Sarcoplasmic Reticulum and Nuclear Envelope Are One Highly Interconnected Ca2+ Store Throughout Cardiac Myocyte
http://www.100md.com
Xu Wu, Donald M. Bers
参见附件。
the Department of Physiology, Loyola University Chicago, Maywood, Ill.
Abstract
Previous ventricular myocyte studies indicated that ryanodine receptors (RyRs) are in the sarcoplasmic reticulum (SR) and are critical in excitation–contraction coupling, whereas the inositol trisphosphate (InsP3) receptors are separately localized on the nuclear envelope (NucEn) and involved in nuclear Ca2+ signaling. Here, we find that both caffeine and InsP3 receptor agonists deplete free [Ca2+] inside both SR and NucEn. Fluorescence recovery after photobleach (FRAP) was measured using the low-affinity Ca2+ indicator Fluo-5N trapped inside the SR and NucEn (where its fluorescence is high because [Ca2+] is &1 mmol/L). After Fluo-5N photobleach in one end of the cell, FRAP occurred, accompanied by fluorescence decline in the unbleached end with similar time constants (&2 minutes) until fluorescence regained spatial uniformity. Notably, SR and NucEn fluorescence recovered simultaneously in the bleached end. Ca2+ diffusion inside the SR-NucEn was also measured. SR Ca2+-ATPase was completely blocked but without acute SR Ca2+ depletion. Then caffeine was applied locally to one end of the myocyte. In the caffeine-exposed end, free SR [Ca2+] ([Ca2+]SR) declined abruptly and recovered partially (=20 to 30 seconds). In the noncaffeine end, [Ca2+]SR gradually declined with a similar , until [Ca2+]SR throughout the cell equalized. We conclude that the SR and NucEn lumen are extensively interconnected throughout the myocyte. Apparent intrastore diffusion coefficients of Fluo-5N and Ca2+ were estimated (&8 μm2 sec–1 and 60 μm2 sec–1). This rapid luminal communication may maintain homogeneously high luminal [Ca2+], ensuring a robust and uniform driving force for local Ca2+ release events from either SR or NucEn.
Key Words: sarcoplasmic reticulum nuclear envelope (NucEn) fluorescence recovery after photobleach (FRAP) caffeine
Introduction
Cardiac myocytes respond to extracellular stimuli by different signaling pathways. An important mechanism in cardiac myocytes is excitation–contraction coupling (ECC).1 Electrical depolarization of the cell membrane opens L-type Ca2+ channels, causing Ca2+ influx. This Ca2+ entry activates ryanodine receptors (RyRs) on the intracellular Ca2+ store sarcoplasmic reticulum (SR), inducing robust SR Ca2+ release. This local Ca2+ -induced Ca2+ release (CICR)2 is the essence of ECC.
Intracellular Ca2+ stores include SR (endoplasmic reticulum [ER] in nonmuscle cells) and nuclear envelope (NucEn). Free [Ca2+] inside these stores can be &1 mmol/L, which is similar to that in the extracellular space.3 The SR is composed of junctional SR (jSR), which is covered by RyRs and faces toward the T tubules, and free SR (fSR), which contains SR-ER Ca2+-ATPase (SERCA).4 The NucEn surrounds the nucleus and has both an inner and an outer membrane and a space in between where [Ca2+] can be millimolar.
Recently, we found that cardiac myocytes use local Ca2+ release from inositol trisphosphate (InsP3) receptors (InsP3R) in the NucEn to respond to the neurohumoral stimuli in excitation–transcription coupling (ETC).5 Endothelin-1 (ET-1) application caused InsP3 to activate local Ca2+ release from the NucEn via InsP3R, which activated calmodulin and CaMKII to phosphorylate histone deacetylase-5 (causing its nuclear export) and activation of transcription.5 Notably, this pathway could only be activated by Ca2+ from local InsP3R (not global Ca2+ transients), presumably because calmodulin and CaMKII physically associate with the InsP3R at the NucEn.5,6 Indeed, in ventricular myocytes, RyRs are mainly in SR, whereas InsP3R are mainly in NucEn.6 Thus, SR and NucEn differ structurally and functionally and may reflect discrete physically different Ca2+ stores. In addition, the relatively rigid sarcomeric organization in striated muscle with a dense protein mesh at the Z-lines raises the possibility that local SR within a given sarcomere could serve its particular sarcomere but not necessarily communicate with neighboring sarcomeres either longitudinally or transversely. The main aim of the present study was to determine whether different SR regions and NucEn constitute discrete separate Ca2+ stores.
In some cell types (nonmuscle cells), it is generally accepted that the outer nuclear membrane and ER have similar characteristics and outer nuclear membrane is continuous with ER membrane including SERCA, whereas the inner nuclear membrane is quite different from the outer nuclear membrane.7,8 However, it is not known whether the lumen of ER/SR and NucEn is subcompartmentalized or continuous. The existence of separated Ca2+ stores has been proposed by some groups,9–11 but rapid diffusion of Ca2+ and other molecules between the lumen of NucEn and ER was also shown in leukemia cells and fibroblasts.12,13
Here we assess diffusion inside the SR and NucEn using the low-affinity Ca2+ indicator Fluo-5N, fluorescence recovery after photobleach (FRAP) and local caffeine applications. We found evidence that the SR and NucEn are highly interconnected and comprise 1 large continuous intracellular Ca2+ store. This will help to understand ECC and ETC, as well as cardiac arrhythmias, because inhomogeneous SR Ca2+ release may contribute to Ca2+ waves.14,15
Materials and Methods
Myocyte Isolation Fluo-5N Loading and Cell Permeabilization
Myocytes were isolated as previously described16 and loaded with Fluo-5N acetoxymethylester (10 μmol/L) for 2 hours and deesterified for 1.5 hour.3 For intact myocytes, the superfusate contained (in mmol/L) NaCl 140, KCl 4, MgCl2 1, CaCl2 2, HEPES 10, glucose 10 (pH 7.4, 23°C). For permeabilization, myocytes were exposed to relaxing solution (in mmol/L: EGTA 0.1, HEPES 10, K-aspartate 120, free MgCl2 1, ATP 5, reduced glutathione 10, phosphocreatine di-Tris 5 [pH 7.4]) and then permeabilized by saponin (50 μg/mL) for 20 seconds. After permeabilization, we used an internal solution (same as relaxing solution, but free [Ca2+] was adjusted to 100 nmol/L and 8% dextran was added to prevent swelling on permeabilization, with pH 7.2). [Ca2+]SR was calculated using pseudoratio method described previously for single-wavelength indicator.5 F0 corresponds to resting [Ca2+]SR, which was taken to be 1 mmol/L.3
Imaging Under Confocal Microscopy
Images were acquired with a Bio-Rad Confocal Microscope5 (488 nm argon laser excitation, emission at 500 nm long pass). Laser power was 100% for photobleach, 4% for line scan images, and 0.5% for whole cell images. Because the laser power for image recording is low, unintended photobleach was negligible. Confocal imaging used a x40 oil immersion objective and temporal resolution of 166 lines per second. ImageJ software was used for image analysis.
For some FRAP experiments, the pinhole diameter was increased to 12 mm (maximal) to obtain fluorescence from the whole cell thickness. However, in caffeine experiments (where release occurs at all depths) and other experiments, pinhole diameter was chosen for confocality (1.2 mm). All fluorescence data were background subtracted.
Simulation of Fluo-5N Diffusion Inside the Ca2+ Store
The cell/SR-NucEn was considered a circular cylinder, with cross-section area A and length L, divided into 20 equal-volume compartments (L) longitudinally (L=L/20). Diffusion was allowed between neighboring compartments (boundary flux was set to 0). For each compartment, Fluo-5N fluorescence or [Ca2+]SR (C) was determined by Fluo-5N or Ca2+ diffusion from/to the adjacent compartments. The initial curve was from experimental data fitted to sigmoid curve. The change in C in compartment n per unit time (Cn/t) is equation
where D is the apparent diffusion coefficient for Fluo-5N or Ca2+, Vn is the volume of compartment n, and Vn=AL. The above applies to the cylindrical cell structure. SR inside the cell has different structure and accounts for only 3% to 5% of the cell volume.2 The SR cross-section area should also be only 3% to 5% of that of the whole cell. However, because Vn=L·A, equation 1 can be independent of A. equation
Thus, D is determined by concentration gradients and L2/t. Note that D (as defined here) will be reduced by contributions of binding of Ca2+ or Fluo-5N inside the SR-NucEn and also tortuosity of diffusional path (ie, noncylindrical actual geometry of the network). At a time point t, C in compartment n (Cn,t) follows: Cn,t=Cn,t–1+Cn,t–1. Equation 2 was solved numerically with t=0.1 second. The value of D was varied to obtain a least square fit to the measured spatiotemporal Cn,t profiles from experiments.
