Abstract
We previously reported that truncation of the N-terminal 79 amino acids of α1D-adrenoceptors (Δ1-79α1D-ARs) greatly increases binding site density. In this study, we determined whether this effect was associated with changes in α1D-AR subcellular localization. Confocal imaging of green fluorescent protein (GFP)-tagged receptors and sucrose density gradient fractionation suggested that full-length α1D-ARs were found primarily in intracellular compartments, whereas Δ1-79α1D-ARs were translocated to the plasma membrane. This resulted in a 3- to 4-fold increase in intrinsic activity for stimulation of inositol phosphate formation by norepinephrine. We determined whether this effect was transplantable by creating N-terminal chimeras of α1-ARs containing the body of one subtype and the N terminus of another (α1ANT-D, α1BNT-D, α1DNT-A, and α1DNT-B). When expressed in human embryonic kidney 293 cells, radioligand binding revealed that binding densities of α1A-or α1B-ARs containing the α1D-N terminus decreased by 86 to 93%, whereas substitution of α1A- or α1B-N termini increased α1D-AR binding site density by 2- to 3-fold. Confocal microscopy showed that GFP-tagged α1DNT-B-ARs were found only on the cell surface, whereas GFP-tagged α1BNT-D-ARs were completely intracellular. Radioligand binding and confocal imaging of GFP-tagged α1D- and Δ1-79α1D-ARs expressed in rat aortic smooth muscle cells produced similar results, suggesting these effects are generalizable to cell types that endogenously express α1D-ARs. These findings demonstrate that the N-terminal region of α1D-ARs contain a transplantable signal that is critical for regulating formation of functional bindings, through regulating cellular localization.
α1-Adrenoceptors (ARs) are heptahelical transmembrane proteins that belong to the G protein-coupled receptor (GPCR) superfamily, which upon agonist binding, stimulate dissociation and activation of α and βγ subunits of Gq/11 (Wu et al., 1992). Three different α1-AR subtypes (α1A, α1B, and α1D) have been cloned that exhibit differences in amino acid sequences and antagonist affinities (Zhong and Minneman, 1999a; Piascik and Perez, 2001). Each subtype is encoded by a different gene, and the human homologs of each α1-AR subtype have been studied in transfected cells. Attempts to identify functional differences between subtypes have proven difficult, because each α1-AR subtype couples to phosphoinositide hydrolysis and increases in intracellular Ca2+ (Perez et al., 1993; Theroux et al., 1996) and to activation of mitogenic pathways (Zhong and Minneman, 1999b). However, α1-AR subtypes do display differences in binding site densities and efficacies in stimulating second messenger pathways (Perez et al., 1993; Esbenshade et al., 1995; Theroux et al., 1996; Zhong and Minneman, 1999b). Of the three α1-AR subtypes, the expression levels and coupling efficiencies of α1D-ARs are substantially lower relative to the α1A-AR and α1B-AR subtypes when transfected into cell lines (Theroux et al., 1996). In addition, attempts to detect α1D-AR protein using radioligand binding in tissues that express significant α1D-AR mRNA levels have proven difficult (Yang et al., 1997), suggesting that the α1D-AR is expressed poorly in vivo. Nonetheless, studies performed using α1D-AR knockout mice clearly demonstrate that α1D-ARs play an important role in the overall regulation of blood pressure (Tanoue et al., 2002a,b).
Structure-function studies suggest that specific structural regions of α1-ARs are responsible for the observed differences in expression and signaling (Greasley et al., 2001). α1-AR subtypes are highly homologous within their transmembrane domains, which form the catecholamine binding pocket (Graham et al., 1996). There is also high conservation among regions of the third intracellular loop, which is primarily responsible for agonist-induced activation of Gq. However, the three α1-AR subtypes display little sequence homology at their C- and N-terminal domains, suggesting these domains may be important in regulating differences in expression and coupling. The properties of C-terminal splice variants (Chang et al., 1998) and mutated α1-ARs (Wang et al., 2000) suggest that the C terminus does not alter receptor expression. In contrast, few studies have investigated the role of the N-terminal regions of the α1-AR subtypes. Previously, we reported that an N-terminal truncation (Δ1-79) of the α1D-AR results in large increases in binding site density relative to full-length α1D-ARs (Pupo et al., 2003). However, whether this phenomenon is due to changes in cellular localization was not examined.
