Phosphoinositide 3-Kinase Isoforms Selectively Couple Receptors to Vascular L-Type Ca²⁺ Channels

In striated and smooth muscles, voltage-gated L-type Ca²⁺ channels represent the major pathway for Ca²⁺ entry and play a crucial role in excitation-contraction coupling. The regulatory role of L-type channels is enhanced by the modulation of this Ca²⁺ current by a large variety of hormones and mediators mainly through protein kinase activation.^1–4 Angiotensin II activates Ca²⁺ entry by stimulating L-type Ca²⁺ channels through Gβγ-sensitive phosphoinositide 3-kinase (PI3K) and protein kinase C (PKC) in venous myocytes^5,6 and through activation of tyrosine kinase and PI3K in A7r5 smooth muscle cells.⁷ Neuronal L-type Ca²⁺ channels are also stimulated in a PI3K-dependent mechanism activated on stimulation of the tyrosine kinase IGF-1 receptor.⁸ Platelet-derived growth factor (PDGF), also acting on a receptor tyrosine kinase (RTK), has been reported to stimulate vascular voltage-gated Ca²⁺ channels.⁹ Moreover, both Ca²⁺ entry and PI3K activation in vascular cells have been reported to be necessary for PDGF-induced and angiotensin II–induced DNA synthesis.^10–13

PI3Ks form an ubiquitously expressed enzyme family that phosphorylates membrane inositol lipids in the D3 position of the inositol ring. Class I PI3Ks are capable of phosphorylating in vitro phosphatidylinositol, phosphatidylinositol-4-phosphate, and phosphatidylinositol-4,5-bisphosphate, whereas in vivo phosphatidylinositol-3,4,5-trisphosphate [PtdIns(3,4,5)P₃] and its metabolites seem the predominant products.^14,15 PI3Ks also possess a protein kinase activity, which is assumed to regulate their own lipid kinase activity for PI3Kα and δ or to regulate the mitogen-activated protein kinase signaling pathways as reported for PI3Kγ.¹⁶ Many cell-surface receptors, including RTKs or G protein–coupled receptors (GPCRs), activate class I PI3Ks, and a large variety of cellular functions are regulated by the lipid products generated by these enzymes.

Four class I enzymes have been identified in humans and other mammals. These are divided into two subclasses (Ia and Ib) on the basis of their noncatalytic subunits. The Ia subgroup consists of the classical p110α and two additional, closely related enzymes, p110β and p110δ. The p110α and p110β isoforms both display a broad tissue distribution in adults, whereas p110δ is mainly expressed in hematopoietic cells. All class Ia enzymes are associated with a p85 adapter subunit to form a heterodimeric complex. All three isoforms of this class are activated by the binding of specific phosphotyrosyl motifs to the two SH2 domains of the regulatory subunits. In addition, activation of PI3Kβ by direct interaction of Gβγ with p110β has been reported.^17–19

The sole member of class Ib, p110γ, is activated by βγ subunits of heterotrimeric G proteins, which are released on activation of 7-transmembrane domain receptors. The p101 subunit, which bears no resemblance to any other known proteins, modulates the sensitivity of the kinase toward Gβγ.^18,20 PI3Kγ has been shown to be expressed in hematopoietic cells as well as in exocrine glands, heart, kidney, and bovine aortic endothelial cells.^20–22

Because there is increasing evidence that different isoforms of the enzyme have distinct functions,^23–25 the goal of the present study was two-fold. First, we determined which of the four class I PI3K isoforms are able to stimulate vascular L-type Ca²⁺ channels by comparing the Ca²⁺ channel current densities in vascular myocytes infused with purified recombinant dimeric PI3Ks through the patch pipette. Second, we studied the recruitment of endogenous class Ia and class Ib PI3Ks by RTK and GPCR. We showed that efficiency of endogenous class Ia PI3Ks in transducing stimulation of Ca²⁺ channels depends on the expression of the β isoform, which increases in cultured myocytes.

Materials and Methods

Cell Preparation

Isolated myocytes were obtained from Wistar rat portal vein by enzymatic dispersion, as described previously.²⁶ Cells were seeded on glass slides in physiological solution and used on the same day or maintained 4 days in primary culture in M199 medium containing 5% FCS, 2 mmol/L glutamine, 1 mmol/L pyruvate, 20 U/mL penicillin, and 20 μg/mL streptomycin.

