Abstract
The antagonist ligand BODIPY-FL-prazosin (QAPB) fluoresces when bound to bovine α1a-adrenoceptors (ARs). Data indicate that the receptor-ligand complex is spontaneously internalized by β-arrestin-dependent endocytosis. Internalization of the ligand did not occur in β-arrestin-deficient cells; was blocked or reversed by another α1 ligand, phentolamine, indicating it to reflect binding to the orthosteric recognition site; and was prevented by blocking clathrin-mediated endocytosis. The ligand showed rapid, diffuse, low-intensity, surface binding, superseded by punctate intracellular binding that developed to equilibrium in 50 to 60 min and was reversible on ligand removal, indicating a dynamic equilibrium. In cells expressing a human α1a-AR-enhanced green fluorescent protein (EGFP) 2 fusion protein, BODIPY-R-558/568-prazosin (RQAPB) colocalized with the fusion, indicating that the ligand gained access to all compartments containing the receptor, and, conversely, that the receptor has affinity for the ligand at all of these sites. The distribution of QAPB binding sites was similar for receptors with or without EGFP2, validating the fusion protein as an indicator of receptor location. The ligand partially colocalized with β-arrestin in recycling and late endosomes, indicating receptor transit without destruction. Organelles containing receptors showed considerable movement consistent with a transportation function. This was absent in β-arrestin-deficient cells, indicating that both constitutive receptor internalization and subsequent intracellular transportation are β-arrestin-dependent. Calculations of relative receptor number indicate that at steady state, less than 30% of receptors reside on the cell surface and that recycling is rapid. We conclude that α1a-ARs recycle rapidly by an agonist-independent, constitutive, β-arrestin-dependent process and that this can transport “α-blockers” into cells carrying these receptors.
We have observed in live smooth muscle cells expressing α1-ARs that a fluorescent antagonist ligand, the prazosin analog QAPB, labels intracellular receptors. This process takes longer than expected of simple diffusion, suggesting an uptake process (Mackenzie et al., 2000). It is known that fluorescent agonists can be internalized while bound to the receptors that they activate (Kallal and Benovic, 2000, 2002; Daly and McGrath, 2003). It is also possible that antagonists of receptor-mediated second messenger regulation can initiate internalization, as shown previously for the cholecystokinin receptor (Roettger et al., 1997). However, an antagonist ligand might be internalized simply by binding to a receptor that is spontaneously recycling. Morris et al. (2004) recently showed that when expressed in rat-1 fibroblasts (R-1Fs), the human α1a-AR can internalize spontaneously by using antibodies to an extracellular epitope to follow the receptors, excluding the possibility of internalization through occupation of the receptive site.
We now show that antagonists such as QAPB can be internalized bound to spontaneously recycling α1a-ARs and analyze the mechanisms involved. QAPB is a prazosin analog in which the furan ring has been substituted by a BODIPY fluorescent moiety. It retains prazosin's high affinity, functional antagonism, and relative selectivity for α1-ARs, and, like prazosin, does not distinguish within the family of three α1-AR subtypes. Importantly for its present use, it fluoresces when bound to the receptor in live cells, yet it remains essentially nonfluorescent in the aqueous phase. This allows quantitative visual assessment of its binding to receptors at equilibrium (Daly et al., 1998). It has been used to measure binding affinity at recombinant and native α1-ARs in different regions of cultured cells and of freshly dissociated cells from human and other mammalian tissue (McGrath et al., 1996, 1999; Daly et al., 1998; Mackenzie et al., 2000). We have quantified the pharmacology of agonist and antagonist interaction at recombinant α1-AR in R-1Fs (Pediani et al., 2000), demonstrating that the receptors are functional and have the predicted pharmacology. This cell line provides the main basis for the present study. We accomplished real-time visualization of antagonist-receptor binding interactions in a concentration range near to the ligand's dissociation constant. Binding was found at intracellular sites as well as on the cell surface. We further established the utility of the fluorescent ligand by demonstrating colocalization of the QAPB-receptor complex and receptors tagged with the autofluorescent green fluorescent protein variant EGFP2. These independent measures of receptor location allowed us to cross-validate the fluorescent ligand as reflecting the localization of receptor binding sites and the EGFP2 as indicating the presence of receptors that could bind antagonists. We investigated the mechanism by which the ligand entered the cell by blocking internalization processes and used vital organelle markers to identify the compartments in which the α1a-AR-ligand fluorescent complex resides. Analysis of the contribution of β-arrestin indicated that α1a-AR spontaneously recycle in a β-arrestin-dependent but agonist-independent manner. The rate of uptake and cellular distribution of the ligand-receptor complex were used to estimate the turnover of the spontaneous recycling process.
Materials and Methods
Materials
Cell culture plastics were supplied by Falcon (BD Biosciences Discovery Labware, Bedford, MA) or PerkinElmer Life and Analytical Sciences (Boston, MA). Dulbecco's modified Eagle's medium (DMEM) with sodium pyruvate, newborn calf serum, l-glutamine, penicillin, streptomycin, active G418 (Geneticin; 50 mg/ml), trypsin/ethylenediamine tetraacetic acid, pCDNA3, and Dulbecco's phosphate-buffered saline (PBS) were purchased from Invitrogen (Paisley, Scotland). The pEGFP2 N2 vector was purchased from PerkinElmer Life and Analytical Sciences (Berkshire, UK). HEPES, sucrose, phenylephrine, phentolamine, concanavalin A (Con A), and filipin were purchased from Sigma Chemical (Poole, Dorset, UK). Prazosin HCl was supplied by Sigma/RBI (Natick, MA), and [3H]prazosin (0.2 nM; specific activity 76 Ci/mmol) was from Amersham Biosciences Inc. (Piscataway, NJ). Hoechst 33342, BODIPY-FL-labeled prazosin (QAPB), BODIPY-R 558/568-labeled prazosin (RQAPB), Lysotracker Red DND-99, and transferrin (Tfn)-Alexa Fluor546 were purchased from Molecular Probes (Eugene, OR). All optical fluorescence filters were purchased from Chroma Technology Corp. (Rockingham, VT).
Stock solutions for each chemical were prepared in distilled water or dimethyl sulfoxide and subsequently aliquoted and stored at -20°C. These stock solutions were diluted to working concentrations in physiological salt solution on each experimental day.
