Excitation-Contraction Coupling Changes during Postnatal Cardiac Development

Abstract

Cardiac contraction is initiated by the release of Ca²⁺ from intracellular stores in response to an action potential, in a process known as “excitation-contraction coupling” (ECC). Here we investigate the maturation of ECC in the rat heart during postnatal development. We provide new information on how proteins of the sarcoplasmic reticulum (SR) and the t-tubules (TTs) assemble to form the structures that support EC coupling during postnatal development. We show that the surface membrane protein, caveolin-3 (Cav3), is a good protein marker for TTs in ventricular myocytes and compared it quantitatively to junctophilin-2 (JP2), a protein found on the SR at sites of SR-TT junctions, or couplons. Although JP2 and Cav3 associate primarily with the SR and TTs, respectively, we found that, they occupy the appropriate sites at maturing structures in synchrony, as visualized with high resolution, quantitative 3-dimensional imaging. We also found the surprising result that while both ryanodine receptor type 2, (RyR2) and JP2 proteins are localized to the same membrane and sub-compartments, they assume their positions at very different rates: RyR2 moves to the SR membrane at the Z-disc very early in development while JP2 only appears in the SR membrane as the TTs mature. Our data suggest that, although RyR2 appears to be prepositioned at the sites ultimately occupied by dyad junctions, JP2 arrives at these sites in synchrony with the development of the TTs at the Z-discs. Finally, we report that EC coupling efficiency changes with development, in concert with these structural changes. Thus we provide the first well-integrated information that links the developing organization of proteins underlying EC coupling (RyR2, DHPR, Cav3 and JP2) to the developing efficacy of EC coupling.

Introduction

Cardiac contraction arises as the cardiac electrical signal, or action potential (AP), spreads throughout the heart and triggers an elevation of calcium ([Ca²⁺]_i) which in turn triggers contraction. The components of this excitation-contraction coupling (ECC) process are under tight spatial and temporal control and mature during development ^1-6. Two central elements of the ECC cascade that change with development and disease are the transverse tubules (TTs) and the sarcoplasmic reticulum (SR) ^7-11. These structures, which underlie ECC, may also de-differentiate in diseased hearts ^{12, 13}.

In mammalian ventricular myocytes, the TT system is incomplete or absent at birth ^{3, 5, 6, 14}. The appearance of the TT system seems to have both a developmental and evolutionary significance. In ventricular myocytes from birds or fish, the cells are remarkably thin (about 5 microns diameter or less) and have no TTs ^15-18. This makes them similar to atrial myocytes of many adult mammals. The TT system in adult mammalian ventricular myocytes is a specialized system of sarcolemmal invaginations which contain ~80% of the total cellular pool of L-type Ca²⁺ channels (dihydropyridine receptors or DHPRs) ^{19, 20}. The junctional sarcoplasmic reticulum (jSR), which contains the SR Ca²⁺ release channels of cardiac muscle (ryanodine receptors, RyR2s), is, localized within 15 nm of the regions of the TTs containing DHPRs in adult ²¹. The functional units that include DHPRs at the TTs and RyR2s at the jSR form functional units known as couplons ²². The opening of even one DHPR at a negative membrane potential in a couplon can trigger Ca²⁺ release from a nearby cluster of RyR2s ^{23, 24}. The activation of a couplon produces a Ca²⁺ spark; the synchronization of many Ca²⁺ sparks by an AP produces the [Ca²⁺]_i transient that activates contraction ²⁵. Approximately 80-85% of Ca²⁺ sparks occur within 0.5 microns of the TTs ^{26, 27}.

Here we examine the development of rat ventricular myocytes from postnatal day 10 (d10) to maturity, with the goal of understanding the morphological changes that underlie the development of the apparatus that mediates ECC under normal conditions, and the malfunctions that occur in heart disease. D10 cardiomyocytes lack a TT system and so are structurally immature, compared to adult myocytes, but they develop rapidly, so, by d20 the TT system is indistinguishable from that of adult myocytes. Even at d10, when the TT system is beginning to develop, RyR2s are aligned at the Z-discs. The time-sequence of postnatal development and subcellular organization of ECC proteins is presented here. The precise juxtaposition of key ECC proteins underlies increasing signaling efficiency, as measured by both the peak of triggered [Ca²⁺]_i and fractional release of SR [Ca²⁺].