Membrane Current

Voltage-clamp and membrane current recordings were made with a standard patch-clamp technique using a List EPC-7 patch-clamp amplifier (Darmstadt-Eberstadt). Whole-cell recordings were performed with patch pipettes having a resistance of 2 to 4 MΩ. Membrane potential and current records were stored and analyzed using P-clamp system (Axon Instruments). L-type Ba²⁺ currents were elicited by depolarizations to +10 mV from a holding potential of −40 mV and measured 4 to 5 minutes after breakthrough into the whole-cell recording mode and were digitally corrected for leakage current. Cell capacitance was determined in each cell tested by imposing 10-mV hyperpolarizing steps from the holding potential (−40 mV). Peak current density was expressed as the maximal amplitude of the Ba²⁺ current per capacitance unit (pA/pF). All experiments were performed at 30±1°C.

Solutions

The physiological solution used to record Ba²⁺ currents contained (in mmol/L) NaCl 130, KCl 5.6, MgCl₂ 1, BaCl₂ 5, glucose 11, and HEPES 10, pH 7.4 with NaOH. The basic pipette solution contained (in mmol/L) CsCl 130, EGTA 10, ATPNa₂ 5, MgCl₂ 2, and HEPES 10, pH 7.3 with CsOH. BSA (0.1%) was added in the pipette solution to increase protein infusion and had no effect by itself on the current charge density. Gβγ proteins were stored in a solution containing (in mmol/L) Tris 20, EDTA 1, CHAPS 11, and β-mercaptoethanol 20. PI3Ks were stored in a solution containing (in mmol/L) Tris 50, NaCl 100, and DTT 10 for the GST-tagged proteins and an additional glutathione 10 for the His-tagged PI3Kβ. These solutions were diluted 50- to 200-fold in the final pipette solution and did not change either the Ba²⁺ charge densities or the peak Ba²⁺ current densities.

Purified Proteins

Construction of recombinant baculoviruses for expression of PI3K subunits was described previously.^21,27,28 For protein expression, cells were incubated at a multiplicity of infection of 1 virus per cell. Subunits of heterodimeric PI3Ks were coexpressed at equal multiplicity of infection numbers in Sf9 cells and purified as previously described.¹⁸ For functional studies, we used GST-p110α and -p110δ fusion proteins and hexa-His–tagged p110β coexpressed with p85α; the p110γ protein was coexpressed with GST-p101 fusion protein. Gβγ-proteins were purified from bovine brain as previously described.²⁹

Western Blot Analysis

Microsomal proteins were prepared from rat portal vein media homogenates and treated with Laemmli sample buffer containing 5% β-mercaptoethanol, boiled for 10 minutes, and separated by SDS-PAGE (10% separating gel with 4% stacking gel). The resolved proteins were transferred to polyvinylidene difluoride (Bio-Rad) membrane (1 hour at 100 V). Polyvinylidene difluoride membrane was then blocked for 1 hour with 3% BSA in PBS complemented with 0.1% Tween 20 (PBS-T) and then incubated overnight with the primary anti-PI3K isoform antibodies at 2 μg/mL. After extensive washes in PBS-T, membranes were incubated for 2 hours with a 0.4-μg/mL dilution of peroxidase-coupled anti-rabbit or anti-goat IgG in PBS-T complemented with 3% BSA. Specific antigen detection was performed using the Bio-Rad Opti 4CN Substrate Kit. Gels were analyzed with KDS1D 2.0 software (Kodak Digital Science). Immunoblot analysis of purified proteins was detailed elsewhere.²⁸

Immunocytochemistry

Myocytes were washed with PBS, fixed with 4% (vol/vol) formaldehyde for 30 minutes at 20°C, and permeabilized in PBS containing 5% FCS and 0.1% (wt/vol) saponin for 30 minutes. Cells were incubated overnight at 4°C plus 1 hour at 20°C with the same solution containing the specific rabbit anti-PI3K isoform antibody at 2 μg/mL. Then cells were washed (4×10 minutes) in PBS containing 5% FCS and 0.1% (wt/vol) saponin and incubated with goat anti-rabbit or donkey anti-goat IgG conjugated to fluorescein isothiocyanate (diluted 1:500) in the same solution for 2 hours at 20°C. After extensive washes (4×10 minutes) in PBS, cells were mounted in Vectashield (Biosys). Images of the stained cells were obtained with a confocal microscope (MRC 1000, Bio-Rad). Cells were compared with each other by keeping acquisition parameters (such as gray values, exposure time, and aperture) constant.