Statistical Analysis
Data are presented as mean±SEM with significance (P<0.05) determined using unpaired 2-tailed Student t test.
Results
Caffeine or InsP3R Agonists can Deplete [Ca2+]SR in Both SR and NucEn
Rabbit ventricular myocytes were loaded with Fluo-5N and permeabilized with saponin. Fluo-5N is a low-affinity Ca2+ indicator (Kd=400 μmol/L),3 only bright where [Ca2+] is very high, as in SR and NucEn (Figure 1A). We also see irregularly distributed bright spots which are caffeine and InsP3-insensitive (most obvious in Figure 6). The Fluo-5N signal was stable when myocytes were left in skinned fiber solution without any treatment (Figure 1A). This shows that [Ca2+] inside SR and NucEn are stable and that no appreciable Fluo-5N is lost over the experimental time course. Caffeine (10 mmol/L) application causes an immediate decline in Fluo-5N signal both in the SR and NucEn (Figure 1B and 1C), with the SR preceding the NucEn. Figure 1D shows SR and NucEn [Ca]SR depletions during electrically evoked Ca transients (0.5 Hz). Again, [Ca2+]SR declined transiently (Ca2+ scraps) in both SR and NucEn. However, in the NucEn, the depletion amplitude was smaller and delayed in both time to nadir and recovery time constant (rec). These results are consistent with RyR-dependent intra-SR Ca depletions driving those in the NucEn, but they do not prove that the stores are connected. On the other hand, if RyRs are only on the SR (not the NucEn),6 and if the SR and NucEn are not connected, then caffeine and twitches should not cause NucEn Ca2+ depletion.
Application of InsP3 (10 μmol/L) or the potent InsP3R agonist adenophostin (10 μmol/L) for 30 minutes also gradually decreased the Fluo-5N signal in both NucEn and SR in permeabilized myocytes (Figure 2A and 2B and time course of Fluo-5N signal changes in Figure 2C). Note that InsP3R activation produced a very much slower less complete Ca2+ depletion of the SR and NucEn. This is consistent with the much lower number of InsP3R (versus RyR) in ventricular myocytes (and lower intrinsic Ca2+ flux rate).2 That is, the very high Ca2+ flux rate via caffeine-activated RyRs completely depletes [Ca2+]SR (despite continued SERCA function). In contrast, the much lower InsP3R flux rate only slowly depletes [Ca2+]SR and results in a new steady state where the SERCA pump and leak rates rebalance at a new low [Ca2+]SR. This explains why the lack of time delay between SR and NucEn signals (similar values), because the InsP3-induced leak rate may be small compared with potentially rapid NucEn-SR Ca2+ equilibration.
Again, if InsP3R are only on the NucEn (not SR) and the 2 pools are not interconnected, this result is unexpected.6 This raises the possibility that the SR and NucEn lumens are highly connected. However, the above phenomena could also happen even if the SR and NucEn were not connected but if there were very small numbers of RyRs on the NucEn and of InsP3R on the SR. These could be functionally effective even if their existence is only near the detection limit of immunocytochemistry, fractionation methods, or Western blot analyses.6 To test whether the lumen of the SR is connected to both the NucEn and distant regions of the SR, we performed the following FRAP and caffeine-induced [Ca2+]SR depletion experiments.
Fluo-5N Diffusion Inside the SR and NucEn: FRAP Studies
Permeabilized, Fluo-5N–loaded myocytes were photobleached on approximately half of the cell longitudinally with the other half left almost unaffected (Figure 3A). Because of dispersion and scatter of illumination, the beam was centered near the left cell end, causing an initial gradient of bleach that was maximal at the end of the cell but gradually declining to be negligible at the cell center (Figure 3A and 3B). Fluorescence in the bleached half (left) decreased immediately on bleach. After bleach, the SR and NucEn fluorescence increased gradually in the bleached half, and simultaneously there was a complementary gradual decrease of fluorescence in the unbleached half (right) until the 2 halves reached the same level (full FRAP; Figure 3A, with longitudinal profile in Figure 3B). On Fluo-5N photobleach, local [Ca2+]SR gradients are neither likely to be significant nor to complicate subsequent FRAP measurements. This is partly because the intrastore Fluo-5N concentration is very low compared with physiological SR Ca2+ buffers (eg, calsequestrin; see Figure II in the online data supplement, available at http://circres.ahajournals.org). Rather, subsequent FRAP is attributable almost entirely to the diffusion of Fluo-5N within the SR and NucEn. Note that in these and subsequent longitudinal profiles, the regular sarcomeric pattern of jSR (with 1.9-μm spacing3; see Figure 3B inset) was intentionally eliminated by smoothing.
Figure 3C shows the time course of FRAP in the bleached half and the decay of fluorescence in the unbleached half. The total fluorescence after bleach (&15% overall bleach) did not change during FRAP, indicating that these measurements are not complicated by changes in [Ca2+]SR. The time constant for FRAP on the left and fluorescence decline on the right were similar (=112.1±9.8 and 115.3±5.7 seconds). The NucEn on the left also demonstrated FRAP, with similar kinetics as the SR (=119.8±7.6 seconds).
This FRAP experiment was also done in the transverse direction (Figure 4). In this image, the upper side of the cell was bleached and transverse FRAP and complementary fluorescence decline in the bottom occurred. The results are qualitatively similar for transverse versus longitudinal FRAP, except that the time constants were faster transversely (&15 seconds) than longitudinally (&110 seconds). This is not surprising because the distance is much less in the transverse direction. Both longitudinal and transverse FRAP were complete and spatially uniform. Because these were permeabilized myocytes and Fluo-5N was trapped inside the SR and NucEn, the recovery of the bleached half can only be attributable to the Fluo-5N diffusion inside the Ca2+ store. In Figure 3A and 3C, both NucEn and SR on the left recover during FRAP, evidence of extensive connection of the SR with both the NucEn and with more distant SR regions.
Because the images in Figures 3 and 4 were confocal, the recovery of fluorescence on the bleached half could be partially attributed to Fluo-5N diffusion from regions above and below the confocal plane (in addition to the other half in the x-y plane). On the other hand, bleach in the Z direction is likely to be substantial (as it is nonconfocal). Nevertheless, to avoid FRAP from above and below the confocal plane, the signal from the whole thickness of the cell was obtained by opening the confocal pinhole maximally (12 mm versus 1.2 mm in confocal images). As in Figure 3, we bleached the left half of the cell (including 1 nucleus), whereas the right half remained relatively unaffected (Figure 5A). The images are fuzzier because of intentional nonconfocality. The fluorescence on the left half dropped immediately by 30% (Figure 5B). The degree of photobleach was only 30% because of the brief laser exposure (less than 1 second to limit cell damage and optimize temporal resolution). FRAP is apparent in both SR and NucEn on the bleached half, accompanied by a gradual and complementary decline on the unbleached half. We compared FRAP kinetics by choosing region of interest (ROI). The values for the recovery of both halves were similar, &2 minutes (Figure 5B). The amplitudes of the post-bleach decrease and increase of fluorescence in the 2 halves were also about the same. The post-bleach total cell fluorescence was constant (Figure 5B). Again, because this was in a permeabilized cell and the only source of Fluo-5N was from the SR and NucEn, this result means that Fluo-5N diffuses from one end of the cell to the other inside the SR-NucEn network. If Fluo-5N came out of the cell, it would diffuse away and we would see loss of total fluorescence. This indicates that the Ca2+ stores of the 2 halves are connected.
At the same longitudinal location, the bleach and recovery of the SR and NucEn were also compared for different ROI (Figure 5C). The SR and NucEn regions were bleached to the same extent and recovered with a similar time constant. Because the recovery of NucEn can only be via Fluo-5N diffusion from the SR, this demonstrates that the SR and NucEn are extensively connected. A region of SR closer to the bleached end of the cell had more extensive bleach (60% versus 30% on average for this half) and also recovered with a slightly longer (see simulations below). The similar FRAP results in Figures 3 and 5 confirm control experiments, indicating photobleach through most of the cell depth.