Construction of Plasmids
Production and subcloning of α1a-AR-EGFP2 involved polymerase chain reaction amplification of the human α1a-AR sequence using the amino-terminal primer 5′-AAA AGG TAC CAT GGT GTT TCT CTC GGG AAA TGC TTC-3′ to introduce a KpnI restriction sequence upstream of the coding sequence. Using the carboxy-terminal primer 5′-AAA AGG ATC CGA CTT CCT CCC CGT TCT CAC TGA GGG-3′, the receptor stop codon was removed, and a BamH1 restriction enzyme site was introduced downstream of the receptor coding sequence. The resulting polymerase chain reaction fragments were then ligated either into the multiple cloning site of the pEGFP2 N2 vector (PerkinElmer Life and Analytical Sciences) or the expression vector pcDNA3 (Invitrogen, Carlsbad, CA) to form the EGFP2-tagged and non-EGFP2-labeled version of the human α1a-AR, respectively. Each construct was fully sequenced before its expression and analysis. β-Arrestin-2-EGFP was obtained from Dr. C. Krasel (Department of Pharmacology, University of Würzburg, Germany).
Cell Culture and Transfection
R-1Fs stably expressing the bovine α1a-AR were grown in DMEM supplemented with 5% (v/v) newborn calf serum, penicillin (100 IU/ml), streptomycin (100 μg/ml), and l-glutamine (1 mM) in a 95% air and 5% CO2 atmosphere at 37°C. Selection was maintained by adding G418 (400 μg/ml) to the growth media.
HEK293T cells were maintained in DMEM supplemented with 0.292 g/l l-glutamine and 10% (v/v) newborn calf serum and incubated at 37°C with 5% CO2. HEK293T cells were grown on glass coverslips to ∼70% confluence before transient transfection with the appropriate human α1a-AR plasmid using LipofectAMINE reagent (Invitrogen). After 4 h, cells were washed twice with Opti-MEM I and cultured in DMEM for 24 h. A total of 3 μg of pEGFP2 N2 vector or pCDNA3 containing the human α1a-AR cDNA was used to transfect each coverslip. Mouse embryo fibroblasts (MEFs) derived from both wild-type and β-arrestin 1 plus β-arrestin 2 knockout animals have been described previously (Kohout et al., 2001) and were obtained from Dr. R. J. Lefkowitz (Duke University, Durham, NC). These were maintained in DMEM supplemented with 0.292 g/l l-glutamine and 10% (v/v) fetal calf serum at 37°C in a 5% CO2 environment. Cells were grown to 60 to 80% confluence before transient transfection on sterile glass coverslips. Transfection (3 μg of cDNA) was carried out using the Amaxa nucleofector system (Cologne, Germany) according to the manufacturer's instructions using program T20.
Radioligand Binding
Competitive α1-AR binding assays were performed as described previously (MacKenzie et al., 2000). In brief, R-1F homogenates (0.05 mg/ml) stably expressing the bovine α1a-AR were incubated in triplicate with the nonselective α1-AR antagonist [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 Tris-HCl assay buffer. Nonspecific binding was determined as radioligand binding in the presence of 10 μM phentolamine. All equilibrations were carried out for 30 min at 25°C, and bound ligand was separated from free ligand by rapid cold vacuum filtration over Whatman GF/C filters (Whatman, Maidstone, UK) using a Brandel cell harvester. Concentrations of displacing agent eliciting 50% displacement of [3H]prazosin (IC50) were interpolated with the use of nonlinear iterative curve-fitting methodologies performed by Prism 4.02 (GraphPad Software Inc., San Diego, CA) and converted into pKi values with the equation of Cheng and Prussoff (1973).
Fluorescence Experiments Undertaken in R-1Fs
Epifluorescence Imaging of QAPB Binding to Spontaneously Recycling Bovine α1a-ARs. R-1Fs stably expressing the bovine α1a-AR were grown on sterile coverslips 24 h before experimentation. Coverslips were mounted into a flow chamber, placed on to the microscope stage, and superfused (5 ml/min) with HEPES-buffered saline solution at room temperature. Cells were illuminated using an ultrahigh-point intensity 75-W xenon arc lamp (Optosource, Cairn Research, Faversham, Kent, UK). A Nikon Diaphot inverted microscope equipped with a 40× oil immersion Fluor objective lens (1.3 numerical aperture) and a monochromator (Optoscan) set at an excitation wavelength of 490 nm (band pass 7 nm) was used to determine the background autofluorescence. After acquiring the background autofluorescence, the cells were superfused with fresh saline supplemented with the fluorescent antagonist ligand QAPB (5 nM) until equilibrium binding of QAPB occurred. QAPB fluorescence emission was detected by a cooled digital charge-coupled device camera (Cool Snap-HQ; Photometrics, Tucson, AZ). MetaFluor imaging software (version 4.6.9; Universal Imaging Corporation, Downingtown, PA) was used for control of the monochromator and chargecoupled device camera and for processing of all the cell image data and movie animations.
Sequential QAPB images (2 × 2 binning) were collected at 5-s intervals to determine the total QAPB fluorescence intensity associated with each cell, and exposure to excitation light was 60 ms/image. QAPB binding to α1a-ARs was time dependently reversed by washing with QAPB ligand-free saline solution, with or without 100 μM phentolamine (the latter was incorporated to prevent rebinding of dissociating ligand and therefore to approximate infinite dilution conditions).
Quantification of the Fraction of QAPB Ligand-Receptor Complexes That Are Intracellular versus Surface at Steady State. Receptor expression at the cell surface (i.e., cell membrane, clathrin-coated pits) was defined as the rapid, initial, diffuse QAPB fluorescence that occurs after ligand addition but before the fluorescent ligand-receptor-containing endocytic vesicles can be visually resolved. Subtraction of the cell surface fluorescence from the total fluorescence allowed us to determine the intracellular fluorescence, and the maximal intracellular fluorescence was normalized to the maximal total QAPB fluorescence measured at steady state.
The rate of development of QAPB-associated or -dissociated fluorescence was determined by fitting one-phase exponential association or dissociation equations (via nonlinear regression), to each data point (GraphPad Software Inc.). The reciprocal of the rate constant was used to determine t1/2 values.
Reversal of QAPB Fluorescence Binding to α1a-ARs by the Competitive α1-Antagonist Ligand Phentolamine. Spontaneously recycling α1a-ARs were pre-equilibrated with QAPB (5 nM) for 75 min at room temperature before an image was acquired to determine the total QAPB fluorescence intensity associated with each R-1F. Sequential QAPB images (2 × 2 binning, 60 ms/image) were collected at 15-s intervals for a further 15 min to determine that equilibrium binding of QAPB had taken place before superfusion of the cells with saline containing QAPB (5 nM) plus phentolamine (10 μM). R-1Fs were equilibrated with this solution until the ligands QAPB and phentolamine had reached their new steady-state receptor occupancies. The remaining QAPB-associated fluorescence was then time-dependently removed by exposing the cells to saline containing the same concentration of QAPB (5 nM) plus a higher concentration of phentolamine (100 μM).