Methods

Developmental Cell Isolation

Cardiac myocytes were isolated from postnatal rats at days 10, 15, or 20 following a protocol modified from Isenberg and Klockner ²⁸. Briefly, animals were heparinized and anesthetized with pentobarbital prior removal of the hearts by thoracotomy. Once removed, hearts were rapidly bathed in ice cold Ca²⁺-free Tyrode’s solution (containing (in mmol/L) 140 NaCl, 0.5 MgCl₂, 0.3 NaH₂PO₄, 5 HEPES, 5.5 glucose, 5 KCl, 0.5 EGTA, pH 7.4) and the aortas were cannulated for Langendorff perfusion. Perfusion to remove blood from the ventricles and circulatory system was at 37°C for 1min with neonatal buffer (NB) solution (containing (in mmol/L) 20 taurine, 70 K-glutamate, 5 pyruvic acid, 40 KCl, 25 HEPES, 1 MgSO₄, 10 KH₂PO₄, 22 glucose, pH 7.4) supplemented with 0.1mM EGTA. Next, hearts were digested for 7-12 min with NB + 0.5-1 mg/mL collagenase type 2 (Worthington) and then perfused again with NB + 0.1 mM EGTA for 2 min to remove any residual collagenase. The ventricles were then removed and gently minced in NB + 0.1 mM EGTA. The tissue pieces were triturated to dissociate the myocytes and the cell suspension was filtered through 200 μm nylon mesh. The isolated cells were kept in NB + 1 mM EGTA for 1hr at 4°C to allow them to recover prior to plating.

Adult Cell Isolation

Adult myocytes were isolated from rats at least 8 weeks old, as described ⁹.

T-tubule Imaging

The sarcolemma and t-tubules (when present) were visualized with the lipophilic fluorescent indicator, Di-8-ANEPPS (Invitrogen Corp. Carlsbad, California). The indicator was initially dissolved to 10 mM in DMSO and applied to freshly isolated myocytes at a final concentration of 10 μM for 5 min at room temperature. Following incubation, the indicator was immediately washed out and the cells were observed by confocal microscopy. Z-stacks were reconstructed into 3-dimensional projections with Volocity version 4.0 (Improvision, Inc.).

Immunofluorescence and Analysis

Freshly isolated myocytes were fixed in 2% paraformaldehyde for 15 min and then washed every 15 min for a total of 4 washes in PBS. Myocytes were incubated in a solution of 1% bovine serum albumin (BSA), 1% normal goat serum (NGS), and 1 mg/mL saponin in phosphate buffered saline (PBS) for 1hr at RT. Cells were incubated with primary antibodies in 1% BSA, 1% NGS, in PBS overnight at 4°C, then washed 4 times every 30 min with 1% BSA, 1% NGS, in PBS, and incubated in secondary antibodies in the same solution for 2 hr at room temperature. After a final series of washes (every 30 min for 3 washes in 1% BSA, 1% NGS in PBS, then 2 times 15 min each in PBS), samples were mounted on glass slides in Slow Fade Lite (Invitrogen Corp.). Negative controls were produced by the same protocol without the use of a primary antibody.

Primary antibodies include monoclonal anti-caveolin 3 (Transduction Labs/BD Biosciences, San Jose, CA), monoclonal anti-ryanodine receptor 2 (Affinity BioReagents, Golden, CO), and polyclonal anti-junctophilin 2 (Zymed, San Francisco, California). Secondary antibodies used here were goat anti-mouse Alexa488 and goat anti-rabbit Alexa633, both from Invitrogen.

Analysis of images for colocalization of pairs of marker proteins was completed using Volocity version 4.0. To compare two proteins in an X-Y image, quantifications of overlap were completed by generating a scatter plot of pixel intensity with one channel (representing one protein) on each axis. After linear regression analysis, a Pearson’s correlation (PC) coefficient was calculated to describe the colocalization or variance of the two proteins being tested ²⁹. More detailed analysis of colocalization was completed by calculating a Manders coefficient for each protein to be localized. This analysis calculates the coefficient of colocalization for each specific marker in the image to quantify the contribution of each marker towards voxels (3-D pixel volumes) containing both markers ³⁰. Briefly, the software records the fluorescence of each marker in a given volume (voxel by voxel) and compares the colocalized fluorescence to the total fluorescence. These values provide the fraction of colocalized protein divided by total protein in a given volume, or the portion of total fluorescence from a single marker that is found in colocalized voxels.

