Interaction of neuronal nitric oxide synthase with alpha₁-adrenergic receptor subtypes in transfected HEK-293 cells

Abstract

Background

The C-terminal four amino acids (GEEV) of human α_1A-adrenergic receptors (ARs) have been reported to interact with the PDZ domain of neuronal nitric oxide synthase (nNOS) in a yeast two-hybrid system. The other two α₁-AR subtypes have no sequence homology in this region, raising the possibility of subtype-specific protein-protein interactions.

Results

We used co-immunoprecipitation and functional approaches with epitope-tagged α₁-ARs to examine this interaction and the importance of the C-terminal tail. Following co-transfection of HEK-293 cells with hexahistidine/Flag (HF)-tagged α_1A-ARs and nNOS, membranes were solubilized and immunoprecipitated with anti-FLAG affinity resin or anti-nNOS antibodies. Immunoprecipitation of HFα_1A-ARs resulted in co-immunoprecipitation of nNOS and vice versa, confirming that these proteins interact. However, nNOS also co-immunoprecipitated with HFα_1B- and HFα_1D-ARs, suggesting that the interaction is not specific to the α_1A subtype. In addition, nNOS co-immunoprecipitated with each of the three HFα₁-AR subtypes which had been C-terminally truncated, suggesting that this interaction does not require the C-tails; and with Flag-tagged β₁- and β₂-ARs. Treatment of PC12 cells expressing HFα_1A-ARs with an inhibitor of nitric oxide formation did not alter norepinephrine-mediated activation of mitogen activated protein kinases, suggesting nNOS is not involved in this response.

Conclusions

These results show that nNOS does interact with full-length α_1A-ARs, but that this interaction is not subtype-specific and does not require the C-terminal tail, raising questions about its functional significance.

Background

α₁-Adrenergic receptors (ARs) are G protein-coupled receptors that mediate some of the actions of norepinephrine and epinephrine. Three human α₁-AR subtypes have been cloned and named α_1A, α_1B and α_1D-ARs[1]. These receptors regulate several important central and peripheral processes, such as neuronal excitability, vascular and nonvascular smooth muscle contraction, and cellular growth and differentiation. The three α₁-AR subtypes are structurally and pharmacologically distinct, but all couple through G_q/11 to cause activation of apparently similar intracellular signaling pathways.

The last four amino acids of the intracellular C-tail of the α_1A-AR, GEEV, matches the motif G(D/E)XV shown previously to interact with the class III PDZ domain of neuronal nitric oxide synthase (nNOS). Experiments using the yeast two-hybrid system showed previously that a protein corresponding to the last 114 amino acids of the rat α_1A-AR (previously referred to as α_1C-AR) interacted strongly with the PDZ domain of nNOS[2]. Since the corresponding amino acids at the C-terminus of α_1B (PGQF) and α_1D-ARs (ETDI) would not be predicted to interact with this PDZ domain, an interaction between α_1A-ARs and nNOS could represent an interaction unique to this subtype.

PDZ domains are protein-binding modules involved in assembly of signaling complexes and subcellular protein targeting[3]. For example, NMDA receptors in cultured cortical neurons associate with nNOS through PSD-95, a protein containing three PDZ domains[4]. Consequently, NMDA receptor activation increases nitric oxide production and neurotoxicity; while suppression of PSD-95 expression inhibits these responses. These results suggest that the PDZ domains of PSD-95 may facilitate the assembly of signaling complexes involving both NMDA receptors and nNOS, and the increases in intracellular Ca2+ caused by NMDA receptor activation may facilitate nNOS activation.

Since α_1A-AR activation also increases intracellular Ca²⁺, we studied the interaction between this receptor and nNOS. We wanted to determine whether full-length α_1A-ARs interact with full-length nNOS, whether the interaction is subtype-specific, and whether it involves the GEEV motif in the C-terminal tail. We co-expressed epitope-tagged full length or C-terminally truncated α₁-ARs with nNOS in HEK-293 cells and examined the ability of anti-Flag and anti-nNOS antibodies to immunoprecipitate both proteins. We found that nNOS does interact with full-length α_1A-ARs, but that it also interacts with other α₁-AR subtypes and β-ARs. In addition, the interaction does not require the C-terminal tail, confirming that it is not specific to the GEEV motif.

