Acute Agonist-mediated Desensitization of the Human α1a-Adrenergic Receptor Is Primarily Independent of Carboxyl Terminus Regulation IMPLICATIONS FOR REGULATION OF α1aAR SPLICE VARIANTS

Despite important roles in myocardial hypertrophy and benign prostatic hyperplasia, little is known about acute effects of agonist stimulation on α1a-adrenergic receptor (α1aAR) signaling and function. Regulatory mechanisms are likely complex since 12 distinct human α1aAR carboxyl-terminal splice variants have been isolated. After determining the predominance of the α1a-1AR isoform in human heart and prostate, we stably expressed an epitope-tagged α1a-1AR cDNA in rat-1 fibroblasts and subsequently examined regulation of signaling, phosphorylation, and internalization of the receptor. Human α1aAR-mediated inositol phosphate signaling is acutely desensitized in response to both agonist and phorbol 12-myristate 13-acetate (PMA) exposure. Concurrent with desensitization, α1aARs in32Pi-labeled cells are rapidly phosphorylated in response to both NE and PMA stimulation. Despite the ability of PKC to desensitize α1aARs when directly activated with PMA, inhibitors of PKC have no effect on agonist-mediated desensitization. In contrast, involvement of GRK kinases is suggested by the ability of GRK2 to desensitize α1aARs. Internalization of cell surface α1aARs also occurs in response to agonist stimulation (but not PKC activation), but is initiated more slowly than receptor desensitization. Significantly, deletion of the α1aAR carboxyl terminus has no effect on receptor internalization or either agonist-induced or GRK-mediated receptor desensitization. Because mechanisms underlying acute agonist-mediated regulation of human α1aARs are primarily independent of the carboxyl terminus, they may be common to all functional α1aAR isoforms.

␣ 1a -Adrenergic receptors (␣ 1a ARs) 1 are G protein-coupled receptors (GPCR) that mediate sympathetic nervous system responses such as smooth muscle contraction and myocardial inotropy (1). NE stimulation of ␣ 1 ARs predominantly activates G q and results in membrane polyphosphoinositide hydrolysis by activation of phospholipase C␤; the resultant second messengers IP 3 and DAG mobilize intracellular calcium and activate protein kinase C (PKC), respectively (2). Three ␣ 1 AR subtypes (␣ 1a , ␣ 1b , and ␣ 1d ) have been cloned and pharmacologically characterized in several expression systems (for review, see Ref. 2). Clinically, activation of ␣ 1a ARs has been implicated in the dynamic component of benign prostatic hyperplasia leading to bladder outlet obstruction and in the development of myocardial hypertrophy (3)(4)(5). Notwithstanding the importance of ␣ 1a ARs in several pathophysiological states, surprising little is known about mechanisms underlying ␣ 1a AR expression and function. Transcriptional mechanisms unique to ␣ 1a ARs have been shown to be important in maintaining full ␣ 1 AR responsiveness to agonist in rat neonatal myocytes where long term (24 -72 h) NE stimulation leads to up-regulation ␣ 1a AR mRNA and receptor protein expression, concurrent with down-regulation of ␣ 1b and ␣ 1d AR subtypes (5,6). While important, such long-term studies do not examine acute regulation of ␣ 1a AR signaling in response to agonist stimulation, giving rise to the question whether ␣ 1a ARs are subject to acute processes such as agonist-induced desensitization and receptor phosphorylation.
Desensitization (dampening of receptor responsiveness to continued agonist exposure) is thought to involve a number of inter-related yet distinct mechanisms occurring at the receptor and more broadly in the signal transduction pathway (7). Receptor desensitization may occur within seconds to minutes of agonist stimulation, and is generally considered to result from receptor uncoupling from downstream effectors due to receptor phosphorylation. Phosphorylation of activated GPCRs by G protein-coupled receptor kinases (GRKs) leads to increased binding of ␤-arrestin to the receptor complex, uncoupling receptors from G proteins and, in some cases, leading to internalization of the phosphorylated receptors. Phosphorylation, desensitization, internalization, and down-regulation (decreased expression) of the ␣ 1b AR subtype in response to either agonist or PMA exposure have been extensively demonstrated in native and cell models (8 -12). Significantly, these processes are mediated through the carboxyl terminus of the ␣ 1b AR. Truncation of its carboxyl terminus significantly decreases agonist-and PMA-mediated phosphorylation, desensitization, and internalization of the ␣ 1b AR, and recent studies have pinpointed critical serine residues within that receptor region that are involved in each of these regulatory processes (8,(11)(12)(13).
In contrast to the ␣ 1b AR, current knowledge regarding agonist-mediated regulation of the ␣ 1a AR subtype is severely lacking. One study of bovine ␣ 1a ARs recently suggested that the ␣ 1a AR, similar to the ␣ 1b subtype, is subject to agonist-induced desensitization and phosphorylation (14). Mechanisms underlying regulation of ␣ 1a ARs, however, are potentially more complex than those of the ␣ 1b subtype. Unlike ␣ 1b and ␣ 1d ARs which are each expressed as a single isoform, several distinct ␣ 1a isoforms have been isolated in addition to the original "wild type" ␣ 1a AR, 12 in humans and 4 in rabbit (15)(16)(17)(18). Although several of these variants give rise to non-functional truncated polypeptides with only six transmembrane domains, there are four fully functional ␣ 1a AR isoforms that exhibit ligand binding and signaling characteristics essentially identical to those of the wild type receptor (15)(16)(17). It is of particular interest that all functional ␣ 1a AR splice variants are identical except at the distal ends of their carboxyl termini, where each differs in sequence and length. The study of ␣ 1a AR regulation in endogenous systems may prove difficult, however, since several ␣ 1a AR carboxyl-terminal splice variants are concurrently expressed in every tissue studied thus far (15)(16)(17). Additionally, expression of severely truncated ␣ 1a AR isoforms can decrease signaling by full-length receptors, possibly through interference with trafficking and membrane localization of full-length receptors (17).
