The cellular mechanism underlying the Frank-Starling law of the heart is myofilament length-dependent activation. The mechanism(s) whereby sarcomeres detect changes in length and translate this into increased sensitivity to activating calcium has been elusive. Small-angle X-ray diffraction studies have revealed that the intact myofilament lattice undergoes numerous structural changes upon an increase in sarcomere length (SL): lattice spacing and the I1,1/I1,0 intensity ratio decreases, whereas the M3 meridional reflection intensity (IM3) increases, concomitant with increases in diastolic and systolic force. Using a short (∼10 ms) X-ray exposure just before electrical stimulation, we were able to obtain detailed structural information regarding the effects of external osmotic compression (with mannitol) and obtain SL on thin intact electrically stimulated isolated rat right ventricular trabeculae. We show that over the same incremental increases in SL, the relative changes in systolic force track more closely to the relative changes in myosin head orientation (as reported by IM3) than to the relative changes in lattice spacing. We conclude that myosin head orientation before activation determines myocardial sarcomere activation levels and that this may be the dominant mechanism for length-dependent activation.
- length-dependent activation
- myofilament lattice spacing
the frank-starling law of the heart describes the relationship between end-diastolic volume and cardiac ejection volume that is the major regulatory system operating on a beat-to-beat basis in the adult heart. The main cellular mechanism that underlies this phenomenon is an increase in the responsiveness of cardiac myofilaments to activating calcium ions at longer sarcomere lengths (SLs). The fundamental mechanism responsible for this increase in calcium sensitivity has been elusive despite considerable experimental scrutiny by various research groups (1, 7, 15, 18, 29).
For the better part of the last twenty years, the prevailing explanation for the increases in peak systolic force with increasing SL was that it was due to a reduction in myofilament spacing (15, 30, 39, 43, 44), i.e., the “lattice spacing hypothesis.” It has been demonstrated many times that an increase in the diastolic or relaxed SL induces a decrease in the spacing between the filaments within the lattice (4, 7, 17, 18, 21, 24, 26); for a thorough review of factors affecting lattice spacing, see Millman (32). With a decreasing myofilament lattice spacing, the probability of forming strong binding cross bridges might be expected to increase, thereby increasing the amount of force production for the same amount of activating calcium (7, 16, 31, 40). This hypothesis was further supported by experiments by Fitzsimons and Moss (14) where they demonstrated that the introduction of nonforce-generating strong binding cross bridges (NEM-S1) was able to shift the calcium sensitivity of cardiac myocytes at short SLs (1.90 μm) to that of myocytes at long SLs (2.25 μm). These data suggest that the attachment of strong binding cross bridges at longer SLs could be a key contributor to the sensitivity of the myofilament lattice to calcium and that the changes in SL could aid in the formation of these attachments. However, recent evidence suggests that rigor cross bridges (similar to those made by NEM-S1) exert a structural impact on titin that is different from that induced by cycling cross bridges (5, 41). Furthermore, we (7) recently showed that blebbistatin, a myosin II ATPase inhibitor that reduces the number of strong binding cross bridges, although it shifted the force/calcium relationship to the right, it did not alter the degree of length-dependent activation (ΔEC50). These results indicate that the increase in calcium sensitivity with increasing SL may not be due to the enhancement of strong binding cross bridges with decreases in lattice spacing but likely involves some other mechanism(s).
Previous work reported from our laboratory (13, 27) suggested that the distance between the thick and thin filaments alone is not sufficient to explain changes in calcium sensitivity under all conditions, indicating that the primary causes of these changes may be due to some structural phenomenon other than lattice spacing. For example, we have shown that the orientation or proximity of the myosin heads to the thin filament, as indicated by the I1,1/I1,0 intensity ratio from equatorial X-ray diffraction patterns [Farman et al. (13)] is better correlated with changes in force and calcium sensitivity than is lattice spacing when skinned muscle is osmotically compressed. Furthermore, we also determined that the attachment of myosin heads to the thin filament in intact muscle does not alter lattice spacing when SL is held constant during the contraction cycle (11). The failure of interfilament lattice spacing to provide a unifying hypothesis, valid under all conditions, for predicting calcium sensitivity motivated the search for some other myofilament lattice structural parameter that varies in a predictable manner when SL is changed and, furthermore, that correlates with changes in the peak systolic force development.
