|Home | About | Journals | Submit | Contact Us | Français|
Conceived and designed the experiments: JL MSI NAB AFD RT DFV DRL. Performed the experiments: JL NAB DRL. Analyzed the data: JL NAB DRL. Contributed reagents/materials/analysis tools: NAB AFD DFV DRL. Wrote the paper: JL DRL AFD.
Here we investigate how ß-adrenergic stimulation of the heart alters regulation of ryanodine receptors (RyRs) by intracellular Ca2+ and Mg2+ and the role of these changes in SR Ca2+ release. RyRs were isolated from rat hearts, perfused in a Langendorff apparatus for 5 min and subject to 1 min perfusion with 1 µM isoproterenol or without (control) and snap frozen in liquid N2 to capture their phosphorylation state. Western Blots show that RyR2 phosphorylation was increased by isoproterenol, confirming that RyR2 were subject to normal ß-adrenergic signaling. Under basal conditions, S2808 and S2814 had phosphorylation levels of 69% and 15%, respectively. These levels were increased to 83% and 60%, respectively, after 60 s of ß-adrenergic stimulation consistent with other reports that ß-adrenergic stimulation of the heart can phosphorylate RyRs at specific residues including S2808 and S2814 causing an increase in RyR activity. At cytoplasmic [Ca2+] <1 µM, ß-adrenergic stimulation increased luminal Ca2+ activation of single RyR channels, decreased luminal Mg2+ inhibition and decreased inhibition of RyRs by mM cytoplasmic Mg2+. At cytoplasmic [Ca2+] >1 µM, ß-adrenergic stimulation only decreased cytoplasmic Mg2+ and Ca2+ inhibition of RyRs. The Ka and maximum levels of cytoplasmic Ca2+ activation site were not affected by ß-adrenergic stimulation.
Our RyR2 gating model was fitted to the single channel data. It predicted that in diastole, ß-adrenergic stimulation is mediated by 1) increasing the activating potency of Ca2+ binding to the luminal Ca2+ site and decreasing its affinity for luminal Mg2+ and 2) decreasing affinity of the low-affinity Ca2+/Mg2+ cytoplasmic inhibition site. However in systole, ß-adrenergic stimulation is mediated mainly by the latter.
Exercise and stress induce the sympathetic “fight or flight” response that increases heart rate and contractility. During this response, increased catecholamine concentrations stimulate cardiac β-adrenergic receptors, resulting in adenylyl cyclase activation, increased cyclic AMP and increased activity of cyclic AMP-dependent protein kinase A (PKA). Increased intracellular [Ca2+] causes Ca2+/calmodulin dependent protein kinase II (CaMKII) autophosphorylation so that it remains active at low [Ca2+] , . PKA and CaMKII phosphorylation increase Ca2+-transport activity that underlies contraction and pacemaking in cardiac cells , –.
Systolic contraction follows activation of sarcolemmal voltage-gated L-type Ca2+ channels during an action potential resulting in Ca2+-influx, which activates cardiac ryanodine receptor (RyR2) calcium release channels on the sarcoplasmic reticulum (SR, major intracellular Ca2+ store). The subsequent rise in cytoplasmic [Ca2+] causes contraction. Diastolic relaxation occurs with cessation of Ca2+ release and Ca2+ sequestration by the SR Ca2+ uptake transporter (sarcoplasmic/endoplasmic reticulum Ca2+-ATPase) .
Three phosphorylation sites have been confirmed by 32P incorporation assays and known to be phosphorylated in vivo; namely S2808/S2809, S2814/S2815 (depending on the species) and S2030 , , . Mass spectroscopy of the RyR2 phosphorylation domain (aa2699–2904) has revealed additional, in vitro phosphorylation sites on RyR2 ,  at S2810, S2811, S2797 and T2876. Present understanding of the contribution of RyR2 to ß-adrenergic stimulation is limited by a lack of knowledge of how RyR2 is regulated by phosphorylation within the cell. Most investigations show increased RyR2 activity in lipid bilayers with in vitro phosphorylation at these sites by exogenous PKA and CaMKII ,  (although  reports that CaMKII inhibits RyR2), increasing RyR2 sensitivity to luminal Ca2+  and decreasing cytoplasmic Mg2+ inhibition . However, it is not clear if exogenous enzymes can replicate the mechanisms of RyR2 phosphorylation at rest or during ß-adrenergic stimulation. Therefore, just how ß-adrenergic signaling alters regulation of RyR2 by luminal and cytoplasmic Ca2+ and Mg2+ is unknown. In this study, we perfused isolated rat hearts with the ß-adrenergic agonist, isoproterenol, then incorporated RyR2 from these hearts into artificial lipid bilayers whilst preserving their phosphorylation state. The response of RyRs to luminal and cytoplasmic Ca2+ and Mg2+ indicated significant and novel changes in RyR2 function associated with increased RyR2 phosphorylation at S2808 and S2814 induced by ß-adrenergic stimulation. Since it is not feasible to measure RyR2 regulation under diastolic [Ca2+] and [Mg2+] (their activity is too low), we used a RyR2 gating model to predict the effects of ß-adrenergic stimulation on RyR2 activity within cells under diastolic conditions.