Ca²⁺ Recording

Freshly isolated myocytes were loaded at room temperature for 30 min with 10μM Fluo-4 AM (stock in 20% pleuronic F127/DMSO). Dye-loaded cells were then perfused with normal Tyrode’s solution and voltage clamped in the whole-cell configuration with a HEKA EPC-9 amplifier. The internal bathing solution consisted of (in mmol/L) 130 CsCl, 5 MgATP, 10 HEPES, and 20 TEA. The extracellular solution during patch experiments contained (in mmol/L) 140 NaCl, 0.5 MgCl₂, 0.3 NaH₂PO₄, 5 HEPES, 5.5 glucose, 1.8 CaCl₂, and 5 CsCl. Once access to the cytoplasm was achieved, the cells were maintained at −80 mV until the protocol called for changes in voltage. SR load was normalized with a voltage protocol including 5 voltage steps from −80 mV to 0 mV at 1 Hz. Voltage-gated Na⁺ channels were inactivated by a ramp from −80mV to −50 mv. From −50 mV, 200 ms voltage steps in 10 mV increments were conducted with simultaneous recording of cell fluorescence.

Imaging

All image acquisition was completed on a Zeiss LSM 510 laser scanning confocal microscope (Zeiss, Inc., Poughkeepsie, NY) with a 63 x water objective and excitation lasers at 488 nm and 633 nm wavelengths. During the acquisition of all two-color images, the multitrack mode was utilized to reduce bleed-through of emitted fluorescence. All X-Y images were captured with a pixel size of no larger than 0.1 μm × 0.1 μm. Z-stack images were recorded with the same X-Y pixel size and a z-step of 0.2 μm.

Statistics

Statistical significance was assessed by ANOVA analysis of all data points with pair-wise comparisons completed as a post-test, with significance defined as a P-value <0.05. Data points represent mean ± SE.

Results

We used postnatal rat ventricular myocytes (d10 – adult) to examine the ontogeny of ECC. Three aspects of the process of development of ECC are relevant to this work: 1. formation of TTs; 2. apposition of TTs and jSR; and 3. maturation of ECC. This work investigates how the alignment of TTs and jSR occurs during development, and underlies the changes in ECC that accompany it.

T-Tubule Formation during Postnatal Development

Although TTs are absent from newborn rat cardiac myocytes ³¹, they can be readily visualized during development with lipophilic indicators such as di-8-ANEPPS ³². Using this dye, we obtained high resolution confocal “z-stack” images of myocytes from rat hearts of each age studied (d10 to adult), to create 3-dimensional representations of t-tubule structures. These images show that the TTs appear to sprout inwards from the SL membrane during development. At day 10 (Fig.1 A, B, C), TTs are barely visible as small membranous invaginations near the cell surface. By day 15 (Fig. 1 D, E, F), TTs can be found throughout the cell volume, but they are sparse and do not appear as punctate elements at the Z-disc, as they do in more mature, d20 (Fig.1 G, H, I) and adult (Fig.1 J, K, L) myocytes.

Figure 1. T-tubule Formation and Alignment in Developing Myocytes.

The sarcolemma and t-tubules of freshly isolated myocytes were labeled with Di-8-ANEPPS. Single X-Y images of myocytes from postnatal days 10 (A), 15 (C), 20 (E), and adult (G) show an increase in labeling of invaginations of the sarcolemma, identified in subsequent experiments as TTs (white scale bars are 10 μm). 3-Dimensional reconstruction of z-stacks (B, D, F, and H) from the same cells emphasizes the changes in the density of these invaginations and the changes in their organization that occur during development. Black scale bars above each image represent 5 μm.