Results

Co-immunoprecipitation of nNOS with HFα_1A-ARs

To study the interaction between α_1A-ARs and nNOS, HEK-293 cells were transfected with rat nNOS and selected with geneticin (400 μg/ml). Western blots using an anti-nNOS antibody showed a strong immunoreactive band of ~170 kDa corresponding to nNOS in stably transfected cells as expected, but little or no signal in untransfected cells (data not shown). Expression of nNOS was similar to that observed with equal amounts of rat brain membrane protein run in parallel, suggesting similar expression levels. HEK-293 cells stably transfected with nNOS were co-transfected with the cDNA encoding HFα_1A-ARs. Expression levels of transiently transfected α₁-ARs in these cells ranged from 100–500 fmol/mg protein, also similar to levels observed in rat brain. Cells were then solubilized, immunoprecipitated with anti-Flag M2 affinity resin, eluted, and blotted with anti-Flag (Fig. 1A) or anti-nNOS antibodies (Fig. 1B). Western blots of anti-Flag immunoprecipitates showed that HFα_1A-ARs migrated as monomers of ~50 kDa (Fig. 1), and also appeared as dimers and trimers, as reported previously[5]. Immunoprecipitation of HFα_1A-ARs with anti-Flag affinity resin resulted in co-immunoprecipitation of nNOS, as revealed by the 170 kDa band detected in immunoblots using anti-nNOS antibody (Fig. 1B). Note that nNOS immunoreactivity was not present in anti-Flag affinity resin immunoprecipitates from solubilized HEK-293 cells not transfected with HFα_1A-ARs (Fig. 1B), showing that co-immunoprecipitation of nNOS requires presence of the tagged receptor construct.

Figure 1

Immunoprecipitation with anti-Flag M2 affinity resin of solubilized HEK-293 cells co-expressing nNOS and HF-tagged α_1A-AR (Flag-A/nNOS) or HEK-293 expressing only nNOS (HEK/nNOS). Cells were solubilized, immunoprecipitated with anti-Flag M2 affinity resin, eluted with Flag peptide, run on SDS-PAGE, and transferred to nitrocellulose as described in Methods. A: Western blot with M2 anti-Flag antibody with arrows showing monomers, dimers and trimers; B: Western blot with anti-nNOS antibody. Blots are representative of at least two other experiments with similar results. IP = immunoprecipitation; IB = immunoblot.

Co-immunoprecipitation of nNOS with HFα_1B- and HFα_1D-ARs

To determine whether the interaction between nNOS and α_1A-ARs was subtype-specific, cells stably expressing nNOS were co-transfected with HFα_1B or HFα_1D-ARs, solubilized, immunoprecipitated with anti-Flag affinity resin, eluted, and blotted with anti-Flag (Fig. 2A) or anti-nNOS antibodies (Fig. 2B). Monomers of the HFα_1B and HFα_1D-ARs migrated as bands of ~65 and ~75 kDa (Fig. 2A), and dimers and trimers were also detected. Surprisingly, nNOS was also co-immunoprecipitated from cells co-expressing nNOS and HFα_1B or HFα_1D-ARs (Fig. 2B), although the GEEV motif predicted to interact with nNOS is not present in either of these subtypes.

Figure 2

Immunoprecipitation with anti-Flag M2 affinity resin of solubilized HEK-293 cells co-expressing nNOS and HF-tagged α_1B- (Flag-B/nNOS) or α_1D-ARs (Flag-D/nNOS) or HEK-293 cells expressing only the HF-tagged α₁-AR subtypes (Flag-B, Flag-D). A: Western blot with anti-Flag antibody; arrows indicate monomers, dimers and trimers of the HF-tagged α_1B- (closed) and α_1D- (open) ARs. B: Western blot with anti-nNOS antibody. Blots shown are representative of at least two other experiments with similar results. IP = immunoprecipitation; IB = immunoblot.