Using the editing capability of MetaFluor, QAPB intensity values derived from groups of R-1Fs were obtained by manually delimiting the profile of the cells and averaging the signals within the delimited region of interest for each time-lapse image captured. These averaged QAPB intensity values were exported into Prism 4.02 (GraphPad Software Inc.) and plotted over time.
Histogram Intensity Plots. Binned pixel intensities (8-bit image of 256 intensities divided into nine intensity bins) of fluorescence histograms representative of the localization of QAPB-labeled α1a-ARs (over time) were computed using the histogram plot profile tool of MetaMorph 6.2.6 software (Universal Imaging Corporation). Background-subtracted image acquired at 0 s was determined as cell autofluorescence, and the pixel intensity of membrane QAPB fluorescence was calculated as the diffuse fluorescence that was observed after 2-min exposure to the ligand. To remove contributions of the plasma membrane signal into the intracellular QAPB signal, the diffuse fluorescence observed at 2 min was subtracted from all subsequently acquired images.
Binned integral counts acquired at various time intervals after addition or removal of the QAPB ligand were measured and expressed as intensity integral counts. QAPB binding to α1a-ARs was detected as a shift to the right of the autofluorescence plot (at 0 s), and the quantified histogram data were used for comparison of intensity in various structures within the image data. Loss of QAPB from cycling α1a-ARs was detected as a progressive shift to the left of the fluorescence intensity measured at equilibrium.
Image Analysis Colocalization
For the analysis of RQAPB-labeled bovine α1a-ARs with β-arrestin-2-EGFP2, a region of no fluorescence adjacent to the cell was used to determine the average background level of fluorescence in the EGFP2 and RQAPB channels. The background amount was then subtracted from each pixel in each channel. Using the editing capability of MetaFluor, QAPB intensity values derived from groups of R-1Fs were obtained by manually delimiting the profile of the cells. Using the editing function a region of interest was manually drawn in the EGFP2 channel image, then selected and transferred to the matched image acquired in the RQAPB channel. Correlation coefficients were evaluated by comparing the amounts of fluorescence measured in each matched pixel of the two different channels using the MetaMorph “correlation plot” application.
Blockade and Reversal of QAPB Fluorescence Binding to Recycling α1a-ARs by Concanavalin A and Hyperosmotic Sucrose, but Not Filipin. To study the role of clathrin-coated vesicles and caveolae in α1a-AR endocytosis, the effects of hyperosmotic sucrose, Con A, and filipin on QAPB binding to recycling α1a-ARs were assessed as follows. R-1Fs were superfused with saline, and a background autofluorescence image was acquired. R-1Fs were then preequilibrated with QAPB (5 nM) for 90 min at room temperature, which was sufficient time for QAPB binding to α1a-ARs to reach equilibrium. An image was then acquired to determine the total QAPB fluorescence intensity. QAPB fluorescence was then completely removed by superfusing the cells with ligand-free saline for 90 min, and a background autofluorescence image was acquired. The R-1Fs were re-exposed to control saline or saline supplemented with Con A (0.25 mg/ml), hyperosmotic sucrose (0.45 M), or filipin (5 μg/ml) for 30 min before re-exposing the cells to QAPB (5 nM) in the continued presence of each individual inhibitor. QAPB intensity values were quantified by manually delimiting the profile of the fibroblast cells and averaging the QAPB fluorescence intensity within the delimited region of interest. All values of QAPB fluorescence intensity were corrected for background autofluorescence and the equilibrium QAPB intensity on this first exposure was normalized to 1. On a second exposure to QAPB, with or without test drugs, fluorescence intensity was measured at 30, 60, and 90 min and normalized to the “first exposure” value. Normalized data from each experimental group were pooled and are presented as mean ± S.E.M. Statistical significance between the means was evaluated using a Student's paired t test.
Multiple Fluorescent Labeling of Recycling and Late Endocytic Vesicles. R-1Fs seeded on to coverslips were rinsed with PBS and then exposed at room temperature to PBS containing either 10 nM QAPB and the recycling endosomal marker Tfn-Alexa Flu- or546 (20 μg/ml) or the late endosomal/lysomal marker Lysotracker Red DND-99 (150 nM). Cells were pulse-labeled with Tfn-Alexa Fluor546 (37°C) or Lysotracker Red (room temperature) plus QAPB (10 nM) for 20 and 70 min, respectively. At the end of each pulselabeling period, cells were rinsed with PBS supplemented only with QAPB (10 nM) and then bathed in this solution for a further 15 min. Approximately 10 min before image acquisition, cells were exposed to PBS containing 10 nM QAPB plus the nuclear DNA-binding dye Hoechst 33342 (10 μg/ml; 10-min incubation at room temperature) to stain cell nuclei. Each coverslip was then washed and bathed in QAPB containing PBS before being imaged. The dyes Hoechst 33342, QAPB, Tfn-Alexa Fluor546, and Lysotracker Red were sequentially excited using the appropriate dichroics/filters to prevent bleed through, and the resultant images were overlaid using MetaMorph software (version 6.2.6).
3D visualization of the colocalization of QAPB ligand-bovine α1a-AR complexes with Tfn-Alexa Fluor546 or Lysotracker Red was achieved by using a Nikon TE2000-E inverted microscope equipped with a z-axis linear encoded stepper motor. Each dye was sequentially excited, and a z-series of images were acquired at 0.22-μm steps to produce individual z-stacks. The z-stack images were then merged and deconvoluted using an iterative and constrained algorithm (Autodeblur software, version 9.3.4; Autoquant Imaging, Watervliet, NY). Triplicate 3D (x-y, x-z, and y-z) maximum projection views were constructed using Autovisualize software (Autoquant Imaging).
Fluorescence Experiments Undertaken in Wild-Type and β-Arrestin Double Knockout Mouse Embryo Fibroblasts. Twenty-four hours after transfection with the α1a-AR-EGFP2 fusion protein, cells were observed under oil immersion, and fluorescence images were acquired at different time intervals in the absence or presence of the RQAPB ligand. During image acquisition, cells were maintained by superfusing them with HEPES-buffered saline solution. EGFP2 and RQAPB were excited using the appropriate wavelength (490 nm for EGFP and 545 nm for RQAPB) and emitter (BP505-545 for EGFP and BP585-625 for RQAPB (Chroma Technology Corp.). 3D imaging was undertaken as described previously, and 3D volumetric measurements of the proportion of surface EGFP2-tagged receptors binding RQAPB (relative to the total EGFP2-tagged α1a-ARs) was determined using Autodeblur software, version 9.3.4 (Autoquant Imaging).