To accurately characterize acute agonist-mediated regulation of human ␣ 1a AR signaling, we directly profiled expression levels of functional full-length ␣ 1a AR splice variants in human heart and prostate and identified ␣ 1a-1 as the predominant ␣ 1a AR isoform in these tissues. We then stably expressed the ␣ 1a-1 AR as an epitope-tagged fusion protein in rat-1 fibroblasts to examine effects of acute (Յ30 min) NE and PMA stimulation on ␣ 1a AR signaling and to subsequently explore the roles of PKC and several GRK family members, receptor phosphorylation, and receptor internalization in acute regulation of ␣ 1a ARs. In parallel we constructed, expressed, and tested a fully functional carboxyl-terminal truncated ␣ 1a AR (whose sequence is common to all functional ␣ 1a AR splice variants) to characterize any regulatory features that may be independent of carboxyl-terminal variation. Our findings suggest that acute agonist-mediated regulation of human ␣ 1a ARs is primarily independent of the carboxyl terminus and could therefore be common to all functional ␣ 1a AR isoforms. Construction of Probes for RNase Protection Assays of Human ␣ 1a AR Carboxyl-terminal Splice Variants-To profile ␣ 1a AR carboxyl-terminal splice variant expression in human heart and prostate, cDNA templates for isoform-specific RNase protection assay probes were constructed. Polymerase chain reaction (PCR) using Pfu polymerase (Roche Molecular Biochemicals, Indianapolis, IN) was used to amplify the 3Ј cDNAs ends of each functional human ␣ 1a AR splice variant: ␣ 1a-1 (wild type), ␣ 1a-2 , ␣ 1a-3 , and ␣ 1a-4 . A common 5Ј amplification primer (containing the internal ␣ 1a AR EcoRV restriction site) located 322 bp upstream of the ␣ 1a AR carboxyl terminus splice site was used in conjunction with variant-specific downstream primers corresponding to distal 3Ј end of coding sequence and containing introduced XbaI restriction sites. PCR products were digested with EcoRV and XbaI and subsequently ligated into a pcDNA3 (Invitrogen, Carlsbad, CA) construct containing the wild type ␣ 1a AR to generate full-length ␣ 1a AR splice variant constructs. Digestion of each splice variant construct with EcoRV and ApaI provided isoform-specific fragments that were subsequently ligated into pGEM7Zϩ (Promega, Madison, WI). Modified sequences of each PCR construct were verified by dideoxy DNA sequencing using the fmol DNA Cycle Sequencing System (Promega).

Materials
RNA Isolation and RNase Protection Assays-Total RNA was extracted from human tissues (prostate, heart) and SK-N-MC cells using RNA STAT-60 (Teltest Inc., Friendswood, TX). Each RNA sample was quantitated spectrophotometrically at 260 and 280 nm and stored at Ϫ70°C until use. Radiolabeled antisense isoform-specific ␣ 1a AR probes were transcribed from linearized cDNA constructs using RNA polymerase T7, and [␣-32 P]UTP as previously described (19). In addition to probes for each ␣ 1a AR splice variant subtype, a control ␣ 1a AR probe (located upstream of carboxyl-terminal splice site in ␣ 1a AR second exon) and a human cyclophilin control probe (Ambion, Inc., Austin, TX) were used; RNase protection assays were conducted as previously described (19).
Construction of HA-tagged ␣ 1a AR and Carboxyl Terminus Truncated Mutants-To facilitate phosphorylation studies, sequence of the hemagglutinin (HA) epitope was added to the amino terminus of the human ␣ 1a-1 AR cDNA using PCR mutagenesis. The 5Ј (sense) 59-mer oligonucleotide (5Ј-AAAAGAATTCATGTACCCATACGACGTCCCAGACTA-CGCCGTGTTTCTCTCGGGAATG-3Ј) contained a synthetic EcoRI restriction site to facilitate cloning, sequence encoding the 9-residue HA epitope (YPYDVPDYA; bold italics) immediately downstream of the ␣ 1a AR start codon, and 19 bases corresponding to bp 4 to 22 of the ␣ 1a AR (underlined). The 3Ј (antisense) primer was 5Ј-GAGCAGCCTCACT-GAGAAGTGCGT-3Ј, corresponding to bases 796 to 763 of the ␣ 1a AR receptor (GenBank TM accession number 4501960). The resulting PCR product was digested with EcoRI and Eco47III and subcloned into the mammalian ␣ 1a AR expression vector pcDNA3:␣ 1a AR; the final construct is called pcDNA3:HA-␣ 1a . A carboxyl-terminal-truncated ␣ 1a AR mutant (pcDNA3:HA-T348) was generated by introducing a STOP codon after Arg 348 of HA-␣ 1a in pcDNA3:HA␣ 1a using the Transformer Site-directed Mutagenesis Kit (CLONTECH, Palo Alto, CA). Modified sequences of each construct were verified by dideoxy DNA sequencing.
Cell Culture, Stable Transfection, and ␣ 1 AR Ligand Binding-Cells were grown in monolayers and maintained in DMEM supplemented with 10% fetal bovine serum, penicillin (100 units/ml), and streptomycin (100 g/ml) in 5% CO 2 at 37°C. For stable expression of receptors, plasmids pcDNA3:HA-␣ 1a and pcDNA3:HA-T348 were individually transfected into rat-1 fibroblasts by electroporation (Gene Pulsar II, Bio-Rad Inc.; optimal conditions were 250 mvolts and 950 F capacity, resulting in a time constant of 12-15 ms). Clones resistant to G418 (800 g/ml) were isolated and tested for receptor expression. Selection was maintained in clonal cell lines with 400 g/ml G418.
Saturation binding and competition analysis of transfected cell membranes was performed with the radiolabeled ␣ 1 AR antagonist [ 125 I]HEAT (300 and 120 pM for saturation and competition binding, respectively) as previously described (20). The ␣ 1 AR agonists norepinephrine and oxymetazoline, and antagonists prazosin and 5-methylurapidil were utilized for competition analysis. Resultant curves were fit using noniterative regression analysis with PRISM software (Graphpad, San Diego, CA).
Ligand binding on intact cells was performed as described by Lattion (8) with a few modifications. Cell monolayers grown in 12-well plates were washed once with DMEM and treated with drugs, washed once again with DMEM and then incubated with 2 nM [ 3 H]prazosin in 0.5 ml of DMEM at either 4°C for 12-15 h to determine cell surface binding or at 37°C for 30 min to determine total cellular receptor content. Phentolamine (10 Ϫ4 M) was used to determine nonspecific prazosin binding. Following binding, cells were washed three times with ice-cold PBS containing 0.1% bovine serum albumin, and scraped in 1 ml of water; 3 H was counted in 7 ml of scintillation mixture using a Wallac 1409 Liquid Scintillation Counter (Wallac, Gaithersburg, MD).
