Acute phosphatidylinositol 4,5 bisphosphate depletion destabilizes sarcolemmal expression of cardiac L-type Ca²⁺ channel Ca_V1.2

Significance

The potent vasopressor angiotensin II is released acutely during blood pressure control, and chronically during heart failure as a physiological strategy to increase cardiac output. We propose a mechanism, whereby angiotensin II signaling in ventricular myocytes stimulates hydrolysis of a membrane phospholipid called PIP₂, triggering Ca_V1.2 channel internalization, and providing a means to acutely tune cellular excitability and modulate cardiac excitation-contraction coupling. Accordingly, this study presents data that supports a novel mechanistic role of PIP₂ as a stabilizer of Ca_V1.2 channel expression on the sarcolemma of ventricular myocytes. Moreover, we provide the evidence that physiological signaling pathways alter cardiac PIP₂ levels, and thus this work has implications for the many cardiac ion channels and exchangers regulated by PIP₂.

Abstract

Ca_V1.2 channels are critical players in cardiac excitation–contraction coupling, yet we do not understand how they are affected by an important therapeutic target of heart failure drugs and regulator of blood pressure, angiotensin II. Signaling through G_q-coupled AT1 receptors, angiotensin II triggers a decrease in PIP₂, a phosphoinositide component of the plasma membrane (PM) and known regulator of many ion channels. PIP₂ depletion suppresses Ca_V1.2 currents in heterologous expression systems but the mechanism of this regulation and whether a similar phenomenon occurs in cardiomyocytes is unknown. Previous studies have shown that Ca_V1.2 currents are also suppressed by angiotensin II. We hypothesized that these two observations are linked and that PIP₂ stabilizes Ca_V1.2 expression at the PM and angiotensin II depresses cardiac excitability by stimulating PIP₂ depletion and destabilization of Ca_V1.2 expression. We tested this hypothesis and report that Ca_V1.2 channels in tsA201 cells are destabilized after AT1 receptor-triggered PIP₂ depletion, leading to their dynamin-dependent endocytosis. Likewise, in cardiomyocytes, angiotensin II decreased t-tubular Ca_V1.2 expression and cluster size by inducing their dynamic removal from the sarcolemma. These effects were abrogated by PIP₂ supplementation. Functional data revealed acute angiotensin II reduced Ca_V1.2 currents and Ca²⁺ transient amplitudes thus diminishing excitation–contraction coupling. Finally, mass spectrometry results indicated whole-heart levels of PIP₂ are decreased by acute angiotensin II treatment. Based on these observations, we propose a model wherein PIP₂ stabilizes Ca_V1.2 membrane lifetimes, and angiotensin II-induced PIP₂ depletion destabilizes sarcolemmal Ca_V1.2, triggering their removal, and the acute reduction of Ca_V1.2 currents and contractility.

Voltage-gated L-type Ca_V1.2 Ca²⁺ channels are depolarization-triggered conduits of Ca²⁺ entry into excitable cells. In the heart, they are critical players in excitation-contraction (EC) coupling, where they provide the initial, predominantly t-tubule localized entry pathway for Ca²⁺ (1). During the peak and plateau phases of the ventricular action potential, Ca_V1.2-conducted Ca²⁺ influx activates a larger release of sarcoplasmic reticulum (SR) Ca²⁺ from clusters of type 2 ryanodine receptors (RyR2) on the juxtaposed junctional SR via Ca²⁺-induced Ca²⁺-release (CICR). The magnitude of local CICR and its summation across thousands of sites within each cardiomyocyte dictates the amplitude of the global Ca²⁺ transient and the strength of the subsequent myocardial contraction. As release of SR Ca²⁺ is graded by the amplitude of Ca²⁺ entry through Ca_V1.2 channels, alterations in the number and/or activity of Ca_V1.2 channels on the t-tubule sarcolemma tunes EC-coupling (2).

Physiological signaling cascades (3–6), post-translational modifications (7–11), and Ca²⁺ itself (12–16) are well-studied regulators of both Ca_V1.2 channel activity and sarcolemmal expression in various tissues, including the heart. However, even though Ca_V1.2 channels are embedded in a lipid environment, and despite the fact that many other ion channels and exchangers are known to be regulated by the membrane phospholipid phosphatidylinositol 4,5 bisphosphate (PI(4,5)P₂, henceforth abbreviated as PIP₂), its potential effects on Ca_V1.2 distribution, activity, and expression in cardiomyocytes have been overlooked. There is accumulating evidence that several subtypes of voltage-gated Ca²⁺ channels are regulated by PIP₂; however, the experiments underlying these findings have largely been carried out in heterologous expression systems (17–20), neurons (21), or pancreatic β-cells (22) and were not done in cardiomyocytes. These previous reports indicate that Ca_V1.2 currents are suppressed upon depletion of PIP₂ downstream of G_q-coupled activation of M₁-muscarinic receptors (M₁R) by oxotremorine-m (Oxo-m) or through activation of a voltage-sensitive 5-phosphatase DR-VSP that removes the 5′ phosphate from PI(4,5)P₂ and converts it to PI(4)P (17, 20). However, how PIP₂ regulates Ca_V1.2 remains unknown.