Couplon Assembly in Developing Myocytes

A couplon is the elementary unit of ECC, composed of one or more DHPRs and a cluster of RyR2 (type 2 ryanodine receptors) that assemble at the junction of a TT and a jSR. To establish and quantify their formation in developing cardiomyocytes, we used antibodies to three proteins that represent distinct components of couplons; caveolin 3 (Cav3), a TT marker; RyR2, a jSR marker, and junctophilin-2 (JP2) a marker of the TT-SR junction Cav3, the primary cardiac isoform of caveolin, associates with cholesterol-rich regions of the surface membranes, termed caveolae, and associates with diverse structural, channel and regulatory proteins ^33-36. Cav3 may also play an important role in organizing and maintaining the TT system in cardiac myocytes ^{37, 38}. JP2 is thought to span the space lying between the TT and the jSR, to connect the two membranes ^{39, 40}. Colocalization of Cav3 with JP2 (Fig. 2) or RyR with JP2 (Fig. 3) in myocytes of different postnatal ages provided a measure of the changing relationship between the TTs and jSR as myocytes mature.

Figure 2. Localization of Caveolin-3 and Junctophilin-2 in Developing Cardiac Myocytes.

Ventricular myocytes were fixed and co-immunostained for caveolin-3 (Cav3) and junctophilin-2 (JP2) at postnatal days 10 (A, B), 15 (D, E), and 20 (G, H) and adult (J, K). Overlay images (F, I, L) provide a comparative view of Cav3 (green) and JP2 (red) protein distribution. Scale bars represent 10 μm.

Figure 3. Localization of Ryanodine Receptor-2 and Junctophilin-2 in Developing Cardiac Myocytes.

Ventricular myocytes were fixed and co-immunostained for ryanodine receptor-2 (RyR2) and junctophilin-2 (JP2) at postnatal days 10 (A, B), 15 (D, E), and 20 (G, H) and adult (J, K). Overlay images (F, I, L) provide a comparative view of RyR2 (green) and JP2 (red) protein distribution. Scale bars represent 10 μm. Inset images in panels F, I, and L show a zoomed region of the cell.

Cav3 and JP2 Co-localize at TTs. Changes in immunofluorescence labeling for Cav3 illustrate how the surface membrane and TT system mature with development (Fig. 2A, D, G, and J). The developmental changes observed with this marker are similar to those observed with di-8-ANEPPS (Fig. 1), but the use of Cav3 as a marker allowed us to examine fixed samples. Comparisons of Cav3 labeling with that for JP2 show that these two proteins change similarly during development (Fig. 2). In d10 myocytes, JP2 immunostaining is largely limited to the SL (Fig. 2B), as is Cav3 (Fig. 2A); there is little or no label for either protein within the volume of the cell (Fig. 2B). The merged image of the Cav3 and JP2 immunofluorescence images from d10 cardiomyocytes shows that the two proteins have very similar distributions (Fig. 2C). We provide quantitative comparisons of these distributions below.

As the rat ventricular myocyte matures and Cav3 is found increasingly at intracellular locations at the TTs, at the level of the Z-discs, it also remains at the surface SL (See Figs. 2A, D, G and J). JP2 reorganizes during development in a somewhat different manner. Like Cav3, it accumulates preferentially where TTs are found within the cell, but it is present at diminishing levels at the SL (Fig. 2E, H, K). Overlay images show that, as Cav3 increasingly marks TTs forming within the volume of the cell, JP2 populates the same or nearby regions (Fig. 2F, I, L). In day 20 and adult myocytes, the distributions of Cav3 and JP2 appear to overlap at the level of the TTs, and, although both are found at the SL, there is relatively little JP2 there compared to Cav3 (Fig. 2I, L). This finding is consistent with the current view that couplons are largely found along the TTs in adult myocytes, and less so at peripheral couplings at the SL ^{20, 21, 24, 26, 27}.

RyR2 and JP2 Co-localize in Adult Myocytes. Even at d10, when TTs are largely unformed in rat ventricular myocytes, RyR2s are present in striated patterns at the level of the Z-discs (Fig. 3A). By contrast, JP2 is primarily found at the surface membrane (Fig. 3B). As discussed above, its limited presence within the interior of the cell appears to be restricted to structures adjacent to TTs. Thus, there is little overlap in the distribution of JP2 and RyR2 at d10 (Fig. 3C). With development, however, JP2 concentrates increasingly at the level of the TT and is present in diminishing amounts at the SL. As this occurs, coincidence of labeling for RyR2 and JP2 increases throughout the volume of the cell, particularly in punctate structures adjacent to the Z-discs near the SL (Fig. 3C, F, I, L). These results suggest that, as myocytes develop, JP2 begins to appear adjacent to the Z-discs in close spatial proximity to the RyR2s and this new population of JP2 increases as TTs mature.