Healthy male adult rats (Sprague-Dawley) were heparinized (2000 U, injection BP), and anesthetized with Isoflurane. Hearts were rapidly removed and immediately perfused via the Langendorff method (see Methods S1 in File S1). Hearts were perfused with Krebs Henseleit buffer for 5 min after which they were either perfused in buffer with or without 1 µM isoproterenol for 1 min, while monitoring heart performance. Following perfusion, hearts were rapidly frozen in liquid nitrogen to preserve phosphorylation and stored at −80°C.
Liquid nitrogen-frozen hearts were homogenized in buffer containing 10 mM imidazole, 0.5 mM DTT, 3 mM sodium azide, 0.29 M sucrose, 1 µg/ml leupeptin, 1 µg/ml pepstatin A, 1 mM benzamidine, 0.5 mM PMSF, 20 mM NaF, pH 6.9 (homogenization buffer). Homogenization was carried in 3×15 s bursts with a rotor-stator homogenizer (HD Scientific) followed by 10 manual strokes of a loose glass/glass Dounce homogenizer. The homogenate was then centrifuged at 7000 rpm for 20 min in Beckman Optima L-100XP Ultracentrifuge. The resulting supernatant was centrifuged at 50000 rpm for 30 min. The pellet from this step was re-homogenized with a tight manual glass/glass Dounce homogenizer in homogenization buffer also containing 0.65 M KCl (storage buffer, pH 6.7), incubated for 30 min and then centrifuged at 7000 rpm for 15 min. The supernatant was centrifuged at 50000 rpm for 1 h, and the resulting pellet was resuspended in storage buffer, snap frozen in liquid nitrogen and stored at −80°C. The NaF present in our buffers was to prevent ongoing activity of phosphatases (PP1 and PP2A) during SR vesicle preparation. The whole procedure was carried out at 4°C.
Lipid bilayers were formed form a mixture of phosphatidylethanolamine and phosphatidylcholine (82 wt/wt) in 50 mg/ml n-decane. Lipid bilayers were formed across a delrin hole (approximately 100 µm diameter), separating two experimental compartments, cis (cytoplasmic) and trans (luminal). For vesicle fusions the cis solution contained (mM): 250 Cs+ (230 CsCH3O3S, 20 CsCl, both from Sigma Aldrich), 1–5 CaCl2 (BDH Chemicals), and 500 Mannitol (Ajax Chemicals); while trans solution contained (mM): 50 Cs+ (30 CsCH3O3S, 20 CsCl), 0.1 CaCl2 and 10 TES. For channel recordings, [Cs+] in trans solution was raised to 250 mM by adding an appropriate volume of 4 M CsCH3O3S. During experiments the composition of the cis solution was altered by local perfusion, which provided solution exchange within ~1 s , and the composition of the trans solutions was altered by addition of aliquots of stock solutions. All experiments were performed at room temperature (21–26°C).
All solutions were made with MilliQ water and were pH buffered with 10 mM TES (N-tris [hydroxymethyl] methyl-2-aminoethanesulfonic acid; ICN Biomedicals) and adjusted to pH 7.4 by CsOH (ICN Biomedicals), using a TPS digital pH meter. Free [Ca2+] ≤3 µM was buffered using 4.5 mM BAPTA (1,2-bis (2-aminophenoxy) ethane- N, N, N’, N’- tetraacetic acid; Invitrogen) and additional 1 mM dibromo BAPTA was used for 10 µM free Ca2+. Sodium citrate was used to adjust free [Ca2+] between 10–50 µM. A Ca2+ electrode (Radiometer) was used to determine free [Ca2+] in the experimental solutions and the purity of Ca2+ buffers and Ca2+ stock solutions using titrations against a 100 mM CaCl2 standard (Fluka).
All recording solutions contained ATP (2 mM, sodium salt from Sigma Chemicals), which buffers Mg2+. The Ca2+ electrode was used to determine the purity of the ATP stock solutions by titrations against the 100 mM CaCl2 standard. Free [Mg2+] was estimated using published association constants  and the program “Bound and Determined” . The free [Mg2+] was confirmed using the fluorescent magnesium indicator, Mag-fura-2 (tetra potassium salt from Molecular Probes). The ratio of fluorescence intensities at 340 and 380 nm were calibrated in the experimental solutions (250 mM Cs+ solutions, see above) also containing 5 µM Mag-fura-2, 4.5 mM BAPTA, (free [Ca2+] <1 µM) and MgCl2 from aliquots of a calibrated stock.
