GPR39 is a G-protein-coupled zinc receptor that protects against diverse effectors of cell death. Its protective activity is mediated via constitutive activation of Gα13 and the RhoA pathway, leading to increased SRE (serum-response element)-dependent transcription; the zinc-dependent immediate activation of GPR39 involves Gq-mediated increases in cytosolic Ca2+ and Gs coupling leading to increased cAMP levels. We used the cytosolic and soluble C-terminus of GPR39 in a Y2H (yeast-2-hybrid) screen for interacting proteins, thus identifying PKIB (protein kinase A inhibitor β). Co-expression of GPR39 with PKIB increased the protective activity of GPR39 via the constitutive, but not the ligand-mediated, pathway. PKIB inhibits protein kinase A by direct interaction with its pseudosubstrate domain; mutation of this domain abolished the inhibitory activity of PKIB on protein kinase A activity, but had no effect on the interaction with GPR39, cell protection and induction of SRE-dependent transcription. Zinc caused dissociation of PKIB from GPR39, thereby liberating it to associate with protein kinase A and inhibit its activity, which would result in a negative-feedback loop with the ability to limit activation of the Gs pathway by zinc.

INTRODUCTION

GPCRs (G-protein-coupled receptors) constitute the largest family of cell-surface transmembrane proteins [1]. They are activated by a wide variety of natural ligands and pharmacological alteration of their signalling constitutes one of the most successful approaches to the treatment of human disease, making them the most targeted protein superfamily in pharmaceutical research [2]. GPR39 (GPCR 39) is a GPCR related to the neurotensin receptors NTR1 and NTR2 [3] and is expressed in the human gastrointestinal tract and brain [4]. GPR39 appears to be involved in pancreatic islet function. When fed on a high-sucrose diet, GPR39-deficient [KO (knockout)] mice had increased fed glucose levels and showed decreased serum insulin levels during an oral glucose tolerance test in the face of unchanged insulin tolerance [5]. A different study using the same GPR39-KO mice came to almost the same conclusions: GPR39-KO mice displayed normal insulin sensitivity, but moderately impaired glucose tolerance during both oral and intravenous glucose tolerance tests and had a decreased plasma insulin response to oral glucose [6].

In addition, GPR39 also appears to play a role in cell death. We found GPR39 to be up-regulated in a hippocampal cell line resistant against diverse stimulators of cell death and showed that its overexpression protected against oxidative and endoplasmic reticulum stress as well as against direct activation of the caspase cascade by BAX overexpression. Silencing GPR39 rendered cells more susceptible to cell death [7]. An array analysis of transcripts induced by GPR39 revealed up-regulation of RGS16 (regulator of G-protein signalling 16), a specific inhibitor of coupling to Gα13 and induction of SRE (serum-response-element)-mediated transcription by the small GTPase RhoA [8]. In line with this, co-expression of GPR39 with RGS16, or dominant-negative RhoA or SRF (serum-response factor), abolished the GPR39 cell protection [7]. Others have reproduced these anti-apoptotic effects of GPR39 using different model systems. When overexpressed in vivo, specifically in pancreatic β-cells under the control of an inducible proinsulin promoter, GPR39 protected β-cells against streptozotocin-induced cell death and thus inhibited non-fasting hyperglycaemia [9]. GPR39 was also found to be frequently up-regulated in human oesophageal squamous cell carcinoma and to possess a strong tumorigenic capacity that leads to increased cell proliferation, foci and colony formation and tumour growth in nude mice, which might be related to its anti-apoptotic activity [10]. Additionally, GPR39 protected against butyrate-induced cell death via up-regulation of clusterin when overexpressed in HT29 colonocytes [11]. GPR39 thus constitutes a novel inhibitor of cell death and may represent a therapeutic target.

GPR39 clearly has a constitutive activity, as overexpression increased SRE-mediated transcription in a concentration-dependent manner [12]. However, zinc ions also elicited a moderate effect on inositol phosphate turnover and CRE (cAMP-response element)-dependent transcription in these cells [12]. In a direct comparison of zinc and obestatin, a peptide erroneously reported to be a ligand for GPR39, only zinc increased inositol phosphate turnover, cAMP production and arrestin mobilization as well as CRE- and SRE-dependent transcriptional activity in GPR39-overexpressing cells [13]. Other groups have also found an increase in Ca2+ mobilization in GPR39-expressing cells by zinc ions, which could be inhibited by phospholipase C inhibitors, suggesting coupling to Gq proteins [14]. It is therefore now generally accepted that GPR39 is a zinc receptor and that the zinc-dependent immediate activation of GPR39, which is probably mediated via Gq and Gs pathways, is distinct from the constitutive activation of SRE-dependent transcription via Gα13 observed after overexpression by us and others [7,12].

In the present study, we used the cytosolic and soluble C-terminus of GPR39 to screen for interacting proteins and identified the cAMP-dependent PKIB [PKA (protein kinase A) inhibitor β]. PKIB was shown to be up-regulated in aggressive castration-resistant prostate cancer, where it enhanced growth and mobility of cancer cells through enhanced phosphorylation of Akt at Ser473 by direct interaction with the PKA catalytic subunit C [15]. We describe in the present paper how PKIB co-expression with GPR39 increases the constitutive activation of GPR39 without interfering with its ligand-mediated function independently of its function as an inhibitor of PKA.

EXPERIMENTAL

Cells, proteins and plasmids

Mouse neuroblastoma (Neuro2a) and hippocampal (HT22) cell lines were maintained in DMEM (Dulbecco's modified Eagle's medium) high glucose (PAA Laboratories) supplemented with 5% FBS (Hyclone). CHO (Chinese-hamster ovary) cells and human HeLa cells were maintained in DMEM supplemented with 10% FBS. All cell lines were propagated in the presence of 1% penicillin/streptomycin (Gibco). ZnCl2 was purchased from Sigma and diluted in distilled water. Chelex® 100 was purchased from Bio-Rad Laboratories. GPR39, PKIB, NECAB2 (N-terminal EF hand calcium-binding protein 2) and mutated PKIB containing the amino acid sequence AAAA instead of the PKA pseudosubstrate domain RRNA, synthesized by Geneart, were cloned into pENTR (Invitrogen) and recombined into the appropriate expression vectors using Gateway technology (Invitrogen). The GPR39 C-terminus was cloned by PCR into pGBT9 (Invitrogen) using the forward primer 5′-GAATTCTCTCTCAACAATTTAGAAG-3′ and the reverse primer 5′-GTCGACTCAAACTTCATGCTCCTG-3′. Human PKIB2 was cloned by PCR into pGEX using the forward primer 5′-GGATCCAGGACAGATTCATC-3′ and the reverse primer 5′-GTCGACTCATTTTTCTTC-3′.

