Mechanically Driven Contour-Length Adjustment in Rat Cardiac Titin’s Unique N2B Sequence
Titin Is an Adjustable Spring
Abstract—The giant elastic protein titin is largely responsible for passive forces in cardiac myocytes. A number of different titin isoforms with distinctly different structural elements within their central I-band region are expressed in human myocardium. Their coexpression has so far prevented an understanding of the respective contributions of the isoforms to myocardial elasticity. Using isoform-specific antibodies, we find in the present study that rat myocardium expresses predominantly the small N2B titin isoform, which allows us to characterize the elastic behavior of this isoform. The extensibility and force response of N2B titin were studied by using immunoelectron microscopy and by measuring the passive force–sarcomere length (SL) relation of single rat cardiac myocytes under a variety of mechanical conditions. Experimental results were compared with the predictions of a mechanical model in which the elastic titin segment behaves as two wormlike chains, the tandem immunoglobulin (Ig) segments and the PEVK segment (rich in proline [P], glutamate [E], valine [V], and lysine [K] residues), connected in series. The overall contour length was predicted from the sequence of N2B cardiac titin. According to mechanical measurements, above ≈2.2 μm SL titin’s elastic segment extends beyond its predicted contour length. Immunoelectron microscopy indicates that a prominent source of this contour-length gain is the extension of the unique N2B sequence (located between proximal tandem Ig segment and PEVK), and that Ig domain unfolding is negligible. Thus, the elastic region of N2B cardiac titin consists of three mechanically distinct extensible segments connected in series: the tandem Ig segment, the PEVK segment, and the unique N2B sequence. Rate-dependent and repetitive stretch-release experiments indicate that both the contour-length gain and the recovery from it involve kinetic processes, probably unfolding and refolding within the N2B segment. As a result, the contour length of titin’s extensible segment depends on the rate and magnitude of the preceding mechanical perturbations. The rate of recovery from the length gain is slow, ensuring that the adjusted length is maintained through consecutive cardiac cycles and that hysteresis is minimal. Thus, as a result of the extensible properties of the unique N2B sequence, the I-band region of the N2B cardiac titin isoform functions as a molecular spring that is adjustable.
During diastole, while the ventricle fills, the myocardium is passively stretched. The rate of filling and the magnitude of the end-diastolic volume are significantly affected by the shape of the passive pressure–volume relation. Titin, a striated muscle–specific giant elastic protein, is largely responsible for the generation of the diastolic force in the cardiac myocyte.1 In addition to influencing ventricular filling, titin also helps maintaining the structural integrity of the contracting sarcomere (for reviews, see References 2 through 72 3 4 5 6 7 ).
Titin’s force is derived from its extensible I-band region, which consists of two main segments: (1) a segment rich in proline (P), glutamate (E), valine (V), and lysine (K) residues (the so-called PEVK segment) and (2) the tandem immunoglobulin (Ig) segments (serially linked Ig-like domains) flanking the PEVK segment. Both segments are expressed in muscle type–specific length variants, with human cardiac muscle coexpressing the low molecular weight N2B isoform and larger isoforms containing the N2A element.8 The molecular mechanism of titin’s elasticity has been investigated in dynamic light-scattering studies on titin in solution,9 mechanical studies on single titin molecules,10 11 12 and immunoelectron microscopic studies on skeletal muscle titin.13 14 15 A model has emerged in which the tandem Ig segments (containing folded Ig-like domains) and the PEVK segment (acting largely as an unfolded polypeptide) behave as serially linked wormlike chains (WLCs). A WLC is a deformable rod whose bending rigidity is expressed in terms of its persistence length (A), a distance within which thermally induced bending orientations are correlated.16 17 The more rigid the polymer, the longer its persistence length (low conformational entropy) and the smaller the force required for its extension (see Materials and Methods). Because the bending rigidity of the tandem Ig segments exceeds that of the PEVK,10 12 13 14 upon stretching the sarcomere, the tandem Ig extends first, followed by extension of the PEVK segment.13 14 15 18
The goal of the present study was to investigate (1) the titin isoform expressed in the rat cardiac myocyte, (2) the extensibility of rat cardiac titin’s elastic segment, and (3) whether the recently established molecular properties of titin in vitro can explain the observed mechanical behavior of the passive myocyte. Rat myocardium was shown in the present study to express predominantly the N2B cardiac titin isoform, allowing us to characterize this isoform without confounding effects from the coexpression of other isoforms. The contour lengths of the tandem Ig and PEVK segments were estimated based on sequence data for N2B titin.8 These contour lengths were used in the simulation of the behavior of the elastic titin segment. Comparing the model predictions with the experimental observations in myocytes revealed that the elastic segment could be extended much further than the predicted contour length. Immunoelectron microscopy revealed that Ig domain unfolding is not a prominent process in contour-length gain. Instead, the contour-length gain is due mainly to extension of titin’s unique N2B sequence, located between the proximal tandem Ig segment and the PEVK segment. Our work indicates that N2B cardiac titin contains three mechanically distinct molecular spring segments: (1) tandem Ig segments, (2) the PEVK segment, and (3) the unique N2B sequence.