Measurement of Total Inositol Phosphate Production-Rat-1 cells stably expressing either HA-␣ 1a or HA-T348 receptors were labeled with [ 3 H]inositol for 20 -24 h with 2.5 Ci/ml in DMEM supplemented with 3% fetal bovine serum, penicillin (100 units/ml), and streptomycin (100 g/ml). After labeling, cells were washed with PBS and when necessary incubated in PBS for various times at 37°C; pretreated cells were exposed to various drugs during this incubation as indicated. Cells were then quickly washed with PBS, placed in PBS with 20 mM LiCl and immediately stimulated by NE addition for the indicated times. NE was added to the cells from ϫ100 stocks in 10 mM ascorbic acid while only ascorbic acid was added to unstimulated cells. Inositol phosphates were extracted as described by Martin (21) and separated using Dowex AG1-X8 anion exchange (formate) columns. After washing the columns twice with water, total inositol phosphates were eluted in 1 M ammonium formate, 0.1 M formic acid, and combined with 15 ml of scintillation mixture; 3 H was quantitated using a liquid scintillation counter.
Final GRK-mediated desensitization assays were performed in transiently transfected cells, since GRK transfection in stable cell lines resulted in unexpected accelerated cell death (a condition not seen in any other assay). All final IP assays were normalized for ␣ 1a AR density and cell count at the time of assay. GRK overexpression was confirmed by Western analysis. Rat-1 or COS-7 cells were plated at 40,000 cells per well in 12-well plates, grown overnight, and transfected with 0.25 g of pcDNA3 containing HA-␣ 1a ARs or HA-T348 with or without 0.25 g of pcDNA1 containing GRK2 or GRK6; the total transfected DNA was kept at 0.75 g (per 40,000 cells) by adding pcDNA3. Following 18 h of transfection, cells were washed with Hank's balanced salt solution, labeled with [ 3 H]inositol in 10% fetal bovine serum, and assayed for total inositol phosphate production in DMEM essentially as described for stable cells.
Photoaffinity Labeling of ␣ 1a ARs-Membranes from cells expressing either HA-␣ 1a or HA-T348 receptors, prepared for ligand binding as described above, were resuspended at a concentration of 1 g of total protein/l. 175 g of total protein were incubated with 6 nM [ 125 I]iodoazidoprazosin (200 l total volume) in the dark at room temperature for 60 min in the presence or absence of prazosin (1 M). Following incubation, samples were UV-irradiated for 10 min, centrifuged at 14,000 ϫ g for 10 min, and resuspended in 6 ϫ SDS-PAGE sample buffer. Proteins were separated by electrophoresis on a 10% SDS-polyacrylamide gel (Protogel, National Diagnostics). 125 I-Labeled receptors were detected by exposure to X-Omat AR film (Kodak) for 24 -48 h. 32 P Labeling and Immunoprecipitation of ␣ 1a ARs in Cells-Equal numbers of rat-1 fibroblasts stably expressing either HA-␣ 1a or HA-T348 receptors were grown to confluency in 6-or 12-well cluster plates. Cell monolayers were washed three times with phosphate-free DMEM, then incubated in the same medium containing 32 P i (0.2 mCi/ml) for 2 h at 37°C. Different drugs were added during this incubation as indicated under "Results." Following incubation, cells were washed three times with ice-cold PBS, solubilized in ice-cold lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1 mM Na 3 VO 4 , 10 mM Na 4 P 2 O 7 , 10 mM NaF, and Complete Protease Inhibitor Mixture (Roche Molecular Biochemicals)), and centrifuged at 14,000 ϫ g for 10 min at 4°C. To reduce background, solubilized proteins were incubated with protein G-agarose (Roche Molecular Biochemicals) for 1 h at 4°C. After removal of nonspecific precipitated material, the supernatant was incubated with 3F10 rat monoclonal antibody (Roche Molecular Biochemicals) against the HA epitope tag at 4°C. After 1 h, protein G-agarose beads were added to the sample mixture and incubation was continued overnight. Isolated immune complexes were washed twice with ice-cold lysis buffer, twice with high salt buffer (50 mM Tris-HCl, pH 7.5, 500 mM NaCl, 0.1% Nonidet P-40, 0.05% sodium deoxycholate), and once with low salt buffer (50 mM Tris-HCl, pH 7.5, 0.1% Nonidet P-40, 0.05% sodium deoxycholate). After removal of buffer, immune complexes were resuspended in 6 ϫ SDS sample buffer and were separated by electrophoresis on a 10% SDS-PAGE gel. After autoradiography, 32 P was quantitated with a PhosphorImager and analyzed using ImageQuant gel image analysis software.
Statistical Analysis-Results are expressed as mean Ϯ S.E. Statistical significance was analyzed by two-factor ANOVA (where appropriate) followed by a two-tailed unpaired t test, with p Ͻ 0.05 considered to be significant.

Profile of Human ␣ 1a AR Isoform Expression in Prostate and Heart
To focus our studies on the full-length functional ␣ 1a AR isoform that is predominantly expressed in human prostate and heart, we directly examined expression of the four fulllength ␣ 1a AR splice variants in these tissues, as well as in SK-N-MC cells (the only currently available human cell line that endogenously expresses ␣ 1a ARs, albeit at very low levels). RNase protection assays (without prior PCR amplification) were performed using ␣ 1a AR probes designed to contain both isoform-specific carboxyl-terminal sequences (WT ␣ 1a-1 , ␣ 1a-2 , ␣ 1a-3 , and ␣ 1a-4 ) and sequences common to all functional ␣ 1a AR isoforms (Fig. 1A). With each probe, levels of individual isoform mRNA expression (isoform-specific fragment) were quantitated relative to all other functional ␣ 1a isoforms (common sequence fragment). As shown in Fig. 1, B and C, the ␣ 1a-1 (wild type ␣ 1a ) predominates in human prostate (81 Ϯ 2%), heart (89 Ϯ 4%), and SK-N-MC cells (85 Ϯ 1%). ␣ 1a-4 mRNA represents 6 -11% of the remaining ␣ 1a AR mRNA pool, whereas ␣ 1a-2 and ␣ 1a-3 are rare in all tissues/cells studied (3-9% combined). Because of its overwhelmingly predominant expression, we subsequently focused our studies on regulation of the wild type ␣ 1a-1 receptor (hereafter referred to as ␣ 1a AR).

