Abstract
Cellular distribution and binding characteristics of native α1-adrenoceptors (ARs) were determined in a live, single, human smooth muscle cell (SMC) with confocal laser scanning microscopy and a fluorescent ligand, BODIPY-FL prazosin (QAPB). This allowed single-cell competitive ligand binding and showed that 40% of α1-AR-binding sites in native cells are intracellular. QAPB had high affinity and acted as a nonselective, competitive antagonist versus [3H]prazosin at cloned human α1a-, α1b-, and α1d-AR subtypes on membrane preparations and whole cells. RS100329 had 70-fold selectivity for α1a-ARs versus α1b- and α1d-ARs, validating its use to identify this subtype. In similar cells QAPB-associated fluorescence provided quantitative data analogous and comparable to [3H]prazosin binding in whole cells. In human, dissociated, prostatic smooth muscle cells QAPB-associated fluorescence binding exhibited specific high-affinity binding properties (FKD = 0.63 ± 0.02 nM), which was 3- to 4-fold higher compared with recombinant cells (FKD = 2.1–2.3 nM). Internal consistency in the data showed that affinity is greater, in general, in membrane preparations than in cells but also greater in the native prostatic tissues or cells than in equivalent recombinant receptors. Fluorescence revealed binding sites both on the plasmalemmal membrane and on intracellular compartments: at all locations RS100329 inhibited QAPB binding identifying the sites as α1A-ARs. Quantitative three-dimensional mapping of QAPB-associated fluorescence binding in native human cells showed that 40% of high-affinity-binding sites was in intracellular compartments. This provides a potential new site for physiological agonism and makes intracellular access a potential differentiator of drug action.
Heptahelical G-protein-coupled receptors play an important role in transducing neurotransmitter signals in the central and autonomic nervous systems. Although they have an established functional location at the plasma membrane with the extracellular ligand recognition site regulating various intracellular effectors (Fonseca et al., 1995), little is known about their subcellular localization in native systems. Tagged recombinant receptors show that they are highly expressed in intracellular regions (Slice et al., 1998; Drmota et al., 1999). However, it is unknown whether this distribution is paralleled in native systems or is an aberration of an artificial system. If the location is significant it could indicate a possible new functional location.
Several lines of evidence suggest that their localization might be of significance. Recent evidence points to intracellular actions, e.g., prostaglandin E2 receptors (Smith et al., 1999) and a recent review points to a number of enigmas that have arisen because it has emerged that heptahelical receptors do not necessarily couple exclusively to G-proteins (Hall et al., 1999). Furthermore, there are various discrepancies between the ligand-binding properties of α1-adrenoceptors (ARs) in membrane preparations and in functional tests of affinity in intact cells (Hieble et al., 1995; Ford et al., 1998).
Cellular distribution of heptahelical receptors in native systems is obscure because the high degree of sequence homology in this large family obstructs the creation of specific antibodies, bedevilling immunohistochemistry. An approach that avoids this and has the advantages of use in live cells and of reporting on ligand-binding properties is fluorescence ligand binding (McGrath et al., 1996a). Provided that they are lipophilic or have a transport mechanism (Al-Damluji et al., 1997), fluorescent ligands can bind to receptive sites irrespective of location, achieve greater specificity for the protein based on pharmacology, allow quantitative competition by other ligands to reveal pharmacological data, and have excellent spatial resolution, thereby allowing subcellular localization.
We therefore set out to investigate the localization and correlation of binding sites with receptors in native cells to test the hypothesis that receptor localization might be a determinant of ligand action. We chose a receptor, ligand type, and cell type of current therapeutic interest. In human prostate smooth muscle, α-blockers are increasingly used to treat benign prostatic hypertrophy, an enormously prevalent and disabling condition in elderly men. However, the mechanisms underlying therapeutic benefit and even the subtype of receptor involved are controversial (Marshall et al., 1995; Ford et al., 1996, 1998; McGrath et al., 1996b).
The benefits of imaging fluorescent ligand binding in live cells or intact tissues have until now been underexploited. Cellular distribution of a ligand can be demonstrated and receptor localization followed in time, such as during desensitization. Conventional radioligand binding can be validated quantitatively on live cells, allowing analysis in realistic and variable experimental conditions. By taking binding to the single-cell level, analysis can be performed on heterogeneous native tissue, cells being isolated either physically or optically. For this purpose, vital confocal microscopy is ideal because it can then be applied to native cells, hence localizing receptors and verifying their identity by analogy with the fluorescence-binding properties of the recombinant receptor subtype.
