present in the fast and slow skeletal muscle isoforms. This extension contains a unique acidic-N′-region, a Xaa-Pro region (12–18) and a bisphosphorylation motif, in which Ser23/24 are substrates for PKA, PKD and the Rho kinases. Bisphosphorylation at Ser23/24 in the N′-extension of cTnI, in response to β-adrenergic stimulation, modulates myofilament Ca2-sensitivity and cross-bridge kinetics (5, 7). The extension is highly conserved among mammals and its functional importance underscored by the finding of two missense mutations within the N′-extension that cause familial hypertrophic cardiomyopathy (FHC) in humans (15, 16). One mutation, FHCA2V (22), is located in the acidic-N′-region and the second, FHCR21C, in the phosphorylation motif (23). Although NMR and modeling studies have shown that the conserved acidic-N′-region of the N′-extension can extend from its position on the N-lobe of cardiac troponin C (cTnC) and interact with the basic inhibitory region of cTnI, the role of the acidic N′-region of cTnI in the modulation of cardiac contractility is not clear. A novel finding reported in the present study is the reduction of rate of contractile function at baseline and β-AR stimulation in cTnIΔ2–11 hearts lacking the acidic N-terminus region of cTnI. The TG mice expressing cardiac-specific cTnIΔ2–11 were allowed to determine the role of acidic-N′-region of cTnI on myocardial function and response to β-AR stimulation. cTnIΔ2–11 maintains the core structure of cTnI and thus wouldn’t modify the troponin complex as reported in previous studies (24, 25). cTnIΔ2–11 hearts, with >95% of cTnI replaced by cTnIΔ2–11, appeared apparently normal. The cTnIΔ2–11 is benign without any obvious increased mortality or detectable cardiovascular pathology. The cTnIΔ2–11 mice were viable and fertile and exhibited normal ventricular weights and heart rates; however, significant differences in contractile function were evident. This data suggests that removal of acidic-N′-region of cTnI wouldn’t affect the heart adversely. Similar findings were observed in cardiac-specific TG mice that express either lack of the N′-extension 2–28 residues of cTnI (24) or replacement of cTnI with the slow skeletal TnI (25), that differ from cTnI by the absence of unique 32 residues at the N′-extension of cTnI. Conversely, the mice (18) and rabbits (26) expressing cTnI146Gly, a FHC mutation that is located within the inhibitory region, resulted in cardiomyocytes disarray, interstitial fibrosis and suffered premature death.
In the present study, the absolute force and Mg2+-ATPase activity data follow the same pattern in cTnIΔ2–11 myofilaments. The expected decrease in maximal force and actomyosin Mg2+-ATPase activity as a result of cTnIΔ2–11 was confirmed in the skinned fiber studies with no change in Ca2+ sensitivity of the cTnIΔ2–11. Our data further show that the Ser23/24 phosphorylation and Ca2+ binding sites are unaltered by incorporation of cTnIΔ2–11. Thus, we think that the deletion alters the orientation of the N′-extension of cTnI, which then affects the interaction between the acidic-N′-region and the inhibitory domain of cTnI with cTnC and actin. In addition, a basic C′-region of cTnI, not observed in the cTnI crystal structure (27) can interact electrostatically with actin (28). Modeling studies suggest that Ca2+-induced translocation of cTnI’s mobile C- terminal domain to actin is triggered by the switch region’s binding to the N′-lobe of cTnC. Thus, in the low Ca2+Δstate, both the inhibitory region and mobile C-terminal domain of cTnI bind to actin, forming an electrostatic clamp that pushes tropomyosin toward the outer domain of actin (28). Our studies indicate that the acidic-N′-region of cTnI plays an active role in modulating the interactions of the inhibitory region and the C-terminal mobile domain of cTnI with α-actin, which could influence the position of tropomyosin binding on actin. Protein-protein interactions within tropomyosin can influence the movement and position of tropomyosin on the actin surface (29).