During G_q-coupled physiological signaling cascades, PIP₂ levels in the plasma membrane (PM) fall due to its phospholipase C (PLC)-mediated hydrolysis, generating inositol trisphosphate (IP₃) and diacylglycerol (DAG). The M₁ receptors targeted in the aforementioned Hille group studies are not expressed in ventricular myocytes but several other G_q-coupled receptors are, including angiotensin type 1 receptors (AT1R). Activated by the peptide hormone and vasoconstrictor agonist angiotensin II (AngII) (23), the AT1R/G_q/PLC signaling cascade is triggered during acute blood pressure control and chronically during hypertension and heart failure (HF) where it is implicated in cardiac hypertrophy and fibrosis (24–26). While AngII and its derivative Ang(1-7) can act on other receptors in the heart including angiotensin type 2 receptors, and Ang(1-7)/Mas receptors (27), we focus here on the effects on AT1R since this is the only pathway of the three that is G_q-coupled and thus is the most likely to directly modulate PIP₂ levels.

AngII has both indirect and direct effects on the heart. Indirectly, the effects of AngII on the vasculature change hemodynamic loading conditions. Directly, AngII has been demonstrated to affect cardiomyocyte L-type Ca²⁺ channel currents (I_Ca) in a species-dependent manner. Accordingly, the regulation of cardiac Ca_V1.2 by acute AngII treatment has been varyingly reported to cause an increase (in rabbit, cat, and cultured neonatal rat) (28–30), or a decrease in (human, chick, and cultured neonatal rat) (6, 31, 32) cardiomyocyte I_Ca. However, the underlying mechanism is unclear. Some have implicated channel phosphorylation by protein kinase C (PKC) as the regulatory driver (32, 33), while others have suggested a role for arachidonic acid and a Ca_Vβ isoform-dependent effect (31). Thus, a unifying theory that explains the I_Ca-depleting effects of AngII on cardiomyocytes remains elusive and is the focus of this study.

We hypothesize that PIP₂ depletion is the driver of the regulatory effects of AngII on cardiac I_Ca. Furthermore, we propose that PIP₂ stabilizes Ca_V1.2 channel expression in the t-tubule sarcolemma and that AngII can acutely depress cardiac excitability by stimulating PIP₂ depletion which destabilizes PM Ca_V1.2 and ultimately leads to their removal from the PM. Using electrophysiology, total internal reflection fluorescence (TIRF) microscopy and single-molecule localization microscopy, biochemistry, and lipid mass spectrometry approaches, our investigation uncovered a novel consequence of AngII signaling, that acute physiological signaling through AngII and the induction of PIP₂ depletion results in destabilization of PM Ca_V1.2 channels, triggering their removal and the acute reduction of I_Ca and contractility in the heart.

Results

Receptor-Stimulated PIP₂ Depletion Destabilizes PM Expression of Ca_V1.2 Channels.

We began our study by testing the hypothesis that PIP₂ stabilizes Ca_V1.2 channel expression in the PM (as illustrated in Fig. 1A) by examining the effects of G_q-coupled receptor-driven PIP₂ depletion on PM Ca_V1.2 channel expression in a reductionist system lacking the complexities of a cardiomyocyte. Accordingly, fluorescence imaging experiments were performed on tsA201 cells transfected with M₁R or AT1R. Cells were additionally transfected with a red fluorescent protein tagged PIP₂ probe based on the pleckstrin homology domain of PLC [PH-PLC∂1-RFP (34)], and a cyan fluorescent protein (CFP) tagged Ca_V1.2 (Ca_V1.2-CFP) to monitor PM PIP₂ and channel expression, respectively. Adapting the approach of Suh et al. (17), M₁R were activated with a saturating concentration of the agonist Oxo-m (10 µM), triggering the G_q/PLC signaling pathway and resulting in PIP₂ hydrolysis. In TIRF time series experiments application of Oxo-m triggered a robust 48.1 ± 4.7% depletion of PH_PLC∂1-RFP from the PM, which upon washout was recovered, demonstrating PIP₂ hydrolysis and resynthesis (Fig. 1 B and C). A 22.4 ± 3.1% decline in PM Ca_V1.2-CFP commenced within ~10 s of PIP₂ depletion followed by a partial recovery upon washout (Fig. 1 D and E). This delay potentially reflects the time required for channel endocytosis and recycling to begin after PIP₂ depletion. Once initiated, the channel removal and recovery phases followed a similar kinetic profile to that of the PH-fluoroprobe (channel: τ_depletion = 49.24 ± 4.54 s, τ_recovery = 118.92 ± 6.16 s; PIP₂: τ_hydrolysis = 21.67 ± 1.85 s, τ_recovery = 139.46 ± 6.83 s). These PIP₂ kinetics measurements agree with prior published values following M₁R activation in tsA201 cells (35, 36). In AT1R expressing tsA201 cells, the application of AngII (100 nM) resulted in a 39.6 ± 3.6% depletion of PH_PLC∂1-YFP signals (PIP₂ τ_hydrolysis = 40.15 ± 17.20 s). This was accompanied by an 11.1 ± 1.6% decline in PM Ca_V1.2-CFP that began ~30 s after the onset of PIP₂ depletion (Fig. 1 F–I). Taken together, these experiments indicate that Ca_V1.2 channel destabilization and removal from the PM occurs after G_qPCR/PLC-mediated PIP₂ hydrolysis.

Fig. 1.