Standard techniques were used for SDS PAGE and Western Blotting (see Methods S1 in File S1 details). Phospho-antibodies were obtained from Badrilla (UK). To establish loading controls, each membrane probed with phospho-antibodies was re-probed with anti-RyR2 (C3–33 from Abcam) after quenching of HRP activity with sodium azide (3% w/v for 3 h at room temperature; see Figure S1 in File S2) . The ratio of binding of phospho- and RyR2- antibodies was used to measure the degree of RyR2 phosphorylation.
Phosphorylation of RyR2 was determined by probing heart SR isolates using phospho-specific antibodies against S2808 and S2814 (pS2808 and pS2814, respectively). We also used an antibody against dephosphorylated S2808 (DepS2808) as a cross check for pS2808. The average degree of S2808 phosphorylation was quantified as both pS2808 immunostaining relative to maximum PKA phosphorylation and also DepS2808 staining (1- DepS2808 staining) relative to minimal phosphorylation in PP1 incubated samples. Similarly, S2814 phosphorylation was expressed as pS2814 immunostaining relative to maximal phosphorylation levels achieved using calcium calmodulin (CaCaM) to activate endogenous CaMKII. Maximal and minimal phosphorylation levels were determined from the time courses of phosphorylation during incubations with PP1, PKA and CaCaM (see Methods S1 in File S1 and Figures S2, S3, and S4 in File S2, respectively).
We tested that our anti-pS2808 and anti-pS2814 antibodies had the same specificity in detecting PKA and CaMKII phosphorylation of RyR2 as seen in a previous study . We subjected PP1-dephosphorylated RyR2 to successive incubations of PKA and CamKII (or in reverse order, see Methods S1 in File S1) and followed the time course of staining by both antibodies (Figures S3 and S4 in File S2). We found that 10 min incubations with PP1 caused maximal decrease in RyR2 phosphorylation at S2808 (Figures S2, S3, and S4 in File S2). Subsequent PKA incubation for 20 min produced near maximal S2808 phosphorylation and also increased antibody staining by S2814 to 20% of maximal levels attainable with CaMKII phosphorylation. Alternatively, CamKII incubation of PP1 dephosphorylated RyRs for 10 min produced maximal S2814 phosphorylation and also increased antibody staining by S2808 to 50% of maximal levels attainable with PKA phosphorylation. Thus, antibodies for S2808 respond to CamKII mediated phosphorylation and those for S2814 respond to PKA phosphorylation as reported previously .
Since fusion of vesicles with lipid bilayers was performed in the absence of phosphatase inhibitors, we determined if RyRs could maintain their phosphorylation levels at S2808 during vesicle fusion by incubating vesicles in the buffers used for vesicle fusion and channel recording. We found no significant decrease in phosphorylation of S2808 during incubation with vesicle fusion buffers over a period of 30 min (Figure S5 in File S2). Since single channel experiments were performed with ATP, we determined whether ATP could phosphorylate S2814 via endogenous CaMKII/calmodulin  SR vesicles from control hearts incubated for 10 min in buffers containing 2 mM ATP, without exogenous CaMKII or calmodulin, did not exhibit increased S2814 phosphorylation (Figure S6 in File S2).
All data are shown as mean ± standard error of the mean (SEM). Significance was calculated by Student’s t test. p<0.05 was considered significant.
All experimental procedures were approved by the University of Newcastle Animal Care and Ethics Committee (A-2009-153).
Mean heart rate was 220±18 bpm (Table S1, hearts 1–14 in File S1) after 5 min perfusion with Krebs-Henseleit buffer. In six experiments, where isoproterenol (1 µM) was added to the perfusion buffer, heart rate increased by 77±9% (n=6) to 348±23 bpm after 1 min (Table S1, hearts 9–14 in File S1).
Western blot analysis using anti-RyR2 and phospho antibodies (Figure 1A) revealed a band corresponding with RyR2 at 560 kDa (full length) and a fainter band corresponding to a 400 kDa, c-terminal fragment similar to that seen in many previous studies (e.g. , ). We measured the effects of isoproterenol perfusion on RyR2 phosphorylation to confirm that RyR2 had been subjected to normal ß-adrenergic signalling processes in the heart. ß-adrenergic stimulation of rat heart was correlated with an increased phosphorylation at S2814 and S2808. The relative levels of phosphorylation at S2814 after exposure to isoproterenol for 1 min increased 4-fold from that detected in control (p=10−5). S2814 phosphorylation in control hearts was not different to RyR2 dephosphorylated by PP1 (Figure 1C, p=0.16). RyRs from control hearts showed a relatively high level of phosphorylation at S2808. Western blots probed with pS2808 and DepS2808 indicated S2808 phosphorylation at 0.69±0.07 and 0.83±0.07 (respectively) of maximal PKA phosphorylation (Figure 1B). Isoproterenol stimulation for 1 min increased S2808 phosphorylation to 0.86±0.07 (p=0.04) as determined by pS2808 staining. However, we could not detect increased phosphorylation using DepS2808. (Since our antibodies to S2030 showed no detectable binding to RyR2 from rat heart, either control or isoproterenol stimulated, S2030 was not examined.) The response of RyR2 phosphorylation to isoproterenol is typical of many previous studies of ß-adrenergic stimulation (see discussion).