Construction, Stable Expression, and Characterization of HA-␣ 1a AR and Carboxyl-Truncated Mutant HA-T348
To facilitate phosphorylation studies of the wild type human ␣ 1a AR, we constructed two specialized mutant ␣ 1a ARs: 1) an amino terminus HA epitope-tagged ␣ 1a AR (HA-␣ 1a ), and 2) a mutated HA-␣ 1a in which the carboxyl terminus was truncated after amino acid 348, eliminating the last 118 amino acids of the ␣ 1a-1 AR (HA-T348). HA-T348 retains expression of the palmitoylated Cys 345 (22), but is truncated well upstream of the ␣ 1a AR carboxyl-terminal splice site present in all func-FIG. 1. Relative expression of ␣ 1a AR carboxyl-terminal splice variant mRNA. A, schematic of probes used in RNase protection assays containing sequences common to all variants as well as isoformspecific sequences. B, representative RNase protection assay results using RNA from human prostate, heart, and SK-N-MC cells with size and identity of ␣ 1a AR isoform-specific (a-1, a-2, a-3, and a-4) and common protected fragments indicated. Exon2 (Ex2) denotes ␣ 1a AR second exon control probe that does not distinguish between isoforms and therefore represents total ␣ 1a AR mRNA pool (shown in unlabeled lane for each tissue). C, quantitation of ␣ 1a AR isoform mRNA expression in human prostate, heart, and SK-N-MC cells. Data are mean Ϯ S.E., n ϭ two to five independent experiments. tional ␣ 1a AR splice variants just after Arg 422 . Thus, the truncated construct represents sequence common to all functional ␣ 1a AR isoforms, regardless of splice variant identity. HA-␣ 1a and HA-T348 constructs were each stably transfected into rat-1 fibroblasts (cells lacking endogenous ␣ 1 ARs), and several stable clones were isolated for each construct with receptor expression levels varying from 0.1 to 1.5 (HA-␣ 1a ) and 0.2 to 2.8 (HA-T348) pmol/mg of total protein. Receptor expression levels were routinely measured to ensure they did not change over time. Unless otherwise stated, clonal lines expressing similar levels of HA-␣ 1a and HA-T348 (1.5 and 1.8 pmol/mg protein, respectively) were utilized in this study. However, assays performed on two additional clones of each cell type with lower levels of receptor expression (HA-␣ 1a , 400 and 200 fmol/mg protein; HA-T348, 700 and 350 fmol/mg protein) demonstrated similar responses of ␣ 1a ARs to agonist-and PMA-induced desensitization (data not shown). This suggests that these phenomena are, to a large degree, independent of receptor density in these cells.
Ligand binding and IP signaling of HA-␣ 1a and HA-T348 were characterized and compared with those of the wild type ␣ 1a AR (untagged, full-length) to determine potential effects of epitope tagging and carboxyl-terminal truncation on the receptor. Binding affinities of both HA-␣ 1a and HA-T348 for agonists (NE and oxymetazoline) and ␣ 1 AR antagonists (prazosin and 5-methylurapidil) are indistinguishable from those of the untagged wild type receptor (Table I). In cells labeled with [ 3 H] inositol and stimulated with varying concentrations of NE, both HA-tagged receptors are capable of stimulating production of total inositol phosphates in a dose-dependent manner to the same extent as wild type ␣ 1a ARs, with EC 50 0.36 Ϯ 0.03 and 0.24 Ϯ 0.05 M for HA-␣ 1a and HA-T348, respectively, compared with 0.31 Ϯ 0.09 M for the wild type ␣ 1a AR. The ϳ100-fold difference between NE EC 50 values and receptor affinities (pK i ) at the ␣ 1a AR is not unique to this receptor. Similar differences between agonist potency over ligand affinity has been demonstrated for other GPCRs (e.g. 2-3-fold increase for full agonists of the m3 muscarinic acetylcholine receptor (23) and a nearly 10,000-fold increase for isoproterenol at the ␤ 2 AR (24)). The ␣ 1 AR antagonist prazosin (1 M) prevents NE-induced increases in IP production above basal values at wild type, HA-tagged, and truncated ␣ 1a ARs (data not shown). Thus, addition of the HA epitope tag to the amino terminus and/or truncation of the ␣ 1a AR carboxyl terminus has little effect on ligand binding or second messenger production of the human ␣ 1a AR.

Acute Desensitization of ␣ 1a ARs
We next examined the rate of total IP accumulation following NE stimulation as a function of time in rat-1 fibroblasts expressing HA-␣ 1a or HA-T348. Total IP production rates of both full-length and truncated receptors are identical (Fig. 2). Stimulation of cells with 10 Ϫ5 M NE induces a rapid rise in IP levels for 2 min, after which point the rate of accumulation decreases. This not only suggests that desensitization of ␣ 1a AR occurs at this early time point but also implies that the carboxyl-terminal tail is not required for acute desensitization. In addition, signaling by both receptors continues to rise at an identical, robust, and nearly linear rate through at least 30 min of receptor stimulation. Continued accumulation of total IPs throughout this time period indicates that the agonist-sensitive pool of [ 3 H]inositol-labeled membrane lipids is not rapidly depleted by agonist treatment in these experiments. It also suggests that non-acute mechanisms of desensitization do not predominate over this period.

Time Course of Acute Agonist-and PMA-induced ␣ 1a AR Desensitization
To address the issue of whether or not human ␣ 1a ARs undergo acute desensitization in response to agonist stimulation, we examined the ability of HA-␣ 1a -expressing rat-1 fibroblasts to respond to a subsequent challenge with NE after initial pretreatment with NE. Cells were treated with 10 Ϫ5 M NE for 2 to 30 min in PBS without LiCl (to allow recycling of IPs generated during pretreatment), washed, and then restimulated with 10 Ϫ5 M NE in the presence of LiCl. NE-stimulated IP responses of pretreated cells were compared with those of untreated cells to determine both the time course and extent of HA-␣ 1a desensitization. Desensitization of HA-␣ 1a to subsequent NE stimulation occurs within 2 min of NE agonist pretreatment, reducing the IP response to Ϸ60 -70% that of naive cells (Fig. 3A, Table II). Longer periods of pretreatment with NE result in reduced, but continued receptor responsiveness; even after 30 min of NE stimulation, signaling by HA-␣ 1a persists at 50% naive levels. Significantly, parallel experiments involving the carboxyl-truncated ␣ 1a AR, HA-T348, revealed that both the rate and extent of agonist-induced desensitization are essentially identical to the corresponding wild type responses (Fig. 3A, Table II). Thus, even very short NE pretreatment periods can induce nearly maximal acute desensitization and this behavior is observed even in the HA-T348 receptor lacking the carboxyl terminus. The modest amount of additional desensitization that occurs between 5 and 30 min also appears similar in the full-length and truncated receptor.  Although the intracellular carboxyl terminus has been implicated in regulation of several GPCRs including the ␣ 1b AR, these data suggest that the carboxyl terminus of the human ␣ 1a AR does not play a similar role in regulating agonist-induced desensitization.