Here we present results from simultaneous small-angle X-ray diffraction and mechanics experiments using intact, electrically stimulated myocardium. Our experiments allowed the correlation of the peak-developed force as a function of SL to 1) lattice spacing (as calculated from the distance between the equatorial reflections), 2) proximity of myosin heads to the thin filaments (as assessed by the equatorial intensity ratio I1,1/I1,0), and 3) orientation of myosin heads perpendicular to the long axis of the thick filament (as assessed by the intensity of the M3 meridional reflection, IM3). We demonstrate, for the first time, that increased orientational ordering of the myosin heads, as reflected in IM3, is a better predictor of peak systolic force development than lattice spacing under these conditions.
MATERIALS AND METHODS
Muscle preparation and solutions.
All experiments were performed according to institutional guidelines concerning the care and use of experimental animals, and all protocols were approved by the Institutional Animal Care and Use Committee of the University of Illinois at Chicago. Male rats (LBNF-1) of ∼275–350 g received intraperitoneal injections of 50 mg/kg pentobarbital sodium and 1.5 ml heparin (13). While the rats were under deep anesthesia, the heart was excised, transferred to a dissection dish, and perfused retrograde with a modified Krebs-Henseleit solution containing (in mM) 118.5 NaCl, 5 KCl, 2 NaH2PO4, 1.2 MgSO4, 10 glucose, 26.4 NaHCO3, and 0.2 CaCl2, as well as 20 2,3-butanedione monoxime to inhibit spontaneous contractions. Only single, isolated, unbranched trabeculae located along the exterior wall connecting the atrioventricular valve to the right ventricular wall were used for experiments.
Mechanics and X-ray diffraction.
X-ray diffraction experiments were performed using the small-angle instrument at the BioCAT X-ray beamline 18ID at the Advanced Photon Source (Argonne, IL). An X-ray wavelength of 0.1033 nm with a specimen to detect distance of ∼2.8 m was used; the incident flux in the full beam was ∼1013 photons/s but was reduced to ∼1012 photons/s using aluminum attenuators to reduce radiation damage. X-ray patterns were collected on a 160 × 80 mm active area Aviex PCCD 16080 (Dexela, London, UK) high-sensitivity charge-coupled device (CCD) detector with 39-μm pixels. Muscles were mounted on a custom X-ray diffraction/muscle mechanics setup designed to fit into the X-ray diffraction instrument. Muscles were attached on the valvular side to a high-speed motor (Cambridge model 308B, ∼1 ms 90% step response) via a hook through the valve end of the muscle. The ventricular end of the muscle was held within a small basket (8), which was mounted on a small metal tube fastened to a KG64 Gueth force transducer, (SI Heidelberg, 0.1–10 mN working range). Both the motor arm and force transducer were attached to three axis positioners that allowed the muscle to be placed inside a 4-mm-wide chamber with thin mylar windows on either side to allow for X-rays to pass through. Muscles were stimulated by field stimulation at 50% above threshold (2 ms, 1 Hz). Samples were maintained between 25 to 26°C for all experiments.
SL was determined using the first order diffraction band from a He-Ne laser captured on a linear CCD with a 2-kHz refresh rate (model PLIN-2605-2, Dexela). Once the muscle was mounted between the force transducer and motor arm, the muscle was stretched to an initial SL of ∼2.2 μm and calcium concentration in the bath was increased to 1 mM. Twitch force was allowed to stabilize for 20–30 min for each experimental condition. Force, length, and X-ray exposure data were collected on a PC running LabView version 8 using custom-designed software. In all experiments the muscles were exposed to a 10-ms X-ray pulse once every 10 s ∼10–15 ms before electrical stimulation. Both the equatorial and meridional patterns were converted to one-dimensional projections using the projection tool in the program FIT2D (19). The intensity data as a function of pixel number was saved as an ASCII format file for later analysis. The peak intensity, widths, and peak separations for the 1,0 and 2,0 equatorial reflections were estimated using a nonlinear least squares fitting procedure as previously described (23). The spacings and intensities of the meridional M3 reflection were fit using the multiple peak fitting routines within the program FIT2D (19). The background was estimated as a third order polynomial, whereas the peak positions, widths, and intensities were estimated assuming a Gaussian peak shape, with changes in the M3 reflection intensity (IM3) used as an indicator of changes in axial cross-bridge orientation. Both structural and mechanical results were expressed as a function of the proportional change in SL to reduce the variability of the measured parameters. Analyses of data were done using Prism 5 (Graphpad Software), and significance was determined at P < 0.05. To assess the impact of a change in sarcomere structure independently from changes in SL, the muscle was returned to the starting SL and new Krebs-Henseleit solution was added containing 1 or 4% mannitol to provide an external osmotic force causing the filament lattice to shrink. This was done to induce a change in the lattice spacing independent of SL changes so as to allow a separation of the impact of the radial position of the myosin heads as well as their orientational ordering from the movement of the myofilament lattice during changes in SL. Peak systolic force was again allowed to stabilize, and X-ray diffraction patterns were obtained as a function of SL as previously described.