The objective of this study was to identify the cellular localization of α1-AR ligand-binding sites on a single, live, human prostate smooth muscle cell (SMC). Previous work by our group has demonstrated the underlying principles using nanomolar concentrations of the fluorescent ligand BODIPY-FL prazosin (QAPB) on recombinant α1-ARs in living cells (McGrath et al., 1996a; Daly et al., 1998). The compound is lipophilic, enters live cells, and hence reports spatial distribution of binding sites throughout the cell, not only at expected functional locations on the plasma membrane. QAPB is much more fluorescent when bound to the receptor than in aqueous solution, indicating receptor occupation under equilibrium conditions (Daly et al., 1998). The compound is nonselective between the three known α1-AR subtypes but, in principle, selective competitive ligands should enable precise identification. This was achieved.
This study is also unique in showing the subcellular distribution of ligand-binding sites in live native cells, in quantitative terms, and in making the validating connection between the binding properties in live cells of recombinant and native receptors. The main technical advance is that the fluorescence binding characteristics in small subcellular regions in live native cells can be identified with a particular recombinant receptor. The biological advance is that the binding properties of receptors in different subcellular regions have been defined, as have the quantitative interaction of ligands at the different sites, producing a new series of possibilities for drug selectivity.
Experimental Procedures
Materials and Drugs
All chemicals were purchased from Sigma-Aldrich Co. (Poole, UK) unless otherwise stated: (R)-A-61603 (Dr. Michael Meyer, Abbott Laboratories, Chicago, IL); BMY7378 (Research Biochemicals, Natick, MA); prazosin HCl, QAPB, and Syto 13 (Molecular Probes, Eugene, OR); phentolamine, 5-methylurapidil, and WB4101 (Research Biochemicals); RS100329 (Dr. Michelson, Roche Bioscience, Palo Alto, CA); doxazosin (Pfizer, Sandwich, UK); and tamsuolsin (YM12617; Yamanouchi Laboratories, Tokyo, Japan).
Cell Culture
Rat-1 fibroblasts stably expressing human α1a-, α1b-, and α1d-AR subtypes were used (Schwinn et al., 1990). Cells were maintained in Dulbecco's modified Eagle's medium containing 10% (v/v) fetal calf serum, 2 mMl-glutamine, 250 μg/ml geneticin (G418), 100 U/ml penicillin, and 100 μg/ml streptomycin (all Life Technologies, Paisley, Scotland) in a humidified atmosphere at 37°C containing 5% CO2 up to passage 30. Human prostatic cells were dissociated from prostate tissue and grown as described above in Dulbecco's modified Eagle's medium containing 50% (v/v) fetal calf serum, 2 mM l-glutamine, 200 U/ml penicillin, and 200 μg/ml streptomycin up to passage 3. Smooth muscle cells were immunohistochemically stained with a cy3 conjugated α-actin monoclonal antibody (Sigma Chemical Co.). The stained specimens were then examined with a Zeiss Axiophot microscope.
Membrane Preparation
Membranes were prepared as described previously (Daly et al., 1998). Briefly, cell pellets or human prostatic chips were resuspended in 5 ml of Tris-HCl buffer (150 mM NaCl, 50 mM Tris-HCl, 0.5 mM EDTA, 1 μg/ml leupeptin, 10 mM benzamidine, 500 μg/ml soya bean trypsin inhibitor, and 10% v/v glycerol, pH 7.4, and homogenized at setting 6 for 3 × 5 s with an ultrapolytron. After two centrifugation steps (600g; 10 min; 4°C), the supernatant fractions were centrifuged at 56,000g for 30 min at 4°C. The resulting membrane pellet was resuspended in 1 ml of ice-cold Tris-HCl buffer and homogenized with a 5-ml Teflon-in-glass homogenizer. The homogenate was processed immediately for protein estimation with a Pierce protein assay kit (Pierce, Rockford, IL) and adjusted to 0.5 mg/ml. Aliquots not used immediately for radioligand-binding studies were stored frozen at −80°C.
