Journal list menu

Volume 455, Issue 3 p. 367-371
Full-length article
Free Access

Identification of a Ca2+/calmodulin-binding domain within the carboxyl-terminus of the angiotensin II (AT1A) receptor

Walter G. Thomas

Corresponding Author

Walter G. Thomas

Molecular Endocrinology Laboratory, Baker Medical Research Institute, P.O. Box 6492, St. Kilda Road Central, Melbourne, 8008, Australia

Corresponding author. Fax: (61) (3) 9521-1362Search for more papers by this author
Luisa Pipolo

Luisa Pipolo

Molecular Endocrinology Laboratory, Baker Medical Research Institute, P.O. Box 6492, St. Kilda Road Central, Melbourne, 8008, Australia

Search for more papers by this author
Hongwei Qian

Hongwei Qian

Molecular Endocrinology Laboratory, Baker Medical Research Institute, P.O. Box 6492, St. Kilda Road Central, Melbourne, 8008, Australia

Search for more papers by this author
First published: 23 July 1999
Citations: 26

Abstract

To identify regulators of the type 1A angiotensin II receptor (AT1A), we investigated the interaction of cellular proteins with a fusion protein containing the rat AT1A receptor carboxyl-terminus. An ∼20 kDa cytoplasmic protein interacted with the fusion protein in a Ca2+-dependent manner and was identified as calmodulin. A control peptide with high affinity for Ca2+/calmodulin and a peptide corresponding to a membrane proximal portion of the AT1A receptor carboxyl-terminus with analogy to known calmodulin-binding sequences were synthesised and tested for calmodulin-binding. Using in vitro binding assays combined with gel shift analysis, we demonstrated the formation of complexes between calmodulin and both peptides, which were Ca2+-dependent and of 1:1 stoichiometry. Affinity gels produced from these peptides also purified calmodulin from cell extracts. These results suggest a novel feedback regulation of the AT1A receptor by Ca2+/calmodulin and identify the membrane proximal region of the carboxyl-terminus as a focal point for interactions important for AT1A receptor function.

1 Introduction

Seven transmembrane-spanning receptors that couple to heterotrimeric guanyl nucleotide-binding proteins (G-proteins) represent a large superfamily of cell surface receptors. These receptors convert a diverse array of sensory and hormonal stimuli into intracellular signals that mediate cellular responses. The general paradigm of G-protein coupled receptor (GPCR) action is that stimulated receptors undergo a conformational change that exposes cytoplasmic domains in the receptor, which bind to and activate G-proteins. The termination of GPCR signalling is mediated by receptor phosphorylation and the recruitment to the receptor of accessory proteins termed arrestins, which prevent further G-protein coupling and direct receptors to clathrin-coated pits for internalisation (reviewed in [1-4]).

Recent reports indicate that proteins other than heterotrimeric G-proteins and arrestins can associate with GPCRs and regulate their function. For example, the Na+/H+ exchanger regulatory factor interacts with the β2-adrenergic receptor through a PDZ motif in the carboxyl-terminus of the receptor [5], endothelial nitric-oxide synthase interacts with the carboxyl-terminus of bradykinin B2 and angiotensin AT1A receptors [6], the small non-heterotrimeric G-proteins ARF and RhoA associate with M3 muscarinic and angiotensin AT1 receptors [7] and the tyrosine phosphatase SHP-1 couples to the sst2 somatostatin receptor [8]. Presumably, such associations increase the repertoire of GPCR actions.

The type 1 (AT1) angiotensin receptor is a GPCR and mediates the important cardiovascular actions of the peptide hormone, angiotensin II (AngII), including vasoconstriction and blood pressure regulation, fluid and electrolyte homeostasis, neuromodulation and cell growth [9]. Two closely related AT1 receptor subtypes (AT1A and AT1B) exist in rodents and stimulation of AT1 receptors by AngII results in Gαq/11 mediated signalling [10], including phospholipase C-β1-dependent production of diacylglycerol and inositol(1,4,5)trisphosphate (IP3) with the resultant activation of protein kinase C and release of calcium from intracellular stores. The actions of raised intracellular calcium are mediated through small acidic calcium-binding proteins, typified by calmodulin (CaM), and AngII stimulation of AT1 receptors activates Ca2+/CaM-dependent kinases and pathways [11-13]. The processes that terminate AT1 receptor signalling are also poorly understood, but may involve the regulatory carboxyl-terminal region, which is targeted for phosphorylation, directs receptor internalisation and may recruit modulatory proteins to the activated receptor [14].