We used X-ray diffraction of intact isolated myocardium to assess whether the orientation of the myosin heads with respect to the thick-filament backbone correlates with twitch force production and whether these parameters correlate with changes in interfilament spacing. In these experiments, external osmotic compression was used to alter the native SL-based lattice spacing, allowing us to distinguish the impact of SL changes on the myofilament spacing from those on myosin head position/orientation.
Effects of SL changes on intact myocardial twitch force and structure.
Figure 1 shows the impact of SL changes on twitch force (Fig. 1A) and structural parameters (Fig. 1, B and C) from electrically stimulated isolated rat myocardium (for brevity, only example fibers for the 0 and 1% mannitol experiments are shown here). A comparison of the Fig. 1A, left (0% mannitol) and right (1% mannitol) reveals no significant impact of external osmotic compression on the SL dependence of normalized force development, except for some apparent slowing of the rates of activation and relaxation. It should be noted here that although developed force was not significantly altered, passive tension was increased with the application of mannitol, especially with the addition of 4% mannitol (data not shown). Furthermore, 4% mannitol substantially reduced peak active force development.
The application of mannitol decreased the intensity of the 1,0 equatorial X-ray diffraction peak at short SL with no further alteration in the 1,0 reflection intensity when the muscle was stretched to 2.2 μm, unlike untreated muscles (Fig. 1B; left). Stretch of the sarcomeres resulted in an increase in IM3 (Fig. 1C, left). Whereas an addition of 1% mannitol to the solution decreased the overall intensity of the M3 meridional peaks, the M3 intensity increased by the same relative amount with increasing SL as in the absence of mannitol (Fig. 1C, right).
As summarized in Fig. 2, changes in SL induced approximately linear changes in three of the measured parameters: lattice spacing (Fig. 2A; absolute changes and A′, relative changes); IM3 (Fig. 2C); and peak twitch force (Fig. 2D) with 0, 1, and 4% mannitol in the bathing solution. Increasing SL in 0% mannitol induced significant decreases in the equatorial intensity ratio (I1,1/I1,0) over the range of the SLs examined. The addition of 1% mannitol, however, induced a decrease in I1,1/I1,0, at the initial short SL compared with that of control muscles. Subsequent increases in SL had little additional affect on I1,1/I1,0 in 1% mannitol. It should be noted that the application of 4% mannitol induced considerable paracrystalline disorder in the sarcomere such that the 1,1 reflections were not visible, thereby precluding measurement of I1,1/I1,0 for this group; however, lattice spacing and IM3 could still be measured.
As expected, the addition of mannitol compressed the starting (short SL) lattice spacing by ∼0.5–0.7 and ∼5 nm in 1 and 4% mannitol, respectively (Fig. 2A). Interestingly, linear regression analysis indicated a significantly increased slope (P < 0.05) for the decrease in myofilament lattice spacing with increasing SL for both 1 and 4% mannitol, indicating that osmotic compression of the isolated trabeculae enhanced the dependence of myofilament lattice spacing on SL. Unlike lattice spacing and equatorial intensity ratio, the impact of SL changes on the changes in twitch force (Fig. 2D) and IM3 (Fig. 2C) were not affected by the compression with 1% mannitol. With 4% mannitol, there was an approximately equal reduction in the increase in twitch force and IM3 with increasing SL, as indicated by the reduced slopes of these relationships (Fig. 2, C and D). As a consequence, the changes in both parameters were matched; that is, even in this highly compressed state, the changes in force and myosin head orientation were still sensitive to changes in SL. From these data, it may be concluded that the changes in not only peak systolic twitch force but also the orientation of the myosin heads (as indicated by IM3) are sensitive to changes in SL.