Y2H screening

Y2H (yeast-2-hybrid) screening of a human brain library was done by RZPD according to protocols published previously [16].

Transfections

High-purity plasmids were prepared using Nucleobond AX 500 columns (Macherey-Nagel). For transient transfections, the cells were grown to 80–90% confluence and transfected with Attractene (Qiagen) or with Lipofectamine™ 2000 (Invitrogen).

Western blotting and co-immunoprecipitation assays

For immunoblotting, cells were transfected with the indicated constructs or treated as indicated and lysed in RIPA buffer (Sigma) containing a protease inhibitor cocktail (Roche). A 30 μg aliquot of the total cell lysate protein was separated on a 4–20% Precise™ protein gel (Thermo Scientific) and transferred on to a nitrocellulose membrane (Invitrogen). To block unspecific binding, the membrane was incubated in PBS-T (PBS and 0.05% Tween 20) containing 4% (w/v) non-fat dried skimmed milk powder (Nestlé) for 1 h at room temperature. After blocking, the membranes were probed with the primary antibody [anti-Myc, 1:1000 dilution; anti-FLAG, 1:1000 dilution; or anti-HA (haemagglutinin), 1:1000 dilution] and the loading control antibody anti-actin (1:3000 dilution; Millipore) in PBS-T and 4% (w/v) non-fat dried skimmed milk powder at 4°C overnight and then washed three times with PBS-T before incubation for 1 h at room temperature with the respective infrared fluorescence-conjugated secondary antibodies (IRdye 800–anti-mouse and IRdye 680–anti-rabbit, 1:30000 dilution; Licor). For detection, the membrane was scanned for infrared fluorescence at 680 and 800 nm with the Odyssey system (LI-COR Biosciences).

Co-immunoprecipitations were performed using the ProFound™ HA and Myc Tag IP/Co-IP Kit (Pierce) according to the manufacturer's instructions. Briefly, CHO cells were transfected with the indicated constructs and lysed 48 h later in M-PER® Mammalian Protein Extraction Reagent (1000 μl/10-cm plate) containing protease inhibitors. The lysate (200 μl; ~500 μg of total protein) was incubated with 10 μl of anti-HA (or anti-Myc for the corresponding assay) antibody agarose slurry at 4°C overnight, eluted with 40 μl of non-reducing sample buffer, and 20 μl of each sample separated by SDS/PAGE (4–20% gel), transferred on to a nitrocellulose membrane and probed with the antibodies as described above.

Fluorescence imaging

Equal numbers of HT22 cells were seeded into 48-well plates with glass coverslips or imaging 96-well plates (BD Falcon™ Microplates). Transfections were carried out 24 h before seeding. For co-localization studies, cells were fixed for 15 min in 2% PFA (paraformaldehyde; Roti®-Histofix 4%) at room temperature and blocked and permeabilized with 0.2% FBS and 0.2% BSA in PBS/0.01% Triton X-100 for 30 min on ice. Nuclei were counterstained with 100 ng/ml DAPI (Life Technologies) for 15 min at room temperature and then washed and embedded in Vectashield™ (Biozol). The cells were imaged using a Zeiss LSM confocal microscope.

Calcium imaging

For zinc studies, transiently transfected cells were incubated overnight in zinc-free medium (Chelex-treated serum) by incubating 50 ml of regular medium with 2.5 g of Chelex® 100 as described above for 1 h with agitation. After filtration of the supernatant, CaCl2 (1.8 mM) and MgSO4 (0.8 mM) were added. To measure the zinc-dependent calcium efflux, cells were loaded with 2 nM fura 2/AM (acetoxymethyl ester; Life Technologies) for 30 min at 37°C, washed twice and incubated in 50 μl of HBSS+/+ (Hanks balanced salt solution) diluted 1:1 with 200 μM ZnCl2. The cells were imaged using a BD Pathway 855 high-content imaging system.

Ratiometric imaging of the AKAR4 sensor

Real-time FRET imaging experiments using the AKAR4 sensor were performed using a fluorescence imaging system built around a Nikon TE200 inverted fluorescence microscope equipped with a Photometrics QuantEM EMCCD (electron multiplying charge-coupled device) camera. Metafluor software (Molecular Devices) was used to control filter wheels (Sutter Instruments) placed in the excitation and emission path and to acquire ratio data. Cells expressing AKAR4 were seeded on to glass coverslips and 24 h later the coverslips were mounted on to a home-built flow-through perfusion chamber before the cells were imaged using a ×60 oil-immersion objective. Cells were bathed in Hepes-buffered Ringer's solution containing 125 mM NaCl, 25 mM Hepes, 10 mM glucose, 5 mM K2HPO4, 1 mM MgSO4 and 1 mM CaCl2 (pH 7.40). FRET emission ratios were acquired every 5–10 s. PKA phosphorylation activity was expressed as the 535 nm/485 nm FRET emission ratios of AKAR4 (440 nm excitation). The fluorescence of mCherry (excitation 585 nm and emission 610 nm) did not interfere with these measurements, as has been reported previously [17].

Viability assays

For toxicity experiments, 8000 Neuro2a or HT22 cells were seeded with 100 μl of medium in 96-well plates. Transfections were carried 24 h before seeding. At 24 h after transfection, tunicamycin, glutamate or vehicle were added to the medium and 24 h later, cells were treated with CTB (CellTiter-Blue; Promega) and the metabolic activity quantified photometrically (Tecan).