To identify modulators of AT1 receptor function, the association of cellular proteins with the AT1A receptor carboxyl-terminus was examined. We report a previously unrecognised interaction between the Ca2+-sensing molecule, CaM, and the carboxyl-terminus of the AT1A receptor, which was dependent upon the membrane proximal region of the receptor tail. This region is a positively charged amphipathic α-helix and as such appears to be an archetypal Ca2+/CaM-binding domain. We propose a novel feedback regulation of the AT1A receptor by Ca2+/CaM that may represent a common mechanism for regulating GPCRs.

2 Materials and methods

2.1 Production of a maltose-binding protein (MBP)-AT1A fusion protein

A MBP fusion protein, containing amino acids 302–359 of the AT1A receptor carboxyl-terminus (MBP-AT1ACT), was made by PCR from the pRc/CMV/AT1A template [15] using sense (5′-CTGCCTGAACCCTurn:x-wiley:00145793:media:feb2s0014579399009047:feb2s0014579399009047-math-si1TACGGCTTTC-3′, bold letters indicate mutations) and antisense (5′-GCCAGTGTGCTurn:x-wiley:00145793:media:feb2s0014579399009047:feb2s0014579399009047-math-si2CTGTCACTCCAC-3′, bold letters indicate mutations) primers, incorporating unique restriction sites (5′ EcoRI and 3′ SalI, underlined). The 216 bp PCR product was digested with EcoRI and SalI and ligated into these sites of the pMal-c2 vector (New England Biolabs). To generate a construct that produces MBP only, the pMal-c2 vector was restriction-digested with EcoRI and XbaI, the overhangs were filled-in using Klenow and the plasmid was re-ligated. This generated an in-frame stop codon (TAG) in the polylinker region. These constructs were confirmed by sequencing.

Plasmid DNA for MBP-AT1ACT and MBP was transformed into the TB-1 strain of Escherichia coli and cultures at A 600 of 0.4–0.5 were induced with IPTG (0.1 mM, 3 h) to stimulate fusion protein production. Bacterial lysates were incubated with amylose resin (New England Biolabs) and the specifically bound fusion proteins eluted with 100 mM maltose. Fusion proteins were >90% pure as judged by SDS-PAGE. Affinity columns were made by applying 1 mg of purified fusion protein to 1 ml of amylose beads and incubating for 30 min at room temperature followed by extensive washing in Hank's buffered salt solution (HBSS).

2.2 Interaction of 35S-labelled cellular proteins with the MBP-AT1ACT fusion protein

Chinese hamster ovary cells (CHO-K1) cells were maintained in α-MEM containing foetal bovine serum (10%) plus antibiotics. At 70% confluence, a 100 mm plate of cells was washed once with 10 ml of DMEM devoid of L-methionine and L-cysteine and the cells were starved in 10 ml of this media for 2 h. Cells were then incubated in 4 ml of media supplemented with 400 μCi/ml Tran35S-Label (∼70% [35S]-L-methionine, ∼15% [35S]-L-cysteine, ICN pharmaceuticals) for 16 h. The cells were washed three times with 10 ml of ice-cold HBSS and lysed in 2 ml of a 50 mM Tris-HCl buffer, pH 7.5, containing 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 1 mM phenylmethylsulfonyl fluoride, 5 μg/ml leupeptin, 1 μg/ml aprotinin and 1 μg/ml pepstatin. The plates were rocked for 30 min at 4°C and cell extracts clarified by centrifugation (14 000×g for 15 min). Typically, 80–90% of [35S] label was incorporated into protein as revealed by precipitation with trichloroacetic acid.

Fusion proteins immobilised on amylose resin (∼50 μg protein on 50 μl of resin) were washed twice with 1 ml of assay buffer (100 mM Tris-HCl, pH 7.5, 200 mM NaCl) and then resuspended in 60 μl of assay buffer containing either 2 mM CaCl2 or 4 mM EGTA. 40 μl of [35S]-labelled CHO-K1 cell extract was added and the suspension was mixed and incubated for 30 min at 22°C. Following pelleting and removal of the supernatant, the affinity beads were washed four times with 500 μl of assay buffer containing either 4 mM EGTA or 2 mM CaCl2. The final pellets were resuspended in 50 μl of SDS-PAGE sample buffer, heated at 95°C for 5 min and resolved by 12% SDS-PAGE. Gels were fixed and dried and apposed to a Fuji type BAS-IIIs phosphoimaging plate for 1–3 days. The imaging plate was read in a FUJIX Bio-imager Analyser BAS1000 to visualise the radioactive proteins.