Primary Smooth Muscle Cell Dissociation
Cells were dissociated by the method of Kamishima and McCarron (1997). Briefly, tissues (human prostate resections) were immediately placed in buffer 1 (147 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, 10 mM HEPES, and 0.1% BSA, pH 7.4). Tissue was washed once in buffer 1, resuspended in buffer 2 (80 mM sodium glutamate, 54 mM NaCl, 5 mM KCl, 1 mM MgCl2, 0.1mM CaCl2, 10 mM HEPES, 10 mM glucose, 0.2mM EDTA, and 0.1% BSA, pH 7.3) with 1.7 mM papain and 0.7 mM dithioerythritol, and incubated at 35°C for 30 min. Tissues were centrifuged at 1200g for 2 min and the supernatant discarded. Tissues were resuspended in buffer 2 with 1.0 mM collagenase II and 1.0 mM hyaluronidase and SMCs were dispersed immediately with a fire-polished Pasteur pipette. Cells were resuspended in buffer 2 and plated onto coverslips.
Radioligand-Binding Studies
Competition Experiments.
Competitive α1-AR-binding assays were performed by incubating fibroblast membranes (0.05 mg/ml) with [3H]prazosin (0.2 nM) in the presence or absence of a range of increasing concentrations of competing ligands in a total volume of 0.5 ml of 50 mM Tris-HCl assay buffer. Nonspecific binding was defined as binding in the presence of 10 μM phentolamine. After equilibrium (30 min at 25°C) bound ligand was separated from free ligand by vacuum filtration over Whatman GF/C filters on a Brandell cell harvester.
Saturation Experiments.
Saturation studies were performed with either cell or human prostate tissue membranes (0.05 or 0.5mg/ml, respectively) or whole cells (1 × 106cells/tube) that were incubated in triplicate with [3H]prazosin (0.05–10 nM; Amersham Corp., Amersham, UK) in 50 mM Tris-HCl assay buffer (50 mM Tris-HCl, 10 mM NaCl, 5 mM EDTA, and 10 mM MgCl2, pH 7.4). The reaction mixture was incubated in a final volume of 0.5 ml for 30 min at 25°C for the measurement of total binding. Nonspecific binding was determined in the presence of 10 μM phentolamine. The reaction was terminated with a Brandell cell harvester and bound ligand was separated from free ligand by vacuum filtration over Whatman GF/C filters. The pellet was washed with 3 × 3 ml of ice-cold assay buffer and bound radioactivity determined by liquid scintillation counting. Specific binding was calculated by subtracting nonspecific from total binding and expressed as either binding sites per cell or femtomoles prazosin bound per milligram protein for membrane fractions.
Data Analysis.
Binding isotherms were fitted to equilibrium fluorescence data and [3H]prazosin saturation data with GraphPad Prism software to determine the dissociation constants, fluorescent (FKD) and radioligand (KD), respectively. Inhibition of specific binding of [3H]prazosin by ligands was analyzed to estimate the IC50(concentration of the drug displacing 50% of specific binding). The inhibitory constant (Ki) was calculated from the IC50 by the equation of Cheng and Prusoff (1973). Binding isotherms from displacement-binding studies were analyzed by a nonlinear least square parametric curve fitting program GraphPad Prism, capable of iterative curve fitting to a single or two-site model. Student's t test was used to determine statistical significance between two groups.
Fluorescence Binding Studies
QAPB-Associated Fluorescence Saturation Binding Studies.
Recombinant cells or native prostate SMCs were grown on coverslips for 24 h before use under the above-mentioned culture conditions. Cultured cells and dissociated cells were treated in the same manner. Coverslips were mounted in a flow chamber and cells were washed three times with HEPES buffer (130 mM NaCl, 5 mM KCl, 20 mM HEPES, 10 mM glucose, 1 mM MgCl2, and 1 mM CaCl2). After baseline was achieved the first concentration of the fluorescent ligand QAPB was added and allowed to equilibrate for at least 5 min. Cumulative concentrations (0.4–10 nM) of QAPB were added to the bath and nonspecific binding was defined as binding in the presence of 10 μM phentolamine.
Competitive Inhibition of QAPB-Associated Fluorescence Binding.
Cells were mounted on a flow chamber bath as described previously and superfused (5 ml/min) with 20 mM HEPES with a peristaltic pump. Inhibition studies were performed with 5 nM QAPB in the absence of competitors for the measurement of total receptor binding. QAPB associated-fluorescence was washed to give a baseline signal and cells were preincubated in the presence of inhibitors for 20 min before addition of 5 nM QAPB. Images of total binding were subtracted from images in the presence of inhibitors with MetaMorph software to show specific receptor-binding sites.
