Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Post-transcriptional control mechanisms participate in the stability and degradation of various cellular RNAs, including those encoding G protein-coupled receptors (Nielsen and Shapiro, 1990; Schiavi et al., 1992; Jackson, 1993). AU-rich elements (AREs), identified in the 3' untranslated regions (UTRs) of these mRNAs, initiate deadenylation as the first step in mRNA turnover (Greenberg et al., 1990; Brewer, 1991; Zubiaga et al., 1995; Xu et al., 1997). AREs may consist of AUUUA pentamers, AU-nonamers, and U-rich elements; specific combinations determine the degree of destabilization function. AREs have been divided into three classes (Xu et al., 1997): class I and II mRNAs contain multiple copies of AUUUA pentamers and display either processive or distributive deadenylation, respectively; class II mRNAs result in the formation of poly(A) intermediates. Class III mRNAs do not contain AUUUA pentamers and undergo processive deadenylation. Clustering of multiple copies (5+) of the AUUUA pentamer seems to be important in the selection of processive versus distributive deadenylation (Xu et al., 1997). The optimal destabilization motif for c-fos mRNAs has been previously determined to be UUAUUUA(U/A)(U/A) (Zubiaga et al., 1995). However, reiteration of the AUUUA pentamer often results in the formation of overlapping nonamers; it is uncertain whether the multiple clustering of the core pentamer or the presence of the overlapping nonamer represents the key determinant for initiation of mRNA degradation (Xu et al., 1997).

Among factors implicated in mRNA stability are several classes of RNA binding proteins (Derrigo et al., 2000). The AU-rich binding/degradation factor (or AUF1), also identified as the hnRNP D protein, is an important factor in the ARE-mediated decay pathway (Zhang et al., 1993; Loflin et al., 1999). AUF1 has been demonstrated to bind to AREs in transcripts encoding G protein-coupled receptors (Zhang et al., 1993, Pende et al., 1996) and undergoes high-affinity binding to AREs as a hexameric protein (DeMaria et al., 1997, Chkheidze et al., 1999). Other hnRNPs, including hnRNP A1 and hnRNP C, also bind to reiterated AUUUA sequences (Hamilton et al., 1993). The Hu proteins, primarily neuronal RNA binding proteins, are composed of four family members, HuA (or HuR), HuB, HuC, and HuD, that recognize AUUUA pentamers or derivative sequences (Ma et al., 1996; Akamatsu et al., 1999). HuR recognizes the UTRs of several transcripts, and has been implicated primarily in RNA stability (Levy et al., 1998; Ford et al., 1999) and most recently in RNA decay (Maurer et al., 1999).

Several G protein-coupled receptor mRNAs, including those encoding the ₁-adrenergic (Izzo et al., 1990), m₁-muscarinic (Lee et al., 1994), and ₂-ARs (Hadcock et al., 1989), undergo agonist-induced decreases in transcript stability. The ₂-AR mRNA contains an AU-rich 3' UTR that can recognize several RNA binding proteins, including 1) the -adrenergic receptor mRNA-binding protein (ARB), a cytosolic protein that requires both AUUUA pentamers and U-rich domains for RNA recognition and 2) the M_r 37,000 RNA degradation factor AUF1 (Port et al., 1992; Huang et al., 1993; Pende et al., 1996). In the presence of -AR agonist, these RNA binding proteins seem to be up-regulated, concomitant with the down-regulation of ₂-AR mRNA. In humans undergoing cardiac failure, AUF1 and norepinephrine levels are increased, concomitant with significant decreases in ₁-AR mRNA and receptor (Pende et al., 1996). Specific AU-rich elements in the ₂-AR 3' UTR, including a 20-bp core sequence, have been demonstrated to be essential for the agonist-mediated destabilization of ₂-AR mRNA (Tholanikunnel and Malbon, 1997; Danner et al., 1998; Tholanikunnel et al., 1999). Using chimeric receptor -globin recombinants, the human ₁-AR mRNA 3' UTR was determined to target -globin mRNA for accelerated decay and was demonstrated to be necessary but not sufficient for agonist-mediated destabilization of transcripts (Mitchusson et al., 1998). Using mass spectrometry, ARB has recently been identified as a composite of HuR and hnRNP A1 (Blaxall et al., 2000). HuR seems to be the predominant RNA binding component of ARB; however, the precise component composition of ARB seems to be dependent on the 3' UTR sequence of the target -AR subtype mRNA (Blaxall et al., 2000).

In this report, we examined the agonist-independent and agonist-mediated regulation of the rat ₁-AR mRNA, endogenously expressed in C6 glioma cells and ectopically expressed in transfectant hamster DDT₁MF2 and rat L6 cells, and have identified factors that may potentially influence post-transcriptional ₁-AR mRNA decay. Our observations are consistent with the interpretation that the HuR proteins may play important roles in the agonist-mediated down-regulation of ₁-AR mRNAs and that expression of this factor may also influence agonist-independent ₁-AR mRNA degradation.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Lines, Antibodies, Expression Recombinants, and Statistical Analyses. The rat C6 glioma (American Type Culture Collection, Manassas, VA), the hamster DDT₁MF2 (American Type Culture Collection), and the rat L6 skeletal muscle myoblast (American Type Culture Collection) cell lines were cultured in Dulbecco's modified Eagle's medium (DMEM), supplemented with 10% heat-inactivated fetal calf serum, 1% penicillin/neomycin/streptomycin, and 1% L-glutamine. Low-glucose (1.5 g/l) DMEM was used to culture the C6 and L6 cell lines. High-glucose (4.5 g/l) DMEM was used to culture the DDT₁MF2 cell line. Antibodies were obtained that recognize HuR, the mammalian elav-like RNA binding proteins (Henry Furneaux, Cornell University School of Medicine, Ithaca, NY), and the heterogeneous nuclear ribonuclear proteins (hnRNP A1; provided by Gideon Dreyfuss, Howard Hughes Medical Institute, University of Pennsylvania School of Medicine, Phildelphia, PA). The monoclonal antibody recognizing the bacterial Flag antigen (anti-Flag M2 monoclonal) was obtained from Sigma (St. Louis, MO). The pSVZeo eukaryotic expression recombinants, containing the complete HuR coding sequence in sense or reverse orientations, were obtained from Andy Levy (Technion Faculty of Medicine, Haifa, Israel). Statistical analyses used to ascertain significance of ₁-AR mRNA half-lives were based on the two-tailed unpaired t test derived from the Microsoft Excel software program (Microsoft, Redmond, WA).

Construction of ₁-AR Recombinants Used for Transfection. We performed a two-step subcloning strategy to obtain the complete ₁-AR coding region and 3' UTR. To obtain the ₁-AR region 23 to +1466, two primers were engineered, a 5' primer containing an XbaI site and ₁-AR sequence from 23 and a 3' primer containing a BamHI site and ₁-AR sequence ending at +1466. Molecular clone pCoding plasmid DNA (50 ng) containing ₁-AR sequence 82 to +1573 was used as template for PCR amplification. To obtain the ₁-AR region +1467 to +2794, two additional primers were engineered, a 5' primer containing a BamHI site and ₁-AR sequence from +1467 and a 3' primer containing an EcoRI site and ₁-AR sequence ending at +2794. Molecular clone pST3000-18 plasmid DNA (50 ng) containing ₁-AR sequence +1165 to +2967 was used as template for PCR amplification. PCRs were conducted using a core buffer consisting of 20 mM Tris-HCl, pH 8.75, 10 mM KCl, 10 mM NH₄SO₄, 2 mM MgSO₄, 0.1% Triton X-100, 100 µg/ml bovine serum albumin; 200 µM concentrations of each dNTP, 100 pmol of each primer, and 2.5 units of Pyrococcus furiosus DNA polymerase (Stratagene, La Jolla, CA). The reaction mixture was subjected to an initial denaturation at 95°C for 1 min, followed by 35 cycles of 95°C for 1 min, 55°C for 1 min, and 68°C for 2 min, with a final extension at 68°C for 20 min followed by a 4°C soak. Each PCR product was then subcloned into pGEM 3Zf(+) (Promega, Madison, WI), to develop large-scale quantities of the insert. Both recombinants were digested with BamHI and EcoRI, and subjected to a three-way ligation into pGEM 3Zf(+). The 2.7-kilobase insert containing the complete ₁-AR coding sequence plus 3' UTR was subcloned into the pcDNA 3.1() eukaryotic expression vector (Invitrogen, Carlsbad, CA). To confirm insert orientation and sequence, a majority of the recombinant was sequenced.