Ca2+ Diffusion Inside the SR and NucEn
We also performed caffeine-induced [Ca2+]SR depletion at one end of the cell to assess Ca2+ diffusion inside the SR and NucEn. This was done in intact cells under special conditions, previously developed and validated,17 where SERCA could be completely blocked but SR stores maintained until a release is triggered. Intact myocytes were paced at 0.5 Hz for 10 beats to preload Ca2+ stores. Then pacing was stopped and perfusion was switched for 90 seconds to a Na+-free, Ca2+-free NT (Li+ replaced Na+ and 10 mmol/L EGTA was included) with 5 μmol/L thapsigargin (TG) (to block the SERCA pump). Using this protocol, [Ca2+]SR was not decreased significantly (Figure 6A and 6B). However, SERCA was completely blocked because, after acute [Ca2+]SR depletion by caffeine (Figure 6C), caffeine washout and then field stimulation [Ca2+]SR could not be detectably refilled (Figure 6D versus 6C). After treatment with TG, the SR Ca2+ content does eventually decline (even in 0 Na+/0 Ca2+ solution), but there is a practical time window (&5 minutes) in which these experiments can be done. Note that the caffeine-insensitive bright spots were also unaffected throughout the procedure.
In cells pretreated with TG 0 Na+/0 Ca2+, as above, we evaluated intra-SR Ca2+ diffusion using local caffeine application. Figure 7A shows local caffeine application at the left end of the cell briefly (1 to 2 seconds) to release Ca2+ from SR and NucEn at this end. There is an acute discernable [Ca2+]SR gradient from the left end to the middle (Figure 7A and 7B). Because SERCA was completely blocked, the only source of Ca2+ recovery at the left end is Ca2+ diffusion within the Ca2+ store from the right. If this recovery happens, SR and NucEn must be connected throughout the myocyte. Indeed, [Ca2+]SR recovery at the left and right ends occurred with similar values (Figure 7C). Additionally, the total [Ca2+]SR inside the Ca2+ store was constant, indicating acute SR Ca2+ release, no net reuptake and no further loss (Figure 7C). In addition, the recovery of SR and NucEn occur simultaneously with similar values (Figure 7D). This again implies that SR and NucEn are highly connected via their lumens.
Simple Model of Fluo-5N Diffusion Inside the SR and NucEn
To further investigate Fluo-5N and Ca2+ diffusion inside the SR and NucEn, we made a simple 1D diffusion model (Figure 8A). Figure 8 (B and E) shows exemplar spatiotemporal data, smoothed by fitting to sigmoid curves (Fluo-5N and Ca2+ diffusion, respectively). C and F in Figure 8 are the corresponding simulated curves. The initial spatial profile (t=0) was set to the smoothed experimental data, and then diffusion of Fluo-5N or Ca2+ was calculated by the diffusion equation with D adjusted to best fit the data. The apparent diffusion coefficients of Fluo-5N (DFluoStore) and Ca2+ (DCaStore) inside the Ca2+ store were estimated at 8 μm2 sec–1 and 60 μm2 sec–1, respectively.
D and G in Figure 8 are theoretical time courses of Fluo-5N or Ca2+ recovery at different longitudinal sites, with values shown in the insets. The value is a U-shaped function with the fastest near the bleach interface and the longest at the ends. This is consistent with experimental data in Figure 5C, where we saw that a site farther to the left exhibited a longer FRAP .
Discussion
The SR and NucEn Are Highly Interconnected in Cardiac Ventricular Myocytes
The data here measuring both Fluo-5N FRAP and Ca2+ diffusion inside the SR, directly demonstrate for the first time that the SR lumen is highly interconnected with the NucEn and also with distant regions of the SR. Moreover, the SR-NucEn throughout the myocyte appears to be a single large continuous Ca2+ storage compartment. Our initial expectation was that the SR within each sarcomere would be relatively isolated from other SR regions. On the other hand, occasional electron micrographs show that jSR can connect SR from 1 side of the Z-line to the other (and to transverse sarcomeres as well),4,18,19 but there has been no prior functional data to indicate the extent to which the SR network is continuous. Clearly it is. Similarly, there have been reports that the outer nuclear membrane is continuous with the ER in some cell types, but no prior functional evidence in cardiac myocytes. Indeed, the apparent difference in Ca2+ release channels (RyR in SR and InsP3R in NucEn) and functional roles (ECC for SR and ETC for NucEn) might lead one to expect that the pools are separate. Clearly this is not the case. These processes and release channels are fueled by the same Ca2+ pool. We were surprised how rapid diffusion appears to be within the SR-NucEn network.
Intra–SR-NucEn Diffusion
In our simple simulation, the estimated values for DCaStore and DFluoStore were 60 μm2 sec–1 and 8 μm2 sec–1, respectively (ie, 7.5 times faster for Ca2+). Because D is inversely related to the square root of molecular weight, the expected ratio is 5.3 times higher D for Ca2+ versus Fluo-5N. Given the many factors that could complicate our simplistic analysis, this is in rather good agreement. The estimated DCaStore is approximately 10-fold slower than expected in aqueous solution (700 μm2 sec–1).20 Main factors that would reduce DCaStore are Ca2+ binding to fixed sites, high viscosity, and tortuosity of the diffusional path. We favor tortuosity as a dominant factor because: (1) both DCaStore and DFluoStore seem reduced roughly in parallel; (2) path-length tortuosity is expected and would affect both similarly; and (3) because intra-SR buffering of Ca2+ and Fluo-5N are likely very different. Indeed, images of the SR network structure would suggest substantial tortuosity. Of course Ca2+ also binds to intra-SR buffers like calsequestrin, but the low-affinity (Kd=0.5 mmol/L)21 and high off-rate would limit the impact on DCaStore. Notably, our DCaStore of 60 μm2 sec–1 is much faster than the apparent DCa estimated in myoplasm (1.4 μm2 sec–1).20 Whereas tortuosity and viscosity may contribute to the low apparent DCa in cytosol, Ca2+ binds to very many, relatively high-affinity (slow off-rate) fixed Ca2+ binding sites in the cytosol (eg, troponin C, SERCA, calmodulin, myosin).2,22 Thus, Ca2+ binding may be the major limitation for Ca2+ diffusion in myoplasm, versus tortuosity inside the SR. The faster Ca2+ diffusion inside the SR may also move more Ca2+ because [Ca2+]SR is 10 000 times higher than in cytosol. Fast intra-SR Ca2+ diffusion may also be important to consider in analyzing propagating waves of CICR in muscle and other cell types.
Ca2+ Wave Initiation and Propagation
Ca2+ waves are thought to initiate when [Ca2+]SR is elevated (perhaps locally), and this initiates a local SR Ca2+ release event (Ca2+ spark).2,23 The wave then propagates via CICR, as released Ca2+diffuses to the next junction where it triggers local SR Ca2+ release and so on. Note that local [Ca2+]SR at the next site must be sufficiently high for RyRs to be activated by the lower local [Ca2+]i (<1 μmol/L) than occurs via local Ca2+ current during normal ECC (>10 μmol/L).2 The rapid equilibration within the SR-NucEn described here would help to keep 1 area from experiencing a higher local [Ca2+]SR (ie, the network buffers local [Ca2+]SR elevation and depletion), thus potentially limiting initiating events for diastolic Ca2+ waves. It may also interfere with wave propagation, if local [Ca2+]SR ahead of the initiating wave could decrease (because of release at the previous region) before release at that next junction is activated by high local [Ca2+]i. That is, lower local [Ca2+]SR could desensitize local RyRs at the next site. Both of these effects might decrease the likelihood of wave initiation and propagation. Of course, these ideas need future experimental tests.
Local elevation of [Ca2+]SR was suggested as a potential mechanism of arrhythmogenesis in hypertrophic heart,15 causing loci prone to wave initiating. Given the present results, this scenario seems unlikely, unless the SR is more fragmented in hypertrophy or heart failure. In preliminary experiments in a rabbit heart failure model,24 we did not uncover any fragmentation of the SR-NucEn using the same FRAP approach used here (not shown). This may merit further exploration in pathophysiological models.
Local SR Ca2+ Depletion During SR Ca2+ Release
Rapid Ca2+ diffusion within the SR (along with buffering by calsequestrin) can enhance SR Ca2+ release through the jSR by minimizing the local SR Ca2+ depletion, which occurs during ECC. Indeed, the first direct measurements of [Ca2+]SR and Ca2+ scraps during normal ECC showed that there was no detectible delay in the time course of local [Ca2+]SR decline in fSR versus jSR.3 This has been confirmed in subsequent measurements,25 consistent with rapid intra-SR diffusion within a sarcomere. Local photolytically induced CICR was shown to have an extremely short refractoriness compared with global ECC,26 and this was ascribed to very rapid repletion of those local SR Ca2+ release sites by neighboring SR regions. Brochet et al25 observed local isolated [Ca2+]SR depletion during Ca2+ sparks (called Ca2+ blinks). They took the detectability of these events as evidence of restricted fSR-jSR diffusion. However, both the extent of [Ca2+]SR depletion and recovery time constant were much smaller for Ca2+ blinks versus Ca2+ scraps or global SR Ca2+ release. These results, we think are consistent with very rapid Ca2+ replenishment from neighboring SR regions as we show here throughout the whole cell.
