A2BAR (A2B adenosine receptor) has been implicated in several physiological conditions, such as allergic or inflammatory disorders, vasodilation, cell growth and epithelial electrolyte secretion. For mediating the protein–protein interactions of A2BAR, the receptor's C-terminus is recognized to be crucial. In the present study, we unexpectedly found that two point mutations in the A2BAR C-terminus (F297A and R298A) drastically impaired the expression of A2BAR protein by accelerating its degradation. Thus we tested the hypothesis that these two point mutations disrupt A2BAR's interaction with a protein essential for A2BAR stability. Our results show that both mutations disrupted the interaction of A2BAR with actinin-1, an actin-associated protein. Furthermore, actinin-1 binding stabilized the global and cell-surface expression of A2BAR. By contrast, actinin-4, another non-muscle actinin isoform, did not bind to A2BAR. Thus our findings reveal a previously unidentified regulatory mechanism of A2BAR abundance.
Extracellular adenosine is a ubiquitous signalling molecule that modulates a wide array of biological and pathological processes [1–3], and it is derived primarily from the metabolism of ATP that is released by cells [4–6]. This release of ATP is triggered by diverse stimuli, including mechanical stress, osmotic challenge, inflammation and tissue damage [7–10]. Extracellular adenosine's biological functions are mediated through its binding to the four subtypes of its cell-surface receptors, i.e. A1, A2A, A2B and A3, each of which exhibits a unique pharmacological profile, tissue distribution, and effector coupling. Adenosine receptors are typical GPCRs (G-protein-coupled receptors) that transmit signals through adenylate cyclase/cAMP and/or PLC (phospholipase C)/Ca2+ pathways. The four adenosine receptors have been cloned from several mammalian species, including humans, and whereas A1, A2A and A2B receptors exhibit extensive sequence similarity between species, A3 receptors show comparatively higher sequence variability .
A2BAR (A2B adenosine receptor) functions as a low-affinity adenosine receptor in several cases [12–14] and is expressed in immune cells, endothelial cells, aortic vascular smooth muscle, intestine and urinary bladder [15,16]. A2BAR has been implicated in mast cell activation and allergic or inflammatory disorders, vasodilation, regulation of cell growth, epithelial electrolyte secretion and modulation of neurosecretion [16–23]. Investigation of A2BAR-interacting proteins holds the key to deciphering A2BAR's function and regulation, and therefore numerous studies have been conducted to identify the receptor's protein-binding partners. Sitaraman et al.  showed that A2BAR associates with ezrin, PKA (protein kinase A) and NHERF-2 (sodium/hydrogen exchanger regulatory factor-2, also known as E3KARP) and thus forms a multiprotein complex for cAMP/PKA modulation of Cl− secretion . A2BAR was also reported to interact with β-arrestin  and SNARE (soluble N-ethylmaleimide-sensitive factor-attachment protein receptor) proteins , which regulated the trafficking of A2BAR. Recently, we found that A2BAR inhibits NF-κB (nuclear factor κB) activation by physically interacting with NF-κB1/p105 and thereby blocking its polyubiquitination and degradation .
Actinins, or α-actinins , represent a class of actin filament-cross-linking proteins that are expressed ubiquitously and play crucial roles in an array of biological processes. In vertebrates, four actinin isoforms have been identified: actinin-1 and -4, which are expressed in almost all cell and tissue types; and actinin-2 and -3, which are primarily expressed in muscle [28,29]. In addition to cross-linking two actin filaments, actinins link other membrane proteins such as receptors and cell adhesion proteins to actin filaments and regulate the functions of these membrane proteins, including their endocytosis and exocytosis [28,29]. In the present paper, we report that actinin-1 binds to the A2BAR C-terminus and thereby stabilizes the cell-surface expression of A2BAR.
MATERIALS AND METHODS
Reagents and plasmids
The following reagents were from commercial sources: all cell culture supplies were from Invitrogen; glutathione–Sepharose 4B matrix was from GE Healthcare; QuikChange® Site-Directed Mutagenesis Kit was from Stratagene; monoclonal anti-actinin antibody clone BM-75.2 (catalogue number A5044), which recognizes both actinin-1 and -4 was from Sigma; polyclonal anti-actinin-1 antibody (catalogue number A0761) was from US Biological; anti-β-actin (catalogue number sc-47778) and anti-GFP (catalogue number sc-9996) antibodies were from Santa Cruz Biotechnology; anti-V5 antibody (catalogue number R960-25) was from Invitrogen; anti-GST antibody (catalogue number 27-4577-01) was from GE Healthcare; horseradish-peroxidase-conjugated donkey anti-mouse and anti-rabbit secondary antibodies were from Jackson Laboratory; and reagents for SDS/PAGE and PVDF membranes were from Bio-Rad Laboratories. All other reagents were from Sigma. Supplementary Figure S1(A) shows the specificity of anti-actinin antibodies used, as assessed by Western blotting.