Cloning of the Hamster Cyclophilin Sequence by Reverse Transcription-Polymerase Chain Reaction. The hamster cyclophilin probe is composed of sequence extending from +63 to +273, relative to the translational start site, and was prepared by reverse transcription-PCR using DDT₁MF2 RNA as template. First strand synthesis was conducted (42°C, 30 min) using 5× first strand buffer (4 µl; 250 mM Tris-HCl, pH 8.3, 375 mM KCl, and 15 mM MgCl₂), 2.5 mM concentrations of each dNTP (2 µl), RNasin (2000 units), 1 mg/ml bovine serum albumin (2.5 µl), 0.1 M dithiothreitol (0.5 µl), reverse transcriptase (100 units; Moloney murine leukemia virus, Invitrogen), DDT₁MF2 RNA (2 µg), and first strand primer (3'GACTCGAGGATCCA-TCGATTTTTTTTTTTTTTTTT5'). PCR was conducted using 10× PCR buffer (10 µl; 500 mM KCl, 100 mM Tris-HCl, pH 9.0, 1% Triton X-100; Bio Basic, Inc., Toronto, ON, Canada), 15 mM MgCl₂ (8 µl), 2.5 mM concentrations of each dNTP (8 µl), Taq polymerase (5 U), first-strand synthesis products (5 µl), and 100 pmol of upstream and downstream primer in a 100-µl total volume. The PCR mixture was subjected to an initial denaturation at 94°C for 1 min, followed by 30 cycles of 94°C for 1 min, 54°C for 45 s, and 72°C for 30 s. Upstream and downstream primers were derived from the Chinese hamster ovary (CHO) cyclophilin sequence available on GenBank (accession number X17105). The CHO and DDT₁MF2 hamster cyclophilin sequences differ in two bases within a common 211 bp sequence (position 93-303); there is a C-to-T conversion at position 235 and an A-to-G conversion at position 275 (in both cases, CHO-to-DDT₁MF2 sequence conversions). The DDT₁MF2 hamster cyclophilin sequence has been deposited in GenBank and has been assigned accession number AY004869.

Construction of the ₁-AR 3' UTR Recombinants and Syntheses of cRNA Probes for RNA Mobility Shift Assays. PCR was used to generate probes encompassing the ₁-AR 3' UTR subregions P1, P2, P3, and P4 (subregions 2187-2306, 2307-2454, 2455-2590, and 2591-2736, respectively), and the control ₁-AR upstream sequence containing subregion 408 to 323, relative to the translational initiation codon. Molecular clone DNA, containing the entire rat ₁-AR gene (50 ng), was used as PCR template. The PCR fragments encompassing the P1, P2, P3, and P4 subregions were individually subcloned by blunt-end ligation into pGEM3Z(+) (Promega). The resulting 3' UTR recombinants, pGEM-P1, pGEM-P2, pGEM-P3, and pGEM-P4, and the upstream recombinant p₁-AR[408,323], were verified by sequence analyses and used to transcribe sense cRNA probes for RNA mobility shift assays. In vitro transcription of cRNA probes is described in Kirigiti and Machida (1999).

Transfection Procedures. pcDNA 3.1 recombinants containing the complete rat ₁-AR coding region and 3' UTR (nucleotide position is from 23 to +2786) were electroporated into the DDT₁ MF2 or L6 cell lines (300 V, 900 µF capacitance; 3.0 × 10⁶ cells in 0.3 ml); transfectants were then subjected to neomycin selection (500 µg/ml G418 [Geneticin; Invitrogen] for 10-14 days) to generate pooled stable clones expressing rat ₁-AR mRNAs. pSVZeo recombinants containing the HuR sequence in either sense or reverse orientation were electroporated into C6 cells using transfection parameters described above. Transfectants were then subjected to zeocin selection (600 µg/ml, product number 45-0430; Invitrogen) for 14 days to generate pooled stable clones; zeocin selection was continued for an additional 2 weeks or more before conducting agonist treatment experiments.

RNase Protection Assays Total RNA was isolated using the QIA RNeasy Mini Kit, a product of QIAGEN, Inc. (Valencia, CA). Methods for RNA quantification, determination of sample integrity, and RNase protection assays (RPA) were conducted as described in Yang and Machida (1999). The probe pCS[82, +273], which contains rat ₁-AR sequence extending from position 82 to +273 (relative to the translational start site), recognize all ₁-AR transcripts and was used to measure total ₁-AR mRNAs in RPA experiments. The hybridization probe used to discriminate the two polyadenylated forms of ₁-AR mRNA in C6 cells was p4Z1 probe [+2084, +2901], which contains ₁-AR sequence extending from +2084 to +2901. Antisense transcripts were generated with T7 RNA polymerase (20 units, 37°C, 1 h) using reaction components described in Yang and Machida (1999).

RNA Mobility Shift Assay and UV-Cross-Linking of RNA-Protein Complexes for SDS-Polyacrylamide Gel Electrophoresis. Preparation of cell extracts and binding reactions were conducted as described in Kirigiti and Machida (1999). RNA mobility shift assay experiments were electrophoresed in 4% nondenaturing polyacryamide gels and subjected to autoradiography. For use in protein size determinations, binding reactions were conducted using cellular extract (20 µg) and 1 × 10⁶ cpm of ³²P-labeled cRNA probe for 20 min at room temperature (in a 50-µl final volume containing 4 mM dithiothreitol, 5 µg of yeast tRNA, 5 µg of heparin, and 50 units of Super RNasin). The reaction mixtures were then placed on ice and subjected to short wave (254 nm) ultraviolet irradiation for 5 min in a cross-linking oven (model FB-UVXL 100; Fisher Biotech, Belmont, WA). Noncovalently bound ribonucleotides were degraded by addition of RNase T1 (1 unit) and RNase A (1 µl of 0.5 µg/µl) for 30 min at 37°C. In some cases, as with the cross-linking associated with the hnRNP A1 analyses, RNase T1 digestion preceded cross-linking of radioactive probe with RNA binding proteins. The reaction mixtures were added to equal volumes of 2× loading buffer (100 mM Tris-HCl, pH 6.8, 200 mM dithiothreitol, 4% SDS, 2% bromphenol blue, and 20% glycerol), boiled for 5 min, and subjected to electrophoresis in a 10% polyacrylamide gel (acrylamide/bis-acrylamide, 29:1; supplemented with SDS). Gels were transferred to nitrocellulose membranes before autoradiography (85°C; 30-72 h) or analysis with the Bio-Rad Molecular Imager FX system.