Materials and Methods
Preparation of Cardiac Myocytes
Care and treatment of animals in the present study were in accordance with protocol 2361 as approved by Washington State University. Single cardiac myocytes were isolated according to the method used by Granzier and Irving.1 Briefly, cells were isolated from rat (male Sprague-Dawley, ≈250 g; Simosen Labs, Gilroy, Calif) heart by perfusing the coronary arteries with oxygenated Krebs solution containing collagenase. The atria were removed and discarded. Ventricles were cut into small pieces, which were repeatedly drawn through a plastic pipette tip to release isolated cells. The majority of myocytes obtained this way had a normal, rod-like shape. The cells were skinned by a 50-minute treatment with 1% Triton X-100 in relaxing solution (in mmol/L: imidazole 40, EGTA 10, magnesium acetate 6.4, sodium ATP 5.9, creatine phosphate 10, potassium propionate 80, DTT 1.0 [pH 7.0 at 21°C]). The detergent was washed out by extensive rinsing with relaxing solution. To prevent degradation, solutions contained protease inhibitors (compare Reference 11 ). Preliminary experiments using intact and skinned trabeculae indicate that the lattice expansion that occurs upon skinning19 does not affect the passive force–sarcomere length (SL) relation (I. Wu, H. Granzier, unpublished observations, 1999), and lattice expansion was therefore ignored in the present study.
Gel Electrophoresis and Western Blotting
Left ventricular myocardium of rat and rabbit and skeletal muscle (human soleus) were frozen in liquid nitrogen, pulverized to a fine powder, and then rapidly solubilized. The samples were analyzed with SDS-PAGE (2% to 12% acrylamide gradient gels) (for details, see Reference 11 ). Western blotting with the anti-titin antibodies specific to the N2A and N2B isoforms was performed as explained in our earlier work.20 Both antibodies are affinity-purified polyclonal antibodies. To raise the N2A antibody (X104–X105), the titin sequence from the N2A splice pathway (base pairs 15607-15957 of human skeletal cDNA entry [EMBL data library accession No. X90569]) was expressed in Escherichia coli. For the N2B antibody (X150-X151), the base pairs 11551 to 11928 of the human cardiac titin cDNA sequence (EMBL data library accession No. X90568) were expressed. This sequence locates within the nonrepetitive unique sequence of the N2B segment (see Figure 1C⇓) that is found in only heart muscle. Protein expression, purification, and antibody production were as described earlier.13 15
The instrumentation and the single-myocyte mechanical protocols used have been described previously.1 Myocytes were added to a temperature-controlled flow-through chamber (volume ≈300 μL) mounted on the stage of a phase-contrast microscope. One end of a single cell was glued to a motor.1 The free end was then bent with a micromanipulator so that the myocyte axis aligned with the microscope optical axis, and the myocyte cross-sectional area was measured. Finally, the free end of the myocyte was glued to a force transducer. Two types of force transducers were used. For experiments with low stretch-release rates (< 1.6 lengths/sec), a model 406A force transducer (Cambridge Technology) was used (resonance frequency ≈75 Hz). To accurately measure force in response to fast stretch-release rates (>1.6 lengths/sec), a model 403A force transducer (Cambridge Technology) was used (resonance frequency ≈400 Hz). To avoid the disturbing effect of meniscus forces, the force transducer was positioned entirely below the solution surface through the side of the chamber. A gravity-based perfusion system was used. Experiments were carried out at 20°C to 22°C.
Using potassium chloride and potassium iodide extraction, we1 previously showed that the force–SL relation of myocytes is largely titin-based with a small amount of force derived from intermediate filaments (IFs). In the present work, extraction experiments revealed that IF-based force at the longest SLs studied was ≈5% of the total force, and that the contribution was less at shorter SLs. Considering the small force contribution of IFs, this component was ignored in the present work.
SL was measured by the method of Granzier and Irving.1 Briefly, phase-contrast images of cells were digitized, 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 ≈30 μm. The traces were processed by discrete Fourier transformation, and the position of the first-order peak in the power spectrum was used to calculate the SL. 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.
