A recombinant adeno-associated viral (rAAV) vector was designed to include tandem cassettes encoding the human β₂AR and enhanced green fluorescent protein (EGFP) genes and was designated AAV-β₂AR/EGFP (Figure 1). To evaluate for efficiency of viral unit transfer into AAV-β₂AR/EGFP infected cells, the detection of EGFP was used as a surrogate or screening marker for β₂AR expression. HEK 293 cells were visualized using epifluorescence microscopy. Approximately 40% of cells infected with AAV-β₂AR/EGFP (200 transducing units/cell) demonstrated green fluorescence (Figure 2), while mock infected cells lacked EGFP expression (data not shown). These results indicate that HEK 293 cells are readily infected with a recombinant AAV and that the EGFP cassette was expressed.

To determine the pharmacologic characteristics of the recombinant β₂AR, we used HEK 293 cells because of their low endogenous expression of β₂AR. We first sought to determine the characteristics of the expressed receptor in saturation binding experiments. [³H]dihydroalprenolol ([³H]DHA) binding to membranes prepared from AAV-β₂AR/EGFP-infected HEK 293 cells was to a single, saturable site that displayed high affinity as shown in a representative Scatchard plot (Figure 3). Separate experiments with four different membrane preparations established a binding site concentration (B_max) of 5.05 ± 1.0 pmol/mg protein (n = 4) and a dissociation constant (K_d) of 0.42 ± 0.1 nM (n = 4). These findings demonstrate [³H]DHA binding to a single population of receptors with affinity expected for wild-type β₂AR 13.

The specificity of [³H]DHA binding was examined in competition binding assays using various adrenergic agonists (Figure 4). In five separate experiments, the rank order potency of agonist binding to membranes prepared from HEK 293 cells infected with AAV-β₂AR/EGFP was isoproterenol (K_i = 1.9 ± 0.7 μM) > epinephrine (K_i = 5.7 ± 2.5 μM) > norepinehrine (K_i = 22.8 ± 7.7 μM) (n = 5). This rank order potency is consistent with a β₂AR interaction.

HEK 293 cells express low levels of β₂AR 14. To determine the capability of AAV-β₂AR/EGFP to increase β₂AR expression in HEK 293 cells, ligand binding assays were employed in AAV-β₂AR/EGFP infected and mock-infected cells grown in DMEM supplemented with 10% FBS. Mock infected HEK 293 cells demonstrated specific binding of [³H]DHA to a single saturable site at a level of 39 ± 11 fmol/10⁶ cells. β₂AR levels were significantly (p < 0.001) increased in AAV-β₂AR/EGFP infected cells to 501 ± 82 fmol/10⁶ cells, representing a 13-fold increase in β₂AR expression levels when comparing AAV-β₂AR/EGFP infected cells to mock-infected cells (Figure 5). To further assess the role of serum source on β₂AR expression in infected HEK 293 cells, we conducted similar studies using 5% CS. In cells cultured in DMEM with 5% CS, background β₂AR expression was lower than in cells grown in 10% FBS, with mock-infected cells showing β₂AR levels of 5.5 ± 3.4 fmol/10⁶ cells. β₂AR levels were significantly increased (p < 0.001) in AAV-β₂AR/EGFP infected cells to 428 ± 95 fmol/10⁶ cells, representing a 77-fold increase in β₂AR levels when comparing AAV-β₂AR/EGFP infected cells to mock-infected cells grown in 5% CS (Figure 5). This dramatic increase in receptor expression when comparing cells grown in 5% CS to those grown in 10% FBS was due to differences in baseline β₂AR expression in mock-infected cells. Interestingly, the absolute level of β₂AR expression after AAV-β₂AR/EGFP infection was not different between culture conditions. Overall, these results indicate that β₂AR levels can be significantly increased in HEK293 cells infected with AAV-β₂AR/EGFP, but that there may be an upper limit for membrane expression of β₂AR in this cell line.

