ABSTRACT
Activation of the two ubiquitous families of protein kinases, protein kinase A (PKA) and protein kinase C (PKC), is thought to be independently coupled to stimulation of Gαs and Gαq, respectively. Live-cell confocal imaging of protein kinase C fluorescent protein fusion constructs revealed that simultaneous activation of Gαs and Gαq resulted in a differential translocation of the conventional PKCα to the plasma membrane while the novel PKCδ was recruited to the membrane of the endoplasmic reticulum (ER). We demonstrate that the PKCδ translocation was driven by a novel Gαs-cyclic AMP-EPAC-RAP-PLCε pathway resulting in specific diacylglycerol production at the membrane of the ER. Membrane-specific phosphorylation sensors revealed that directed translocation resulted in phosphorylation activity confined to the target membrane. Specific stimulation of PKCδ caused phosphorylation of the inositol-1,4,5-trisphosphate receptor and dampening of global Ca2+ signaling revealed by graded flash photolysis of caged inositol-1,4,5-trisphosphate. Our data demonstrate a novel signaling pathway enabling differential decoding of incoming stimuli into PKC isoform-specific membrane targeting, significantly enhancing the versatility of cyclic AMP signaling, thus demonstrating the possible interconnection between the PKA and PKC pathways traditionally treated independently. We thus provide novel and elementary understanding and insights into intracellular signaling events.
INTRODUCTION
G protein-coupled receptors (GPCR) translate incoming extracellular stimuli into a variety of intracellular signaling cascades by regulating the subunits Gα and Gβγ of heterotrimeric G proteins (1). One of the main classes of G proteins contains Gαq subunits and activates phospholipase C β (PLCβ). This results in hydrolysis of the membrane phospholipid phosphoinositol-(4,5)-bisphosphate (PIP2) into diacylglycerol (DAG) and inositol-1,4,5-trisphosphate (InsP3), which in turn regulates intracellular Ca2+ (2). Another ubiquitous Gα protein subunit is Gαs, which activates adenylate cyclases (ACs) and increases intracellular levels of the diffusible second messenger cyclic AMP (cAMP).
Following rises in the concentration of DAG, Ca2+, and cAMP, two important serine/threonine kinase families, the protein kinases C (PKCs) and protein kinases A (PKA) are stimulated. Activity of PKCs is evoked by signaling lipids and/or Ca2+ (3–5), while PKA is activated by cAMP (6). Nevertheless, a putative cross talk between these two pathways, especially a response of PKCs to changes in the concentration of cAMP, could not be established within 3 decades of intensive studies (7).
The family of PKC kinases comprises 10 isoforms: Ca2+-sensitive or conventional and phorbol ester-responsive PKCs (cPKCs) α, βI, βII, and γ; Ca2+-insensitive but DAG- and phorbol ester-responsive novel PKCs (nPKCs) ε, δ, η, and θ; and atypical PKC isoforms (aPKCs) ζ and ι, which are regulated by different lipid derivatives and are not responsive to Ca2+ or phorbol esters (4, 5, 8, 9).
Since the kinase domains of PKCs share similar substrate specificities (10, 11), the specificity of target protein phosphorylation by members of the abundant PKC family results from specific targeting rather than from the substrate specificity of the active site (8, 12–14).
Upon physiological stimuli, the mobilization of intracellular Ca2+ triggers cPKCs' translocation to the plasma membrane. The further association of cPKCs to DAG on the plasma membrane endows the kinase with competent activity (15, 16). However, the mechanisms of how exactly nPKCs decode incoming physiological information and downstream to which G-proteins nPKCs can respond are still rather unclear. An earlier report on cultured neurones using immunofluorescence analysis provided evidence for a possible link between cAMP production and PKCε activation (17).
Utilizing confocal microscopy of living cells, we addressed the following questions. (i) Which particular membranes are targeted by nPKCs (PKCδ) following physiological activation of endogenous G-protein-coupled signaling pathways? (ii) What are the underlying molecular signaling cascades leading to specific membrane recruitment of nPKCs? (iii) Is specific translocation translated into specific phosphorylation at the target membrane when comparing cPKCs and nPKCs? (iv) What are the physiological consequences of specific membrane targeting of nPKCs?
MATERIALS AND METHODS
Cell culture and transfection.HEK293 cells were cultured as described previously (18). NanoJuice (Novagen, USA) was used according to the producer's recommendations for the transfection of all plasmids (see below). Cells were assessed 48 h after transfection. When using small interfering RNA (siRNA) approaches, the cells were transfected using Lipofectamine 2000 (Invitrogen, USA) with the protocol recommended by the manufacturer. Cells were investigated 72 h after transfection.
Solutions and chemical compounds.Experiments were performed at room temperature (20 to 22°C) in Tyrode's solution, comprising 135 mM NaCl, 5.4 mM KCl, 2 mM MgCl2, 1.8 mM CaCl2, 10 mM glucose, and 10 mM HEPES adjusted to pH 7.35 with NaOH. For UV flash photolysis of caged InsP3, a modified Tyrode's solution was applied, containing 137 mM LiCl, 5.4 mM KCl, 2 mM MgCl2, 10 mM EGTA, 10 mM glucose, and 10 mM HEPES, adjusted to pH 7.35 with LiOH. Here, extracellular Na+ and Ca2+ were omitted to inhibit plasma membrane Na+/Ca2+ exchange activity.