Potential PKC Involvement in Acute Agonist-induced Desensitization of ␣ 1a ARs
Agonist stimulation of ␣ 1a ARs leads to activation of the second messenger-dependent kinase PKC, a kinase that could potentially feed back to phosphorylate and desensitize the receptors. To examine a possible role of PKC in ␣ 1a AR desensiti-zation, we utilized the active phorbol ester PMA to stimulate PKC directly. PMA pretreatment of cells expressing HA-␣ 1a or HA-T348 effectively reduces the IP response of both full-length and truncated receptors to a subsequent stimulation by NE, although HA-␣ 1a is desensitized to a greater extent than HA-T348 at each time point (Fig. 3B, Table II). PMA-induced desensitization is dose-dependent, with maximal effects achieved with 10 Ϫ7 M PMA (data not shown). Similar to the time course of NE desensitization, PMA-induced desensitization is rapid, with maximal desensitization occurring within 2 min of PMA treatment. Thus PKC could potentially play a role in agonist-mediated desensitization of ␣ 1a ARs. It is worth noting, however, that while carboxyl-truncated ␣ 1a AR is less sensitive to PMA-mediated desensitization than the full-length receptor, the truncated receptors sensitivity to agonist-mediated desensitization is not different from the full-length ␣ 1a AR. This suggests that second messenger activation of PKC probably does not serve in an agonist-mediated desensitization feedback loop.
To explore this topic further, we utilized the specific PKC inhibitor bisindolylmaleimide I (GF 109203X) to block PKC activity. If PKC plays a substantial role in agonist desensitization, pretreatment of ␣ 1a AR expressing cells with the PKC inhibitor should reduce the desensitizing effects of NE pretreatment. Initial dose-response experiments determined that 1 M bisindolylmaleimide I is sufficient to completely inhibit PMA-mediated effects on ␣ 1a AR signaling (data not shown). When [ 3 H]inositol-labeled cells expressing either full-length and truncated ␣ 1a ARs were treated with bisindolylmaleimide I complete inhibition of PMA-induced desensitization of both receptors was seen (Fig. 4). In contrast, treatment with the PKC inhibitor does not affect the degree to which either receptor is desensitized in response to NE stimulation. These data indicate that although direct PKC activation can lead to desensitization of ␣ 1a ARs, it does not play a significant feedback role in agonist-induced desensitization of ␣ 1a ARs.

GRK-mediated Desensitization of ␣ 1a ARs
After concluding that PKC does not function in a significant feedback capacity in agonist-mediated ␣ 1a AR desensitization, we examined the ability of members of the GRK family to affect ␣ 1a AR signaling. In these experiments, IP accumulation was assessed in rat-1 cells transiently expressing HA-␣ 1a or HA-T348 with or without coexpressed GRK 2 and GRK6 (Fig. 5). Coexpression with GRK2 desensitized both HA-␣ 1a AR and HA-T348 identically, resulting in 63.5 Ϯ 2.4% and 63.8 Ϯ 2.6% of the activity of receptor alone, respectively (Fig. 5, A and B). In contrast, coexpression of GRK6 with HA-␣ 1a AR resulted in no desensitization (Fig. 5C). The ability of GRK2 to desensitize HA-␣ 1a AR was also confirmed in COS-7 cells where it was found to be similar to that observed for HA-␣ 1b AR (data not shown). Although perhaps coincidental, it is worth noting that GRK-mediated desensitization (Fig. 5) and NE-induced (compare with Table II and Fig. 3) desensitization were approximately the same. The identical desensitization observed for full-length and truncated ␣ 1a AR reinforces the evidence presented above indicating that the carboxyl terminus plays little if any role in desensitization of this receptor in rat-1 cells.

Potential Mechanisms Underlying NE-and PMA-induced ␣ 1a AR Desensitization
Phosphorylation of Full-length and Truncated ␣ 1a ARs in Cells-Discovery of rapid desensitization of human ␣ 1a ARs led us to explore possible mechanisms underlying this phenomenon. Rapid phosphorylation of GPCRs by GRKs and/or second messenger-activated kinases such as PKC is thought to lead to receptor desensitization; therefore, we desired to examine phosphorylation of full-length and truncated ␣ 1a ARs in intact cells. Experiments designed to identify ␣ 1a AR proteins showed that photoaffinity labeling of membranes expressing HA-␣ 1a with [ 125 I]azidoprazosin yields a single diffuse band centered ϳ60 kDa (Fig. 6, lane 1). This corresponds to the size of the glycosylated wild type ␣ 1a AR reported previously (25). Labeling of membranes derived from cells expressing HA-T348 yields a diffuse band of ϳ48 kDa, corresponding to the expected gel mobility of the truncated receptor (Fig. 6, lane 2). Prazosin (1 M) blocks labeling of both peptides, confirming specificity of the photoaffinity labeling reaction and identification of ␣ 1a ARs (Fig. 6, lanes 3 and 4).