Correlation and Colocalization Analysis. We used statistical measures of colocalization to assess the relationships between Cav3 and JP2, and between RyR2 and JP2. Pearson’s coefficients (PC) were calculated from myocytes at different postnatal ages that had been co-immunostained for each pair of proteins (Fig. 4A and B). Consistent with the observation that Cav3 and JP2 have very similar patterns of distribution, PCs calculated for Cav3 and JP2 showed no significant differences between them at any of the time points we tested (Fig. 4A). In sharp contrast to this result, PC values calculated for the relationship between RyR2 and JP2 increased during development (Fig. 4B). These results indicate that JP2 and Cav3 become concentrated in intracellular structures as the TT system develops, and that as they do so, their spatial distribution becomes very similar to that of the RyR2.

Figure 4. Correlation and Colocalization Analysis of TT-SR Junction Components.

A linear correlation analysis was completed to quantify the relationship between Cav3 and JP2 (A) or RyR and JP2 (B). Pearson’s coefficients (PCs) calculated for the relationship between Cav3 and JP2 showed no significant difference between any of the time points tested (P-value > 0.05) (A). In contrast, PCs calculated for RyR and JP2 showed a statistically significant (P-value < 0.05) increase at each subsequent time point although there was not significance between day 20 and adult (P-value > 0.05) (B). Further quantification of the spatial relationship between RyR2 and JP2, colocalization analysis was completed by performing a Manders analysis (C). Displayed are the coefficients of colocalization for RyR at regions either near the sarcolemma (open bars) or within the cytosol (closed bars). (Asterisk denotes statistical significance with P-value < 0.05)

Whereas the determination of PCs provides a useful statistic of overall correlation of two labeled proteins, analysis using the methods of Manders et al ³⁰ provides more specific assessment of colocalization throughout the volume of the cell or in selected regions (Fig. 4C) by calculating the contribution of each molecule to colocalized volumes. We used this method to analyze the specific voxel content of RyR2 and JP2 visualized in high resolution z-stacks of immunolabeled cells. The contribution of JP2 to colocalization with RyR2 was consistently ~90% (or a coefficient of ~0.9) across all ages and regions (data not shown), but the contribution of RyR2 to colocalization increased throughout development. In d10 myocytes, the contribution of RyR to colocalization was greater near the SL (~33%) than in the middle (~14%) of the cell. By d15, RyR contribution to colocalization in the cytosol (~35%) was equivalent to colocalization near the SL (~30%). As development continued, the contribution of RyR2 to colocalization with JP2 increased in both subcellular regions until it reached ~79% in the cytosol of adult myocytes (Fig. 4C).

CICR and EC Coupling during Development

The literature suggests that I_Ca-triggered SR Ca²⁺ release should increase during postnatal development, as cellular organization matures ^{14, 41}. We studied ECC in postnatal cardiomyocytes by imaging [Ca²⁺]_i signals and simultaneously recording I_Ca using a whole cell patch clamp method.

I_Ca and the triggered [Ca²⁺]_i transient. We investigated the relationship between L-type I_Ca and [Ca²⁺]_i by loading myocytes with Fluo-4 AM and then patch clamping them in the whole cell configuration. Fig. 5 shows the voltage-gated I_Ca and the triggered [Ca²⁺]_i for a voltage step from −50mV to 0mV. Pooled data from a range of voltage steps (−40mV to +50mV) show that current density (pA/pF) increased significantly from day 10 to day 15 (Fig. 5A, B, and F). Current densities in myocytes at d15 and d20 were not significantly different (Fig. 5B, C, and F). The highest current density was recorded from adult myocytes (Fig. 5F). Although current density increased from day 10 to day 15, there was no difference in the [Ca²⁺]_i (ΔF/F₀) recorded at either age (Fig. 5E). [Ca²⁺]_i recorded from either d20 or adult myocytes was significantly greater than [Ca²⁺]_i from either d10 or d15 myocytes.

Figure 5. Ca²⁺-induced Ca²⁺ Release.