Whole-Cell Image Analysis.
A Noran Odyssey real-time laser scanning confocal microscope was used in conjunction with a Nikon Diaphot (inverted) microscope. The 488-nm line (515 band pass) of an argon-ion laser was used throughout. To maximize the signal detection at very low concentrations of the fluorescent ligand a 50- to 100-μm slit was used in fluorescence-binding experiments. A Nikon 40× oil immersion objective (NA 1.3) was used throughout.
With cell autofluorescence a suitable group of cells was selected and the focal plane fixed by locking the focus motor. The system was then set to acquire images (64 frame averages; 2.56-s exposure) at 1-min intervals. Universal Imaging's MetaMorph software was used to select the cell area and intensity was measured over time for each cell.
Isosurface Models.
Serial confocal z-sections of QAPB (10 nM) binding to a freshly dissociated human prostate SMC was acquired at 0.5-μm steps with a 15-μm slit after a 10-min incubation period. The nuclear dye Syto 13 was added (1 μM) for 5 min at room temperature and washed three times with HEPES buffer before acquisition of an identical series of confocal z-sections. MetaMorph was used to threshold/segment either diffuse low-intensity fluorescence (intensities 5–80) or clusters of intracellular fluorescence (intensities 80–255) to create two masks. Black values 0 to 5 were considered background noise and eliminated from the data set. The masks were subtracted from the original data to reveal the fluorescence related to either the cell surface or intracellular sites.
The isosurface module of IMARIS (Bitplane AG, Zurich, Switzerland) three-dimensional image processing software was used to construct wire frame models of cellular components. Briefly, the full confocal data set was surveyed to determine the range of intensities (gray values) within the volume (0–255; black-white). An isovalue (intensity) was chosen and the software constructed a map of all the chosen values in the volume. The points were joined together with vectors that form a wire frame and a surface texture was applied with the opacity set to allow visualization of other surfaces. In this way, structures or varying intensities can be assigned different surfaces and colors and a range of different fluorescent-binding intensities can be depicted as a complex isosurface model. For the purposes of this article, we have chosen to show two isosurfaces that delineate the cell surface receptors (low isovalue, gray) or intracellular binding sites (high isovalue, yellow), respectively. A third isosurface created for the nucleus was added to the previous data set. More detailed visualization can be seen on our Web site athttp://www.cardiovascular.org
Results
Affinity of QAPB for Subtypes of Human α1-ARs Assessed versus [3H]Prazosin Binding to Membrane Preparations: Selectivity of Competitors for Subtype Identification
The affinity of the fluorescent ligand QAPB for the three known human recombinant α1-ARs was first assessed in standard membrane preparation assays versus [3H]prazosin. The affinities of ligands to be used in identifying subtypes in subsequent experiments also were assessed.
Radioligand Saturation Binding.
High-affinity, saturable-binding sites for [3H]prazosin, consistent with α1-ARs, were found on crude membranes prepared from cells expressing α1-AR subtypes (Fig.1a–c). Affinities calculated from saturation analysis were similar in all three subtypes (KD = 0.4–0.5 nM) with [3H]prazosin binding to a single class of binding site. Maximal receptor capacities at all of the three subtypes were approximately 1000 fmol/mg (Table1). In human prostatic crude membrane homogenates, [3H]prazosin bound to an apparent single class of binding sites with an affinity of 0.15 ± 0.04 nM and receptor density of 6.6 ± 0.48 fmol/mg of protein (Fig. 1d).
Competitive Inhibition of Radioligand Binding.
Consistent with the pharmacology of α1a-ARs, prazosin, RS100329, WB4101, YM12617, and 5-methylurapidil all displayed subnanomolar affinity in competition studies (Table2). The α1a-AR selective agonist (R)-A-61603 had 30- to 50-fold higher affinity at α1a-ARs versus α1b- or α1d-ARs. Competitive inhibition of [3H]prazosin by QAPB was similar for all subtypes, displaying up to 10-fold lower affinity compared with prazosin (Fig. 2). Although affinities at the α1a-AR subtype were found to be slightly higher than at the α1d-AR subtype, QAPB was a relatively nonselective competitive antagonist at α1-AR-binding sites. The only ligand found to discriminate between α1b-AR subtypes and the α1a- or α1d-subtypes was 5-methylurapidil, which displaced [3H]prazosin with a 5-fold lower affinity at the α1b-AR-binding site. Of the two drugs used clinically for benign prostatic hypertrophy, doxazosin was found to be nonselective between the three α1-AR subtypes, whereas tamsulosin (YM12617) had a 50-fold higher affinity at α1a-ARs versus α1b-ARs. The α1a-AR antagonist RS100329 demonstrated high affinity by displacing specific [3H]prazosin binding to α1a-AR membranes with a pKi of 9.7 and displaced binding to α1b-AR and α1d-AR membranes with low affinity (pKi = 8.0; Fig. 2). Conversely, BMY7378 displaced specific binding to α1d-AR membranes with a pKi of 10.5 compared with a pKi of 7.1 at α1a-ARs (Fig. 2).