Our measurements here are over a spatial scale that makes [Ca2+]SR gradients readily detectable. This places a lower limit on DCaStore (60 μm sec–1) within a single sarcomere. We scaled down our simple 1D model to a half-sarcomere. We decreased [Ca2+]SR instantaneously from 1 to 0.5 mmol/L only in the junctional 50-nm compartment (corresponding to a maximal local jSR release). With our DCaStore value of 60 μm2 sec–1, the [Ca2+]SR gradient throughout that 1-μm half-sarcomere within 2 ms is less than 3% (0.96 to 0.99 mmol/L) versus 25% if we had used DCa reported for myoplasm (1.4 μm2 sec–1; see supplemental Figure I). These results are consistent with the lack of detectable jSR-fSR [Ca2+]SR gradient during Ca2+ scraps. This simple model is a starting point for further considerations.
Uniform Ca2+ Distribution Is Physiologically Significant
The spatially uniform [Ca2+]SR provides a spatially uniform driving force for SR Ca2+ release throughout the myocyte during ECC. This may be important in producing homogeneous Ca2+ release and contractile activation throughout the ventricular myocyte. The network also ensures fidelity of local SR or NucEn Ca2+ release events involved in local Ca2+ signaling. That is, the entire network of SERCA pumps in the SR-NucEn help maintain this [Ca2+] gradient for all its uses. The results in Figure 2 show that InsP3 or adenophostin reduces [Ca2+]SR everywhere, even if the primary functional targets and locus of release are at the NucEn. The incomplete [Ca2+]SR depletion indicates that the leak generated by strong InsP3R activation is less than the overall forward SERCA pumping rate at 100 nmol/L [Ca2+]i (&20 μmol/L cytosol per second) or more than 100 times slower than SR Ca2+ release via RyR during a twitch or caffeine-induced Ca2+ transient).2 A potential cost of this rapid equilibration is that any pathophysiological SR Ca2+ leak could deplete [Ca2+]SR everywhere in the cell and limit both RyR and InsP3R-mediated Ca2+ release.
In conclusion, the cardiac SR is highly interconnected to both NucEn and more distant SR throughout the entire myocyte. Fast Ca2+ diffusion within this SR-NucEn store may stabilize local free [Ca2+]SR to provide uniform driving force for RyR and InsP3R. Homogeneous [Ca2+]SR may ensure uniform and synchronous SR Ca2+ release and contractile activation.
Acknowledgments
We thank Drs Eckard Picht and Fei Wang for helpful discussions and Jayme O’Brien and Karl Hench for technical assistance.
Sources of Funding
Supported by NIH grants HL30077 and HL64724 (D.M.B.) and an American Heart Association Fellowship (to X.W.).
Disclosures
None.
Footnotes
Original received April 19, 2006; revision received May 31, 2006; accepted June 13, 2006.
References
Bers DM. Cardiac excitation-contraction coupling. Nature. 2002; 415: 198–205. [Order article via Infotrieve]
Bers DM. Excitation-Contraction Coupling and Cardiac Contractile Force. 2nd ed. Dordrecht, The Netherlands: Kluwer Academic Publishers; 2001.
Shannon TR, Guo T, Bers DM. Ca2+ scraps: local depletions of free [Ca2+] in cardiac sarcoplasmic reticulum during contractions leave substantial Ca2+ reserve. Circ Res. 2003; 93: 40–45.
Franzini-Armstrong C. Structure of sarcoplasmic reticulum. Fed Proc. 1980; 39: 2403–2409. [Order article via Infotrieve]
Wu X, Zhang T, Bossuyt J, Li X, McKinsey TA, Dedman JR, Olson EN, Chen J, Brown JH, Bers DM. Local InsP3-dependent perinuclear Ca2+ signaling in cardiac myocyte excitation-transcription coupling. J Clin Invest. 2006; 116: 675–682.
Bare DJ, Kettlun CS, Liang M, Bers DM, Mignery GA. Cardiac type 2 inositol 1,4,5-trisphosphate receptor: interaction and modulation by calcium/calmodulin-dependent protein kinase II. J Biol Chem. 2005; 280: 15912–15920.
Terasaki M, Slater NT, Fein A, Schmidek A, Reese TS. Continuous network of endoplasmic reticulum in cerebellar Purkinje neurons. Proc Natl Acad Sci U S A. 1994; 91: 7510–7514.
Gerasimenko OV, Gerasimenko JV, Tepikin AV, Petersen OH. Calcium transport pathways in the nucleus. Pflugers Arch. 1996; 432: 1–6. [Order article via Infotrieve]
Hashimoto S, Bruno B, Lew DP, Pozzan T, Volpe P, Meldolesi J. Immunocytochemistry of calciosomes in liver and pancreas. J Cell Biol. 1988; 107: 2523–2531.
Golovina VA, Blaustein MP. Spatially and functionally distinct Ca2+ stores in sarcoplasmic and endoplasmic reticulum. Science. 1997; 275: 1643–1648.
Button D, Eidsath A. Aequorin targeted to the endoplasmic reticulum reveals heterogeneity in luminal Ca2+ concentration and reports agonist- or InsP3-induced release of Ca2+. Mol Biol Cell. 1996; 7: 419–434.
Subramanian K, Meyer T. Calcium-induced restructuring of nuclear envelope and endoplasmic reticulum calcium stores. Cell. 1997; 89: 963–971. [Order article via Infotrieve]
Mogami H, Nakano K, Tepikin AV, Petersen OH. Ca2+ flow via tunnels in polarized cells: recharging of apical Ca2+ stores by focal Ca2+ entry through basal membrane patch. Cell. 1997; 88: 49–55. [Order article via Infotrieve]
Venetucci L, Trafford AW, Eisner DA. Illuminating sarcoplasmic reticulum calcium. Circ Res. 2003; 93: 4–5.
Song LS, Pi Y, Kim SJ, Yatani A, Kudej RK, Zhang Q, Cheng H, Hittinger L, Ghaleh B, Vatner DE, Lederer WJ, Vatner SF. Paradoxical cellular Ca2+ signaling in severe but compensated canine left ventricular hypertrophy. Circ Res. 2005; 97: 457–464.
Bassani JW, Bassani RA, Bers DM. Calibration of indo-1 and resting intracellular [Ca2+]i in intact rabbit cardiac myocytes. Biophys J. 1995; 68: 1453–1460. [Order article via Infotrieve]
Bassani JW, Bassani RA, Bers DM. Twitch-dependent SR Ca2+ accumulation and release in rabbit ventricular myocytes. Am J Physiol. 1993; 265: 533–540.
Forbes MS, Hawkey LA, Jirge SK, Sperelakis N. The sarcoplasmic reticulum of mouse heart: its divisions, configurations, and distribution. J Ultrastruct Res. 1985; 93: 1–16. [Order article via Infotrieve]
Ogata T, Yamasaki Y. High-resolution scanning electron microscopic studies on the three-dimensional structure of the transverse-axial tubular system, sarcoplasmic reticulum and intercalated disc of the rat myocardium. Anat Rec. 1990; 228: 277–287. [Order article via Infotrieve]
Kushmerick MJ, Podolsky RJ. Ionic mobility in muscle cells. Science. 1969; 166: 1297–1298.
Mitchell RD, Simmerman HK, Jones LR. Ca2+ binding effects on protein conformation and protein interactions of canine cardiac calsequestrin. J Biol Chem. 1988; 263: 1376–1381.
Shannon TR, Wang F, Puglisi J, Weber C, Bers DM. A mathematical treatment of integrated Ca2+ dynamics within the ventricular myocyte. Biophys J. 2004; 87: 3351–3371.
Lukyanenko V, Subramanian S, Gyorke I, Wiesner TF, Gyorke S. The role of luminal Ca2+ in the generation of Ca2+ waves in rat ventricular myocytes. J Physiol. 1999; 518: 173–186.
Pogwizd SM, Schlotthauer K, Li L, Yuan W, Bers DM. Arrhythmogenesis and contractile dysfunction in heart failure: roles of sodium-calcium exchange, inward rectifier potassium current, and residual beta-adrenergic responsiveness. Circ Res. 2001; 88: 1159–1167.
Brochet DX, Yang D, Di Maio A, Lederer WJ, Franzini-Armstrong C, Cheng H. Ca2+ blinks: rapid nanoscopic store calcium signaling. Proc Natl Acad Sci U S A. 2005; 102: 3099–3104.
