Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Receptor signaling through guanine nucleotide binding proteins (G proteins) is a major mechanism of intercellular communication. Interfering with this process through agonist or antagonist action at the binding site of GPCRs is the mechanism of many therapeutic agents. Several observations, however, suggest that targeting G protein signaling mechanisms downstream of the receptor could be beneficial. First, GPCRs are able to couple to G protein in the absence of agonist (Neubig et al., 1988), and mutations (both naturally occurring and induced) result in constitutive activation of receptors (Kjelsberg et al., 1992; Samama et al., 1993). Such mutated receptors strongly activate the G protein without agonist being present and can result in human disease (Shenker et al., 1993; Parma et al., 1993). Inhibition of responses distal to the receptor would thus be an effective strategy for blocking such pathologic responses. Second, there are several processes that activate G proteins but don't involve classical GPCRs. These include the wasp venom peptide mastoparan (Higashijima et al., 1988), the IGF II receptor (Nishimoto et al., 1989) and amyloid transmembrane precursor protein (Okamoto et al., 1995). Finally, the G proteins represent both a convergence point and a divergence point in signaling. Multiple receptors can activate a G protein (Neer, 1994), and one G protein can activate multiple effectors, possibly through separate actions of  and subunits (Clapham and Neer, 1993).

The angiotensin 1a (AT1a) receptor couples primarily to phosphoinositide hydrolysis via PLC activation through a non-pertussis toxin-sensitive (G_q/11 type) G protein (Johnson and Garrison, 1987). This system is also activated by many other GPCRs, including alpha-1 adrenergic (Wu et al., 1992) and endothelin (Jouneaux et al., 1994) receptors. The AT2 receptor subtype that preferentially binds PD 123319 has been shown recently to couple to G_i-mediated ion channel regulation (Kang et al., 1994) and phosphotyrosine phosphatase inhibition (Takahasi et al., 1994; Kambayashi et al., 1993). The AT1aR also couples to adenylyl cyclase inhibition via a pertussis toxin-sensitive G protein (Pobiner et al., 1991) and to dihydropyridine-sensitive voltage-dependent Ca⁺⁺ channels via a non-G_q/11 G protein mechanism (Ohnishi et al., 1992).

The data identifying which angiotensin receptor intracellular domains determine G protein coupling and specificity have been conflicting. The second and third loops and the C-terminus of the AT1aR were implicated in activation of G_q by site-directed mutagenesis (Ohyama et al., 1992), whereas a chimera study identified largely the third intracellular loop (Wang et al., 1995). Synthetic peptide studies suggested a role for only the third loop and the C-terminal region in activation of G_i (Shirai et al., 1995), whereas a recent mutagenesis study implicated i2 as well (Shibata, 1996).

Thus we have utilized DNA-mediated delivery of receptor fragments, as pioneered by Lefkowitz and colleagues (Hawes et al., 1994; Luttrell et al., 1993), to shed more light on the AT1aR coupling mechanisms and as a first step toward the development of inhibitors of G_q activation. We show that expression of the i2 or i3 loop of the AT1aR inhibits angiotensin-dependent activation of PLC by the AT1aR, a result that supports a role for both regions in G_q activation. These data were previously presented in abstract form (Thompson et al., 1995).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. HEK-293 cells were from American type culture collection (ATCC) and COS-7 cells were a gift from Dr. Bill Pratt (University of Michigan). Dulbecco's modified essential medium (DMEM) was from Irvine Scientific, and fetal bovine serum from BIOWhittaker. Lipofectamine reagent and Opti-Mem were from BRL Life Technologies. Dowex AG1-X8 anion exchange resin was from BIO-RAD. Phenylmethylsulfonyl fluoride, aprotinin, benzamidine, bovine serum albumin (BSA, Fraction V, A6003), angiotensin II, Sar¹, Ile⁸-angiotensin II, and saralasin and other cell culture reagents were from Sigma. All other chemicals were reagent grade or better.

Construction of a mammalian expression plasmid containing DNA encoding the rat vascular AT1a receptor. DNA corresponding to nucleotides +1 to 1089 of the rat vascular AT1a receptor (Murphy et al., 1995) was generated using the PCR. A cDNA encoding the rat AT1a receptor, isolated from a rat kidney cDNA library, was used as the template. The PCR product was cloned to the HindIII/XhoI sites of pCDM8 (Invitrogen). This plasmid construct (herein designated pAT1aR) was sequenced (Sanger dideoxy chain termination) to confirm its identity to the cloned receptor.