All chemical compounds were of research grade: ATP, bryostatin 1, forskolin, 8-pCPT-2′-O-Me-cAMP, 3-isobutyl-1-methylxanthine (IBMX), U37122, phorbol myristate acetate (PMA) (all from Sigma-Aldrich, Germany), and m-3M3FBS and Gö6983 (both from Tocris, United Kingdom). From appropriate stock solutions, the desired experimental solutions were generated directly prior to the experiments. In control experiments, the inability of the vehicles to cause cellular activation was tested (data not shown).
The human siRNA structures Epac1 siRNA (catalog no. L-007676-00-0005), PLCE1 siRNA (catalog no. L-004201-00-0005), scrambled siRNA (catalog no. D-001810-01-05), and transfection indicator siGLO Red (catalog no. D-001630-02-05) were purchased from Thermo Scientific (Germany).
Fluorescently labeled proteins.Human PKCδ fused at the C terminus with enhanced green fluorescent protein (eGFP) in pcDNA3 was a kind gift from Michael Schaefer (19). Human PKCα was C-terminally fused to enhanced yellow fluorescent protein (eYFP) or DsRed2 in the pcDNA3 plasmid as previously described (18). Human PKCα fused to C-terminal TagRFP-T (20) was constructed in a pCR259 vector.
The endoplasmic reticulum (ER)-targeting sequence of calreticulin at the 5′ end of DsRed2 and the KDEL ER retention sequence at the 3′ end of DsRed2 were cloned from the commercial pDsRed2-ER (Clontech, USA) vector and inserted into the pCR259 vector (21). As a Golgi body target, we used the N-terminal 81 amino acids of human beta 1,4-galactosyltransferase C-terminally fused to monomoric red fluorescent protein (mRFP) (in pCR259). As a mitochondrial target, we used subunit VIII of human cytochrome c oxidase C-terminally fused to mRFP (in pCR259) (21).
The CEPAC, cAMP-detecting fluorescence resonance energy transfer (FRET) biosensor mCerulean-Epac(δDEP-CD)-mCitrine, was a kind gift of Andre Zeug (22, 23).
The CKAR and pmCKAR plasmids were a kind gift of Alexandra Newton (24). The erCKAR construct was generated by fusing the transmembrane domain of cytochrome b (5) to the 3′ end of CKAR, resulting in a construct in which CKAR is oriented toward the cytosol. The ER-targeting sequence was a kind gift of Nica Borgese (25).
The pCMV2-FLAG-RapGAP1 plasmids, encoding Rap1GAP protein, were a kind gift from Lawrence Quilliam (26). The constructs of Rap1GAP and PKCδ-eGFP were subcloned into pBI-CMV1 (Clontech, USA), the bidirectional promoter vector, at multiple cloning sites 1 and 2, respectively, resulting in pBI-Rap1GAP-PKCδ-eGFP, in which Rap1GAP expression was indicated by PKCδ-eGFP after transfection.
PKCδ mutagenesis.Human PKCδ dominant negative mutation PKCδK378R T507E (27, 28) was generated in the DNA plasmid PKCδ-TagRFP-T by site-directed mutagenesis at the 378th and 507th amino acids using the primers 5′-TGGAGAGTACTTTGCCATCAGGGCCCTCAAGAAGGATGTG-3′ and 5′-AGAGCCGGGCCAGCGAGTTCTGCGGCACCC-3′. All constructs were confirmed by sequencing.
Image acquisition and analysis.Solutions were changed using a custom-built solution changer using gravity-driven flow. For most of the experiments, we used a Leica TCS SP5 II (Leica, Germany) with a resonating x-scanner. The objective was a 63×, 1.4- numerical-aperture (NA) Plan-Apo oil immersion lens (Leica, Germany). Excitation was achieved with appropriate wavelengths of an argon ion laser (Lasos, Germany) or a solid-state laser (Melles Griot, France) of 488 nm for eGFP or 561 nm for mRFP, DsRed2, and TagRFP-T. The respective emissions were captured at 495 to 550 nm for eGFP/eYFP and 600 to 750 nm for mRFP, DsRed2, and TagRFP-T. Images (1,024 by 1,024 pixels) were recorded at 0.2 frame per second (fps).
The structure of the ER is among the most dynamic structures in a living cell (for examples, see references 29, 30, and 31) and displays dynamic rebuilding in the time domain of seconds. Therefore, choosing a typical and representative static region of interest (ROI) for the fluorescence over time plots was indeed rather difficult. ROIs were chosen in the particular subcellular region that displayed a high and rather constant ER “content” over the entire recording time. Nevertheless, in all cases the ROIs included both ER and cytosol and consequently represented an average of the two compartments. As can be seen for most of the panels, the fluorescence trace averaged from the ER region was often much more noisy than the cytosolic traces owing to the fact that in the former ER tubules were constantly entering and leaving the ROI. For each cell analyzed, at least 4 or 5 of such ROIs were studied to evaluate the typical behavior of translocation, and the one with the smallest contribution of ER fluctuations over time was chosen for display. Only after collecting sufficient amounts of cells (as described in the text and the figure legends) was the final conclusion drawn. Additionally, in order to ensure maximum objectivity and unbiased analyses, around 50% of the experiments were analyzed by a second researcher in a blinded manner following the guidelines outlined above. This second round of blinded analysis always came to the same conclusions supporting our approach.