Fluorescence Experiments Undertaken in HEK293T Cells
Dual Hoechst and QAPB Epifluorescence Imaging. HEK293T cells transiently transfected with the non-EGFP2-tagged version of the human α1a-AR were rinsed several times in PBS and then exposed to QAPB (10 nM) for 70 min at room temperature. Cell nuclei were stained before visualization. QAPB and Hoechst were sequentially excited and detected as detailed previously. Sequential images (no binning) were collected at 15-s intervals.
Dual EGFP2 and RQAPB Confocal Imaging. Cells were plated onto glass coverslips (22 mm), and after a 24-h growth period, they were transiently transfected with the EGFP2-tagged version of the human α1a-AR. Transfected cells were cultured overnight, and 24 h later, growth medium was removed and replaced with fresh PBS. Transfected cells were then pre-equilibrated with fresh PBS containing 10 nM RQAPB (a red fluorescent variant of the antagonist ligand QAPB) for 75 min. The RQAPB-treated cells were mounted onto an imaging chamber, and a Zeiss 510 PASCAL laser scanning confocal inverted microscope equipped with a 63× oil immersion Plan Fluor Apochromat objective lens (1.4 numerical aperture) was used to visualize each fluorescent probe. Using the appropriate laser lines (i.e., 488 nm for EGFP2 and 543 nm for RQAPB) and emission filters (i.e., BP505-530 for EGFP2 and LP570 for RQAPB), sequential images were acquired to determine the total EGFP2 and RQAPB fluorescence emission intensity associated with each transfected cell. Overlay images were created as outlined previously.
Solutions. HEPES-buffered saline solution contained 130 mM NaCl, 5 mM KCl, 1 mM CaCl2, 1 mM MgCl2, 20 mM HEPES, and 10 mM d-glucose, pH adjusted to 7.4 using NaOH. Tris-HCl assay buffer composition was 150 mM NaCl, 50 mM Tris-HCl, 10 mM MgCl2, and 5 mM EDTA, pH 7.4.
Results
Affinities of Compounds Measured by [3H]Prazosin Competition Binding
Binding affinity estimates for the ligands used in this study were determined in R-1Fs stably expressing the bovine α1a-AR to set the quantitative context for visualization of ligand-ligand competition. Competition curves in membrane preparations between [3H]prazosin and both prazosin and its fluorescent analog QAPB were monophasic and exhibited Hill coefficients that were close to unity (Fig. 1). Prazosin and QAPB displayed subnanomolar affinity for the receptor. The pKi value of phentolamine was 8.5, and this ligand was used at high concentration (10 μM) to determine nonspecific binding.
Direct Visualization of the Fluorescent Ligand-α1a-AR Complex
Does Fluorescent Ligand Binding Coincide with Receptors? Nontransfected HEK293T cells showed no fluorescence on exposure to QAPB (10 nM) for 75 min (Fig. 2A, i and ii). Similar exposure of cells transfected with the α1a-AR produced fluorescence predominantly in intracellular punctate vesicles, with a lower level reflecting cell surface localization (Fig. 2B). This is consistent with the concept that unbound QAPB does not fluoresce and that binding to the α1a-AR occurs with high affinity. In cells transfected with the α1a-AR-EGFP2 fusion protein, EGFP2 autofluorescence had a similar distribution (Fig. 2C, i) to QAPB in α1a-AR (no EGFP2) cells (compare with Fig. 2B). In the α1a-AR-EGFP2-expressing cells, exposure to the red version of QAPB produced red fluorescence (Fig. 2C, ii) that colocalized with the green fluorescence of EGFP2 (Fig. 2C, iii). This indicates that 1) on incubation with live cells, QAPB binds to, and thus indicates the location of, all the α1a-AR population; 2) the ligand gains access to all compartments containing the EGFP2-labeled α1a-AR; 3) the receptor has affinity for the ligand at all of these sites (because unbound ligand does not fluoresce); 4) the EGFP2 tag does not alter receptor distribution or the movement of the organelles that they inhabit; and 5) QAPB did not change the cellular disposition of EGFP2-tagged receptors, unlike agonists, which can cause a visible shift toward internalization. The similarity of the distribution of EGFP2 and QAPB also confirmed that the relative lack of cell surface QAPB binding is an accurate reflection of the proportion of the receptor population at the plasma membrane rather than a technical artifact. As in HEK293T cells, QAPB labeling of the α1a-AR stably expressed in R-1F cells confirms the intracellular localization of the ligand-receptor complex in a second cell type. Data are presented as a 3D reconstruction [Fig. 2D, x-y view (i), x-z view (ii), and y-z view (iii); cell nuclei shown in blue and green punctate spots represent QAPB binding].
Transiently transfected HEK293T cells were suitable for demonstrating features of cell distribution and colocalization, but because they had variable expression of the transfected receptors, they were less suitable for quantitative analysis of fluorescence ligand binding than stably transfected fibroblast lines.
Time Course of Formation, Spatial Pattern, and Continuous Movement of Ligand-α1a-AR Complexes. The equilibrated cells discussed above had indicated that QAPB gains access to all α1a-AR in live cells. The time course of the appearance (and disappearance) of fluorescence allowed analysis of the processes involved. Time-lapse epifluorescence images over 75 min were constructed in R-1F cells stably transfected with wild-type receptor and using QAPB at a concentration (5 nM) near its equilibrium constant (Fig. 3). Events were most easily seen and could be readily plotted by taking a two-dimensional view that keeps the fluorescence from the whole cell thickness in the focal plane.
Within 1 to 2 min, the cells were rapidly “covered” with ligand, as shown by low level diffuse fluorescence corresponding to binding to receptors on, or in proximity to, the cell membrane (Fig. 3A, ii). Slightly later, noticeable within 3 to 4 min of ligand application, higher intensity fluorescent signals occurred in punctate intracellular vesicles that were scattered throughout the cell cytoplasm (Fig. 3A, iii). Very soon after this (5-10 min), higher intensity fluorescent signals started to occur in larger punctate structures near the nucleus (Fig. 3A, iv), and their intensity increased steadily to equilibrium at 55 to 60 min (Fig. 3A, v). The pattern of fluorescence therefore evolves from a small particulate distribution toward larger structures with time. Asterisk superimposed on image i of Fig. 3A represents the link to supplemental QuickTime Movie 1 (created from assembled sequential time-lapse images of image i acquired at 5-s intervals).
Video sequence of Movie 1 revealed that some of the fluorescent objects displayed short-range bidirectional motion, whereas others were inert or moved vast distances across the body of the R-1F. Thus, the fluorescent organelles containing α1a-ARs moved continuously, and this occurs in the absence of agonist stimulation.