DelPrincipe F, Egger M, Niggli E. Calcium signaling in cardiac muscle: refractoriness revealed by coherent activation. Nat Cell Biol. 1999; 1: 323–329. [Order article via Infotrieve]
the Department of Physiology, Loyola University Chicago, Maywood, Ill.
Abstract
Previous ventricular myocyte studies indicated that ryanodine receptors (RyRs) are in the sarcoplasmic reticulum (SR) and are critical in excitation–contraction coupling, whereas the inositol trisphosphate (InsP3) receptors are separately localized on the nuclear envelope (NucEn) and involved in nuclear Ca2+ signaling. Here, we find that both caffeine and InsP3 receptor agonists deplete free [Ca2+] inside both SR and NucEn. Fluorescence recovery after photobleach (FRAP) was measured using the low-affinity Ca2+ indicator Fluo-5N trapped inside the SR and NucEn (where its fluorescence is high because [Ca2+] is &1 mmol/L). After Fluo-5N photobleach in one end of the cell, FRAP occurred, accompanied by fluorescence decline in the unbleached end with similar time constants (&2 minutes) until fluorescence regained spatial uniformity. Notably, SR and NucEn fluorescence recovered simultaneously in the bleached end. Ca2+ diffusion inside the SR-NucEn was also measured. SR Ca2+-ATPase was completely blocked but without acute SR Ca2+ depletion. Then caffeine was applied locally to one end of the myocyte. In the caffeine-exposed end, free SR [Ca2+] ([Ca2+]SR) declined abruptly and recovered partially (=20 to 30 seconds). In the noncaffeine end, [Ca2+]SR gradually declined with a similar , until [Ca2+]SR throughout the cell equalized. We conclude that the SR and NucEn lumen are extensively interconnected throughout the myocyte. Apparent intrastore diffusion coefficients of Fluo-5N and Ca2+ were estimated (&8 μm2 sec–1 and 60 μm2 sec–1). This rapid luminal communication may maintain homogeneously high luminal [Ca2+], ensuring a robust and uniform driving force for local Ca2+ release events from either SR or NucEn.
Key Words: sarcoplasmic reticulum nuclear envelope (NucEn) fluorescence recovery after photobleach (FRAP) caffeine
Introduction
Cardiac myocytes respond to extracellular stimuli by different signaling pathways. An important mechanism in cardiac myocytes is excitation–contraction coupling (ECC).1 Electrical depolarization of the cell membrane opens L-type Ca2+ channels, causing Ca2+ influx. This Ca2+ entry activates ryanodine receptors (RyRs) on the intracellular Ca2+ store sarcoplasmic reticulum (SR), inducing robust SR Ca2+ release. This local Ca2+ -induced Ca2+ release (CICR)2 is the essence of ECC.
Intracellular Ca2+ stores include SR (endoplasmic reticulum [ER] in nonmuscle cells) and nuclear envelope (NucEn). Free [Ca2+] inside these stores can be &1 mmol/L, which is similar to that in the extracellular space.3 The SR is composed of junctional SR (jSR), which is covered by RyRs and faces toward the T tubules, and free SR (fSR), which contains SR-ER Ca2+-ATPase (SERCA).4 The NucEn surrounds the nucleus and has both an inner and an outer membrane and a space in between where [Ca2+] can be millimolar.
Recently, we found that cardiac myocytes use local Ca2+ release from inositol trisphosphate (InsP3) receptors (InsP3R) in the NucEn to respond to the neurohumoral stimuli in excitation–transcription coupling (ETC).5 Endothelin-1 (ET-1) application caused InsP3 to activate local Ca2+ release from the NucEn via InsP3R, which activated calmodulin and CaMKII to phosphorylate histone deacetylase-5 (causing its nuclear export) and activation of transcription.5 Notably, this pathway could only be activated by Ca2+ from local InsP3R (not global Ca2+ transients), presumably because calmodulin and CaMKII physically associate with the InsP3R at the NucEn.5,6 Indeed, in ventricular myocytes, RyRs are mainly in SR, whereas InsP3R are mainly in NucEn.6 Thus, SR and NucEn differ structurally and functionally and may reflect discrete physically different Ca2+ stores. In addition, the relatively rigid sarcomeric organization in striated muscle with a dense protein mesh at the Z-lines raises the possibility that local SR within a given sarcomere could serve its particular sarcomere but not necessarily communicate with neighboring sarcomeres either longitudinally or transversely. The main aim of the present study was to determine whether different SR regions and NucEn constitute discrete separate Ca2+ stores.
In some cell types (nonmuscle cells), it is generally accepted that the outer nuclear membrane and ER have similar characteristics and outer nuclear membrane is continuous with ER membrane including SERCA, whereas the inner nuclear membrane is quite different from the outer nuclear membrane.7,8 However, it is not known whether the lumen of ER/SR and NucEn is subcompartmentalized or continuous. The existence of separated Ca2+ stores has been proposed by some groups,9–11 but rapid diffusion of Ca2+ and other molecules between the lumen of NucEn and ER was also shown in leukemia cells and fibroblasts.12,13
Here we assess diffusion inside the SR and NucEn using the low-affinity Ca2+ indicator Fluo-5N, fluorescence recovery after photobleach (FRAP) and local caffeine applications. We found evidence that the SR and NucEn are highly interconnected and comprise 1 large continuous intracellular Ca2+ store. This will help to understand ECC and ETC, as well as cardiac arrhythmias, because inhomogeneous SR Ca2+ release may contribute to Ca2+ waves.14,15
Materials and Methods
Myocyte Isolation Fluo-5N Loading and Cell Permeabilization
Myocytes were isolated as previously described16 and loaded with Fluo-5N acetoxymethylester (10 μmol/L) for 2 hours and deesterified for 1.5 hour.3 For intact myocytes, the superfusate contained (in mmol/L) NaCl 140, KCl 4, MgCl2 1, CaCl2 2, HEPES 10, glucose 10 (pH 7.4, 23°C). For permeabilization, myocytes were exposed to relaxing solution (in mmol/L: EGTA 0.1, HEPES 10, K-aspartate 120, free MgCl2 1, ATP 5, reduced glutathione 10, phosphocreatine di-Tris 5 [pH 7.4]) and then permeabilized by saponin (50 μg/mL) for 20 seconds. After permeabilization, we used an internal solution (same as relaxing solution, but free [Ca2+] was adjusted to 100 nmol/L and 8% dextran was added to prevent swelling on permeabilization, with pH 7.2). [Ca2+]SR was calculated using pseudoratio method described previously for single-wavelength indicator.5 F0 corresponds to resting [Ca2+]SR, which was taken to be 1 mmol/L.3
Imaging Under Confocal Microscopy
Images were acquired with a Bio-Rad Confocal Microscope5 (488 nm argon laser excitation, emission at 500 nm long pass). Laser power was 100% for photobleach, 4% for line scan images, and 0.5% for whole cell images. Because the laser power for image recording is low, unintended photobleach was negligible. Confocal imaging used a x40 oil immersion objective and temporal resolution of 166 lines per second. ImageJ software was used for image analysis.
For some FRAP experiments, the pinhole diameter was increased to 12 mm (maximal) to obtain fluorescence from the whole cell thickness. However, in caffeine experiments (where release occurs at all depths) and other experiments, pinhole diameter was chosen for confocality (1.2 mm). All fluorescence data were background subtracted.
Simulation of Fluo-5N Diffusion Inside the Ca2+ Store
The cell/SR-NucEn was considered a circular cylinder, with cross-section area A and length L, divided into 20 equal-volume compartments (L) longitudinally (L=L/20). Diffusion was allowed between neighboring compartments (boundary flux was set to 0). For each compartment, Fluo-5N fluorescence or [Ca2+]SR (C) was determined by Fluo-5N or Ca2+ diffusion from/to the adjacent compartments. The initial curve was from experimental data fitted to sigmoid curve. The change in C in compartment n per unit time (Cn/t) is equation
where D is the apparent diffusion coefficient for Fluo-5N or Ca2+, Vn is the volume of compartment n, and Vn=AL. The above applies to the cylindrical cell structure. SR inside the cell has different structure and accounts for only 3% to 5% of the cell volume.2 The SR cross-section area should also be only 3% to 5% of that of the whole cell. However, because Vn=L·A, equation 1 can be independent of A. equation
Thus, D is determined by concentration gradients and L2/t. Note that D (as defined here) will be reduced by contributions of binding of Ca2+ or Fluo-5N inside the SR-NucEn and also tortuosity of diffusional path (ie, noncylindrical actual geometry of the network). At a time point t, C in compartment n (Cn,t) follows: Cn,t=Cn,t–1+Cn,t–1. Equation 2 was solved numerically with t=0.1 second. The value of D was varied to obtain a least square fit to the measured spatiotemporal Cn,t profiles from experiments.