Luciferase assays

Neuro2a and HT22 cells were transiently transfected in 48-well plates with the indicated luciferase reporter plasmids, a SRE-Renilla or EGFP control plasmid and the indicated expression constructs. At 48 h after transfection, cells were washed and lysed in 200 μl of cell lysis buffer (Promocell). The lysate was then centrifuged at 12000 g for 1 min and 20 μl of supernatant was transferred into a white 96-well microtiter plate. Directly before measurement, 100 μl of luciferase assay buffer (Promega) was injected into each well. Luminescence was measured by a Genios Pro microplate reader (Tecan) and integrated for 10000 ms. Normalization was performed by either EGFP fluorescence or dual luciferase assays before 20 μl of lysate was transferred to a second white microtiter plate and 40 μl of Renilla assay enhancer solution was added to the wells. Coelenterazine (Promocell) was then injected into the well in the appropriate assay buffer and luminescence was measured as described above.

Co-localization assays

HT22 cells were plated in a six-well tissue culture plate and grown for 24 h. Cells were transiently transfected with pKIB-mCherry and GPR39–GFP using Lipofectamine™ 2000 Transfection Reagent. At 24 h later, 75000 cells were seeded into 10-mm coverslips and grown in zinc-free medium, as described above, for another 24 h. Medium was replaced by zinc-free medium and incubated with 200 μM zinc or vehicle for 2 or 15 min. Then the cells were washed with PBS and fixed with 4% PFA for 20 min at room temperature. Afterwards the cells were washed three times with PBS. Cell nuclei were stained with DAPI for 3 min and washed again twice with PBS. Coverslips were removed from the culture plate, washed with distilled water and fixed on to a microscope slide with mounting medium (Dako). Images were taken with a confocal microscope (SP5; Leica) and the correlation coefficient quantitated using the van Steensel approach within the JACoP software (ImageJ; http://imagej.nih.gov/ij/).

Statistical analysis

Results are shown as means±S.E.M. and the statistical significance assessed using two-tailed Student's t test or ANOVA, followed by the Dunnet's test as indicated.

RESULTS

Identification of GPR39-interacting proteins by Y2H screening

In order to identify proteins that might interact with GPR39, we screened a human brain library for interacting proteins using the Y2H technique with the C-terminus as bait, as this soluble intracellular part is the major interaction scaffold of GPCRs. Such an approach has been successfully used before for other GPCRs (reviewed in [18]). Our screen resulted in three main potential interacting proteins. A total of 24 clones corresponded to NECAB2, two to PKIB and four to CLIC4 (chloride intracellular channel 4). CLIC4 clones were not in frame and were not considered further. NECAB2 clones contained different parts of the ORF and the two PKIB clones encompassed the complete frame.

NECAB2 has been identified previously in a Y2H screen as an interactor of the adenosine A2a receptor C-terminus [19]. When co-expressed with the adenosine A2a receptor, NECAB2 decreased cell-surface expression, attenuated ligand-dependent internalization and enhanced receptor-mediated ERK1/2 (extracellular-signal-regulated kinase 1/2) phosphorylation [19]. In contrast, PKIB has never been identified in a Y2H screen. PKIB is one of three specific inhibitors of the catalytic subunit of cAMP-dependent protein kinase (also known as PKA). Inhibition is caused by binding of a pseudosubstrate site to the catalytic subunit of PKA. It has also been demonstrated that complexes of the PKIs (protein kinase inhibitors) and the catalytic subunit of PKA are more rapidly exported out of the nucleus in a temperature- and ATP-dependent process. The complex of PKIB with the catalytic subunit C of PKA was recently shown to enhance phosphorylation of Akt [15]. PKIs therefore have two conserved domains: a nuclear export signal and a pseudosubstrate of PKA. It is noteworthy that PKIB is approximately 40-fold less potent than its homologues PKIα and PKIγ, suggesting that it has other functions besides the inhibition of PKA (reviewed in [20]).

GPR39 interacts and co-localizes with PKIB at the plasma membrane

We cloned both potential interacting proteins, NECAB2 and PKIB, in-frame with an N-terminal FLAG tag, and overexpressed them together with the soluble c-Myc-tagged GPR39 C-terminal domain, which was used as bait in the Y2H screen, in CHO cells. FLAG–PKIB, but not FLAG–NECAB2 was unequivocally co-immunoprecipitated with the anti-Myc antibody (Figure 1A). It was not possible to use full-length GPR39 because GPCRs, including GPR39, are notoriously difficult to detect in immunoblots. We therefore cloned mCherry-tagged PKIB and NECAB2 in order to visualize their intracellular localization and possible co-localization with an EGFP-tagged full-length GPR39. When co-expressed in human HeLa cells, both potential interacting proteins were found mainly in the cytosol; however, PKIB, but not NECAB2, co-localized with GPR39 at the plasma membrane (Figure 1B). We therefore focused on PKIB in all subsequent experiments. It must be mentioned, however, that we later found out that we had erroneously used a NECAB2 splice variant that lacked a significant part of the N-terminus including the GPCR-interacting domain described previously [19]. We can therefore not rule out that full-length NECAB2 interacts with GPR39 similarly to its previously reported interaction with other GPCRs.

GPR39 interacts and co-localizes with PKIB at the plasma membrane

Figure 1
GPR39 interacts and co-localizes with PKIB at the plasma membrane

(A) FLAG-tagged PKIB and NECAB2 were co-expressed with Myc-tagged GPR39 C-terminus (G39CT) in CHO cells and immunoprecipitated with anti-Myc-coupled beads. Immunoblots (IB) demonstrate the input. Lower panel shows the immunoblot of immunoprecipitated samples (IP) stained simultaneously with anti-Myc and anti-FLAG antibodies that can be separated with different secondary antibodies, but are shown here as a merged picture. (B) mCherry-tagged PKIB, but not NECAB2, lacking its N-terminal domains co-localizes with full-length GPR39–EGFP at the plasma membrane (arrow). Scale bar, 10 μm.

