Titin Develops Restoring Force in Rat Cardiac Myocytes
When relaxed after contraction, isolated cardiac myocytes quickly relengthen back to their slack length. The molecular basis of the force that underlies passive relengthening, known as restoring force, is not well understood. In a previous study of titin's elasticity in cardiac myocytes, we proposed that titin/connectin develops restoring force, in addition to passive force. This study tested whether titin indeed contributes to the restoring force in cardiac myocytes. Skinned rat cardiac myocytes in suspension were shortened by ≈20%, using Ca2+-independent shortening, followed by relaxation. Cells were observed to relengthen until they reached their original slack sarcomere length. However, the ability to relengthen was abolished after cells had been treated for 12 minutes with trypsin (0.25 μg/mL, 20°C). Gel electrophoresis showed that this treatment had degraded titin without clearly affecting other proteins, and immunoelectron microscopy revealed that the elastic segment of titin in the I band was missing from the sarcomere. Restoring force was also directly measured, before and after trypsin treatment. Restoring force of control cells was −61±20 μg (per cell) at a sarcomere length of 1.70 μm. Comparison of our results with those of activated trabeculae indicated that a large fraction of restoring force of cardiac muscle originates from within the myocyte. Restoring force of myocytes was found to be depressed after titin had been degraded with trypsin. We conclude that cardiac titin indeed develops restoring force in shortened cardiac myocytes, in addition to passive force in stretched cells, and that titin functions as a bidirectional spring. Our work suggests that at the level of the whole heart, part of the actomyosin-based active force that is developed during systole is harnessed by titin, allowing for elastic diastolic recoil and aiding in ventricular filling.
When the heart contracts, a restoring force is recruited that, during diastole, allows the heart to restore its original shape and aids ventricular filling.1 This property of the heart can also be observed in isolated cardiac myocytes; when relaxed after contraction, myocytes will relengthen back to their slack length.2 The molecular basis for the restoring force underlying this passive relengthening is not well understood.3 4 Our previous immunoelectron microscopic study on the endosarcomeric protein titin (also named connectin) of cardiac myocytes led to the prediction that titin develops restoring force.5 The aim of the present study was to test this prediction.
Titin is a giant protein that spans from the Z line to the M line in sarcomeres of both cardiac and skeletal muscles.6 7 8 9 10 Titin has elastic properties and is likely to develop passive force in sarcomeres stretched beyond their slack length.11 12 13 14 Our recent study on rat cardiac myocytes indicates that only a small segment of the I-band domain of titin is elastic and that in sarcomeres that are slack (≈1.85 μm), the elastic titin segment is highly folded.5 Based on this work, a two-stage mechanism of passive-force development in the heart was proposed, in which (1) between sarcomere lengths of ≈1.85 μm and ≈2.0 μm, titin's elastic segment straightens, developing entropic force, and (2) at lengths longer than ≈2.0 μm, the same segment elongates by unfolding of molecular subdomains. An interesting prediction of our previous work is that sarcomere shortening to lengths below slack also results in straightening of the elastic titin segment, giving rise to an entropic force that opposes shortening (restoring force) and that tends to bring sarcomeres back to their slack length.
Accordingly, in the present study, we tested the hypothesis that titin contributes to restoring force. We have followed the relengthening of single cardiac myocytes that is normally observed when unattached shortened cells are relaxed. It was found that degrading the elastic segment of titin, by limited tryptic proteolysis, abolished the ability of the myocytes to relengthen. We also directly measured the restoring force of short cell segments, using a method similar to that of Fabiato and Fabiato.15 Our observations indicate that a large part of the measured restoring force of myocytes originates from titin. Thus, titin might function as a bidirectional spring that develops passive force in sarcomeres stretched beyond their slack length and restores force in ones shortened below this length.
Materials and Methods
Preparation of Cardiac Myocytes
Single cardiac myocytes were isolated according to the method used by Granzier and Irving.14 Briefly, cells were isolated from the rat (male Sprague-Dawley, ≈250 g) heart by perfusing the coronary arteries with oxygenated Krebs' solution containing collagenase. Atria were subsequently removed, and ventricles were cut into small pieces that were gently drawn several times through a plastic pipette. This released a large number of isolated cells, the majority of which had a normal rodlike shape. Cells were then washed extensively and skinned for 50 minutes in relaxing solution with 1% Triton, followed by washing with relaxing solution only. To prevent titin degradation, all solutions contained protease inhibitors (compare with Granzier and Irving14 ).