Both immunoelectron microscopy (IEM) and immunofluorescence techniques were used. The details of the IEM methods have been published previously.21 22 Briefly, cells were glued in the stretched state to the bottom of a minichamber that was used for immunolabeling, fixing, and embedding of the cells. The time between stretching and fixing of cells varied from 15 to 60 minutes. The anti-titin antibodies T12, 9D10, Ti102, and N2B (X150-X151) were used. (For the source of the T12, 9D10, and Ti102 antibodies, see Reference 2020 .) Their binding sites in the titin sequence are shown in Figure 1C⇑. Immunofluorescence was performed on stretched single rat cardiac myofibrils according to Linke et al.13
Actin Extraction Using Gelsolin
To obtain thin-filament free myocytes, actin was extracted by using a gelsolin fragment. The extraction method is explained in Trombitás and Granzier.23
The elastic region of titin was modeled as two WLCs with different elastic properties in series: the tandem Ig segment and the PEVK segment. For a WLC, the external force (F) is related to the chain’s extension (z), as in References 16 and 1716 17 : where A is the persistence length (a measure of the chain’s bending rigidity), kB is the Boltzmann constant, T is the absolute temperature, and L is the contour length (length of the chain when completely straight). The contour length of the tandem Ig segments was calculated from the number of Ig domains times the mean domain spacing of 5 nm of a native (folded) Ig domain in a completely extended chain (see Trombitás et al15 ). Thus, for the cardiac N2B isoform, the contour length of the first tandem Ig segment is 75 nm and that of the second is 140 nm (for further details, see Reference 1515 ). The persistence length of the native tandem Ig segment was taken as 15 nm.9 The contour length of the PEVK segment was obtained from the number of amino acid residues (163) times the maximal residue spacing in an unfolded polypeptide (0.38 nm). Accordingly, the PEVK segment was assumed to behave as a completely unfolded polypeptide with a contour length of 60 nm. The persistence length of the PEVK segment was taken as 1.3 nm, based on recent single-molecule mechanical measurements on N2B titin.24
Because the tandem Ig and PEVK segments are connected in series, they bear equal forces. Therefore, for a given F, the extension of each segment (zIg and zPEVK) can be calculated. We may then calculate, for that F, the total extension of titin’s elastic segment (zIg+zPEVK). By adding the total length of nonextensible sarcomeric components (1800 nm; see References 20 and 2120 21 ), the SL (zIg+zPEVK+1800 nm) can be calculated.
The extension of a WLC approaches L as F−1/2 approaches zero.10 16 At high extensions, F−1/2 is a linear function of z and extrapolates to the length axis at the contour length. We used the feature of linearity to identify WLC behavior in our myocyte mechanical measurements and to estimate the contour length of titin’s elastic segment in stretched myocytes. Increased contour length beyond the length estimated by the sequence of the elastic segment (see Results) was assumed to arise from conversion of folded domains into unfolded polypeptide. The effect of unfolding on the force versus SL relation was simulated by assuming that the persistence length of the unfolded polypeptide was 1.3 nm (based on our recent single-molecule mechanical experiments on N2B titin24 ) and adding the contour length of the unfolded segment to that of the PEVK segment.
Scaling Measured Myocyte Force to the Single Molecule
To scale the single myocyte measurements down to the single titin molecule, the measured force per myocyte was divided by the cross-sectional area of the cells. Considering that only part of the cross-sectional area contains myofibrils, skinned cells were prepared for electron microscopy, and cross sections were made and photographed. Analysis of the cross sections indicated that 81±9% of the cells (n=5) was taken up by myofibrils. This value was used to convert passive stress per unit area of cell into stress per unit area of myofibril. The number of thick filaments per μm2 myofibril was taken as 540, and the number of titin molecules per half-thick filament as six. For additional details, see Granzier and Irving.1
Composition of Titin’s Elastic Region in Rat Myocardium
It has been shown in human myocardium that the N2B and N2A titin sequences located between the proximal tandem Ig and the PEVK segments are differentially expressed.8 Our SDS-PAGE analysis of rabbit myocardium is consistent with coexpression of two major titin isoforms because it revealed two T1 bands: a higher- and a lower-mobility band (Figure 1A⇑). In contrast, rat myocardium contains a single T1 band, which comigrates with the bottom T1 band of the rabbit (Figure 1A⇑). To test for the identity of the single T1 band in rat, Western blot experiments were performed with the anti-titin antibodies N2B and N2A, raised against unique sequences found in cardiac N2B and N2A isoforms, respectively (Materials and Methods). The rat cardiac titin isoform, corresponding to the lower T1 band of rabbit, was found to strongly react with the N2B but not with the N2A antibody. Although these findings do not exclude expression of a small amount of N2A titin that is below the detection limit of our Western blot and SDS-PAGE methods, it is clear that rat myocardium expresses predominantly the N2B titin isoform.
The composition of the elastic region of rat cardiac titin can be established from the known binding sites in the N2B titin sequence of the T12 and Ti102 antibodies that demarcate the extensible region of titin in rat cardiac myocytes.22 This indicates that the elastic region of the N2B titin isoform contains ≈45 Ig/fibronectin-like domains, the 572-residue unique sequence in the N2B element and a 163-residue PEVK segment (Figure 1C⇑). A schematic diagram showing our vision of how the elastic titin segment is assembled from two WLCs (tandem Ig and PEVK segments) is shown in Figure 1C⇑ (bottom).
