Titin is the largest protein that has been identified, spanning from the Z-disk to the M-band of the cardiac sarcomere (LeWinter et al., 2007). Acting as a molecular spring in the I-band, titin contributes to passive tension as sarcomeres are stretched and influences diastolic suction, or elastic recoil at short sarcomere lengths (Granzier and Irving, 1995; Helmes et al., 1996; Wu et al., 2000). Titin compliance is primarily dependent upon differential splicing, resulting in isoforms of different lengths (Labeit and Kolmerer, 1995; Freiburg and Gautel, 1996; Wu et al., 2000). RNA Binding Motif Protein 20 (RBM20) suppresses differential titin splicing such that wild-type mice (WT) predominantly express the stiffer N2B titin isoform and homozygous Rbm20^ΔRRM mice express a very large, more compliant N2BA titin isoform (Guo et al., 2013; Li et al., 2013; Methawasin et al., 2014).

Actin-myosin cross-bridge behavior is regulated by intracellular [Ca²⁺] and sarcomere length, both of which are constantly changing throughout the heartbeat (for reviews see Tobacman, 1996; Cooke, 1997; Gordon et al., 2000; Kobirumaki-Shimozawa et al., 2014). Previous studies have shown that increased N2BA expression reduces passive tension (Fukuda et al., 2003; Makarenko et al., 2004; Nagueh et al., 2004; Hanft et al., 2014) which can compromise maximal Ca²⁺-activated tension production and reduce Ca²⁺-sensitivity of the tension-pCa relationship (Fukuda et al., 2001, 2003; Hanft et al., 2014; Methawasin et al., 2014). Increased myocardial compliance in Rbm20^ΔRRM mice and rats also demonstrated an attenuated Frank-Starling response (Methawasin et al., 2014; Ait-Mou et al., 2016). We have recently shown that cross-bridge cycling kinetics slowed at longer sarcomere length due to slowing of MgATP binding and MgADP release (Tanner et al., 2015). This led to the hypothesis that increased sarcomeric compliance in Rbm20^ΔRRM hearts could affect cross-bridge cycling kinetics differently at short vs. long sarcomere lengths, which may provide an explanation for compromised myocardial function in Rbm20^ΔRRM vs. WT myocardium.

To test this hypothesis we measured tension-pCa relationships, and cross-bridge kinetics at 1.9 and 2.2 μm sarcomere length in skinned papillary muscle strips from WT and Rbm20^ΔRRM mice. We found increased titin compliance in the Rbm20^ΔRRM strips resulted in decreased maximal tension, depressed Ca²⁺-sensitivity of the tension-pCa relationship, and slowed MgADP release compared to WT strips at each sarcomere length. As sarcomere length increased from 1.9 to 2.2 μm sarcomere length, Rbm20^ΔRRM strips showed a minimal increase in maximal tension Ca²⁺-sensitivity of the tension-pCa relationship, while WT strips demonstrated a robust increase in tension and Ca²⁺-sensitivity of the tension pCa relationship. These findings suggest that titin compliance influences sarcomere-length dependent activation of contraction and cross-bridge nucleotide handling rates, influencing myocardial function more greatly at longer sarcomere length.

There were no obvious differences in sarcomere organization or monitored sarcomere length in skinned papillary muscle strips from WT and Rbm20^ΔRRM mice (Figure 1A). As skinned myocardial strips were Ca²⁺-activated from pCa 8.0 to pCa 4.8, steady-state, isometric tension developed in a sigmoidal manner that was fit to a 3-parameter Hill equation (Figures 1B–F, Table 1). These tension-pCa relationships are shown two different ways, where: (i) where absolute tension values (=measured force values normalized to cross-sectional area of each myocardial strip; Figures 1C,D) illustrate the total tension produced by the strip (i.e., both the passive tension value at pCa 8.0 plus the Ca²⁺-activated active tension values), and (ii) developed tension values illustrate the Ca²⁺-activated tension produced by the strip (i.e., absolute tension minus the passive, relaxed tension value at pCa 8.0; Figures 1E,F).