Receptor-stimulated PIP₂ depletion destabilizes plasma membrane Ca_V1.2 expression. (A) Illustration of our hypotheses where PIP₂ stabilizes membrane expression of Ca_V1.2 (Left) and G_q-coupled receptor-stimulated PIP₂ hydrolysis destabilizes Ca_V1.2 expression triggering their endocytosis (Right). (B) Representative TIRF images showing localization of the PIP₂ biosensor RFP-PH-PLC∂1 before and after 10 µM Oxo-m treatment in M₁R and Ca_V1.2-expressing tsA201 cells (n = 12). (C) Time-course (mean ± SEM) of the Oxo-m stimulated changes in normalized RFP-PH-PLC∂1 fluorescence emission (F/F₀) in the TIRF footprint. (D) TIRF images of Ca_V1.2-CFP in the same cell before and after Oxo-m, and (E) the average time-course of changes in Ca_V1.2-CFP fluorescence emission (F/F₀) over the course of the experiments. (F) Representative TIRF images and (G) average time-course of intensity changes of YFP-PH-PLC∂1 before and after 100 nM AngII treatment in AT1R and Ca_V1.2-expressing tsA201 cells (n = 13). (H) TIRF images of Ca_V1.2-CFP in the same cell and, (I) average time-course of the experiments.

Recruitment of a Lipid Phosphatase to the PM Leads to Ca_V1.2 Endocytosis.

G_q-coupled receptor signaling cascades generate/activate several second messengers and downstream effectors, e.g., arachidonic acid and protein kinase C (PKC), that could themselves affect Ca_V1.2 channel expression. To isolate the specific effects of PIP₂ depletion on PM Ca_V1.2 expression levels, we utilized a chemical dimerization strategy to deplete PIP₂ in a targeted manner that would be independent of receptor signaling. In this previously described system, addition of rapamycin is used to dimerize an FKBP-rapamycin binding (FRB) domain-linked membrane anchor (Lyn₁₁-FRB), to an FK506 binding protein (FKBP) that is coupled to a 4′, 5′ lipid phosphatase enzyme (FKBP-4′ 5′ phosphatase) (37). This irreversible dimerization recruits cytosolic FKBP-4′, 5′ phosphatase to the PM anchor where it removes the 4′ and 5′ phosphate groups from PIP₂ to produce PI (Fig. 2A). To control for any rapamycin-specific effects, we initially recruited an enzymatically dead lipid phosphatase to the PM of transfected tsA201 cells and observed no reduction in green fluorescent protein (GFP) tagged Ca_V1.2 (Ca_V1.2-GFP) (Fig. 2 B and C). In contrast, PM recruitment of a catalytically active phosphatase produced a 12.4 ± 1.7% reduction in PM Ca_V1.2-GFP in response to rapamycin-induced PIP₂ depletion (Fig. 2 D and E). Together, these experiments show that PIP₂ depletion destabilizes PM expression of Ca_V1.2-GFP.

Fig. 2.

Receptor-independent PIP₂ depletion via chemical translocation of a lipid phosphatase destabilizes plasma membrane Ca_V1.2 expression. (A) Illustration of the FKBP–FRB rapamycin-dependent dimerization system used to irreversibly recruit an enzymatically dead (control; Left) or an active (Right) lipid phosphatase to the PM. The active enzyme depletes PIP₂ by metabolizing it into PI. The cartoon graphically illustrates the testable prediction that this would destabilize PM expression of Ca_V1.2 and trigger their endocytosis. (B) TIRF images and (C) average Ca_V1.2-GFP (F/F₀) time-course showing PM Ca_V1.2-GFP expression in transfected tsA201 cells before and after rapamycin (1 µM) induced recruitment of enzymatically dead 5′-lipid-phosphatase (n = 7). (D) TIRF images and (E) average Ca_V1.2-GFP (F/F₀) time-course showing modulation of PM Ca_V1.2-GFP expression in transfected tsA201 cells before and after rapamycin-induced recruitment of active 4′,5′-phosphatase (pseudojanin-FKBP; n = 6).

PIP₂ Depletion Triggered Ca_V1.2 Endocytosis Is Dynamin-Dependent.

AT1R are known to undergo agonist-promoted internalization in a process that involves β-arrestin and dynamin-dependent endocytosis of the receptors (38). Thus to determine whether AngII-stimulated internalization of Ca_V1.2-GFP occurred simply due to receptor-adjacent channels being caught up in AT1R sequestration, we stimulated the receptors with [Sar1,Ile4,Ile8]-AngII (SII-AngII), a β-arrestin-biased agonist of AT1R (39). SII-AngII binding to AT1R induces β-arrestin recruitment but fails to engage G-proteins thus allowing discrimination between an internalization that requires β-arrestin and one that is stimulated by G_q signaling (Fig. 3A). In proof-of-reagent experiments, confocal imaging performed on AT1R-mCherry and PH_PLC∂1-CFP expressing tsA201 cells revealed AT1R-mCherry internalization in response to SII-AngII (Fig. 3B), but PIP₂ depletion only occurred during AngII application (Fig. 3 C and D). As PIP₂ hydrolysis occurred, PH_PLC∂1-CFP lost its binding partner in the PM, was released from membrane sites and in turn accumulated in the cytosol. These results illustrate the β-arrestin bias of SII-AngII and its failure to engage G-proteins. Importantly, TIRF experiments on cells expressing Ca_V1.2-GFP revealed no change in TIRF fluorescence of the channel with the SII-AngII treatment. Moreover, after a washout period, these same cells responded to AngII, leading to a 12.7 ± 2.3% reduction in channel fluorescence (Fig. 3E). Collectively, these results indicate that AngII-stimulated Ca_V1.2 channel internalization does not occur due to β-arrestin-dependent sequestration of neighboring AT1R.