While it is clear from Fig. 2 that changes in lattice spacing and IM3 are correlated to the changes in SL, it is not possible from these data alone to determine which parameter is the primary cause for the changes in twitch force. To determine which structural parameter correlated best with peak twitch force, we plotted the percent changes in lattice spacing, I1,1/I1,0 equatorial intensity ratio, and IM3 as a function of the percent change in developed force as shown in Fig. 3. The underlying assumption is that if peak force is directly related to a given structural parameter, the data collected with different degrees of osmotic compression should all fall on the same curve. Figure 3A shows that within each individual osmotic group, (0, 1, and 4% mannitol), the relative changes in the lattice spacing appear to vary nearly linearly with the relative changes in peak twitch force. If it is assumed that all three data sets should be described by a single linear relationship, then lattice spacing and peak force are poorly correlated (r2 = 0.089), suggesting that these factors are not tightly coupled to each other. As with lattice spacing, the changes in the I1,1/I1,0 within each osmotic group (0 and 1% mannitol) also appear to vary linearly with changes in peak systolic force (Fig. 3B). But when the data for both 0 and 1% mannitol are combined, the degree of linear correlation with twitch force is not strong (r2 = 0.69), indicating that there may not be a single relationship between I1,1/I1,0 and force. This suggests that although radial movements of the myosin heads away from the thick filaments may influence subsequently developed force (13), this relationship may be indirect. In contrast, when the percent changes of IM3 from all three osmotic groups (0, 1, 4% mannitol) are plotted against the relative changes in peak systolic force, a single linear relationship emerges where IM3 is highly correlated to peak systolic twitch force. (r2 = 0.874) (Fig. 3C). These results indicate that changes in the orientation of the myosin heads, as indicated by IM3, bear the same relationship to the changes in peak systolic force whether or not mannitol is present.
The combination of focused high-intensity X-rays and precise timing has allowed us to assess comprehensive information regarding the structure of the diastolic myofilament lattice just before electrical stimulation in isolated rat myocardium. Based on our observations, we conclude that a key mechanism underlying the Frank-Starling law of the heart may be an increase, upon an increase in SL, in the orientational ordering of the myosin heads perpendicular to the thick filament backbone in resting muscle before activation.
It has been known for some time (9, 12, 22, 38) that the IM3 from X-ray diffraction patterns is a sensitive indicator of the average orientation of the myosin heads with respect to the thick filament backbone. The intensity is expected to be weakest when the cross bridges are at oblique angles or distributed over a wide range of axial angles; conversely, IM3 increases as the population of cross-bridges becomes more perpendicular to the filament axis (25). Most of the studies to date have been concerned with axial movements of the myosin heads in the molecular mechanism of muscle contraction. However, the notion that alterations in the resting myosin head orientation might participate in force modulation mechanisms in cardiac muscle is a new concept for which an interpretive framework does not yet exist.
Nevertheless, there is a growing literature reporting studies of the indirect flight muscle of insect model systems that does indeed suggest that the orientational ordering of the myosin heads under resting conditions may impact subsequent active force production. When insect flight muscle [either Drosophila (9) or Lethocerus (37)] is passively stretched before active contraction, the M3 meridional reflection increases, indicating an improvement in the orientational ordering of the myosin heads (or a subpopulation thereof) that may be instrumental in the stretch-activation response (37). Taken together, these studies suggest that stretch alone, with no exogenous stimulation, might improve the orientational ordering of the myosin heads by increasing the degree of phosphorylation of the myosin light chains. AL-Khayat et al. (3), based on their modeling study of relaxed insect flight muscle, argued that an arrangement of myosin heads perpendicular to the filament long axis minimizes the volume that needs to be searched by the myosin heads before it can find a stereo-specific binding site under contracting conditions, thereby enhancing the probability of productive acto-myosin interactions and hence active force. Our data suggest that this may indeed be the case in cardiac muscle. Alternatively, electron microscopic studies of isolated thick filaments from a variety of species, including mammalian cardiac muscle (45) as well as a number of modeling studies of X-ray patterns (3, 36), indicate that under relaxed conditions while one myosin head of a given pair is in an extended configuration, the other is wrapped the filament making presumably inhibitory contacts with other heads, either intramolecularly (3) or intermolecularly (36). We do not know to what extent the relaxed configuration seen in isolated thick filaments from cardiac muscle (45) pertains to twitching muscle in diastole. Titin-based passive tension has been shown to influence length-dependent activation in studies that showed that stiffer transgenic titins resulted in more pronounced length-dependent activation (28). The possibility needs to be considered, therefore, that the putative inhibitory contacts between heads could be affected by thick-filament strain induced by the higher passive tension at longer lengths, which would allow heads to be released and form more favorable actomyosin interactions, presumably with improved orientational order perpendicular to the thick filament backbone.