Affinity of QAPB for Subtypes of Human α1-ARs Assessed versus [3H]Prazosin Binding to Intact Cell Preparations
The affinities for [3H]prazosin sites of QAPB were then assessed on whole-cell preparations harboring the three recombinant α1-ARs because this is the modality in which fluorescent binding to native receptors will be assessed.
Radioligand Saturation Binding.
In saturation experiments in whole cells expressing the three α1-AR subtypes [3H]prazosin binding was specific, saturable, and high affinity (Fig. 1). [3H]Prazosin bound to cells expressing α1a-, α1b-, and α1d-ARs with a single class of binding sites with dissociation constants 1.3 ± 0.9, 1.4 ± 0.2, and 1.4 ± 0.4 nM, respectively. Receptor number (Bmax) was also similar between the three subtypes with a mean of 6754 binding sites/cell (Table 1).
Fluorescent Ligand Saturation Binding on Intact Recombinant Cells
QAPB-associated fluorescence binding was both time- and concentration-dependent in recombinant cells and native cells, binding to a single class of α1-ARs (Fig.3). Incubation with QAPB produced minimal background fluorescence with a working range of 0.4 to 10 nM and fluorescence in nontransfected cells was negligible. In the presence of phentolamine (10 μM), QAPB binding was significantly inhibited and the residue was used to calculate nonspecific binding. Nonspecific binding was less than 10%, demonstrating high specificity for α1-ARs. Specific binding isotherms were similar for QAPB binding to whole cells for all three subtypes (Fig. 1, a–c). The calculated affinities for fluorescent ligand binding of QAPB to whole cells were within 4 to 5 times less than those for QAPB in competition with [3H]prazosin in radioligand-binding studies to membrane preparations and 2 to 3 times less than its affinities versus [3H]prazosin on whole cells (Table 1).
Saturation-binding isotherms show that all three α1-AR subtypes expressed similar QAPB-associated maximal fluorescence (75.9–77.7 arbitrary units), expressed as average intensity of the fluorophore. This correlates with [3H]prazosin binding (Fig. 1, a–c), which shows that comparable receptor numbers were expressed in all three recombinant cells (5987–6728 sites/cell) and membranes (878–951 fmol/mg of protein) prepared from these cells.
Diffuse fluorescence was detected mainly on the plasma membrane at low QAPB concentrations (<2 nM) in all three α1-AR subtypes. Intracellular fluorescence was detected at higher concentrations, localized to clusters of saturable intensity around the nuclear membrane. This was particularly apparent in the α1b-AR subtype (Fig. 3). The nonspecific binding was predominantly diffuse and located on the plasma membrane.
Fluorescent Ligand Saturation Binding on Freshly Dissociated Human Prostate SMCs
QAPB bound to a single class of α1-ARs on smooth muscle cells dissociated from the human prostate (Fig. 1d). QAPB-associated fluorescence binding was saturable, of high affinity (FKD = 0.63 ± 0.02 nM), and up to 10 nM was almost completely abolished in the presence of 10 μM phentolamine (Fig. 3h). The fluorescent affinity for QAPB binding to the native cell was 4 times less than calculated from radioligand-binding studies on crude membrane homogenates prepared from human prostatic “chips” (KD = 0.15 ± 0.04 nM; Fig. 1d). This lower affinity by fluorescence measurement is similar to the equivalent relationship found in recombinant cells, i.e., comparing radioligand binding to membranes with fluorescent ligand binding to cells. However, the absolute values of the affinities by either fluorescence or radioligand binding in the human prostate SMC were almost 4 times higher than on recombinant cells expressing α1a-ARs.