Immunoprecipitation of UV-Cross-Linked RNA-Protein Complexes. In some cases, to enhance the autoradiographic signal of immunoprecipitated proteins, 5-fold additional probe and 3-fold additional protein were used in UV-cross-linking experiments. After UV-cross-linking of RNA-protein complexes and RNase A/T1 digestion, 1 to 2 µl of specific antibody was added to the reaction mix and allowed to incubate overnight at 4°C. Phosphate-buffered saline (PBS) was added to the mixture to bring the final volume to 200 µl. Protein A slurry (20-40 µl of a 1:1 mixture of PBS and protein A-Sepharose 4B beads; Sigma) was then added, and allowed to mix for 2 h at 4°C. The protein A beads were then washed three times by gentle centrifugation (2000g, 2 min) in PBS, and then resuspended in 6× loading buffer, before heating at 100°C for 2 min. RNA-protein complexes released from the protein A beads were then loaded into SDS-polyacrylamide gels. Electrophoresis and autoradiographic or molecular imaging analyses were conducted as described above.

Electrotransfer to Nitrocellulose Membranes and Immunoblotting Procedures. Proteins were electro-transferred onto nitrocellulose membranes, and blots were blocked and prepared for immunostaining using procedures described in Kirigiti et al. (2000) Primary antibodies include the mouse monoclonals to the mammalian elav-like protein HuR (1:5000 dilution) or to the heteroribonuclear protein hnRNP A1 (1:1000 dilution). The primary antibodies were then incubated for 2 to 5 h at room temperature with gentle agitation. The secondary antibodies for the mouse monoclonals were anti-mouse horseradish peroxidase (1:10,000 dilution used in both cases). Membranes were then washed four times with 1× Tris-buffered saline/Tween 20, and then developed using the SuperSignal Western pico chemiluminescence substrate (Pierce Chemicals, Madison, WI).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Short-Term Isoproterenol Agonist Treatment Induces Initial Up-Regulation Followed by Reduction of ₁-AR mRNA Levels in C6 Glioma Cells. Consistent with previous determinations (Hosoda et al., 1994; Rydelek-Fitzgerald et al., 1996), isoproterenol treatment (10 µM) of C6 glioma cells induces an initial short-term up-regulation of ₁-AR mRNA levels, followed by reduction of transcripts and return to basal levels (Fig. 1, A and B). Using RNase protection assays and the pCS[82, +273] probe, which contains sequences present in all transcribed ₁-AR mRNAs, we observed a 1.5- to 2-fold elevation of ₁-AR mRNAs at 80 to 100 min after agonist addition. This was followed immediately by a diminution in ₁-AR mRNAs, reaching a 50% reduction from its peak levels within 20 to 40 min and steady-state residual levels within 180 min after agonist addition (Fig. 1, A and B). Multiple repeats of this experiment were conducted (n = 3), and all exhibited similar trends. In our experiments, agonist is removed and replaced with fresh culture medium following a short-term (60 min) isoproterenol treatment. In both Hosoda et al. (1994) and Rydelek-Fitzgerald et al. (1996), an initial up-regulation in ₁-AR mRNA levels was observed 1 h after isoproterenol treatment; this was followed by ₁-AR mRNA down-regulation to below basal levels during 4 h of continuous isoproterenol treatment. Thus, the conditions between our experiments and reports of Hosoda et al. (1994) and Rydelek-Fitzgerald et al. (1996) are not directly comparable. We believe that the additional isoproterenol treatment may trigger down-regulatory mechanisms, perhaps by transcriptional repression or by post-transcriptional mechanisms described in this study, that may account for the increased reduction of ₁-AR mRNA levels.

View larger version (37K): [in this window] [in a new window]   Fig. 1.   1-AR transcripts obtained from isoproterenol-treated (60 min) C6 cells. Total RNA (15 µg) was hybridized to probe pCS[82, +273] (contains rat 1-AR sequence extending from 82 to +273) or the probe p4Z1 (contains rat 1-AR sequence extending from +2084 to +2901) and pIB15.158 (the 158-bp coding fragment derived from the rat cyclophilin gene). A and C, RPA of representative experiments examining 1-AR transcripts obtained from isoproterenol-treated (60 min) C6 cells, using probe pCS[82, +273] and probe p4Z1, respectively. Steady-state 1-AR mRNA levels at 240, 120, and 60 min (left to right in each panel) before agonist addition were also collected and analyzed by RPA. Samples collected before agonist treatment seem to be equivalent to the time 0 sample for the agonist-induction experiments. Note that 1-AR mRNA levels may fall below pretreatment levels upon further extended sampling. B and D, statistical evaluation of multiple repeats of these experiments (n = 3 and n = 4 for A and C, respectively). B shows an evaluation of RPAs using probe pCS[82, +273] (as in A) and D shows an evaluation of RPAs using probe p4Z1 (as in C). The 1-AR mRNA levels are normalized with corresponding cyclophilin mRNA levels at each time point, followed by a series normalization of all ratios with the time 0 ratio being assigned the numerical value of 1.0.

Actinomycin D Treatment Experiments Indicate That the Decline of ₁-AR Transcripts during the Reduction Phase after Short-Term Agonist-Induction Is Not Accelerated by Enhanced Transcript Degradation. Actinomycin D was applied at the peak of the up-regulatory phase after short-term agonist-induction (at 60 min), in an attempt to measure ₁-AR mRNA stability during the reduction phase of the biphasic curve (Fig. 2A). In control experiments, actinomycin D was applied to C6 cell cultures that were not treated with agonist. The transcript half-lives for ₁-AR mRNAs are not statistically different between the short-term agonist-treated or untreated C6 cell series (t_1/2 = 63 ± 2.3 min versus t_1/2 = 65 ± 4.3 min, respectively; p = 0.375; Table 1 and Fig. 2, B and C), indicating that the decline of ₁-AR mRNAs occurring during the reduction phase after short-term agonist treatment is not accelerated by enhanced transcript degradation. Figure 2B illustrates the semilogarithmic plot of ₁-AR mRNA levels versus time after initiation of actinomycin D treatment. We observed a rapid first-stage decay of ₁-AR mRNA levels during the initial 120 min after initiation of actinomycin D treatment, followed by a second-stage leveling of ₁-AR mRNA levels. Figure 2C illustrates a best-fit analysis of the first stage decline of ₁-AR mRNA levels, using a first-order decay equation; these equations were also used to determine ₁-AR mRNA half-lives under each experimental condition. In other short-term agonist/actinomycin D experiments using the p4Z1 RPA probe, the half-lives for ₁-AR transcripts containing either the distal or proximal polyadenylation sites were also not statistically different in C6 cells, compared with corresponding agonist-independent half-lives determined for both polyadenylated transcripts (Y. B. and C. A. M., unpublished observations)

View larger version (24K): [in this window] [in a new window]   Fig. 2.   A and D, 1-AR mRNA half-lives in C6 cells undergoing short-term (60 min; A) or chronic (24 h; D) isoproterenol treatment. Illustrations depict representative RNase protection experiments. For all experiments, after either short-term or chronic isoproterenol treatment, actinomycin D was added to cells to inhibit de novo transcription. The hybridization probes used in these experiments were pCS[82, +273] and pIB15.158. B and E, normalization analyses of 1-AR mRNA levels. Normalization analyses for short-term agonist treatment in C6 cells (representative experiment shown in A) and chronic agonist treatment in C6 cells (representative experiment shown in D) are illustrated in B and E, respectively. Normalization calculations of 1-AR mRNA levels against corresponding cyclophilin mRNA levels were conducted as described under Materials and Methods. Replicate experiments (n = 3) were conducted for all experimental series. Values were plotted on a semilogarithmic plot of normalized 1-AR mRNA levels versus time after initiation of actinomycin D treatment. C and F, first-order decay of 1-AR mRNAs. First-order decay analyses for short-term agonist treatment in C6 cells and chronic agonist treatment in C6 cells are illustrated in C and F, respectively. These panels illustrate statistical analyses (two-tailed unpaired t tests) of the first stage decline of 1-AR mRNA levels described in B and E, by fitting data into a first-order decay equation. Data time points to develop first-order decay equations for the short-term agonist treatment experiment in C6 cells and chronic agonist treatment experiment in C6 cells were taken up to 100 and 90 min, respectively. First-order decay equations were derived and then used to determine 1-AR mRNA half-lives under each agonist treatment condition.