The expression vectors pcDNA4-A2BAR-V5 and pcDNA4-A2BAR-Myc were generated by subcloning A2BAR cDNA from the vector pRK5-A2BAR (kindly provided by Dr S.V. Sitaraman, Emory University, Atlanta, GA, U.S.A.). To generate GST-fusion proteins, we used PCR to amplify the C-terminus of A2BAR and incorporate appropriate restriction enzyme sites, and then ligated the fragment into pGEX6p-1 vector (GE Healthcare). The expression vectors of human actinin-1–EGFP (exon19a splice variant) and HA (haemagglutinin)-tagged human actinin-4 (exons 19a and 8a splice variant, HA–actinin-4) were kindly provided by Dr Marc D. Basson (Michigan State University, East Lansing, MI, U.S.A.) and Dr Ben Margolis (University of Michigan, Ann Arbor, MI, U.S.A.) respectively. All constructs were verified by DNA sequencing.
Cell culture and transfection
HEK (human embryonic kidney)-293T and COS7 cells were maintained in DMEM (Dulbecco's modified Eagle's medium) supplemented with 10% (v/v) FBS and 100 units/ml penicillin/streptomycin under 5% CO2 at 37°C. All transfections were performed using the Lipofectamine™ 2000 kit (Invitrogen), following the manufacturer's instructions. Experiments were performed at 36–48 h after transfection.
Mutagenesis of A2BAR
Single amino acid mutations in residues 293–302 of A2BAR–V5 were generated using the QuikChange® mutagenesis kit according to Stratagene's protocol. All ten amino acids were individually mutated to alanine, and all mutations were confirmed by DNA sequencing.
GST pull-down assays
GST and GST-fusion proteins were produced in Escherichia coli strain BL21 and then purified using the glutathione–Sepharose 4B matrix (according to the manufacturer's instructions). Purified proteins were resolved by SDS/PAGE to verify their size and purity.
In pull-down assays, GST or GST-fusion proteins were bound to glutathione–Sepharose and then cell lysates were added to the beads and incubated overnight at 4°C. Subsequently, the beads were washed three times with 3 ml of PBS containing 0.2% Triton X-100, and then the bound proteins were eluted and analysed by Western blotting.
Immunoprecipitation and Western blotting
We mixed 100–200 μg of whole-cell extracts with a 25 μl bed volume of Protein A–Sepharose beads (Santa Cruz Biotechnology) and 1 μg of various antibodies, and the total volume for incubation was then adjusted to 400 μl with a co-IP (co-immunoprecipitation) lysis buffer (20 mM Hepes, 175 mM NaCl, 0.25% NP-40, 10% glycerol, 1 mM EDTA, 1 mM DTT and 1 mM PMSF, pH 7.6 adjusted with NaOH) supplemented with 1× Complete Protease Inhibitor Cocktail (Roche). After incubation overnight at 4°C, the beads were spun down and the supernatants were discarded, and then the beads were washed three times with the co-IP lysis buffer (700 μl) and the immunocomplexes were collected for Western blot analysis, as described previously .
Non-radioactive pulse–chase assays were performed using the Click-iT labelling and Click-iT detection kits (Life Technologies), according to the manufacturer's instructions. Briefly, at 24 h after transfection, HEK-293T cells grown in six-well culture dishes were washed with PBS and then incubated with methionine-free DMEM for 30 min at 37°C. Next, AHA (L-azidohomoalanine) was added to the culture medium to a final concentration of 50 μM, and, after incubation for 1 h at 37°C, the cells were washed three times with PBS and then incubated in MEM (minimal essential medium) (containing 10% FBS) for various chase times. The cells were then lysed with RIPA buffer (50 mM Tris/HCl, pH 8.0, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate and 0.1% SDS) supplemented with 1× Complete Protease Inhibitor Cocktail , and the supernatants were collected and transferred into another Eppendorf tube for detection of labelled proteins. We mixed 60 μl of AHA-labelled proteins (from a single well of cells) with 100 μl of Click-iT reaction buffer Component A containing the biotin–alkyne and then 10 μl of Components C and D was added, and after incubating the samples with end-over-end rotation for 20 min at room temperature, we added 600 μl of methanol, 150 μl of chloroform and 400 μl of Milli-Q water. Next, the samples were centrifuged at 16000 g for 5 min and the upper aqueous phase was removed, and then 450 μl of methanol was added to the lower and middle phase. After brief vortex-mixing, proteins were precipitated through centrifugation at 16000 g for 5 min, and the pellets were washed with methanol, precipitated again and dissolved using 90 μl of RIPA buffer. We combined three samples of dissolved pellets (each from one well of cells in a six-well culture dish) and mixed the proteins with 60 μl of NeutrAvidin slurry and incubated the mixtures at 4°C overnight. Lastly, the beads were washed three times with PBS containing 0.5% Triton X-100 and 0.1% PMSF, and the captured proteins were eluted by mixing the beads with 80 μl of fresh SDS/PAGE loading buffer (plus 50 μl of 50 mM DTT) and incubating with end-over-end rotation for 2 h at room temperature. The eluted proteins were examined through Western blotting.
