“Shaping” of cell signaling via AKAP-tethered PDE4D: Probing with AKAR2-AKAP5 biosensor
PKA, a key regulator of cell signaling, phosphorylates a diverse and important array of target molecules and is spatially docked to members of the A-kinase Anchoring Protein (AKAP) family. AKAR2 is a biosensor which yields a FRET signal in vivo, when phosphorylated by PKA. AKAP5, a prominent member of the AKAP family, docks several signaling molecules including PKA, PDE4D, as well as GPCRs, and is obligate for the propagation of the activation of the mitogen-activated protein kinase cascade from GPCRs to ERK1,2.
Using an AKAR2-AKAP5 fusion “biosensor”, we investigated the spatial-temporal activation of AKAP5 undergoing phosphorylation by PKA in response to β-adrenergic stimulation. The pattern of PKA activation reported by AKAR2-AKAP5 is a more rapid and spatially distinct from those “sensed” by AKAR2-AKAP12. Spatial-temporal restriction of activated PKA by AKAP5 was found to “shape” the signaling response. Phosphatase PDE4D tethered to AKAP5 also later reverses within 60 s elevated intracellular cyclic AMP levels stimulated by β-adrenergic agonist. AKAP12, however, fails to attenuate the rise in cyclic AMP over this time. Fusion of the AKAP5 PDE4D-binding-domain to AKAP12 was found to accelerate a reversal of accumulation of intracellular cyclic AMP.
AKAPs, which are scaffolds with tethered enzymes, can “shape” the temporal and spatial aspects of cell signaling.
KeywordsAKAP5AKAP12AKAR2β-adrenergic receptorPDE4DProtein kinase AScaffoldTethered
A-kinase anchoring protein
Protein kinase A
Cyan fluorescent protein
Dominant interfering peptide
Green fluorescent protein
g protein-coupled receptor
Human embryonic kidney cells
Mitogen-activated protein kinase
Cyclic AMP phosphodiesterase
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis
Yellow fluorescent protein (in these studies the variant citrine)
T/P395: AKAP5 146-427, dominant interfering peptide which does not bind the cell membrane, can dock PDE but not PKA.
Scaffold proteins that dock the regulatory subunits (i.e., RI//RII) of cyclic AMP (cyclic AMP)-dependent protein kinase A (PKA, A-kinase) constitute a family of 40+ members. These so-called “A-Kinase Anchoring Proteins” (AKAPs) have been revealed as essential to cellular signaling . AKAPs dock PKA and act as molecular “tool boxes” capable of targeting the protein kinases [e.g., PKA as well as protein kinase C (PKC) and the Src-family of tyrosine kinases], phosphoprotein phosphatases (such as protein phosphatase-2B), and adaptor molecules to microdomains in cells [2–4]. AKAPs likely nucleate supermolecular signaling complexes that facilitate productive interactions among protein kinases, phosphoprotein phosphatases, phosphodiesterases (PDE), ion channels, adaptor molecules (like Grb2), and G protein-coupled receptors (GPCR) . Two AKAPs, AKAP5 (also known as AKAP75/79) and AKAP12 (also known as gravin, AKAP 250, and SSECKS) dock to β-adrenergic receptors [2, 5–9]. Human epidermoid carcinoma A431 and human embryonic kidney cells (HEK) 293 cells express several AKAPs, including AKAP5 and AKAP12, and a full complement of the prototypic GPCR, notably the β2-adrenergic receptor [10–13].
Detailing the spatial-temporal dynamics of molecules that actively participate in a signaling pathway during signaling has been a challenge . Immunohistochemical and imaging of auto fluorescently-tagged molecules cannot distinguish between molecules actively participating in signaling versus the entire cellular complement of the molecule. Study of the subset of these molecules actively involved in the signaling in live cells is especially challenging, as they constitute a vanishing small percentage of the full cellular complement. “Biosensors” capable of acting as reporters in real time for activation of a protein kinase (e.g., AKAR/AKAR2 reporters of PKA activation) offer a powerful, novel strategy targeting spatial-temporal features of signaling catalyzed by PKA . AKAR2 is a biosensor for PKA-catalyzed phosphorylation, composed of a phosphoamino acid-binding sequence of the Forkhead Homology domain and a citrine, a less pH-sensitive variant of enhanced yellow fluorescent protein (YFP) [14, 15]. PKA-catalyzed phosphorylation of AKAR2 consensus sequence provokes binding of the phosphoamino acid-binding domain, closing the proximity of CFP and YFP, provoking fluorescence energy transfer (FRET) that can be measured and imaged in real time in live cell [14, 15].