Experiments using UV flash photolysis of caged Ca2+ were performed as described earlier (18). Briefly, cells were loaded with 7.5 μM NP-EGTA-AM (Invitrogen, Germany). Ca2+ was photoreleased by a bright UV flash (TILL Photonics, Germany). Uncaging was performed with simultaneous recording of confocal image series on an inverted microscope (TE-2000U; Nikon, Germany) using an oil immersion objective (40×, 1.3-NA S-Fluor; Nikon, Germany). The microscope was attached to a two-dimensional (2D)-array scanner VT infinity (VisiTech International, United Kingdom) equipped with an electron-multiplying charge-coupled-device (EMCCD) camera, iXon 887 (Andor Technology, United Kingdom). The whole setup was integrated and controlled by VoxCellScan software (VisiTech International, United Kingdom). Images (256 by 256 pixels) were recorded at 10 fps.
UV flash photolysis of caged InsP3 was performed as described above for caged Ca2+. Cells were loaded with 1 μM Fluo4 AM (Invitrogen, Germany) and 0.5 μM membrane-permeable caged InsP3, ci-InsP3/PM (SiChem, Germany). Caged InsP3 was photouncaged by a UV flash (TILL Photonics, Germany). Images (120 by 128 pixels) were recorded at 2 fps.
The FRET assays were accomplished on an inverted microscope (TE-2000U; Nikon) equipped with an oil immersion objective (20×, 0.75-NA Pan-Fluor), a CCD camera (Imago QE; TILL Photonics, Germany) and an image splitter (OptoSplit II; Cairn Research, United Kingdom). A monochromator (Polychrome IV; TILL Photonics, Germany) was utilized to generate the desired excitation wavelength of 430 nm for cyan fluorescent protein (CFP) and mCerulean, 512 nm for YFP and mCitrine, and the emitted light was separated into CFP (470/24 nm) and YFP (535/30 nm) channels by a dichroic mirror. Control of the setup and the experiment was achieved with TILLVision software (TILL Photonics, Germany). Images (160 by 240 pixels) were recorded at 0.1 fps.
The acquired images were stored using OMERO software (OME, Dundee, United Kingdom) (32) and further analyzed using custom-built macros in ImageJ (W. Rasband, NIH, USA). The fluorescence over time in regions of interest was transferred into IGOR software (Wavemetrics, USA) and further processed using custom-written macros. When necessary, images were denoised using a wavelet-based algorithm (33) built into an ImageJ plug-in.
When necessary, self-ratio images were calculated as previously described (18). Color-coded 2D images were constructed in ImageJ from 16-bit or 8-bit gray scale images by applying the appropriate look-up tables. Image correlation analysis was based on Pearson's coefficient approach using the ImageJ plug-in, JACoP (F. P. Cordelieres, Institut Curie, France).
To avoid interference of the FRET measurements from cell movement, bleaching, and spectral excitation and emission cross talk, we calculated the apparent FRET efficiency instead on the CFP/YFP fluorescence ratios according to a method published previously (34) using the following formula:
The final figure design was produced in Adobe Illustrator software (Adobe Inc., USA).
Statistical analysis.Statistical analysis was performed with Prism 5 (GraphPad, USA) using unpaired t tests after testing for Gaussian distribution using the D'Agostino-Pearson omnibus normality test. Bar graphs are displayed as means ± standard errors of the means (SEM).
RESULTS
PKCα and PKCδ target different intracellular membranes.The distribution and the translocation of cPKCs and nPKCs in response to a wide range of stimuli have been investigated in numerous reports (35–41). Nevertheless, the way these two important subfamilies of PKCs indeed interpret a sole stimulus differentially has not been addressed so far. To thoroughly investigate such translocation regimes in living cells in response to cellular stimulation, we coexpressed members of the cPKC and nPKC subfamily fused to monomeric fluorescent proteins of different colors. In our hands, application of low PMA concentrations (500 nM) resulted in stereotyped and irreversible plasma membrane and nuclear envelope (for PKCδ) recruitment (Fig. 1A). In contrast, stimulation of the cells with ATP resulted in transient targeting of PKCα to the plasma membrane, while PKCδ was concomitantly recruited to intracellular membranes, most probably the ER, albeit on a slower time scale (Fig. 1B). The table presented in Fig. 1C summarizes and extends such findings for a range of stimulation regimes: PMA, ATP, and flash photolysis of caged Ca2+ (exemplified data can be found in Fig. S1 in the supplemental material). To identify the nature of the intracellular membranes that PKCδ targeted, we coexpressed PKCδ and ER-targeted DsRed2. Under resting conditions, the analysis depicted limited colocalization (Fig. 1D, upper panels; note the shallow relationship indicated by the dashed line), but following ATP stimulation this correlation increased substantially as indicated by the increased steepness of the dashed yellow line (Fig. 1D, lower panels). Other organelles such as mitochondria and the Golgi apparatus did not display this particular colocalization pattern with PKCδ (see Fig. S2 in the supplemental material).