Under maximally activated conditions, myocardial strips with both WT and Rbm20^ΔRRM genotypes displayed greater absolute tension at 2.2 vs. 1.9 μm sarcomere length (Figures 1C,D). Relaxed tension values (pCa 8.0) were also greater at the longer sarcomere length in both genotypes (Table 1). Developed tension was greater at 2.2 vs. 1.9 μm sarcomere length from pCa 5.8–4.8 in myocardial strips from WT mice (Figure 1E). However, in myocardial strips from the Rbm20^ΔRRM mice, developed tension was only greater at 2.2 μm sarcomere length at pCa 5.5 and 5.4 (Figure 1F). Thus, Ca²⁺-sensitivity of the tension-pCa relationship increased with sarcomere length in the WT strips (by ~0.08 pCa units), but this sarcomere length-dependent increase in Ca²⁺-sensitivity of tension was lost in Rbm20^ΔRRM strips (Table 1). At 2.2 μm sarcomere length, WT strips also displayed greater Ca²⁺-sensitivity of tension than Rbm20^ΔRRM strips (by ~0.07 pCa units). In WT strips, the Hill coefficient (n_H) for the tension-pCa relationship was smaller at 2.2 vs. 1.9 μm sarcomere length, indicating reduced cooperativity at longer sarcomere length (Table 1). At 2.2 μm sarcomere length, n_H was smaller for WT strips than Rbm20^ΔRRM strips; there were no differences in n_H between genotypes at 1.9 μm sarcomere length.

In both genotypes under relaxed conditions (pCa 8.0), elastic moduli values were greater at 2.2 vs. 1.9 μm sarcomere length for all frequencies >1.5 Hz (Figures 2A,C). Viscous moduli values were also greater at longer sarcomere length at frequencies >51 Hz in WT strips and frequencies >54 Hz in Rbm20^ΔRRM strips (Figures 2B,C, respectively). Under activated conditions (pCa 4.8, 5 mM MgATP), elastic moduli values were greater at longer sarcomere length for frequencies above 145 Hz in WT (Figure 3A), and frequencies >22 Hz in Rbm20^ΔRRM strips (Figure 3C). In addition to these moduli differences, there was a consistent shift toward lower frequencies for the overall elastic moduli-frequency relationship at longer sarcomere length; this shift toward lower frequencies was larger for Rbm20^ΔRRM strips vs. WT strips. Under activated conditions, viscous moduli were not different at any particular sarcomere length in the WT strips (Figure 3B), and viscous moduli were greater at 2.2 vs. 1.9 μm sarcomere length at frequencies between 9.5 and 54 Hz in the Rbm20^ΔRRM strips (Figure 3D). There was also a consistent shift toward lower frequencies for the overall viscous moduli-frequency relationship at longer sarcomere length; this shift toward lower frequencies was larger for Rbm20^ΔRRM strips vs. WT strips. Altogether these data indicate greater myocardial viscoelasticity at longer sarcomere length under relaxed and activated conditions, although the influence of titin compliance was minimal as there were no significant effects of genotype in the moduli-frequency relationships (Figures 2, 3). The length-dependent shifts toward lower frequencies in the moduli-frequency relationships at pCa 4.8 indicate slower cross-bridge cycling as sarcomere length increased for both genotypes, although this slowing was greater for Rbm20^ΔRRM.

Moduli values were fit to Equation (1) to extract model parameters related to viscoelasticity, cross-bridge binding, and cross-bridge kinetics as the skinned strips were titrated toward rigor (5.0–0.05 mM MgATP, pCa 4.8). These model parameters are plotted against [MgATP] in Figure 4, with p-values listed in the left panel for each parameter that demonstrated significant main effects or interactions from the mixed-model analysis. As [MgATP] was titrated toward rigor, A values increased and k values decreased for both genotypes, suggesting increased viscoelastic myocardial stiffness that became more elastic (vs. viscous) due to greater cross-bridge binding as MgATP decreased (Figures 4A–D). In both genotypes, A values were greater and k values were smaller at 2.2 vs. 1.9 μm sarcomere length, which represents greater myocardial viscoelasticity due to a combination of: (i) passive elements of the sarcomere being stretched or extended more at 2.2 vs. 1.9 μm sarcomere length and (ii) greater binding of slower-cycling cross-bridges at 2.2 vs. 1.9 μm sarcomere length. For both genotypes, the values for B and C increased as [MgATP] was titrated toward rigor and the magnitudes for C increased at 2.2 vs. 1.9 μm sarcomere length (Figures 4E,F), also suggesting greater cross-bridge binding at longer sarcomere length.

As [MgATP] decreased, cross-bridge attachment rate (2πb, Figures 4I,J) slowed in both genotypes. The significant MgATP × genotype interaction suggests cross-bridge attachment rate was more sensitive to [MgATP] in WT than in Rbm20^ΔRRM strips, although cross-bridge attachment rates were not different at 2.2 vs. 1.9 μm sarcomere length. Similarly, as [MgATP] decreased toward rigor, cross-bridge detachment rate (2πc, Figures 4K,L) slowed in both genotypes. Cross-bridge detachment rates were also slower at 2.2 vs. 1.9 μm sarcomere length for both genotypes. The significant genotype effect on cross-bridge detachment rate suggests that Rbm20^ΔRRM strips displayed a slower cross-bridge detachment rate than WT strips across the entire [MgATP] range, although this statistic was primarily driven by the slowest detachment rates occurring for Rbm20^ΔRRM strips at 2.2 μm sarcomere length. Again the significant MgATP × genotype interaction suggests that cross-bridge detachment rate was more sensitive to [MgATP] in WT than in Rbm20^ΔRRM strips.