Passive Properties of Cardiac Myocytes and Comparison With Simulated WLC Curves
To determine the mechanical properties of titin’s elastic segment in single rat cardiac myocytes, relaxed cells were stretched at a constant rate to a predetermined maximum length. Then the process was reversed to obtain the release half-cycle. During this protocol, the SL and the force (F) were measured. Between consecutive stretch-release cycles, the myocyte was allowed to rest at the slack length (zero force) for 15 minutes. Imposing the rest period ensured reproducibility of the mechanical results. From the F and SL data, F versus SL plots were constructed. From such F versus SL plots (Figure 2A⇓), several initial observations can be made. (1) In the early part of stretch, a nonlinear force response is seen. (2) As the sarcomere is stretched to longer lengths (≈2.2 μm), the force response deviates from the initial nonlinearity and increases almost linearly with increasing SL. (3) It was possible to stretch the sarcomere to lengths greater than that allowed by the contour length of titin’s elastic segment (tandem Ig segments with folded Ig domains plus unfolded PEVK segment, SL ≈2.35 μm). (4) During release, the force response deviates from that seen during stretch. In the initial part of release, force drops rapidly and continues to decrease in a highly nonlinear fashion with decreasing SL. (5) For a given SL, the stretch-force is always greater than the release force. Thus, a force hysteresis is observed. (6) Hysteresis is greater at longer maximum SLs. Hysteresis begins to increase greatly at ≈2.1 μm and continues to increase significantly with further increase in maximum SL (Figure 2A⇓, inset, open symbols). Hysteresis is not affected by extracting thin filaments using gelsolin (see Figure 2A⇓, inset, closed symbols).
According to the analysis by Kellermayer et al,10 the single titin molecule may be modeled as a WLC. The WLC theory can be more reliably and reproducibly applied to the release curves than to the stretch curves. The release data obtained in the present study on cardiac myocytes can indeed be well fit with WLC curves (Figure 3B⇓ and below). However, when the force response of myocytes measured during stretch is scaled down to the single titin molecule level, significant deviations from the prediction based on the WLC model of titin’s elastic segment can be seen, both at SLs less than and greater than ≈2.2 μm. Below ≈2.2 μm SL, the measured force exceeds the predicted one. In contrast, above ≈2.2 μm SL, the measured force is much less than predicted (Figure 2B⇑).
The lower than predicted forces at SLs above 2.2 μm could conceivably be caused by an increase in titin’s contour length due to structural transitions (eg, unfolding) within titin’s elastic segment. The increased contour length may be reflected in characteristics of the release curve. Previous work on single titin molecules10 showed that at the beginning of release, a reversible WLC curve is seen with a contour length that reflects domain unfolding that took place during the preceding stretch. Refolding of these domains occurs only as the force is sufficiently lowered by shortening the molecule.10 The reversibility of the WLC curve at the beginning of the release indicates that the contour length stays constant, and its adjustment (increase/decrease) does not take place. Reversibility can be tested in partial release–restretch experiments. Accordingly, we tested whether contour-length adjustment is absent during the initial part of the release of cardiac myocytes, by partially releasing the stretched myocytes followed by restretching.
The force measured during restretching the myocyte after partial release retraced the force curve obtained during release, ie, hysteresis is absent (Figure 3A⇑). The reversibility of the force curve during this initial part of the release indicates that the structural constraints that were broken during stretch and that gave rise to contour-length adjustment did not reform during the partial release. This observation supports previous suggestions that for refolding to occur in the single titin molecule, the external force must be decreased to low levels; before that, refolding is precluded.10 Whether the extensible titin segment during the initial part of the release behaves as a WLC was tested by plotting F−1/2 versus SL for the release curve. The F−1/2 versus SL curve obtained during release reveals WLC behavior, because the experimental data linearly approached the SL axis (Figure 3B⇑). Linear extrapolation to the length axis gives the SL at which the contour length of titin’s elastic segment is reached. Subtracting from the extrapolated SL the nonextensible sarcomeric components (1800 nm) and dividing the result by two (to account for the two half sarcomeres) will yield the contour length of titin’s elastic segment. The contour length thus obtained increased with increasing maximum experimental SL (Figure 3C⇑, left-hand scale).
To estimate the effect of this contour-length gain on the simulated F−SL curve, we assumed that the adjustment results from an increase in the length of titin’s unfolded region and determined the predicted F–SL relation of the WLC model accordingly (for details, see Materials and Methods). Results were compared with the average measured stretch force scaled down to the single-molecule level. Findings indicate that when contour-length gain is taken into account, predicted and measured forces at SLs >2.2 μm agree more closely (Figure 2B⇑, broken and thin solid lines, respectively).
In summary, our data suggest that the elastic region of titin behaves as a WLC with a contour length that increases with the amplitude of stretch. The length gain of the elastic segment may be explained by a progressive increase in the length of the unfolded segment. Increasing the length of the unfolded segment allows the sarcomeres to be stretched to long lengths while avoiding the extremely steep increase of force expected if contour-length adjustment were to be absent.
Time Course of Contour-Length Recovery
Absence of hysteresis in partial release–restretch experiments (Figure 3A⇑) suggests that contour-length recovery does not take place at high force. To investigate the process of contour-length recovery at low force, we determined the rate of hysteresis recovery during successive stretch-release cycles. The myocyte was stretched from its slack length to a predetermined amplitude (see inset of Figure 4A⇓); the cell was then released to the slack length in ≈2 ms, and, after a pause of preset duration, the slow stretch/rapid release was repeated. Figure 4A⇓ shows an example in which a cell was stretched to a SL of ≈2.4 μm, an amplitude that is likely to result in contour-length adjustment, and in which the pause durations were 0, 12, and 200 seconds. Recovery occurs slowly and extends into the pause phase of the protocol. If recovery is completed during the pause, then hysteresis is recovered, and the successive stretch curve retraces the previous one. If, however, recovery is incomplete, then hysteresis is not recovered completely, and the second stretch curve lies below the first. The examples in Figure 4A⇓ show that recovery is completed only after a 200-second pause. The time course of hysteresis recovery can be fitted with a logarithmic function, an example of which is shown in Figure 4B⇓. Experiments were also conducted in which the amplitude of SL stretch was varied. This indicated that the time for complete recovery increases with increasing amplitude of stretch. For SL amplitudes below ≈2.0 μm, force recovery was completed within several seconds, but for longer SLs, the time needed for 90% recovery ranged from 100 seconds at SL 2.15 μm to ≈400 seconds at SL 2.42 μm (Figure 4C⇓).