Statistical Analysis
Data are presented as mean±SEM with significance (P<0.05) determined using unpaired 2-tailed Student t test.
Results
Caffeine or InsP3R Agonists can Deplete [Ca2+]SR in Both SR and NucEn
Rabbit ventricular myocytes were loaded with Fluo-5N and permeabilized with saponin. Fluo-5N is a low-affinity Ca2+ indicator (Kd=400 μmol/L),3 only bright where [Ca2+] is very high, as in SR and NucEn (Figure 1A). We also see irregularly distributed bright spots which are caffeine and InsP3-insensitive (most obvious in Figure 6). The Fluo-5N signal was stable when myocytes were left in skinned fiber solution without any treatment (Figure 1A). This shows that [Ca2+] inside SR and NucEn are stable and that no appreciable Fluo-5N is lost over the experimental time course. Caffeine (10 mmol/L) application causes an immediate decline in Fluo-5N signal both in the SR and NucEn (Figure 1B and 1C), with the SR preceding the NucEn. Figure 1D shows SR and NucEn [Ca]SR depletions during electrically evoked Ca transients (0.5 Hz). Again, [Ca2+]SR declined transiently (Ca2+ scraps) in both SR and NucEn. However, in the NucEn, the depletion amplitude was smaller and delayed in both time to nadir and recovery time constant (rec). These results are consistent with RyR-dependent intra-SR Ca depletions driving those in the NucEn, but they do not prove that the stores are connected. On the other hand, if RyRs are only on the SR (not the NucEn),6 and if the SR and NucEn are not connected, then caffeine and twitches should not cause NucEn Ca2+ depletion.
Application of InsP3 (10 μmol/L) or the potent InsP3R agonist adenophostin (10 μmol/L) for 30 minutes also gradually decreased the Fluo-5N signal in both NucEn and SR in permeabilized myocytes (Figure 2A and 2B and time course of Fluo-5N signal changes in Figure 2C). Note that InsP3R activation produced a very much slower less complete Ca2+ depletion of the SR and NucEn. This is consistent with the much lower number of InsP3R (versus RyR) in ventricular myocytes (and lower intrinsic Ca2+ flux rate).2 That is, the very high Ca2+ flux rate via caffeine-activated RyRs completely depletes [Ca2+]SR (despite continued SERCA function). In contrast, the much lower InsP3R flux rate only slowly depletes [Ca2+]SR and results in a new steady state where the SERCA pump and leak rates rebalance at a new low [Ca2+]SR. This explains why the lack of time delay between SR and NucEn signals (similar values), because the InsP3-induced leak rate may be small compared with potentially rapid NucEn-SR Ca2+ equilibration.
Again, if InsP3R are only on the NucEn (not SR) and the 2 pools are not interconnected, this result is unexpected.6 This raises the possibility that the SR and NucEn lumens are highly connected. However, the above phenomena could also happen even if the SR and NucEn were not connected but if there were very small numbers of RyRs on the NucEn and of InsP3R on the SR. These could be functionally effective even if their existence is only near the detection limit of immunocytochemistry, fractionation methods, or Western blot analyses.6 To test whether the lumen of the SR is connected to both the NucEn and distant regions of the SR, we performed the following FRAP and caffeine-induced [Ca2+]SR depletion experiments.
Fluo-5N Diffusion Inside the SR and NucEn: FRAP Studies
Permeabilized, Fluo-5N–loaded myocytes were photobleached on approximately half of the cell longitudinally with the other half left almost unaffected (Figure 3A). Because of dispersion and scatter of illumination, the beam was centered near the left cell end, causing an initial gradient of bleach that was maximal at the end of the cell but gradually declining to be negligible at the cell center (Figure 3A and 3B). Fluorescence in the bleached half (left) decreased immediately on bleach. After bleach, the SR and NucEn fluorescence increased gradually in the bleached half, and simultaneously there was a complementary gradual decrease of fluorescence in the unbleached half (right) until the 2 halves reached the same level (full FRAP; Figure 3A, with longitudinal profile in Figure 3B). On Fluo-5N photobleach, local [Ca2+]SR gradients are neither likely to be significant nor to complicate subsequent FRAP measurements. This is partly because the intrastore Fluo-5N concentration is very low compared with physiological SR Ca2+ buffers (eg, calsequestrin; see Figure II in the online data supplement, available at http://circres.ahajournals.org). Rather, subsequent FRAP is attributable almost entirely to the diffusion of Fluo-5N within the SR and NucEn. Note that in these and subsequent longitudinal profiles, the regular sarcomeric pattern of jSR (with 1.9-μm spacing3; see Figure 3B inset) was intentionally eliminated by smoothing.
Figure 3C shows the time course of FRAP in the bleached half and the decay of fluorescence in the unbleached half. The total fluorescence after bleach (&15% overall bleach) did not change during FRAP, indicating that these measurements are not complicated by changes in [Ca2+]SR. The time constant for FRAP on the left and fluorescence decline on the right were similar (=112.1±9.8 and 115.3±5.7 seconds). The NucEn on the left also demonstrated FRAP, with similar kinetics as the SR (=119.8±7.6 seconds).
This FRAP experiment was also done in the transverse direction (Figure 4). In this image, the upper side of the cell was bleached and transverse FRAP and complementary fluorescence decline in the bottom occurred. The results are qualitatively similar for transverse versus longitudinal FRAP, except that the time constants were faster transversely (&15 seconds) than longitudinally (&110 seconds). This is not surprising because the distance is much less in the transverse direction. Both longitudinal and transverse FRAP were complete and spatially uniform. Because these were permeabilized myocytes and Fluo-5N was trapped inside the SR and NucEn, the recovery of the bleached half can only be attributable to the Fluo-5N diffusion inside the Ca2+ store. In Figure 3A and 3C, both NucEn and SR on the left recover during FRAP, evidence of extensive connection of the SR with both the NucEn and with more distant SR regions.
Because the images in Figures 3 and 4 were confocal, the recovery of fluorescence on the bleached half could be partially attributed to Fluo-5N diffusion from regions above and below the confocal plane (in addition to the other half in the x-y plane). On the other hand, bleach in the Z direction is likely to be substantial (as it is nonconfocal). Nevertheless, to avoid FRAP from above and below the confocal plane, the signal from the whole thickness of the cell was obtained by opening the confocal pinhole maximally (12 mm versus 1.2 mm in confocal images). As in Figure 3, we bleached the left half of the cell (including 1 nucleus), whereas the right half remained relatively unaffected (Figure 5A). The images are fuzzier because of intentional nonconfocality. The fluorescence on the left half dropped immediately by 30% (Figure 5B). The degree of photobleach was only 30% because of the brief laser exposure (less than 1 second to limit cell damage and optimize temporal resolution). FRAP is apparent in both SR and NucEn on the bleached half, accompanied by a gradual and complementary decline on the unbleached half. We compared FRAP kinetics by choosing region of interest (ROI). The values for the recovery of both halves were similar, &2 minutes (Figure 5B). The amplitudes of the post-bleach decrease and increase of fluorescence in the 2 halves were also about the same. The post-bleach total cell fluorescence was constant (Figure 5B). Again, because this was in a permeabilized cell and the only source of Fluo-5N was from the SR and NucEn, this result means that Fluo-5N diffuses from one end of the cell to the other inside the SR-NucEn network. If Fluo-5N came out of the cell, it would diffuse away and we would see loss of total fluorescence. This indicates that the Ca2+ stores of the 2 halves are connected.
At the same longitudinal location, the bleach and recovery of the SR and NucEn were also compared for different ROI (Figure 5C). The SR and NucEn regions were bleached to the same extent and recovered with a similar time constant. Because the recovery of NucEn can only be via Fluo-5N diffusion from the SR, this demonstrates that the SR and NucEn are extensively connected. A region of SR closer to the bleached end of the cell had more extensive bleach (60% versus 30% on average for this half) and also recovered with a slightly longer (see simulations below). The similar FRAP results in Figures 3 and 5 confirm control experiments, indicating photobleach through most of the cell depth.
Ca2+ Diffusion Inside the SR and NucEn
We also performed caffeine-induced [Ca2+]SR depletion at one end of the cell to assess Ca2+ diffusion inside the SR and NucEn. This was done in intact cells under special conditions, previously developed and validated,17 where SERCA could be completely blocked but SR stores maintained until a release is triggered. Intact myocytes were paced at 0.5 Hz for 10 beats to preload Ca2+ stores. Then pacing was stopped and perfusion was switched for 90 seconds to a Na+-free, Ca2+-free NT (Li+ replaced Na+ and 10 mmol/L EGTA was included) with 5 μmol/L thapsigargin (TG) (to block the SERCA pump). Using this protocol, [Ca2+]SR was not decreased significantly (Figure 6A and 6B). However, SERCA was completely blocked because, after acute [Ca2+]SR depletion by caffeine (Figure 6C), caffeine washout and then field stimulation [Ca2+]SR could not be detectably refilled (Figure 6D versus 6C). After treatment with TG, the SR Ca2+ content does eventually decline (even in 0 Na+/0 Ca2+ solution), but there is a practical time window (&5 minutes) in which these experiments can be done. Note that the caffeine-insensitive bright spots were also unaffected throughout the procedure.