Figure 1
GPR39 interacts and co-localizes with PKIB at the plasma membrane

(A) FLAG-tagged PKIB and NECAB2 were co-expressed with Myc-tagged GPR39 C-terminus (G39CT) in CHO cells and immunoprecipitated with anti-Myc-coupled beads. Immunoblots (IB) demonstrate the input. Lower panel shows the immunoblot of immunoprecipitated samples (IP) stained simultaneously with anti-Myc and anti-FLAG antibodies that can be separated with different secondary antibodies, but are shown here as a merged picture. (B) mCherry-tagged PKIB, but not NECAB2, lacking its N-terminal domains co-localizes with full-length GPR39–EGFP at the plasma membrane (arrow). Scale bar, 10 μm.

PKIB enhances GPR39-mediated protection against oxidative stress by increasing constitutive, but not ligand-mediated, signalling

Oxidative glutamate toxicity in the hippocampal cell line HT22 is a model for endogenous oxidative stress. In these cells, extracellular glutamate inhibits the glutamate/cysteine antiporter system and eventually leads to glutathione depletion and cell death by oxidative stress [2123]. When overexpressed in HT22 cells, GPR39 protects against glutamate toxicity, endoplasmic reticulum stress and direct activation of the apoptotic cascade by overexpression of BAX [7]. Interestingly, PKIB also protected against glutamate, which was, however, synergistically enhanced by co-expression with GPR39, providing support for the idea of a co-operative rather than solely additive mechanism (GPR39 25%, PKIB 33%, GPR39+PKIB 63% survival at 20 mM glutamate; Figure 2A).

PKIB enhances GPR39-mediated protection against oxidative stress by increasing constitutive but not ligand-mediated signalling

Figure 2
PKIB enhances GPR39-mediated protection against oxidative stress by increasing constitutive but not ligand-mediated signalling

(A) HT22 cells (5000) were transfected with the indicated constructs, seeded after 24 h into 96-well plates and subjected to the indicated concentrations of glutamate again 24 h later. Viability was quantitated 16 h later with the CTB reagent and normalized. (B) HT22 cells were transfected with the indicated constructs and an SRE-luciferase reporter plasmid. Values are given as fold of luminescence over the empty vector. *P<0.05 compared with the vector as determined by ANOVA and the Dunnet's post-test. (C) HT22 cells were transfected with the indicated constructs and mCherry and kept in zinc-free medium for 24 h. Then the cytosolic Ca2+ concentration was measured in mCherry-positive cells as fura 2/AM ratio in response to 200 μM zinc. (D) The histogram represents the mean area under the curve±S.D. for three independent experiments. n.s., not significant.

Figure 2
PKIB enhances GPR39-mediated protection against oxidative stress by increasing constitutive but not ligand-mediated signalling

(A) HT22 cells (5000) were transfected with the indicated constructs, seeded after 24 h into 96-well plates and subjected to the indicated concentrations of glutamate again 24 h later. Viability was quantitated 16 h later with the CTB reagent and normalized. (B) HT22 cells were transfected with the indicated constructs and an SRE-luciferase reporter plasmid. Values are given as fold of luminescence over the empty vector. *P<0.05 compared with the vector as determined by ANOVA and the Dunnet's post-test. (C) HT22 cells were transfected with the indicated constructs and mCherry and kept in zinc-free medium for 24 h. Then the cytosolic Ca2+ concentration was measured in mCherry-positive cells as fura 2/AM ratio in response to 200 μM zinc. (D) The histogram represents the mean area under the curve±S.D. for three independent experiments. n.s., not significant.

We next investigated the effect of co-expression on the ability of GPR39 to increase SRE-dependent transcription, which is the presumed mechanism of cell protection [7]. We overexpressed empty vector, GPR39+vector, PKIB and GPR39+PKIB together with an SRE luciferase reporter construct and quantitated the emitted light. The effect of GPR39 is concentration-dependent [12]; we therefore chose concentrations of GPR39 that result in only a minor SRE induction, to avoid a ceiling effect. Similarly to the effect on cell death, PKIB was capable of increasing GPR39-mediated SRE induction (Figure 2B). PKIB alone had no effect on SRE-mediated transcription, suggesting that other mechanisms participate in the cell protection mediated by this protein without GPR39. Zinc is the endogenous ligand of GPR39 and causes Ca2+ release from the endoplasmic reticulum [24]. To clarify whether PKIB also affects this ligand-mediated function of GPR39, we co-transfected HT22 cells with GPR39 or GPR39 and PKIB together with an mCherry empty vector in order to identify transfected cells and quantified the Ca2+ released by 200 μM zinc by single-cell Ca2+ imaging using the ratiometric dye fura 2/AM. To rule out an effect due to zinc being present in the medium, we first incubated the cells overnight in zinc-free medium to which CaCl2 and MgSO4 had been added. Here, PKIB had no effect on the amount of Ca2+ released after zinc treatment (Figures 2C and 2D). In the cell death and luciferase experiments, FLAG-tagged or untagged PKIB was used, as mCherry-tagged PKIB had opposite effects probably due to the bulky mCherry moiety at its N-terminus (results not shown).

On the basis of our co-immunoprecipitation assays and confocal microscopy, we conclude that PKIB interacts with GPR39. Additionally, single-cell Ca2+ imaging, as well as luciferase and viability assays, suggest that PKIB synergistically enhances the effects of GPR39 on SRE induction and cell protection, but has no effect on the ligand-mediated activity of GPR39.