The setup has been described in detail by Granzier and Irving.14 Briefly, cells were added to a temperature-controlled flow-through chamber (volume, ≈300 μL) that was attached to the stage of a phase-contrast microscope. One end of a single cell was glued to a motor; its other end, to a force transducer. To reduce force artifacts caused by the effect of fluid level fluctuation on the force transducer, the transducer entered the chamber horizontally, via an opening in the chamber wall. A gravity-based perfusion system was used, which had an ≈10-second delay between switching to a new perfusate and the arrival of this perfusate at the cell and an additional ≈30-second delay before the new perfusate had completely replaced the old chamber content. Rigor force was measured during continuous perfusion. Peak-to-peak force noise was typically ≈40 μg. Passive force and restoring force were measured in the absence of perfusion in order to reduce peak-to-peak force noise to ≈20 μg.
To obtain reproducible results, we used a computer-controlled mechanics workstation designed to impose identical stretch/release protocols, while force and length signals were concomitantly sampled with high spatial and temporal resolution. Stretch protocols consisted of a stretch to a predetermined amplitude at a constant velocity, immediately followed by a release at the same constant speed. Stretch and release were each completed in 45 seconds.
Sarcomere length was measured by the method of Granzier and Irving.14 Briefly, phase-contrast images of cells were digitized using a computer, and density traces along the long axis of the cell were measured. Density traces were restricted typically to a width of ≈4 μm and a length of either ≈50 μm (unattached cells) or ≈15 μm (cell segments used for restoring force measurements). The traces were processed by discrete Fourier transformation, and the position of the first-order peak in the power spectrum was used to calculate sarcomere length. Three different regions of the cell (encompassing 50% to 75% of the cell width) were analyzed, and these results were averaged to obtain a single representative value.
Measurement of Restoring Force
Relaxed cell segments were first buckled, and the chamber was then perfused with rigor solution. After a delay, sarcomeres shortened until the buckle disappeared. The perfusion buffer was then switched from rigor to relaxing solution. The sarcomere length of the relaxed cell was ≈1.60 to 1.70 μm. We subsequently stretched the relaxed cell to a sarcomere length of ≈2.05 μm while measuring force. The force measured below the slack sarcomere length (≈1.85 μm) was defined as restoring force, and the force measured above slack was defined as passive force. The restoring force measurement protocol is depicted graphically in Fig 1⇓.
Myocytes in Suspension
Mild trypsin digestion has been used to specifically degrade titin and study its involvement in passive force development,12 13 14 and we used this method to study whether titin develops restoring force. Cells in relaxing solution were centrifuged, and the pellet (volume, ≈0.5 mL) was resuspended in 13 mL leupeptin-free rigor solution (mmol: imidazole 40, EGTA 10, NaN3 5, potassium proprionate 90, and dithiothreitol 1, pH 7.0 at 20°C). This procedure was repeated five times, which diluted both leupeptin and ATP by a factor of >105. Trypsin (Sigma, type III, 12 700 BAEE U/mg) was made up fresh in leupeptin-free rigor solution, at a concentration of 0.5 μg/mL. The rigorized cells were divided equally over two tubes. The cell suspension in one tube was diluted with an equal volume of trypsin (final concentration, 0.25 μg/mL). The cells in the other tube were diluted with an equal volume of leupeptin-free rigor solution and functioned as controls. Both tubes were then incubated at 20°C. At regular time intervals, samples were taken from the tubes and were quickly added to tubes on ice containing relaxing solution with leupeptin, phenylmethylsulfonyl fluoride, and diisopropyl fluorophosphate to final concentrations of 0.25, 0.5, and 5 mmol/L, respectively. This inhibited trypsin and relaxed the cells from rigor. The average sarcomere length of relaxed cells was determined for 10 randomly chosen cells for each sample. Sarcomere length measurement took place after the cells had been in relaxing solution for at least 30 minutes. The remainder of each sample was solubilized and electrophoresed to establish the effect of trypsin on the protein composition of the cells.