We investigated the concentration dependencies of cytoplasmic and luminal Ca2+ and Mg2+ regulation of RyR2 at a bilayer potential of −40 mV where the four identified Ca2+/Mg2+ regulation mechanisms in RyR2 are best revealed , . In initial experiments, the cytoplasmic bath contained 100 nM Ca2+ (diastolic [Ca2+]) with 2 mM ATP and the luminal bath contained 0.1 mM Ca2+. Figure 2A (Traces 1 and 2) shows representative single channel recordings from two experiments on RyRs from control hearts performed under these conditions. RyR2 from isoproterenol stimulated hearts showed similar gating kinetics, albeit with higher opening rates (Figure 2A, Traces 3 and 4).
RyR2 was dephosphorylated to assess whether changes in RyR2 activity after ß-adrenergic stimulation were due to increased RyR2 phosphorylation. The activity of RyR2 from control and isoproterenol stimulated hearts was measured with 100 nM and 10 µM cytoplasmic [Ca2+]. RyR2 was dephosphorylated in-bilayer by adding PP1 (7–10 units) to the cis bath for 5 min either by local perfusion or by aliquot addition of PP1 . Then the cis chamber was perfused with solutions lacking PP1 and with cytoplasmic [Ca2+] of 100 nM or 10 µM and the same channel was recorded again (recordings shown in Figure 2B). In 8 of 8 experiments, PP1 reduced RyR2 opening rate at 100 nM cytoplasmic Ca2+ (Figure 2C). Prior to PP1 incubation, RyRs from isoproterenol treated hearts (n=4) had significantly higher opening rates than RyRs from control hearts (n=4). However, after PP1 incubation, there was no significant difference in opening rate between the two groups. These results suggest that ß-adrenergic stimulation increases RyR2 opening rate due to an increased channel phosphorylation. PP1 inhibition of RyR2 from control hearts suggests a significant phosphorylation-mediated RyR2 stimulation in the absence of excess ß-adrenergic stimulation. At 10 µM Ca2+ (n=7), PP1 did not change either opening rate or mean open duration (p=0.86 or 0.13, respectively). Figure 2D shows the relative differences in RyR gating parameters of three groups of RyRs; namely, 1) control RyRs that were incubated with PP1 for 5 min, 2) RyRs from control hearts and 3) RyRs from isoproterenol stimulated hearts. (Absolute values, numbers of experiments and p-values are given in Table 1). ß-Adrenergic stimulation of RyRs in control heart led to a 2.5-fold increase in open probability (Po) due to an increase in RyR opening rate whereas PP1 incubation led to a 15-fold decrease in Po due to decreases in both opening rate and mean open duration.
RyR2 from control and isoproterenol stimulated hearts were strongly activated by micromolar cytoplasmic Ca2+. Increasing cytoplasmic Ca2+ from 0.3 µM to 10 µM decreased RyR2 mean closed duration and increased mean open duration (Figure 3A, Traces 1–4). Po exhibited activation at µM Ca2+ (Figure 3B) and inhibition at mM Ca2+ (below). Hill equations were fitted to the data (not shown), using the Hill parameters in Table 2. Isoproterenol stimulation for 1 min increased Po by 10-fold at low [Ca2+] (Pmin) with no significant effect on Ka for activation or maximal Ca2+ activation (Pmax), suggesting that the Ca2+ sensitivity of the cytoplasmic Ca2+ activation site of RyR2 was not affected by isoproterenol. The Ca2+ dependencies of Po were also reflected in RyR2 opening rate (Figure 3C). At sub µM Ca2+, isoproterenol stimulation caused a 10-fold increase in opening rate and this difference became less as cytoplasmic Ca2+ increased. In these experiments, there was no difference in the mean open duration between the isoproterenol and control groups (p=0.218, Figure 3D) however in the luminal Ca2+ experiments, differences were observed (see below).