I_Ca and [Ca²⁺]_i were simultaneously recorded with a whole cell patch clamp method and confocal imaging using freshly isolated myocytes loaded with fluo-4. Whole cell Ca²⁺ transients were recorded during 10mV steps from a holding potential of −50mV. Representative line scan images show [Ca²⁺]_i recorded during a voltage step to 0mV for each age animal tested (A-D). Beneath the line scan image is a profile of total fluorescence (arbitrary units) from the line scan image and the simultaneously recorded, raw I_Ca (pA). Pooled data for the peak of [Ca²⁺]_i (E) showed no difference between day 10 (◆) (n=6) and day 15 (■) (n=5) myocytes. Although the results obtained from day 20 myocytes (▲) (n=7) were significantly different from these earlier time points, it was not different from values recorded from adult (●) (n=6) myocytes. The corresponding peak I_Ca density (I_Ca normalized to cell capacitance, pA/pF) showed no difference between days 15 and 20, but both were significantly greater than day 10 and less than adult.

Ca²⁺ content in the SR during development. The amount of Ca²⁺ within the SR is expected to change with development between postnatal day 10 and adulthood, as cardiomyocytes get larger and the organization of the SR matures. To evaluate total SR [Ca²⁺], we measured the peak of caffeine-induced (10 mM) [Ca²⁺]_i transients (Fig. 6B). Ca²⁺ measurements (given as ΔF/F_o) surprisingly showed that there was no change in SR Ca²⁺ content between postnatal d10 and d20 (Fig. 6A). Caffeine-induced [Ca²⁺]_i transients recorded in adult myocytes were significantly greater (1.5 fold) than in d20 myocytes, however, suggesting that the ability of the SR to store Ca²⁺ increases after d20. To estimate the maturity of ECC in developing myocytes, we calculated the fractional release of triggered [Ca²⁺]_i elicited by a voltage step from −50 mV to 0 mV (Fig. 6C). In day 10 and day 15 myocytes only ~33% of total SR [Ca²⁺] was released by depolarization, while day 20 and adult myocytes released ~60%.

Figure 6. SR Ca²⁺ Content and Triggered Fractional Ca²⁺ Release.

Cell size was measured at each developmental time point by recording cell capacitance (pF) (A). Capacitance recordings, calculated by measuring the capacitative current induced by a rapid 10mV hyperpolarizing pulse, shows that cell size was largely unchanged between postnatal day 10 (n=6) and day 15 (n=5). Cells were slightly larger at day 20 (n=7) and considerably larger in adult myocytes (n=6). To measure SR [Ca²⁺] load (B), freshly isolated ventricular myocytes were loaded with the fluorescent Ca²⁺ indicator, fluo-4. V_m was held constant at −80 mV by whole cell patch clamp while the myocytes were rapidly perfused with a solution of 10 mM caffeine, to induce SR Ca²⁺ release. The recorded peak of the caffeine induced [Ca²⁺]_i transient (ΔF/F₀) was considered to be a measure of the approximate SR [Ca²⁺] load. There was no statistical difference between transient peaks recorded in days 10 (n=9) and 15 (n=8) myocytes, while those recorded in day 20 (n=10) were larger (p<0.001) and [Ca²⁺]_i transients recorded in adult (n=5) myocytes were significantly (P < 0.05) larger than at all earlier time points. To measure the efficiency of ECC in developing myocytes, fractional SR Ca²⁺ release was calculated by comparing triggered [Ca²⁺]_i transients to those induced by caffeine application (C). Fractional Ca²⁺ release was smaller in day 10 (n=6) and day 15 (n=8) myocytes at ~33%, than day 20 (n=8) and adult (n=5) myocytes at ~60% (p<0.05).