Fitting the 2πc-MgATP relationship to Equation (2) (solid lines in Figures 4K,L) provides an estimate of the cross-bridge rates of MgADP release (k_−ADP) and MgATP binding (k_+ATP). The MgADP release rate slowed with increased titin compliance for the Rbm20^ΔRRM fibers, and there was a length-dependent slowing of k_−ADP at longer sarcomere length for both genotypes (Table 2). For WT fibers, increasing sarcomere length from 1.9 to 2.2 μm slowed MgADP release by 12% (p = 0.015 using a t-test). For Rbm20^ΔRRM fibers k_−ADP slowed 22% as sarcomere length increased from 1.9 to 2.2 μm, showing about twice as much length-dependent slowing of k_−ADP for Rbm20^ΔRRM than WT. However, increased titin compliance in the Rbm20^ΔRRM strips led to slower rates of MgADP release at both sarcomere lengths (13 and 23% slower at short and long sarcomere length, respectively), compared to WT k_−ADP values. The cross-bridge rate of MgATP binding did not differ with genotype or with sarcomere length (Table 2). These finding suggest that increased compliance of the myofilament lattice slows cross-bridge cycling kinetics, primarily due to slower MgADP dissociation from cross-bridges.

Computational simulations of muscle contraction have demonstrated that mechanical characteristics of the sarcomere (i.e., filament, cross-bridge, and titin compliance; compliance = stiffness⁻¹) influence the dynamics of cross-bridge binding and tension generation in a muscle fiber (Daniel et al., 1998; Martyn et al., 2002; Chase et al., 2004; Campbell, 2006, 2009, 2016; Sheikh et al., 2012; Tanner et al., 2012a, 2014). These mathematical models predict that increasing sarcomeric compliance diminishes steady-state tension, slows the apparent rate of tension development, slows cross-bridge cycling rates, and can impact the rate of tension relaxation as well. As RMB20^ΔRRM mice express more of the compliant N2BA titin isoform than the WT (Guo et al., 2013; Methawasin et al., 2014), these transgenic animals represent a useful model system to test some of these model predictions and directly assess the role of titin compliance in length-dependent tension production and ensemble cross-bridge behavior in skinned myocardial strips. In this study we observed that increased titin compliance in Rbm20^ΔRRM fibers diminished steady-state tension, reduced Ca²⁺-sensitivity of the tension-pCa relationship, and slowed cross-bridge detachment rate due to slowed MgADP dissociation from strongly-bound cross-bridges. The effects of titin compliance were sarcomere length-dependent, showing almost no length-dependent tension response in Rbm20^ΔRRM strips, in contrast to the robust length-dependent increase in maximal tension and Ca²⁺-sensitivity of the tension-pCa relationships between 1.9 and 2.2 μm sarcomere length in WT strips. This length-dependent activation response was eliminated in Rbm20^ΔRRM strips despite a slowed cross-bridge detachment rate as sarcomere length increased, which would be expected to enhance thin-filament activation at 2.2 μm sarcomere length due to strong cross-bridge binding (Bremel and Weber, 1972; Wang and Fuchs, 1994; Metzger, 1995; Fitzsimons and Moss, 1998; Smith et al., 2009; Terui et al., 2010; Li et al., 2014). Empirical findings in this study support previous computational simulations predicting the important role that sarcomeric compliance plays in muscle contraction and further suggests that titin mechanics affect length dependent activation of contraction, perhaps by altering how tension propagates throughout the sarcomere to influence thin-filament activation.