Identification of Native α1-AR Subtype on Cultured Prostate Smooth Muscle Cells by Competitive Inhibition of QAPB-Associated Fluorescence
Cultured prostate SMCs at passage 3 bound QAPB in a concentration-dependent manner and fluorescence was inhibited in the presence of 10 μM phentolamine (Fig. 4). In these cultured cells QAPB affinity (FKD = 0.8 ± 0.2 nM) was similar to freshly dissociated cells (Table 1).
Cultured prostatic SMCs were equilibrated at 5 nM QAPB. Fluorescence developed with single-phase kinetics reaching a plateau after a 30-min incubation (Fig. 4a). Fluorescence was washed off followed by preincubation with 1 nM RS100329 (Ki = 0.22 ± 0.07 nM). On addition of 5 nM QAPB the reappearance of fluorescence was significantly inhibited (Fig. 4b) and image subtraction revealed specific binding sites (Fig. 4c) both on the plasmalemmal membrane and in intracellular clusters around the nuclear membrane. After competitive inhibition of QAPB-associated fluorescence, cells were stained with a cy3 conjugated α-actin monoclonal antibody to verify SMC origin (Fig. 4d).
Subcellular Distribution: Three-Dimensional Visualization of α1-ARs
Equilibration of prostatic SMCs with low concentrations of QAPB (<2 nM) revealed fluorescence predominantly associated with the cell membrane. As with the recombinant cells, higher concentrations identified intracellular binding sites around the nucleus (Fig. 3).
The subcellular distribution of binding sites in a prostate SMC was visualized in two dimensions in Fig. 5a. An extended focus was created from the original data set. A mask was prepared by thresholding all the pixels with a gray scale value between 5 and 80 (Fig. 5b; red) and shows the fluorescence mainly associated with the cell surface (Fig. 5c). A second mask was created of pixels with gray scale values between 80 and 255 (Fig. 5d; red) and this mask was used to highlight the pixels in the original data set showing only the clustered intracellular fluorescence (Fig. 5e). These two data sets were then used to create the isosurface model in Fig. 5, f and g.
To determine the subcellular location of α1-ARs in a native human SMC a series of confocal “optical” sections was captured once a steady state between QAPB and the receptor had been achieved. The construction of a three-channel isosurface model offers easy visualization and interaction with the data set. In Fig. 5f, QAPB-associated fluorescence (10 nM) is visualized in one color for that near to the cell membrane (Fig. 5f; gray) and in another color for that which is separated from this membrane-bound domain and located inside the cell (yellow). A third color is used to show the position of the cell nucleus (red), which had been stained with a nuclear dye. In Fig. 5f, the surface receptors have been rendered partly transparent to allow sight of the intracellular receptors and nucleus. This shows that there is a covering of binding sites on the plasma membrane but distinct clusters inside the cell, particularly two large regions in proximity to the nucleus. Figure 5g shows another view of the whole cell in three dimensions with the outer membrane layer sliced to allow full visualization of the intracellular binding sites and the nucleus.
The total volume of the prostate SMC was calculated with IMARIS software and was plotted against gray scale intensity values (0–255) where zero is black and 255 is white (Fig. 5h). The number of voxels gives a measure of the volume of the cell and its components and the integrated fluorescence from the three-dimensional isomodel was calculated as the sum of the intensities multiplied by the number of voxels in the data set. The surface and associated regions of the cell outside of the intracellular compartments constituted 86% of the total volume and contained 60% of the fluorescence (Fig. 5c; intensities 5–80). The intracellular component (Fig. 5e; intensities 80–255) was 14% of the total cell volume and contained 40% of the total cell fluorescence. A second peak of high intensity was observed between gray scale values 160 and 200, indicating a cluster of very high intensity (Fig. 5a; yellow-red).
Discussion
It was our goal to show the spatial distribution of α1-ARs and their binding properties in subcellular regions of live, single native cells. In rat-1 fibroblasts expressing cloned human α1-ARs the affinity of QAPB versus [3H]prazosin was similar across the subtypes with 2- to 8-fold lower affinity than prazosin (Fig. 1). This confirmed that QAPB is a high-affinity nonsubtype-selective competitive antagonist at α1-AR-binding sites.
Saturation binding in live cells QAPB measured by fluorescence compared with [3H]prazosin binding produced similar results in all three clones (Table 1). The maximal receptor capacity (Bmax) judged by [3H]prazosin binding to crude membrane fractions or whole cells and maximal fluorescence binding were each similar for the three subtypes in agreement with the findings ofSchwinn et al. (1990). The consistency of fluorescence and radioligand binding on whole cells opens the possibility of expressing number of sites per unit of fluorescence.