A number of previous reports have indicated that some GPCRs are sequestered in intracellular compartments in the absence of agonist-mediated internalization. For example, α2C-ARs are maintained in the endoplasmic reticulum during steady-state conditions after heterologous expression and are not recycled between the endoplasmic reticulum and the plasma membrane (Daunt et al., 1997; Edwards et al., 2000). These receptors are presumably inaccessible to agonist and therefore unlikely to mediate functional responses in vitro. Imaging of green fluorescent protein (GFP)-tagged α1-ARs in human embryonic kidney (HEK)293 cells found that α1A- and α1B-ARs are localized on the plasma membrane, and the majority of α1D-ARs are located within intracellular compartments (Mackenzie et al., 2000; Chalothorn et al., 2002). It is not yet clear whether these receptors bind ligand or are functional, and the mechanisms responsible for preventing surface expression are unknown. Similar problems with surface expression have been found with other GPCRs, including the large family of sensory receptors, in particular the odorant (Buck, 2000), and bitter taste (Chandrashekar et al., 2000) receptors, which to date, have only been able to be expressed at the plasma membrane after modification of their N-terminal domains. Thus, these studies clearly indicate the importance of understanding the role of the N terminus in regulating the expression of functional GPCRs at the plasma membrane.
The goal of this study was to determine the role of the α1D-AR N terminus in regulating both expression of α1-AR binding sites and α1-AR subcellular localization. Using confocal imaging of GFP-tagged constructs and sucrose gradient density gradient fractionation, we examined the role of the α1-AR N termini in cellular localization. In addition, we created a series of N-terminal chimeric receptors in which the N-terminal domains of the receptors were swapped. Finally, these studies involved examining α1-ARs expressed in a number of different cell types, to determine whether the observed effects are generalizable.
Materials and Methods
Materials. Materials were obtained from the following sources: cDNAs for the α1A-AR, and human α1A-, α1B-, and α1D-AR C-terminally tagged GFP constructs in pEGFP-N3 were generously provided by Dr. Gozoh Tsujimoto (National Children's Hospital, Tokyo, Japan), the human α1B-AR cDNA was from by Dr. Dianne Perez (Cleveland Clinic, Cleveland, OH), and the human α1D-AR cDNA was cloned in our laboratory (Esbenshade et al., 1995); BE 2254 was a gift from Dr. Giuseppe Romeo (Universita di Catania, Catania, Italy); HEK293 and Phoenix producer cells were from American Type Culture Collection (Manassas, VA); (-)-arterenol bitartrate (noradrenaline), prazosin, Dowex-1, Dulbecco's modified Eagle's medium, penicillin, streptomycin, FLAG peptide, anti-FLAG M2 affinity resin, and horseradish peroxidase-conjugated anti-FLAG M2 antibody were from Sigma-Aldrich (St. Louis, MO); BMY 7378 was from Sigma/RBI (Natick, MA); carrier-free Na125I, enhanced chemiluminescence reagent was from Amersham Biosciences Inc. (Chicago, IL); myo-[3H]inositol was from American Radiolabeled Chemicals (St. Louis, MO); Precast Tris-Glycine gels were from Novex (Carlsbad, CA); QuikChange site-directed mutagenesis kit was from Stratagene (Cedar Creek, TX); serum and trypsin were from Invitrogen (Carlsbad, CA); [3H]prazosin was from PerkinElmer Life Sciences (Boston, MA); and Superfect and Polyfect transfection reagents were from QIAGEN (Valencia, CA). Rat aortic smooth muscle cells (RASMs) were kindly donated by Dr. T. J. Murphy (Emory University, Atlanta, GA).