Interaction of C-terminally truncated HFα₁-ARs with nNOS

The role of the C-terminus of α₁-ARs in the interaction with nNOS was examined by truncation. Stop codons were introduced approximately 20 amino acids after the predicted 7th transmembrane domain, at a conserved glutamine (α_1A, Gln 344; α_1B, Gln 366) or an adjacent arginine (α_1D, Arg 418). Truncated HFα₁-ARs were transfected into HEK-293 cells stably expressing nNOS, and cells were solubilized, immunoprecipitated with anti-Flag M2 affinity resin, and blotted with anti-Flag (Fig. 3A) or anti-nNOS antibodies (Fig. 3B). The monomeric truncated receptors migrated with molecular masses ~25% lower than that of the full length receptors (Fig. 3A), and higher order oligomers were also apparent as observed with full-length receptors. Specific immunoreactivity to anti-nNOS antibody was also detected in these immunoprecipitates (Fig. 3B), showing that the C-terminal cytoplasmic tail of the HFα₁-ARs is not required for interaction.

Figure 3

Immunoprecipitation with anti-Flag M2 affinity resin of solubilized HEK-293 cells expressing C-terminally truncated HF-tagged α₁-AR subtypes (Flag-trA, Flag-trB and Flag-trD) and nNOS. A: Western blot with anti-Flag M2 antibody; arrows indicate positions of monomers and dimers. B: Western blot with anti-nNOS antibody; arrow indicates nNOS. Blots shown are representative of at least two other experiments with similar results. IP = immunoprecipitation; IB = immunoblot.

Co-immunoprecipitation of HF-tagged α₁-ARs with anti-nNOS antibody

We also examined the ability of anti-nNOS antibodies to co-immunoprecipitate full length and truncated HFα₁-ARs. HEK-293 cells expressing nNOS were transfected with each receptor construct and harvested after 48–72 hr. Samples were solubilized, incubated with anti-nNOS antibody, immunoprecipitated with Protein A agarose, and blotted with anti-Flag antibody. This procedure resulted in a strong non-specific band migrating at ~50 kDa (approximately the size of the HFα_1A-AR), probably representing IgG heavy chains. For comparison, parallel samples were immunoprecitated with anti-Flag affinity resin and loaded on the same gel. Fig. 4A shows that anti-nNOS antibodies caused co-immunoprecipitation of all three full length and truncated HFα₁-ARs. Note that neither protein A agarose alone, nor anti-nNOS antibody plus Protein A agarose, caused immunoprecipitation of HFα₁-ARs in cells not expressing nNOS (data not shown). This indicates that HFα₁-ARs do not nonspecifically interact with anti-nNOS antibody and/or protein A agarose.

Figure 4

Immunoprecipitation with anti-nNOS antibody of solubilized HEK-293 cells co-expressing nNOS and HF-tagged full length (Flag-A, Flag-B, Flag-D) or C-terminally truncated (Flag-trA, Flag-trB, FlagtrD) α₁-AR subtypes. For comparison, immunoprecipitation of HF-tagged receptors with anti-Flag M2 affinity resin is also shown. Monomers of each full-length and truncated subtype are indicated by arrows on the left of each blot. Blots shown are representative of at least two other experiments with similar results. IP = immunoprecipitation; IB = immunoblot.

Co-immunoprecipitation of nNOS with Flag-tagged β₁- and β₂-ARs

To further examine the specificity of this interaction, HEK-293 cells stably expressing nNOS were co-transfected with Flag-tagged β₁ or β₂-ARs, solubilized, immunoprecipitated with anti-Flag affinity resin or anti-nNOS antibody plus Protein A agarose, and blotted with anti-Flag (Fig. 5A) or anti-nNOS antibodies (Fig. 5B). Fig. 5A shows that both Flag-tagged β₁ and β₂-ARs migrated as monomers (β₁ at ~70 kDa and β₂ at ~50 kDa) as well as oligomers (data not shown), and that immunoprecipitation of nNOS caused co-immunoprecipitation of both Flag-tagged β₁ or β₂-ARs. Fig. 5B shows that nNOS was also observed following immunoprecipitation with anti-Flag M2 affinity resin in cells transfected with either Flag-tagged β₁ and β₂-ARs.