Chemicals

PDGF-BB was from Sigma-RBI. Angiotensin II was from Neosystem Laboratories. The doubly tyrosine-phosphorylated peptide used in this study (CGGY(P)MDMSKDESVDY(P)VPMLDM) was based on that of the human platelet-derived growth factor receptor³⁰ and was kindly donated by Dr Andreas Steimeyer (Schering AG, Berlin, Germany). A nonphosphorylated peptide from the same supplier was used as a control and had no effect. Rabbit and goat anti-p110 isoform antibodies (sc-1331 and sc-7174 for p110α, sc-603 and sc-7175 for p110β, sc-7176 for p110δ, and sc-7177 for p110γ) and rabbit anti-p85 antibody (sc-423) were from Santa Cruz Biotechnology (Santa Cruz, Calif). The fluorescein isothiocyanate–conjugated goat anti-rabbit IgG and donkey anti-goat IgG were from Interchim-Jackson Immunoresearch Laboratories (Montluçon, France).

Data Analysis

All values are given as mean±SEM. A Student’s t test was performed to estimate the significance of the differences between mean values. A value of P<0.05 was considered significant.

Results

Effects of PI3K Isoforms on L-Type Ca²⁺ Channels

We investigated the capability of all four PI3K isoforms to stimulate vascular L-type Ca²⁺ channels by intracellular infusion of purified recombinant heterodimeric enzymes that have been checked for their enzymatic activities, as reported in a previous study.¹⁸

First, we used relatively high concentrations (1 nmol/L) of PI3K to stimulate the L-type Ca²⁺ channel current in freshly isolated portal vein myocytes. As illustrated in Figure 1, intracellular infusion of all four heterodimeric isoforms through the patch pipette resulted in an increase (about 2-fold) of the maximal peak current density measured 4 to 5 minutes after breakthrough by the whole-cell recording mode. This effect was not observed when the PI3Ks were boiled (95°C, 30 minutes) before intracellular application, thus showing that the stimulation of the current was strictly related to the integrity of the protein (Figure 1). These control experiments together with the measurements of current densities in cells infused with the protein buffers alone (data not shown) attested to the absence of any nonspecific effect. The peak Ca²⁺ channel current densities were 5.2±0.8 pA/pF (n=5) in control cells, 4.8±0.9 pA/pF (n=5) in cells infused with the buffer of the GST-tagged proteins (PI3Kα, PI3Kδ, and PI3Kγ), and 5.1±0.6 pA/pF (n=5) in cells infused with the buffer of His-tagged proteins (PI3Kβ).

• Download figure
• Download PowerPoint

Although a dose-response of PI3K-induced stimulation of Ca²⁺ channels was not fully achieved because of the difficulty to control the exact amount of PI3K in the cytoplasm of the patch-clamped cells, we noted a clear difference between 0.1 nmol/L and 1 nmol/L of PI3Ks in the pipette solution, because 0.1 nmol/L PI3Ks failed to increase the Ca²⁺ channel current densities (Figure 2).

• Download figure
• Download PowerPoint

Gβγ- or Phosphotyrosyl-Activated Effects of PI3K Isoforms

Class I PI3Ks have been shown to be activated in vitro by Gβγ or phosphotyrosyl peptides (Ptyr) resembling an intracellular domain of the PDGF receptor. We have previously shown that 400 nmol/L Gβγ maximally stimulates the Ca²⁺ channel current in vascular myocytes.⁶ In this study, we show that in freshly isolated vascular myocytes, Ptyr was unable to stimulate Ca²⁺ channel currents (Figure 2A). In contrast, when this peptide was coinfused with a low concentration (0.1 nmol/L) of any of the p85-associated p110 subunits, it did increase the Ca²⁺ channel peak current density (Figure 2B). As controls, the unphosphorylated peptide (Tyr) failed to stimulate any PI3K-mediated effect on L-type Ca²⁺ channel, and when the Ptyr was coinfused with low concentrations of PI3Kγ, the mixture did not affect the Ca²⁺ channel peak current density (Figure 2B).