2.3 Preparation of peptide affinity media

A peptide corresponding to the putative CaM-binding domain of the AT1A receptor (CLGKKFKKYFLQLLKYIPPK-amide, referred to as 1ACT, for AT1A carboxyl-terminus peptide) and the model CaM-binding peptide (CLKLKKLLKLLKKLLKLG-amide, referred to as CBCP, for CaM-binding control peptide) were synthesised (Chiron mimotopes) with a amino-terminal cysteine to allow immobilisation on an iodoacetate-activated gel (SulfoLink Coupling gel, Pierce). 5 mg of each peptide was dissolved in 1 ml of 50 mM Tris-HCl, 5 mM EDTA, pH 8.5 and mixed with 2 ml Sulfolink Coupling gel for 45 min at room temperature. Unbound peptide was removed by extensive washing with 50 mM Tris-HCl, 5 mM EDTA, pH 8.5 and uncoupled iodoacetate groups in the gel were inactivated with a 50 mM cysteine wash. Finally, the column was washed with 1 M NaCl, equilibrated with 50 mM Tris-HCl, pH 7.5 and stored as a 50% suspension.

The method for identifying proteins that interact with the peptide affinity beads was the same as that described above for the AT1A carboxyl-terminus fusion protein. Some SDS-PAGE gels were also Western-blotted to a PVDF membrane (Polyscreen, NEN) and probed with a monoclonal antibody to CaM (Upstate Biotechnology) and visualised using Renaissance ECL (NEN). As a control, a sample (1 ng) of pure bovine CaM (Sigma) was run in parallel to confirm the immunodetection process.

2.4 In vitro binding assay and gel shift analysis

Gel shift analysis of complexes formed between CaM and 1ACT or CBCP, using urea-polyacrylamide gel electrophoresis, was performed as described by Erickson-Viitanen and Degrado [16]. Reactions (30 μl) containing 300 pmol (∼5 μg) of pure CaM and increasing amounts (0–3000 pmol) of the peptides 1ACT or CBCP, in 100 mM Tris-HCl, pH 7.5, 4 M urea and either 0.1 mM CaCl2 or 1 mM EGTA, were incubated at 22°C for 30 min. 15 μl of a 50% glycerol/0.1% bromphenol blue solution was added to each reaction and the complexes resolved in 12.5% acrylamide gels containing 4 M urea and either 1 mM CaCl2 or 1 mM EGTA. The gels were fixed and stained with Coomassie blue.

3 Results and discussion

The carboxyl-terminus of the AT1 receptor is important for receptor function. It interacts with and activates G-proteins [17, 18] and other signalling molecules [19, 20], indicating a contribution to receptor activation, while the identification of phosphorylation sites [21, 22] and internalisation motifs [23, 24] suggests a key role in receptor desensitisation. To identify possible regulators of AT1A receptor function, we investigated the interaction of [35S]-labelled cellular proteins with a purified MBP fusion protein containing the entire receptor carboxyl-terminus. As shown in Fig. 1, in the presence Ca2+, we observed the interaction of a small (∼20 kDa) cytoplasmic protein with the MBP-AT1ACT fusion protein coupled to amylose resin. This interaction was not observed with the MBP protein alone (not shown) or when Ca2+ was chelated with excess EGTA (Fig. 1). Based on its size (∼20 kDa) and the Ca2+-dependence of the interaction, we wondered whether this protein was CaM or a similar member of the small acidic EF-hand containing family of proteins that mediate the effects of intracellular Ca2+. Given that one of the primary cellular responses to activation of the AT1 receptor by AngII is the elevation of intracellular Ca2+ levels and the activation of Ca2+/CaM-dependent pathways [11-13], we wanted to determine if this interacting protein was CaM and to identify the site of interaction within the AT1A carboxyl-terminus.

figure image
Ca2+-dependent interaction of proteins with the rat AT1A carboxyl-terminus. Shown (top left) is a schematic representation of the 359 amino acid, seven transmembrane-spanning AT1A receptor. A maltose-binding fusion protein (MBP-AT1ACT) was constructed that contained the entire AT1A receptor carboxyl-terminus (from Phe300 to Glu359). This fusion protein was purified and immobilised on amylose resin for use as an affinity matrix to identify proteins that interact with the AT1A receptor tail. In the presence of Ca2+ (+), but not EGTA (−), a small ∼20 kDa [35S]-labelled protein (arrow) was found to associate with the fusion protein. This experiment was repeated three times with similar results.