Phosphorylation of ␣ 1a ARs was examined in rat-1 fibroblasts stably expressing HA-tagged full-length or truncated receptors that were equilibrated with inorganic 32 P to label intracellular pools of ATP. Following drug treatment, cells were lysed and solubilized and the HA epitope-tagged receptors were immunoprecipitated and subjected to SDS-polyacrylamide gel electrophoresis. Treatment of cells expressing HA-␣ 1a or HA-T348 with either NE or PMA clearly resulted in increased phosphorylation of the receptors relative to untreated cells (Fig. 6). This phosphorylation response was both dose-dependent (data not shown) and time-dependent (see below). Less phosphorylation was observed for HA-T348 than for the HA-␣ 1a (Fig. 6). This difference is quantitated in carefully paired time courses presented below. Both HA-␣ 1a and HA-T348 display some basal phosphorylation, although basal phosphorylation of HA-T348 is reduced relative to HA-␣ 1a and is not apparent at the exposure level shown in Fig. 6. Pretreatment of cells with prazosin before NE stimulation prevents agonist-mediated increases in phosphorylation of the receptor (data not shown). The selective ␤-AR antagonist propranolol (10 Ϫ4 M) does not block NE-stimulated 32 P incorporation into the HA-␣ 1a , consistent with the absence of ␤-ARs in rat-1 cells; furthermore, treatment with forskolin/isobutylmethylxanthine (10 Ϫ4 M/10 Ϫ3 M) does not change basal phosphorylation of ␣ 1a ARs, ruling out a possible role for cAMP-dependent protein kinase A (PKA) in either basal or agonist-mediated ␣ 1a AR phosphorylation (data not shown). These data demonstrate that ␣ 1a ARs are phosphorylated in response to both agonist stimulation and direct activation of PKC by PMA. In addition, stimulation with either NE or PMA causes receptor phosphorylation at sites outside of the carboxyl tail of the ␣ 1a AR.
To characterize the temporal correlation between ␣ 1a AR desensitization and receptor phosphorylation, we examined receptors from 32 P-labeled rat-1 fibroblasts expressing either full-length or truncated receptors that were treated with NE or PMA for various periods of time (Fig. 7). In a manner that parallels time courses of both NE-and PMA-induced desensitization of ␣ 1a ARs, both drugs increase HA-␣ 1a phosphorylation on a rapid time scale. Near-maximal HA-␣ 1a phosphorylation in response to NE stimulation is achieved within 2 min, with a slight increase at 5 min, and is sustained throughout 30 min of treatment (Fig. 7A). PMA induces maximal phosphorylation of HA-␣ 1a within the first 2 min of treatment, with PKC-mediated phosphorylation remaining at maximal levels after 30 min of treatment as well (Fig. 7A). Thus, the time frame of phosphorylation of ␣ 1a AR is closely correlated with the temporal occurrence of acute receptor desensitization.
Interestingly, NE-induced phosphorylation of the truncated receptor does display somewhat different behavior than full-length receptor. As for HA-␣ 1a , phosphorylation of HA-T348 is maximal at about 2 min of receptor stimulation during which time 32 P incorporation into the receptor rises 8.9-fold over basal levels (Fig. 7B). However, after 5 min receptor phosphorylation begins to decrease until a steady level of 4.2-fold over basal is achieved after 20 min of receptor stimulation. It is important to remember that HA-T348 remains maximally desensitized during this entire period (refer to Fig. 3A). In addition, quantification of Fig. 7 data (in arbitrary pixel units) suggests that NE-induced phosphorylation of the truncated receptor is approximately half that observed for full-length receptor. This observation could indicate that truncation has eliminated potential phosphorylation sites. However, other explanations are possible; for example, decreased solubilization or immunoprecipitation of HA-T348 compared with full-length receptor. Even if NE stimulation induces phosphorylation of sites in the carboxyl-tail of ␣ 1a AR, the absence of any apparent role for the carboxyl-terminal tail in desensitization (Figs. 2, 3, and 5) suggests the sites do not participate in controlling agonist-mediated desensitization. This does not preclude potential roles of other phosphorylation sites in other regulatory processes.
The time course of PMA effects on HA-T348 phosphorylation is similar to that of the full-length receptor. Maximal 32 P incorporation was achieved between 2 and 5 min of treatment with PMA, continuing through 30 min (Fig. 7B). Quantitation indicates that PMA-induced phosphorylation of HA-T348 is about half that observed for the full-length receptor. In contrast to agonist-mediated desensitization, the carboxyl-terminal tail of ␣ 1a AR does influence PMA-induced desensitization (see Fig. 3B above). This fact is consistent with phosphorylation of the tail playing a role is PKC mediated ␣ 1a AR desensitization. Nevertheless, PMA-induced phosphorylation and partial desensitization of HA-T348 suggests regions other than the carboxyl-terminal tail also participate in these events. As for full-length ␣ 1a ARs, the overall time frame of both NE and PMA-induced phosphorylation of HA-T348 receptors is closely correlated with the temporal occurrence of acute receptor desensitization.
Agonist-induced Internalization of ␣ 1a ARs-Sequestration of receptors from the extracellular surface into various intracellular compartments may occur as a component of both agonistdependent and agonist-independent desensitization and/or resensitization processes. To determine whether receptor sequestration plays a role in acute desensitization of ␣ 1a ARs, we performed radioligand binding on intact rat-1 fibroblasts expressing HA-␣ 1a or HA-T348 with the lipophilic ␣ 1 AR antagonist [ 3 H]prazosin at 4°C, as previously described. Ligand binding at this temperature has been shown to inhibit partitioning of lipophilic prazosin across cell membranes, thus preventing recognition of receptors located intracellularly and allowing quantitation of cell surface receptors; ligand binding at 37°C, on the other hand, allows detection and quantitation of total cellular receptors (26). Effects of NE treatment were similar in cells expressing full-length ␣ 1a ARs and those expressing the truncated receptor. During at least 10 min of agonist stimulation, there is no measurable decrease in the number of cell surface ␣ 1a ARs (Fig. 8A). Between 10 and 30 min of agonist exposure, however, surface receptor numbers decrease by 25% and continue to slowly decline over the next 60 min. Total numbers of either full-length or truncated ␣ 1a ARs do not decline in response to NE or PMA stimulation over this same time period, indicating that loss of surface receptors following NE exposure is due to internalization and not degradation of receptors (data not shown). Thus, internalization of ␣ 1a ARs occurs at a much slower rate than does either phosphorylation or desensitization of the receptors and, therefore, cannot be responsible for rapid ␣ 1a AR desensitization that is seen within the first several min of agonist exposure.
Interestingly, PMA treatment has no effect on the number of full-length or truncated ␣ 1a ARs at the cell surface throughout 90 min of exposure (Fig. 8B). The inability of PMA-mediated stimulation of PKC to induce internalization of ␣ 1a ARs stands in stark contrast to its ability to induce phosphorylation and dampen signaling of these receptors, indicating that these are sharply distinct processes for the ␣ 1a AR. This also suggests that increased receptor phosphorylation alone does not induce ␣ 1a AR internalization, but that other factors are necessary to initiate this process. These ␣ 1a AR regulatory processes are clearly not induced by PKC activity. Furthermore, when cells stably expressing ␣ 1a ARs are treated with the PKC inhibitor bisindolylmaleimide I prior to agonist stimulation, there is no effect on the rate or extent of agonist-mediated internalization of HA-␣ 1a ARs (data not shown). These data provide further evidence that, similar to agonist-mediated ␣ 1a AR desensitization, agonist-induced internalization of ␣ 1a ARs occurs through a PKC-independent pathway.