PKCα and PKCδ are differentially targeted upon ATP stimulation. (A and B) Translocation of PKCs in HEK293 cells expressing PKCα and PKCδ fused to fluorescence proteins. Representative images (left) and plots of the relative fluorescence over time (right) for the regions of interest (ROI) encircled in the left panels during an experiment using 300 nM PMA (A) or 100 μM ATP (B). Color coding of the traces is explained in the inset between panels A and B. Light gray areas denote PKCα, while the darker areas denote PKCδ. Similar data were obtained in 56 cells from 18 independent experiments. (C) Comparison of PKCα and PKCδ translocation targets for 3 different experimental stimulation regimes utilizing (from left to right) PMA (500 nM), ATP (100 μM), and UV flash photolysis of NP-EGTA. (D) After ATP stimulation, PKCδ translocated to the ER. HEK293 cells expressing PKCδ-eGFP and ER-targeted DsRed2 before (upper row) and following (lower panels) ATP stimulation. The right panels represent the colocalization analysis before and during ATP stimulation. Typical examples for 20 cells from 3 different experiments. Scale bars, 10 μm.
We assumed that for PKCδ targeting, DAG production was the key step in the underlying signaling mechanism and therefore investigated the molecular signaling cascade that led to this ATP-dependent ER membrane recruitment.
ER recruitment of PKCδ is mediated by a cAMP-EPAC-PLCε signal pathway.Because it was reported that P2Y receptors can stimulate not only Gαq but also Gαs proteins (42), we suggested a cAMP-dependent signaling cascade comprising cAMP-EPAC-Rap-PLCε, whose activation results in ER-targeted DAG production (Fig. 2A; see also Fig. S3A in the supplemental material).
Activation of P2Y receptor results in cAMP production and ER recruitment of PKCδ. (A) Cartoon depicts the proposed signaling cascades for cPKC and nPKC translocation in HEK293 cells. The upper gray region highlights processes at the plasma membrane and the lower region those at the ER membrane. (B to D) In HEK293 cells expressing PKCδ-eGFP, direct activation of adenylate cyclase (AC) with forskolin or ATP and simultaneous inhibition of phosphodiesterases with IBMX (PDE) (B and D) resulted in ER recruitment of PKCδ. Application of IBMX alone does not cause any redistribution. Left images depict PKCδ-eGFP distribution prestimulation (labeled “1”) and late during stimulation (labeled “2”). Fluorescence traces were calculated from regions of interest labeled with matching colors in the prestimulation images. The vertical dotted lines indicate the time points at which the leftmost images were taken. (E and F) HEK293 cells were transfected with the cAMP sensor CEPAC and cAMP-detecting FRET biosensor mCerulean-Epac(δDEP-CD)-mCitrine and subjected to various stimulation regimes indicated above the panels. The substances applied are depicted with corresponding colors. Note that a drop of the apparent FRET efficiency of this sensor highlights an increase of cAMP. Traces represent apparent FRET efficiency calculated from the global cellular fluorescence (see Materials and Methods for details). The panels to the right illustrate exemplified apparent FRET images color coded for the time point indicated in the traces at the three time points detailed. The figures are representative of at least 20 cells from 7 experiments. Scale bars, 10 μm.
To initially investigate the principal presence of the appropriate gene products, we performed reverse transcription (RT)-PCR analysis of HEK293 cells for P2Y11, PKCs, EPAC1, EPAC2, and the Rap isoforms 1a, 1b, 2a, and 2b (see Fig. S3C to E in the supplemental material). In subsequent experiments, we verified the involvement of our proposed signaling cascade (Fig. 2 and 3). For all of our experiments, we employed PKCδ translocation to the ER as the cellular readout. We employed the P2Y11 selective inhibitor NF157 (43, 44) to verify the involvement of this P2Y11 receptor and found that in a majority of cells tested, 10 μM NF157 indeed abrogated the ATP-induced PKCδ translocation (see Fig. S4 in the supplemental material).
The cAMP-EPAC-PLCε signal pathway mediates the ER recruitment of PKCδ. (A) Cartoon depicts the proposed signaling cascade for nPKC translocation in HEK293 cells following increases in intracellular cAMP. (B to E) For all panels, the blue traces denote the fluorescence at an ER region, and the red traces represent cytosolic fluorescence (regions of interest marked at the first time point). Numbers in the confocal images correspond to time points highlighted by vertical dashed lines in the corresponding traces. Either red lines (inhibition of the particular protein) or a green line (activation of the particular protein) links the panels to different members of the signaling pathway, leading to recruitment and activation of nPKCs. (B) EPAC activation by membrane-permeable cAMP analogue. (C) EPAC downregulation by RNAi against EPAC1. Note that the right panel in the upper row depicts the siGLO-Red fluorescence for detecting successfully transfected HEK293 cells. For experiments on scrambled RNA, see Fig. S6B in the supplemental material. (D) PLCε downregulation by specific RNAi. The rightmost panel depicts siGLO-Red fluorescence for detecting successfully transfected HEK293 cells. For experiments on scrambled RNA, see Fig. S6A in the supplemental material. (E) Inhibition of Rap activity by coexpressing PKCδ-eGFP and RapGAP in a bidirectional plasmid vector. Note that all green cells express both proteins. Data are representative of at least 22 cells from 6 experiments.