Ca2+-binding to CaM leads to a conformation change that allows specific binding to sequences within target proteins, including enzymes, channels and receptors [25]. Ca2+/CaM-binding motifs in target proteins commonly involve a preponderance of basically charged residues often within regions predicted to form amphipathic α-helices [25, 26]. Using computer modelling, we predicted [24] that the 54 amino acid cytoplasmic carboxyl-terminus of the AT1A receptor starts with Leu305 and that the region from Lys307 to Ile320 has a high probability of forming a positively-charged amphipathic α-helix. This prediction was subsequently confirmed by Franzoni et al. [27] using synthetic peptides corresponding to this region examined by circular dichroism and nuclear magnetic resonance. Shown in Fig. 2 is the helical wheel projection of this region, which reveals an amphipathic nature characteristic of a CaM-binding site. For comparison, the helical wheel projection of the model CaM-binding peptide (CBCP, K d=∼2 nM for CaM, [28]) and an alignment of the two peptide sequences is shown.

figure image
The proximal AT1A carboxyl-terminus is a positively charged amphipathic helix. Shown are an alignment of residues 305–327 of the rat AT1A carboxyl-terminus and a leucine/lysine repeating peptide that forms a perfect amphipathic helix with nM affinity for calmodulin [28]. Identical and similar residues are indicated by lines and dashed lines, respectively. Shown below are the helical wheel representations for the AT1A receptor carboxyl-terminus peptide (termed 1ACT) and the CBCP (termed CBCP). Note, the similarity in charge and amphipathicity of the receptor sequence and the control peptide.

To determine if this basic amphipathic helical region of the AT1A receptor carboxyl-terminus was sufficient to affinity purify the ∼20 kDa Ca2+-dependent protein and could function as a possible CaM-binding site, we coupled the synthetic peptides (1ACT and CBCP) to SulphoLink gel matrix and used these peptide affinity gels to identify interacting [35S]-labelled proteins. As shown in Fig. 3, the CBCP affinity gel purified an ∼20 kDa protein and this association was lost when Ca2+ was chelated with EGTA. Similarly, the 1ACT affinity matrix enriched a cytoplasmic protein of ∼20 kDa in a Ca2+-dependent manner. Western blot analysis, using a monoclonal CaM antibody that does not cross-react with other small Ca2+-binding proteins (S100α, S100β, tropinin C and parvalbumin), identified this protein as CaM (Fig. 4).

figure image
Ca2+-dependent interaction of proteins with 1ACT and CBCP peptide affinity gels. The 1ACT and CBCP peptides were covalently attached through an introduced N-terminal cysteine to SulphoLink coupling gel. The resulting affinity gels were used to identify interacting [35S]-labelled proteins. In the presence of Ca2+ (+), but not EGTA (−), an ∼20 kDa [35S]-labelled protein (arrow) was found to associate with both the 1ACT and CBCP affinity gels. This experiment was performed four times with similar results.
figure image
Western blot analysis identifies the 20 kDa interacting protein as CaM. CHO-K1 cell extracts were purified over the 1ACT and CBCP affinity matrices, separated by SDS-PAGE and Western-blotted with an monoclonal CaM antibody. 1 ng of purified bovine CaM was run as a positive control (arrow). Immunoreactive CaM, with a migration similar to the control CaM, was bound in a Ca2+-specific manner by both peptide affinity gels. This experiment was performed three times with similar results.

To determine if the peptide 1ACT (residues 305–327 of the rat AT1A receptor) forms a high affinity complex with Ca2+/CaM, polyacrylamide gel electrophoresis in the presence of 4 M urea was performed. This procedure identifies only high affinity interactions (K d<100 nM) because 4 M urea dissociates low affinity and non-specific complexes [16]. The stoichiometry of binding was determined by incubating a constant amount of pure CaM (300 pmol) with increasing concentrations (0–3000 pmol) of either 1ACT or the control (CBCP) synthetic peptides. Shown in Fig. 5 are Coomassie blue-stained gels of the gel shift titration experiment. In the presence of Ca2+, the addition of either CBCP or 1ACT resulted in a decreased mobility (upward shift) of CaM indicating the formation of high affinity complexes between CaM and the peptides. The gel shift was complete at equimolar amounts of CaM and peptide (300 pmol) and further addition of peptide did not result in additional bands, which would suggest multivalent complexes. No complexes were formed when Ca2+ was chelated by EGTA, regardless of the amount of peptide added. These data indicate a single, high affinity, Ca2+-dependent binding site for CBCP and 1ACT on CaM.