View this table: [in this window] [in a new window]   TABLE 1 1-AR mRNA half-life determinations Analyses were based on the two-tailed unpaired t test. Generalized equation to determine half-life: y = (intercept point on y-axis) × exp (mx) where y = percentage of 1-AR mRNA remaining; m = slope of line; and x = time (in min). Statistical power of significance between A and B, p = 0.375; between C and D, p = 0.013; between E and F, p = 0.001; between G and H, p = 0.028; between I and J, p = 0.23; between K and L, p = 0.049; between E and G, p = 0.35; between C and I, p = 0.009. All half-life determinations are expressed in minutes; "iso" refers to isoproterenol. C6 cells were either untreated (A and C) or treated for 60 min (B) or for 24 h (D) with isoproterenol. DDT1MF2 and L6 cells were stably transfected with the 1-AR expression recombinant; transfectant cells were either untreated (E and G) or treated for 24 h with isoproterenol (F and H). C6 cells were stably transfected with HuR expression recombinants containing inserts in either sense (I and J) or reverse orientation (K and L); transfectant cells were either untreated (I and K) or treated for 24 h with isoproterenol (J and L). All cultures were treated with actinomycin D to inhibit de novo transcription before RNA extraction and RPA analyses.

Chronic Isoproterenol Treatment Results in Accelerated Reduction of ₁-AR Transcripts in C6 Cells. To address the possibility that chronic agonist treatment was required for induction of potential RNA degradation factors and for the resulting decrease in ₁-AR transcript half-life, we chronically treated C6 cells with isoproterenol (10 µM) for 24 h, before actinomycin D addtion and RPA analyses (Fig. 2D). In these experiments, ₁-AR transcripts retain half-lives (t_1/2) of 41 ± 5.9 min, statistically less (34.9% reduction) than the mRNA half-lives determined for either the untreated C6 cell series (t_1/2 = 63 ± 1.7 min) or during the reduction phase for the short-term isoproterenol treatment (t_1/2 = 63 ± 2.3 min; p = 0.013; Table 1 and Fig. 2, E and F). These results are consistent with the hypothesis that chronic agonist treatment of C6 cells results in the induction or increased induction of RNA degradation factors that can directly affect ₁-AR mRNA destabilization.

Rat ₁-AR Transcripts Expressed in Native C6 Cells or Hamster DDT₁MF2 and Rat L6 Cell Transfectants Retain Differential mRNA Half-Lives. DDT₁MF2 is a well-characterized hamster smooth muscle cell line that expresses only ₂- and not ₁-ARs; this cell line has been used extensively as a model to examine the agonist-mediated down-regulation of ₂-ARs transcripts and have been used as a source of extracts for protein interaction with the hamster ₂- and human ₁-AR UTRs (Port et al., 1992; Huang et al., 1993; Tholanikunnel and Malbon, 1997; Tholanikunnel et al., 1999). Danner and Lohse (1997) have previously examined the agonist-mediated down-regulation of ₂-AR mRNAs in the DDT₁MF2 and C6 cell lines. In these studies, the ₂-AR mRNA half-life in the DDT₁MF2 cell line ranged from 2 to 12 h (in suspension or monolayer cultures, respectively), that were reduced by 50% in the presence of agonist; the ₂-AR mRNA half-life in the C6 cell line was approximately 55 min and was not affected by agonist treatment. In these studies, the ₂-AR mRNA half-life was assessed after 30 min of agonist treatment, similar to the short-term isoproterenol treatment (60 min) and assessment of ₁-AR mRNA half-life conducted in our experiments above (Fig. 2A); the rat ₂-AR mRNA half-life results are similar to our rat ₁-AR mRNA half-life determinations (t_1/2 = 63 ± 2.3 min; Table 1 and Fig. 2, D and E). In addition, Hosoda et al. (1995) have determined the ₂-AR mRNA half-life in C6 cells to be approximately 60 min; in these studies, ₂-AR mRNA half-life was not influenced by short-term exposure to agonist. Danner and Lohse (1997) proposed that the agonist-mediated ₂-AR mRNA down-regulation in the DDT₁MF2 and C6 cell lines were moderated by distinct cell-specific mechanisms, potentially involving both transcriptional and post-transcriptional controls

View larger version (32K): [in this window] [in a new window]   Fig. 3.   A and D, rat 1-AR mRNA half lives in DDT1MF2 (A) and L6 (D) transfectants undergoing chronic (24 h) isoproterenol treatment. Illustration depicts representative RNase protection experiment. The hybridization probes used in this experiment series were pCS[82, +273] and the 211-bp hamster cyclophilin probe (A) or pIB15.158 (D). B and E, normalization analyses of 1-AR mRNA levels. Normalization calculations of 1-AR mRNA levels against corresponding cyclophilin mRNA levels were conducted as described under Materials and Methods. Replicate experiments (n = 3 or 4 for B and E, respectively) were conducted. Values were plotted on a semilogarithmic plot of normalized 1-AR mRNA levels versus time after initiation of actinomycin D treatment. C and F, first-order decay of 1-AR mRNAs. These panels illustrate statistical analyses (two-tailed unpaired t tests) of the first-stage decline of 1-AR mRNA levels described in B and E, respectively, by fitting data into first-order decay equations. Data time points to develop first-order decay equations for the chronic agonist treatment experiment in DDT1MF2 and L6 transfectants were taken up to 120 min. First-order decay equations were derived and then used to determine 1-AR mRNA half-lives.

Multiple Cellular Proteins Recognize the ₁-AR 3' UTR, Including Sequences Surrounding the Polyadenylation Signals. With the use of RNA mobility shift assays and RNA probes recognizing discrete subregions of the ₁-AR 3' UTR, we detected potential RNA binding protein(s) derived from extracts of C6 glioma cells (Fig. 4). Our initial strategy was to develop probes encompassing two basic subregions of the ₁-AR 3' UTR (probe T1, position 2150-2441; probe T2, 2465-2736), delineating the 3' ends of ₁-AR transcripts containing either the proximal or distal polyadenylation sites. However, probes of this size were too large for effective use in electrophoretic mobility shift assays (Kirigiti and Machida, 1999; also P. K. and C. A. M., unpublished observations). To generate more easily discernible mobility shifts, smaller probes encompassing the ₁-AR 3' UTR were developed (probe P1, position 2187-2306; probe P2, position 2307-2454; probe P3, position 2455-2590; probe P4, position 2591-2736) (see map and sequence in Fig. 4, A and B). There seems to be multiple mobility shift complexes identified for the P2, P3, and P4 subregion probes (Fig. 4C), supporting the hypothesis that multiple protein interactions may occur for each subregion. We also conducted additional RNA mobility shift experiments using the P3 subregion probe and employing increasing dosages of cellular extract. In this experiment, as the amounts of cellular extract increases, the proportion of the observed RNA-protein complexes shifts to the higher molecular mass complex (Fig. 4D), indicating that there are at least two protein interaction sites on the P3 subregion.