This visual interpretation was confirmed quantitatively by binning the pixels according to their intensity (Fig. 3C) and then displaying the contents of each bin as a time course (Fig. 4A). In this way, the binding to receptors over the whole cell surface can be distinguished from binding to the punctate intracellular structures, which represent a smaller fraction of the image but are brighter. The lowest intensity bin (bin 1), which contains the black pixels and the background fluorescence, is omitted from the graphs. Bin 2 is the lowest intensity bin that shows specific fluorescence. This shows the rapid acquisition of the binding signal over the whole cell surface. It reached a peak at between 2 and 3 min, and at this time, there were no high-intensity signals because it was too early for the ligand to have reached the intracellular structures. However, after 3 min the higher intensity bins started to increase in sequence as the punctate intracellular structures occurred. A consequence of this was that the number of pixels in bin 2 decreased as the punctate structures took over their pixels. The time course of the increase in the number of pixels in the highest intensity bins indicates the slower acquisition of binding by the intracellular structures until ligand binding reached a steady state (at 40-60 min; Figs. 3A, image v; and 4A). Note that the total number of pixels in the “brighter bins” (bins 3-8) remained constant after 10 min, but the number in the brightest bins (bins 5-8) increased at the expense of the intermediate bins (bins 3 and 4); i.e., all the intracellular structures were fully revealed at 10 min, but they did not become saturated until 60 min.
Separation of α1a-ARs in Intracellular Pools and at the Cell Surface in R-1Fs
Receptor expression on the cell surface can be estimated from the rapid, initial, diffuse binding that occurs after addition of ligand but before the intracellular organelles can be visually resolved. This occurred within 2 min (Fig. 3A, ii), indicating a rapid association rate for the ligand- and surface-located receptors. The sum of the cellular fluorescence at this point in time thus represents the total of the cell surface receptors. When equilibrium occurs, in a 2D image that visualizes everything in the cell, the image of the intracellular binding is superimposed upon this surface image. Subtraction of this image from the equilibrium image thus leaves the image of the intracellular binding (Fig. 3B). Because the surface image is constant, this subtraction can be applied to a time series of images to leave a much enhanced view of the mobile intracellular organelles (Fig. 3B) and proved a remarkably effective way of separating the images of the surface and intracellular ligand-receptor complex.
This information could then be used quantitatively to assess relative numbers of receptors on the surface and inside cells. Because the fluorescence at equilibrium represents the total receptor population, the proportion of receptors on the surface can be calculated from the ratio of binding at 2 min and at equilibrium, giving a surface population of 30% and intracellular of 70%. When the surface binding was subtracted from the binding at each time point a graph of the rate of binding detected in intracellular sites could be constructed (Fig. 4C; compare solid black line with broken line trace). From this graph, a rate constant for the appearance of intracellular binding could be estimated. The resulting t1/2 of 12.4 min did not differ significantly from the equivalent value calculated from the total fluorescence [i.e., 12.5 min (2)].
Reversal of Binding by Washing or the Competitive Antagonist Phentolamine
Washout. The binding of QAPB was reversible when the ligand was withdrawn from the superfusate (Figs. 3D and 5B). Viewing the time lapse or movie showed a steady fade of all parts of the image, but it was less easy to analyze visually than the “on” phase. Quantitative analysis showed essentially a reversal of the histogram intensity pattern compared with the on phase (a progressive shift toward lower intensity bins) (Fig. 4B). However, in contrast to the fast initial increase in low-intensity pixels in the on phase binding, the initial event was a slow but steady fall in the number of pixels across the high-intensity bins, suggesting that outgoing receptor-ligand complexes were replaced by vacant receptors (there is a corresponding increase in the number of low-intensity pixels in bin 2 as the punctate structures disappear). Although this is not dramatic on the histogram plot, because these pixels are of high intensity, the fall in overall fluorescence was substantial and rapid (Fig. 5B). The overall pattern was therefore of loss from the inside, with low-intensity surface binding last to disappear, as would be expected if ligand-receptor complexes steadily returned to the surface.
Phentolamine. Although washing with saline produced a rapid fall in QAPB fluorescence, the t1/2 for the disappearance of fluorescence, 22.0 min (Table 1 was slower than for its appearance. This could be accounted for by reassociation of ligand because washing in the presence of an excess of phentolamine (100 μM; n = 3) increased the rate of loss of QAPB fluorescence by a factor of 2; t1/2 = 10.7 min (Table 1; Fig. 5B, pooled data points, dark gray line). Thus, the true rates of uptake and loss are very similar, at 12.5 and 10.7 min, respectively.
Phentolamine was able to displace fluorescent ligand even in the continued presence of QAPB (5 nM) in the saline. This was concentration-dependent (Fig. 5): 10 μM phentolamine was not fully effective (Fig. 5A, iii) but 100 μM was (Fig. 5A, iv, and B), thus establishing that all detectable binding was specific and that the fluorescent receptor-ligand complexes were susceptible to competitive ligands. Displacement in the continued presence of QAPB by phentolamine (100 μM) was slower than washing with or without phentolamine (Table 1; Fig. 5B, light gray line, t1/2 = 25.6 min; n = 4), as expected if some bound ligand continued to be internalized before equilibration with phentolamine had been attained.
Estimation of Receptor Cycling Time and Residence at Cell Surface
If the receptor population is spontaneously cycling, the average time at the surface should be the same proportion of the total cycling time. The cycling time cannot be calculated accurately from these data, but its limits can. Because halfsaturation is achieved in 12.5 min (Table 1), it is a reasonable working assumption that one inward leg of the cycle must occur within this time even if all receptors carried a bound ligand molecule (and the equilibrium binding data suggest that there would be on the order of 90% saturation, provided that the receptors were located at the surface long enough to equilibrate). This is supported by the rate of loss of fluorescence when the ligand was withdrawn in an excess of competitor, which was very similar; t1/2 = 10.7 min (Table 1). Together, these data allow the estimate that a full cycle is less than 25 min. Therefore, the average residence at the cell surface should be less than 7.5 min (30% × 25 min).
Is Endocytosis Involved in Internalization of Ligand?
The hypothesis that the QAPB ligand is internalized in association with recycling receptors was validated by blocking clathrin-coated pit/vesicle formation endocytosis using either Con A or hypertonic sucrose (Fonseca et al., 1995). It was first shown, quantitatively, that QAPB binding was reversible and reproducible (Fig. 6, open columns). Con A (0.25 mg/ml) or hyperosmotic sucrose (0.45 M) significantly inhibit the binding of QAPB (Fig. 6, black and gray columns) to recycling α1a-ARs. After 30-min exposure to Con A or sucrose, QAPB fluorescence was inhibited by 84 ± 6% (p < 0.001; n = 5) and 85 ± 3% (p < 0.001; n = 5), respectively. This inhibition was maintained over 90 min with only a small further accumulation of fluorescence (Fig. 6). Filipin (5 μg/ml), an inhibitor of caveolar endocytosis, had no effect at any time point (shown at 90 min in Fig. 6B, dotted black column).