Trypsin Digestion Combined With Mechanics
Single cells were first mechanically characterized in leupeptin-containing relaxing solution. Leupeptin was then removed by flushing 80 mL of leupeptin-free relaxing solution through the chamber. Next, trypsin was added by washing 20 mL of trypsin-containing relaxing solution (0.25 μg trypsin/mL) through the chamber, followed by a 12-minute incubation at 20°C. The reaction was stopped by perfusing the chamber with 10 mL relaxing solution containing 40 μg/mL leupeptin. Restoring force– and passive force–sarcomere length relationships for each myocyte were determined before and after trypsin treatment.
Immunolabeling and Electron Microscopy
The effect of tryptic digestion on the elastic segment of titin was studied using immunoelectron microscopy on single cardiac myocytes. The anti-titin antibody 9D10, developed by Dr Greaser, was obtained from the Developmental Studies Hybridoma Bank maintained by the Department of Biological Sciences, University of Iowa, Iowa City (under contract NO1-HD-2-3144 from the National Institute of Child Health and Human Development). Antibody characterization and further technical details can be found in Granzier et al.5
Cells were solubilized and electrophoresed using 2.5% to 12% acrylamide gradient gels16 that were stained with Coomassie blue.
Results are given as mean±SD. For multiple comparisons in the relengthening experiments, an ANOVA followed by Dunnett's test was used.
Titin Degradation by Trypsin
The role of titin in restoring force development was studied by using a mild trypsin treatment on cardiac myocytes to degrade titin. To prevent titin degradation during the cell isolation protocol, leupeptin was added to all solutions, and leupeptin was removed from the cell suspension before starting the trypsin treatment. To assess the integrity of titin before tryptic digestion, we carried out gel electrophoresis of leupeptin-free cells. Titin consisted of a doublet: T1 and T2 (see Fig 2⇓ controls); T2 is generally considered a degradation product of T1.6 The T1-to-T2 ratio of leupeptin-free control cells was somewhat reduced relative to that of leupeptin-containing control cells (compare results of Fig 2⇓ with those of Granzier and Irving14 ), indicating that these cells may contain a small amount of degraded titin.6 Nevertheless, these cells behaved normally and could be reversibly shortened and relaxed (see below).
When trypsin was added to the cells, titin was rapidly digested. After a 12-minute trypsin treatment, T1 was completely degraded, and two new bands appeared on the gels, one below T2 and another below myosin heavy chain (MHC) (Fig 2⇑). Extending the trypsin treatment to 30 minutes resulted in some degradation of T2 and an increase of the two degradation products (Fig 2⇑, lane 3). Inspection of the gels revealed that the only detectable protein degraded was titin, and other proteins were not affected, confirming earlier studies.12 13 14 This conclusion was supported by our observation that trypsin-treated cells vigorously contracted in activating solution (data not shown) and by results from our restoring force measurement protocols (see below), which indicated that trypsin did not significantly affect rigor force (Fig 3⇓, Table).
Tryptic digestion of titin can only be used to study whether titin develops restoring force if trypsin degrades the elastic I-band region of titin. Therefore, we performed immunoelectron microscopy and labeled the middle of the elastic region of the titin molecule with the anti-titin antibody 9D10 and the ends of this region with the anti-titin antibodies T12 and Ti-102 (compare with Reference 5). We found that after the trypsin treatment, 9D10 failed to label the cells (Fig 4A⇓), whereas both the T12 and Ti-102 epitopes were still present, although they were not labeled as well as in the control (Fig 4B⇓). We conclude that a mild trypsin treatment degrades the elastic region of the titin molecule, making trypsin treatment a suitable method for studying the role of titin in restoring force development.