Figure 4A shows RyR2 activation by increasing luminal Ca2+ from 10 to 100 µM. Figures 4B–D show the luminal [Ca2+]-dependencies of the Po, opening rate (ko) and mean open time (To) of control and isoproterenol treated RyR2 with diastolic (100 nM) cytoplasmic [Ca2+]. In the absence of luminal Ca2+, RyR2 from control and stimulated groups had a mean Po=0.003 and exhibited a bell-shaped luminal [Ca2+]-dependence with peak activity occurring in the presence of ~100 µM. Peak Po of RyRs from the isoproterenol-stimulated hearts was 4-fold larger (Po=0.045) than control (Figure 4B, curves show fits of Hill equations using Hill parameters in Table 2). The only Hill parameter that showed a significant difference between RyRs from control and stimulated hearts was Pmax. Therefore, isoproterenol-stimulation enhances the sensitivity of RyR2 to changes in luminal Ca2+ concentration without altering the Ka for activation or the Ki for inhibition. This enhanced sensitivity was also reflected by increases in both RyR2 opening rate and mean open duration (Figures 4C & D, curves show fits of RyR2 gating model using parameters in Table S3 in in File S1). Since measurements of opening rate are carried out on closed channels (opening rate=1/mean closed duration), the increase in the luminal Ca2+-dependence of opening rate unambiguously indicates an increased effect of Ca2+ binding to the luminal Ca2+ activation site of RyR2 , .
Cytoplasmic and luminal Mg2+ inhibition of RyR2 at diastolic cytoplasmic [Ca2+] (<1 µM) was previously shown to be due to Mg2+ competing with Ca2+ at cytoplasmic and luminal Ca2+ activation sites; (at higher cytoplasmic [Ca2+] cytoplasmic Mg2+ inhibits RyR2 by mimicking Ca2+ at the low affinity inhibition site  and this is dealt in the following section). Mg2+ inhibition via the cytoplasmic and luminal Ca2+ activation sites was investigated here by measuring Mg2+ inhibition in the presence of 100 nM cytoplasmic Ca2+ (Figures 5 and and6).6). Cytoplasmic and luminal Mg2+ inhibition of Po were fitted with Hill equations (not shown; Hill parameters in Table 3). RyR2 gating model is compared with the data in Figures 5, ,6,6, and and77 (parameters in Table S3 in File S1, note that model fits did not minimise least squares differences to data in Figure 5 because fits were optimised for all data in Figures 5, ,6,6, and and77 using the same parameters). Cytoplasmic Mg2+ inhibition was no different (p=0.56) in RyRs from control and stimulated hearts (Figure 5A). However, adrenergic stimulation reduced the effects of luminal Mg2+ inhibition of Po (Figure 6A), increasing the Ki (p=0.003) by 4.5-fold compared with control (Table 3). The decreased luminal Mg2+ inhibition seen after adrenergic stimulation was reflected in increased mean open durations (Figure 6D) and a decreased effect of Mg2+ on the opening rate (Figure 6C). For reasons outlined in the previous paragraph, the alleviation of Mg2+ effect on the opening rate indicates a modulation of a luminal facing site on the RyR.
Cytoplasmic Ca2+ and Mg2+ can inhibit RyR2 by binding to a low affinity divalent binding site . This site underlies Ca2+ inhibition and Mg2+ inhibition at mM concentrations, seen when cytoplasmic [Ca2+] exceeds 10 µM. Figure 7A shows Ca2+ inhibition of RyRs from control and stimulated hearts. Hill equations were fitted to the data, using Hill parameters in Table 2. Isoproterenol alleviated Ca2+ inhibition of RyRs by increasing Ki from 1.5 to 2.5 mM, indicating decreased Ca2+ inhibition via the low affinity divalent binding site.
We also investigated effects of adrenergic stimulation on Mg2+ inhibition via the Ca2+ inhibition site by measuring Mg2+ inhibition of Po in the presence of 10 µM cytoplasmic Ca2+ (Figure 7B). We found that RyR2 from stimulated hearts showed less Mg2+ inhibition (at 1 mM Mg2+) than control, indicating decreased Mg2+ inhibition via the low affinity divalent binding site.
Finally, we measured the Ca2+ activation properties of rat RyR2s in the presence of physiological concentrations of cytoplasmic Mg2+ (1 mM) and in the presence of 1 mM luminal Ca2+ (Figure 7C). Under these conditions, the Ka for cytoplasmic Ca2+ was 80 and 65 µM for RyRs from control and stimulated hearts, respectively. Although RyR2 from stimulated hearts showed higher mean Po than those from control hearts over the range pCa 4–5, none of the differences reached significance in these experiments.
Our observations are considered in the framework of a kinetic model originally developed for gating for sheep RyR2 that incorporates luminal and cytoplasmic regulation , . We adapted the model to the rat RyR and use it to gain insight into the mechanisms of ß-adrenergic regulation of rat RyR2 by intracellular Ca2+ and Mg2+ and to predict the effect of these mechanisms on RyR activity in the cell.