Discussion

ECC in the mature adult heart relies on the amplification of a relatively small current flowing through the DHPRs, which triggers RyRs in the closely apposed SR membrane to open, but neonatal myocytes which continue to develop their ECC organization after birth rely on different mechanisms. During embryonic development and through the time of birth, cardiac myocytes in rodents appear to rely on an influx of Ca²⁺ from not only the DHPR, but also T-type Ca²⁺ channels and the Na-Ca²⁺ exchanger (NCX) ⁴². SERCA is expressed at much lower levels early in development than in adults, so removal of elevated [Ca²⁺]_i produced by Ca²⁺ entry and SR Ca²⁺ release is more dependent on NCX and, presumably, the plasmalemmal Ca²⁺ ATPase ⁴³. We have studied the subcellular changes in organization and efficiency of ECC in postnatal rat cardiomyocytes as they mature and show that ECC increases with the formation and maturation of a TT system and TT-SR couplons. We show that, although the RyR2s of the SR are organized at the Z-discs and are fully capable of Ca²⁺ release early in development, the efficiency by which they are triggered by DHPRs in the TTs changes dramatically during development. More specifically, we show the surprising result that Cav3 which is found at the TTs colocalizes with the jSR protein JP2 throughout development. However, the dramatic increase in ECC efficiency that is a hallmark of the mature adult ventricular myocyte, does not occur until JP2 and RyR2 colocalize. Our results provide the first time course of the coordinated structural and functional maturation of couplons during postnatal development in rodent heart.

Our results reveal the relationship between the distribution of proteins involved in excitation-contraction coupling and the functional changes that occur with development during the postnatal period in the rat heart. The transverse tubule system is absent at birth and first appears in an extremely primitive form at approximately postnatal day 10³. The TT system matures over the first three postnatal weeks, when it attains the architecture seen in the adult heart. Although RyR2 organizes at the Z-discs developmentally early in postnatal rat hearts, proteins of the TT (Cav3 and others) appear intracellularly at later times, consistent with their association with sarcolemmal invaginations visualized with di-8-ANEPPS. Our results show that the efficiency of ECC seen in the adult heart requires complete development and alignment of TT and RyR in the SR, and that intermediate stages of development, in which TT and SR development are incomplete, are accompanied by less efficient coupling of excitatory influx of Ca²⁺, via L-type Ca²⁺ channels, to the RyR-mediated release of Ca²⁺ from the SR.

Caveolin and T-tubule Formation

Caveolins were first identified as structural proteins of caveolae, cholesterol-rich, flask shaped subdomains of the plasma membrane⁴⁴. Cav3 is also believed to play a role in the formation of TTs in cardiac and skeletal muscle during development, but its absence due to homologous recombination does not prevent the formation of TTs in developing mouse heart ^{37, 38, 45, 46}. In agreement with others, our results show that Cav3 is present at both the SL and developing and mature TTs in cardiac myocytes ^{37, 38}. Although its role in TTs is unknown, Cav3 provides a useful tool for tracking the formation and maturation of TTs during cardiac development. Similar to results with Di-8-ANEPPS, antibodies to Cav3 only labeled primitive invaginations in day 10 postnatal cardiac myocytes, In older myocytes, both markers revealed the stages of TT formation and their transverse alignment with respect to the contractile apparatus. These results were consistent with our finding that L-type Ca²⁺ current increased from postnatal day 10 to day 15, when TTs form but are not yet aligned transversely.

Our detection of Cav3 in key locations during development of both the TT system and its enrichment at those locations during maturation of ECC suggest a role in normal cardiac development or function. For example Cav3 may modulate the activity of Na/K ATPase, which would likely affect membrane potential and, indirectly, the size of the Ca²⁺ store in the SR ⁴⁶. Recent work by Shibata and colleagues has also suggested that variations in caveolae may enable or blunt Na⁺ channel function ^{47, 48}. Further Cav3 colocalizes with both RyR2 and DHPR, but not NCX, in cardiac myocytes, suggesting that Cav3 is found at dyads ^{39, 47, 48}. Reduction in caveolae by exposure to increasing concentrations of beta-methyl cyclodextrin, which depletes cells of cholesterol, leads to a dose-dependent decrease in the frequency and size of individual Ca²⁺ release events (Ca²⁺ “sparks”) in both smooth and cardiac myocytes ⁴⁹.