In particular, we initially addressed whether ATP-mediated stimulation of P2Y receptors indeed activated Gαs signal transduction and consequently resulted in a cAMP-mediated ER recruitment of PKCδ (Fig. 2). Direct activation of cAMP production (forskolin), while simultaneously inhibiting its metabolism (IBMX) as illustrated in Fig. 2B, resulted in a robust loss of cytosolic fluorescence (red trace) and accumulation of the PKCδ construct at the ER membrane (blue trace). Interestingly, inhibition of the metabolic degradation of cAMP by IBMX alone as illustrated in Fig. 2C did not cause PKCδ translocation, indicating that the cAMP production was not constitutively active to a degree that could cause PKCδ recruitment. Inhibition of cAMP metabolism together with ATP stimulation resulted in a comparable PKCδ recruitment as direct activation of the AC (compare Fig. 2B and D). In the next series of experiments, we thought of providing direct evidence for cAMP production by employing the FRET-based cAMP sensor CEPAC, cAMP-detecting FRET biosensor mCerulean-Epac(δDEP-CD)-mCitrine (22, 23) as shown in Fig. 2E and F. In analogy to the recruitment patterns observed in Fig. 2B to D, we found cAMP production (denoted as a drop in apparent FRET efficiency) for coapplication of forskolin and IBMX (Fig. 2F, red trace), ATP and IBMX (Fig. 2E, blue trace), and ATP alone (Fig. 2E, black trace). While the former two stimulation regimes resulted in rapid and sustained cAMP levels, application of ATP alone caused a readily detectable but transient cAMP production (Fig. 2E, black trace). Application of IBMX alone was unable to accumulate sufficient cAMP (Fig. 2E, red trace). From these data, we concluded that indeed, ATP stimulation of P2Y receptors caused production of cAMP and this increase in intracellular cAMP was a key step for the recruitment of PKCδ to the ER membrane. A more direct way of testing the latter possibility is to bypass cAMP production by using the membrane-permeable cAMP analogue with high specificity for EPAC activation (8-pCT-2″-O-Me-cAMP) (45–47). As illustrated in Fig. 2F (red trace), application of this cAMP analogue indeed resulted in its robust intracellular increase as measured with the FRET probe.
Moreover, application of this cAMP analogue alone caused recruitment of PKCδ to the ER membrane as depicted in Fig. 3B, indicating that the intracellular cAMP sensor EPAC might be involved in the ER recruitment process. Interestingly, this cAMP analogue did not cause detectable Ca2+ signaling in HEK293 cells expressing PKCα-eGFP (see Fig. S7 in the supplemental material). To verify the EPAC involvement, we used EPAC1-specific siRNA (48) that we mixed with the siRNA transfection indicator siGLO-Red (49) and subsequently analyzed only those HEK293 cells that showed robust siRNA transfection with a red color (Fig. 3C, upper right panel). As can be depicted from Fig. 3C in HEK293 cells with downregulated EPAC1, ATP failed to cause PKCδ recruitment to the ER, while direct panspecific activation of PLCs resulted in ER recruitment (Fig. 3C, panel 3 and rightmost traces). In a series of control experiments, we tested scrambled RNA interference (RNAi) and found that it did not abrogate ATP-induced PKCδ recruitment to the ER (see Fig. S6B in the supplemental material).
Application of the panspecific PLC-activator m-3M3FBS (50) was sufficient to induce PKCδ recruitment (see Fig. S5A in the supplemental material), while simultaneous inhibition of PLCs by U73122 suppressed translocation (see Fig. S5A in the supplemental material). The application of the biologically inactive stereoisomer o-3M3FBS failed to induce any translocation (data not shown). To substantiate the involvement of PLCε further, we downregulated PLCε expression by siRNA in HEK293 cells and found that the downregulation of PLCε expression suppressed ATP-dependent PKCδ recruitment to the ER (Fig. 3D) but direct activation of PKCδ by 12 nM bryostatin 1 (see below) still induced PKCδ to translocate, indicating that despite PLCε downregulation, PKCδ recruitment was still possible (Fig. 3D, panel 3 and traces below it). These results supported our notion that PLCε and ER membrane-delimited production of DAG were involved in the recruitment process (Fig. 3A, DAGER).
Rap proteins have been described to bridge the signaling gap between EPAC and PLCε (51–53). We therefore further substantiated the direct involvement of Rap proteins by overexpressing RapGAP, a negative and highly specific regulator of Rap activity. To easily identify RapGAP-expressing HEK293 cells, we cloned the DNA for both RapGAP and PKCδ-eGFP into a bidirectional vector so that all cells expressing PKCδ-eGFP (green fluorescence in Fig. 3E) also expressed RapGAP. We found that RapGAP effectively suppressed downstream signaling to PLCε by stimulating the intrinsic GTPase activity of Rap (26). As illustrated in Fig. 3E, the expression of RapGAP abrogated PKCδ translocation in cells stimulated with ATP (Fig. 3E, left traces) but did not inhibit the cellular responses to m-3M3FBS in the same cells (Fig. 3E, right traces).
The data presented so far strongly indicated that ATP-mediated P2Y receptor activation evoked two different signaling pathways (Fig. 2A). Pathway 1 involved the activation of Gαq, resulting in the hydrolysis of PIP2 and InsP3 and in production of DAG at the plasma membrane (this pathway was responsible for the recruitment of PKCα [Fig. 1A and 2, left part of the cartoon]). Pathway 2 activated Gαs and caused cAMP production, which in turn stimulated PLCε and DAG production at the ER membrane, generating the signaling lipid that recruited PKCδ to the ER (Fig. 2, right part of the cartoon).