The effect of PKIB on GPR39-mediated SRE induction is independent of its PKA-inhibiting activity

PKIB inhibits protein kinase A by direct interaction with its pseudosubstrate domain (reviewed in [20]) and this complex increases phosphorylation of Akt in LNCaP cells [15]. To investigate whether the positive effect on SRE-mediated transcription is a downstream effect of PKA inhibition or a direct effect on GPR39 and its Gα13 coupling, we cloned PKIB lacking the pseudosubstrate domain RRN, which we replaced with AAA, and tested its loss of function. HeLa cells were either transiently transfected with the PKA sensor AKAR4 [25] alone or co-transfected with mCherry-tagged PKIB or the PKA-insensitive mutant mCherry–PKIB–ΔPKA. The next day, equal numbers of AKAR4 and PKIB-expressing cells were mixed and seeded on to glass coverslips. During these experiments we chose, in the same visual field, cells expressing AKAR4 alone or cells that were positive for both AKAR4 and one of the mCherry-tagged PKIB constructs. As shown in Figure 3(A), PKIB–ΔPKA did not alter the cellular PKA activity measured by AKAR4, whereas wild-type PKIB completely abolished PKA activity (n=3 experiments for each PKIB and PKIB–ΔPKA). Overexpressed PKIB also attenuated forskolin-induced CRE-mediated transcription in luciferase assays, which was increased instead by PKIB–ΔPKA (Figure 3B). This demonstrates further that deleting this domain indeed inhibits the PKA-inhibiting activity of PKIB. PKIB–ΔPKA was, however, still able to interact with GPR39, as shown by co-immunoprecipitation (Figure 3C), suggesting that the GPR39-interacting domain of PKIB is distinct from its pseudosubstrate domain.

The effect of PKIB on GPR39-mediated SRE induction is independent of its PKA-inhibiting activity

Figure 3
The effect of PKIB on GPR39-mediated SRE induction is independent of its PKA-inhibiting activity

(A) Cells were transfected with AKAR4 (continuous trace) and AKAR4 together with PKIB (dashed trace) or the PKA-blind mutant ΔPKA–PKIB (dotted trace), both tagged with mCherry. Cells expressing AKAR4 alone promptly responded to treatment with maximal stimuli using the broad adenylyl cyclase activator forskolin (FSK) combined with the phosphodiesterase inhibitor 3-isobutyl-methylxanthine (IBMX). In cells expressing PKIB, the AKAR4 response was completely abolished, but was unaltered by overexpression of the PKIB mutant lacking the PKA-blocking domain. Both AKAR4 and AKAR4 plus ΔPKA–PKIB responses to FSK and IBMX treatment were reversible by the addition of the broadly used PKA inhibitor H89, proof of the specificity of signal (n=6 experiments). (B) Cells were transfected with the indicated constructs and treated with forskolin for 2 h before CRE transcription was quantified by dual luciferase assays. (C) Immunoblot (IB) showing Myc-tagged G39CT with FLAG-tagged PKIB or mutated ΔPKA–PKIB co-transfected into CHO cells. Immunoprecipitation (IP) shows interaction independent of the presence of the PKA inhibitory domain. Both panels show immunoblots stained simultaneously with anti-Myc and anti-FLAG antibodies that can be separated with different secondary antibodies, but are shown here as a merged picture. (D) Neuro2a cells were transfected with the indicated constructs and SRE-mediated transcription quantified by dual luciferase assays. Results are mean±S.D. for three independent experiments done in quadruplicate. DPKA, ΔPKA; EV, empty vector; n.s., not significant. *P<0.05.

Figure 3
The effect of PKIB on GPR39-mediated SRE induction is independent of its PKA-inhibiting activity

(A) Cells were transfected with AKAR4 (continuous trace) and AKAR4 together with PKIB (dashed trace) or the PKA-blind mutant ΔPKA–PKIB (dotted trace), both tagged with mCherry. Cells expressing AKAR4 alone promptly responded to treatment with maximal stimuli using the broad adenylyl cyclase activator forskolin (FSK) combined with the phosphodiesterase inhibitor 3-isobutyl-methylxanthine (IBMX). In cells expressing PKIB, the AKAR4 response was completely abolished, but was unaltered by overexpression of the PKIB mutant lacking the PKA-blocking domain. Both AKAR4 and AKAR4 plus ΔPKA–PKIB responses to FSK and IBMX treatment were reversible by the addition of the broadly used PKA inhibitor H89, proof of the specificity of signal (n=6 experiments). (B) Cells were transfected with the indicated constructs and treated with forskolin for 2 h before CRE transcription was quantified by dual luciferase assays. (C) Immunoblot (IB) showing Myc-tagged G39CT with FLAG-tagged PKIB or mutated ΔPKA–PKIB co-transfected into CHO cells. Immunoprecipitation (IP) shows interaction independent of the presence of the PKA inhibitory domain. Both panels show immunoblots stained simultaneously with anti-Myc and anti-FLAG antibodies that can be separated with different secondary antibodies, but are shown here as a merged picture. (D) Neuro2a cells were transfected with the indicated constructs and SRE-mediated transcription quantified by dual luciferase assays. Results are mean±S.D. for three independent experiments done in quadruplicate. DPKA, ΔPKA; EV, empty vector; n.s., not significant. *P<0.05.

Using this tool, we were able to clarify whether the induction of SRE signalling by PKIB was mediated via its effect on PKA. We co-expressed GPR39 with untagged PKIB or PKIB–ΔPKA and quantitated SRE induction, which showed no difference between wild-type and mutant PKIB (Figure 3D). Not surprisingly, PKIB–ΔPKA also had no effect on zinc-induced Ca2+ release (results not shown).

These results imply that the effects of PKIB on the constitutive activity of GPR39 are independent of PKIB's effect on PKA.

Zinc treatment dissociates PKIB and GPR39

We hypothesized that zinc causes dissociation of PKIB from GPR39, thereby attenuating its constitutive activity. In this model, cytosolic PKIB then associates with PKA and inhibits its activity, which would result in a negative-feedback loop with the ability to limit activation of the Gs pathway by zinc. In order to prove this hypothesis, we again co-transfected GPR39 and PKIB in HT22 cells, incubated the cells overnight in zinc-free medium to which CaCl2 and MgSO4 had been added, and stimulated the receptor with 200 μM of its endogenous ligand zinc. Then, the cells were fixed in PFA after 2 or 15 min and analysed by confocal microscopy. To avoid investigator bias and because we did not observe an obvious difference with the naked eye, we employed van Steensel's cross-correlation coefficient (reviewed in [26]) to quantitate the differences in GPR39–PKIB interaction in zinc-treated cells. This revealed a significant reduction in the cross-correlation coefficient after 2, but not after 15, min of exposure to zinc (Figure 4). This implies that binding of its ligand zinc alters the conformation of GPR39 causing dissociation of PKIB.