JP2 and RyR Organization

The junctophilin family of proteins links the plasma membrane (PM) to the endoplasmic/sarcoplasmic reticulum (ER/SR) in many types of cells. JPs have a transmembrane domain that anchors them to the ER/SR. Their cytoplasmic sequences bind to the SL or TT ^{39, 50}. In skeletal muscle, JP1 and JP2 localize to triad junctions, where they mediate the adhesion of TTs and the terminal cisternae of the SR, but only JP2 is found in heart ^{51, 52}. JP2 localization changes from a peripheral distribution near the SL in immature cardiac myocytes to a concentration in the triad junctions of the SR, aligned with TTs, in mature myocytes. Although the timing of this redistribution generally parallels that of TT formation, it is slightly delayed, Our results suggest that, during postnatal cardiac development in the rat, JP2 increasingly co-localizes with the RyR2 clusters as the TTs invaginate and acquire their mature organization, consistent with the idea that the mature organization of JP2 depends on the intersection of the forming TTs and concentrations of RyR2, already established in the SR compartments localized around Z-discs. JP2 and its interactions with proteins of the TT are likely to play a significant role in the health of cardiac muscle. For example, JP2 associates with Cav3, and their interaction is down-regulated in cardiomyopathy. Furthermore, mutations in JP2 occur in some patients suffering from hypertrophic cardiomyopathy^{51, 53, 54}. Our findings support an important role for JP2 in the maturation of the structures underlying ECC.

Maturation of EC-Coupling

Our results suggest that the SR in postnatal cardiac myocytes may not fully mature until rats approach 3 weeks of age. Bassani and Bassani (2002) showed that cardiac myocytes from newborn rats relied less on the contribution of SERCA for removal of [Ca²⁺]_i than adult myocytes (75% vs. 92%) ⁵⁵. Our data show that SR [Ca²⁺] load, as measured by peak response to caffeine, is lower at postnatal days 10 and 15 than in day 20 and adults. Measuring the peak response of Fluo-4 to caffeine application is a less quantitative method for measuring SR [Ca²⁺] load than methods used by others ⁵⁶, but it greatly simplifies interpretation of the caffeine response and has been recently used by others as an approximation of SR Ca²⁺ load ^{57, 58}. Although largely a qualitative measure, the results of this experiment support the idea that postnatal maturation of the SR and its ability to store Ca²⁺ is a slow and ongoing process.

The slow maturation of SR [Ca²⁺] load parallels that of the triggered [Ca²⁺]_i transient (Fig. 5), where there is an increase in L-type Ca²⁺ current but no change in the triggered SR [Ca²⁺] release. The implications of this change in current density are of particular significance for the maturation of ECC, as it represents the functional impact of TT formation without couplon formation. Since there are fewer couplons formed in younger cardiac myocytes compared to adults (Fig. 3, 4), any influx of Ca²⁺ ions that occurs will trigger SR Ca²⁺ release at very few local sites. The fractional release of SR Ca²⁺ expected under these conditions should therefore be reduced. We found that the fractional release of SR Ca in myocytes at 10-15 days after birth, to be ~33%, compared to ~60% in adult cells.

Not only does the absence of couplons early in development have an impact on the fractional release of SR Ca²⁺, but it also affects rise time or rate of rise for the peak of the [Ca²⁺] transient. As is apparent in the linescan images shown in Fig. 5, myocytes that lack TT-SR junctions (days 10 and 15) reach the peak of the [Ca²⁺] transient later than more developed myocytes (day 20 and adult). Moreover, more developed myocytes will have more synchronous DHPR-triggered SR Ca²⁺ release while less developed myocytes will rely heavily on [Ca²⁺] transient propagation to trigger SR Ca²⁺ release. The slower rise time of the [Ca²⁺] transient in less mature myocytes is not prima facia evidence that the SR cannot be readily triggered by local [Ca²⁺]_i. Indeed, when the triggering mechanism is similar, the rise time of the [Ca²⁺]_i transient is similar as found for caffeine-induced [Ca²⁺] transients (data not shown).

Our results in adult myocytes confirms the observations of Delbridge, et al. (1997), who calculated an SR fractional release value of 55.4 ± 9.4% for a single action potential ⁵⁹. The combination of lower current density, the inefficient local release due to a lack of couplons, and the small SR Ca load all contribute to the lower SR Ca²⁺ release and the immature state of ECC in early postnatal cardiac development.

Finally, our results suggest that the terminal cisternae of the postnatal rat cardiomyocyte is assembled from at least two distinct membrane components, One forms early, assembles around the Z-discs and contains RyR2; the other forms later, as TTs invaginate, and contains JP2. The factors coordinating the assembly of these two components to form the mature terminal cisternae of the SR, as well as those that link the terminal cisternae to the other compartments of the SR, must still be identified.