Binding of agonist to the β₂AR results in adenylyl cyclase activation and conversion of ATP to cyclic AMP 15. To evaluate the ability of the recombinant β₂AR to activate early receptor signaling pathways, isoproterenol-stimulated cyclic AMP accumulation was measured in HEK 293 cells infected with AAV-β₂AR/EGFP (Figure 6). Cells were treated with the phosphodiesterase inhibitor, IBMX, at the time of isoproterenol treatment to maximize the cyclic AMP signal. In mock-infected (control) cells, cyclic AMP accumulation was 4.83 ± 0.42 nmoles/dish in the absence of isoproterenol and 63.4 ± 9.6 nmoles/dish in the presence of isoproterenol, representing a 13-fold increase in cyclic AMP accumulation in isoproterenol-treated, mock infected cells. In AAV-β₂AR/EGFP infected cells, cyclic AMP accumulation increased from 4.69 ± 0.84 nmoles/dish in the absence of isoproterenol stimulation to 314 ± 43 nmoles/dish in the presence of isoproterenol, representing a 67 fold increase in cyclic AMP accumulation. The increase in cyclic AMP production in AAV-β₂AR/EGFP infected cells was significantly different from control, mock infected cells (p < 0.05). These data indicate that in addition to binding agonists with the expected pharmacologic specificity, the recombinant β₂AR was capable of interacting with downstream intracellular signaling proteins to stimulate cyclic AMP accumulation.

Previous reports indicate that ligand-induced trafficking of the β₂AR begins in the early endosome 1416. Through further intracellular signaling, the internalized β₂AR is then either recycled to the plasma membrane or is committed to a degradation pathway terminating in the lysosome 17. To determine if the recombinant β₂AR expressed from AAV-β₂AR/EFGP retains receptor trafficking in HEK 293 cells, receptor distribution was assessed using a polyclonal antibody to the cytoplasmic tail of the β₂AR labeled with a Texas Red fluorochrome. Recombinant receptors were localized to the cell surface after treatment with vehicle alone, with minimal evidence for intracellular distribution (Figures 7, Panel A). Following isoproterenol treatment for 20 minutes, recombinant β₂AR were observed to move from the cell surface to small, punctate intracellular vesicles with minimal surface expression noted (Figure 7, Panel B). Following isoproterenol treatment for 24 hours, recombinant β₂AR were noted to traffick to both large and small, perinuclear vesicles as would be expected for wild-type receptors following prolonged agonist exposure (Figure 7, Panel C). Additionally, images obtained after 24 hour agonist treatment suggest that some receptors were located on the plasma membrane possibly due to efficient recycling mechanisms as is seen with native β₂AR 18 or due to the abundance of expressed β₂AR. These results indicate that agonist induced trafficking of recombinant β₂AR remains intact with ligand-induced internalization of receptor but with retention of some cell surface expression, even after prolonged agonist exposure. These results further suggest that an added benefit of recombinant β₂AR expression is persistence of β₂AR on the cell surface in the continuing presence of agonist.

A recombinant adeno-associated viral (rAAV) vector was designed to include tandem cassettes encoding the human β₂AR and enhanced green fluorescent protein (EGFP) genes and was designated AAV-β₂AR/EGFP (Figure 1). Cassettes containing the β₂AR and EFGP genes, both driven by CMV promoters, were cloned into pAV53-LR, a plasmid vector containing the internal terminal repeats (ITRs) from AAV (provided by Dr. Juinyan Dong, Medical University of South Carolina, Charleston, SC). Briefly, the β₂AR gene was PCR-amplified from human genomic DNA using a forward primer (5'CATATA

AAGCTT-

GGATCC

GCATGC

GCATGC

The human embryonic kidney cell line, HEK 293, was used for all experiments. HEK 293 cells were grown in DMEM supplemented with 10% fetal bovine (FBS). Supplemental studies assessing the role of growth media on receptor expression were conducted using 5% calf serum (CS) in place of FBS. HEK 293 cells at a cell density of 0.25 × 10⁶ cells/well in 6 well plates were transduced by addition of AAV-β₂AR/EGFP (200 transducing units/cell in 1 ml of media per well). Approximately 16 hrs after initial viral application, 1 ml of growth media was added to each well. Assays to determine β₂AR expression levels and function were performed on day 4 following infection.