Constructs. To construct N-terminal α1-AR chimeras, cDNAs spanning the full-length coding sequences for the α1A-, α1B-, and α1D-ARs containing N-terminal FLAG epitopes in the mammalian expression plasmid pDT were subjected to site-directed mutagenesis using the QuikChange kit (Stratagene). MluI restriction sites (ACGGCT) were created in the N-termini close to the putative first transmembrane domain by polymerase chain reaction using specific primers to substitute amino acids TR for S24K25 (70TCCAAG75) in α1A-, T43R44 in α1B- (127ACCAGG132) (silent) and A94Q95 in α1D-AR (280GCGCAG285) and sequenced. pDT vectors containing mutated α1-AR cDNAs were subjected to digestion with MluI and EcoRI, and the products were separated on an agarose gel and extracted. Isolated cDNA fragments were ligated to create α1A- (α1ANT-D) and α1B-ARs (α1BNT-D) with the α1D-AR N terminus and the α1D-ARs with the α1A- (α1DNT-A) or α1B-AR N terminus (α1DNT-B). Δ1-79α1D N-truncated mutants were generated by polymerase chain reaction using specific primers as described previously (Pupo et al., 2003). To create GFP-tagged constructs, cDNAs for GFP-α1D-, Δ1-79α1D-, and α1DNT-B were digested with EcoRI and AgeI, fragments isolated, and N-terminal portions of the Δ1-79α1D- and α1DNT-B-AR constructs religated to the C-terminal portion of the GFP-α1D-AR construct contained in the pEGFP-N3 vector. To create GFP-tagged α1BNT-DARs, GFP-α1B- and α1BNT-D-ARs were digested with EcoRI and AscI, fragments were isolated and the N-terminal portion of α1BNTD-AR was ligated to the C-terminal portion of GFP-α1B-AR contained in the pEGFP-N3 vector and verified by restriction analysis.
Cell Lines. HEK293 and RASM cells were propagated in Dulbecco's modified Eagle's medium with sodium pyruvate supplemented with 10% heat-inactivated fetal bovine serum, 100 μg/ml streptomycin, and 100 U/ml penicillin in a humidified atmosphere with 5% CO2 (Theroux et al., 1996; Wang et al., 1997). Confluent plates were subcultured at a ratio of 1:3. CHO-K1 cells were propagated in Ham's F-12 medium supplemented with 10% fetal bovine serum and 200 μg/ml G418 (Wang et al., 2000).
Transfections. For HEK293 cultures, cells were transfected with 10 μg of cDNAs using Superfect transfection reagent, and stably transfected cells were selected with G418 (400 μg/ml). RASMs were transfected with infectious retroviral supernatants harvested from transfected Phoenix producer cells generated by a helper virus-free protocol as described previously (Abbott et al., 2000). CHO-K1 cells were transfected using PolyFect transfection reagent, and cells were selected for resistance to G418 (800 μg/ml).
Radioligand Binding. Confluent 150-mm plates were washed with phosphate-buffered saline (PBS; 20 mM NaPO4, 154 mM NaCl, pH 7.6) and harvested by scraping. Cells were collected by centrifugation, homogenized with a Polytron, centrifuged at 30,000g for 20 min, and resuspended in 1× buffer A (25 mM HEPES, 150 mM NaCl, pH 7.4) supplemented with protease inhibitors (1 mM benzamidine, 3 μM pepstatin, 3 μM phenylmethylsulfonyl fluoride, 3 μM aprotinin, 3 μM leupeptin, and 5 mM ethylenediamine tetraacetic acid). Radioligand binding sites were measured by saturation analysis of specific binding of the α1-adrenoceptor antagonist radioligand 125I-BE 2254 (20-800 pM). Nonspecific binding was defined as binding in the presence of 10 μM phentolamine. The pharmacological specificity of radioligand binding sites was determined by displacement of 125I-BE 2254 (50-70 pM) by prazosin and BMY 7378, and data were analyzed using nonlinear regression (Theroux et al., 1996).
Laser Confocal Microscopy. HEK293 cells transiently transfected with GFP-tagged constructs were grown on sterile coverslips, fixed for 30 min with 2% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4, and rinsed several times with PBS containing 0.5% normal horse serum. Coverslips were then mounted onto slides with Vectashield mounting medium (Vector Laboratories, Burlingame, CA). Cells were scanned with an LSM 510 laser scanning confocal microscope (Carl Zeiss, Heidelberg, Germany) as described previously (Volpicelli et al., 2001). GFP fluorescence was excited using an argon laser at a wavelength of 488 nm. Z-scans were performed with 1 μm as the average slice size. Each image is the average of 16 scans.