Zinc treatment dissociates PKIB and GPR39

Figure 4
Zinc treatment dissociates PKIB and GPR39

HT22 cells were transfected with GPR39–GFP and PKIB–mCherry and kept in zinc-free medium for 24 h before the addition of 200 μM zinc or vehicle for the indicated time before fixing with PFA. Then, the van Steensel's cross-correlation coefficient of n=6 cells was analysed using ImageJ and pooled. A higher coefficient corresponds to more co-localization. Scale bar, 5 μM.

Figure 4
Zinc treatment dissociates PKIB and GPR39

HT22 cells were transfected with GPR39–GFP and PKIB–mCherry and kept in zinc-free medium for 24 h before the addition of 200 μM zinc or vehicle for the indicated time before fixing with PFA. Then, the van Steensel's cross-correlation coefficient of n=6 cells was analysed using ImageJ and pooled. A higher coefficient corresponds to more co-localization. Scale bar, 5 μM.

DISCUSSION

In the present study, we have attributed a novel function to the PKA-inhibiting protein PKIB. We found that PKIB binds to GPR39 and specifically augments its RhoA-mediated constitutive effects on SRE-mediated transcription. This function is mainly responsible for the cell death-inhibiting effect of GPR39 [7], which explains why PKIB co-expression also increased the cytoprotective effects of GPR39. However, PKIB by itself also had a substantial SRE-independent effect against cell death, as it also protected when co-expressed with empty vector alone, without any effect on SRE-mediated transcription. The cytoprotective function of PKIB conforms with the fact that PKIB was found to be up-regulated specifically in castration-resistant and aggressive prostate carcinomas [15]. In addition, the enhanced growth and mobility of prostate carcinoma cells transfected with PKIB [15] may well be explained by enhanced RhoA-mediated signalling that regulates the actin and adhesion dynamics that control cell migration also in prostate carcinoma [27]. PKIB was also recently found to be down-regulated in human embryonic stem cell lines carrying the mutant huntingtin, which causes Huntington's disease and leads to striatal neurodegeneration [28].

Our results demonstrate that this novel function of PKIB is clearly independent of its ability to inhibit PKA activity. PKIB is approximately 40-fold less potent than its homologues PKIα and PKIγ [20], which already leads to the hypothesis that this protein may have additional functions. Ligand-mediated activation of GPR39 by zinc increases cAMP production [13], which then causes dissociation of the regulatory subunits of PKA (also known as cAMP-dependent protein kinase) and nuclear translocation of the catalytic subunit of PKA, where it exerts its kinase activity. We demonstrate that zinc causes dissociation of PKIB from GPR39, thereby attenuating its constitutive activity. We hypothesize that, once liberated, PKIB associates with the catalytic subunit of PKA, which inhibits its activity and causes a more rapid export out of the nucleus [20]. Once cAMP levels return to normal, PKIB probably dissociates from PKA and returns to the plasma membrane to re-associate with GPR39 and enhance its constitutive ligand-independent activity. This would result in a negative-feedback loop with the ability to limit excessive activation of GPR39 by its ligand zinc (summarized in Figure 5).

Depiction of the presumed function of PKIB at the plasma membrane or the cytosol

Figure 5
Depiction of the presumed function of PKIB at the plasma membrane or the cytosol

(A) Unstimulated and (B) zinc-treated cells.

Figure 5
Depiction of the presumed function of PKIB at the plasma membrane or the cytosol

(A) Unstimulated and (B) zinc-treated cells.

According to our model, cells would be able to regulate the two functions of GPR39 by the amount of PKIB present in the cell. Theoretically, the abundance of PKIB should match that of GPR39, which, like most GPCRs, is a rather rare protein. Interacting proteins should also be expressed to some degree in the same tissue. To our knowledge, no data showing tissue expression of GPR39 at the protein level have been published. RNA-based methods of quantification are hampered by the fact that the 3′ exon of the GPR39 gene overlaps with an antisense gene called LYPD1 (Ly-6/PLAUR domain-containing 1). Taking this into account, two independent groups have demonstrated that GPR39 is mainly expressed in the gastrointestinal tract, including the liver and pancreas as well as in the kidney and adipose tissue [29,30], with no expression in the brain. In contrast, others have shown expression of GPR39 in the brain [31,32]. PKIB was first identified from testis [33] and later shown by quantitative Northern blotting to be mainly expressed in the heart, brain, placenta, liver, skeletal muscle, kidney and pancreas [34], which matches the presumed expression pattern of GPR39.

Taken together, our results assign a novel function to the PKA-inhibiting protein PKIB in the fine-tuning of constitutive against ligand-mediated signal transduction mediated by the GPCR GPR39. Future research will show whether this applies only to this peculiar receptor or whether PKIB also modulates the function of other receptors.

Abbreviations

     
  • AM

    acetoxymethyl ester

  •  
  • CHO

    Chinese-hamster ovary

  •  
  • CLIC4

    chloride intracellular channel 4

  •  
  • CRE

    cAMP-response element

  •  
  • CTB

    CellTiter-Blue

  •  
  • DMEM

    Dulbecco’s modified Eagle’s medium

  •  
  • GPCR

    G-protein-coupled receptor

  •  
  • GPR39

    GPCR 39

  •  
  • HA

    haemagglutinin

  •  
  • KO

    knockout

  •  
  • NECAB2

    N-terminal EF hand calcium-binding protein 2

  •  
  • PFA

    paraformaldehyde

  •  
  • PKA

    protein kinase A

  •  
  • PKI

    protein kinase inhibitor

  •  
  • PKIB

    PKA inhibitor β

  •  
  • RGS16

    regulator of G-protein signalling 16

  •  
  • SRE

    serum-response element

  •  
  • Y2H

    yeast-2-hybrid

AUTHOR CONTRIBUTION

Axel Methner conceived the study, made the Figures and wrote the paper. Teresa Schacht performed the co-localization studies. Konstantinos Lefkimmiatis performed the AKAR4 experiments. Zsuzsa Kovacs performed most other experiments. Ann-Kathrin Herrmann and Philipp Albrecht contributed to many experiments that did not end up in the paper and helped write the paper. All authors critically revised and approved the final paper.