The RyR2 gating model considers Ca2+/Mg2+ regulation of RyR2 in terms of four Ca2+ sensing sites (L-, A-, I1- and I2-sites, see Figure 8A). The experimental evidence for each of these sites is reviewed elsewhere , . Two Ca2+-activation sites in the luminal (L-site) and cytoplasmic domains (A-site) of RyR2 have Ca2+ affinities of 8 µM and 2 µM, respectively. These sites trigger a common gating mechanism to produce synergistic activation by luminal and cytoplasmic Ca2+. The cytoplasmic domain also possesses two inhibitory sites with Ca2+ affinities of 0.4 µM (I1-site) and 1.5 mM (I2-site). Magnesium, which competes with Ca2+ at the L-, A- and I2-sites, inhibits RyRs and shapes the Ca2+-dependent activation of RyRs . Luminal Ca2+ and Mg2+ can pass through open channels and act at cytoplasmic facing sites.
RyR opening rates and mean open durations (Figures 3, ,4,4, ,5,5, and and6)6) were fitted by the model (solid and dashed lines; fit parameters in Tables S2 and S3 in File S1). The model accounted for most of the key features of the data with the exception of the high RyR opening rate observed in the presence of 100 µM cytoplasmic Ca2+ (Figure 3C). None the less, the model did provide good predictions of Po and To under this condition. Changes in RyR2 activity with adrenergic stimulation were reflected in changes in four parameters associated with three RyR2 sites, namely: 1) The opening rate in response to Ca2+ binding to the luminal Ca2+ activation site (L-site) increased 20-fold, accounting for the increased opening rate at low cytoplasmic [Ca2+] (Figure 3C) and increased luminal Ca2+ activation of opening rate (Figure 4C). 2) There was a 4-fold decrease in the Mg2+ affinity of the L-site, accounting for the decrease in luminal Mg2+ inhibition (Figure 6C), 3) RyR2 closing via the high affinity Ca2+ inactivation site (I2-site) was slowed 2-6-fold, accounting for the increase in mean open duration (Figure 4D). 4) The affinity of the cytoplasmic Ca2+/Mg2+ inhibition site (I1-site) was reduced by 40%, accounting for the reduced Ca2+ inhibition (Figure 7A) and Mg2+ inhibition in the presence of 10 µM cytoplasmic Ca2+ (Figure 7B). Interestingly, we found no effect of adrenergic stimulation on the cytoplasmic Ca2+ activation site (A-site). Model fits to the data indicated a 25% change in the Mg2+ affinity of the A-site which accounted for the systematic, though non-significant increase in RyR activity in Figure 7C.
We then predicted how the activity of RyRs might respond to ß-adrenergic stimulation within the cell using assumptions for the cell interior that have been reviewed and justified elsewhere . Briefly, we assumed that 1) the free Mg2+ in cytoplasm and lumen is 1 mM, 2) Ca2+ and Mg2+ fluxes through RyR2 are slightly larger in the presence of intracellular [K+] than with 250 mM Cs+ and that 3) intracellular Ca2+ buffering is substantially weaker and slower in the cell than in our bilayer solutions. Assumptions 2 and 3 affect the feed-through parameters, XA and XI in the model (see Table S3 in File S1).
The model predicts that under systolic conditions where cytoplasmic [Ca2+] exceed 1 µM, ß-adrenergic stimulation causes a 3-fold increase in RyR2 Po (Figure 8B). This is similar to a 2-fold stimulation of Po measured at systolic levels of Ca2+ and Mg2+ (Figure 7C). Under diastolic conditions (cytoplasmic [Ca2+] ~100 nM), the model predicts a RyR2 open probability in the range 10−7 to 10−6 which is far too small to be measured by single channel recording. Under these conditions, ß-adrenergic stimulation increases RyR2 Po by 20-fold, which was due to an increase in the ability of luminal Ca2+ to activate the RyR (Figure 8C).
We explored how the modulation of the L-, I1- and I2-sites during ß-adrenergic stimulation contribute to increasing RyR2 Po under both diastolic (Figure 8D) and systolic conditions (Figure 8E). In diastole, the model predicted that the increased action of Ca2+ binding to the L-site contributes a 3-fold increase in RyR2 activity during ß-adrenergic stimulation (Figure 8D, L-site (Ca2+)), decreased Mg2+ inhibition at the L-site contributed another 50% increase (Figure 8D, L-site (Ca2++Mg2+)) and reduced cytoplasmic Mg2+ inhibition contributed another 3-fold increase in activity (Figure 8D, full; I1-site and A-site contributed ~1.7-fold each). The I2- and A-sites made no significant contribution to ß-adrenergic stimulation. In systole, the only significant contribution to ß-adrenergic stimulation came from a decrease in cytoplasmic Mg2+ inhibition (Figure 8E).
Here, we report the first characterization of steady-state changes in RyR2 regulation by cytoplasmic and luminal Ca2+ and Mg2+ in response to acute ß-adrenergic stimulation of the heart by isoproterenol. We find that ß-adrenergic stimulation is associated with a marked change in RyR2 regulation by intracellular Ca2+ and Mg2+ and an increase in the level of phosphorylation of S2808 and S2814 on the RyR. ß-adrenergic-induced changes in RyR function were reversed when RyR2 was dephosphorylated by incubation with PP1, indicating that ß-adrenergic-induced changes were mediated by phosphorylation.