We have sought a hybrid approach that combines the power of the AKAR2 biosensor with a desire to measure events catalyzed by PKA, confined intentionally to single AKAPs, which are PKA substrates [16–19]. PKA catalyzed phosphorylation of AKAP5 is central to its function, mediating the mitogen-activated protein kinase cascade. Phosphorylation of AKAP12 is essential to the ability of AKAP12 to dephosphorylate, resensitize and recycle GPCRs . PKA-catalyzed phosphorylation per se regulates AKAP scaffolds, including those that dock ion channels and/or GPCRs [1, 6, 8, 16]. Earlier we reported the engineering and use of the first AKAP biosensor, AKAR2-AKAP12 . Herein, a novel fusion protein of the biosensor AKAR2 with the AKAP5 is created and characterized. AKAP5 is organized similarly to AKAP12, but is only one-quarter of the size of AKAP12. This novel fusion protein retains all the functional capability of the native AKAP5 and successfully acts as a biosensor. Activation of AKAR2-AKAP5 in live cells undergoing β-adrenergic activation of mitogen-activated protein kinase (MAPK) cascade can be monitored in live cells. We explore the impact of tethered PKA and tethered PDE4D on the “shape” of the cellular responses mediated by AKAP5, using AKAP12 to test findings obtained with the smaller AKAP.
Construction, expression, and function of AKAR2-AKAP5 fusion biosensor
Knock-down (KD) of AKAP5 with antisense morpholinos targeting the 5′-UTR of the human AKAP5 suppressed expression of endogenous AKAP5 (Figure 1B). Expression of the AKAR2-AKAP5 fusion biosensor, in contrast, was not targeted by the morpholinos (Figure 1B). KD of endogenous AKAP5 attenuated the activation of ERK1,2 in response to stimulation with the β-adrenergic agonist isoproterenol [ISO, 10 μM](Figure 1C, D). In control cells (i.e., untreated with morpholinos antisense to AKAP5), ISO stimulated a rapid, ~2.0-fold increase in activated, phosphorylated ERK1,2. The response was quantified in immunoblots stained with activation- and phospho-specific antibodies to ERK1,2. The cellular content of ERK1,2 was assayed by immunoblotting and staining with anti-ERK1,2 antibodies (bottom, Figure 1C). Expression of AKAR2-AKAP5 in the AKAP5-deficient cells fully rescued of the beta-adrenergic activation of ERK1,2 (Figure 1D). The time-course of β-adrenergic-induced ERK1,2 activation was rapid (observed within 1 min of addition of ISO), peaked within 2–3 min of β-adrenergic agonist, and was returned to baseline within ~10 min. Thus, AKAR2-AKAP5 is readily expressed and fully functional with respect to activation of ERK1,2 by β-adrenergic agonist.
AKAP5 and AKAP12 differentially localize
In vivo biosensing of PKA-catalyzed phosphorylation of AKAPs
Human epidermoid carcinoma A431 cells are a well-characterized cell line [11–13, 28, 29]. In A431 cells, FRET signaling of the same phosphorylation events derived from AKAR2-AKAP5 was markedly delayed in comparison to the signal derived in HEK cells (Figure 3A, 3B). The amplitude of the peak FRET signal sampled from AKAR2-AKAP5, as well as the delay in reaching and escaping from the ISO-stimulated activation were muted. FRET derived from AKAR2-AKAP12 in response to stimulation of A431 cells in response to ISO, in contrast, was similar to that observed in the HEK cells. The FRET signal derived from AKAR2 itself displayed an even slower, progressive increase in response to ISO. FRET signals from AKAR2 and AKAR2-AKAP12 showed no reversal. In cell lines expressing the inactive mutants AKAR2(T/P), AKAR2-AKAP5(T/P), and of AKAR2-AKAP12(T/P) no FRET signal was detected in response to stimulation with ISO (Figure 3A and B; data not shown).
Fidelity of biosensing with AKAP phosphorylation and function
HEK cells expressing the AKAR2-AKAP12 biosensor likewise were tested for fidelity with phosphorylation and function. This larger AKAP displays many of the same docking partners, including PKA, PKC, and β2-adrenergic receptors as its 3-fold smaller AKAP member, AKAP5. FRET signal from the two AKAP biosensors displayed the same response to chemical inhibitors of PKA and block of RI/II binding (Figure 4A, B). Inhibition of PKA activity blocked FRET signal of both biosensors in response to ISO. Treatment of cells to block PKA RI/II subunit docking to each AKAP also abolished the FRET signal in response to challenging the cells to ISO (Figure 4B).