We thank Jin Zhang for AKAR4 and Susanne Thomsen for technical assistance.

FUNDING

This work was supported by the Dr Kurt und Irmgard Meister-Stiftung and the Dr Robert Pfleger-Stiftung.

References

References
1
Civelli
O.
The state of GPCR research in 2004
Nat. Rev. Drug Discov.
2004
, vol. 
3
 (pg. 
575, 577
-
626
)
2
Drews
J.
Drug discovery: a historical perspective
Science
2000
, vol. 
287
 (pg. 
1960
-
1964
)
[PubMed]
3
Joost
P.
Methner
A.
Phylogenetic analysis of 277 human G-protein-coupled receptors as a tool for the prediction of orphan receptor ligands
Genome Biol.
2002
, vol. 
3
 pg. 
RESEARCH0063
 
[PubMed]
4
McKee
K. K.
Tan
C. P.
Palyha
O. C.
Liu
J.
Feighner
S. D.
Hreniuk
D. L.
Smith
R. G.
Howard
A. D.
Van der Ploeg
L. H.
Cloning and characterization of two human G protein-coupled receptor genes (GPR38 and GPR39) related to the growth hormone secretagogue and neurotensin receptors
Genomics
1998
, vol. 
46
 (pg. 
426
-
434
)
5
Tremblay
F.
Richard
A.-M. T.
Will
S.
Syed
J.
Stedman
N.
Perreault
M.
Gimeno
R. E.
Disruption of G protein-coupled receptor 39 impairs insulin secretion in vivo
Endocrinology
2009
, vol. 
150
 (pg. 
2586
-
2595
)
[PubMed]
6
Holst
B.
Egerod
K. L.
Jin
C.
Petersen
P. S.
Østergaard
M. V.
Hald
J.
Sprinkel
A. M. E.
Størling
J.
Mandrup-Poulsen
T.
Holst
J. J.
, et al. 
G protein-coupled receptor 39 deficiency is associated with pancreatic islet dysfunction
Endocrinology
2009
, vol. 
150
 (pg. 
2577
-
2585
)
[PubMed]
7
Dittmer
S.
Sahin
M.
Pantlen
A.
Saxena
A.
Toutzaris
D.
Pina
A.-L.
Geerts
A.
Golz
S.
Methner
A.
The constitutively active orphan G-protein-coupled receptor GPR39 protects from cell death by increasing secretion of pigment epithelium-derived growth factor
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
7074
-
7081
)
[PubMed]
8
Johnson
E. N.
Seasholtz
T. M.
Waheed
A. A.
Kreutz
B.
Suzuki
N.
Kozasa
T.
Jones
T. L. Z.
Brown
J. H.
Druey
K. M.
RGS16 inhibits signalling through the G alpha 13-Rho axis
Nat. Cell Biol.
2003
, vol. 
5
 (pg. 
1095
-
1103
)
[PubMed]
9
Egerod
K. L.
Jin
C.
Petersen
P. S.
Wierup
N.
Sundler
F.
Holst
B.
Schwartz
T. W.
β-Cell specific overexpression of GPR39 protects against streptozotocin-induced hyperglycemia
Int. J. Endocrinol.
2011
, vol. 
2011
 pg. 
401258
 
[PubMed]
10
Xie
F.
Liu
H.
Zhu
Y.-H.
Qin
Y.-R.
Dai
Y.
Zeng
T.
Chen
L.
Nie
C.
Tang
H.
Li
Y.
, et al. 
Overexpression of GPR39 contributes to malignant development of human esophageal squamous cell carcinoma
BMC Cancer
2011
, vol. 
11
 pg. 
86
 
[PubMed]
11
Cohen
L.
Azriel-Tamir
H.
Arotsker
N.
Sekler
I.
Hershfinkel
M.
Zinc sensing receptor signaling, mediated by GPR39, reduces butyrate-induced cell death in HT29 colonocytes via upregulation of clusterin
PLoS ONE
2012
, vol. 
7
 pg. 
e35482
 