Fig. 3.

AngII/AT1R stimulated Ca_V1.2 endocytosis is dependent on dynamin, not β-arrestin. (A) Illustration of our experimental design where SII AngII was used to stimulate β-arrestin recruitment but failed to activate G-proteins (Left). (Right) CFP-tagged PH-probes were used to monitor PM PIP₂ levels, while Ca_V1.2-GFP and AT1R-mCherry were used to track the localization of the channels and receptors before and after application of agonists. (B, Top) Representative confocal images showing AT1R-mcherry localization in tsA201 cells, and (Bottom) accompanying time-course showing AT1R-mCherry “particle” tracking in the PM (blue line) and cytosolic (red line) compartments of this cell before and after application of the β-arrestin-biased AT1R agonist, SII AngII (10 µM; n = 8). (C) Confocal images showing localization of the PIP₂ biosensor PH-PLC∂1-CFP before and during application of SII AngII (Top), or AngII (Bottom). (D) Plot profiles of normalized PH-PLC∂1-CFP fluorescence intensity across the cell in each condition at the line-regions of interest (ROIs) indicated in panel C. (E) TIRF images showing Ca_V1.2-GFP localization in a representative tsA cell treated first with SII AngII, and subsequently with 100 nM AngII. (Right) Corresponding mean ± SEM time courses of the normalized Ca_V1.2-GFP intensity (n = 14). (F) Representative TIRF images and corresponding mean ± SEM time-courses showing Ca_V1.2-GFP intensity in tsA201 cell TIRF footprints during application of AngII in the presence (Top), and absence (Bottom) of the dynamin-inhibitor dynasore (10 µM; n = 5).

Next, to test the hypothesis that AngII-stimulated Ca_V1.2 channel endocytosis was dynamin-dependent (SI Appendix, Fig. S1), we pre-treated Ca_V1.2-GFP and AT1R expressing tsA201 cells with the dynamin inhibitor dynasore (80 µM). In TIRF imaging experiments, this treatment prevented AngII-stimulated Ca_V1.2 removal from the PM (Fig. 3F). The washout of dynasore and subsequent reapplication of AngII to the same cells revealed the recovery of the Ca_V1.2 channel removal response (Fig. 3F). Channel endocytosis was determined to be Ca²⁺ independent, as it persisted when Ba²⁺ was substituted for Ca²⁺ (SI Appendix, Fig. S2). These results confirm that AT1R-stimulated PIP₂ depletion destabilizes Ca_V1.2 anchoring in the PM and triggers dynamin-dependent endocytosis of the channel in tsA201 cells.

AngII Reduces Ca_V1.2 Cluster Size and Sarcolemma Expression in Ventricular Myocytes.

Turning our attention to primary cells from native tissues, we set out to determine whether a similar AngII-triggered Ca_V1.2 endocytosis exists in freshly isolated ventricular myocytes. If a similar phenomenon exists in these cells, then a testable prediction is that acute treatment with AngII should reduce Ca_V1.2 channel expression at the sarcolemma. That hypothesis was rigorously tested using a two-pronged approach. First, we examined the nanoscale distribution of immunostained Ca_V1.2 channels using super-resolution single molecule localization microscopy (SMLM). Ca_V1.2 channel cluster areas in myocytes acutely treated with AngII were 30.9 ± 4.4% smaller on average compared to controls (Fig. 4 A and B). The number of events per pixel within the cell-occupied area of the localization map provides an indication of the density of labeling and the number of channels. This measure was also reduced by 47.9 ± 14.8% in the AngII-treated cells versus controls (Fig. 4C) suggesting that acute AngII treatment reduced the number of Ca_V1.2 channels in the t-tubule sarcolemma. Similar results were obtained from female ventricular myocytes suggesting this response is conserved across both sexes (SI Appendix, Fig. S3). It is also noteworthy that this stimulated decrease in Ca_V1.2 channel clustering and expression appears confined to the t-tubules as crest-localized Ca_V1.2 clusters were not significantly altered by acute AngII (SI Appendix, Fig. S4).

Fig. 4.