Sucrose Density Gradient Centrifugation. CHO-K1 cells stably expressing FLAG-tagged full-length human α1-AR subtypes were grown to confluence on 150-mm plates, washed, and lysed by scraping in ice-cold hypotonic lysis buffer (1 mM Tris, pH 7.4, 140 mM NaCl). The lysate was layered onto a discontinuous sucrose density gradient consisting of 1.7 ml of 15% sucrose (w/v), 5.0 ml of 30% sucrose, and 2.5 ml of 60% sucrose. Samples were centrifuged at 28,000 rpm for 65 min at 4°C using an SW41 rotor in an L8-70 refrigerated ultracentrifuge (Beckman Coulter Inc., Fullerton, CA. Fractions of 1 ml each were collected, and aliquots were subjected to radioligand binding with 1.2 to 1.4 nM [3H]prazosin in Tris binding buffer (20 mM Tris, pH 7.4, 2 mM MgCl2, 140 mM NaCl), as described previously (Wang et al., 2000). Nonspecific binding was calculated as binding in the presence of 100 μM phentolamine.
Measurement of [3H]InsP Formation. Accumulation of [3H]InsP was determined in confluent 96-well plates. Cells were prelabeled with myo-[3H]inositol for 48 h, and the production of [3H]InsP was determined by modification of a protocol described previously (Wilson et al., 1990). After prelabeling, medium containing [3H]inositol was removed, and 100 μl of Krebs' buffer (129 mM NaCl, 5.5 mM KCl, 2.5 mM CaCl2, 1.2 mM NaH2PO4, 1.2 mM MgCl2, 20 mM NaHCO3, 11 mM glucose, 0.029 mM Na2EDTA) containing 10 mM LiCl was gently added to each well. To stimulate [3H]InsP formation, cells were incubated with increasing concentrations of NE in Krebs' uffer for 60 min. After 60 min, the reaction was stopped by the addition of 40 mM Tris-formate,and samples were ultrasonicated for 10 s. Each sample was then separated by anion exchange chromatography, and the amount of [3H]InsP formation in each sample was quantified. Total myo-[3H]inositol incorporation in each sample was determined by removing 5-μl aliquots before extraction. [3H]InsP formation was then calculated as [3H]InsP formation (cpm)/total [3H]Ins incorporated (cpm) ± S.E.M.
Immunoprecipitation/Immunoblotting. HEK293 cells expressing FLAG-tagged full-length α1-ARs or FLAG-tagged N-terminal chimeras were harvested by scraping in ice-cold PBS and washed by repeated centrifugation and homogenization. Cell lysates were solubilized, immunoprecipitated with anti-FLAG M2 resin and probed using anti-FLAG M2 antibodies to detect protein as described previously (Pupo et al., 2003).
Data Analysis and Statistics. Radioligand binding and sucrose density fractionation data were calculated as means ± S.E.M. and statistical comparisons used GraphPad Prism software (GraphPad Software Inc., San Diego, CA).
Results
Differential Distribution of GFP-Tagged Human α1-AR Subtypes. Previous reports have suggested that α1-AR subtypes show different subcellular distributions (Hirasawa et al., 1997; Mackenzie et al., 2000; Chalothorn et al., 2002). We obtained full-length human α1-AR subtypes with GFP tags at their C termini (Hirasawa et al., 1997; Chalothorn et al., 2002), transiently transfected them into HEK293 cells, and visualized their distribution by confocal microscopy. Figure 1A shows that both GFP-α1A and GFP- α1B-ARs showed a primarily surface localization, whereas Fig. 1B shows that GFP-α1D-ARs were found almost exclusively in intracellular compartments, as reported previously (Chalothorn et al., 2002).
N-Terminal Truncation of the α1D-AR Changes Cellular Localization. Because α1D-ARs are primarily located at intracellular sites after heterologous expression, where they presumably cannot be stimulated by endogenous hydrophilic agonists, we hypothesized that N-terminal truncation might increase binding site density by allowing localization of these receptors to the plasma membrane. To investigate this, N-terminal fragments were swapped between untagged and C-terminal GFP-tagged α1-AR subtypes to create cDNA constructs of the full-length human α1D- and Δ1-79α1D-AR containing the C-terminal GFP-tag. Each receptor cDNA was transiently transfected into HEK293 cells, and cells were fixed and examined using confocal microscopy. N-Terminal truncation caused a dramatic translocation of the majority of α1D-ARs from intracellular compartments to the plasma membrane (Fig. 1B), with essentially no intracellular accumulation observed.