Construction of plasmids encoding fragments of the rat AT1aR and hamster alpha-1b adrenergic receptors. Plasmid constructs containing DNA sequences encoding regions of the rat AT1a angiotensin receptor (fig. 1) were prepared as described below. In order to maximize expression of these DNA constructs, methionine (for translation initiation) and glycine codons were added upstream of the receptor-specific sequences. The initiation site was engineered within the context of a Kozak (or ribosomal binding) sequence. Downstream of the receptor-specific sequences we placed a termination codon, followed by the SV40 poly A tail and 3' UTR, both provided by the pCDM8 vector.

View larger version (41K): [in this window] [in a new window]   Fig. 1.   Seven transmembrane topology of the rat AT1aR and location of minigene constructs. The sequence of the AT1aR is shown. Amino acid residues included in the first, second and third intracellular loops and in the C-term for which DNA constructs (minigenes) were prepared are indicated in gray.

Cell culture and transfection using Lipofectamine reagent. HEK-293 and COS-7 cells were grown in DMEM supplemented with 10% fetal bovine serum, 20 mM glutamine, 100 units/ml penicillin and 0.1 mg/ml streptomycin. When cells in 100-mm dishes were at 60% to 80% confluence, they were rinsed with serum- and antibiotic-free DMEM. Transfections were carried out according to the manufacturer's instructions. Briefly, the indicated amounts of DNA were mixed with 6 or 12 µl of Lipofectamine reagent per microgram of DNA for 45 min in Opti-Mem media (Gibco-BRL). The mixture was then added to cells with gentle swirling, followed by incubation under standard culture conditions (5% CO₂, 37°C, in a humidified incubator) for 5 to 6 hr. At that time the medium was supplemented with 10 volumes of DMEM containing 11% fetal bovine serum, penicillin (100 U/ml) and streptomycin (0.1 mg/ml), and the incubation was continued for 18 hr under standard culture conditions. Cells were then subcultured for an additional 48 hr in 6- or 12-well plates for phosphoinositide (PI) assays or in a 60-mm dish for binding studies. In a limited number of experiments, cells were not subcultured but were allowed to continue growing in 100-mm dishes until harvesting.

IP production. Measurement of IP_x release was done as described (Dudley et al., 1990). Briefly, cells were incubated with 2 to 4 µCi/ml ³H-myoinositol for 24 to 30 hr in DMEM. Cells were rinsed with medium containing 0.1% BSA, and 10 mM LiCl then incubated in the same medium for 15 min at room temperature. AII and appropriate drugs were added, and cells were incubated at 37°C in a CO₂ incubator for 30 min. The medium was aspirated, and ice-cold 5% TCA was added. The TCA-soluble material was aspirated and IP_x were isolated by passing the extracts over Dowex AG1-X8 columns followed by batch elution with 2 × 2 ml of 1 M ammonium formate and 0.1 M formic acid (Dudley et al., 1990). The eluates were combined and counted by liquid scintillation spectroscopy in 16 ml of Beckman Ready Gel.

Membrane preparation. Membranes were prepared according to the method of Huang et al. (Huang et al., 1990). Briefly, 3 days after transfection, cells were rinsed twice in cold phosphate-buffered saline. Then 5 ml of hypotonic buffer (1 mM Tris, pH 7.4) with protease inhibitors (1 mM phenylmethyl sulfonyl fluoride, 10 U/ml aprotinin and 10 mM benzamidine) was added for 15 to 20 min, and cells were harvested by scraping with a rubber policeman. They were then pelleted at 30,000 × g for 30 min. Pellets were recovered, resuspended in a small volume of TME (50 mM Tris, 10 mM MgCl₂ and 1 mM EGTA, pH 7.6) and subsequently homogenized with 10 to 15 strokes of a glass/Teflon homogenizer. Membranes were then snap-frozen in liquid nitrogen and stored at 70°C for 1 to 6 weeks before use in binding assays.

Binding assays. Membranes were incubated with the angiotensin receptor antagonist [¹²⁵I]Sar¹, Ile⁸-AII in 50 mM Tris, pH 7.4, 10 mM MgCl₂, 0.1% BSA at a protein concentration of 10 to 90 µg/tube. Radioligand concentration was 0.5 nM for single-point determinations and 0.125 to 4 nM for saturation curves. Each point was determined in duplicate, and nonspecific binding was assessed with 10 µM unlabeled AII. Binding was allowed to proceed for 60 min at room temperature. Then samples were diluted with 4 ml of wash buffer (50 mM Tris, pH 7.4, 0.1% BSA) and filtered on Whatman glass-fiber filters (GC50) presoaked in wash buffer. Filters were washed twice with 4 ml of wash buffer, and bound radioactivity was assessed by gamma counting. Nonspecific binding generally accounted for 10% to 15% of total binding.