In cells pretreated with TG 0 Na+/0 Ca2+, as above, we evaluated intra-SR Ca2+ diffusion using local caffeine application. Figure 7A shows local caffeine application at the left end of the cell briefly (1 to 2 seconds) to release Ca2+ from SR and NucEn at this end. There is an acute discernable [Ca2+]SR gradient from the left end to the middle (Figure 7A and 7B). Because SERCA was completely blocked, the only source of Ca2+ recovery at the left end is Ca2+ diffusion within the Ca2+ store from the right. If this recovery happens, SR and NucEn must be connected throughout the myocyte. Indeed, [Ca2+]SR recovery at the left and right ends occurred with similar values (Figure 7C). Additionally, the total [Ca2+]SR inside the Ca2+ store was constant, indicating acute SR Ca2+ release, no net reuptake and no further loss (Figure 7C). In addition, the recovery of SR and NucEn occur simultaneously with similar values (Figure 7D). This again implies that SR and NucEn are highly connected via their lumens.
Simple Model of Fluo-5N Diffusion Inside the SR and NucEn
To further investigate Fluo-5N and Ca2+ diffusion inside the SR and NucEn, we made a simple 1D diffusion model (Figure 8A). Figure 8 (B and E) shows exemplar spatiotemporal data, smoothed by fitting to sigmoid curves (Fluo-5N and Ca2+ diffusion, respectively). C and F in Figure 8 are the corresponding simulated curves. The initial spatial profile (t=0) was set to the smoothed experimental data, and then diffusion of Fluo-5N or Ca2+ was calculated by the diffusion equation with D adjusted to best fit the data. The apparent diffusion coefficients of Fluo-5N (DFluoStore) and Ca2+ (DCaStore) inside the Ca2+ store were estimated at 8 μm2 sec–1 and 60 μm2 sec–1, respectively.
D and G in Figure 8 are theoretical time courses of Fluo-5N or Ca2+ recovery at different longitudinal sites, with values shown in the insets. The value is a U-shaped function with the fastest near the bleach interface and the longest at the ends. This is consistent with experimental data in Figure 5C, where we saw that a site farther to the left exhibited a longer FRAP .
Discussion
The SR and NucEn Are Highly Interconnected in Cardiac Ventricular Myocytes
The data here measuring both Fluo-5N FRAP and Ca2+ diffusion inside the SR, directly demonstrate for the first time that the SR lumen is highly interconnected with the NucEn and also with distant regions of the SR. Moreover, the SR-NucEn throughout the myocyte appears to be a single large continuous Ca2+ storage compartment. Our initial expectation was that the SR within each sarcomere would be relatively isolated from other SR regions. On the other hand, occasional electron micrographs show that jSR can connect SR from 1 side of the Z-line to the other (and to transverse sarcomeres as well),4,18,19 but there has been no prior functional data to indicate the extent to which the SR network is continuous. Clearly it is. Similarly, there have been reports that the outer nuclear membrane is continuous with the ER in some cell types, but no prior functional evidence in cardiac myocytes. Indeed, the apparent difference in Ca2+ release channels (RyR in SR and InsP3R in NucEn) and functional roles (ECC for SR and ETC for NucEn) might lead one to expect that the pools are separate. Clearly this is not the case. These processes and release channels are fueled by the same Ca2+ pool. We were surprised how rapid diffusion appears to be within the SR-NucEn network.
Intra–SR-NucEn Diffusion
In our simple simulation, the estimated values for DCaStore and DFluoStore were 60 μm2 sec–1 and 8 μm2 sec–1, respectively (ie, 7.5 times faster for Ca2+). Because D is inversely related to the square root of molecular weight, the expected ratio is 5.3 times higher D for Ca2+ versus Fluo-5N. Given the many factors that could complicate our simplistic analysis, this is in rather good agreement. The estimated DCaStore is approximately 10-fold slower than expected in aqueous solution (700 μm2 sec–1).20 Main factors that would reduce DCaStore are Ca2+ binding to fixed sites, high viscosity, and tortuosity of the diffusional path. We favor tortuosity as a dominant factor because: (1) both DCaStore and DFluoStore seem reduced roughly in parallel; (2) path-length tortuosity is expected and would affect both similarly; and (3) because intra-SR buffering of Ca2+ and Fluo-5N are likely very different. Indeed, images of the SR network structure would suggest substantial tortuosity. Of course Ca2+ also binds to intra-SR buffers like calsequestrin, but the low-affinity (Kd=0.5 mmol/L)21 and high off-rate would limit the impact on DCaStore. Notably, our DCaStore of 60 μm2 sec–1 is much faster than the apparent DCa estimated in myoplasm (1.4 μm2 sec–1).20 Whereas tortuosity and viscosity may contribute to the low apparent DCa in cytosol, Ca2+ binds to very many, relatively high-affinity (slow off-rate) fixed Ca2+ binding sites in the cytosol (eg, troponin C, SERCA, calmodulin, myosin).2,22 Thus, Ca2+ binding may be the major limitation for Ca2+ diffusion in myoplasm, versus tortuosity inside the SR. The faster Ca2+ diffusion inside the SR may also move more Ca2+ because [Ca2+]SR is 10 000 times higher than in cytosol. Fast intra-SR Ca2+ diffusion may also be important to consider in analyzing propagating waves of CICR in muscle and other cell types.
Ca2+ Wave Initiation and Propagation
Ca2+ waves are thought to initiate when [Ca2+]SR is elevated (perhaps locally), and this initiates a local SR Ca2+ release event (Ca2+ spark).2,23 The wave then propagates via CICR, as released Ca2+diffuses to the next junction where it triggers local SR Ca2+ release and so on. Note that local [Ca2+]SR at the next site must be sufficiently high for RyRs to be activated by the lower local [Ca2+]i (<1 μmol/L) than occurs via local Ca2+ current during normal ECC (>10 μmol/L).2 The rapid equilibration within the SR-NucEn described here would help to keep 1 area from experiencing a higher local [Ca2+]SR (ie, the network buffers local [Ca2+]SR elevation and depletion), thus potentially limiting initiating events for diastolic Ca2+ waves. It may also interfere with wave propagation, if local [Ca2+]SR ahead of the initiating wave could decrease (because of release at the previous region) before release at that next junction is activated by high local [Ca2+]i. That is, lower local [Ca2+]SR could desensitize local RyRs at the next site. Both of these effects might decrease the likelihood of wave initiation and propagation. Of course, these ideas need future experimental tests.
Local elevation of [Ca2+]SR was suggested as a potential mechanism of arrhythmogenesis in hypertrophic heart,15 causing loci prone to wave initiating. Given the present results, this scenario seems unlikely, unless the SR is more fragmented in hypertrophy or heart failure. In preliminary experiments in a rabbit heart failure model,24 we did not uncover any fragmentation of the SR-NucEn using the same FRAP approach used here (not shown). This may merit further exploration in pathophysiological models.
Local SR Ca2+ Depletion During SR Ca2+ Release
Rapid Ca2+ diffusion within the SR (along with buffering by calsequestrin) can enhance SR Ca2+ release through the jSR by minimizing the local SR Ca2+ depletion, which occurs during ECC. Indeed, the first direct measurements of [Ca2+]SR and Ca2+ scraps during normal ECC showed that there was no detectible delay in the time course of local [Ca2+]SR decline in fSR versus jSR.3 This has been confirmed in subsequent measurements,25 consistent with rapid intra-SR diffusion within a sarcomere. Local photolytically induced CICR was shown to have an extremely short refractoriness compared with global ECC,26 and this was ascribed to very rapid repletion of those local SR Ca2+ release sites by neighboring SR regions. Brochet et al25 observed local isolated [Ca2+]SR depletion during Ca2+ sparks (called Ca2+ blinks). They took the detectability of these events as evidence of restricted fSR-jSR diffusion. However, both the extent of [Ca2+]SR depletion and recovery time constant were much smaller for Ca2+ blinks versus Ca2+ scraps or global SR Ca2+ release. These results, we think are consistent with very rapid Ca2+ replenishment from neighboring SR regions as we show here throughout the whole cell.