To confirm these results using a different technique that measures radioligand binding sites rather than protein expression, we isolated subcellular fractions using sucrose density gradient fractionation, as described previously (Wang et al., 1997; Toews, 2000). Each fraction was then subjected to radioligand binding analysis using the α1-AR-selective antagonist [3H]prazosin. As shown in Fig. 2A, the majority of α1A- and α1B-AR binding sites (40-42%) were located in the “heavy” fractions (fractions 8 and 9) containing plasma membrane, whereas the full-length α1D-AR binding sites (>44%) were located in the “light” fraction (fractions 3 and 4) containing intracellular vesicles. However, N-terminal truncation of the α1D-AR increased the total amount of binding and dramatically shifted the binding sites from the “light” intracellular membranes (19%) fraction to the heavy plasma membrane fraction (57%) (Fig. 2B). These fractionation results support those obtained in GFP fluorescence confocal studies, suggesting that the results are not due to artifacts of tags or experimental techniques. Thus, these data indicate that in addition to increasing binding site density, N-terminal truncation causes an almost completely quantitative translocation of α1D-ARs to the plasma membrane.
N-Terminal Truncation of α1D-ARs Increases the Intrinsic Activity of NE in Stimulating [3H]InsP Formation. We examined whether N-terminal truncation of α1D-ARs was associated with increases in the ability of NE to stimulate [3H]InsP formation. Concentration-response curves for NE stimulation of [3H]InsP formation were generated in HEK293 cells stably expressing either FLAG-tagged full-length α1D-ARs or FLAG-tagged Δ1-79α1D-ARs (Fig. 3). Compared with full-length α1D-ARs, N-terminal truncation increased the intrinsic activity of NE by 3.3-fold, without causing a significant increase in potency, supporting the hypothesis that N-terminal truncation increases the formation of functional binding sites.
The α1D-AR N Terminus Contains a Transplantable Signal That Decreases Binding Site Density. To determine whether the effect of the N terminus was transplantable, a series of α1-AR chimeric receptors were constructed in which N-terminal regions were swapped between subtypes (Fig. 4). HEK293 cells were then transfected with either full-length or N-terminal α1-AR chimera cDNA constructs and selected with G418. Cell membranes were harvested and used for radioligand binding studies with the α1-AR antagonist 125I-BE 2254. In agreement with previous studies (Theroux et al., 1996), saturation experiments demonstrated that recombinant α1A- and α1B-ARs were highly expressed, whereas recombinant α1D-ARs were expressed poorly (Table 1). Interestingly, both the α1A- and α1B-AR chimeric receptors containing the α1D-AR N terminus displayed almost 10-fold decreases in binding site density relative to the full-length receptors (Fig. 5, A and B), whereas the α1D-AR chimeras containing the α1A- or α1B-AR N-terminal regions resulted in a 2- to 3-fold increase in binding site densities in comparison with the full-length α1D-AR (Fig. 5, C and D). To support these findings, truncation of the initial 38 amino acids of the α1B-AR N terminus (Δ1-38α1B) did not alter the binding site density in comparison with full-length α1B-ARs (data not shown), suggesting that the observed differences in binding site density were a result of transplantable signal within the α1D-AR N terminus and not a result of N-terminal truncation.
From saturation and competition radioligand binding experiments, we detected no significant differences in the affinity of the nonselective α1-AR antagonists 125I-BE 2254 or prazosin for full-length or chimeric α1-ARs (Table 1). In addition, the α1D-AR-selective antagonist BMY 7378 bound with high affinity to α1DNT-A and α1DNT-B chimeric and full-length α1D-ARs, whereas it had a lower affinity for α1A-, α1B-, α1ANT-D, and α1BNT-D chimeric receptors, although these showed some variations. As previously reported, the N-terminal FLAG-tag did not alter the affinity values for 125I-BE 2254, prazosin or BMY 7378 for any α1-AR subtype (Vicentic et al., 2002). Together, these data suggest that the α1D-AR N terminus contains a signal that can be transplanted onto the α1A- and α1B-AR subtypes, causing a decrease in expression of binding sites without altering the pharmacological properties of the expressed receptors.