Protein assays. Protein assays were performed according to the method of Lowry et al. (1951), using BSA as standard.

Data and statistical analysis. For presentation of dose-response curves, the IP_x released are shown as means of multiple AII concentration-response curves. Each experiment was normalized to the maximum response obtained before averaging. The maximum responses obtained correlated with receptor densities (data not shown). EC₅₀ values from each control and paired minigene experiment were compared by paired t test with Instat (GraphPad Software). Significance values from the four different conditions (i1, i2, i3 and C-term) were corrected for multiple comparisons by the Bonferroni correction.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

In order to determine the optimal DNA concentrations for transfection studies and to show that the IP_x release was due to the introduced AT1a receptor, we transfected pAT1aR DNA at 0, 1, 3, 10 and 30 µg into HEK-293 cells as described in "Materials and Methods." AII-stimulated IP_x release is shown in figure 2. The amount of IP_x released increased to a maximum (4.5-fold stimulation) at 10 µg of plasmid DNA. Further increases in DNA to 30 µg resulted in a decrease to 1.8-fold stimulation. This was due to a decrease in receptor expression, cell number and cell viability at higher levels of DNA probably because of lipofecatmine toxicity. On the basis of this result, we used 3 µg of pAT1aR DNA in subsequent experiments. This amount of DNA resulted in approximately 450 fmol/mg of AT1aR expressed (fig. 2 inset; table 1). It also provided a 2- to 3-fold stimulation of IP_x release, while permitting use of 7 µg of cotransfected minigene or control vector to maintain the optimal 10 µg of total DNA to avoid toxicity from the higher amounts of lipofectamine required by the larger amount of DNA.

View larger version (54K): [in this window] [in a new window]   Fig. 2.   Dependence of IPx release on transfected AT1aR cDNA. HEK 293 cells were transfected with the indicated amounts of pAT1aR DNA as described in "Materials and Methods." After 48 hr, cells were labeled with [3H] inositol for measurement of IPx release. At 72 hr after transfection, IPx release was determined in the presence of 10 µM AII as described in "Materials and Methods." Data shown are means ± S.E.M. of three experiments. Inset shows saturation binding of [125I]Sar, Ile AII to membranes prepared from cells transfected with 3 µg of AT1aR DNA. Specific binding is indicated with nonspecific binding accounting for 10% to 15% of total at 2.5 nM.

Cotransfection of the pAT1aR DNA with the different intracellular fragment constructs or control DNA gave somewhat variable receptor densities (table 1 and below). However, there were no statistically significant differences between expression of the AT1aR with control vector and with the minigenes for loops AT1a/i1, i2 and i3². The levels of AT1aR expression with the AT1a/C-term and 1b/i3 minigenes were lower than control but comparable to one another. The K_d values for [¹²⁵I]Sar¹ Ile⁸-AII binding to membranes obtained from cells expressing combinations of the various constructs were not significantly different from each other (table 1).

There was no change in basal IP_x release upon expression of any of the receptor peptides (data not shown). Dose-response curves for AII-stimulated IP_x release are shown in figure 3 for cells cotransfected with the pAT1aR and either the AT1a/i1, AT1a/i2, AT1a/i3, or AT1a/C-term minigene constructs. The i2 and i3 constructs resulted in consistent and statistically significant increases of the angiotensin concentration needed for 50% stimulation of IP_x release (5- and 3-fold, respectively). In contrast, the i1 and c-term constructs had no significant effect on the EC₅₀ for AII.

View larger version (38K): [in this window] [in a new window]   Fig. 3.   Effect of coexpression of single intracellular loop constructs on AII-stimulated PLC activity in human 293 cells. Cells were cotransfected with 3 µg AT1a DNA and 7 µg of either control vector pCDM8 (solid line) or the indicated minigene (dashed line: AT1a/i1, AT1a/i2, AT1a/i3 or AT1a/C-term). IPx release in the presence of the indicated concentrations of AII was determined as described in "Materials and Methods." Data are expressed as the fraction of maximal AII-stimulated IPx release; points and error bars indicate means ± S.E.M. The numbers of experiments were 3 (AT1a/i1), 12 (AT1a/i2), 8 (AT1a/i3) and 4 (AT1a/C-term), and each individual dose-response curve included a paired vector control. Paired Student's t tests were performed to determine the significance of the change in EC50 for each construct. The resulting P values and fold-shifts in the dose-response curves are indicated on the graphs.