DISCUSSION
One unique feature of ␣ 1a AR expression is the presence of at least 12 receptor isoforms in various species, as opposed to single isoform expression of ␣ 1b and ␣ 1d AR family members. The four full-length functional human isoforms (WT ␣ 1a-1 , ␣ 1a-2 , ␣ 1a-3 , and ␣ 1a-4 ) differ in amino acid sequence at their distal carboxyl termini. Although the physiological significance of the ␣ 1a AR isoforms remains unclear particularly since their pharmacologic binding properties and second messenger coupling are identical, differences between their carboxyl-terminal amino acid sequences suggest that each ␣ 1a AR isoform may be differentially regulated through this receptor region. For example, 4 potential PKC sites are found in the carboxyl terminus of the ␣ 1a-1 , 3 each in isoforms ␣ 1a-2 , ␣ 1a-3 , and ␣ 1a-4 (15,16). Studies of ␣ 1a AR desensitization in endogenous systems face several potential difficulties, however, since all four functional ␣ 1a ARs and numerous nonfunctional truncated receptor isoforms are found in all tissues studied thus far. To focus on the most abundant ␣ 1a AR isoform in human prostate and heart, we profiled expression of functional ␣ 1a AR isoform mRNAs in these tissues. ␣ 1a-1 is clearly the predominant splice variant in human prostate and heart, as well as in SK-N-MC cells. Although Chang et al. (16) originally suggested that the ␣ 1a-4 mRNA predominates in human prostate (based on reverse transcriptase-PCR experiments), our direct RNase protection assay results clearly demonstrate much higher expression of the ␣ 1a-1 isoform in both prostate and heart. Therefore, we stably transfected epitope-tagged human ␣ 1a-1 AR cDNA in rat-1 fibroblasts to create a system in which acute regulation of the ␣ 1a-1 AR isoform can be studied without confounding coexpression of other splice variants.
Our studies provide convincing evidence that human ␣ 1a ARs are subject to acute agonist-induced desensitization of IP signaling in response to NE stimulation. Desensitization of ␣ 1a ARs is rapid (occurring within 2 min) and is characterized by a rightward shift and lowering of the NE dose-response curve (see Table II). We demonstrate that ␣ 1a AR phosphorylation is significantly increased within 2 min of agonist stimulation and closely correlates with acute agonist-mediated desensitization of ␣ 1a ARs. This agonist-induced process is most likely mediated through actions of at least one GRK as overexpression of GRK2 is able to diminish ␣ 1a AR-mediated IP signaling. These data also suggest that receptor phosphorylation plays an important role in acute agonist regulation of ␣ 1a AR signaling, as has been proposed for GPCRs in general. It is possible that agonist-induced desensitization and phosphorylation of ␣ 1a ARs may be concurrent yet independent events; indeed, the causal effect of protein phosphorylation on receptor desensitization has not been definitively demonstrated for any GPCR. However, mutagenic elimination of protein phosphorylation sites in several GPCRs (including ␣ 1b ARs) does provide compelling evidence that the elimination of receptor phosphorylation sites results in significantly reduced receptor desensitization (8,12).
Internalization of GPCRs from the cell surface may also serve to desensitize receptor populations by removal of receptors from agonist exposure and prevention of further signaling. Agonist stimulation does induce internalization of human ␣ 1a ARs (as demonstrated by loss of cell surface receptor binding). However, since this does not occur until after at least 10 min of agonist exposure, it cannot be responsible for rapid agonist-mediated ␣ 1a AR desensitization that is seen after only 2 min. This finding corresponds with observations that receptor internalization does not play a role in rapid desensitization of several other GPCRs, including the ␤ 2 AR (27), m3-muscarinic receptor (28), and ␣ 1b AR (8).
Agonist-mediated desensitization of ␣ 1a ARs occurs independently of PKC effects, as demonstrated by the failure of PKC inhibition to affect this process. A similar lack of phospholipase C involvement in agonist-mediated desensitization of several other phospholipase C-coupled receptors including substance P (29), m3-muscarinic (30), thromboxane (31), and ␣ 1badrenergic receptors (8) has also been demonstrated. In our hands, human ␣ 1a ARs are nevertheless subject to PKC-mediated desensitization as phorbol ester (PMA) pretreatment substantially decreased IP production resulting from subsequent NE stimulation of the ␣ 1a AR. PMA treatment also leads to a rapid increase in ␣ 1a AR phosphorylation that closely correlates with the onset of PMA-induced desensitization. In this context, it is noteworthy that PMA treatment does not induce internalization of ␣ 1a ARs from the cell surface. Conversely, PKC inhibition does not affect agonist-mediated internalization of ␣ 1a ARs from the cell surface. These data support the hypothesis that exposure to agonist and PMA induces distinct and possibly separate ␣ 1a AR regulatory responses. Although our data do not support a specific role of PKC "feedback" in agonistinduced desensitization of ␣ 1a ARs, this second messenger-dependent kinase may play a role in modulating the ability of ␣ 1a ARs to signal following stimulation of other PLC-coupled receptors within the same cell. This hypothesis is supported by recent demonstrations that stimulation of endothelin ET A receptors or bradykinin B 2 receptors in rat-1 fibroblasts leads to a marked increase in phosphorylation of stably expressed ␣ 1b ARs (32,33).
