Arg92Leu-cTnT Alters the cTnC-cTnI Interface Disrupting PKA-Mediated Relaxation

Abstract

Background: Impaired left ventricular relaxation, high filling pressures, and dysregulation of Ca²⁺ homeostasis are common findings contributing to diastolic dysfunction in hypertrophic cardiomyopathy (HCM). Studies have shown that impaired relaxation is an early observation in the sarcomere-gene-positive preclinical HCM cohort, which suggests the potential involvement of myofilament regulators in relaxation. A molecular-level understanding of mechanism(s) at the level of the myofilament is lacking. We hypothesized that mutation-specific, allosterically mediated, changes to the cTnC (cardiac troponin C)-cTnI (cardiac troponin I) interface can account for the development of early-onset diastolic dysfunction via decreased PKA accessibility to cTnI.

Methods: HCM mutations R92L-cTnT (cardiac troponin T; Arg92Leu) and Δ160E-cTnT (Glu160 deletion) were studied in vivo, in vitro, and in silico via 2-dimensional echocardiography, Western blotting, ex vivo hemodynamics, stopped-flow kinetics, time-resolved fluorescence resonance energy transfer, and molecular dynamics simulations.

Results: The HCM-causative mutations R92L-cTnT and Δ160E-cTnT result in different time-of-onset diastolic dysfunction. R92L-cTnT demonstrated early-onset diastolic dysfunction accompanied by a localized decrease in phosphorylation of cTnI. Constitutive phosphorylation of cTnI (cTnI-D₂₃D₂₄) was sufficient to recover diastolic function to non-Tg levels only for R92L-cTnT. Mutation-specific changes in Ca²⁺ dissociation rates associated with R92L-cTnT reconstituted with cTnI-D₂₃D₂₄ led us to investigate potential involvement of structural changes in the cTnC-cTnI interface as an explanation for these observations. We probed the interface via time-resolved fluorescence resonance energy transfer revealing a repositioning of the N-terminus of cTnI, closer to cTnC, and concomitant decreases in distance distributions at sites flanking the PKA consensus sequence. Implementing time-resolved fluorescence resonance energy transfer distances as constraints into our atomistic model identified additional electrostatic interactions at the consensus sequence.

Conclusions: These data show that the early diastolic dysfunction observed in a subset of HCM is attributable to allosterically mediated structural changes at the cTnC-cTnI interface that impair accessibility of PKA, thereby blunting β-adrenergic responsiveness and identifying a potential molecular target for therapeutic intervention.

Keywords: calcium; cardiomyopathy, hypertrophic; fluorescence resonance energy transfer; molecular dynamics simulation; phosphorylation; sarcomeres; troponin.

Figure 1.

Cardiac function assessed via 2D echocardiography at early (1-2 months) and late (4-6 month) time points in Non-Tg and cTnT-linked models of HCM (R92L-cTnT, Δ160E-cTnT). A: Systolic function assessed by % Ejection Fraction and B: % Fractional Shortening, C: Total wall thickness (Td, mm), D: Diastolic function assessed by E/e’, E: Early diastolic transmittal flow velocity (E, mm/s). F: Early diastolic mitral annulus velocity (e’, mm/s), and F: Isovolumetric relaxation time (IVRT, msec). An N = 9-15 was used, mean ± S.E.M is shown. A 2-way ANOVA with Tukey correction was used as described in the Supplemental Materials. Means, exact group sizes, and adjusted p-values are presented in Table S1 and S2. Parameters with significant p-values between time points are noted directly on the figure. * vs Non-Tg early, # vs Non-Tg late, † vs R92L-cTnT early, ‡ vs R92L-cTnT late.

Figure 2.

Phosphorylation status of cTnI and PLB in cardiac homogenates and cTnI in human CTFs. A: Representative western blots of phosphorylated cTnI-S₂₃S₂₄ (pTnI) ratio to cTnI, and the Ser-16 phosphorylation of PLB (pPLB) to total PLB, performed on ventricular homogenates before (−ISO) and following (+ISO) isoproterenol stimulation in 4–6-month-old Non-Tg, R92L-cTnT, and Δ160E-cTnT mice. B: Summary of pTnI/TnI and C: pPLB/PLB ratios before and following isoproterenol stimulation. D: Representative ProQ and Coomassie stains. Response of CTFs to active enzyme from 5 to 30 minutes shown as a E: percentage of WT response and F: absolute relative response. Created with Biorender.com. Values are reported as mean ± S.E.M., N = 10 hearts per group and N = 4 reconstitutions. A 2-way ANOVA with Tukey and Kruskal-Wallis with Dunn correction were used respectively as described in the Supplemental Materials. Adjusted p-values are presented in Table S3 and S4. B and C: * p<0.05 vs Non-Tg -ISO, ## p<0.01 vs Non-Tg +ISO, ‡ p<0.05 vs R92L +ISO. E and F: * p < 0.05 vs WT CTF. p-values for comparisons between −ISO and +ISO are noted directly on the figure.