View larger version (37K): [in this window] [in a new window]   Fig. 4.   A, description of 1-AR cRNA probes used in RNA mobility shift experiments. B, nucleotide sequence of the 1-AR 3' UTR subregion probes P1, P2, P3, and P4. C, RNA mobility shift experiments using the subregion probes P1, P2, P3, and P4. For each subregion analyses, probe was bound to C6 cellular extract (1.0 µg) and loaded into polyacrylamide gels. Electrophoresis was conducted in 5% polyacrylamide gels in Tris-glycine EDTA at 150 V (25 mA) for 3 h. D, RNA mobility shift experiment using the subregion probe P3 with increasing amounts of cellular extract. In this experiment, varying amounts of cellular extract (0.5, 1.0, 2.0, and 4.0 µg) obtained from C6 glioma cells were used in independent binding reactions. Binding reactions were conducted using 2.5 × 104 cpm P3 probe for 15 min at room temperature.

UV-Cross-Linking and Denaturation of ₁-AR 3' UTR-Protein Complexes Identifies a Major M_r 39,000 Binding Protein and Several Minor Binding Components. C6 cellular proteins were bound to ₁-AR probe pTail 1, which contains ₁-AR sequence extending from +2084 to +2901; probe-protein complexes were then subjected to UV-cross-linking, denaturation, and SDS-PAGE. This analysis identified a major M_r 39,000 binding protein and several minor components M_r 90,000, 82,000, 72,000, 63,000 and 48,000 (Fig. 5A). The presence of RNasin in our binding/UV-cross-linking reactions increases the electrophoretic resolution of the cross-linked components (Fig. 5A) but creates an experimental artifact (Fig. 5A; compare lanes containing probe with or without RNasin, and in the absence of protein extract, but subjected to UV-cross-linking). We conjecture that undetermined amounts of RNasin and the RNA probe have become UV-cross-linked and that subsequent RNase digestion of cross-linked components results in the radioactive tagging of small amounts of RNasin molecule. Hence, we are unable to rule out the possibility that the M_r 63,000 protein band may represent RNasin protein artifact. When the ₁-AR probe pTail 1 was bound to cellular extracts from the C6, DDT₁MF2, and L6 cell lines, binding protein patterns were very similar (Fig. 5B). Notable differences in binding pattern included variations in the intensities of the M_r 39,000 band, highest in the C6 cell line and faint in the DDT₁MF2 and L6 cell lines, and the absence of the M_r 48,000 band and the appearance of a new M_r 56,000 band in the DDT₁MF2 cell line (Fig. 5, A and B).

View larger version (41K): [in this window] [in a new window]

UV-Cross-Linking/Immunoblot Analyses Indicate That the Major M_r 39,000 Molecule Is the Mammalian Elav-Like Protein HuR. When cross-linked proteins were electro-transferred onto nitrocellulose membranes and the blot was immunostained with HuR antibodies, we observed coincident immunostaining with the M_r 39,000 protein (Fig. 5C). This molecular mass determination is in good agreement with the calculated M_r 36,069 for HuR.

UV-Cross-Linking/Immunoblot Analyses Indicate That the Rat ₁-AR 3' UTR Probe Also Weakly Recognizes hnRNP A1. Parallel UV-cross-linking and immunoblot experiments were also conducted with heteroribonuclear protein A1 antibodies (Fig. 5D); immunoblot experiments demonstrate two immunoreactive bands positioned immediately above and below the M_r 39,000 protein, consistent in size with the 43- and 36- kDa hnRNP A1 proteins characterized in human lymphocytes (Hamilton et al., 1993). Careful laser scanning and alignment of the radioactive bands in this autoradiogram with the immunostained bands in the immunoblot revealed that neither of these two hnRNP A1 bands were coincident with the electrophoretic position of the M_r 39,000 protein (Fig. 5C; see black lines denoting alignment positions of the hnRNP A1 proteins).

UV-cross-linking and denaturation of RNA-protein complexes for polyacrylamide gel electrophoresis and electro-transfer and staining of nitrocellulose membranes with HuR and hnRNP A1 antibodies. RNA complexes using the 1-AR probe pTail 1, which contains 1-AR sequence extending from +2084 to +2901 and C6 cellular extracts (25 µg) were prepared as described under Materials and Methods. Complexes were irradiated with UV light for 5 min on ice using a UV-cross-linking oven and treated with RNase A and T1. Samples were then denatured and subjected to SDS-PAGE. Proteins from the polyacrylamide gels were electro-transferred onto nitrocellulose membranes; blots were subsequent immunostained with primary and secondary antibodies as described under Materials and Methods. For all panels, the sizes of the relative molecular mass markers (in kilodaltons) are illustrated on the left side of each panel. In A, parallel binding/UV-cross-linking experiments were conducted in the presence or absence of RNasin. Note presence of experimental artifact in RNA probe lane (in the presence of RNasin). In B, binding/UV-cross-linking experiments, all conducted in the presence of RNasin were conducted using extracts from C6, DDT1MF2, and L6 cells. Duplicate UV-cross-linking reactions for C and D were electrophoresed in the same gel. Cross-linked proteins were transferred to a single nitrocellulose membrane and membranes were subsequently cut to conduct parallel HuR and hnRNPA1 immunostaining reactions. Immunostained blots using primary antibodies recognizing HuR (C) and hnRNPA1 (D) are displayed. Note boxes which denotes coincident localization of the immunoreactive HuR band with the Mr 39,000 cross-linked protein (C and F). Lines in D denote the electrophoretic positions of the hnRNPA1 immunoreactive proteins on the original autoradiogram displaying the cross-linked proteins. For experiment described in E, formation, UV-cross-linking and denaturation of RNA-protein complexes, SDS-PAGE, and electrotransfer of proteins are as described under Materials and Methods, with modifications concerning the addition of RNase T1 before conducting the cross-linking reactions. The blot in E was immunostained with antibodies recognizing hnRNPA1. The Mr 44,000 and 37,000 cross-linked proteins are perfectly aligned with corresponding immunoreactive hnRNPA1 molecules (see arrows in E). F and G contain binding/UV-cross-linking experiments using the P3 subregion probe (containing rat 1-AR 3' UTR sequence +2455 to +2590; F) and a nonrelevant control RNA probe (containing 1-AR upstream sequences 408 to 323; G). Fig. 5.

UV-Cross-Linking/Immunoprecipitation Analyses Verify Interaction of the ₁-AR 3' UTR Probe to HuR and hnRNP A1. To augment the UV-cross-linking/immunoblot analyses, we conducted additional experiments using direct immunoprecipitation of UV-cross-linked proteins obtained from C6 cells and antibodies recognizing HuR and hnRNP A1. The UV-cross-linking/immunoblot analyses provided an initial determination in the recognition of the ARE-binding proteins to the rat ₁-AR 3' UTR; however, this assay could not distinguish ARE-binding proteins that were not bound directly to the probe, but present as an unbound single or cross-linked multimeric protein present in the original binding reaction. Using this immunoprecipitation approach, we verified direct interaction of the ₁-AR 3' UTR probe with HuR and hnRNP A1 obtained from C6 cell extracts (Fig. 6). The immunoprecipitated HuR protein seems to be electrophoretically coincident with the M_r 39,000 protein identified by UV-cross-linking (compare lanes in Fig. 6) and with the HuR molecule identified in cellular extracts by Western blot analyses (Fig. 7). In addition, both the M_r 44,000 and 37,000 hnRNP A1 molecules were immunoprecipitated from UV-cross-linked proteins obtained from C6 cells (Fig. 6), consistent with the UV-cross-linking/immunoblot results using the hnRNP A1 antibody (Fig. 5E) and with the sizes of the two molecular forms identified by direct Western blot analyses of C6 cellular extracts (Fig. 7). Negative control experiments, which do not demonstrate presence of immunoprecipitated proteins, include the use of nonrelevant antibody, such as the monoclonal antibody recognizing the bacterial Flag antigen, or the systematic removal of individual components in the immunoprecipitation procedure (Fig. 6). UV-cross-linking/immunoprecipitation experiments using cell extracts from the DDT₁MF2 line also resulted in the recovery of the HuR protein and the two molecular forms of hnRNP A1 (X. L. and C. A. M., unpublished observations).