Confocal microscopy analysis
At 12 h after seeding on to coverslips, HEK-293T cells were transfected with GFP-tagged actinin-1 and CFP-tagged A2BAR or CFP mock vectors , and were examined 36 h later using a Zeiss LSM7 DUO confocal microscope (excitation and emission wavelengths: CFP, 435 and 485 nm; GFP, 488 and 510 nm).
At 36 h after transfection, HEK-293T cells were stimulated with 10 μM NECA (5′-N-ethylcarboxamidoadenosine) or 2 μM BAY 60-6583 for 8 min at 37°C. The cells were lysed with 0.1 M HCl, and the cAMP concentration in the lysates was measured using a cAMP immunoassay kit (Biomedical Technologies).
RNA oligonucleotides were synthesized by Dharmacon; the actinin-1 siRNA (5′-CACAGAUCGAGAACAUCGAAG-3′), actinin-4 siRNA (5′-GGGAGAAGCAGCAGCGCAA-3′), and non-targeting siRNA (catalogue number D-001210-01-05) were inserted into pSuper vector (Oligoengine). HEK-293T cells grown on plates to 70–80% confluence were transfected with the actinin-1, actinin-4 or non-targeting siRNA using Lipofectamine™ 2000 in FBS-free DMEM, and after 6 h, the cells were transferred to normal culture medium and grown for 30 h before collection.
All data are expressed as means±S.E.M. The methods of statistical analysis are indicated in the Figure legends.
F297A and R298A mutations of A2BAR markedly suppress its protein expression
Previously, we determined that the A2BAR C-terminal region from residues Arg293 to His302 was necessary for binding to NF-κB1/p105 , and therefore we mutated each residue in this region to alanine in an attempt to map the precise binding site of NF-κB1/p105. WT (wild-type) and mutant A2BARs were tagged with the V5 tag to facilitate their detection, because A2BAR is expressed at low levels  and no reliable anti-A2BAR antibody is commercially available . Unexpectedly, we found that the expression level of the F297A and R298A mutant proteins was drastically reduced or even eliminated, but that, in marked contrast, the expression of the other eight mutants was unchanged relative to control (Figure 1A). In agreement with these Western blot data, the F297A and R298A mutations substantially lowered cAMP generation following A2BAR stimulation with the agonist NECA (Figure 1B). Collectively, these results indicated that the mutations F297A and R298A reduce A2BAR protein expression in the cell.
F297A and R298A mutations lower the protein expression of A2BAR
One possibility raised by the aforementioned results was that the two mutations increase A2BAR degradation. To test this, we assessed the biosynthesis and degradation of the two mutant receptors by performing non-radioactive pulse–chase assays. In line with our hypothesis, the biosynthesis of F297A and R298A mutants was normal, but their half-lives (<2 h) were considerably shorter than that of WT A2BAR (>4 h) (Figure 2A); these results indicated that both mutations substantially accelerate A2BAR degradation.
F297A and R298A mutations increase A2BAR degradation
Next, to ascertain whether the two mutant receptors undergo proteasomal or lysosomal degradation, we tested the effects of proteasomal and lysosomal inhibitors on the expression of the mutants. Treatment with the proteasomal inhibitor MG132 increased the expression of F297A and R298A mutants 2–4-fold but only modestly affected the expression of WT and D296A-A2BAR (Figures 2B and 2C), which suggested that the F297A and R298A mutations augment proteasome-mediated degradation of WT A2BAR. By contrast, treatment with the lysosomal inhibitor chloroquine produced no effect on the expression of either F297A or R298A mutant (Figure 2D); the treatment, however, was effective in a positive-control experiment (results not shown).