The effect of slow hysteresis recovery on the F versus SL relation during repeated stretch-release protocols was investigated by subjecting the myocyte to stretch-release cycles with no pause between the individual cycles. Repeated stretch-release cycles did not have a significant effect on the F versus SL curve during release. However, a dramatic effect of repeated stretch-release cycles was the collapse of the F−SL curve during stretch onto the release curve, resulting in a large reduction of force hysteresis (Figure 5A⇓). The hysteresis recovered after a 20-minute rest period at the slack length (Figure 5A⇓). Repeated stretch-release cycles modestly reduced peak forces at SLs longer than ≈2.2 μm (Figure 5B⇓) with a time course that was similar to that of the hysteresis decrease (Figure 5C⇓). Hysteresis (determined as the area under the stretch curve minus the area under the release curve) decreased significantly with the number of stretch-release cycles, with ≈50% reduction in 5 cycles (Figure 5C⇓).
In summary, after a stretch release to a long SL, hysteresis recovery is a slow process, and, so is, by inference, contour-length recovery. Slow contour-length recovery minimizes hysteresis in repeated stretch-release cycles.
Effect of Stretch Rate on Contour-Length Gain
Single-molecule mechanical studies on titin revealed that increasing the stretch rate reduces the contour-length length gain through unfolding, as a result of the increase in the force at which unfolding takes place with an increase in stretch rate.11 Conceivably, contour-length gain in the cardiac myocyte is similarly reduced by increasing the stretch rate. If indeed the length gain by the elastic segment is reduced under these conditions, a given SL can be reached only at a greater fractional extension of the elastic segment, hence at a greater force. The relationship between force and stretch rate was studied by subjecting the myocyte to a series of stretch-release cycles with increasing stretch rates. Figure 6A⇓ shows F versus SL stretch curves obtained at different rates of stretch, spanning a 250-fold range (0.03 to 8 μm per sarcomere per second). The relationship between force at the maximum experimental SL and the logarithm of stretch rate is shown in Figure 6B⇓. The force for a given SL increases with stretch rate. The effect of stretch rate becomes more pronounced as the SL exceeds ≈2.1 μm. For example, increasing stretch rate ≈250-fold at a SL of 2.45 μm increased the force by ≈50% whereas doing so at a SL of 2.05 μm increased force by only 10%. Plotting the corresponding release curves of Figure 6A⇓ as F−1/2 versus SL showed that at a SL of 2.45 μm, increasing the stretch rate ≈250-fold reduces the contour length of the extensible titin segment by ≈30 nm. The effect of the ≈30-nm reduction in contour length was calculated, from the WLC theory, to yield ≈45% increase in force—a value similar to the measured value.
In summary, the stretch force of the myocyte at a given SL increases with the stretch rate in a logarithmic manner. The increase in force with increase in stretch rate results from a reduction in contour-length gain of titin’s elastic segment.
Sequence Within Titin Underlying Contour-Length Adjustment
To determine where along the elastic segment of titin contour-length gain takes place, IEM was used. Using the antibodies T12 and Ti102 (see Figure 1C⇑), it was shown previously that at SLs <3.0 μm, rat cardiac titin’s elastic segment is restricted to the sequence demarcated by the T12 and Ti102 epitopes.20 Thus, contour-length gain does not result from recruiting titin from the A-band or from near the Z-line to titin’s extensible segment; instead it results from within the segment demarcated by the T12 to Ti102 epitopes. It is unlikely that the PEVK segment is a major source of contour-length gain because studies with the anti-titin antibody 9D10, which labels the full length of this segment, indicate that the PEVK does not extend much at SLs beyond ≈2.3 μm,25 whereas large contour-length gain takes place at these SLs (Figure 3C⇑). To explore the site of contour-length adjustment, the T12, 9D10, and Ti102 antibodies were used to measure the extension of (1) the proximal tandem Ig segment plus the unique N2B element (T12 to edge of 9D10 labeled region) and (2) the distal tandem Ig segment (edge of 9D10 labeled region to Ti102), both as a function of SL. The predicted extension of the tandem Ig segments was also calculated using the serially linked WLC model (Figure 1C⇑; see Materials and Methods).
The predicted extension of the proximal tandem Ig segment was close to the measured T12-9D10 distance at SLs from ≈1.8 to 2.1 μm, whereas at longer SLs, the measured extensions exceeded the predicted extension (Figure 7B⇓). For example, at a SL of 2.9 μm, the measured length was 340±44 nm (n=13), or 265 nm longer than the predicted maximal length of the proximal tandem Ig segment of 75 nm. The predicted extension of the distal tandem Ig segment tracked reasonably well with the measured extensions at SLs between ≈1.8 to ≈2.4 μm, although with further stretch, the segment extended slightly beyond the predicted maximal contour length of 140 nm (Figure 7C⇓).