The α1D-AR N Terminus Contains a Transplantable Signal That Controls Cellular Location. We also determined whether the N-terminal signal of the α1D-AR controlling surface expression was transplantable onto other subtypes. If so, α1-AR chimeras containing the α1D-AR N terminus would be concentrated at intracellular regions, and chimeras with other α1-AR N termini would be expressed at the plasma membrane. To examine this hypothesis, C-terminal GFP-tagged α1BNT-D and α1DNT-B chimeras were constructed. As in previous experiments, cDNAs for each chimera were transiently transfected into HEK293 cells, which were then fixed and examined using confocal microscopy. Figure 6 shows that α1BNT-D chimeric receptors displayed similar cellular localization patterns as full-length α1D-ARs, with dense fluorescence in intracellular regions. In addition, the α1B-AR N terminus was able to confer an α1B-AR localization pattern, because α1DNT-B chimeric receptors were primarily localized at the plasma membrane (Fig. 6). Therefore, these data indicate that the α1D-AR N terminus contains a transplantable signal that can alter the localization of α1-ARs in addition to preventing expression of functional binding sites.
To ensure that the observed differences in binding site densities and cellular localization patterns were not caused by errors in protein transcription or caused by proteolysis, membranes prepared from HEK293 cells stably transfected with FLAG-tagged α1B-, α1BNT-D, α1D-, or α1DNT-B subtypes were immunoprecipitated and immunoblotted using anti-FLAG antibodies. Protein bands of the expected size were observed for both full-length α1B- and α1D-ARs and their corresponding N-terminal chimeras (Fig. 7), suggesting that the observed differences in this study are not due to nonspecific artifacts.
N-Terminal Truncation Allows α1D-AR Localization to the Plasma Membrane and Increases Binding Site Density in RASMs. The studies above were performed in cell lines that do not endogenously express α1-ARs (HEK293 and CHO-K1). Because these cells may lack the necessary components required to ensure proper folding and expression of α1D-ARs, we conducted similar experiments in RASMs. cDNAs for full-length and Δ1-79α1D-ARs containing C-terminal GFP tags were retrovirally transfected into RASMs, which normally express α1D-ARs in vivo but gradually lose their α1-AR expression with cell passaging. As observed in HEK293 cells, 125I-BE 2254 saturation binding demonstrated the binding density of Δ1-79α1D-ARs (Bmax = 308 ± 135) to be significantly higher than the binding density of full-length α1D-ARs (Bmax = 77 ± 49) (Fig. 8A). No specific binding was observed in untransfected RASMs (data not shown). 125I-BE 2254 was bound with high affinity to both GFP-tagged α1D-ARs (KD = 0.05 ± 0.03 nM) and Δ1-79α1D-ARs (KD = 0.14 ± 0.05 nM), suggesting that the C-terminal GFP-tag did not alter the pharmacological properties of these receptors, as reported previously (McCune et al., 2000). In addition, confocal microscopy revealed results somewhat comparable to those found in HEK293 cells, with N-terminal truncation altering the cellular localization of α1D-ARs from intracellular compartments to the plasma membrane (Fig. 8B). However, N-terminal truncation of the α1D-AR only caused partial redistribution to the plasma membrane in RASMs, suggesting that differences in cell phenotypes can affect receptor localization and trafficking. However, these data show that the effects of the α1D-AR N terminus on expression of binding sites and cellular localization are not limited to a single cell type and occur in vascular cells that normally express α1-ARs.
Discussion
Of the three α1-AR subtypes, the α1D-AR has been the least studied due to difficulties in obtaining significant expression levels and to its poor coupling to signal transduction pathways (Theroux et al., 1996). However, recent experiments performed in α1D-AR knockout models suggest that this α1-AR subtype plays an important role in the overall regulation of blood pressure (Tanoue et al., 2002b), stressing the functional importance of this subtype. In addition to α1D-ARs, several other class I GCPRs demonstrate problems in surface expression, including both odorant and bitter taste GCPRs (Krautwurst et al., 1998; Chandrashekar et al., 2000). Several studies have attempted to increase the expression of these receptors through N-terminal modification. Previously, we demonstrated that truncation of the proximal 79 amino acids of the α1D-AR N terminus resulted in significant increases in binding site density (Pupo et al., 2003). However, we did not investigate how the α1D-AR N terminus regulates binding site density. In this study, we tested the hypothesis that the α1D-AR N terminus regulates expression of functional receptors through controlling cellular localization. Two different methods, including confocal imaging of GFP-tagged α1-ARs and sucrose density gradient fractionation led to the same conclusion; that the α1D-AR N terminus prevents the expression of functional α1D-ARs at the plasma membrane. To confirm these results, we determined whether the α1D-AR N-terminal effects on cellular location could be transplantable on to the α1A-AR and α1B-AR subtypes. To test this, we created a series of α1-AR N-terminal chimeric receptors, including α1A- and α1B-ARs containing the α1D-AR N terminus and α1D-ARs containing either the α1A-or α1B-AR N-terminal domains. These α1-AR constructs were expressed in HEK293 cells, and their Bmax and KD values were determined using 125I-BE 2254 radioligand binding in isolated membranes. We found that the presence of the α1D-AR N terminus on either the α1A-AR or α1B-AR resulted in significant decreases in binding site density without altering the pharmacological properties of these receptors. Thus, these studies clearly demonstrate the N terminus of the α1D-AR serves an important role in regulating the expression of functional receptors at the plasma membrane.