Actinin-1, but not actinin-4, physically interacts with A2BAR
A previous study suggested that the C-terminus (amino acids 293–321) of A2AAR (A2A adenosine receptor), an adenosine receptor isoform that is highly homologous with A2BAR in primary sequence, binds to actinins, although the actinin isoform involved was not clearly identified . Intriguingly, the N-terminal half (amino acids 293–310) of this region (amino acids 293–321) of A2AAR shares 56% identity and 72% similarity with amino acids 295–312 of A2BAR, the sequence stretch that contains Phe297 and Arg298 (Figure 3A). Therefore we hypothesized that Phe297 and Arg298 participate in actinin binding and that actinin binding is crucial for the stability of A2BAR.
Actinin-1, but not actinin-4, physically interacts with A2BAR C-terminus
We first tested whether Phe297 and Arg298 are involved in actinin binding. GST–A2BAR-C (GST fused with C-terminal amino acids 293–332 of A2BAR) but not GST alone co-precipitated endogenous actinin from HEK-293T cell extracts (Figure 3B), which suggested that A2BAR-C interacts with actinins. Actinin was detected using an anti-actinin antibody that recognizes both actinin-1 and -4, which are non-muscle actinin isoforms. More importantly, the interaction was eliminated by the F297A and R298A mutations, which indicated that Phe297 and Arg298 are involved in actinin binding (Figure 3B).
To test specifically whether actinin-1 or actinin-4 interacts with A2BAR-C, we used GST–A2BAR-C to pull down actinin-1 or acinin-4 ectopically expressed in HEK-293T cells. Actinin-1–GFP but not HA–actinin-4 was pulled down by GST–A2BAR-C but not GST alone, and this interaction was independent of GFP (Figure 3C); these results suggest that actinin-1, but not actinin-4, interacts with A2BAR-C. The interaction between actinin-1 and A2BAR was confirmed further using co-IP assays. Agreeing with the pull-down result, actinin-1–GFP but not HA–actinin-4 or GFP co-precipitated with Myc-tagged A2BAR (Figure 3D); and the interaction of actinin-1–GFP and A2BAR–Myc was not affected by 8 min of pre-treatment with an A2BAR agonist (2 μM BAY 60-6583) (Supplementary Figure S1B). Furthermore, we did not see any changes in total A2BAR expression upon agonist stimulation (Supplementary Figure S1C), which argues against a change in the interaction of A2BAR with actinin-1 because actinin-1 interaction modulates A2BAR expression (see below).
To examine the co-localization of A2BAR and actinin-1 in live cells, HEK-293T cells were co-transfected with actinin-1–GFP and either CFP-tagged A2BAR or CFP alone. Actinin-1–GFP co-localized with CFP–A2BAR at the cell surface (Figure 3E), implying that actinin-1 and A2BAR can interact at or near the cell surface.
Actinin-1 increases the protein expression of A2BAR
Next, to examine whether actinin-1 binding is crucial for A2BAR stability, HEK-293T cells were co-transfected with A2BAR and actinin-1, actinin-4 or control vectors. Whereas actinin-1 expression increased A2BAR protein expression substantially, actinin-4 expression exerted almost no effect in this regard (Figure 4A). In agreement with these results, cAMP generation by A2BAR was significantly increased by actinin-1, but not actinin-4 (Figure 4B).
Actinin-1 increases the total and cell-surface expression of A2BAR
In a complementary set of experiments, we assessed the effect of knocking down actinin-1 or actinin-4 on the activity of endogenous A2BAR. As mentioned earlier, no reliable anti-A2BAR antibody is currently available for detecting endogenous A2BAR expression by means of Western blotting ; thus, as an alternative indicator of endogenous A2BAR expression, we examined cAMP generation stimulated by BAY 60-6583, a highly selective A2BAR agonist. Our results showed that knockdown of actinin-1, but not actinin-4, halved the cAMP production induced by BAY 60-6583 (Figure 4C). Considering all our other data described above, these later results suggest that actinin-1 stabilizes endogenous A2BAR expression. We recognize, however, that actinin-1 manipulation could also indirectly affect A2BAR activity via a potential effect on the level or activity of one of the several isoforms of adenylate cyclase. This, however, could constitute a new line of investigation.
Lastly, we examined the effect of actinin-1 on the cell-surface expression of A2BAR, which is more biologically relevant than the receptor's total protein level. The results of surface biotinylation experiments showed that the expression of actinin-1, but not actinin-4, substantially increased the cell-surface levels of A2BAR (Figure 4D). By contrast, actinin-1 did not affect the cell-surface expression of unrelated membrane proteins such as transferrin receptors.