The four PI3K heterodimers were then assayed for their Gβγ sensitivity. As previously described, Gβγ stimulated L-type Ca²⁺ channel activity in a concentration-dependent manner,⁶ and even a low concentration (50 nmol/L) of Gβγ slightly increased the current in some cells, although the mean increase in Ca²⁺ channel current density was not significant (Figure 3A). Therefore, the study of the effect of exogenous Gβγ associated with various exogenous PI3Ks on Ca²⁺ channel currents was performed after pretreatment of the cells with 100 nmol/L wortmannin for 1 hour to inhibit Gβγ-activated endogenous PI3Ks⁶ (Figure 3A). In these conditions and when associated with PI3Kβ or PI3Kγ, 50 nmol/L Gβγ maximally stimulated Ca²⁺ channel currents, whereas it was ineffective when associated with PI3Kα or PI3Kδ (Figure 3B). Because PI3Kβ was also activated by Ptyr, we examined the combined effects of Gβγ and Ptyr but we did not note additional stimulation of Ca²⁺ channels when compared with the effects of one stimulus or the other (Figure 3B).

• Download figure
• Download PowerPoint

Effect of Cell Culture on PI3K-Induced Stimulation of L-Type Ca²⁺ Channels

Because the role of PI3Ks in transducing proliferative signals has been well-documented, we compared the Gβγ- and Ptyr-mediated stimulation of Ca²⁺ channels in freshly isolated and cultured vascular myocytes. As mentioned above, in freshly isolated myocytes (day 0), intracellular infusion of 400 nmol/L Gβγ increased the peak Ca²⁺ channel current densities to the same extent as 1 nmol/L of heterodimeric PI3Kγ, whereas 1 μmol/L Ptyr was without effect (Figure 4, left). In contrast, in myocytes cultured for 4 days (day 4), Ca²⁺ channel current densities were similarly increased by Ptyr, Gβγ, and PI3Kγ (Figure 4, right), whereas the unphosphorylated peptide (Tyr) was ineffective. Moreover, pretreatment of the cells with wortmannin (100 nmol/L for 1 hour) before patch-clamp experiments abolished Gβγ- as well as Ptyr-mediated stimulations of Ca²⁺ channels (n=12; data not shown).

• Download figure
• Download PowerPoint

In agreement with these effects of Gβγ and Ptyr peptide, we have previously shown that in freshly isolated and cultured myocytes, angiotensin II stimulates Ca²⁺ channel current via Gβγ and PI3K.^5,6 In this study, we show that extracellular application of PDGF-BB, a RTK ligand, failed to increase the Ca²⁺ channel current in freshly isolated myocytes (Figure 5), whereas it did significantly increase peak current density in cultured myocytes by 39±4.2% (n=7; Figure 5). The ineffectiveness of PDGF-BB in freshly isolated myocytes was not related to the absence of PDGF receptors on the membrane of the cells after enzymatic dissociation, because when the cells were infused with low concentrations (0.1 nmol/L) of either PI3Kα or PI3Kβ, external application of PDGF-BB produced a significant increase in peak Ca²⁺ channel current density (5.6±1.8 pA/pF in control cells versus 10.1±2.3 pA/pF in PDGF-BB–stimulated cells, n=12; data not shown).

• Download figure
• Download PowerPoint

Effect of Cell Culture on PI3K Isoform Expression

Different patterns of expression of PI3K isoforms could provide an explanation for the absence of effect of PDGF-BB and Ptyr in freshly isolated myocytes. Therefore, we performed immunostaining experiments with antibodies specifically targeting each p110 catalytic subunit on freshly isolated and cultured portal vein myocytes.