The results of Figure 7⇑ indicate that contour-length adjustment takes place between T12 and 9D10 epitopes, a region that includes the proximal tandem Ig segment and the N2B element (Figure 1C⇑). To determine which of these segments gives rise to contour-length adjustment, experiments were performed with the N2B antibody (X150-X151; see Materials and Methods), which was raised against the unique N2B sequence located just outside the proximal tandem Ig segment (Figure 8C⇓, inset). Examples of IEM images of rat cardiac myocytes labeled with the N2B antibody are shown in Figure 8A⇓. The antibody labels a well-defined epitope in the I-band. The distance between this epitope and the T12 epitope was measured in sarcomeres stretched to different lengths (Figure 8C⇓, red symbols). Initially, the segment tracked the predicted length of the proximal tandem Ig segment (Figure 8C⇓, curved line); at SLs >2.1 μm, it exceeded the predicted length, and a plateau was attained in sarcomeres longer than ≈2.3 μm (Figure 8C⇓, Table⇓). The plateau value (≈85 nm) is slightly larger than the predicted contour length of the proximal tandem Ig segment (75 nm). Immunofluorescence on single myofibrils labeled with N2B and 9D10 (Figure 8B⇓ and 8C⇓) confirmed and extended the IEM results.
The segment between the N2B epitope and the edge of the 9D10 labeled region contains a large part of the unique N2B sequence as well as two Ig domains, I18 and I19 (Figure 8C⇑, inset). The length of this segment can be obtained from the T12-9D10 length measurements (Figure 8C⇑, open circles) minus the T12-N2B length measurements (Figure 8C⇑, red symbols). The obtained values indicate that this segment has a near zero length at SLs <2.2 μm, whereas at longer SLs, it extends (Figure 8C⇑, Table⇑). It is unlikely that this extension results from Ig domain unfolding (I18 and I19) because superimposing results with an antibody against I18 (from Gautel et al26 ) with 9D10 data indicates that the distance between I18 and 9D10 does not increase with SL (Figure 8C⇑, compare open circles and closed black circles; see also legend to Figure 8⇑). Instead, our findings provide evidence that the unique N2B sequence is extensible. Thus, a major source of contour-length gain in titin’s elastic segment is derived from the unique N2B sequence.
In the beating heart, the myocytes, hence the sarcomeres within them, are subjected to repeated stretch-release cycles. During such a cycle, the passive force developed in the sarcomere changes: (1) during stretch (diastole), passive force increases because sarcomeric elastic elements are extended; (2) during release (systole), passive force decreases because the stretched elastic elements relax. The shape of the passive F–SL relation, which is determined by the properties of these elastic elements, is expected to influence ventricular filling during diastole and ventricular emptying during systole. It has previously been demonstrated that titin is the major sarcomeric elastic element that contributes to passive force of the cardiac sarcomere, with, as a minor source, the intermediate filaments.1 With novel insights about the mechanical features of the single, isolated titin molecule in hand,10 11 12 we explored the mechanical behavior of titin in situ, in the cardiac sarcomere, by subjecting the single isolated myocyte to a wide range of stretch-release protocols. We found that titin’s elastic segment gains length during stretch, as a result of extension of the unique N2B sequence, and that both the rate and magnitude of the preceding mechanical perturbation significantly influence the shape of the passive F−SL relation.
Composition of Titin’s Extensible Segment in Rat Cardiac Myocytes
In human myocardium, the N2B and N2A titin sequences located between the proximal tandem Ig and the PEVK segments are differentially expressed.8 In contrast, our results (Figure 1B⇑) indicate that rat myocardium expresses predominantly the N2B titin isoform. The composition and contour length of the extensible region of rat cardiac titin can thus be established from the known binding sites in the N2B titin sequence of the T12 and Ti102 antibodies that demarcate the extensible region of titin in rat cardiac myocytes.22 In N2B titin, this region contains tandem Ig segments with a combined contour length of ≈215 nm (assuming all Ig domains are folded) and a PEVK segment (assuming to be completely unfolded) with a contour length of ≈60 nm (Figure 1B⇑ and 1C⇑). In addition, to tandem Ig and PEVK segments, the extensible segment of N2B titin also contains a unique sequence of 572 residues located between the proximal tandem Ig segment and the PEVK segment (see Figure 1B⇑). Our findings indicate that this sequence is extensible as well (Figure 8C⇑). Thus, the extensible segment of rat cardiac titin contains three types of subsegments: the tandem Ig segments, the PEVK segment, and the unique N2B sequence.
Does Contour-Length Gain Result From Ig Domain Unfolding?