View larger version (26K): [in this window] [in a new window]   Fig. 6.   UV-cross-linking and immunoprecipitation of RNA-protein complexes. RNA complexes using the 1-AR probe pTail 1, which contains 1-AR sequence extending from +2084 to +2901 and C6 cellular extracts (25 µg) were prepared as described under Materials and Methods. Complexes were irradiated with UV light for 5 min on ice using a UV-cross-linking oven and treated with RNase A and T1. Specific antibody (1-2 µl of anti-HuR and anti-hnRNP A1) was added to the reaction mix and allowed to incubate for 1 h at 4°C. PBS was added to the mixture to bring the final volume to 200 µl. Protein A slurry (20-40 µl) was then added and allowed to mix overnight at 4°C. The protein A beads were then washed twice by gentle centrifugation and then resuspended in 1× loading buffer before heating at 100°C for 5 min. RNA-protein complexes released from the protein A beads were then loaded into SDS-polyacrylamide gels. Negative control experiments include the use of nonrelevant antibody, such as the monoclonal antibody recognizing the bacterial Flag antigen, or the systematic removal of individual components in the immunoprecipitation procedure. Note that all control experiments, including probe lane, have undergone UV-cross-linking and RNase digestion. Molecular mass markers (in kilodaltons) are displayed in the left margin.

View larger version (22K): [in this window] [in a new window]   Fig. 7.   Immunoblot analyses of constitutive and isoproterenol-induced HuR (A and C) and hnRNPA1 (B) expression in C6 and DDT1MF2 cell lines. C6 and DDT1MF2 cells were treated with isoproterenol for 24 h. Cellular extract (25 µg) were subjected to polyacrylamide gel electrophoresis, electrotransfer to nitrocellulose membranes, and immunoblot analyses using antibodies as described under Materials and Methods. Untreated controls (equivalent to zero hour sample) and isoproterenol-treated samples are displayed. Note high constitutive expression of the Mr 44,000 hnRNPA1 protein in C6 cells versus DDT1MF2 cells. Experiments were conducted in replicate (n = 4 for agonist-mediated induction of HuR in C6 cells, A; n = 3 for agonist-mediated induction of HuR in DDT1MF2 cells, C) and are quantified in Table 2. D, Immunoblot analyses comparing constitutive levels of HuR in C6, DDT1MF2, and L6 cell extracts. Preparation of cell extracts, SDS-PAGE and immunoblot analyses using the HuR antibody were conducted as described under Materials and Methods. Band intensities were examined using the Bio-Rad Molecular Imager FX system. The average intensities (and standard deviations) for constitutive HuR expression are: C6, 3376 ± 176; DDT1MF2, 2634 ± 61.9; and L6, 2374 ± 93.2.

Immunoblot Analyses Indicate That HuR Is Constitutively Expressed in the C6 Cell Line and Undergoes Further Induction after Isoproterenol Treatment. We systematically examined constitutive (or agonist-independent) and isoproterenol-induced levels of the RNA binding proteins HuR and hnRNP A1 in the rat C6 and hamster DDT₁MF2 cell lines. The M_r 39,000 HuR becomes elevated upon chronic isoproterenol treatment in both C6 and DDT₁MF2 cells (Fig. 7, A and C; also Table 2). These observations are consistent with the interpretation that the induction of this factor after isoproterenol treatment may lead to increased agonist-mediated ₁-AR transcript degradation. In addition, using immunoblot analyses conducted on the same blot, we have determined that the constitutive levels of HuR in C6 cells are statistically higher (1.28×) than corresponding levels in DDT₁MF2 and L6 cells (Fig. 7D).

C6 Stable Transfectants Containing HuR Expression Recombinants Exhibit Reduction in ₁-AR mRNA Half-Lives. To provide additional information concerning the role of HuR in ₁-AR mRNA degradation, we transfected expression recombinants containing the complete HuR coding sequence into the C6 cell line, and isolated stable transfectants. As a transfection control, we also conducted transfections using HuR expression recombinants containing inserts in reverse orientation. C6 cells transfected with the sense HuR expression recombinant exhibit a reduced ₁-AR mRNA half-life (t_1/2 = 52 ± 2.7 min; 13.3 to 17.5% reduction; p = 0.009) compared with either nontransfected C6 cells (t_1/2 = 63 ± 1.7 min) or C6 cells transfected with the control HuR recombinant (t_1/2 = 60 ± 2.9 min) (Fig. 8 and Table 1). Western blot analyses indicate that HuR levels are elevated approximately 10.8 to 19.9% in the sense HuR transfectants compared with nontransfected C6 cells (Y. B. and C. A. M., unpublished observations). In stable transfectants containing the control HuR recombinant, after chronic isoproterenol treatment for 24 h, the ₁-AR mRNA half-lives were significantly reduced (t_1/2 = 60 ± 2.9 min versus t_1/2 = 50 ± 4.4 min; absence and presence of chronic isoproterenol treatment, respectively; 16.7% reduction), verifying the capacity of these cells to undergo agonist-mediated ₁-AR mRNA down-regulation (Table 1). The agonist-mediated reduction in ₁-AR mRNA half-lives in the stable transfectants containing the sense HuR expression recombinant seems to be statistically equivalent with the ₁-AR mRNA half-life measured in the absence of isoproterenol (Table 1).

View larger version (27K): [in this window] [in a new window]   Fig. 8.   A, rat 1-AR mRNA half lives in C6 transfectants containing HuR expression recombinants and subjected to chronic (24 h) isoproterenol treatment. Illustration depicts representative RNase protection experiment. The hybridization probes used in this experimental series were pCS[82, +273] and pIB15.158. Eukaryotic expression recombinants containing HuR inserts in sense or reverse orientation were transfected into C6 cells. B, normalization analyses of 1-AR mRNA levels. Normalization calculations of 1-AR mRNA levels against corresponding cyclophilin mRNA levels were conducted as described under Materials and Methods. Replicate experiments (n = 4) were conducted. Values were plotted on a semilogarithmic plot of normalized 1-AR mRNA levels versus time after initiation of actinomycin D treatment. C, first-order decay of 1-AR mRNAs. This panel illustrates statistical analyses (two-tailed unpaired t tests) of the first stage decline of 1-AR mRNA levels described in B, by fitting data into first-order decay equations. Data time points to develop first-order decay equations for the chronic agonist treatment experiment in C6 transfectants were taken up to 120 min. First-order decay equations were derived and then used to determine 1-AR mRNA half-lives.