We did not see any consistent changes in the maximum PLC response between minigene and control transfections. In some experiments the minigene samples gave a greater maximum response, whereas in others the control samples gave a bigger maximum response. This was related to differences in receptor expression between the two samples; the maximum response was roughly correlated with receptor density but was not dependent on the coexpressed minigene (data not shown).

To be sure that the EC₅₀ changes that we observed were not due to alterations in the receptor density in these cells, we plotted the negative logarithm of EC₅₀ vs. receptor density for experiments with AT1a/i2. This plot shows that receptor expression was rather variable but that the EC₅₀ shifts were independent of receptor density (fig. 4). We attempted to demonstrate expression of the minigene peptides by Western blotting, but because of the quality of available antibodies against intracellular determinants of the AT1aR, we were not able to demonstrate expression of either the i3 or the C-term peptides (data not shown).

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Effect of i2 minigene to reduce AII potency is not due to decreased receptor expression. The values of log EC50 for AII-stimulated IPx release in cells transfected with a control vector (filled squares) or the i2 minigene vector (open squares) are plotted against receptor density determined by [125I]Sar, Ile AII binding. Each point represents a single experiment. The heavy lines are linear regressions of the data, and the curves are 95% confidence intervals of the linear regressions.

To determine whether simultaneous coexpression of the i2 and i3 constructs might result in synergistic inhibition of PLC responses, we cotransfected 3 µg of pAT1aR with 3.5 µg of AT1a/i2 and 3.5 µg of AT1a/i3. The combined loops caused a 7-fold right shift in the angiotensin concentration-response curve (fig. 5). This is slightly larger than that observed with 7 µg of i2 alone, but it does not suggest a strongly synergistic response because it is within the range of effects seen with the i2 and i3 constructs alone.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Additive effect of AT1a/i2 and AT1a/i3 minigenes on AII-stimulated PLC activity. HEK 293 cells were cotransfected with 3 µg of AT1a receptor cDNA and 3.5 µg of AT1a/i2 plus 3.5 µg of AT1a/i3. Controls were the same as for the previous minigene experiments. Data are expressed as the fraction of maximal AII-stimulated IPx release as indicated in figure 2. Data are mean ± S.D. from a single experiment performed in triplicate. This experiment was replicated once with similar results.

View larger version (22K): [in this window] [in a new window]   Fig. 6.   The i3 loop from the alpha-1b adrenergic receptor also inhibits AII-stimulated PLC activity. Human 293 cells were cotransfected with 3 µg of AT1aR cDNA and 7 µg of 1b/i3 or empty vector (control). Data are expressed as a fraction of maximum PI stimulation as described in figure 2 and in "Materials and Methods." Data are the means of three separate minigene and two separate control experiments ± S.D.

We also wanted to determine whether an i3 fragment from another G_q-coupled receptor could also block signaling by the AT1aR. Thus we performed similar experiments with a plasmid construct encoding the third intracellular loop of the alpha-1b adrenergic receptor (1b/i3). This 1b/i3 construct is similar to that reported by Hawes et al. (1994). As with the AT1a/i2 and i3 constructs, cotransfection of the 1b/i3 minigene increased the EC₅₀ for AII stimulation of PLC (fig. 6, P = .08).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

We report here the use of minigenes encoding fragments of the AT1aR to assess the role of different intracellular domains in PLC activation. The second (AT1a/i2) and third (AT1a/i3) intracellular loop constructs inhibit AT1aR responses, whereas i1 and C-term constructs do not. These data provide further support for a role of the i2 as well as the i3 loop of the AT1aR in coupling G_q.

The use of site-directed antibodies (Strosberg, 1985; Strader et al., 1983; Couraud et al., 1981), mutagenesis (Ostrowski et al., 1992), synthetic peptides (Dalman and Neubig, 1991; Palm et al., 1989; König et al., 1989; Cheung et al., 1991; Munch et al., 1991; Taylor et al., 1996), chimeras and, most recently, cellular expression of receptor fragments (Hawes et al., 1994; Luttrell et al., 1993) has provided strong evidence for a role of the third intracellular loop in coupling to G protein. Our results are fully consistent with this conclusion.