Figure 3:

Cardiac function assessed via 2D echocardiography at early and late time points in Non-Tg and R92L-cTnT mice bred to express phosphomimetic cTnI. A: Systolic function assessed by % Ejection Fraction and B: % Fractional Shortening, C: Total wall thickness (Td, mm), D: Diastolic function assessed by E/e’, E: Early diastolic transmittal flow velocity (E, mm/s). F: Early diastolic mitral annulus velocity (e’, mm/s), and F: Isovolumetric relaxation time (IVRT, s). An N = 8-15 was used, mean ± S.E.M is shown. A 2-way ANOVA with Tukey correction was used as described in the Supplemental Materials. Means, exact group sizes, and adjusted p-values are presented in Table S5 and S6. Parameters with significant p-values between time points are noted directly on the figure. * vs Non-Tg early, # vs Non-Tg late, † vs R92L-cTnT early, ‡ vs R92L-cTnT late.

Figure 4:

Cardiac function assessed via 2D echocardiography at early and late time points in Non-Tg and Δ160E-cTnT mice bred to express phosphomimetic cTnI. A: Systolic function assessed by % Ejection Fraction and B: % Fractional Shortening, C: Total wall thickness (Td, mm), D: Diastolic function assessed by E/e’, E: Early diastolic transmittal flow velocity (E, mm/s). F: Early diastolic mitral annulus velocity (e’, mm/s), and F: Isovolumetric relaxation time (IVRT, msec). An n = 7-15 was used, mean ± S.E.M is shown. A 2-way ANOVA with Tukey correction was used as described in the Supplemental Materials. Means, exact group sizes, and adjusted p-values are presented in Table S7 and S8. Parameters with significant p-values between time points are noted directly on the figure. * vs Non-Tg early, # vs Non-Tg late, † vs Δ160E-cTnT early, ‡ vs Δ160E-cTnT late.

Figure 5:

Ca²⁺ dissociation rates from reconstituted CTFs. A: Representative traces. B: Rate of Ca²⁺ dissociation in WT, WTDD, R92L-cTnT, RLDD, Δ160E-cTnT, and ΔDD CTFs. Values are expressed as mean ± S.E.M. A 2-way ANOVA with Tukey correction for MC was used as described in the Supplemental Materials. * p < 0.05 vs WT, #### p < 0.0001 vs baseline (without phosphomimetic) within genotype, †††† p < 0.0001 vs WTDD, ‡ vs RLDD. Adjusted p-values are presented in Table S9.

Figure 6:

Structure of the WT cTnC-cTnI Interface determined by TR-FRET and MD. Distance distributions of the WT cTnC-cTnI interface generated via TR-FRET between A: cTnC84C to cTnIA9C, B: cTnC84C to cTnIA17C, C: cTnC84C to cTnIA28C. Means of the distances and FWHMs obtained and adjusted p-values are presented in Table S10 and S11. D: MD of the WT cTnC-cTnI interface. E: Zoom-in of box outlined in panel D. F: Relevant side chains interaction labeled with their respective side chain amino acid and site (eg. Ser23), solid line indicates single interaction, PKA site in bold. Created with Biorender.com.

Figure 7:

TR-FRET flanking the PKA recognition sites in CTFs for WT, R92L-cTnT, and Δ160E-cTnT. For TR-FRET pair cTnC84C-cTnIA17C - A: Distances (Å), B: FWHM (Å), and C: Averages distance distribution plots. For FRET pair cTnC84C-cTnIA28C - D: Distances (Å), E: FWHM (Å), and F: Averages distance distribution plots. Values are expressed as mean ± S.E.M., N = 9-14 per group. A 2-way ANOVA with Tukey correction was used as described in the Supplemental Materials. Means of distances and FWHMs obtained and adjusted p-values are presented in Table S12 and S13. * p< 0.05, ** p<0.01 vs WT -Ca; # p<0.05, ## p<0.01 vs R92L-cTnT -Ca; †p<0.05 vs Δ160E-cTnT -Ca.

Figure 8:

Interactions at the cTnC-cTnI interface for R92L-cTnT and Δ160E-cTnT produced via MD. A: MD of the R92L-cTnT cTnC-cTnI interface as in Figure 7D with the R92L-cTnT overlayed in green over the WT (magenta). B: Zoom-in of the box outlined in panel A. Side chain interactions shown are those altered by the inclusion of R92L-cTnT. C: Relevant side chain interactions labeled with their respective side chain amino acid and site (eg. Ser23), solid line indicates single interaction, dashed indicates 2 interactions, PKA site in bold. D: MD of the Δ160E-cTnT cTnC-cTnI interface with the Δ160E-cTnT overlayed in orange. E: Zoom-in of the box outlined in panel F, interactions shown are those of WT as no interactions are present for Δ160E-cTnT. F: Relevant side chain interactions presented as in panel C. Created with Biorender.com.