Constitutive ₁-AR mRNA Levels in HuR Stable Transfectants Are Reduced to Levels below Those Observed in Isoproterenol-Treated C6 cells. We conducted RNase protection analyses comparing basal and isoproterenol-induced levels of ₁-AR mRNAs in C6 cells, the DDT₁MF2 and L6 transfectants ectopically expressing ₁-AR mRNAs and the C6 transfectants containing HuR expression recombinants (Table 3). Analyses for all cell lines were conducted on the same gel to obtain relative and comparable values of ₁-AR mRNA levels. This experiment was conducted multiple times (n = 4), with the value obtained for basal ₁-AR mRNA levels in the C6 cell line normalized as the unit measure 1. In these experiments, cell cultures were subjected to 24 h of isoproterenol treatment and then harvested immediately. This experiment measures the composite ₁-AR mRNA level and is a reflection of transcription activation/repression and post-transcriptional degradation. As expected, after chronic agonist treatment, the levels of ₁-AR mRNAs in C6 cells retained a reduced value of 0.84 ± 0.03 compared with basal uninduced levels (ascribed a unit value of 1). Interestingly, when C6 cells are transfected with the sense HuR expression recombinant, basal ₁-AR mRNAs levels in these transfectant cells retained a value of 0.63 ± 0.02, statistically reduced from the isoproterenol-induced ₁-AR mRNA levels observed in nontransfectant C6 cells (0.84 ± 0.03). When the sense HuR transfectants were treated chronically with isoproterenol, ₁-AR mRNA levels remained at 0.60 ± 0.02, indicating that the effect of HuR expression in the uninduced condition was potentially near saturation. In stable transfectants containing the control HuR recombinant in reverse orientation, basal ₁-AR mRNA levels were 1.15 ± 0.01, statistically higher than the basal level observed in uninduced C6 cells (normalized unit value of 1). Although untested, this may indicate the potential effect of antisense RNA hybridization occurring with endogeneous HuR transcripts, resulting in the overall diminution of HuR expression. Overall, these experiments support the hypothesis that HuR plays a role in the agonist-independent and agonist-mediated down-regulation of ₁-AR mRNAs.

View this table: [in this window] [in a new window]   TABLE 3 1-AR mRNA levels in C6 cells and transfectants Stable transfectants were generated using expression recombinants containing the complete rat 1-AR cDNA, including 3' UTR and HuR in either sense or reverse orientation. Chronic isoproterenol treatment was 10 µM for 24 h. RNA was harvested immediately after isoproterenol treatment. RNA measurements are a reflection of composite transcriptional and post-transcriptional activity. 1-ARmRNA levels were determined for all cell types and treatment conditions by RNase protection analyses. Analyses for each experiment were conducted on a single gel to obtain comparative values. This experiment was repeated multiple times (n = 4). Intensity values were quantified using the system described under Materials and Methods. Values for each cell type and condition were averaged and then normalized using the intensity value identified in the uninduced C6 cell line, which was ascribed a unit measure of 1.00. Standard deviations were calculated using Microsoft Excel.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We have used an array of experimental tools, including RNase protection assays, RNA mobility shifts, immunoblotting and immunoprecipitation experiments, and UV-cross-linking/denaturation analyses, to develop a basic understanding of the molecular determinants underlying agonist-mediated degradation of ₁-AR mRNAs. We have accumulated evidence that the mammalian elav-like protein HuR and the heteroribonuclear protein hnRNP A1 interact with the rat ₁-AR 3' UTR and that the induction of HuR by isoproterenol may play a role in the agonist-mediated modulation of ₁-AR mRNA half-life.

We were originally uncertain why the rat ₁-AR mRNA half-life in C6 cells (t_1/2, 63 - 65 min; Fig. 2) was much shorter than the reported half-lives of the hamster ₂-AR mRNA in DDT₁MF2 cells (Danner and Lohse, 1997). Furthermore, after chronic isoproterenol treatment of DDT₁MF2 cells, the reported half-life of the hamster ₂-AR mRNA underwent a significant reduction, to either 50 min in the case of suspension cultures or 6.5 h in the case of monolayer cultures (Danner and Lohse, 1997). In this same study, isoproterenol had no effect on ₂-AR mRNA stability in C6 cells, although agonist treatment reduced transcript levels by approximately 50%.

In our studies, the agonist-independent half-life for the rat ₁-AR mRNA was perceptibly longer (t_1/2 = 63 ± 1.7 min) than the half-life determined after chronic agonist treatment (t_1/2 = 41 ± 5.9 min), indicating that chronic isoproterenol induction can trigger a significant decrease in ₁-AR mRNA stability in C6 cells. Our experimental results for the effects of isoproterenol treatment on ₁-AR mRNA levels in C6 cells are similar to determinations reported previously for ₂-AR mRNA in the same cell line. We hypothesized that cellular environment, dictating differences in expression of key RNA degradation factors in the C6 and DDT₁MF2 cell lines could account for the agonist-mediated observed differences in both the agonist-independent and -mediated -AR mRNA stability. We tested this hypothesis by transfecting an expression vector encoding the rat ₁-AR into hamster DDT₁MF2 cells and creating a stable transfectant pool. In these experiments, the agonist-independent half-life of the rat ₁-AR transcript (t_1/2 = 109 ± 4.0 min; Fig. 3 and Table 1) in transfectant DDT₁MF2 cells was much longer than in its endogenous environment in C6 cells (t_1/2 = 63 ± 1.7 min) and seemed to be similar to the half-life determined for the hamster ₂-AR mRNA in the DDT₁MF2 cell line. In addition, in the DDT₁MF2 cell transfectants, chronic isoproterenol treatment resulted in decreases in rat ₁-AR mRNA half-life (t_1/2 = 83 ± 5.1 min). Similar results were also obtained using the rat L6 cell transfectants that ectopically express ₁-AR mRNAs. Using the L6 transfectants, the ₁-AR mRNA half-lives, determined both in the absence and presence of chronic isoproterenol treatment, were comparable with half-lives observed in DDT₁MF2 transfectants (Fig. 3 and Table 1). These observations provide evidence that functional ₁-AR mRNAs are expressed in the DDT₁MF2 and L6 transfectant cell lines, and that the rat ₁-AR mRNAs undergo differential agonist-mediated degradation rates dependent on cell environment.

The rat ₁-AR mRNA has two functional polyadenylation signals at +2450 and +2732 (Searles et al., 1995). In Fig. 1, ₁-AR transcripts containing either the proximal (short-form transcript) or the distal (long-form transcript) polyadenylation signal seem to undergo simultaneous and equivalent induction after agonist treatment, consistent with the common transcription and promoter/enhancer apparatus affecting these two polyadenylated mRNAs.

We recognize that the reduced levels of the shorter polyadenylated ₁-AR transcript is at the extreme lower limits of measurement for the RPA procedure; with this caveat and by using repeated evaluations (n = 4), we can detect no appreciable difference in stability of the two polyadenylated ₁-AR mRNAs (Y. B. and C. A. M., unpublished observations). Other investigators have confirmed the presence of two human ₁-AR transcripts, using consensus polyadenylation signals at +2469 and +2751 (Ellis and Frielle, 1998). In these studies, using ventricular heart RNA isolated from patients with idiopathic dilated cardiomyopathy, both polyadenylated transcripts underwent substantial decreases compared with nonfailing controls, with the larger transcript undergoing significantly greater decreases (2.1- and 2.7-fold decreases for the shorter and longer transcripts, respectively). These investigators conjectured that the differences in transcript levels were caused by differences in mRNA stability mediated by unique AREs present only in the longer transcript, but did not experimentally rule out potential transcriptional regulators. We cannot eliminate the potential role of sequence differences between the rat and human ₁-AR 3' UTR and the differences in cell environment (C6 cells versus cells in the left ventricular heart) for creating these distinctly different observations.