One major new finding of this study is that expression of the second intracellular loop of the AT1aR blocks receptor-stimulated PLC activity. This effect was manifested as a reduction in the potency of the agonist AII to stimulate [³H]IP_x release. A study by Ohyama et al. utilizing site-directed mutagenesis also implicated the second loop of the AT1aR in G protein coupling (Ohyama et al., 1992). In contrast, Wang et al. reported that the second intracellular loop of AT1aR is not important for G protein specificity when stably expressed in Chinese hamster ovary cells (Wang et al., 1995). Their experiments utilized chimeras of AT1a and AT2 receptors to induce AII-stimulated c-fos expression and Ca⁺⁺ mobilization. Thus the i2 loop may be involved in receptor-G protein coupling but may not be important as a determinant of G protein specificity. Shirai et al. tested the ability of synthetic peptides based on the AT1aR sequence to activate purified G_o and G_i and found that only the i3N and C-term peptides resulted in G protein activation (Shirai et al., 1995). Differences between that study and ours include their use of G_o and G_i rather than G_q and their assessment of G protein activation rather than blockade of receptor-G protein coupling. A similar divergence among different responses was seen in our previous study showing that a second intracellular loop peptide from the alpha-2a adrenergic receptor binds to G protein but fails to block its function as assessed by GTPase activity (Dalman and Neubig, 1991). Additional evidence for a role of loop i2 in G protein coupling came from peptide studies with rhodopsin by Konig and associates (König et al., 1989), chimeric studies of AR, muscarinic and endothelin receptors (Takagi et al., 1995; Wong et al., 1990) and site-directed mutagenesis studies of muscarinic receptors (Moro et al., 1993).

Our observation that loop i3 of AT1aR inhibits receptor-stimulated PLC activity is consistent with the findings of Luttrell et al. (1993) and Hawes et al. (1994) for the ₁AR, ₂AR, M₁-muscarinic and M₂-muscarinic receptors. In contrast to their results with the ₁AR/i2 and C-term minigenes, we observed a clear effect of the AT1a/i2 construct and no effect of the AT1a/C-term construct. Also, our results showed a reduction of potency in the AII concentration-response curves, indicative of a reversible, competitive process, whereas they reported decreases in the maximum response.

We observed that coexpression of the alpha-1b adrenergic receptor third intracellular loop with the AT1aR also decreased angiotensin II potency in activation of PLC. This cross-reactivity between the angiotensin 1a receptor and the 1b receptor fragment is consistent with an effect at the G protein level. The cross-reactivity reported by Hawes et al. (1994) was different. Specifically, their _1B/i3 minigene did not block PLC activation by the M₁ muscarinic receptor, but the M₁/i3 minigene did block _1BAR function.

Because we were concerned that variable levels of receptor expression could contribute to the decreased agonist potency, we examined EC₅₀ values as a function of receptor density. The EC₅₀ shifts produced by the i2 loop did not depend on receptor density (fig. 4). Similar results were obtained for the i3 loop (not shown). This result is consistent with an action of the fragments downstream of the receptor.

We were somewhat surprised that there was no significant increase in basal IP_x release upon expression of the AT1aR i3 loop or C-term, because they have been shown to activate G proteins in vitro (Shirai et al., 1995). It is likely that the efficacy of synthetic peptides as G protein agonists is lower than that of receptors. Previous studies with synthetic alpha-2A adrenergic receptor peptides showed stimulation of GTPase with purified G_i or G_o (Wade et al., 1996), but this was not as robust as the stimulation by mastoparan. Also, there was no receptor peptide-mediated stimulation of GTPase in platelet membranes despite good receptor-mediated stimulation (Dalman and Neubig, 1991). Specifically for the AT1aR, we have recently shown that synthetic peptides corresponding to the AT1a/i3 but not the AT1a/i2 region do activate ion channel responses in a neuronal microinjection system (Zhu et al., 1997). The effectiveness of the i3 peptide in that system may reflect the relatively high concentration used (50-100 µM) and perhaps signal amplification that permits weaker responses to be detected.

In conclusion, our data indicate that the AT1aR intracellular loops can be used to map receptor-G protein contact sites. Both the second and the third intracellular loops appear to interact with G protein, whereas the first intracellular cytoplasmic loop does not. These data confirm the utility of intracellular expression of receptor fragments and could prove useful in blocking responses due to constitutively activated receptors.