In Drosophila, it has been shown that increased orientational disorder in the myosin heads (as indicated by reduced IM3) (12) in mutant flies where the phosphorylation sites on the regulatory light chain (RLC) are replaced by nonphosphorylatable alanines is associated with impaired flight ability with lower wing-beat frequency (10). These results were interpreted to suggest that phosphorylation of the RLC promotes head orientations that allow optimal interactions with the thin filament. Thus, while the RLC in Drosophila is constitutively phosphorylated, in other species, where phosphorylation state varies under physiological conditions, the affect of phosphorylation on the orientation of the myosin heads may well provide a regulatory mechanism. While measurements of IM3 as a function of phosphorylation state of the RLC are not yet available from mammalian muscle, there is evidence that phosphorylation state of the RLC indeed affects the performance of human (34, 42), rat (6, 20, 35), guinea pig (2), and rabbit myocardium (33). Reduced light chain phosphorylation, and presumably less favorable myosin head orientation, found in failing human myocardial tissue was associated with reduced calcium sensitivity compared with normal, donor tissue, and this could help explain the reduced functional activity observed in heart failure (42). In accordance with these results, it has been demonstrated that phosphorylation of myosin light chain, by myosin light chain kinase, increases maximally activated force and calcium sensitivity in healthy, skinned “demembranated” trabeculae isolated from rat ventricular tissue (6, 35) and human atrial tissue (34). A possible connection between light chain phosphorylation and SL changes is suggested by results from Hidalgo et al. (20) who reported that altering the passive tension of isolated rat hearts increased the degree of phosphorylation of myosin light chains in epicardial tissue, demonstrating that alterations in SL can induce alterations in the phosphorylation state of the myosin light chain. Likewise, Ait Mou et al. (2) reported an increase in RLC phosphorylation upon stretch in isolated guinea pig myocardium, a result that was recently confirmed in rabbit myocardium (33).
There is ambiguity in interpreting IM3 changes, since the heads could all be “ordered” (i.e., have the same angle) but display an orientation that is not perpendicular to the thick and thin filaments or, conversely, the average angle could be perpendicular to the long axis but have a significant dispersion around this angle. Both of these conditions result in a reduction of IM3 by similar amounts; that is, these conditions cannot be distinguished without additional information. We therefore have used changes in “orientational ordering” to encompass these possibilities, and the reader should use caution in over-interpreting the observed changes.
We have demonstrated a strong correlation between the orientation of the myosin heads and peak twitch force development. We propose that sarcomere strain results in favorable alignment of myosin heads at increased SL, resulting in enhanced force development at a given level of activator calcium concentration, possibly by modulation of myosin light chain phosphorylation. Full elucidation of such mechanisms may require detailed analysis of two-dimensional diffraction patterns from isolated myocardium as a function of SL in conjunction with measurements of myosin light chain phosphorylation state as a function of SL. Whatever the fundamental stretch sensitive mechanism is, it is insensitive to the concomitant changes in lattice spacing as SL is increased, as shown in Fig. 4.
This study was supported, in part, by National Heart, Lung, and Blood Institute Grant HL-75494 (to P. P. de Tombe). Use of the Advanced Photon Source was supported by the U.S. Department of Energy, Basic Energy Sciences, Office of Science, under contract No. W-31-109-ENG-38. BioCAT is a National Institutes of Health-supported Research Center RR-08630.
No conflicts of interest, financial or otherwise, are declared by the author(s).
The content is solely the responsibility of the authors and does not necessarily reflect the official views of the National Center for Research Resources or the National Institutes of Health.
- Copyright © 2011 the American Physiological Society