Partially purified membrane preparations were obtained from AAV-β₂AR/EGFP infected HEK 293 cells, cultured in DMEM with 10% FBS, by differential centrifugation as previously described 13. Briefly, cells were washed with ice-cold phosphate-buffered saline (PBS) and scraped into ice-cold PBS with a rubber policeman. The cells were centrifuged at 250 × g for 5 minutes, resuspended in assay buffer (50 mM pH 7.4 Tris-HCl, 2 mM MgCl₂) and homogenized with a glass-glass homogenizer followed by sonication (5–10 second bursts at setting 6) with a Tekmar Model AS1 Sonic Disrupter. The nuclei were removed by centifugation at 600 × g for 10 minutes. Membranes were obtained from the resulting supernatant by centrifugation at 30,000 × g for 15 minutes, then resuspended in assay buffer and centrifuged again. The final pellets were resuspended in assay buffer, aliquoted, and stored at -80°C. Protein concentrations of membrane preparations were determined by the method of Bradford 25 using bovine serum albumin as the standard. [³H]Dihydroalprenolol (DHA) (Dupont-NEN, Boston, MA; specific activity = 120 Ci/mmole) was used to identify β₂AR as previously described 13. In saturation experiments, aliquots of HEK 293 cell membranes (final concentration in assay tube = 70 μg/ml) were incubated with 7 different concentrations of [³H]DHA ranging from approximately 0.05 to 5 nM. In competition experiments, membrane aliquots were incubated with approximately 1 nM [³H]DHA and increasing concentrations of the competitors isoproterenol, epinephrine, and norepinehrine (range 10^-9 to 10^-4 moles/liter). Nonspecific binding was defined with 0.1 μM (-)-propranolol. Data from saturation experiments were analyzed using LIGAND 26. Inhibition constants were calculated using the method of Cheng and Prusoff 27.

The effects of β-agonist treatment on β₂AR expression were determined by growing AAV-β₂AR/EGFP infected HEK 293 cells in DMEM containing 10% FBS. To determine the impact of a less enriched media on β₂AR expression, infected HEK 293 cells were also cultured in DMEM with 5% CS. [³H]DHA was used in ligand binding assays to determine β₂AR levels as previously described 13. Approximately 1.2 × 10⁶ cells/ml were incubated in triplicate with a single saturating concentration of [³H]DHA (~5 nM). Nonspecific binding was defined with 0.1 μM (-)-propranalol.

Both AAV-β₂AR/EGFP infected and mock-infected HEK 293 cells were cultured in DMEM with 10% FBS then in serum-free media overnight. For cyclic AMP determination, cells were then treated either with vehicle or 10 μM isoproterenol and the phosphodiesterase inhibitor, isobutylmethylxanthine (IBMX, 250 μM), for 10 minutes. Cellular cyclic AMP levels were determined by radioimmunoassay using the Biotrak CAMP Assay System (Amersham Life Sciences, Arlington Heights, IL).

For fluorescence microscopy, HEK 293 cells were cultured in DMEM with 10% FBS at a density of 2.5 × 10⁵ cells/well on glass coverslips, infected with AAV-β₂AR/EGFP and treated on day 4 with vehicle or 10 μM isoproterenol for 24 hrs at 37°C. Cells were fixed with 1% paraformaldehyde at the designated time intervals. Efficiency of cell infection was evaluated through imaging of green fluorescence as an indicator of EGFP expression. β₂ARs were detected with a rabbit polyclonal antibody specific to the cytoplasmic tail of the human β₂AR (1:500 dilution; Bethyl Laboratories, Montgomery, TX) and Texas Red-labeled (red fluorescence) goat anti-rabbit IgG antibody (1:200 dilution; T2767; Molecular Probes, Eugene, OR) as previously described 28. Fluorescence imaging was then performed with a Zeiss Axiovert digital deconvolution microscope (Carl Zeiss, Inc., Thornwood, NY). For EGFP expression, cells were visualized using epifluorescence microscopy with a 100× oil objective. For β₂AR detection, images were collected using a 100× oil objective in Z-stacks then digital deconvolution was performed using AxioVision 3.1 (Carl Zeiss, Inc.). Images were then converted to tagged-image files (tiff) for comparison.

Data are presented as the mean ± S.E.M. Comparisons between groups were made by using one-way analysis of variance (ANOVA) with Newman-Keuls post hoc testing. The 0.05 level of probability was accepted as significant. Computations were performed using the SigmaStat software package (Jandel Scientific, San Rafael, CA).