Imaging PKA-catalyzed activation of AKAP5
PDE shaping of AKAP5 phosphorylation
Several members of the AKAP family dock cyclic AMP phosphodiesterases (PDE), in addition to protein kinase and phosphoprotein phosphatases. By degrading cyclic AMP, PDEs have been reported to “condition” cellular microenvironments [34, 35]. PDEs particularly members of the PDE4 family play central roles in reversing the activation of substrate proteins by PKA [35–40]. AKAP5 has been reported to tether PDEs in multiple ways [7, 35, 41], whereas AKAP12 displays no putative PDE-binding site. To test role of PDE in AKAP5 we adopted three interrelated approaches. The first was to create a dominant interfering peptide (DIP) from the AKAP5 sequence that would sequester PDE via the tethering site, but one that would not localize to the cell membrane. The first step was to delete all three N-terminal positively-charged domains (i.e., PCD1, PCD2, and PCD3) that localize AKAP5 to the cell membrane [26, 27], providing a “sink” to which PDE can tether away from the membrane. Next, a point mutation at the RII binding domain of this 146-427 AKAP5 fragment was inserted. Mutation of the RII binding site insured that the fragment would not be a “sink” for PKA. This mutated 146-427 peptide was tested to ascertain if it could function to sequester PDE away from AKAP5. HEK cells were stably transfected to express an HA-tagged version of the N-terminally-truncated, mutated AKAP5 146-427 (termed “Δ1-145,T/P395”). An expression vector harboring AKAR2-AKAP5 was transfected into either wild-type cells or cells expressing Δ1-145,T/P395. In immunoblots stained with anti-PDE4D5, cells expressing GFP-tagged AKAR2-AKAP5 display less bound PDE4D5 when co-transfected with Δ1-145,T/P395 peptide (Additional file 1: Figure S1A), i.e., the presence of the DIP reduces the amount of PDE tethered to AKAP5.
A second approach made use of the specific PDE4 inhibitor Rolipram (Figure 6B). Treating cells with Rolipram would block the catalytic activity of PDE, including that tethered to AKAP5. In the absence of Rolipram, ISO stimulated a rapid, transient peak of FRET signal from AKAR2-AKAP5. Inhibiting PDE activity did not alter the initial rapid peak of FRET signal, but rather blocked the reversal of the signal that rapidly follows in the absence of PDE inhibitor. Since Rolipram can inhibit several PDEs and only PDE4D5 is the member tethered to AKAP5 uniquely, we sought to ask a more precise question about the role of this PDE in AKAP5 function. Making use of siRNA specifically targeting PDE4D5, the expression of PDE4D5 was knocked-down (KD) (Figure 6C). Suppression of PDE4D5, like Rolipram, did not alter the initial peak of AKAR2-AKAP5-derived FRET signal, but effectively abolished the reversal of the signal which reflects loss of PKA-catalyzed phosphorylation of the AKAP.
AKAP5-mediated ERK1,2 activation: role of PDE
Biosensors can offer unparalleled insights into the temporal and spatial features of signaling in live cells. The AKAR2 biosensor enables the detection in real time of PKA-catalyzed phosphorylation events in live cells [15, 42, 43]. We successfully engineered AKAR2 to the N-terminus of AKAP5, providing a biosensor that reports from the local environment, in cis orientation with respect to AKAP5. The FRET-based signal generated from the AKAP5 biosensor displayed a rapid activation in response to beta-adrenergic stimulation. Within less than 10 s the FRET signal, reflecting in situ PKA-catalyzed phosphorylation of AKAP5, reached peak amplitude. PKA-catalyzed phosphorylation of AKAP5 underwent rapid reversal. PCDs confine AKAP5 to the cell membrane, in close proximity to the membrane-embedded adenylylcyclase that generates cyclic AMP. The localization of AKAP5 was highly uniform, decorating the inner leaflet of the cell membrane. AKAP12, in contrast, is sequestered in close proximity to the cell membrane, but after activation migrates away from the membrane with the GPRCs undergoing resensitization and recycling catalyzed by AKAP12 .