Perhaps our most significant finding is that mechanisms underlying acute agonist-dependent regulation of human ␣ 1a AR signaling appear to function independently of the receptor carboxyl terminus. This is demonstrated by equal sensitivity of the truncated and full-length ␣ 1a AR to NE-induced desensitization. In addition, GRK2 desensitizes signaling by the truncated and full-length receptor equally, suggesting that the agonist-mediated pathway of desensitization remains intact in the absence of the ␣ 1a AR carboxyl terminus. On the other hand, sensitivity of the truncated receptor to PMA-induced desensitization is less than that of the full-length ␣ 1a AR. These data provide further evidence for the hypothesis that agonist-and PMA-induced desensitization proceed through separate, distinguishable mechanisms, and indicate that although the ␣ 1a AR carboxyl terminus plays a substantial role in PMA-mediated desensitization, it is not required in the agonist-mediated pathway. This highlights a significant regulatory difference between human ␣ 1a ARs and the ␣ 1b AR subtype as the carboxylterminal of the ␣ 1b AR plays an indispensable role in mediating both agonist-and PMA-induced phosphorylation and desensi-tization (8,11,12). In addition, our results place the human ␣ 1a AR as the only G q -coupled member of a very small group of GPCRs whose mechanisms of desensitization are primarily independent of carboxyl terminus regulation (e.g. follitropin receptor (34), ␣ 2 ARs (35), and the angiotensin II receptor (36)).
As acute desensitization of truncated ␣ 1a ARs predicts, NE and PMA treatment each lead to significantly increased phosphorylation of the truncated ␣ 1a AR. However, HA-T348 is modestly less phosphorylated than the full-length receptor under all conditions. This is not an unexpected result, given that potential phosphorylation sites were eliminated with truncation of the ␣ 1a AR after Arg 348 . For example, the human ␣ 1a-1 AR contains 4 putative PKC phosphorylation sites within the carboxyl terminus as well as several potential sites in other intracellular regions of the receptor, including 1 in the second intracellular loop and 4 in the third intracellular loop. Thus truncation eliminates half of the potential PKC phosphorylation sites, perhaps explaining why the truncated ␣ 1a AR cannot be desensitized as completely as the full-length receptor. Of course the truncated receptor still displayed substantial PMAmediated desensitization concurrent with rapid phosphorylation strongly suggesting involvement of sites not in the tail.
Possible sites of GRK-mediated ␣ 1a AR phosphorylation are more difficult to predict, since consensus sequences for receptor recognition and phosphorylation by individual GRKs are not as clearly defined. Early studies suggested that GRK2 and GRK3 preferentially phosphorylate serines or threonines in proximity to acidic amino acids (37) and Fredericks et al. (38) has observed that several receptors that are phosphorylated by GRKs contain a pair of acidic residues on the amino-terminal side of the phosphorylated residue. The human ␣ 1a AR does not contain acidic pairs of residues, however, several potential GRK phosphorylation sites (represented by serines and threonines in the vicinity of acidic residues) are found within the carboxyl terminus and, more interestingly, within intracellular loops of the human ␣ 1a AR. Experiments presented here indicate that elimination of some potential GRK phosphorylation sites by truncation does not affect the ability of the ␣ 1a AR to undergo acute agonist-mediated desensitization even if overall receptor phosphorylation is decreased. Furthermore, agonist-induced phosphorylation of the truncated ␣ 1a AR occurs concomitant with acute desensitization. These data are consistent with the hypothesis that phosphorylation of only a discrete subset of amino acids (retained within the truncated human ␣ 1a AR) regulates agonist-mediated desensitization of the ␣ 1a AR, although other sites may be phosphorylated that do not affect receptor desensitization. It is worth noting that the agonistinduced phosphorylation increase (over basal) was higher for HA-T348 (8.9-fold) than for full-length HA-␣ 1a AR (4.5-fold). It could certainly be the case that fold increases in phosphorylation can influence desensitization as much as absolute phosphorylation levels. If so, the fact that decreasing HA-T348 phosphorylation remains at 4.2-fold over basal even at the longest NE exposure times (30 min), may explain continued maximal desensitization. On the other hand, at times far removed from acute desensitization it is probable that mechanisms directed at downstream elements of the IP cascade are functioning.
While our manuscript was in preparation, a study of phosphorylation of the bovine ␣ 1a AR stabily expressed in rat-1 fibroblasts was published by Vazquez-Prado and colleagues (14). Both studies observe that exposure of the ␣ 1a AR to NE causes rapid receptor phosphorylation (1 to 2 min), generating a 5-10-fold increase in phosphorylation compared with basal. Vazquez-Prado and colleagues (14) also suggested that phosphorylation of ␣ 1a AR occurred primarily in the carboxyl-termi-nal tail; however, the evidence presented was indirect. While phosphorylation of the chimeric ␣ 1ba AR (consisting of the unphosphorylatable transmembrane region of ␣ 1b AR (12) and carboxyl terminus of ␣ 1a AR) does imply phosphorylation of the terminus, it does not demonstrate this phosphorylation is dominant or functionally significant. Indeed, no functional assays were performed using the chimeric protein. On the other hand, data presented in our current study are clear and direct; we demonstrate that the truncated ␣ 1a AR (HA-T348) is phosphorylated in response to NE stimulation in the same rapid manner as full-length receptor and furthermore, has identical desensitization properties. Although, phosphorylation of the carboxyl terminus may occur, it is not involved in agonistmediated desensitization as deletion of the tail did not change any characteristic of desensitization.
In conclusion, our findings demonstrate that ␣ 1a AR signaling is subject to acute desensitization in response to NE as well as PMA exposure. In agreement with current paradigms, acute desensitization appears to be mediated by receptor phosphorylation since both occur simultaneously within 2 min of NE exposure. The GRK family of kinases is probably responsible for agonist-mediated phosphorylation as PKC was not involved and GRK2 was capable of desensitizing the ␣ 1a AR. Study of a truncated ␣ 1a AR provides definitive evidence that ␣ 1a AR phosphorylation is not localized to the carboxyl terminus alone, but that other receptor regions are also phosphorylated (presumably one or more intracellular loops). In addition, characterization of the functional consequences of carboxyl-terminal truncation of the human ␣ 1a AR also shows that this receptor region is not necessary for full agonist-induced receptor desensitization and internalization. Not only does carboxyl-terminal independent regulation of ␣ 1a ARs represent a significant mechanistic difference between ␣ 1a and ␣ 1b AR family members; it also suggests a common regulatory mechanism between ␣ 1a AR splice variants. Although differences may be discovered between ␣ 1a AR isoform regulation upon closer examination of individual ␣ 1a AR splice variants, our data suggest a common behavior is possible and perhaps likely. Our findings also underscore the importance of examining receptor-specific regulation, particularly in cases where distinct differential agonistmediated regulatory processes are known to occur such as within the ␣ 1 AR family. The challenge now is to identify specific amino acid residues that are involved in human ␣ 1a AR desensitization and begin to unravel the relationship of acute desensitization to other regulatory pathways such as receptor trafficking and gene transcription.