According to our mechanical measurements, above ≈2.2 μm SL titin’s elastic segment extends beyond the combined contour length of tandem Ig and PEVK segments (Figure 3C⇑). This contour-length gain does not result from the unfolding of Ig domains in the distal tandem Ig segment because this segment extends little at SLs >≈2.2 μm (Figure 7C⇑). This conclusion is consistent with biophysical studies of genetically engineered Ig domains from the distal tandem Ig segment27 28 that found that these domains are very stable. Contour-length gain arises instead from the region between the T12 and 9D10 epitopes (Figure 7B⇑), a region that includes Ig domains (proximal tandem Ig segment) and the unique sequences of the N2B element. The proximal tandem Ig segment is contained in the region demarcated by T12 and N2B epitopes (Figure 1B⇑). The distance between these epitopes extends beyond the predicted contour length of the proximal tandem Ig segment by ≈10 nm (Figure 8C⇑), which may be explained by unfolding of a just a single Ig domain. However, it is also possible that the longer than expected T12-N2B distance results from the short unique sequence contained within this region (Figure 8C⇑, inset). Future research with antibodies that more precisely demarcate the proximal tandem Ig segment will be required to resolve this issue. It is clear from the present work, however, that Ig domain unfolding is, at best, of limited significance in explaining the large contour-length adjustment that takes place in titin at SLs >2.2 μm. Although the titin elasticity model based on Ig domain unfolding, proposed by Erickson,29 is attractive and has had great impact, our studies of the largest titin isoform in skeletal muscle (human soleus titin; Trombitás et al15 ) and the smallest isoform in cardiac muscle (the present study) show that large-scale Ig domain unfolding is unlikely under physiological conditions.
The Unique N2B Sequence as Source of Contour-Length Gain
Results indicate that a prominent site of contour-length adjustment in rat cardiac myocytes is the unique N2B sequence. The unique N2B segment C-terminal of the N2B epitope contains between 310 and 436 residues (depending on where the N2B epitope is located within the 126-residue-long unique N2B fragment that was used to raise the antibody), and our results indicate that this segment has a short end-to-end length at SLs <≈2.1 μm (Figure 8C⇑). It is likely therefore that in absence of external force, the unique sequence has a compact structure. Our findings that this 310- to 436-residue unique sequence extends to ≈70 nm at 2.4 μm SL in rat myocytes (Figure 8C⇑, Table⇑) and the preliminary observations in other species where extensions of 150 nm were attained (K. Trombitás, S. Labeit, H. Granzier, unpublished data, 1999) suggest that average residue spacings can be achieved that are only compatible with a fully unfolded polypeptide. Thus, it is likely that when the sarcomere is extended beyond a length of ≈2.2 μm, the unique N2B sequence unfolds from a compact state to ultimately a completely unfolded and extended polypeptide.
Unfolding-based extension of the unique N2B sequence is consistent with the mechanical behavior of the myocyte. We found that increasing the stretch rate increased the force at the maximum experimental SL, a phenomenon that is most pronounced at long SLs where extension of the unique N2B sequence takes place (Figure 6⇑). This stretch-rate dependence of force indicates that N2B extension in the cardiac myocyte involves a rate-limited process. Previous observations on single titin molecules have shown that the unfolding within titin is a kinetic process that gives rise to a linear relation between the unfolding force (the force at which domain unfolding takes place in titin) and the logarithm of stretch rate.11 Because we also find a linear relationship between peak force and logarithm of stretch rate in isolated myocytes (Figure 6B⇑), it is likely that unfolding within the unique N2B sequence is responsible for the length gain of titin’s elastic segment.
During stretch and release, the cardiac myocytes display force hysteresis: for a given SL, the force measured during stretch is greater than the force measured during release. The presence of hysteresis in thin filament–extracted myocytes (Figure 2A⇑, inset) suggests that titin–thin filament interaction does not greatly contribute to force hysteresis; rather, the hysteresis arises from within titin itself. Hysteresis begins to increase significantly as the SL exceeds ≈2.1 μm and continues to increase with increasing SL (Figure 2A⇑). The onset of significant hysteresis is close to the onset of extension of the unique N2B sequence, suggesting that hysteresis depends on the kinetics of unfolding/refolding within the unique N2B sequence.
The reproducibility of the stretch curve after a sufficient pause indicates that unfolding is a completely reversible process in the myocyte (Figures 2A⇑ and 4⇑). However, refolding takes a relatively long time to be completed. The time course of refolding is reflected in the force recovery curves of Figure 4⇑. The time course can be fit with a logarithmic function, indicating that most of the refolding is rapid but that complete refolding occurs more slowly (Figure 4A⇑ and 4B⇑). Because complete refolding is a slow process, part of the N2B sequence is still in the unfolded state at the completion of the stretch-release cycle. If a new stretch-release cycle is initiated immediately, the N2B sequence will contain preunfolded regions. Therefore, the force generated in the immediately following stretch is less then in the previous one, bringing the stretch curve closer to the release curve, and reducing hysteresis (Figure 5⇑).
As a result of unfolding of the unique N2B sequence, sarcomeres can be stretched to long lengths, whereas slow refolding during release gives rise to strong hysteresis in the first stretch-release cycle, but it minimizes hysteresis in subsequent cycles (Figure 5⇑). If parts of the extensible region of titin unfold and refold each time the myocyte is stretched and released, an amount of energy equal to the area within the hysteresis loop would be wasted as heat. Minimizing passive force hysteresis in cells that are stretch-released repetitively (Figure 5C⇑), as is the case in the beating heart, may be physiologically significant.