The role of one particular docking partner, PDE, was of keen interest. Recent studies have revealed that the microenvironment of cyclic AMP signaling are shaped by PDEs that degrade the cyclic nucleotide [7, 35, 44, 45]. We made use of a dominant interfering peptide fragment of AKAP5 (Δ1-145,T/P395) to sequester cellular PDE. Expression of the Δ1-145,T/P395 DIP did not alter the rapid peak of FRET signal derived from AKAP5 biosensor in response to ISO, but rather abolished the rapid reversal and transient character of the AKAP5 response. Knocking down of PDE4D5 and chemical inhibition of PDE4 also abolished the rectification of the PKA-catalyzed phosphorylation of AKAP5. Unlike AKAP5, AKAP12 does not display a binding site for PDE. PDE tethered through β-arrestin has been proposed as an alternative for AKAP12 biology . When tethered to AKAP12 via the insertion of the AKAP5 PDE binding site, PDE4D5 is shown to accelerate deactivation, shaping the reversal much like that of AKAP5 (Figure 8).
Materials and methods
Phusion® High-Fidelity DNA polymerase was purchased from New England Biolabs (Ipswich, MA). All DNA restriction enzymes employed in this study were purchased from New England Biolabs (Ipswich, MA). Antibodies AKAP5 (cat. #: sc-17772), β-Actin (cat. # sc-47778) and GFP (cat. #: sc-8334) were purchased from Santa Cruz (Santa Cruz, CA); HA (cat. #: 11867431001) was purchased from Roche (Indianapolis; IN); and anti-PDE4D5 antibodies (cat. #: GTX14626) were obtained from GeneTex (Irvine, CA). DMEM with 4.5 g/L glucose, sodium pyruvate without L-glutamine, and phenol red were purchased from Cellgro (Manassas, VA). InCELLect™ AKAP St-Ht31 Inhibitor Peptide and St-Ht31P Control Peptide were purchased from Promega (Madison, WI). The PKA inhibitors H89, KT5720, and the β-adrenergic agonist ISO were obtained from Sigma-Aldrich, Inc (St. Louis, MO). DNA ligase, transfection reagent Lipofectamine® and all siRNAs employed in these studies was purchased from Invitrogen (Carlsbad, CA). The 35 mm glass bottom dishes with 13 mm wells were obtained from Cell E&G LLC (Huston, TX). The Nikon Eclipse TE300 microscope and a cool charge-coupled device digital camera ORCA-ER were purchased from Hamamatsu (Bridgewater, NJ). METAFLUOR 7.2 software was from Olympus (Center Valley, PA). The filter changer Lambda 10-3 and the Lambda DG-4 were purchased from Shutter Instrument (Novato, CA). Antisense morpholino oligonucleotides (morpholinos) were designed, synthesized, and purified by Gene Tools, LLC (Philomath, OR).
Construction of plasmids
Construction of AKAR2-AKAP12 and AKAR2(T391A)-AKAP12 plasmids was described earlier . AKAR2 was engineered into the N-terminus of human AKAP12. Cellular targeting and the biological activity of the AKAP12 were unaffected by attachment of the AKAR2 moiety . The cellular distribution of AKAR2-fused AKAP12 (Figure 2) and that of the endogenous AKAP12 as well as the GFP-tagged AKAP12 are identical to earlier published studies [2, 6, 16, 19].
AKAR2-AKAP5 was constructed and engineered into an expression plasmid using the same strategy. AKAR2 sequence was amplified using PCR with primers containing HindIII and EcoRI at N- or C- termini, respectively. AKAP5 sequence was amplified with primers containing EcoRI and NotI found at N- or C- termini, respectively by using Phusion® High-Fidelity DNA polymerase (New England Biolabs; Ipswich, MA). PCR products were subcloned simultaneously into pcDNA 3.1+ vector, which then was digested with HindIII and NotI. To construct mutant AKAR2(T/P)-AKAP5 the Threonine-to-Proline substitution mutant of AKAR2 [which was generated before ] was amplified, digested, and subcloned into pcDNA3.1+ vector in tandem with the AKAP5 sequence. The targeting and biological activity of AKAP5 was unaffected by fusion with AKAR2 biosensor (Figure 1). The cellular distribution of AKAR2-fused AKAP5, endogenous AKAP5, and GFP-fused AKAP5 in the current studies are almost identical to those reported earlier by others [22, 24, 25, 49].
AKAR2-AKAP12-[Δ1-145,T/P395] was created by fusing AKAP5 which is harboring the point mutation at PKA binding site and has been truncated (i.e., the first 145 amino acids deleted) making use of NotI-XhoI restriction sites in the DNA.