Acute AngII application reduces t-tubular Ca_V1.2 cluster area and PM expression in a PIP₂-depletion–dependent manner. (A) Representative SMLM localization maps showing Ca_V1.2 channel distribution in the t-tubules of mouse ventricular myocytes with (Right) or without (Left) AngII-stimulation. Boxes indicate the location of the enlarged regions. (B) Aligned dot plots showing mean Ca_V1.2 channel cluster areas, and (C) events per pixel as an indicator of channel expression density (control: N = 4, n = 14; AngII: N = 3, n = 17). (D) Representative western blot image of the total and biotinylated cell surface fraction of Ca_V1.2, internal control alpha-tubulin, and total protein ponceau stain, in untreated control, and AngII-treated ventricular myocyte lysates. (E) Aligned dot plots showing relative Ca_V1.2 surface expression (N = 13, n = 5). (F, Top) The homeostatic phospholipid metabolism pathway with species scaled to illustrate the observed effects of AngII. (Bottom) Histograms summarizing UPLC-MS/MS measurements of PI, PIP, PIP₂, and PIP₃ in samples of untreated control or AngII-stimulated whole heart lysates, (N = 3 for each). (G) SMLM localization maps showing Ca_V1.2 channel distribution in control and AngII-treated myocytes supplemented with PIP₂ (control + PIP₂N = 3, n = 13; AngII + PIP₂N = 3, n = 16). (H) Aligned dot plots showing mean Ca_V1.2 channel cluster area and (I) the events per pixel in each condition. Error bars indicate SEM. Statistical analyses on data summarized in B, C, F, H, and I were performed using unpaired, two-tailed Student’s t-tests. Data in E were compared using a paired, two-tailed Student’s t-test.

Next, in the second prong of our approach, we biotinylated membrane proteins from isolated myocytes and quantified the biotinylated (plasma membrane-localized) fraction of the a₁1.2 Ca_V1.2 channel subunit that bound to NeutrAvidin beads in pull-down assays via immunoblotting. Here we determined that surface a₁1.2 in the AngII-treated myocytes was reduced by 58.8 ± 13.8% in comparison to controls (Fig. 4 D and E). Altogether, these complementary results using two separate techniques reveal that acute treatment of mouse ventricular myocytes with AngII leads to reduced expression of Ca_V1.2 channels at the PM.

Whole-Heart Phospholipid Species’ Levels Are Altered by Acute AngII Treatment.

To establish that AT1R/G_q/PLC-driven PIP₂ depletion drives these changes in Ca_V1.2 channel distribution, it was necessary to quantify phosphoinositide species in hearts with and without acute AngII-treatment. A simplified view of phosphoinositide metabolism is shown in Fig. 4 F, Top. We measured each of these species using phospholipid mass spectrometry and found that Langendorff perfusion of hearts with 100 nM AngII for just 5 min generated significant alterations in the levels of PIP₂ and its precursor PIP (Fig. 4 F, Bottom). Specifically, we observed a 62.8 ± 22.4% decrease in PIP and a 44.2 ± 13.0% decrease in PIP₂ (Fig. 4 F, Bottom) after AngII. These results suggest that acute, physiological elevations in AngII in the heart result in PIP and PIP₂ depletion.

AngII-Stimulated Ca_V1.2 Endocytosis in Cardiomyocytes Requires PIP₂ Depletion.

To address the question of whether the reductions in t-tubule sarcolemma-localized Ca_V1.2 were driven by PIP₂ depletion-triggered endocytosis, SMLM was performed on cardiomyocytes supplemented with PIP₂. We reasoned that if PIP₂ stabilizes Ca_V1.2 channel expression then boosting PIP₂ concentration in the membrane might reduce or eliminate the AngII-stimulated endocytosis. In line with that prediction, PIP₂ supplementation prevented AngII-mediated decreases in cluster area and expression (Fig. 4 G–I). Furthermore, myocyte treatment with the β-arrestin-biased AT1R agonist SII-AngII failed to trigger the endocytosis response (SI Appendix, Fig. S5). As discussed above, SII-AngII bound AT1R cannot engage G-proteins to activate PLC-mediated hydrolysis of PIP₂, thus these results reinforce the idea that AngII stimulation of AT1R triggers Ca_V1.2 channel endocytosis in a process that requires PIP₂ depletion.

Ca_V1.2 Channels Endocytosed Downstream of AT1R Activation Are Stored in Endosomes.

We next sought to visualize the dynamics of AngII-stimulated Ca_V1.2 endocytosis in live cells. To accomplish this, we utilized transduced myocytes isolated from mice 2 wk after their inoculation with a retro-orbital injection of the AAV9-Ca_Vβ₂a-paGFP. The transduced expression construct encodes an auxiliary subunit that finds and binds endogenous pore-forming α_1C-subunits on a 1:1 basis and thus acts as a fluorescent biosensor of Ca_V1.2 channels as has been extensively verified in other studies (4, 5). After photoactivation with 405 nm light, clusters of GFP-tagged channels were visualized and a time series of experiments was performed to track biosensor-tagged channels before and during application of AngII via 150 nm penetration depth TIRF microscopy, an imaging technique that allows detection of the surface and initial portion of the t-tubule sarcolemma. Image analysis was used to identify and quantify subpopulations of channels that were inserted, removed, or static during AngII treatment relative to the control period (Fig. 5 A and B). Our results indicated a strong bias toward channel removal during AngII treatment (Fig. 5B). Since the population of channels at the membrane at any given time is dictated by the balance between their insertion and removal, enhanced removal relative to insertion will lead to reduced expression of channels at the membrane. Accordingly, we found a 28.8 ± 5.2% reduction in biosensor-tagged channels in the TIRF footprints of myocytes after AngII (Fig. 5C). This evident dynamic shift toward channel internalization invites the question, where are these channels going?

Fig. 5.