This principle was explored in a heterogeneous native tissue. QAPB-associated fluorescent binding demonstrated specific, high-affinity α1-ARs were present on human prostate SMCs, both freshly dissociated and in culture up to three passages (Table 1). Cultured cells were used for quantitative, competitive analysis of the subtype because this provides a homogeneous population of stable cells, allowing duplication of prolonged protocols.
The compound that discriminated best between the α1a-AR and the other two subtypes was RS100329 (Fig. 2). RS100329 at 1 nM is a concentration 5 times higher than its inhibitory constant versus α1a-ARs but it is still 10 times lower than versus α1b-ARs or versus α1d-ARs and thus should displace only α1a-ARs. This concentration of RS100329 inhibited approximately 50% of fluorescence from both the plasma membrane and intracellular areas of high density (Fig. 4c) around the nuclear membrane and was consistent with a concentration-dependent competitive inhibition. The potency of RS100329 showed that receptors in human prostate smooth muscle have similar competitive ligand-binding pharmacology to recombinant α1a-ARs. According to the International Union of Pharmacology Committee on Receptor Nomenclature they should be classified as the native functional equivalent, α1A-ARs (Hieble et al., 1995).
Classical radioligand assays use membrane preparations that eliminate some nonspecific binding but they do not reflect the true binding of ligands to receptors on intact, live cells nor deal with tissue heterogeneity. The fluorescent ligand approach is applicable to cells with low receptor number or to cells from heterogeneous populations, allowing an estimation of both affinities and receptor densities for receptor subtypes on single, live cells.
Prostatic tissue illustrates the insight provided by fluorescence binding. The maximal receptor number by radioligand binding on membrane homogenates from prostate tissue is low (Bmax = 6.6 ± 0.48 fmol/mg of protein) compared with membranes from recombinant cells (meanBmax 911 fmol/mg of protein). This is due to dilution of the SMCs that express the receptors, rather than to a low receptor density on relevant cells. The fluorescent ligand method indicates the level of receptor expression on the cells of interest and also provides information on their subcellular distribution.
Previous information on α1-AR subcellular localization was available only for recombinant receptors because it relied on epitope or green fluorescent protein tags (Fonseca et al., 1995; Hirasawa et al., 1996, 1997). This led to the hypothesis that differences in receptor distribution might allow subtypes to be distinguished. Hirasawa et al. (1997) suggested that, in COS-7, the recombinant α1a-AR had a relatively greater proportion located intracellularly than did the α1b-AR. In this study, QAPB bound to cell membrane and intracellular perinuclear sites in all three recombinant subtypes expressed in rat-1-fibroblasts although it was more obvious around the perinuclear area of the cell in the α1b-AR subtype. This provides another example of different subtypes adopting different locations in the same host cell. However, there was no general rule as to the localization of particular subtypes.
Remarkable similarities are observed between the intracellular distribution of binding sites in recombinant and native systems. We found that in prostatic SMCs 40% of specific binding sites were intracellular, particularly around the nucleus, which could represent binding in the Golgi and may include both newly synthesized and recycling stores of receptors. This demonstrates that the considerable quantity of intracellular binding sites previously shown in recombinant cells is not an artifact of artificial systems but a natural phenomenon.
The ability of an antagonist ligand to bind competitively to an intracellular site in live cells, demonstrated herein for the first time, raises the possibility that they are involved in signal transduction. Agonists also may bind to the receptor and activate it. Other G-protein-coupled receptors (EP3 and EP4) normally thought to be localized to the plasma membrane have been shown to be activated in the nuclear envelope by their natural ligand prostaglandin E2 (Smith et al., 1999). Our finding that α1-AR subtypes bind ligands in intracellular domains suggests that this might be a general property of heptahelical receptors.
For drug action this finding has several important implications. If intracellular receptors transduce signals then activators or blockers may act at these differentially according to whether they can penetrate the cells, either by diffusion/lipophilicity or whether they are substrates for transmembrane carriers; this may explain the previously unknown function of many carriers for neuroendocrines. Noradrenaline is taken up and accumulated by SMCs that express α1-ARs (Avakian and Gillespie, 1968). The transporter responsible, the extraneuronal monamine transporter, formerly known as (Uptake2), has been cloned (Gründemann et al., 1998). The physiological basis for the transporter has never been established although, when it is blocked, the metabolism of noradrenaline is attenuated, which has given rise to the general concept that the purpose of the carrier is to eliminate noradrenaline from the extracellular space. The presence of intracellular α1-ARs capable of binding ligands opens the possibility that a function of the transporter is to deliver catecholamines to intracellular receptors. Because the transporter is regulated by corticosteroids this may point to this function involving the regulation of cell metabolism and, possibly, cell growth.