Our findings are consistent with, and substantially extend previous observations that RyR2 from ß-adrenergic stimulated mouse hearts that had a higher opening rate than unstimulated hearts , ,  at diastolic cytoplasmic Ca2+ (150 nM) and supra-physiological luminal Ca2+ (50 mM). However, the degree of RyR activation seems to depend on the presence of ATP and Mg2+. In the absence of ATP and Mg2+, isoproterenol stimulation caused a 15-fold increase in RyR2 activity and less than a 2-fold increase in the presence of ATP . Another study  reported that exercise caused a 5-fold increase in RyR activity in the presence of Mg2+. Here, we present the first investigation of how ß-adrenergic stimulation alters the various Ca2+ and Mg2+ regulation mechanisms in RyR2 and their relative roles in physiological Ca2+ release from the SR. We identify four actions of ß-adrenergic stimulation on RyR2 gating which we associate with three mechanisms: #1) a 3- to 5-fold increase in RyR2 activation by luminal Ca2+ (Figure 4) and decreased RyR2 inhibition by luminal Mg2+ (Figure 6); both actions being attributable to changes in the luminal Ca2+ binding site (L-site) . #2) diminished Mg2+ inhibition at mM concentrations (Figure 7) attributable to decreased affinity of the I1-site and possibly the A-site, and #3) increased RyR2 mean open durations (Figure 4D), attributable to a decreased rate of cytoplasmic Ca2+ inactivation (I2-site). The cytoplasmic Ca2+ activation, associated with the cytoplasmic Ca2+ activation site (A-site; ) was not altered by ß-adrenergic stimulation. This novel data show that changes in luminal Ca2+ activation make the dominant contribution to increases in Po observed at low cytoplasmic [Ca2+] whereas at higher [Ca2+], the dominant contribution comes from alleviation of cytoplasmic Mg2+ inhibition.
By fitting the data obtained here from rat RyR2 with a model previously used to account for Ca2+ and Mg2+ regulation of sheep RyR2  and extrapolating this model to physiological ionic and Ca2+ buffering conditions, we predict that different combinations of these mechanisms underlie stimulation of RyR2 activity in diastole and systole. In diastole, ß-adrenergic stimulation causes a 20-fold increase in RyR Po, mainly due to mechanisms #1 and #2 whereas in systole, it causes a smaller, 3-fold increase in Po, mediated by mechanism #2.
Western Blot analysis shows that ß-adrenergic stimulation of intact hearts induces a 4-fold increase in S2814 phosphorylation, which is at the high end of the range reported by previous studies of perfused hearts , . In addition, S2814 phosphorylation occurs within 60 s of ß-adrenergic stimulation suggesting a role in the acute changes in cardiac function induced by isoproterenol. Several studies show that S2814 phosphorylation occurs as a result of the raised cytoplasmic [Ca2+] resulting from increased heart rate during adrenergic stimulation which activates CaMKII , .
Western blots with pS2808 antibody indicated that S2808 was substantially phosphorylated before ß-adrenergic stimulation and that ß-adrenergic stimulation increased phosphorylation at this site (Figure 1). We used the DepS2808 antibody that detected dephosphorylated S2808 as a cross check for the pS2808 experiments. The experiments using DepS2808 confirmed the high basal levels of S2808 phosphorylation but not the effects of ß-adrenergic stimulation due to the low levels of antibody staining in these experiments. Our pS2808 data confirm antibody measurements of basal and ß-adrenergic-induced phosphorylation levels of S2808 in rat perfused hearts , quiescent cardiomyocytes ,  and other species , , , , . However, studies using 32P incorporation assays report larger relative increases in S2808 phosphorylation that imply lower basal phosphorylation levels than we report. Exercise in mice  and isoproterenol administered to rats  induced 3-fold increases in PKA phosphorylation in RyR2.
We did not examine phosphorylation at S2030 because we could not detect pS2030 antibody binding to RyR2. Previous characterization of several antibodies by other workers also failed to reliably measure S2030 phosphorylation . However, ß-adrenergic-induced changes in phosphorylation at S2030 were detected by others in mouse and rat using antibodies not available to us , . Potential changes in phosphorylation at S2030 would not affect our main conclusions regarding how RyR2 gating changes in response to ß-adrenergic stimulation.
A picture is emerging where phosphorylation at any of the three known phosphorylation sites on the RyR2 increases its activation by luminal Ca2+. The ß-adrenergic-induced increase in luminal Ca2+ activation, that correlated in this study with increased phosphorylation of S2814 and S2808, is similar to that reported by constitutive phosphorylation of S2030 in recombinant S2030D mice ,  and PKA-induced phosphorylation of S2808 . ß-adrenergic stimulation of the heart is substantially reduced in S2808A (dephosphorylated) transgenic mice , . These apparently conflicting findings could be reconciled if multiple phosphorylation sites have a synergistic effect on RyR2 activity and that adrenergic stimulation requires phosphorylation at both S2814 and S2808. This would not be surprising given the close proximity of S2808 and S2814 in the RyR2 sequence.