AngII promotes Ca_V1.2 channel removal from cardiomyocyte sarcolemmas and sequestration in early endosomes. (A) Representative TIRF images of GFP fluorescence emission from Ca_Vβ_2a-paGFP transduced cardiomyocytes before (Top) and after 100 nM AngII (Bottom) (N = 5, n = 14). Images illustrating stable, inserted, and endocytosed channel populations are shown to the Right. Time courses of the changes in Ca_Vβ_2a-paGFP intensity in individual ROIs (indicated by circles on TIRF images) are represented below. (B and C) Histograms summarizing the percentage of static, inserted, and removed channel populations after AngII (B) and the percentage change of TIRF-footprint Ca_Vβ_2a-paGFP fluorescence after AngII (C). (D) Two-color Airyscan super-resolution images showing distributions of Ca_V1.2 and EEA1⁺ early endosomes in representative control (Left) and AngII-stimulated (Right) myocytes. Binary colocalization maps (Bottom) display pixels in which Ca_V1.2 and endosomal expression fully overlap. (E) Histogram summarizing the percentage colocalization between EEA1 and Ca_V1.2 in control (N = 3, n = 14) and AngII-stimulated (N = 3, n = 13) conditions. Statistical analyses on data summarized E was performed using unpaired, two-tailed Student’s t-tests. Error bars indicate SEM.

Based on our prior work (4), we hypothesized that internalized channels may populate an endosomal reservoir. To determine the acute fate of these channels in response to AngII, we examined the distribution of Ca_V1.2 on EEA1-positive early endosomes, Rab7-positive-late endosomes, and Rab11-positive recycling endosomes. Two-color Airyscan super-resolution images were taken on control or AngII-treated AMVMs immunostained for Ca_V1.2 and these endosomal markers. Colocalization analysis revealed a 10.7 ± 4.2% increase in Ca_V1.2 localized on early endosomes treated with AngII (Fig. 5 D and E). However, there was no significant difference in the percent of Ca_V1.2 on recycling endosomes (SI Appendix, Fig. S6 A and B) and late endosomes (SI Appendix, Fig. S6 C and D) between the two groups. This suggests that after only 5 min of AngII signaling, internalized channels are gathered on early endosomes but have not yet been shuttled toward recycling or degradation pathways.

AngII Attenuates I_Ca, PIP₂, and Ca²⁺ Transient Amplitudes in Cardiomyocytes.

We hypothesized that the apparent deficit in PM Ca_V1.2 channel expression and clustering would ultimately decrease functional output. To test that, we performed perforated patch clamp experiments and recorded I_Ca from freshly isolated myocytes, before and during acute AngII-treatment. Accordingly, application of AngII caused an ~20% decrease in I_Ca density that occurred with a τ of 0.69 min (Fig. 6 A and B). If this AngII-induced current reduction is triggered by PIP₂ depletion and destabilization of Ca_V1.2, then a testable prediction is that PIP₂ hydrolysis should occur on a similar timescale. To test this, we transduced ventricular myocytes with adenovirus-packaged RFP-PH_PLCδ1. After 48 h in culture, the transduced PH-probe biosensor was predominantly localized to the sarcolemma of the cardiomyocytes (Fig. 6C). Time series confocal experiments revealed an AngII-induced reduction in membrane localization of RFP-PH_PLCδ1 due to PIP₂ hydrolysis and a subsequent recovery upon washout as PIP₂ was resynthesized (Fig. 6D). The τ of PIP₂ hydrolysis was, at 0.42 min, slightly faster than that of the AngII-triggered current reduction. Superimposition of the current decay onto the PH-probe measurements revealed the initiation of PIP₂ hydrolysis slightly precedes the onset of I_Ca decay.

Fig. 6.

AngII depression of ventricular myocyte Ca_V1.2 currents and Ca²⁺ transient amplitude occurs coincidently with PIP₂ depletion. (A) Representative L-type I_Ca elicited every 6 s by a 300 ms depolarization step from −40 mV to 0 mV. Black traces show the current before, and blue traces the current after AngII. (B) Diary plot of normalized Ca²⁺ current density summarizing the results from N = 4, n = 6 cells. (C) Confocal images of a cardiomyocyte transduced with Ad-RFP-PH_PLCδ1 before and during AngII. (D) Time course of the changes in membrane RFP-PH_PLCδ1 intensity (F/F₀) in the same cell before, and during AngII and its recovery on washout. Representative of N = 3, n = 4 cells. Time course of the changes in I_Ca from panel B are superimposed for ease of comparison. (E) Representative EFS-evoked Ca²⁺ transients recorded under control (black traces) and AngII-stimulated conditions (blue traces). Arrows indicate EFS pulse application at 1 Hz. Corresponding line scans from the fluo4-AM-loaded cells appear below. (F) Histogram summarizing Ca²⁺ transient amplitude (N = 3, n = 16). Data were analyzed using paired Student’s t-tests. Error bars indicate SEM.

Finally, to investigate the effects of AngII on EC-coupling, we recorded Ca²⁺ transients from freshly isolated myocytes, before and during acute AngII-treatment. Accordingly, AngII elicited a a 15.8 ± 3.3% reduction in EFS-stimulated Ca²⁺ transient amplitude (Fig. 6 E and F). In line with our hypothesis, these data collectively show that acute AngII-stimulation tunes and diminishes EC-coupling and does so on a timescale that closely aligns with that of PIP₂ hydrolysis.