Is β-Arrestin Necessary for the Internalization or Trafficking of α1a-ARs?
This was tested in MEFs devoid of β-arrestin 1/2. The α1a-AR-EGFP2 fusion protein was predominantly expressed inside these cells in a clumped and clustered vesicular arrangement (Fig. 7A, i and ii, green, white arrows) that presumably reflects the well appreciated limitation of many transiently transfected cells to effectively mature and deliver receptors to the cell surface.
The asterisk on image i in Fig. 7A represents the link to Movie 2 prepared from assembled images of image i recorded at 15-s intervals. The video sequence of Movie 2 indicated that the intracellular, clumped pool of α1a-AR-EGFP2 was motionless, in contrast with observations in wild-type MEF cells (see below). Despite this, in both sets of MEF cells, some of the α1a-AR-EGFP2 fusion protein was localized at the cell surface (Fig. 7A, i and ii, green, red arrows). In the β-arrestin 1/2-deficient cells, after 75-min exposure to RQAPB (10 nM), the intracellular α1a-AR-EGFP2 had not bound the RQAPB ligand (Fig. 7B, overlay image iv, white arrows). RQAPB did bind on the cell surface, where it colocalized with EGFP2-labeled surface α1a-ARs (Fig. 7B, iv, red arrow, yellow represents colocalization). Asterisk on top of overlay image iv of Fig. 7B represents the link to Movie 3 showing a 3D y-z view of this image. Movie 3 clearly indicates the RQAPB ligand-α1a-AR-EGFP2 complex (yellow) only on the surface of the MEF cell. The volume of α1a-AR-EGFP2- (surface plus intracellular) and RQAPB-labeled α1a-AR-EGFP2 complex (surface only) was 1823 and 1232 μm3, respectively. The volume of intracellular EGFP2-tagged α1a-ARs not labeled with RQAPB was 590 μM3 (i.e., 32% of the total available EGFP2-tagged receptors). The relative fraction of surface to intracellular α1a-ARs (at steady state), was 68 and 32%, respectively. This indicates that RQAPB did not gain access to the intracellular pool of α1a-ARs in β-arrestin-negative MEF cells, demonstrating further that the ligand is not able to diffuse efficiently across the cell surface. By contrast, in wild-type MEF cells that express β-arrestins, the α1a-AR-EGFP2 recycled (movie not shown), and after pre-equilibration with RQAPB (10 nM) for 75 min, it was observed that the RQAPB ligand colocalized with both cell surface (red arrow) and intracellular (white arrow) EGFP2-tagged α1a-ARs (Fig. 7C, overlay image iv, yellow represents colocalization). Asterisk on overlay image iv of Fig. 7C represents the link to Movie 4 showing a 3D y-z view of this image. The video sequence of Movie 4 clearly shows the RQAPB ligand-α1a-AR-EGFP2 complex (yellow) inside and on the cell surface. As anticipated from the above-mentioned description, nontransfected wild-type MEFs displayed no RQAPB fluorescence after 75-min exposure (Fig. 7C, compare image i with ii and iii.). These studies also support the hypothesis that RQAPB has to bind to cell surface receptors before its internalization and entry into the cell.
Are α1a-AR- and β-Arrestin Found Together?
To answer this, R-1F cells were used that expressed both β-arrestin 2-EGFP and the bovine α1a-AR. This resulted in nonuniform cytoplasmic localization of β-arrestin 2-EGFP with more intense β-arrestin-2-related fluorescence expressed in punctate vesicles (Fig. 8A, i, green vesicles some shown by white arrows). In contrast to this pattern, introduction of β-arrestin 2-EGFP into otherwise nontransfected R-1Fs resulted in a uniform cytoplasmic distribution of the tagged β-arrestin 2 (data not shown). Asterisk overlaid on image i of Fig. 8A represents the link to Movie 5 prepared from reconstructed sequential time-lapse 2D images of image i recorded at 15-s intervals. Movie 5 revealed that the organelles containing β-arrestin 2-EGFP showed rapid movement around the cells similar to that found for organelles carrying α1a-ARs (Movie 5; Fig. 8A, i). No significant change in the localization or movement of β-arrestin 2-EGFP2 occurred upon exposure (15 min) to the agonist phenylephrine (3 μM) (Fig. 8A, compare image ii with iii). The same was found for the antagonist/inverse agonists prazosin (3 μM), phentolamine (3 μM), and RQAPB (10 nM) (data not shown).
When α1a-ARs were labeled with RQAPB (10 nM), it partly colocalized with β-arrestin 2-EGFP (Fig. 8B, iv; yellow/orange punctate vesicles, some illustrated by white arrows). The degree of colocalization was quantified by plotting the amount of background-subtracted green EGFP fluorescence from the pixels in a region of interest (Fig. 8C, red rectangle superimposed on image ii and iii), against the amount of background red fluorescence in the corresponding pixels of the matched region in the RQAPB image. Correlation coefficients were quantified that described the degree by which EGFP and RQAPB fluorescence at each pixel within the region varied from a perfect correlation of 1.00. Calculated correlation coefficient was 0.91, indicating that RQAPB-labeled bovine α1a-ARs colocalized very efficiently with β-arrestin-2-EGFP. Together, these data demonstrate partial, spontaneous association of α1a-AR with β-arrestin-2EGFP- and indicate that α1a-AR recycling is a β-arrestin-dependent, ligand-independent process.
Which Endosomal Compartments Does the Ligand-α1a-AR Complex Enter?
To visualize and determine the identity of the intracellular compartments where the QAPB-receptor complexes resided, we examined potential colocalization with fluorescent reagents that can selectively identify different subcellular compartments. Tfn-Alexa Fluor546, a recycling endosome fluorescent marker for endocytosis via clathrin-coated pits (Anderson, 1998; Gagnon et al., 1998), was used to identify the recycling pathway. Merged images, presented in Fig. 9A, iii, show that after 40 min, fluorescence corresponding to α1a-AR-QAPB ligand sites (Fig. 9A, i, green punctates) partially overlapped with the fluorescent recycling marker (Fig. 9C, iii, yellow perinuclear vesicles, some indicated by white arrows). Some of the receptor-QAPB ligand binding sites were spatially distinct from the recycling endosomal Tfn receptors (Fig. 9C, ii, red punctates). A 3D x-z view of bovine α1a-AR-QAPB ligand complexes partially colocalizing with recycling compartments (yellow punctates, white arrows) is shown in image iv of Fig. 9A.