[PubMed]
12
Holst
B.
Holliday
N. D.
Bach
A.
Elling
C. E.
Cox
H. M.
Schwartz
T. W.
Common structural basis for constitutive activity of the ghrelin receptor family
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
53806
-
53817
)
[PubMed]
13
Holst
B.
Egerod
K. L.
Schild
E.
Vickers
S. P.
Cheetham
S.
Gerlach
L.-O.
Storjohann
L.
Stidsen
C. E.
Jones
R.
Beck-Sickinger
A. G.
, et al. 
GPR39 signaling is stimulated by zinc ions but not by obestatin
Endocrinology
2007
, vol. 
148
 (pg. 
13
-
20
)
[PubMed]
14
Yasuda
S.-I.
Miyazaki
T.
Munechika
K.
Yamashita
M.
Ikeda
Y.
Kamizono
A.
Isolation of Zn2+ as an endogenous agonist of GPR39 from fetal bovine serum
J. Recept. Signal Transduct. Res.
2007
, vol. 
27
 (pg. 
235
-
246
)
[PubMed]
15
Chung
S.
Furihata
M.
Tamura
K.
Uemura
M.
Daigo
Y.
Nasu
Y.
Miki
T.
Shuin
T.
Fujioka
T.
Nakamura
Y.
, et al. 
Overexpressing PKIB in prostate cancer promotes its aggressiveness by linking between PKA and Akt pathways
Oncogene
2009
, vol. 
28
 (pg. 
2849
-
2859
)
[PubMed]
16
Mohr
K.
Koegl
M.
High-throughput yeast two-hybrid screening of complex cDNA libraries
Methods Mol. Biol.
2012
, vol. 
812
 (pg. 
89
-
102
)
[PubMed]
17
Lefkimmiatis
K.
Srikanthan
M.
Maiellaro
I.
Moyer
M. P.
Curci
S.
Hofer
A. M.
Store-operated cyclic AMP signalling mediated by STIM1
Nat. Cell Biol.
2009
, vol. 
11
 (pg. 
433
-
442
)
[PubMed]
18
Milligan
G.
White
J. H.
Protein–protein interactions at G-protein-coupled receptors
Trends Pharmacol. Sci.
2001
, vol. 
22
 (pg. 
513
-
518
)
[PubMed]
19
Canela
L.
Luján
R.
Lluís
C.
Burgueño
J.
Mallol
J.
Canela
E. I.
Franco
R.
Ciruela
F.
The neuronal Ca2+-binding protein 2 (NECAB2) interacts with the adenosine A2A receptor and modulates the cell surface expression and function of the receptor
Mol. Cell. Neurosci.
2007
, vol. 
36
 (pg. 
1
-
12
)
[PubMed]
20
Dalton
G. D.
Dewey
W. L.
Protein kinase inhibitor peptide (PKI): a family of endogenous neuropeptides that modulate neuronal cAMP-dependent protein kinase function
Neuropeptides
2006
, vol. 
40
 (pg. 
23
-
34
)
[PubMed]
21
Lewerenz
J.
Klein
M.
Methner
A.
Cooperative action of glutamate transporters and cystine/glutamate antiporter system Xc− protects from oxidative glutamate toxicity
J. Neurochem.
2006
, vol. 
98
 (pg. 
916
-
925
)
[PubMed]
22
Lewerenz
J.
Letz
J.
Methner
A.
Activation of stimulatory heterotrimeric G proteins increases glutathione and protects neuronal cells against oxidative stress
J. Neurochem.
2003
, vol. 
87
 (pg. 
522
-
531
)
[PubMed]
23
Lewerenz
J.
Albrecht
P.
Tien
M.-L. T.
Henke
N.
Karumbayaram
S.
Kornblum
H. I.
Wiedau-Pazos
M.
Schubert
D.
Maher
P.
Methner
A.
Induction of Nrf2 and xCT are involved in the action of the neuroprotective antibiotic ceftriaxone in vitro
J. Neurochem.
2009
, vol. 
111
 (pg. 
332
-
343
)
[PubMed]
24
Saadi
R. A.
He
K.
Hartnett
K. A.
Kandler
K.
Hershfinkel
M.
Aizenman
E.
SNARE-dependent upregulation of potassium chloride co-transporter 2 activity after metabotropic zinc receptor activation in rat cortical neurons in vitro
Neuroscience
2012
, vol. 
210
 (pg. 
38
-
46
)
[PubMed]
25
Depry
C.
Allen
M. D.
Zhang
J.
Visualization of PKA activity in plasma membrane microdomains
Mol. Biosyst.
2011
, vol. 
7
 (pg. 
52
-
58
)
[PubMed]
26
Bolte
S.
Cordelières
F. P.
A guided tour into subcellular colocalization analysis in light microscopy
J. Microsc.
2006
, vol. 
224
 (pg. 
213
-
232
)
[PubMed]
27
Schmidt
L. J.
Duncan
K.
Yadav
N.
Regan
K. M.
Verone
A. R.
Lohse
C. M.
Pop
E. A.
Attwood
K.
Wilding
G.
Mohler
J. L.
, et al. 
RhoA as a mediator of clinically relevant androgen action in prostate cancer cells
Mol. Endocrinol.
2012
, vol. 
26
 (pg. 
716
-
735
)
[PubMed]
28
Feyeux
M.
Bourgois-Rocha
F.
Redfern
A.
Giles
P.
Lefort
N.
Aubert
S.
Bonnefond
C.
Bugi
A.
Ruiz
M.
Deglon
N.
, et al. 
Early transcriptional changes linked to naturally occurring Huntington's disease mutations in neural derivatives of human embryonic stem cells
Hum. Mol. Genet.
2012
, vol. 
21
 (pg. 
3883
-
3895
)
[PubMed]
29
Moechars
D.
Depoortere
I.
Moreaux
B.
de Smet
B.
Goris
I.
Hoskens
L.
Daneels
G.
Kass
S.
Ver Donck
L.
Peeters
T.
, et al. 
Altered gastrointestinal and metabolic function in the GPR39-obestatin receptor-knockout mouse
Gastroenterology
2006
, vol. 
131
 (pg. 
1131
-
1141
)
[PubMed]
30
Egerod
K. L.
Holst
B.
Petersen
P. S.
Hansen
J. B.
Mulder
J.
Hökfelt
T.
Schwartz
T. W.
GPR39 splice variants versus antisense gene LYPD1: expression and regulation in gastrointestinal tract, endocrine pancreas, liver, and white adipose tissue
Mol. Endocrinol.
2007
, vol. 
21
 (pg. 
1685
-
1698
)
[PubMed]
31
Jackson
V. R.
Nothacker
H.-P.
Civelli
O.
GPR39 receptor expression in the mouse brain
NeuroReport
2006
, vol. 
17
 (pg. 
813
-
816
)
[PubMed]
32
Chorin
E.
Vinograd
O.
Fleidervish
I.
Gilad
D.
Herrmann
S.
Sekler
I.
Aizenman
E.
Hershfinkel
M.
Upregulation of KCC2 activity by zinc-mediated neurotransmission via the mZnR/GPR39 receptor
J. Neurosci.
2011
, vol. 
31
 (pg. 
12916
-
12926
)
[PubMed]
33
Beale
E. G.
Dedman
J. R.
Means
A. R.
Isolation and regulation of the protein kinase inhibitor and the calcium-dependent cyclic nucleotide phosphodiesterase regulator in the Sertoli cell-enriched testis
Endocrinology
1977
, vol. 
101
 (pg. 
1621
-
1634
)
[PubMed]
34
Zheng
L.
Yu
L.
Tu
Q.
Zhang
M.
He
H.
Chen
W.
Gao
J.
Yu
J.
Wu
Q.
Zhao
S.
Cloning and mapping of human PKIB and PKIG, and comparison of tissue expression patterns of three members of the protein kinase inhibitor family, including PKIA
Biochem. J.
2000
, vol. 
349
 (pg. 
403
-
407
)
[PubMed]