Irrespective of whether the intracellular receptors are active in signal transduction they may act as a buffer for bound ligands, consequently affecting the kinetics of binding to extracellular receptors by altering ligand bioavailability. This could produce distortion in pharmacological analysis as well as a new basis for patterns of drug selectivity.
In this study a number of important cross-correlations emerged that explain some controversial issues in the literature. First, there is higher affinity, by a factor of 2 to 3, for membrane preparations compared with cells, namely, [3H]prazosin binding for recombinant α1a-AR membranes versus cells. Second, the human prostate had approximately 3 times higher affinity than did the recombinant α1a-ARs measured either by [3H]prazosin on membranes or QAPB fluorescence on cells. Because competition analysis suggested that the prostate receptors were α1a-ARs and they had the highest affinity of the three subtypes for QAPB, we cannot escape the conclusion that prostate-binding sites have higher affinity than recombinant receptors.
Taking into consideration that membranes have twice the affinity of cells and that prazosin has twice the affinity of QAPB at α1a-ARs, this is all consistent with the prostatic membranes having the highest affinity in the study for [3H]prazosin binding and a value of 5 to 6 times that for QAPB at prostatic cells. Therefore, there is an internal consistency in the data that shows that affinity is generally greater in membrane preparations than in cells but also is greater in the native prostatic tissues or cells compared with recombinant receptors. This clears up quantitative discrepancies in the literature in binding data from cloned whole cells, membranes, and membranes prepared from native prostate tissues (Kenny et al., 1994; Takeda et al., 1997;Williams et al., 1999). It also corroborates, for native prostatic tissue, the hypothesis of Ford et al. (1998) that receptors expressed in cells show lower binding affinities than those in membrane preparations. This contributes to the debate on whether receptors should be defined as a different subtype (α1H-AR or α1L-AR) on the basis of prazosin affinity, supporting the contention of Ford et al. (1998) that absolute affinity can vary according to the system in which it is measured and that, under the appropriate conditions, α1L-AR are indistinguishable from α1A-AR.
The higher absolute values for binding to native prostatic tissue membranes or cells indicate that there is a particular factor that determines relatively high affinity. This hypothesis requires further analysis in other native tissues to tell whether it is a feature of all native tissues, per se, or is unique to the prostate. If the latter is true it could be extremely important in directing drug selectivity. There has not previously been an adequate explanation for why α-blockers should be effective against adrenergic stimuli to prostate without causing postural hypotension due to blockade of vascular α1-ARs, which have similar receptor subtypes (Marshall et al., 1995; Ford et al., 1996, 1998; McGrath et al., 1996). A cell-type-specific modification of affinity could be the explanation. This feature could not have been detected without single-cell quantitative fluorescence binding. The molecular basis for the modification is not yet known but seems to indicate a cell-specific post-translational modification or associated factor, which enhances binding in the native receptors.
Acknowledgments
We acknowledge the excellent technical support of Ruth Murdoch and Dr. Ian Montgomery. Human tissue was kindly provided by Fletcher Dean and Professor David Kirk, Gartnavel General Hospital, Glasgow. We are grateful to Dr. Debra Schwinn, Duke University, North Carolina, for the gift of cells harboring recombinant α1-AR subtypes and for generous advice.
Footnotes
-
Send reprint requests to: Dr. J. F. Mackenzie, Autonomic Physiology Unit, Division of Neuroscience & Biomedical Systems, Institute of Biomedical & Life Sciences, West Medical Bldg., University of Glasgow, Glasgow, G12 8QQ, UK. E-mail:j.f.mckenzie{at}bio.gla.ac.uk
-
↵1 This study was supported by Pfizer, Medical Research Council, and the Prostate Research Campaign, United Kingdom.
- Abbreviations:
- AR
- adrenoceptor
- SMC
- smooth muscle cell
- QAPB
- BODIPY-FL prazosin
- FKD
- fluorescentKD
- Received January 4, 2000.
- Accepted April 18, 2000.
- The American Society for Pharmacology and Experimental Therapeutics