Lysotracker Red DND-99, a highly selective red fluorescent acidotrophic probe for labeling acidic organelles (Bucci et al., 2000), was used to identify acidic late endosomes. After 90 min, Lysotracker Red fluorescence had extensive overlap with the intracellular structures binding QAPB (Fig. 9B, iii, yellow punctates, some shown by white arrows, compare images i, ii, and iii in Fig. 9B), implying that α1a-AR ligand complexes were trafficked to the late endosomal/lysomal network. The 3D x-z colocalization view of bovine α1a-AR-QAPB ligand complexes fusing with late/lytic endosomes is shown in Fig. 9B, iv (yellow punctates, white arrows).
Discussion
Visualizing the receptor/ligand complex in real-time demonstrates that the α1a-AR is spontaneously internalized by endocytosis when expressed in a range of cells, including R-1Fs. The receptor continuously recycles, spending a relatively small fraction of time at the cell surface. Most of the time, each receptor molecule is present in intracellular organelles that shuttle between the cell surface and perinuclear compartments, including early and late endosomes. Both receptor internalization and the formation and movement of the mobile intracellular organelles are dependent on β-arrestins but not on receptor activation.
Cycling is continuous and rapid, with complete turnover within 50 to 60 min. This has implications for understanding the “purpose” and properties of recycling and the dynamics of drugs acting at the receptors. The rate of agonist-induced internalization is not known in the case of α1a-AR, but it is more resistant to desensitization than α1b-AR (Vazquez-Prado et al., 2000). Agonist-induced internalization, as judged by disappearance of cell surface receptors, is undetectable for tagged α1a-ARs in R-1Fs after 10 min (Price et al., 2002) and achieves small reductions even after 60 min of high concentrations of agonists (Stanasila et al., 2003; Morris et al., 2004). These net shifts of location may of course disguise an increase in recycling, with only the net shift in location being detected. Thus, constitutive recycling is the phenomenon more likely to occur in physiological conditions.
Agonist-induced recycling of α1-AR has yet to be demonstrated at physiological concentrations of natural agonists or in native cells. When the α1a-AR is localized in native cells such as smooth muscle or hepatocytes, it has consistently shown a predominant intracellular location coupled with a less dominant plasmalemmal location (Mackenzie et al., 1998, 2000; Hrometz et al., 1999; McGrath et al., 1999; Deighan et al., 2004; Miquel et al., 2005; for review, see Daly and McGrath, 2003). This could result from modulation by low levels of physiological agonist activation. Spontaneous α1a-AR recycling may maintain an internal reserve pool of receptors that can sustain continuous agonist-induced signaling in the face of stimulation-induced attrition of the surface population, as suggested by Morris et al. (2004).
Our current advance is the ability to visualize and follow the receptor-ligand complex in real time. EGFP2-tagged receptors show total receptor localization but provide no information on movement within this population. Fluorescent antagonist ligands show the most recently labeled receptors that have internalized. Ligands access the interior of cells only if bound to endocytosing receptors because ligands enter only cells that possess both the receptor and β-arrestins.
The rate of appearance of intracellular fluorescence provides a time and concentration profile of the internalizing receptors from which rates of entry and of receptor recycling turnover can be estimated. This is analogous to the established use of antibodies to extracellular epitope tags for the study of internalization (Morris et al., 2004), but, additionally, can be followed in real time, and, in conjunction with other vital dyes, can provide information on coincident dynamic events. The entire population recycles within 25 min, so the average residence time of the 30% at the cell surface is less than 7.5 min, even without stimulation.
The rapidity of these events has implications for receptor signaling and regulation. It cannot be assumed that internalized receptors are in a desensitized state that is incapable of signaling, although interaction with a β-arrestin will prevent concurrent interaction with G protein. An emerging literature indicates a distinct set of signaling functions for β-arrestins, particularly in the regulation of the kinetics and effectiveness of mitogen-activated protein kinase signaling cascades (Shenoy and Lefkowitz, 2003; Lefkowitz and Whalen, 2004). Although receptors are able to internalize with agonists bound, the low pH of recycling and other endosomes encourages dissociation, and it is likely that only high-affinity agonist ligands (e.g., peptides) remain bound throughout the cycle. Likewise, the paradigm of agonist-mediated internalization may require modification because agonists may simply accelerate spontaneous endocytosis rather than initiate it. A change in the balance of surface/intracellular receptors will occur only when the inward rate exceeds the rate of return to the surface, and this concept can be modeled mathematically (Koenig and Edwardson, 1997; Koenig, 2004).
Real-time visualization of movement of endosomal ligand-receptor complexes can allow the study of recycling in native cells, which was until now impractical because of the need for epitope tags. Receptors in mobile endosomes have been visualized previously only by using GFP-tagged G protein-coupled receptors (Koenig and Edwardson, 1997; Kallal and Benovic, 2002). Because an identical pattern of distribution and movement is found with the ligand or EGFP, their colocalization provides cross-validation of both techniques. Neither the EGFP2 tag nor the bound ligand slows endosomal movement or changes receptor distribution.
In cells devoid of β-arrestins, immobility of EGFP2-tagged receptor showed that β-arrestins are needed for the genesis of the mobile endosomes, a key part of the recycling system. Without them, receptors remained immobile in a fixed reticular endosomal compartment. Nevertheless, the proportion of receptors at the surface in β-arrestin-deficient MEF cells was similar to wild-type MEF cells, so β-arrestins are necessary for recycling, but recycling does not regulate partitioning of the receptor population.
Vesicles containing ligand-receptor complexes moved to and from the plasma membrane, indicating a ferrying function. This was visible for QAPB-labeled receptors in both R-1Fs and HEK293T cells and for EGFP2-labeled receptor in HEK293T cells. These movements are therefore an inherent property of the cells and not a phenomenon instigated by receptor occupation or cell type. The development with time of the intensity of these fluorescent structures when the cells were exposed to QAPB is akin to development of a photograph. Before uncovering the key role of endocytosis, it might have been assumed that ligand diffused into the cell, where it bound to internally located immobile receptors. Internalization of QAPB can now be understood in terms of endocytosis of a QAPB-receptor complex. In endocytosis, the basic units of transport are small structures, beyond the resolution of the microscope (e.g., caveoli or endocytic vesicles), that reversibly fuse with small visible structures to exchange a cargo of plasma proteins, including the ligand-receptor complex. These small fluorescent “transit” endosomes then move further into the cell and reversibly fuse with larger perinuclear endosomes, delivering ligand-receptor complexes.