Recruitment of PKCα and PKCδ results in phosphorylation activity at their specific subcellular target membranes.In the following, we aimed to investigate whether different translocation patterns were translated into elevated activation levels of the PKC molecule and translocation-equivalent phosphorylation activity.
Figure 4A illustrates that application of 12 nM bryostatin 1, a specific activator of nPKCs (54, 55), resulted in an isoform-specific translocation of PKCδ to the ER membrane while the coexpressed PKCα construct did not show any redistribution. We point out that at higher concentrations (>25 nM) bryostatin 1 was also able to induce translocation of PKCα as we evaluated in a preliminary study (data not shown), but at 12 nM we did not find any cell for which PKCα displayed plasma membrane translocation. In the following, we made use of the differential recruitment regime for PKCα and PKCδ described above to investigate phosphorylation activity at the plasma and ER membrane. For this, we employed the C kinase activity reporter (CKAR) backbone, a PKC-specific phosphorylation sensor (13, 24) in HEK293 cells without PKC overexpression. In combination with specific anchors for the plasma membrane (pmCKAR, Fig. 4Ba, left panel) and ER membrane (erCKAR, Fig. 4Ba right panel), which ensured cytosolic facing of the phosphorylation sensor (25, 56), we were able to specifically express the phosphorylation sensor at these two membranes (Fig. 4Ba).
PKC translocation to the target membrane induces phosphorylation of target membrane-specific kinase reporters. (A) Bryostatin 1 specifically recruits PKCδ to the ER. Left images show the fluorescence distribution of PKCδ and PKCα at the time points given. The traces depict plots of the intracellular fluorescence over time for PKCα and PKCδ. Regions of interest are indicated in the images taken at t of 0 min. (B) (a) Subcellular distribution of PKC phosphorylation sensors pmCKAR and erCKAR. (b and c) Stimulation of HEK293 cells expressing pmCKAR or erCKAR with PMA (b) and bryostatin 1 (c). Increases in phosphorylation are indicated by decreases in the apparent FRET efficiency. PKC inhibitor Gö6983 abrogated phosphorylation (29 cells from 8 independent experiments). (C) Stimulation of HEK293 cells expressing erCKAR at various points of the suggested signaling cascade (compare with Fig. 2). Changes in the probe's apparent FRET efficiency following application of a membrane-permeable cAMP analogue, ATP, and bryostatin 1 are shown. PKC inhibitor Gö6983 abrogated phosphorylation (at least 18 cells from 5 independent experiments). (D) HEK293 cells overexpressing PKCδ display an increased phosphorylation level at the ER following bryostatin 1 stimulation (26 to 28 cells from 5 independent experiments). ***, P < 0.001. Scale bars, 10 μm.
The application of PMA resulted in an increased phosphorylation activity confined to the plasma membrane (Fig. 4Bb, green trace; note that a decrease of the FRET efficiency indicates increased phosphorylation), while changes in phosphorylation at the ER were absent (Fig. 4Bb, red trace). Treatment with the panspecific PKC inhibitor Gö6983 (57) diminished the CKAR signal, supporting the involvement of PKCs. In contrast to that, application of 12 nM bryostatin 1 caused specific PKCδ recruitment to the ER membrane (Fig. 4A) and consequently failed to produce phosphorylation activity at the plasma membrane (Fig. 4Bc, green trace) but caused augmented kinase activity at the ER (Fig. 4Bc, red trace). Specific phosphorylation at the ER was confirmed following ATP stimulation (Fig. 4C, blue trace) and the EPAC-specific cAMP analogue (Fig. 4C, green trace), strongly suggesting that PKCδ was indeed contributing to the ATP-mediated erCKAR signals. This finding was further substantiated by significantly increased erCKAR signals following overexpression of PKCδ (Fig. 4D).
In the following, we investigated a possible phosphorylation target of PKCδ at the ER membrane to substantiate the putative physiological relevance of ER-targeted PKCδ activity.
The recruitment of PKCδ induced desensitization of the InsP3R and dampening of Ca2+ signaling.The inositol-1,4,5-trisphosphate receptor (InsP3R) reportedly contains numerous phosphorylation sites, including consensus sequences for PKC-dependent phosphorylation that have been functionally confirmed by in vitro PKC phosphorylation assays (58, 59). To investigate whether ER targeting of PKCδ indeed resulted in an altered InsP3R behavior, we loaded HEK293 cells with Fluo4 and probed ATP-dependent Ca2+ responses under control conditions and following specific preactivation of PKCδ (Fig. 5A). We characterized the global Ca2+ signals by fitting the relaxation phase of the Ca2+ transient with a monoexponential decay function (dashed lines in Fig. 5Aa) and investigated the characterizing decay time constants statistically (Fig. 5Ab). The green trace in Fig. 5Aa depicts a typical global Ca2+ transient in response to the application of 50 μM ATP. Prestimulation of HEK293 cells with bryostatin 1 resulted in a substantially shortened Ca2+ transient (red trace in Fig. 5Aa and red column in Fig. 5Ab). A possible explanation for this finding is a decreased InsP3 sensitivity of the InsP3 receptor. To test this hypothesis, we set up an assay that allowed direct probing of the InsP3R's sensitivity by using graded flash photolysis of caged InsP3 (60) with and without prestimulation of PKCδ.