Role of Titin in Unloaded Relengthening of Short Myocytes
Suspensions of relaxed cells were brought into rigor by washing out ATP with rigor solution (see “Materials and Methods”). This resulted in sarcomere shortening of ≈20% (Fig 5A⇓). Such shortening has been observed by others,17 who ascribed it to Ca2+-independent crossbridge cycling that occurs at [MgATP] in the range of 0.1 to 100 μmol/L. Thus, as [MgATP] is gradually decreased from that in relaxing solution to that in the rigor solution, shortening will occur as long as the [MgATP] is within this range. We found that Ca2+-independent shortening was completely reversible upon addition of relaxing solution (Fig 5A⇓). Furthermore, by observing single cells that were glued at one end to the bottom of the mechanics chamber, we established that the Ca2+-independent shortening followed by relaxation could be induced repeatedly in the same cell without affecting the length after relaxation (data not shown). Ca2+-independent shortening is ideal for studying titin, because shortening can be induced without the need to add Ca2+, reducing the risk of unwanted titin degradation by Ca2+-dependent proteases.6 14
When short myocytes in rigor were treated with trypsin, the potential to relengthen upon relaxation was reduced (Fig 5B⇑). A 6-minute treatment led to a significant decrease in relengthening (using Dunnett's multiple comparisons procedure with 0-minute trypsin treatment as control; P<.05, one-tailed test; n=20; pooled data from two experiments), whereas a 12-minute treatment resulted in cells that did not relengthen at all. Results were identical when relaxed cells were treated with trypsin, followed by Ca2+-independent shortening and relaxation (data not shown). Thus, the restoring force that relengthens shortened myocytes is abolished by trypsin-induced titin degradation.
Measurement of Restoring and Passive Forces
We adopted the protocol of Fabiato and Fabiato15 and obtained relaxed cells below slack by inducing Ca2+-independent shortening in a buckled cell segment (≈10 sarcomeres long) followed by relaxation (protocol further explained in Fig 1⇑). The relaxed cells, which had a sarcomere length of ≈1.6 μm, did not rebuckle but instead pushed on the force transducer. Thus, this protocol made it possible to obtain relaxed cells below their normal slack length, confirming findings by Fabiato and Fabiato. The cells were subsequently slowly stretched while restoring force (sarcomere length, <1.85 μm) and passive force (sarcomere length, >1.85 μm) were measured.
The restoring force–sarcomere length curve typically had a relatively shallow slope at sarcomere lengths between ≈1.60 and ≈1.70 μm and an approximately linear slope of decreasing force between ≈1.70 and slack (1.85 μm). Above slack, passive force at first increased linearly and then, at lengths above ≈2.0 μm, more exponentially. Around the slack length, restoring and passive forces followed the same linear relationship and could be fit with a straight line. Results from two representative cells are shown in Fig 6⇓. The average results are listed in the Table⇓.
We then studied whether restoring and passive forces were trypsin sensitive. A 12-minute trypsin treatment decreased both the restoring and passive forces by ≈40% and ≈60%, respectively (Fig 7⇓). It is known from previous work14 that complete removal of titin from cardiac myocytes results in an ≈90% reduction of passive force and that the remaining force is derived from intermediate filaments. The measured reduction in passive force after 12-minute trypsin treatment suggests that ≈70% of titin had been degraded, under the conditions that existed in the mechanical experiments. This is in contrast to our gel electrophoresis and relengthening studies (Figs 2 and 5⇑⇑), from which we expected that the 12-minute trypsin treatment would have completely degraded titin and fully abolished restoring force. To test whether the measured residual restoring force resulted from incomplete titin degradation, we repeated the mechanical experiment on cells for which the trypsin treatment was extended beyond 12 minutes. Those experiments were unsuccessful, however, because the cells showed structural damage upon rigor force development. Hence, we restricted our analysis to cells treated with trypsin for 12 minutes, where structural damage was not detected and rigor force was unaltered (Table⇑). The discrepancy between the relengthening and mechanical experiments will be addressed in “Discussion.”
The shape of the restoring force–sarcomere length curve was altered by trypsin. The linear range of the curve extended to shorter lengths than in the control curves (Fig 7⇑). We found that the decrease in restoring force after the trypsin treatment was less than the decrease in passive force (Fig 7⇑, Table). The decrease in restoring force was ≈30% to ≈40% (sarcomere length, 1.70 to 1.75 μm), and the corresponding value for passive force was ≈50% to ≈60% (sarcomere length, 1.95 to 2.00).