Movements of transit vesicles and their transactions with the perinuclear complexes are made visible by the high concentrations of receptor-ligand complexes on their resolvable surfaces. The intensity per pixel shows receptor density to be 5 times higher than on the cell surface, indicating a high degree of physical marshaling and cargo selection. In all these intracellular transactions, the ligand and binding site are restricted to the internal face of endosomes, so that even if they dissociate, the ligand molecules will be restricted to the compartment and will be likely to reassociate. QAPB leaves the cell again on its removal from the extracellular environment, with a time course similar to its endocytic uptake, so it seems likely that it leaves the cell by the reverse process, bound to receptors recycling back to the cell surface.
Many α1-AR antagonists are inverse agonists, including the quinazolyl-piperazine (e.g., prazosin, doxazosin, and alfuzosin) family to which QAPB belongs (García-Sáinz and Torres-Padilla, 1999; Price et al., 2002). However, the wild-type α1a-AR receptor has low levels of constitutive activity (Rossier et al., 1999; Price et al., 2002). Furthermore, antagonists, including prazosin and QAPB, do not produce either a depression or a rise of resting intracellular Ca2+ concentration in these R-1F cells (Pediani et al., 2000), eliminating inverse agonism and partial agonism. In addition, Morris et al. (2004) showed that prazosin did not affect spontaneous endocytosis of the hemagglutinin-tagged α1a-AR, and our data with phentolamine are consistent with competition and displacement of QAPB, rather than blockade of recycling.
Endocytosis was established as the entry mechanism for ligand-receptor complex using two independent agents (hyperosmotic sucrose and Con A) that block clathrin-mediated internalization (Daukas and Zigmond, 1985; Heuser and Anderson, 1989), and involvement of β-arrestins was confirmed by the absence of ligand uptake in β-arrestin-deficient MEF cells and by the partial colocalization of β-arrestin-2-GFP with the ligand/receptor. The generalization that interactions between receptor and β-arrestins require agonist-mediated phosphorylation of the receptor no longer holds. Such phosphorylation frequently increases the affinity of such interactions rather than defining them. For example, interactions between β-arrestins and the protease-activated receptor-1 are independent of receptor phosphorylation (Chen et al., 2004) as are such interactions with the BLT1 leukotriene B4 receptor (Jala et al., 2005). However, the interactions of β-arrestins with components of the machinery of clathrin-mediated endocytosis are central to the delivery of β-arrestin/receptor complexes to clathrin-coated pits (Ferguson et al., 1996). Filipin, an inhibitor of caveolar endocytosis, had no major effect on uptake of QAPB. These results show that endocytosis via clathrin-coated vesicles is responsible for the ingress of QAPB.
The receptors initially enter in a similar manner as Tfn, and after a longer period penetrate endocytotic compartments beyond the limit of Tfn's ingress. Here, they are trafficked to more acidic vesicles: in the late endosomes, the receptors then go through a second stage of sorting and either are targeted to lysosomes for degradation or are slowly recycled back to the cell surface (Mellman, 1996; Resat et al., 2003). Their dynamic equilibrium prevents easy distinction between early, late, and lysosome compartments, so we cannot estimate their relative proportions or the amount of receptor degradation. However, saline washing eliminated fluorescent ligand from cells, excluding accumulation in a compartment for long-term storage. A high proportion of the receptors were found in endocytic vesicles in both cells types used. This might be attributed to “overexpression” in these recombinant systems, but it is very similar to native systems such as smooth muscle cells from rat blood vessels and human prostate that express α1-ARs (Mackenzie et al., 1998, 2000; McGrath et al., 1999; Deighan et al., 2004).
Spontaneous endocytosis responsible for internalization of the α1a-AR and its antagonist ligand cargo is a new paradigm for uptake of antagonist drugs. It alters the potential mechanisms involved when whole cells are used in radioligand binding assay. Order of addition of drugs to cells becomes a potential source of variability, and time to equilibrium should be longer than if all binding were on the cell surface. When “membrane” preparations are used, many receptors will derive from intracellular membranes, and this may influence their properties, given the greater densities and the presence of different cofactors. Based on susceptibility to internalization, receptors are categorized into those that internalize/recycle constitutively or upon binding agonist (Milligan and Bond, 1997). However, receptors internalizing with antagonists attached do not fit this scheme. Agonist-independent G protein-coupled receptor internalization has been observed for the angiotensin AT1A receptor and cholecystokinin receptors (Hein et al., 1997; Tarasova et al., 1997; Anborgh et al., 2000). For angiotensin AT1A receptor, β-arrestin association with receptor, without agonist, is responsible for agonist-independent loss of cell surface receptor (Anborgh et al., 2000). The α1a-AR can now be added to the “spontaneously recycling”, β-arrestin-dependent category. How widely this applies among other G protein-coupled receptors remains to be seen. The cellular uptake of α-blockers at low concentrations also introduces a new concept into their mechanism of action in treating conditions such as benign prostatic hyperplasia. Intracellular concentration of drug at receptor-dense sites raises potential intracellular actions that require investigation and introduces new pharmacokinetic issues.
In conclusion, α1a-AR recycle rapidly by agonist-independent, constitutive, β-arrestin-dependent processes. Receptors spend most of their time in endosomal recycling compartments and have a relatively short residence at the cell surface. Antagonists that bind to the receptive site are internalized with the receptor and transported through the recycling process, attaining saturation binding to intracellular receptors.
Acknowledgments
We thank Professor Gwyn Gould (Biochemistry and Molecular Biology, Institute of Biomedical and Life Sciences, University of Glasgow) for helpful discussions.
Footnotes
-
D.C. was supported by a Medical Research Council Ph.D. studentship and J.F.C. was supported by the Ann B. McNaught Bequest.
-
ABBREVIATIONS: α1-AR, α1-adrenoceptor; QAPB, BODIPY-FL prazosin; R-1F, rat-1 fibroblast; EGFP, enhanced green fluorescent protein; DMEM, Dulbecco's modified Eagle's medium; PBS, phosphate-buffered saline; Con A, concanavalin A; RQAPB, BODIPY-R 558/568 prazosin; HEK, human embryonic kidney; MEF, mouse embryo fibroblast; Tfn, transferrin; 3D, three-dimensional; 2D, two-dimensional; GFP, green fluorescent protein.
-
↵s⃞ The online version of this article (available at http://molpharm.aspetjournals.org) contains supplemental material.
- Received October 19, 2004.
- Accepted December 21, 2004.
- The American Society for Pharmacology and Experimental Therapeutics