Bryostatin 1 stimulation results in a PKCδ-dependent desensitization of the InsP3R toward InsP3. (A) Fluo4-loaded HEK293 cells were challenged with 50 μM ATP in the absence of bryostatin 1 and following bryostatin 1 incubation (green and red traces and bars, respectively). The decaying Ca2+ oscillations were fitted with monoexponential equations (dashed lines) (a) and statistically analyzed (bar graphs) (b) (107 transients from 4 independent experiments). (B) Experiments in naive HEK293 cells loaded with membrane-permeable caged InsP3 and Fluo4. (a) Arrows and dashed lines indicate time points of UV flash delivery with one-quarter (1st) and full (2nd) energy. Control conditions are shown as green traces and bars (analysis of the amplitudes depicted in the right bars). (b) Experiments performed 15 min after incubation with bryostatin 1 or bryostatin 1 and Gö6983 in red and blue, respectively (78 to 156 cells from 7 independent experiments). (C) Experiments in PKCδ-overexpressing HEK293 cells using an experimental approach similar to that described for panel Ba. Upper traces, cells expressing PKCδ; lower traces, cells expressing dominant negative (DN) PKCδ (PKCδ-DN-TagRFP-T). (b) ***, P < 0.001; 69 to 106 cells from 6 independent experiments.
In control experiments, we established that 25% UV flash energy was just above the threshold required to induce global Ca2+ transients, while 100% was far above this threshold and produced a consistently high response. We thus used both of these UV flash energy levels to assess the InsP3 sensitivity of the InsP3R.
While two consecutive flashes with 25% and 100% energy produced two robust Ca2+ responses under control conditions with comparable amplitudes (green traces in Fig. 4Ba and green columns in Fig. 4Bb), prestimulation of the cells with 12 nM bryostatin 1 reduced the first Ca2+ response by more than 70% and thus significantly reduced the amplitude ratio (A1/A2) (red traces in Fig. 5Ba and red bars in Fig. 5Bb). This reduction was inhibited by the simultaneous preincubation of HEK293 cells with bryostatin 1 and the PKC blocker Gö6983, which almost completely restored the control situation (blue traces in Fig. 5Ba and blue bars in Fig. 5Bb).
To verify that the bryostatin 1-induced reduction in the sensitivity of InsP3R to InsP3 was mediated by PKCδ, we used a dominant negative approach, expressing a PKCδ mutant with a “dead” kinase domain (PKCδ-DN [28, 61]). While the overexpression of wt-PKCδ did not change the ability of bryostatin 1 to modulate the InsP3 sensitivity of InsP3R (upper traces in Fig. 5Ca; red and green bars in Fig. 5Cb), the overexpression of the dominant negative mutant of PKCδ abrogated the effect of bryostatin 1 (lower traces in Fig. 5Ca; hatched columns in Fig. 5Cb).
Taken together, these results strongly indicate that the bryostatin 1-induced recruitment of PKCδ reduced the InsP3 sensitivity of InsP3R and shortened the Ca2+ oscillations following ATP stimulation. This behavior was most likely a result of PKCδ-mediated phosphorylation of InsP3R, highlighting the possible physiological relevance of the ER targeting of PKCδ.
DISCUSSION
Our study explicitly revealed a novel signaling link between stimulation of Gαs, cAMP production, and activation of PLCε, starting with the activation of Gαs, production of the soluble second messenger cAMP, cAMP-mediated activation of EPAC, and Rap-mediated PLCε-dependent DAG production at the ER membrane, which eventually recruits and activates PKCδ, a member of the subfamily of nPKCs. We identified that endogenous DAG production on the ER membrane via the EPAC-Rap-PLCε signal pathway was the key step to differentiate the activation patterns of two PKCs, PKCα and PKCδ, following physiological stimulation of purinergic receptors, which coupled to two G-protein-dependent signaling cascades, via Gαs and via Gαq. Furthermore, we revealed that the highly specific recruitment of PKCδ to the ER membrane allows phosphorylation-mediated adjustments of the InsP3's sensitivity toward its natural ligand, InsP3, with significant consequences for cellular Ca2+ signaling.
From our data, we conclude that application of phorbol esters to activate PKCs obliterates the physiological response patterns of different PKC isoforms (18, 62–64). In our hands, the stimulation of cells with PMA evoked stereotypic PKC translocation patterns to the membrane, where the phorbol ester enriched first. Since PKCδ was also expressed in the nucleus, later, accumulation of PMA in the nuclear envelope also recruited nucleoplasmic PKCδ to the nuclear membrane system. Compared to DAG, the slow metabolism of artificial phorbol esters (65) and the extremely high affinity of PMA to C1 domains (39, 66–68) abolish target discrimination of PKCα and PKCδ (Fig. 1A). Moreover, while the use of phorbol esters results in PKC-target membrane interaction that lasts for tens of minutes, physiological responses of PKCs are often only very brief (e.g., less than a second for cPKCs [18]). Strong binding of PKCs to membrane-bound phorbol esters results in high association and low dissociation rates and thus favors artificially long PKC-target protein interactions. Such interactions and the associated phosphorylation will occur at phosphorylation sites that during brief, physiological interactions might not be accessible for the kinase domains of PKCs, thus displaying errant phosphorylation patterns and misactivated signaling cassettes or even causing misconceptions of cellular signaling.