Ser23/24 phosphorylation within the N′-extension of cTnI results in a reduction in myofilament Ca2+ sensitivity, an increase in cross-bridge cycling, and enhanced binding of cTnI to the thin filament (7, 8, 30, 31). Studies in the intact heart have demonstrated a significant role of cTnI phosphorylation for both afterload dependence of ejection and relaxation (32) as well as force frequency modulation (30). To determine the functional consequences of the acidic-N′-region upon β-AR stimulation, we examined contraction and relaxation under baseline conditions and during β-AR stimulation. Incorporation of cTnIΔ2–11 into the sarcomere reduces baseline values for dP/dtmax and dP/dtmin, indicating a negative regulatory mechanism of cardiac function followed by decreased maximal force and Mg2+-ATPase activity. Moreover, during chronic long-term β-AR stimulation with ISO, contractile function was blunted in the cTnIΔ2–11 mouse hearts. We were able to exclude the possibility that alterations in either phosphorylation levels or expression levels of the cMyBP-C, myosin light chains and PLN were compensating for a reduction in contractile function in the cTnIΔ2–11 hearts. Consistent with altered interactions of mutant cTnI with the cTnC and actin, TG mice that express a cTnI which lacks residues 2–28 (24) and the cTnI146Gly (18) showed enhanced contractility with impaired relaxation at the whole heart level.
NMR data and sequence analyses indicate a loosely structured N′-extension with a propensity for a helical region surrounding the bisphosphorylation motif (20–24), followed by a helical C-terminal region in residues 25–30 (Rosevear et al., unpublished data). An extended poly(L-proline)II helix (11–19) appears to serve as a rigid linker that aids in positioning the acidic-N′-region. In this conformation, the N′-extension of cTnI interacts with the N′-lobe of troponin C (cTnC), modulating myofilament Ca2+-sensitivity. Introduction of a negative charge at Ser23/24 via phosphorylation weakens its interaction with the N′-lobe of cTnC by repositioning the extension. NMR studies (9, 11, 19, 20, 33), binding studies (34), deletion mutagenesis (21), and cross-linking studies (35) have shown that the N′-extension of cTnI interacts with the N′-lobe of cTnC. Specifically, these studies defined residues 19–30 as the minimal region of the N′-extension necessary for interacting with the N′-lobe of cTnC. Residues surrounding the inactive Ca2+ binding site I in cTnC may comprise the potential interaction site with the N′-extension. NMR studies further showed that interaction of the cardiac N-extension alters N′-lobe conformational equilibria in cTnC toward a more active/open conformation capable of binding the switch region of cTnI (9, 20, 34). Bisphosphorylation at Ser23/24 in the N′-extension results in the loss of N′-lobe interactions, shifting the N′-lobe conformational equilibrium toward a more closed conformation, resulting in decreased Ca2+ sensitivity. This mechanism utilizes the unique isoform differences in both cTnC and cTnI, providing a molecular switch for modulating cardiac muscle contraction.
The roles of the conserved Xaa-Pro and acidic-N′-region in the N′-extension remain relatively undefined. Recent structural studies from one of our laboratories (P.R.) on the phosphorylation of Ser23/24 showed that the Xaa-Pro region forms an extended polyproline helix (unpublished data), possibly providing a rigid spacer for interaction of the acidic-N′-region with basic regions within the troponin complex, inducing a bending in the rod-like cTnI at the end that interacts with the cTnC/cTnT component (36). Furthermore, we constructed atomic models for troponin that show the conformational transition induced by bisphosphorylation, based on the structure of the bisphosphorylated N′-extension, the X-ray crystal structure of the cardiac troponin core (27), and the uniform density models of the troponin components derived from neutron contrast variation data (36). Therefore, we hypothesize that the acidic-N′-region interacts with the basic residues in its inhibitory region, competing with actin for the inhibitory region of cTnI or the second actin-binding region of cTnI (Fig. 8) and is responsible for altering cross-bridge kinetics.
Our data are consistent with the hypothesis that removal of the acidic-N′-region from cTnI prevents interaction of this region with the inhibitory domain of cTnI, permitting it to associate more readily with actin. As a result, enhancing the contact between actin and the inhibitory domain of cTnI might ultimately diminish cardiac contractility and maximal force of contraction. In support of our proposed model, the cTnIΔ2–11 myofilaments showed decreased maximal absolute force and Mg2+-ATPase activity, leading to reduction of contractile function at baseline and during β-AR stimulation. Taken together, the data indicate that the interaction of the acidic-N′-region with the inhibitory region of cTnI provides a novel mechanism by which the acidic-N′-region modulates cross-bridge kinetics and regulates actomyosin interactions. Additional biochemical and structural studies are warranted to gain insight into the molecular function of the acidic-N′-region and its roles in cardiac muscle acidosis and regulation of contraction.
This research was supported by National Institutes of Health grants HL69799, HL60546, HL52318, HL60546, and HL56370 (J.R.) and by the American Heart Association, Ohio Valley Affiliate (S.S.) and United States Department of Defense Grant ARO MURI DAAD 19-02-1-0027 (PRR).