First, we verified the specificity of the anti-p110 subunit antibodies by Western blot analysis, showing that each antibody specifically recognizes only one of the four recombinant p110 catalytic subunits (Figure 6A). The same antibodies have then been used to determine the p110 subunits expressed in rat portal vein myocytes. Western blot performed on portal vein media homogenates revealed that only p110α and p110γ were expressed (Figure 6B). The absence of p110β was confirmed in Western blot by using a second anti-p110β antibody (sc-603; data not shown) and by immunostaining. As illustrated in Figure 7, freshly isolated myocytes expressed only two isoforms of PI3Ks, namely p110α and p110γ subunits, whereas p110β subunit was not detected by FITC-labeled immunostainings. When cells were cultured for 4 days, the p110β subunit was clearly immunostained by two different antibodies (sc-7175, Figure 7 and sc-603, data not shown) in addition to p110α and p110γ. We did not use the antibody targeting p110δ in immunostaining, because Western blot analysis revealed that this antibody weakly recognized several bands but no p110 in rat portal vein homogenates (Figure 6B) and because expression of PI3Kδ has been reported to be restricted to hematopoietic cells.²⁷

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Because one member of the class Ia PI3K (p110α) was expressed in freshly isolated myocytes that did not display PTyr-induced Ca²⁺ channel stimulation, we verified that a class Ia regulatory subunit was present in these cells. Immunostainings performed with an antibody targeting the regulatory subunits (anti-p85) revealed that at least one of the regulatory subunits was expressed at day 0 as well as day 4 (Figure 7), suggesting that a lack of regulatory subunit could unlikely explain the lack of PDGF-BB–mediated effect on freshly isolated myocytes.

To determine causality between the appearance of PI3Kβ in cultured cells and the ability of PDGF-BB to stimulate L-type Ca²⁺ channel in these cells, we performed experiments with anti-PI3K antibodies. When the anti-PI3Kβ antibody was infused into the intracellular medium through the patch pipette, PDGF-BB was not able to stimulate Ca²⁺ channels (Figures 8A and 8B). In contrast, infusion of anti-PI3Kα (Figure 8B) or boiled anti-PI3Kβ antibody (not shown) did not affect the stimulation of Ca²⁺ channel by PDGF-BB. The efficiency of anti-PI3Kα antibody to inhibit recombinant PI3Kα-mediated stimulation of Ca²⁺ channels was shown in Figure 8C.

• Download figure
• Download PowerPoint

Discussion

Our results show L-type Ca²⁺ channel regulation by different PI3Ks depending on the agonist and on the stage of cell culture. The main findings are that (1) all four recombinant class I PI3Ks are potentially able to stimulate vascular L-type Ca²⁺ channels; (2) the endogenously expressed PI3Kα does not transduce L-type channels stimulation; and (3) expression of PI3Kβ in cultured myocytes is responsible of Ptyr-mediated and PDGF-BB–mediated regulation of L-type channels by PI3Ks.

The first finding is supported by a series of experiments where recombinant heterodimeric isoforms of PI3K were assayed for their ability to increase Ca²⁺ channel current density. Intracellular infusion of either a high concentration of PI3K or a low concentration of Gβγ- or PTyr-stimulated PI3K led to a maximal increase in peak Ca²⁺ channel current density. These results show that all four isoforms are potentially able to stimulate Ca²⁺ channels. However, because the four class I PI3K isoforms possess both lipid kinase and protein kinase activities,³¹ the present study does not help to discriminate whether PI3Ks stimulate Ca²⁺ channel through a protein phosphorylation mechanism or through the second messenger PtdIns(3,4,5)P₃.

Receptor-induced regulation of PI3Ks has been classically assigned to the noncatalytic subunit of the enzymes. In the present study, we confirm that the Ca²⁺ channel stimulation by p85-associated p110 subunits is increased by PTyr whereas the p101-associated p110 subunit–mediated effect is increased by Gβγ. However, because low concentrations of p85α/p110β are fully able to stimulate L-type Ca²⁺ channels once activated by Gβγ, this suggests that the p101 subunit usually associated with p110γ is not required to confer the Gβγ sensitivity of the p110 catalytic subunits. Our results are in agreement with studies showing that the p101 subunit is not required for p110γ activity in vitro²⁸ and that p110β or p110γ activation by Gβγ occurs regardless of the noncatalytic subunit it is associated with.¹⁸ In our cellular system, Gβγ-activated exogenous PI3Kβ and PI3Kγ stimulate L-type Ca²⁺ channel activity to the same extent, suggesting that PI3Kβ may represent the GPCR-activated PI3K counterpart in cells that do not express PI3Kγ.