Mounting evidence suggests that actinins interact with numerous membrane proteins, and further that the interactions with actinins regulate the trafficking of these membrane proteins and thereby affect their global and cell-surface expression [28,33]. For instance, actinin-2 interacts physically with SK2 channels and enhances their cell-surface expression, possibly by inhibiting the channel's internalization and/or promoting its recycling from recycling endosomes . Notably, the C-terminus of A2AAR was also reported to interact with actinins, although the exact actinin isoform involved was not determined ; interestingly, the authors suggested that the interaction with actinins may promote A2AAR internalization, because deletion of the C-terminus of A2AAR and disruption of actin filaments by using cytochalasin D impaired the internalization of A2AAR . The results of the present study show that the interaction of A2BAR with actinin-1 increases the cell-surface expression of A2BAR; this suggests that actinin-1 either reduces the internalization or increases the recycling of A2BAR, which is similar to the effect of actinin-2 on SK2 channels  or the effect of cytoskeletal proteins on other membrane proteins . Moreover, our study showed that Phe297 and Arg298 in A2BAR are essential for the receptor's interaction with actinin-1, thus it would be of interest to test whether these two amino acid residues, which are conserved in A2AAR, are also required for the interaction of A2AAR with actinins, particularly actinin-1, and whether the interaction with actinin-1 also increases the global and cell-surface expression of A2AAR.
An early study suggested that A2AAR interacts with A2BAR and forms a hetero-oligomer and that this interaction enhances the cell-surface expression of A2BAR . Because both A2AAR and A2BAR interact with actinins, the homodimer of actinin-1 or the heterodimer of actinin-1 with another actinin isoform  might mediate the dimerization of A2AAR and A2BAR and thereby promote the surface expression of A2BAR. This mechanism is clearly not mutually exclusive with the mechanism by which actinin-1 mediates the interaction of A2BAR with actin filaments and thereby modulates the trafficking and cell-surface expression of A2BAR. The ability of actinins to play multiple roles related to a single protein has been reported previously: actinin-2 not only interacts with SK channels and modulates their surface expression , but also couples SK channels with L-type Ca2+ channels to form a functional complex .
In the present study, the actinin-1 isoform used for ectopic expression was the exon19a splice variant (see the Materials and methods section), which is a Ca2+-sensitive isoform: when the intracellular Ca2+ level is high, Ca2+ binds to the calmodulin-like domain in this variant and thereby weakens the cross-link between actinin-1 and actin filaments . Thus, if this actinin isoform interacts with A2BAR in vivo and its up-regulation of A2BAR expression requires cross-linking with actin filaments, A2BAR expression would be attenuated by an elevation of intracellular Ca2+. This could form a negative-feedback loop for regulating A2BAR activation, because the activation of A2BAR might increase intracellular PLC/Ca2+ signalling . This negative-feedback regulation could provide another mechanism for A2BAR desensitization besides arrestin-mediated receptor desensitization .
Notably, we found that actinin-1, but not actinin-4, interacts with A2BAR, although human actinin-4 shares 82% identity with actinin-1. Previous studies have suggested that the cellular localization and function of actinin-1 and actinin-4 are not redundant. For example, actinin-4, but not actinin-1, appears to be required for the normal formation and function of glomeruli in the kidney , and actinin-4 is the predominant actinin isoform reported to be associated with cancer . Moreover, actinin-4 has been widely suggested to be capable of translocating to the nucleus and functioning there as a transcriptional regulator . Intriguingly, actinin-4 binds to NF-κB subunits p65 and p50 and acts as a co-activator for NF-κB transcription factors . It will therefore be of interest to ascertain whether actinin-1 also binds to NF-κB proteins, including NF-κB1/p105, and how this binding affects the physical and functional interactions between NF-κB1/p105 and A2BAR .
Ying Sun, Wenbao Hu and Pingbo Huang conceived the study and wrote the paper. Pingbo Huang directed the study. Ying Sun and Wenbao Hu performed and analysed most of the work. Xiaojie Yu contributed partially to the data in Figure 3 and Supplementary Figure S1, and writing the paper. Zhengzhao Liu performed the cell-surface biotinylation experiment. Robert Tarran and Katya Ravid provided reagents, and participated in conceptual discussion of experiments and data, and in the final writing of the paper.
We thank Fengqiang Sun for generating A2BAR mutants and contributing to the initial experiments, and Ka-lun So for technical assistance.
This work was supported by the Hong Kong Research Grants Council [grant numbers GRF660913 and GRF16102415 (to P.H.)], the Shenzhen Innovation Committee of Science and Technology, China [grant number JCY20130401144532136 (to Y.S.)] and the National Heart, Lung, and Blood Institute (NHLBI) [grant number HL93149 (to K.R.)].
These authors contributed equally to this work.