Cardiac Titin as an Adjustable Spring
It is well established that tandem Ig and PEVK segments extend sequentially13 14 15 : extension is first dominated by straightening of the tandem Ig segments and at longer SLs by that of the PEVK polypeptide. Sequential extension can be simulated well by modeling the PEVK and tandem Ig segments as serially linked WLCs with different persistence lengths (bending rigidities).15 However, when slack cardiac myocytes are stretched, developed force is initially slightly higher than predicted by the serially linked WLC model (Figure 2B⇑). We propose that this results from weak noncovalent interactions between different regions of titin (see thin black lines in Figure 9⇓) that form when titin is in a “contracted” state in slack sarcomeres. These interactions are disrupted during stretch and reform only slowly during release, giving rise to hysteresis at short SLs. Furthermore, when cells are stretched to SLs >≈2.2 μm, contour-length gain takes place within titin’s extensible region, as a result of extension of the unique N2B sequence (Figure 9⇓ “adjustable spring segment”), decreasing the fractional extension of tandem Ig and PEVK segments and explaining why force at SLs longer than 2.2 μm is much lower than predicted by the serially linked WLC model. Indeed, when taking into account contour-length gain, the force–SL relation predicted by the model more closely simulates measured forces (Figure 2B⇑). Thus, the unique N2B sequence provides cardiac N2B titin with a third source of extensibility (in addition to straightening of the tandem Ig segments with folded Ig domains and extension of the largely unfolded PEVK segment) that allows sarcomeres to be stretched to long lengths.
Contour-length gain in the unique N2B sequence during stretch and slow recovery during release represent two levels of adjustment within titin that determine the shape of the F−SL curve, as well as the amount of hysteresis in repeated stretch-release cycles. Thus, N2B titin may be viewed as an adjustable spring. This property allows N2B titin to accommodate long SLs and to adjust titin’s efficient working range with minimal energy loss due to hysteresis. Considering that human myocardium coexpresses length variants of titin,8 titin’s adjustable spring property may also allow the efficient working range of isoforms to be adjusted to each other.
This work was supported by grants from the National Institutes of Health National Heart, Lung, and Blood Institute (HL61497 and HL62881) to H.G., the Human Frontier Science Program (to S.L.), the Fortschungsfond fur Klinische Medizin Mannheim (to S.L.), the Deutsche Fortschungsgemeineschaft (La668/5-1 to S.L. and SFB 320 to S.L. and W.A.L.), and the Hungarian Science Foundation, OTKA F025353 (to M.K.). H.G. is an established investigator of the American Heart Association.
- Received February 9, 1999.
- Accepted April 19, 1999.
- © 1999 American Heart Association, Inc.
Labeit S, Kolmerer B, Linke WA. The giant protein titin. Emerging roles in physiology and pathophysiology [see comments]. Circ Res. 1997;80:290–294.
Trinick J. Titin as a scaffold and spring. Cytoskeleton Curr Biol. 1996;6:258–260.
Maruyama K. Connectin/titin, giant elastic protein of muscle. FASEB J. 1997;11:341–345.
Labeit S, Kolmerer B. Titins: giant proteins in charge of muscle ultrastructure and elasticity [see comments]. Science. 1995;270:293–296.
Kellermayer MS, Smith SB, Granzier HL, Bustamante C. Folding-unfolding transitions in single titin molecules characterized with laser tweezers [see comments]; [published erratum appears in Science. 1997;277:1117]. Science. 1997;276:1112–1116.
Rief M, Gautel M, Oesterhelt F, Fernandez JM, Gaub HE. Reversible unfolding of individual titin immunoglobulin domains by AFM [see comments]. Science. 1997;276:1109–1112.
Linke WA, Ivemeyer M, Mundel P, Stockmeier MR, Kolmerer B. Nature of PEVK-titin elasticity in skeletal muscle. Proc Natl Acad Sci U S A. 1998;95:8052–8057.
Linke WA, Stockmeier MR, Ivemeyer M, Hosser H, Mundel P. Characterizing titin’s I-band Ig domain region as an entropic spring. J Cell Sci. 1998;111:1567–1574.
Trombitás K, Greaser M, Labeit S, Jin JP, Kellermayer M, Helmes M, Granzier H. Titin extensibility in situ: entropic elasticity of permanently folded and permanently unfolded molecular segments. J Cell Biol. 1998;140:853–859.
Bustamante C, Marko JF, Siggia ED, Smith S. Entropic elasticity of λ-phage DNA. Science. 1994;265:1599–1600.
Trombitás K, Jin JP, Granzier H. The mechanically active domain of titin in cardiac muscle. Circ Res. 1995;77:856–861.
Trombitás K, Granzier H. Actin removal from cardiac myocytes shows that near Z line titin attaches to actin while under tension. Am J Physiol. 1997;273:C662–C670.
Kellermayer M, Smith S, Bustamante C, Granzier H. A molecular segment in titin is an adjustable spring. Biophys J. 1999;76:A9. Abstract.
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.