Because of the common AU-rich characteristics of the ₁- and ₂-AR 3' UTR sequences, including the presence of several AUUUA pentamers, and from evidence that RNA binding proteins may influence the post-transcriptional regulation of ₂-AR mRNAs, we hypothesized that the agonist-induced down-regulation of ₁-AR mRNAs is mediated, at least in part, by post-transcriptional destabilization of transcripts via interaction with RNA degradation molecules. However, transcriptional repression of ₁-AR mRNAs may also occur, by interaction of the inducible cAMP early repressor (ICER) with the cyclic AMP response element (CRE site) in the ₁-AR promoter region. Our studies with collaborators have demonstrated that ICER interacts with the CRE at position 1314 to 1307 in the rat ₁-AR gene and that cotransfection experiments using an ICER expression vector in C6 cells represses transcription of the rat ₁-AR promoter-luciferase recombinant containing the CRE (Rydelek-Fitzgerald et al., 1996). We believe that the transcriptional repression results identified in Rydelek-Fitzgerald et al. (1996) is consistent with our current findings, where agonist-mediated post-transcriptional control mechanisms occur after long-term (24 h) but not short-term (60 min) agonist treatment. As shown in Fig. 1, we determined that enhanced post-transcriptional degradation of ₁-AR mRNAs does not occur after short-term isoproterenol treatment. Thus, we believe that our current findings are complementary with Rydelek-Fitzgerald et al. (1996) concerning transcriptional repression. In addition, differences in the basal levels of the polyadenylated rat or human ₁-AR mRNAs may be caused by differences in the translational efficiency of the two polyadenylated transcripts. Two subtypes of the ₂-ARs (the _2a/d-ARs and _2c-ARs) retain dramatically different translational efficiencies; this difference is caused by a 288-bp sequence in the 3' UTR of the _2c-AR mRNA (Yang et al., 1997).

UV-cross-linking and denaturation analysis of the pTail 1-C6 protein complexes identified a major M_r 39,000 binding protein and several minor components M_r 90,000, 82,000, 72,000, 63,000, and 48,000 (Fig. 5A). When the ₁-AR probe pTail 1 was bound to cellular extracts from the C6, DDT₁MF2, and L6 cells, binding protein patterns were very similar (Fig. 5A). The notable differences were variations in the intensities of the M_r 39,000 band, highest in the C6 cell line and faint in the DDT₁MF2 and L6 cell lines; the absence of the M_r 48,000 band; and the appearance of a new M_r 56,000 band in the DDT₁MF2 cell line.

Immunoblot and immunoprecipitation analyses of the UV-cross-linked proteins indicate that the predominant M_r 39,000 protein is HuR (Fig. 5C and Fig. 6). In addition, when partial RNase digestion is used for the UV-cross-linking/hnRNP A1 immunoblot analyses (Fig. 5, A and B), allowing conservation of weaker probe-protein cross-linking products to occur, we observe additional minor protein components of M_r 44,000 and 37,000; these molecules are immunoreactive to antibodies recognizing hnRNP A1. Our immunoblot analyses indicate that HuR levels in the C6 cells are statistically higher (1.28 X) than corresponding levels observed in DDT₁MF2 and L6 cells (Fig. 7D; also Table 2). Thus, differences in the intensity of the M_r 39,000 binding protein observed in the UV-cross-linking of the ₁-AR 3' UTR probe to distinct cell extracts (as shown in Fig. 5B) can be partially attributed to abundance differences in HuR. We hypothesize that these intensity differences may also be caused by variations in binding avidity of the HuR molecule, perhaps conferred by post-translational structural modifications (e.g., phosphorylation) or by the requirement of additional cofactors. Four HuR ligands, known as SET and SET, pp32, and APRIL (acidic protein rich in leucine), have been recently identified and seem to influence the ability of HuR to recognize ARE-containing mRNAs (Brennan et al., 2000); thus, the existence of other cofactors or binding molecules that can influence the avidity of HuR to recognize ₁-AR mRNAs is a distinct possibility. These experiments infer that HuR abundance and/or HuR recognition to the rat ₁-AR mRNA 3' UTR seem to correlate with differences in basal transcript stability. These observations also support the interpretation that other RNA binding molecules, including hnRNP A1, can recognize elements within the rat ₁-AR 3' UTR.

Our observations are consistent with the report by others (Blaxall et al., 2000), who find that both HuR and hnRNP A1 can undergo induction by isoproterenol and bind to distinct -AR 3' UTRs with differential affinities. In the ARB complex, HuR seems to be the predominant RNA binding component; the precise ratio of the HuR and hnRNP A1 binding components are dependent on the 3' UTR sequence of the target -AR subtype mRNA (Blaxall et al., 2000). In these studies, HuR and hnRNP A1 bind to human ₁-AR 3' UTR sequences, whereas only HuR is found to bind to hamster ₂-AR 3' UTR sequences. In our experiments, both HuR and two molecular forms of hnRNP A1 are identified directly in cross-linking/immunoblot and immunoprecipitation analyses using cellular extracts and the rat ₁-AR 3' UTR probe (Fig. 5, C and E; Fig. 6). Our evidence indicates that the two molecular forms of hnRNP A1 recognize the rat ₁-AR 3' UTR, but with perhaps weaker avidity compared with HuR. HuR seems to be induced in isoproterenol-treated C6 and DDT₁MF2 cells (Fig. 7, A and C; Table 2) and is constitutively expressed at higher levels in the C6 cell line (Fig. 7D and Table 2). In addition, the M_r 40,000 hnRNP A1 molecule is constitutively expressed at high levels in C6 cells, and undergoes agonist-mediated induction (Fig. 7B; also Table 2).

To test the hypothesis that HuR played a role in ₁-AR mRNA down-regulation, we transfected eukaryotic expression recombinants containing HuR inserts in sense or reverse orientation into C6 cells. C6 stable transfectants containing the sense HuR recombinant expressed elevated levels of HuR, and exhibited a shorter agonist-independent ₁-AR mRNA half-life and lower basal ₁-AR mRNA levels compared with the corresponding mRNA degradation rate and levels observed in nontransfected C6 cells (Fig. 8 and Tables 1 and 3). Furthermore, after chronic isoproterenol treatment, the ₁-AR mRNA half-life and levels did not seem to be statistically altered (Fig. 8; Tables 1 and 3), indicating that constitutive HuR expression in the stable transfectants were potentially saturating and that isoproterenol induction had no additional effect. The constitutive ₁-AR mRNA decay rate in the HuR sense transfectant was not maximally accelerated to the rate observed in isoproterenol-treated nontransfected C6 cells, but the ₁-AR mRNA levels in the HuR sense transfectant seem to be depressed to a much-reduced basal level. C6 stable transfectants containing the control HuR recombinant in reverse orientation exhibited an agonist-independent ₁-AR mRNA half-life that was statistically equivalent to the degradation rate observed in nontransfected C6 cells. Observations described in this manuscript are consistent with the interpretation that the rat ₁-AR 3' UTR retains high avidity to HuR, and that HuR may both play a role in the agonist-independent and agonist-mediated post-transcriptional degradation of the ₁-AR mRNAs. Thus, work conducted in this study and by Rydelek-Fitzgerald et al. (1996) implicates the existence of at least three regulatory mechanisms influencing ₁-AR mRNA levels in C6 cells: 1) an initial, acute up-regulation induced by transcriptional activation, 2) an intermediate down-regulation moderated by transcriptional repression, potentially by ICER, and 3) a chronic down-regulation mediated by post-transcriptional degradation mechanisms, potentially influenced by HuR, a component of ARB, and/or by other RNA binding proteins.