The lack of additivity of Gβγ- and PTyr-activated effects in increasing PI3Kβ-induced stimulation of Ca²⁺ channels contrasts with previous in vitro studies, showing a synergistic activation of lipid kinase activity when both stimuli were coapplied.^17,18 This difference might be attributable to an intrinsic limitation of Ca²⁺ channel facilitation or to a limitation of the endogenous phosphatidylinositol-4,5-bisphophate that serves as a substrate for the production of PtdIns(3,4,5)P₃ by PI3K.

The second main conclusion of the present study points out that the stimulatory mechanism of Ca²⁺ channels by PI3K can be differently regulated by hormones and mediators in function of the culturing stage of the cells. This statement is supported by results showing that in freshly isolated myocytes, PTyr and PDGF-BB were unable to stimulate Ca²⁺ channels except when low concentrations of exogenous PI3Ks were infused into the cytoplasm. In contrast, Gβγ and angiotensin II stimulate Ca²⁺ channels in the same cell batches without needing exogenous PI3Ks.⁶ Although we showed that in addition to PI3Kγ, PI3Kβ can fully activate L-type Ca²⁺ channels when activated by Gβγ only, the PI3K isoform involved in the Gβγ-activated and angiotensin II–activated effects in native myocytes is believed to be PI3Kγ. Arguments supporting this proposal are that the β isoform is not expressed in freshly isolated myocytes and crude portal vein media as revealed by immunostaining and Western blot analysis and that an anti-PI3Kγ antibody specifically inhibits the Gβγ-induced and angiotensin II–induced stimulation of Ca²⁺ channels.³²

In this study, we show that PTyr and PDGF-BB failed to stimulate L-type channels in freshly isolated vascular myocytes, although the PTyr-activatable class Ia PI3Kα is expressed as revealed by immunocytochemistry and by Western blot analysis. This could be attributable to a negative regulation of p110α either by Ruk-like adaptor proteins that have been shown to inhibit p85α-associated p110 lipid kinase activity³³ or by some isoforms of the regulatory subunit, ie, p85β, p55α, or p55γ, which are recognized by the same anti-p85 antibody as p85α but transduce very weak, if any, activation of PI3Ks.³⁴ An alternative possibility would be that enzymatic dissociation alters PDGF receptor coupling efficiency, but this is unlikely, because PTyr is also not able to stimulate Ca²⁺ channels through PI3Kα in freshly isolated myocytes. Finally, one may speculate that the efficiency and subcellular compartmentalization of endogenous PI3Kα do not allow stimulation of Ca²⁺ channels.

Because we noticed that p110β subunit expression differs in tissue extract and freshly isolated vascular myocytes versus cultured myocytes, we hypothesized that the appearance of endogenous PI3Kβ might be responsible for the PTyr-mediated and PDGF-BB–mediated effects on L-type Ca²⁺ channels observed in cultured cells. Damage of PI3Kβ during the cell isolation process is unlikely, because Western Blot analysis performed on tissue extract led to the same result as immunostaining in freshly isolated myocytes, ie, the absence of p110β subunit. Our hypothesis has been confirmed by using the anti-PI3K antibodies in functional experiments, showing that anti-PI3Kβ, but not anti-PI3Kα, inhibits PDGF-BB–induced stimulation of Ca²⁺ channels. During intracellular infusion of exogenous PI3K, the loss of specificity of PI3Kβ versus PI3Kα in transducing RTK- and PTyr-mediated stimulation of L-type Ca²⁺ channels argues in favor of specific subcellular localization of endogenous PI3Ks. Referring to recent studies where PI3Kα and β isoforms have been reported to have different roles in RTK-mediated signaling,^24,35 it is likely that the PI3Kα expressed in vascular cells may be involved in other cellular responses than Ca²⁺ channel stimulation.

In conclusion, we show that Ca²⁺ channel activity can potentially be enhanced by all four types of class I PI3K but that, physiologically, this cellular response involves specific isoforms depending on the agonist acting on the cells and the culture stage of the cells. Our results also point out a specificity of endogenous PI3Kβ versus endogenous PI3Kα in transducing stimulation of Ca²⁺ channels in cultured cells.

Footnotes