Discussion

Our data show that PIP₂ stabilizes cardiac Ca_V1.2 channel PM lifetime and that AngII-induced PIP₂ depletion destabilizes PM Ca_V1.2 clusters, triggers their dynamin-dependent endocytosis, results in the acute reduction of whole-cell I_Ca, and decreases Ca²⁺ transient amplitudes, which diminishes myocardial excitability and contractility. We provide evidence for this phenomenon using both a heterologous expression system and primary ventricular myocytes. In transfected tsA201 cells, both G_qPCR-mediated and receptor-independent PIP₂ depletion triggered Ca_V1.2 removal from the PM. In myocytes, AngII/AT1R signaling triggered sarcolemmal PIP₂ hydrolysis, reduced Ca_V1.2 expression and cluster area, and resulted in the stimulated removal of Ca_V1.2 from the PM to early endosomes. PIP₂ reinforcement stabilized Ca_V1.2 channel clusters and their sarcolemmal expression. β-arrestin biased AT1R agonism also eliminated this response, specifically implicating G_q activation. Moreover, dynamin inhibition abrogated the internalization response, suggesting internalization proceeded via dynamin-dependent endocytosis.

The finding that Ca_V1.2 endocytosis is dynamin-dependent agrees with published work from several other groups although none were testing the triggering of this event by AngII-mediated PIP₂ depletion (12, 40, 41). In our hands, endocytosed Ca_V1.2 were sequestered into early endosomes but not recycling or late endosomes. Previous work from our group identified an endosomal reservoir of Ca_V1.2 channels on early and recycling endosomes that can be mobilized to the t-tubule sarcolemma of cardiomyocytes during periods of acute stress via Rab4-dependent fast and Rab11-dependent slow recycling pathways (4), enhancing Ca_V1.2 cluster areas and channel expression in t-tubule sarcolemma (5), and facilitating a positive inotropic response to tune cardiac performance to meet physiological demands. Results from the current study reveal activation of AT1R has essentially the opposite effect, invoking an image of a rheostatic system whereby cardiac EC-coupling can be tuned up or down by manipulating sarcolemmal expression of Ca_V1.2 channels via selective activation of different receptor populations. While our work rules out the possibility that the acute Ca_V1.2 internalization response we report occurs due to their entrapment in β-arrestin-dependent sequestration of nearby AT1R, others have shown that longer, 1 h applications of AngII produce Ca_V1.2 internalization that does involve β-arrestin (6). We found that AngII-stimulated Ca_V1.2 internalization occurred predominantly in t-tubules and not in the sarcolemma crest. It is possible that AT1R are more closely associated with t-tubule Ca_V1.2 channels than with crest populations, creating signaling microdomains. However, to our knowledge, there are no specific antibodies against AT1R and thus this idea remains speculative.

Prior studies have indicated that neuronal Ca_V1.2 channel endocytosis occurs upon Ca²⁺ influx through the channels as the incoming Ca²⁺ binds to calmodulin (CaM) forming Ca²⁺. CaM that competes with α-actinin for a common binding site on the α_1C C-terminal (12). In that model, binding of α-actinin to an α_1C surface-targeting motif promotes Ca_V1.2 channel expression on the PM. Ca²⁺.CaM outcompetes α-actinin, displacing it from the channel and favoring endocytosis. Our model describes what appears to be a separate process since it occurs independently of Ca²⁺ influx, as shown by its persistence with Ba²⁺ substitution. We find that PIP₂ exerts a stabilizing influence on Ca_V1.2 surface expression and that its depletion by either G_q-receptor stimulation or PM recruitment of a lipid phosphatase favors channel endocytosis. This implies that PIP₂ must bind to the channel complex. Rosetta structural modeling has predicted the presence of at least one putative PIP₂ binding site formed by four arginines within the I-II loop α-helix region of Ca_V1.2 (42). Interestingly, neutralization of these residues removed PM binding of the linker and decreased on-gating current (Q_on), revealing that there were fewer functional channels at the PM. Thus, there appears to be a role for PIP₂ in supporting calcium channel activity and expression at the cardiac sarcolemma. Intriguingly, the putative site is found on the I-II loop that also contains the gain-of-function Timothy syndrome Ca_V1.2 mutation site at G406R (G436R in mice). Replacing a neutral glycine with a positively charged arginine might further enhance binding to negatively charged phospholipids at this interface. In our model, enhanced PIP₂ binding would favor stabilization of Ca_V1.2 channels at the PM and supporting that idea, we have previously reported that G436R mutant gain-of-function Ca_V1.2 channels form extremely large, pronounced clusters in ventricular myocytes (43). In Ca_V2.2 channels, high-resolution cryo-EM structures in the presence of a pore-blocking pain-killer called Ziconotide revealed a PIP₂ binding site that stabilizes the down conformation of voltage sensitive domain II (44). This would favor the closed state of the channel and limit Ca²⁺ influx. However, over 50% of the residues in that site are Ca_V2-specific, so it remains to be determined whether a homologous site exists in Ca_V1.2.