figure image
Complex formation between CaM and the synthetic peptides 1ACT and CBCP in the presence of 4 M urea. Purified CaM (300 pmol per tube) was incubated with increasing amounts (0, 75, 150, 300, 600, 1500 and 3000 pmol) of the peptides 1ACT or CBCP in the presence of 4 M urea and either Ca2+ or EGTA. Complexes were resolved by polyacrylamide gel electrophoresis in gels containing 4 M urea and either Ca2+ or EGTA to match the samples. The gels were stained with Coomassie blue. Complex formation was observed as a shift in the migration of CaM, in the presence of Ca2+, for both peptides. No CaM shift, at any concentration of added peptide, was observed in presence of EGTA (bottom panels). This experiment was performed four times with similar results.

Our observation of a high affinity CaM-binding site in the proximal region of the AT1A receptor carboxyl-terminus has important implications for receptor function. We propose the following feedback loop between the AT1A receptor and Ca2+/CaM: AngII-stimulated AT1A receptors promote release of Ca2+ from intracellular stores, which generates activated Ca2+/CaM complexes. These complexes then feedback to bind the basically charged, amphipathic helical region encompassing residues 305–320 within the AT1A receptor carboxyl-terminus. The proposed CaM feedback is analogous to the feedback pathways reported for other Ca2+-mobilizing receptors and channels (e.g. the inhibition of the N-methyl-D-aspartate receptor [29], the CaM regulation of the transient receptor potential-like channel and the negative regulation of phototransduction, which occurs at the level of the rhodopsin receptor [30]).

The role of this putative CaM-AT1A receptor feedback loop is unknown, but the proximal carboxyl-terminus (residues 305–322) has been implicated in a number of important AT1A receptor functions. Foremost, this region is involved in binding and activating heterotrimeric G-proteins [17, 18] and point mutations of Tyr312, Phe313 and Leu314 uncouple the receptor from G-protein-dependent signalling. It is therefore tempting to speculate that once the AngII-stimulated AT1A receptor has activated and released the heterotrimeric G-protein to initiate signalling, the subsequent Ca2+ transient results in Ca2+/CaM recruitment to the G-protein-binding site to prevent re-association of G-protein with the receptor. In such a scenario, Ca2+/CaM-binding would be predicted to produce a form of receptor desensitisation. We have also reported that mutations of hydrophobic amino acids at the end of this amphipathic α-helix dramatically reduce AT1A receptor internalisation [24] and Gaborik et al. [31] have linked the region Lys307–Lys311 to cell surface expression, suggesting that important determinants of receptor trafficking are associated with this helical region. Interestingly, Ca2+/CaM has been implicated in the general process of receptor endocytosis and, more specifically, in the internalisation of another GPCR, the serotonin receptor [32]. Whether, CaM-binding to the AT1A receptor is involved in receptor internalisation remains to be determined. Finally, the sequence Tyr319–Pro322 reportedly binds phospholipase C-γ1 and a complex of SHP-2 phosphatase/Jak2 kinase [19, 20], suggesting a mechanism for the activation of growth factor-like tyrosine kinase signalling pathways by the AT1A receptor [10, 33]. Clearly, this region of the carboxyl-terminus is multifaceted, which suggests a complex hierarchy of interactions central to AT1A receptor function.

In conclusion, we have identified, for the first time, the presence of a CaM-binding site within the carboxyl-terminus of the G-protein-coupled AT1A receptor. This high affinity, Ca2+-dependent binding domain is a basic amphipathic α-helix and therefore bears all the hallmarks of a genuine CaM regulatory site. The recent identification of Ca2+/CaM-binding sites within the carboxyl-terminus of another seven transmembrane-spanning GPCR, namely the metabotropic glutamate receptor subtype 5 [34], predicts that Ca2+/CaM-binding to, and modulation of, GPCRs may be a widespread phenomenon. CaM can also regulate the GPCR kinases [35] which phosphorylate these receptors and CaM-dependent protein kinases phosphorylate and modulate the activity of the arrestin proteins [36], which bind to phosphorylated GPCRs and lead to their desensitisation and internalisation. Together, these data identify CaM as a primary regulator of GPCR function.

Acknowledgements

A National Health and Medical Research Council of Australia Institute Block Grant to the Baker Medical Research Institute and a National Heart Foundation of Australia Grant-in-Aid (G97M4935) to W.G.T supported this work. We thank Drs. Robert Andrews and Kathleen Curnow for critical reading of the manuscript.