Regulation of RyR2 by phosphorylation and luminal Ca2+ and Mg2+ is controversial. Our finding that the effect of ß-adrenergic stimulation strongly correlates with CaMKII phosphorylation at S2814 and that its effect on RyR2 activity depends on the ionic conditions used in bilayer experiments appears to reconcile some conflicting reports using exogenous CamKII to phosphorylate RyRs. For example, the ß-adrenergic-induced increase in RyR2 luminal Ca2+ activation in the presence of 100 nM cytoplasmic Ca2+ was found in other studies of S2814 phosphorylation by CaMKII  and the phospho-mimetic S2814D mutation . The contradictory findings that CaMKII had either no effect on RyR2 activity , or caused a decrease in activity  were obtained using cytoplasmic Ca2+ in excess of 1 µM. Under such conditions we also find that ß-adrenergic stimulation also had no effect on RyR2 activity. We could not detect an effect of adrenergic stimulation on RyR2 inhibition by cytoplasmic Mg2+ in the presence of 100 nM cytoplasmic Ca2+; a phenomenon associated with the A-site . However, ß-adrenergic stimulation did reduce Mg2+ inhibition at high cytoplasmic Ca2+ (associated with the I1-site); consistent with the finding that exogenous CaMKII removes Mg2+ inhibition with 5–10 µM Ca2+ , . Hain et al.  found that activation of endogenous CaMKII caused inhibition of RyRs that were previously dephosphorylated by PP1. In our study, ß-adrenergic activation of RyR2 associated with CamKII activity could be because RyR2 in our experiments were not dephosphorylated prior to stimulation.
The aim of this study was to identify the effects of acute ß-adrenergic stimulation on RyR2 function that occur in a timeframe that similar to the rate of onset of ß-adrenergic stimulation in the heart. To achieve this it was necessary adopt the in vitro Langendorff perfusion method so that hearts could be snap frozen within one minute of administering isoproterenol. (Flash freezing methods have been found to accurately capture the phosphorylation state of RyR2 ). Sustained perfusion (15 min) of isolated hearts appears to reduce phosphorylation levels of RyR2 compared to hearts in situ (c.f. Fig 1A,C in ). Therefore, control hearts in our study were perfused for just long enough for heart rate to reach a steady level (5 min). This approach would minimise run-down in RyR2 phosphorylation during perfusion but at the cost of ensuring that hearts have fully recovered from explantation.
We find that the binding of antibodies for S2808 and S2814 to RyRs that were incubated with PKA and CamKII phosphorylation indicated that S2808 was mainly phosphorylated by PKA and S2814 by CamKII. In addition, these kinases appeared to phosphorylate the alternate sites to 20–50%, confirming previous antibody experiments  indicating that PKA and CamKII act on both sites. However, given that 32P incorporation experiments find that PKA and CamKII are each highly specific for S2808 and S2814, respectively , we can’t rule out the possibility that our antibodies are cross-reacting with both phosphorylated serines.
In conclusion, ß-adrenergic stimulation induces changes in at least three Ca2+ and Mg2+ regulation mechanisms in RyR2. In bilayer experiments, these changes produce changes in channel activity that depended on the experimental conditions, which can account for conflicting reports of phosphorylation-induced changes in RyR2 function. In the cell, our RyR2 gating model predicts that ß-adrenergic stimulation of RyR2 is 20-fold in diastole but only 3-fold in systole; reflecting the different mechanisms underlying stimulation in each case.
Supplementary material detailing the methods and the model for Ca2+/Mg2+ dependent gating of RyR2. The file gives additional information for methods of 1) perfusion of isolated rat hearts, 2) data acquisition and analysis, 3) SDS PAGE and Western Blotting, 4) in-vitro Exogenous PP1, PKA and endogenous CaMKII activity assays and 5) a summary of heart rates and phosphorylation levels determined by Western Blots for each heart used in this study (Table S1). Additional details of the model are included in Tables S2 and S3.
Figures of supporting information on the analysis of phosphorylation of RyR2 in Western Blots. These figures demonstrate 1) the feasibility of reprobing Western Blots, 2) the time course of PP1 incubation of SR vesicles, 3) the time course of binding by antibodies to phospho-S2808 and phospho-S2814 in response to PKA and CamKII incubation, 4) that endogenous phosphatases do not alter S2808 phosphorylation in bilayer experiments and 5) that ATP does not alter S2814 phosphorylation in bilayer experiments.
We wish to thank Paul Johnson for his assistance with the single channel recording and Peter Dosen for his assistance with perfusion of isolated rat hearts.