Upon transfection into HEK293 cells, the α1-AR subtypes display characteristic binding properties and are able to stimulate phosphoinositide hydrolysis and increases in intracellular Ca2+ levels (Theroux et al., 1996), but they display differences in binding densities and coupling efficiencies, with α1A > α1B > α1D. The mechanisms regulating differences in α1-AR subtype binding densities and coupling efficiencies are unknown. In this and previous studies, we have shown that the N terminus of the α1D-AR is primarily responsible for regulating binding site density and coupling efficiency, because truncation of the proximal 79 amino acids of the human α1D-AR N terminus resulted in 6- to 13-fold increases in the binding site density and increased coupling to Ca2+ release and inositol 1,4,5-trisphosphate production by ∼3-fold (Pupo et al., 2003) compared with the full-length α1D-AR. Although the amino acid sequences within the transmembrane and third intracellular loops of these receptors are highly homologous (Graham et al., 1996), there is very little sequence homology at their C- and N-terminal domains. Studies using mutated α1B-ARs (Wang et al., 2000) and α1A-ARs (Price et al., 2002) have suggested that the C terminus does not affect receptor expression but instead plays differential roles in α1-AR desensitization and internalization after agonist exposure. The N-terminal regions of the α1-AR subtypes differ greatly in length (α1D is 95 aa, α1B is 45 aa, and α1A is 25 aa), and our studies suggest that the long N terminus of the α1D-AR contributes to preventing expression of functional receptors at the plasma membrane. Previous reports have suggested the N-terminal domains of other receptors play a significant role in regulating surface expression, although these studies generally found that the N terminus promotes receptor expression. In direct contrast, our results suggest that the N terminus of the α1D-AR reduces receptor expression. For example, truncation of the initial 64 amino acids from the N terminus of the endothelin B receptor decreases expression by ∼15-fold (Kochl et al., 2002). In addition, removal of nine residues from the N terminus of the GluRI subunit prevents surface expression of the AMPA ligand-gated ion channel by promoting sequestration within the endoplasmic reticulum (Xia et al., 2002). Addition of a cleavable signal sequence from the influenza hemagglutinin onto the β2-AR N terminus resulted in significant increases in binding site density (Guan et al., 1992). Studies involving olfactory and bitter taste GPCRs have been greatly hindered due to their inability to be expressed in transfected cell systems. However, addition of the first 20 to 39 amino acids of the rhodopsin receptor N terminus to these receptors promotes significant increases in binding site expression and translocation to the plasma membrane (Krautwurst et al., 1998; Chandrashekar et al., 2000). Our results also support an important role for the N terminus in expression and trafficking of the α1D-AR; although in this case, the N terminus reduces receptor expression, apparently by promoting intracellular retention.