Figure 5

Immunoprecipitation with anti-nNOS antibody or anti-Flag M2 affinity resin of solubilized HEK-293 cells co-expressing nNOS and Flag-tagged β₁- or β₂-ARs. A: Western blot with anti-Flag antibody; B: Western blot with anti-nNOS antibody. Blots shown are representative of at least two other experiments with similar results. IP = immunoprecipitation; IB = immunoblot.

Effect of l-NAME on α_1A-AR-induced ERK activation in PC12 cells

It is known that nitric oxide produced by nNOS is required for PC12 cell differentiation induced by nerve growth factor (NGF) and that treatment of PC12 cells with NGF induces nNOS expression[6]. Since α_1A-AR stimulation also activates ERKS and induces differentiation in PC12 cells stably transfected with this subtype[7], we investigated the effects of l-NAME, an inhibitor of NOS, on norepinephrine-induced ERK phosphorylation in PC12 cells stably transfected with HFα_1A-ARs. Fig. 6 shows that treatment of HFα_1A-PC12 cells with high concentrations of l-NAME did not block ERK phosphorylation induced by norepinephrine, or by UTP, EGF or NGF, suggesting that nitric oxide is not required for mitogenic signals in this cell line.

Figure 6

Effect of L-NAME (500 μM) on UTP (100 μM), norepinephrine (NE, 100 μM), NGF (100 ng/ml) and EGF(100 ng/ml) induced ERK phosphorylation in PC12 cells. HFα_1A-transfected PC12 cells were serum-starved for 2 hr, pretreated with L-NAME for 30 min, and exposed to agonist for 15 min before harvesting. ERK activation was determined by Western blotting with dual phospho-specific ERK antibodies as described in Methods. Although the blot shown is overexposed, similar strength of signals were seen with NE, NGF and EGF after shorter exposures.

Discussion

We evaluated the specificity and functional importance of the reported interaction of the C-terminus of α_1A-ARs with the PDZ domain of nNOS. Previous work using a yeast two hybrid assay showed that the C-terminal 114 amino acids of rat α_1A-ARs (referred to by the previous name of α_1C-) strongly interacted with residues 1–111 of nNOS[2]. The bradykinin B2 receptor, also a G protein coupled receptor, has been shown to bind directly to the oxygenase domain of nNOS and form an inhibitory complex[8], and it has been proposed that nNOS is released and activated upon receptor stimulation. The domain of the bradykinin B2 receptor that interacts with nNOS is in the C-tail shortly after the predicted 7^th transmembrane domain, and spatially similar but structurally dissimilar domains of the rat angiotensin AT1 receptor and human endothelin-1 ETB receptors have been proposed to block endothelial NOS (eNOS) activity, possibly through a similar mechanism[9]. Therefore we wanted to determine whether there was a specific interaction between full-length α_1A-ARs and nNOS in intact cells.

Co-immunoprecipitation experiments showed that epitope-tagged α_1A-ARs do interact with full-length nNOS when expressed together in HEK-293 cells. Immunoprecipitation of HFα_1A-ARs from cells stably expressing nNOS caused co-immunoprecipitation of nNOS. Similarly, immunoprecipitation of nNOS caused co-immunoprecipitation of HFα_1A-ARs. This interaction appeared to be specific, since nNOS was not immunoprecipitated by anti-Flag affinity resin in cells not transfected with tagged receptors, and tagged receptors were not immunoprecipitated by anti-nNOS antibody in cells not expressing nNOS.