We investigated whether titin develops restoring force in rat cardiac myocytes. Restoring force was studied in unattached cells that underwent Ca2+-independent shortening and subsequent relaxation. We found that the ability of cells to relengthen to their slack length was abolished if titin was degraded by trypsin. Restoring force was also directly measured by gluing short segments of cells, containing ≈10 sarcomeres, to a motor and a force transducer and using a protocol adapted from Fabiato and Fabiato.15 At a sarcomere length of 1.70 μm, restoring force was −61±20 μg, and a 12-minute trypsin treatment reduced the absolute value by ≈30% to ≈40%. Below, we discuss these results vis-a`-vis the recently proposed model of restoring force development by titin.5
As found in other studies,12 13 14 a mild trypsin treatment quickly degraded titin without clearly affecting contractile proteins (Figs 2 and 3⇑⇑). It cannot be excluded, however, that minor proteins (desmin, nebulette, etc) are affected by trypsin and develop restoring force but that their proteolysis went undetected because of the limited sensitivity of the gel-based detection system of proteolysis. Another issue is that the involvement of titin in restoring force development can only be addressed using trypsin, if trypsin degrades the elastic segment of the titin molecule. The only previous study addressing this issue is by Yoshioka et al,12 who reported that in skeletal muscle one of the stripes labeled with anti-connectin (titin) antiserum is missing after trypsin treatment and that this stripe is located at the A/I junction. We found that the anti-titin antibody 9D10, which labels the I-band domain of cardiac titin in control cells, fails to label the sarcomere after trypsin treatment (Fig 4⇑). Furthermore, the regions of the titin molecule just outside the elastic domain label weakly with the T12 and Ti-102 anti-titin antibodies (Fig 4⇑). Thus, it is likely that trypsin cleaves both ends of the elastic I-band domain of titin and that this domain (≈700 kD in cardiac muscle10 ) subsequently diffuses out of the myocytes, which would explain thereby its absence from our gels (Fig 2⇑). Diffusion of high-molecular-weight titin fragments from the I band of skeletal muscle myofibrils after tryptic digestion has been reported by Astier et al.18 The titin segment located between the Z line and the T12 epitope that remains in the trypsin-treated sarcomere (Fig 4B⇑) has a mass of ≈200 kD10 and may correspond to the degradation product that is seen on gels just below the MHC (Fig 2⇑). In conclusion, a mild trypsin treatment preferentially degrades the elastic domain of titin in the I band, making it a good tool for studying the contribution of this domain to the mechanical properties of cardiac myocytes.
Titin degradation is also known to occur by endogenous proteases.6 To prevent titin degradation, cells were isolated in solutions that contained protease inhibitors.14 That the presence of these inhibitors is critical for keeping titin intact can be concluded from the control cells shown in Fig 2⇑. In these experiments, protease inhibitors were removed, in order to allow for the subsequent trypsin treatment, by centrifuging the cell suspension and resuspending the pellet of cells in solution without protease inhibitors (see “Materials and Methods”). It is likely that during this ≈100-minute protocol endogenous proteases degraded titin to some degree, as can be concluded from comparing the T1-to-T2 ratio of control cells in Fig 2⇑ with results in Granzier and Irving.14 It is also noteworthy that most of the previous work on restoring force (for review, see Brady4 ) has been carried out without awareness of the existence and sensitivity of titin. Without protease inhibitors in these previous studies, titin degradation is likely to have occurred.
The restoring force decreased by 30% to 40%, and the passive force decreased by ≈60% (Fig 7⇑, Table) after a 12-minute trypsin treatment. Granzier and Irving14 showed that ≈90% of the passive force is titin based, which implies that in our mechanical experiments titin degradation was incomplete. This is in contrast to the relengthening experiments, where, after 12-minute trypsin treatment, titin was completely degraded (Fig 2⇑) and the ability to relengthen was abolished (Fig 5⇑). An explanation may be that when the tryptic digestion started, cells had a different history in the two types of experiment. As for cells attached to the mechanics setup, protease inhibitors were removed by flushing the chamber with inhibitor-free solution, and this was a much faster procedure than the one used for cell suspensions (≈5 minutes versus ≈100 minutes). Thus, cells used for mechanical measurements were exposed to low levels of leupeptin only for a short time; therefore, they were less susceptible to titin degradation before trypsin treatment. This may explain why a subsequent 12-minute trypsin treatment is insufficient to completely degrade titin in the mechanical setup but is sufficient in the relengthening experiments.