GPCRs were originally assumed to be primarily, if not exclusively, located on the cell surface where they would be easily accessible to hydrophilic ligands. However, it is now known that many GPCRs, when expressed in heterologous systems, are sequestered in intracellular compartments. The odorant receptor family is unable to reach the cell surface when heterologously expressed in recombinant systems (Buck, 2000). Immunocytochemical localization studies have identified the α2C-AR to be predominantly intracellular in both Madin-Darby canine kidney II cells (Wozniak and Limbird, 1996) and HEK293 cells (von Zastrow et al., 1993), primarily within the endoplasmic reticulum (Daunt et al., 1997); thus, they are unlikely to be accessible to their hydrophilic ligands. In the current studies, we confirmed previous reports that α1D-ARs are primarily located at intracellular sites, as determined from confocal microscopy of GFP-tagged receptors and [3H]prazosin radioligand binding to isolated subcellular fractions. This agrees with previous studies reporting intracellular α1-ARs using fluorescent imaging of GFP-tagged α1A-ARs in COS-7 (Hirasawa et al., 1997) and GFP-tagged α1D-ARs in HEK293 cells (Chalothorn et al., 2002). In addition, radioligand binding studies performed with a fluorescent form of the α1-AR-selective antagonist prazosin revealed that approximately 40% of α1-ARs in smooth muscle are found intracellularly (Mackenzie et al., 2000). Several explanations for the existence of intracellular pools of GPCRs have been proposed, including the possibility that these pools act as reservoirs for receptors available for translocation to the plasma membrane or that they may comprise specific microdomains important in signaling (Edwards et al., 2000). An alternate explanation may include the existence of binding partners that recognize the α1D-AR N terminus and act to prevent proper folding and retention, and possibly degradation of α1D-ARs within the endoplasmic reticulum. Although misfolded or partially degraded α1D-ARs can still be recognized using fluorescence imaging or immunostaining techniques, these methods are incapable of making distinctions between protein and functional receptors. In fact, our data suggests that the intracellular α1D-ARs may not represent a pool of functional receptors, because Δ1-79α1D-ARs displayed significant increases in binding site density and coupling efficiency over full-length α1D-ARs, yet α1D-AR protein expression levels are greater then that observed for Δ1-79α1D-ARs (Pupo et al., 2003).
Finally, we examined α1D-AR subcellular localization and expression in multiple cell types, because not all cell phenotypes may contain the necessary signals or binding partners required to transport a target receptor to the plasma membrane. We found that α1D-ARs are expressed almost exclusively intracellularly in HEK293, CHO-K1, and RASM cells, suggesting that such intracellular localization is generalizable. Thus, heterologously expressed α1D-ARs are primarily intracellular in many cell types, although the functional significance of this phenomenon is still unknown.
The findings from these experiments may have implications for α1D-AR function in intact physiological systems. To date, examples of norepinephrine-mediating responses through α1D-AR stimulation have been limited to contraction of rat aorta (Piascik et al., 1995), mesenteric (Hussain and Marshall, 2000), and carotid and pulmonary arteries (Hussain and Marshall, 1997) and stimulation of protein and mRNA synthesis in rat aorta (Chen et al., 1995). However, recent studies performed on knockout mice suggest that the α1D-AR plays an important role in overall control of blood pressure and development of hypertension (Tanoue et al., 2002b) as well as ventricular contraction and cardiac inotropy (Turnbull et al., 2003). In our studies, we examined the effect of N-terminal truncation on α1D-AR expression and localization in isolated cardiovascular cells, RASMs. Although the increases in binding site density and plasma membrane translocation were not as dramatic as in HEK293 cells, we found that N-terminal truncation increased expression and cell surface localization in RASMs also, suggesting that our findings are relevant in cardiovascular cells as well.
In summary, our findings suggest the N-terminal domain of α1D-ARs plays a crucial role in two important processes, the expression of functional binding sites and their localization at the plasma membrane. These results may be applicable to other GPCRs that are difficult to express and/or show primarily intracellular localization such as the α2C-AR, odorant and taste receptors, and provide valuable insights into the structure-function relationships of this important receptor family.
Acknowledgments
We are grateful to Gozoh Tsujimoto, Dianne Perez, Giuseppe Romeo, and T. J. Murphy for donation of cells and reagents; to Laura Volpicelli, Howard Rees, and Alan Levey for help with confocal microscopy; to James Loss and T. J. Murphy for aid with retroviral transfection; and to George Rogge for technical assistance.
Footnotes
-
This study was supported by National Institutes of Health Grants NS-21325 (to K.P.M.) and GM-34500 (to M.L.T.).
-
DOI: 10.1124/jpet.103.060509.
-
ABBREVIATIONS: AR, adrenoceptor; GPCR, G protein-coupled receptor; GFP, green fluorescent protein; HEK, human embryonic kidney; BE 2254, 2-[[β-(4-hydroxyphenyl)ethyl]aminomethyl]-1-tetralone; BMY 7378, 8-[2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl]-8-azaspiro[4,5]decane-7,9-dione; NT, N terminus; RASM, rat aortic smooth muscle; CHO, Chinese hamster ovary; PBS, phosphate-buffered saline; InsP, inositol phosphate; NE, norepinephrine; aa, amino acid.
- Received September 24, 2003.
- Accepted December 3, 2003.
- The American Society for Pharmacology and Experimental Therapeutics