These results support an interaction between nNOS and α_1A-ARs, which could be due to the previously reported interaction of the receptor C-terminus and the PDZ domain of nNOS[2]. However, this hypothesis is weakened by the unexpected observation that nNOS also co-immunoprecipitates with both α_1B and α_1D-ARs. There is little or no homology between the C-terminal sequences of α_1A, α_1B and α_1D-ARs, and neither α_1B nor α_1D-ARs contain the GEEV motif predicted to mediate the interaction between α_1A-ARs and nNOS. However, nNOS was found to co-immunoprecipitate with both HFα_1B- and HFα_1D-ARs after co-expression in HEK-293 cells. These interactions could be observed by blotting for the tagged receptors after immunoprecipitation with anti-nNOS antibody, or by blotting for nNOS after immunoprecipitation of the tagged receptors with anti-Flag antibody. Direct comparison of α_1A, α_1B and α_1D-ARs in the same experiment showed similar degrees of interactions of all three subtypes with nNOS (data not shown), further demonstrating that this interaction is not specific to the α_1A subtype.

We examined the role of the C-terminal tail in this interaction by constructing receptors in which the C-terminus was truncated. Studies with HFα₁-AR subtypes with short (~20 aa) C-terminal tails suggested that the C-terminal tails are not required for interaction with nNOS. Following transfection into nNOS-expressing cells, immunoprecipitation of all three C-terminally truncated HFα₁-AR constructs caused co-immunoprecipitation of nNOS similar to that observed with full-length receptors. Although PDZ domains of some proteins, such as PSD-95 and syntrophin, can bind internal peptide sequences that fold as β-fingers and mimic canonical C-terminal peptides[10, 11], there is very little homology between the intracellular loops of α₁-AR subtypes, making it unlikely that there is a common internal amino acid sequence involved in interaction with nNOS.

Since we found nNOS to co-immunoprecipitate with all three α₁-AR subtypes, we also determined whether it would directly associate with β-ARs. Flag-tagged β₁ and β₂-ARs were transfected into cells stably expressing nNOS, and after solubilization and immunoprecipitation with anti-Flag M2 affinity resin we again found co-immunoprecipitation of nNOS. Since β₁ and β₂-ARs show no sequence homology to α₁-AR subtypes in their intracellular domains, this further supports the conclusion that interaction with nNOS is not localized to discrete intracellular domains.

Conclusions

Our data suggest that full-length α_1A-ARs do interact with nNOS; however this interaction is not subtype-specific since α_1B- and α_1D-ARs showed similar interactions. The interaction did not require the receptor C-terminus, and similar interactions were observed with β₁ and β₂-ARs. This data does not support a proposed specific interaction between the α_1A-AR C-terminus and the nNOS PDZ domain suggested by studies with fusion proteins. Studies on α_1A-ARs in transfected PC12 cells showed no role for nitric oxide in mitogenic signaling, also raising questions about the functional significance of this interaction.

Methods

Materials

HEK-293 cells were purchased from ATCC. PC12 cells were obtained from Cindy Miranti and Michael Greenberg (Harvard Medical School, Boston, MA, USA). The cDNA encoding rat nNOS was from Dr. Thomas Michel (Harvard Medical School, Boston, MA), the human α_1A-AR cDNA [12] from Dr. Gozoh Tsujimoto (National Children's Hospital, Tokyo, Japan), the human α_1B-AR cDNA[13] from Dr. Dianne Perez (Cleveland Clinic), and the human α_1D-AR cDNA[14] was cloned in our lab. Materials were obtained from the following sources: Dulbecco's modified Eagle's medium (DMEM); L-NAME; norepinephrine ((-)-arterenol); streptomycin, penicillin, Flag peptide, anti-Flag M2 affinity resin, HRP-conjugated anti-Flag M2 antibody and goat anti-rabbit HRP-conjugated secondary antibodies (Sigma, St Louis, MO); geneticin; n-Dodecyl-β-D-maltoside (Calbiochem); anti-nNOS rabbit polyclonal antibody, Protein A-agarose resin (Santa Cruz); rabbit polyclonal dual phospho-specific anti-ERK antibody, PNGase F (New England Biolabs); and ECL reagent (Amersham).