A striking feature of the measured restoring force–sarcomere length curve is that at sarcomere lengths shorter than ≈1.7 μm, restoring force was relatively independent of sarcomere length (Fig 6⇑). Others (References 15, 19, and 20 and P.H. Backx and H.E.D.J. ter Keurs, unpublished data, 1989) have reported that the restoring force and stiffness increase steeply when the sarcomere length is decreased below ≈1.7 μm and ascribed this to thick-filament compression. This compressive force is apparently absent in our experiments, indicating that thick filaments were buckled. Thick-filament buckling has indeed been observed on electron micrographs of short sarcomeres (Reference 6 and H. Granzier and K. Trombita´s, unpublished data, 1996). That thick filaments may be able to withstand high compressive forces in activated trabeculae,20 and not in our relaxed myocytes, may be due to crossbridges that during activation cross-link thin and thick filaments, thereby increasing the resistance of thick filaments to buckling.
The restoring force at a sarcomere length of 1.70 μm was, on average, 61 μg (Table⇑). This value corresponds to ≈3.5 mN·mm−2, assuming that the average cross-sectional area of the cells was 170 μm2.14 This result can be compared with the estimate of restoring force in intact trabeculae made by Backx,20 who obtained, at a sarcomere length of 1.70 μm, a value of ≈5 mN·mm−2. Restoring force in the present study could have been underestimated because of viscoelasticity, which manifests itself as a decrease in force with time (stress-relaxation) and in a speed dependence of measured force. In the study of Backx, the restoring force was estimated within seconds after the trabeculae had attained their short length, and restoring force may have decreased by a small amount only. In the present study, however, restoring force was measured after the cell had been at the short length for ≈4 minutes. In preliminary experiments we estimated that at sarcomere lengths close to the slack length, restoring and passive forces undergo <15% of stress relaxation in 4 minutes (M. Helmes and H. Granzier, unpublished data, 1996). This indicates that the viscous component of the restoring force is small and that our restoring force measurements are minimally affected by viscoelasticity. This conclusion is supported by our finding that restoring force and passive force do not significantly vary with the speed of stretch (M. Helmes and H. Granzier, unpublished data, 1996). Restoring force may have been underestimated in the present study, however, because of thick-filament buckling that is likely to occur in relaxed myocytes with sarcomere lengths of ≈1.70 μm and shorter but not in activated cells (see above). In our model of restoring force development by titin (see below and Fig 8⇓), the distance between the end of the anchoring segment of titin (≈100 nm from the Z line) and the end of the thick filament determines the level of restoring force developed by titin. Thick-filament buckling will decrease this distance and lower restoring force accordingly. Despite the differences in the protocols used to measure restoring force, it is clear that this force of myocytes is only slightly less than that of intact trabeculae, indicating that the restoring force of cardiac muscle arises to a large extent from structures within the myocyte.
The studies on the effect of trypsin on measured force revealed that restoring force is less sensitive to trypsin than passive force (Fig 7⇑, Table). On the basis of the trypsin-induced reduction in passive force, we estimated that in our mechanical experiments ≈70% of titin was degraded, and the measured reduction in restoring force of ≈40% suggests that ≈60% of the restoring force is titin based. As for the source of the remaining 40%, candidates are the several nontitin structures that have previously been proposed to develop restoring force, such as intermyofibrillar structures, thin filaments, Z bands, and compressed crossbridges.3 4 15 21 22 It is difficult, however, to explain why nontitin-based restoring force did not relengthen the unattached cells after tryptic digestion of titin (Fig 5B⇑). We believe that it is possible that titin is the sole contributor to the measured restoring force but that our measurements underestimated its contribution to restoring force. As explained above, buckling of thick filaments will result in an underestimation of titin-based restoring force. Our finding that the linear range of the restoring force–sarcomere length curve after trypsin treatment extends to shorter sarcomere lengths (Fig 7⇑) indicates that thick-filament buckling is not as severe after tryptic digestion, and the underestimation of restoring force will therefore be less. Hence, the trypsin sensitivity of restoring force and the titin-based component of restoring force will be underestimated.