Our observations that increased titin compliance in the Rbm20^ΔRRM strips reduced maximal tension values and decreased Ca²⁺ sensitivity of tension agree with previous findings that suggest greater N2BA titin isoform expression depresses maximum tension production (Makarenko et al., 2004; Lewinter et al., 2010; Patel et al., 2012; Hanft et al., 2014; Methawasin et al., 2014). Our measurements also show that effects of titin compliance on Ca²⁺-activated tension are sarcomere length-dependent, supporting previous studies showing that increased titin compliance depresses tension more significantly at longer sarcomere length (Fukuda et al., 2003; Methawasin et al., 2014). This is most evident by the similar tension-pCa relationships at 1.9 μm sarcomere length among both genotypes (Figure 1; Table 1), with a robust length-dependent increase in Ca²⁺-activated tension production as sarcomere length increased to 2.2 um for WT strips that did not occur for Rbm20^ΔRRM strips. These data suggest that length-dependent activation of contraction and the slope of the ascending limb of the sarcomere-length vs. Ca²⁺-activated tension relationship may depend upon mechanical characteristics of titin. This implies that dynamic processes related to cross-bridge cycling kinetics, thin-filament activation, and tension development within the sarcomere may be influenced by the mechanical characteristics of titin.

Cross-bridge detachment rates slowed as sarcomere length increased from 1.9 to 2.2 μm among both genotypes, but the slowing was more pronounced for the Rbm20^ΔRRM strips. For WT strips, slower myosin detachment at 2.2 μm sarcomere length effectively enhances cross-bridge contributions to thin-filament activation to augment tension production and Ca²⁺-sensitivity of the tension-pCa relationship. Previous studies have linked greater Ca²⁺-affinity of troponin C and greater opening of the N-terminus of troponin C with increases in strong cross-bridge binding (Hofmann and Fuchs, 1987; Wang and Fuchs, 1994; Terui et al., 2008, 2010; Smith et al., 2009; Li et al., 2014), and our current findings in WT strips and rat papillary muscle strips (Tanner et al., 2015; Pulcastro et al., 2016) imply this cooperative activation pathway becomes stronger at longer sarcomere lengths. The MgADP release rate (k_−ADP) was 13% slower at 1.9 μm and 23% slower at 2.2 μm sarcomere length in Rbm20^ΔRRM strips, compared to WT strips, which would be expected to slow cross-bridge detachment and stabilize, or amplify thin-filament activation more greatly in Rbm20^ΔRRM strips. However, slower cross-bridge detachment rates did not enhance tension nor length-dependent activation of contraction in Rbm20^ΔRRM strips with greater titin compliance. Thus, cross-bridge contributions to thin-filament activation and increased Ca²⁺-affinity of troponin C may require titin interacting with the thin-filament or titin transmitting tension between the thick and thin-filament. Increased titin compliance in Rbm20^ΔRRM strips may compromise this titin interaction or tension transmission pathway, thereby depressing Ca²⁺-activated tension production and length-dependent activation of contraction.

Some muscle mechanics studies use large amplitude release-restretch protocols (~15% muscle length) to assess the cross-bridge rate of tension redevelopment (k_tr), in comparison to the low amplitude strains used for stochastic length perturbation analysis (<0.15% muscle length). Skinned myocardial strips from WT and Rbm20^ΔRRM strips mice showed no differences in sarcomere length-dependent k_tr under maximally Ca²⁺-activated conditions (Methawasin et al., 2014). Previous studies using skinned myocardium from rats expressing the more compliant N2BA titin isoform have shown mixed reports of slower and faster k_tr values as sarcomere length increased (Patel et al., 2012; Hanft et al., 2014), compared to wild-type controls that predominantly express the stiffer N2B titin isoform. Herein we measured cross-bridge kinetics as [MgATP] varied, which allowed us to estimate cross-bridge rates of MgADP release (k_−ADP) or MgATP binding (k_+ATP; Table 2). As the rate of MgADP release limits cross-bridge detachment in a muscle fiber (Siemankowski et al., 1985), the ~12% slowing in k_−ADP from 1.9 to 2.2 μm sarcomere length drives the length-dependent slowing of cross-bridge detachment in WT strips. However, the length-dependent slowing of k_−ADP was nearly twice as great in Rbm20^ΔRRM strips (~22%) and k_−ADP was also slower at each sarcomere length when titin compliance increased. There was not a significant difference between cross-bridge MgATP binding rates at 1.9 vs. 2.2 μm sarcomere lengths for either genotype. These data support our previous observations that slowed MgADP release rate is the predominate step of the cross-bridge cycle that is responsible for the length-dependent slowing of cross-bridge kinetics (Tanner et al., 2015; Pulcastro et al., 2016). We do not think these slowed nucleotide handling kinetics in Rbm20^ΔRRM strips stem from any α-to-β myosin heavy chain isoform shift, because Methawasin et al. (2014) reported solely α-myosin heavy chain expression in both of these mouse lines. Moreover, these data also demonstrate that titin compliance influences sarcomere length-dependent cross-bridge nucleotide handling rates, and the effects of titin on cross-bridge kinetics become greater as sarcomere length increases.