Our measurements here are over a spatial scale that makes [Ca2+]SR gradients readily detectable. This places a lower limit on DCaStore (60 μm sec–1) within a single sarcomere. We scaled down our simple 1D model to a half-sarcomere. We decreased [Ca2+]SR instantaneously from 1 to 0.5 mmol/L only in the junctional 50-nm compartment (corresponding to a maximal local jSR release). With our DCaStore value of 60 μm2 sec–1, the [Ca2+]SR gradient throughout that 1-μm half-sarcomere within 2 ms is less than 3% (0.96 to 0.99 mmol/L) versus 25% if we had used DCa reported for myoplasm (1.4 μm2 sec–1; see supplemental Figure I). These results are consistent with the lack of detectable jSR-fSR [Ca2+]SR gradient during Ca2+ scraps. This simple model is a starting point for further considerations.
Uniform Ca2+ Distribution Is Physiologically Significant
The spatially uniform [Ca2+]SR provides a spatially uniform driving force for SR Ca2+ release throughout the myocyte during ECC. This may be important in producing homogeneous Ca2+ release and contractile activation throughout the ventricular myocyte. The network also ensures fidelity of local SR or NucEn Ca2+ release events involved in local Ca2+ signaling. That is, the entire network of SERCA pumps in the SR-NucEn help maintain this [Ca2+] gradient for all its uses. The results in Figure 2 show that InsP3 or adenophostin reduces [Ca2+]SR everywhere, even if the primary functional targets and locus of release are at the NucEn. The incomplete [Ca2+]SR depletion indicates that the leak generated by strong InsP3R activation is less than the overall forward SERCA pumping rate at 100 nmol/L [Ca2+]i (&20 μmol/L cytosol per second) or more than 100 times slower than SR Ca2+ release via RyR during a twitch or caffeine-induced Ca2+ transient).2 A potential cost of this rapid equilibration is that any pathophysiological SR Ca2+ leak could deplete [Ca2+]SR everywhere in the cell and limit both RyR and InsP3R-mediated Ca2+ release.
In conclusion, the cardiac SR is highly interconnected to both NucEn and more distant SR throughout the entire myocyte. Fast Ca2+ diffusion within this SR-NucEn store may stabilize local free [Ca2+]SR to provide uniform driving force for RyR and InsP3R. Homogeneous [Ca2+]SR may ensure uniform and synchronous SR Ca2+ release and contractile activation.
Acknowledgments
We thank Drs Eckard Picht and Fei Wang for helpful discussions and Jayme O’Brien and Karl Hench for technical assistance.
Sources of Funding
Supported by NIH grants HL30077 and HL64724 (D.M.B.) and an American Heart Association Fellowship (to X.W.).
Disclosures
None.
Footnotes
Original received April 19, 2006; revision received May 31, 2006; accepted June 13, 2006.
References
Bers DM. Cardiac excitation-contraction coupling. Nature. 2002; 415: 198–205. [Order article via Infotrieve]
Bers DM. Excitation-Contraction Coupling and Cardiac Contractile Force. 2nd ed. Dordrecht, The Netherlands: Kluwer Academic Publishers; 2001.
Shannon TR, Guo T, Bers DM. Ca2+ scraps: local depletions of free [Ca2+] in cardiac sarcoplasmic reticulum during contractions leave substantial Ca2+ reserve. Circ Res. 2003; 93: 40–45.
Franzini-Armstrong C. Structure of sarcoplasmic reticulum. Fed Proc. 1980; 39: 2403–2409. [Order article via Infotrieve]
Wu X, Zhang T, Bossuyt J, Li X, McKinsey TA, Dedman JR, Olson EN, Chen J, Brown JH, Bers DM. Local InsP3-dependent perinuclear Ca2+ signaling in cardiac myocyte excitation-transcription coupling. J Clin Invest. 2006; 116: 675–682.
Bare DJ, Kettlun CS, Liang M, Bers DM, Mignery GA. Cardiac type 2 inositol 1,4,5-trisphosphate receptor: interaction and modulation by calcium/calmodulin-dependent protein kinase II. J Biol Chem. 2005; 280: 15912–15920.
Terasaki M, Slater NT, Fein A, Schmidek A, Reese TS. Continuous network of endoplasmic reticulum in cerebellar Purkinje neurons. Proc Natl Acad Sci U S A. 1994; 91: 7510–7514.
Gerasimenko OV, Gerasimenko JV, Tepikin AV, Petersen OH. Calcium transport pathways in the nucleus. Pflugers Arch. 1996; 432: 1–6. [Order article via Infotrieve]
Hashimoto S, Bruno B, Lew DP, Pozzan T, Volpe P, Meldolesi J. Immunocytochemistry of calciosomes in liver and pancreas. J Cell Biol. 1988; 107: 2523–2531.
Golovina VA, Blaustein MP. Spatially and functionally distinct Ca2+ stores in sarcoplasmic and endoplasmic reticulum. Science. 1997; 275: 1643–1648.
Button D, Eidsath A. Aequorin targeted to the endoplasmic reticulum reveals heterogeneity in luminal Ca2+ concentration and reports agonist- or InsP3-induced release of Ca2+. Mol Biol Cell. 1996; 7: 419–434.
Subramanian K, Meyer T. Calcium-induced restructuring of nuclear envelope and endoplasmic reticulum calcium stores. Cell. 1997; 89: 963–971. [Order article via Infotrieve]
Mogami H, Nakano K, Tepikin AV, Petersen OH. Ca2+ flow via tunnels in polarized cells: recharging of apical Ca2+ stores by focal Ca2+ entry through basal membrane patch. Cell. 1997; 88: 49–55. [Order article via Infotrieve]
Venetucci L, Trafford AW, Eisner DA. Illuminating sarcoplasmic reticulum calcium. Circ Res. 2003; 93: 4–5.
Song LS, Pi Y, Kim SJ, Yatani A, Kudej RK, Zhang Q, Cheng H, Hittinger L, Ghaleh B, Vatner DE, Lederer WJ, Vatner SF. Paradoxical cellular Ca2+ signaling in severe but compensated canine left ventricular hypertrophy. Circ Res. 2005; 97: 457–464.
Bassani JW, Bassani RA, Bers DM. Calibration of indo-1 and resting intracellular [Ca2+]i in intact rabbit cardiac myocytes. Biophys J. 1995; 68: 1453–1460. [Order article via Infotrieve]
Bassani JW, Bassani RA, Bers DM. Twitch-dependent SR Ca2+ accumulation and release in rabbit ventricular myocytes. Am J Physiol. 1993; 265: 533–540.
Forbes MS, Hawkey LA, Jirge SK, Sperelakis N. The sarcoplasmic reticulum of mouse heart: its divisions, configurations, and distribution. J Ultrastruct Res. 1985; 93: 1–16. [Order article via Infotrieve]
Ogata T, Yamasaki Y. High-resolution scanning electron microscopic studies on the three-dimensional structure of the transverse-axial tubular system, sarcoplasmic reticulum and intercalated disc of the rat myocardium. Anat Rec. 1990; 228: 277–287. [Order article via Infotrieve]
Kushmerick MJ, Podolsky RJ. Ionic mobility in muscle cells. Science. 1969; 166: 1297–1298.
Mitchell RD, Simmerman HK, Jones LR. Ca2+ binding effects on protein conformation and protein interactions of canine cardiac calsequestrin. J Biol Chem. 1988; 263: 1376–1381.
Shannon TR, Wang F, Puglisi J, Weber C, Bers DM. A mathematical treatment of integrated Ca2+ dynamics within the ventricular myocyte. Biophys J. 2004; 87: 3351–3371.
Lukyanenko V, Subramanian S, Gyorke I, Wiesner TF, Gyorke S. The role of luminal Ca2+ in the generation of Ca2+ waves in rat ventricular myocytes. J Physiol. 1999; 518: 173–186.
Pogwizd SM, Schlotthauer K, Li L, Yuan W, Bers DM. Arrhythmogenesis and contractile dysfunction in heart failure: roles of sodium-calcium exchange, inward rectifier potassium current, and residual beta-adrenergic responsiveness. Circ Res. 2001; 88: 1159–1167.
Brochet DX, Yang D, Di Maio A, Lederer WJ, Franzini-Armstrong C, Cheng H. Ca2+ blinks: rapid nanoscopic store calcium signaling. Proc Natl Acad Sci U S A. 2005; 102: 3099–3104.
DelPrincipe F, Egger M, Niggli E. Calcium signaling in cardiac muscle: refractoriness revealed by coherent activation. Nat Cell Biol. 1999; 1: 323–329. [Order article via Infotrieve]
您现在查看是摘要介绍页，详见ORG附件。