Cell culture

HEK-293 cells were propagated in Dulbecco's Minimal Essential Medium with sodium pyruvate, 10% heat inactivated fetal bovine serum, 100 U/l streptomycin, and 100 U/l penicillin at 37°C in a humidified atmosphere with 5% CO₂. Confluent plates were subcultured at a 1:3 ratio. PC12 cells were propagated in Dulbecco's Minimal Essential Medium containing 4.5 g/l glucose, 1.4% glutamine, 20 mM Hepes, 100 U/l streptomycin, 100 U/l penicillin, 10% donor horse serum, and 5% fetal bovine serum. For measurement of ERK phosphorylation, 35 mm dishes of PC12 cells were seeded at a density of 600,000 cells/2 ml.

Transfections

Receptor coding sequences were generated by PCR, sequenced, and subcloned into the mammalian expression plasmid pDT containing sequential N-terminal hexahistidine and FLAG (HF) epitopes as previously described[5]. HEK-293 cells (150 mm plates) were transfected with 50 μg cDNA encoding the rat isoform of nNOS by calcium phosphate precipitation, and stably transfected cells selected with geneticin (400 μg/ml). cDNAs encoding each of the HF-human α₁-ARs were transfected into parental HEK-293 cells or cells stably transfected with nNOS by calcium phosphate precipitation and cells harvested 48–72 h later. The density of HFα₁-ARs was measured by specific binding of [¹²⁵I]-HEAT[7], and ranged from 100–500 fmol/mg protein.

Immunoprecipitation

HEK-293 cells expressing HF-tagged α₁-ARs, nNOS, or both, were harvested by scraping and fractionated by repeated centrifugation and homogenization. Cell lysates (1–2 mg protein) were solubilized in 1X buffer (25 mM Hepes and 150 mM NaCl, pH 7.4) with 2% n-Dodecyl-β-D-maltoside for 90 min at 4°C in buffer A (25 mM HEPES, 150 mM NaCl, pH 7.4) supplemented with protease inhibitors (aprotinin 2 μg/ml, leupeptin 2 μg/ml, pepstatin 2 μg/ml, benzamidine 2 μg/ml, PMSF 2 mM, and EDTA 50 mM). Solubilized samples were centrifuged, the supernatant diluted 10-fold with buffer A containing protease inhibitors, and incubated with 100–200 μl anti-Flag M2 affinity resin for 90 min at 4°C with gentle rotation[5]. Alternatively, the supernatant was incubated with 5 μl of anti-nNOS rabbit polyclonal antibody (200 μg/ml) for 90 min at 4°C and then incubated with 20 μl Protein A-agarose overnight at 4°C. Immunoprecipitated material was recovered by centrifugation and washed at least 4 times with buffer A containing protease inhibitors. After washing, samples immunoprecipitated with anti-Flag affinity resin were eluted with 100 to 200 μl buffer A containing 400 μg/ml Flag peptide, while samples immunoprecipitated with anti-nNOS antibody were eluted with 40μl of 2X Laemmli loading buffer. All samples were deglycosylated after immunoprecipitation by treatment with 1 μl PNGase F for 2 h at room temperature. Aliquots of 30 μl were separated by 4–20% SDS-PAGE, transferred to nitrocellulose, and blotted with anti-Flag M2 antibodies conjugated to HRP (1:600) or anti-nNOS rabbit polyclonal antibodies (1:1,000) followed by goat anti-rabbit HRP-conjugated secondary antibodies (1:15,000). Proteins were visualized by ECL.

ERK phosphorylation in PC12 cells

Confluent PC12 cells stably transfected with HF-α_1A-ARs[15] (~1 pmol/mg of protein) were serum-starved for 2 h before use, and incubated with or without l-NAME (500 μM/30 min). Cells were then incubated with norepinephrine or other agonists for 15 min and lysed with Laemmli sample buffer. Cell lysates were centrifuged and proteins (10 μg/lane) were separated by SDS-PAGE and transferred to nitrocellulose membranes. Phosphorylation of ERKs was detected by immunoblotting using a 1:1000 dilution of rabbit polyclonal dual phospho-specific ERK antibodies with HRP-conjugated goat anti-rabbit IgG as a secondary antibody and visualized by ECL.