Under relaxed conditions, both viscoelastic mechanical stiffness (Figure 2) and steady-state tension values (Table 1) were greater at longer sarcomere length, without any differences between the two genotypes. These differences stem from passive elements of the sarcomere being stretched or extended more greatly at 2.2 vs. 1.9 μm sarcomere length [i.e., titin and collagen (Granzier and Irving, 1995)]. We had anticipated that length-dependent increases in relaxed stiffness and tension would be greater for WT vs. Rbm20^ΔRRM, similar to previous observations using skinned myocytes (Methawasin et al., 2014). However, Methawasin et al. (2014) also showed greater collagen expression in Rbm20^ΔRRM vs. heterozygous Rbm20 knockout mice, which could be a compensatory mechanism to increase myocardial stiffness as titin compliance decreased in the homozygous mice. Given that skinned myocardial strips encompass some component of passive stiffness due to collagen that isn't present in isolated myocytes, it is possible that the mechanical characteristics of collagen, rather than titin, are dominating our relaxed muscle mechanics measurements. While previous studies suggested that greater passive tension values are correlated with greater Ca²⁺-activated tension production and length-dependent activation of contraction (Fukuda et al., 2001, 2003), our measurements do not support this mechanism driving length-dependent activation because relaxed stiffness and tension values were similar at each sarcomere length for both genotypes.

Thick-to-thin-filament spacing consistently decreases as sarcomere length increases in skinned and intact muscle preparations (Matsubara and Millman, 1974; Irving et al., 2000; Konhilas et al., 2002; Smith et al., 2009). Mechanical characteristics of titin influence this lattice spacing vs. sarcomere length relationship, showing that increased titin compliance can increase myofilament lattice spacing and affect the relationship between lattice spacing and sarcomere length (both increasing and decreasing the slope of this relationship; Cazorla et al., 2001; Fukuda et al., 2001, 2003, 2005; Irving et al., 2011). In addition, recent measurements show smaller myofilament lattice spacing values in Rbm20 knockout rat myocardium at both short and long sarcomere length, compared to wild-type controls (Ait-Mou et al., 2016). Cross-bridge cycling rates have been shown to slow as thick-to-thin-filament spacing decreased in vertebrate and invertebrate muscle fibers that were osmotically compressed with Dextran (Krasner and Maughan, 1984; Kawai and Schulman, 1985; Smith et al., 2009; Tanner et al., 2012b) and with increases in sarcomere length in skinned (Adhikari and Wang, 2004; Tanner et al., 2015; Pulcastro et al., 2016), and intact (Milani-Nejad et al., 2013) cardiac muscle preparations. Therefore, increases in sarcomere length will accompany decreases in thick-to-thin-filament spacing, which could contribute to slower cross-bridge detachment at longer sarcomere length for both genotypes.

While reduced lattice spacing may slow cross-bridge cycling, this does not translate into increased length-dependent activation of contraction in Rbm20^ΔRRM strips (Hanft et al., 2014; Methawasin et al., 2014). Therefore, slowed cross-bridge cycling kinetics may not be the primary mechanism responsible for increasing Ca²⁺-activated tension at long sarcomere length (Patel et al., 2012), particularly when titin compliance increases from normal. Perhaps, titin interacts with the thin-filament to influence thin-filament activation and length-dependent activation of contraction, either directly or by influencing load (or strain) borne by thin-filament proteins (Terui et al., 2008; Hanft et al., 2014). This titin-thin-filament activation pathway may be suppressed with the more compliant titin in Rbm20^ΔRRM fibers, because titin is less taut and cannot effectively transmit tension between the M-band and Z-disks to maintain tension throughout the sarcomere. Thus, increases in cross-bridge duty ratio due to slowed detachment kinetics in Rbm20^ΔRRM fibers do not necessarily translate into the greater tension production due to a compromised capacity to generate tension or distribute tension throughout a more compliant sarcomere. Altogether, this would diminish ventricular function, and may scale with the expression ratio between the more compliant N2BA titin isoform and the stiffer N2B titin isoform. These impaired mechanisms of thin-filament activation and tension production may contribute to cardiac dysfunction and the associated cardiomyopathies in humans, rats, and mice bearing RBM20 mutations that influence titin splicing (Makarenko et al., 2004; Nagueh et al., 2004; Guo et al., 2012; Methawasin et al., 2014).

HP, PA, MM, and BT participated in performing the experiments and data collection. BT, WD, and HG conceived and designed the experiments. HP, PA, and BT analyzed the data. All authors helped interpret the data, write, and revise the manuscript, and have approved the final version of this manuscript.