During Ca2+ activation at short sarcomere lengths on the ascending limb of the force–sarcomere length relation, restoring force opposes contractile force, and measured active force will therefore be equal to the level of active force that is actually produced minus the restoring force. The restoring force measured in the present study at a sarcomere length of 1.70 μm corresponds to 6% to 10% of the maximal active force produced by rat skinned cardiac myocytes at their optimal length.15 23 24 25 However, the level of restoring force may have been underestimated in the present study because of thick-filament buckling (see above). Thus, restoring force is an important factor to consider in explaining the low levels of measured active force on the ascending limb of the force-length relation of cardiac muscle.3 26 27
Model of Passive Force and Restoring Force Development by Titin
The model is based on previous immunoelectron microscopic studies6 28 in which monoclonal anti-titin antibodies were used that recognize different epitopes in the sarcomere. Results indicated that only a small segment of the titin molecule is elastic and that in slack sarcomeres this segment is highly folded and in an entropically optimal state (Fig 8A⇑). In the model, this elastic segment develops passive force in response to sarcomere stretch in two stages. When slack sarcomeres are stretched, the elastic domain of titin first straightens (Fig 8C⇑), and an entropic force develops. Upon further stretch, the elastic titin segment elongates by unfolding of subdomains, and a conformational force develops (Fig 8D⇑). The elastic segment of human cardiac titin consists of repeating immunoglobulin-like domains, each containing ≈100 residues, and a unique domain of ≈160 residues rich in proline (P), glutamate (E), lysine (K), and valine (V), referred to as the PEVK domain.10 It has been suggested that the PEVK domain may unfold first, because it is likely to be less stable than the immunoglobulin-like domains.10 29 This ≈160-residue domain can elongate to a maximal length of ≈60 nm, assuming a maximal residue spacing of 3.8 Å.30 Thus, PEVK unfolding may account for passive force development when sarcomeres are stretched ≈0.1 μm beyond the length at which the elastic segment is straight (≈2.0 μm), whereas further stretch will result in unfolding of immunoglobulin-like domains. In summary, passive force development in cardiac myocytes is a two-stage process that consists of straightening and unfolding of titin.
In our immunoelectron microscopic studies of sarcomeres below slack length (Reference 5 and K. Trombita´s and H. Granzier, unpublished data, 1996), we found evidence that the elastic segment of titin straightens, similar to what occurs during stretch. (This evidence is based on our observation that the order of titin epitopes reverses around the slack length: compare the order of the schematically indicated epitopes of Fig 8B and 8C⇑⇑.) Therefore, we proposed that sarcomere shortening to lengths below slack results in a titin-based entropic restoring force (Fig 8B⇑). A prediction of the model is that when a certain length change is imposed on the sarcomeres, the degree of titin straightening will be similar when the cell is shortened versus stretched, and the restoring force will have the same magnitude as the passive force. Although at lengths shorter than ≈1.7 μm the magnitude of restoring force is less than that of passive force at lengths longer than ≈2.0 μm, which may be due to thick filament crumpling at such short lengths (see above), we found that close to the slack length, the restoring force-sarcomere length curve is indeed a mirror image of the passive force–sarcomere length relation (Figs 7 and 8⇑⇑), as predicted by the model.
In conclusion, cardiac titin may function as a bidirectional spring that develops passive force in stretched sarcomeres and restoring force in shortened sarcomeres. At the level of the whole heart, part of the actomyosin-based active force that is developed during systole may be harnessed by titin, allowing for elastic diastolic recoil and aiding in ventricular filling.