Our study provides a quantitative analysis of phosphoinositide species in the heart and reveals that PIP₂ levels measurably change during physiological G_q-protein coupled receptor signaling in cardiomyocytes. Specifically, we show that levels of PIP₂ and its precursor PIP are both significantly reduced by acute application of AngII. Further experiments on cultured adult mouse ventricular myocytes transduced with an adenovirus-packaged fluorescent PIP₂ biosensor (PH-RFP) revealed the kinetics of AngII-stimulated sarcolemmal PIP₂ depletion. The hydrolysis of PIP₂ was concordant with the observed reduction in I_Ca supporting the idea that PIP₂ depletion destabilizes Ca_V1.2 channels. The effects of G_q/PLC-dependent pathway activation were previously studied in guinea pig ventricular myocytes but in contrast to our results using different experimental approaches, the tested agonists (endothelin, phenylephrine, and adenosine) elicited no detectable change in total cardiac PIP₂ (45). Given that PIP₂ is a modulator of many cardiac ion channels and transporters including Ca_V1.2 (17), KCNQ1/E1 (46), K_ATP and Na⁺-Ca²⁺ exchanger [NCX (47)], our findings have a broad impact. It remains to be determined whether PIP₂ exerts a similar stabilizing effect on Ca_V1.2 in other excitable tissues, including smooth muscle, neurons, and pancreatic islets. Specifically in the heart, elevated production of AngII is associated with cardiac hypertrophy, remodeling, and HF (48–51). Future studies should examine the effects of chronic AngII on phosphoinositide populations in the heart and the ramifications for cardiac ion channel function.

Our finding that PIP₂ depletion destabilizes Ca_V1.2 expression potentially explains prior results indicating voltage-gated Ca²⁺ channel current depletion by activation of M₁Rs or voltage-sensitive phosphatase (DR-VSP) (17, 20). In favor of this stabilization hypothesis, PIP₂ supplementation is known to reduce P/Q-type Ca_V2.1 channel rundown in Xenopus oocytes, while sequestering PIP₂ with antibodies increased rundown (19). Rundown describes the poorly understood process of ion channel current decline observed in whole-cell and excised-patch clamp recordings over time. The endocytosis of Ca_V1.2 channels from the membrane has also been proposed as a factor underlying rundown (12). The supplementation of patch bathing or cell dialyzing solutions with MgATP, CaM, or crude cytoplasmic extracts from cardiac cells (52) have all been reported to reduce I_Ca rundown (53, 54). MgATP supports PIP₂ resynthesis by activating lipid kinases (19, 46, 55), while Mg²⁺ can guard against PIP₂ hydrolysis by a charge shielding effect (56). CaM has also been reported to promote PIP₂ generation in cardiac cell membranes although the underlying mechanism underlying this effect is not understood (57). Accordingly, each of these additives support PIP₂ synthesis or maintenance and appear to stabilize Ca_V1.2 channel currents, supporting our hypothesis.

Prior studies have shown PIP₂ suppression of Ca_V currents dependent on the associated Ca_Vβ-subunit isoform with the palmitoylated Ca_Vβ_2a conferring the least sensitivity to PIP₂ depletion (58). Our model implies the palmitoylation may provide an additional membrane anchor that stabilizes Ca_V1.2 and reduces its sensitivity to PIP₂ depletion. To that end, AngII inhibition of Ca_V1.2 currents has been shown to be less robust when co-expressed with Ca_Vβ_2a versus other Ca_Vβ subunits (31). In the heart, the most abundant isoform is Ca_Vβ_2b, a splice variant that lacks the consensus site for palmitoylation (59, 60) and thus cardiac Ca_V1.2 channels may be vulnerable to PIP₂ depletion. Dynamic imaging of Ca_Vβ_2a-paGFP transduced myocytes herein revealed a bias toward channel endocytosis with acute AngII-treatment but if palmitoylation of Ca_Vβ_2a does in fact confer additional stability that reduces the modulatory effects of PIP₂ depletion on PM Ca_V1.2 expression levels, it is possible that these results represent an underestimation of the true effects in vivo.

In summary, our results show that PIP₂ stabilizes Ca_V1.2 channel expression at the PM and that acute AngII can depress cardiac EC-coupling by stimulating PIP₂ depletion and destabilization of PM Ca_V1.2, leading to their endocytosis from the PM. Given that PIP₂ is present in every PM of every cell in the body, our findings have broad ramifications not only for cardiac physiology but for the myriad excitable cells that express Ca_V1.2.

Materials and Methods

All animal handling and procedures adhered to the NIH Guide for the Care and Use of Laboratory Animals (UC Davis) and were approved by the local Institutional Animal Care and Use Committee. Ventricular myocytes were isolated from 3 to 6-mo-old C57BL/6J mouse hearts via retrograde Langendorff perfusion as previously described (4, 5) and used in microscopy, Ca²⁺ transient, and/or patch-clamp electrophysiology experiments. In surface biotinylation experiments, isolated myocytes from untreated and AngII-treated hearts were biotinylated, lysed, and probed for protein content using western blot as previously described (61, 62). Phosphoinositide species were quantified using ultra-high-pressure liquid chromatography coupled to tandem mass spectrometry (UPLC-MS/MS) as previously described (63, 64). N and n represent the number of animals and number of cells, respectively. Data are reported as mean ± SEM. Datasets were compared using paired or unpaired Student’s t tests, one-way ANOVAs, or two-way ANOVAs with Tukey’s multiple comparisons post-hoc tests as stated in the figure legends. P < 0.05 was considered statistically significant. Detailed methods can be found in the SI Appendix.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.