We thus conclude that cells can direct PKCs to specific target membranes, e.g., by membrane-delimited production of DAG, here DAGER. This signaling lipid is produced in response to ATP-mediated activation of P2Y receptors and their ability to signal downstream by interacting with both Gαs and Gαq. While activation of Gαq results in DAG production at the plasma membrane (mediated by PLCβ) and InsP3-mediated Ca2+ release from the ER (69, 70), we identified a complex signaling cascade following stimulation of Gαs-mediated AC activity (Fig. 2). The diffusible second messenger cAMP binds to EPAC (52, 71) and subsequently activates Rap proteins at the ER membrane (72, 73), which in turn stimulate PLCε activity, hydrolyzing PIP2 and generating DAGER (53, 74). This membrane-delimited second messenger eventually recruits PKCδ to the ER membrane.
Even though protein kinase A (PKA) is the most prominent downstream effector of cAMP, our results support the existence of an additional novel mechanism for simultaneous activation of nPKCs via an EPAC-Rap-PLCε signal pathway. In this cascade, cAMP plays an essential role because its increase triggers the production of ER-specific DAG and eventually recruits PKCδ to the ER membrane. When employing the EPAC-based cAMP sensor CEPAC, we found that stimulation of HEK293 cells with ATP indeed resulted in a cAMP production that was sufficient for PKCδ recruitment, although its concentration appeared much lower than cAMP production evoked by direct AC stimulation with forskolin (Fig. 2E and F). This might be attributable to various reasons: coupling of ATP-evoked Gαs stimulation might not be very efficient or only targeting subpopulations of ACs, and ATP can reportedly also activate Gαi-coupled P2Y receptors, e.g., P2Y2, P2Y12, and P2Y13, subsequently decreasing the cAMP production (75). Moreover, we interestingly found that in the absence of the phosphodiesterase (PDE) inhibitor, ATP evoked only a transient cAMP increase. Although such a transient response could result from receptor desensitization (76, 77) or coactivation of Gαi, the finding that the process of PKCδ recruitment to the ER membrane persisted for a much longer time appears to be incompatible with such explanations. We suggest here that such a sustained recruitment response actually argues strongly for the involvement of PLCε that contains a CDC25 domain and thus might by itself amplify and prolong the signal from Rap1 (78, 79). In addition, the PKCδ accumulation on the ER membrane also demonstrated a specific DAGER production at a distinct intracellular membrane because the C1 domain of PKCδ can be seen as a DAG sensor. Due to its high sensitivity to DAG (80), it was recently even employed as a sensor for DAG production (81).
The targeted DAG production also allows differentiation between cPKC and nPKC activation. (i) Using organelle-targeted phosphorylation sensors to report targeted phosphorylation activity, we demonstrated that PMA resulted in sole plasma membrane phosphorylation activity, independent of the isoform. In contrast to that, activation of physiological signaling cascades led to the expected and versatile ER membrane-restricted phosphorylation activity (Fig. 4). Thus, the observed recruitment resulted in “opening” of the PKC molecule, release of the pseudosubstrate domain from the kinase activity center, and kinase activity (8, 16). Noteworthy, the specific targeting of PKCα and PKCδ induced restricted phosphorylation events at their respective target membranes, which not only highlights spatial specificity but also supports the importance of spatiotemporally restricted PKC activity as a major determinant for substrate specificity. (ii) ER-specific recruitment of PKCδ resulted in dampened InsP3-dependent global Ca2+ signaling, ATP-induced Ca2+ transients displayed significantly shorter lifetimes (Fig. 5). PKCδ accumulation at the ER membrane will help to limit the positive feedback in InsP3-dependent Ca2+ release from the ER by reducing the InsP3 sensitivity of the InsP3R. The principle of the possibility of InsP3R phosphorylation by PKC and the resulting decrease in the InsP3R′s InsP3 sensitivity has been reported numerous times (see, for example, references 58, 59, and 82). It appears important to note here that such a negative-feedback loop works only when, in addition to Gαq, Gαs proteins are also activated. Moreover, modulation of global Ca2+ signals such as their duration or longevity of underlying Ca2+ oscillations has been shown to be effective in changing the activity levels of Ca2+-regulated genes (83, 84).
In the current report, we demonstrate the importance of a novel signal cascade from cAMP to nPKCs that eventually transduces plasma membrane stimuli to phosphorylation of target proteins at the ER membrane such as the ubiquitously expressed InsP3 receptor-modulating Ca2+ signaling. Such a signaling pathway not only significantly enhances the versatility of cAMP signaling, it also allows simultaneous conversion of different signaling pathways onto distinct members of the PKC family and provides novel and elementary understanding and insights into physiological and pathological cellular responses.
ACKNOWLEDGMENTS
X.H., L.K., and P.L. acknowledge financial support by the DFG (LI 753/6-1 & SFB1027, TPA5 to P.L.) and funding by the HOMFOR program of the Medical faculty and science start-up programs of the Saarland University.
FOOTNOTES
- Received 9 January 2014.
- Returned for modification 6 February 2014.
- Accepted 8 April 2014.
- Accepted manuscript posted online 14 April 2014.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/MCB.00040-14.
- Copyright © 2014, American Society for Microbiology. All Rights Reserved.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