Cells were transfected with an expression vector encoding one of the following: AKAR2; AKAP5; AKAR2-AKAP5 or AKAR2(T/P)-AKAP5. Samples isolated from whole-cell lysates of the transfected cells were separated by SDS-PAGE and transferred to nitrocellulose membranes. Blots were probed with polyclonal antibodies specific for AKAP5 (1/1000 dilution), HA (1/1000 dilution) and GFP (1/1000 dilution). Western blotting technique used in this study is described elsewhere [50, 51]. Antisense morpholino oligonucleotides (morpholinos) specified to the 5′ untranslated region (UTR) of human AKAP5 were employed only in the experiments whose results are displayed in Figure 1. The expression of AKAP5 was suppressed in HEK 293 cells by use of antisense morpholinos which targeted the 5′ untranslated region of human AKAP5, as described earlier . In all other cases, knock-down of protein expression was accomplished by use of siRNAs. Where employed, antisense morpholinos were designed, synthesized, and purified by Gene Tools. At 24 h later these morpholinos-treated cells were divided into two batches, one group was transfected with empty plasmid and the other with plasmid encoding AKAR2-AKP5 fusion protein. At 72 h after initial morpholino transfection AKAP5-deficient and wild-type cells were stimulated with 10 μM ISO. Samples obtained from whole-cell lysates were subsequently subjected to SDS-PAGE and the resolved proteins transferred to nitrocellulose membranes. Blots were probed with antibodies specific for either pERK or pan-ERK. The efficiency of the antisense morpholinos in suppressing AKAP5 was complete and established by immunoblotting with anti-AKAP5 antibody (Santa Cruz; Santa Cruz, CA). β-actin was stained with antibodies (1/1000) is used as an internal control for sample loading in immunoblotting. Representative immunoblots displayed show the efficiency of suppression of the target proteins using either antisense morpholinos or siRNAs (Figure 1B and 9B).
Confocal fluorescence microscopy
Cells were plated on 35 mm glass bottom dishes with 1 3 mm well produced by Cell E&G LLC. At 24 h later, cells were transfected with plasmids encoding one of the following: AKAR2, AKAR2(T/P), AKAR2-AKAP5, AKAR2(T/P)-AKAP5, AKAR2-AKAP12 or AKAR2(T/P)-AKAP12. At 24 h later, cells were washed twice with DMEM supplemented with 4.5 g/L glucose, sodium pyruvate free of L-glutamine and phenol red (Cellgro, Manassas, VA). Cells were stained with Hoechst 33258 (diluted 1:5000 from stock solution of 10 mg/ml) for 30 min in a humidified atmosphere of 5% CO2 and 95% air maintained at 37°C. Microscopy was performed on a Zeiss LSM 510 scanning laser confocal microscope, as described .
Live cell imaging
Live cell imaging was performed as described . Fluorescence ratio imaging was performed on a Nikon Eclipse TE300 microscope and a cool charge-coupled device digital camera ORCA-ER controlled by METAFLUOR 7.2 software (Olympus, Center Valley, PA). Emission ratios were obtained by using a 420DF20 excitation filter, a 450DRLP dichroic mirror, and two emission filters (475DF40 for cyan and 535DF25 for citrine) controlled by filter changer Lambda 10-3 (Shutter Instrument, Novato, CA). The Lambda DG-4 (Shutter Instrument, Novato, CA) illumination system was employed as the light source. Cells were transfected with the indicated plasmids and maintained as described above. At 24 h post transfection, cells were washed twice with DMEM with 4.5 g/L glucose and sodium pyruvate without L-glutamine and phenol red. Cells were treated at time 0 with 10 μM ISO. The ratios of yellow-to-cyan were recorded at every 5 s by METAFLUOR software and normalized by dividing all ratios by the emission ratio before stimulation. Similarly, pseudocolor images which show the change in FRET in response to ISO (10 μM) stimulation at various time points recorded at every 5 s by METAFLUOR software.
Recycling of β2-Adrenergic Receptor
A431 cells were stable transfected with Δ1-145,T/P395. AKAP12, AKAP5 or PDE4D5 were suppressed, where indicated, by specific siRNAs. The desensitization, internalization, resensitization/recycling assays were performed as described elsewhere .
We would like to thank to Dr. Roger Tsien (University of California, San Diego) for providing to us the AKAR2 construct. These studies are generously supported by a United States Public Health Services award (DK25410-33) from the National Institutes of Health (NIDDK).
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License ( http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.