This study was supported by a Grant-in-Aid of the American Heart Association (Washington State Affiliate, Inc), a grant from the National Institute of Arthritis and Musculoskeletal and Skin Disease (AR-42652), and a grant from the Whitaker Foundation for Biomedical Research. We gratefully acknowledge Dr Jin's generosity in providing us with the anti-titin antibody Ti-102. We express our gratitude to Drs Campbell, Kellermayer, and Slinker for critical reading of various drafts of the manuscript and to Ms Stockman for superb technical assistance.
Reprint requests to Henk Granzier, Department of Veterinary and Comparative Anatomy, Pharmacology, and Physiology, Washington State University, Pullman, WA 99164-6520. E-mail email@example.com.
- Received April 19, 1996.
- Accepted July 12, 1996.
Gilbert JC, Glantz SA. Determinants of left ventricular filling and of the diastolic pressure-volume relation. Circ Res. 1989;64:827-852.
Brady A. Cardiac myocyte mechanics. Physiol Rev. 1991;71:413-424.
Labeit S, Kolmerer B. Titins, giant proteins in charge of muscle ultrastructure and elasticity. Science. 1995;270:293-296.
Yoshioka T, Higuchi H, Kimura S, Ohashi K, Umazume Y, Maruyama K. Effects of mild trypsin treatment on the passive tension generation and connectin splitting in stretched skinned fibers from frog skeletal muscle. Biomed Res. 1986;7:181-186.
Higuchi H. Changes in contractile properties with selective digestion of connectin (titin) in skinned fibers of frog skeletal muscle. J Biochem. 1992;111:291-295.
Fabiato A, Fabiato F. Dependence of calcium release, tension generation and restoring forces on sarcomere length in skinned cardiac cells. Eur J Cardiol. 1976;4(suppl):13-27.
Nichols CG, Lederer WJ. The role of ATP in energy-deprivation contractures in unloaded rat ventricular myocytes. Can J Physiol Pharmacol. 1989;68:183-194.
Astier C, Labbe´ J, Roustan C, Benyamin Y. Effects of different enzymic treatments on the release of titin fragments from rabbit skeletal myofibrils. Biochem J.. 1993;290:731-734.
Iwazumi T. A new method to attach single cardiac myocyte to transducers. Biophys J.. 1987;51:479a. Abstract.
Backx PH. The Force-Sarcomere Length Relationship in Rat Cardiac Muscle. Calgary, Canada: University of Calgary; 1989. Thesis.
Ter Keurs HEDJ, Rijnsburger WH, van Heuningen R. Restoring forces and relaxation of rat cardiac muscle. Eur Heart J. 1980;1(suppl A):67-80.
Goldstein MA, Schroeter JP, Michael LH. Role of the Z-band in the mechanical properties of the heart. FASEB J. 1991;5:2167-2174.
Schweitzer NK, Moss RL. Determinants of loaded shortening velocity in single cardiac myocytes permeabilized with α-hemolysin. Circ Res. 1993;73:1150-1162.
McDonald KS, Moss RL. Osmotic compression of single cardiac myocytes eliminates the reduction in Ca2+ sensitivity of tension at short sarcomere length. Circ Res. 1995;77:199-205.
Palmer RE, Brady AJ, Roos KP. Mechanical measurements from isolated cardiac myocytes using a pipette attachment system. Am J Physiol. 1996;270(Cell Physiol 39):C697-C704.
Kentish JC, ter Keurs HEDJ, Ricciardi L, Bucx JJJ, Noble MIM. Comparison between the sarcomere length-force relations of intact and skinned trabeculae from rat right ventricle: influence of calcium concentrations on these relations. Circ Res. 1986;580:755-768.
Kentish JC, Stienen JM. Differential effects of length on maximum force production and myofibrillar ATPase activity in rat skinned cardiac muscle. J Physiol (Lond). 1994;475.1:175-184.
Trombita´s K, Jin J-P, Granzier H. The mechanically active domain of titin in cardiac muscle. Circ Res. 1995;77:856-861.
Erickson HP. Reversible unfolding of fibronectin type III and immunoglobulin domains provides the structural basis for stretch and elasticity of titin and fibronectin. Proc Natl Acad Sci U S A. 1994;91:10114-10118.