A Dominant-Negative Strategy for Studying Roles of G Proteins in Vivo *

G proteins play a critical role in transducing a large variety of signals into intracellular responses. Increasingly, there is evidence that G proteins may play other roles as well. Dominant-negative constructs of the α subunit of G proteins would be useful in studying the roles of G proteins in a variety of processes, but the currently available dominant-negative constructs, which target Mg2+-binding sites, are rather leaky. A variety of studies have implicated the carboxyl terminus of G protein α subunits in both mediating receptor-G protein interaction and in receptor selectivity. Thus we have made minigene plasmid constructs that encode oligonucleotide sequences corresponding to the carboxyl-terminal undecapeptide of Gαi, Gαq, or Gαs. To determine whether overexpression of the carboxyl-terminal peptide would block cellular responses, we used as a test system the activation of the M2 muscarinic receptor activated K+ channels in HEK 293 cells. The minigenes were transiently transfected along with G protein-regulated inwardly rectifying K+ channels (GIRK) into HEK 293 cells that stably express the M2 muscarinic receptor. The presence of the Gαi carboxyl-terminal peptide results in specific inhibition of GIRK activity in response to agonist stimulation of the M2 muscarinic receptor. The Gαi minigene construct completely blocks agonist-mediated M2 mAChR K+ channel response whereas the control minigene constructs (empty vector, pcDNA3.1, and the Gα carboxyl peptide in random order, pcDNA-GαiR) had no effect on agonist-mediated M2 muscarinic receptor GIRK response. The inhibitory effects of the Gαi minigene construct were specific because overexpression of peptides corresponding to the carboxyl terminus of Gαq or Gαs had no effect on M2 muscarinic receptor stimulation of the K+channel.

G proteins play a critical role in transducing a large variety of signals into intracellular responses. Increasingly, there is evidence that G proteins may play other roles as well. Dominant-negative constructs of the ␣ subunit of G proteins would be useful in studying the roles of G proteins in a variety of processes, but the currently available dominant-negative constructs, which target Mg 2؉ -binding sites, are rather leaky. A variety of studies have implicated the carboxyl terminus of G protein ␣ subunits in both mediating receptor-G protein interaction and in receptor selectivity. Thus we have made minigene plasmid constructs that encode oligonucleotide sequences corresponding to the carboxyl-terminal undecapeptide of G␣ i , G␣ q , or G␣ s . To determine whether overexpression of the carboxyl-terminal peptide would block cellular responses, we used as a test system the activation of the M 2 muscarinic receptor activated K ؉ channels in HEK 293 cells. The minigenes were transiently transfected along with G protein-regulated inwardly rectifying K ؉ channels (GIRK) into HEK 293 cells that stably express the M 2 muscarinic receptor. The presence of the G␣ i carboxyl-terminal peptide results in specific inhibition of GIRK activity in response to agonist stimulation of the M 2 muscarinic receptor. The G␣ i minigene construct completely blocks agonistmediated M 2 mAChR K ؉ channel response whereas the control minigene constructs (empty vector, pcDNA3.1, and the G␣ carboxyl peptide in random order, pcDNA-G␣ i R) had no effect on agonist-mediated M 2 muscarinic receptor GIRK response. The inhibitory effects of the G␣ i minigene construct were specific because overexpression of peptides corresponding to the carboxyl terminus of G␣ q or G␣ s had no effect on M 2 muscarinic receptor stimulation of the K ؉ channel.
Many biologically active molecules transduce their signals through heptahelical receptors coupled to heterotrimeric guanine nucleotide-binding proteins (G proteins). 1 G proteins play important roles in determining the specificity and temporal characteristics of a variety of cellular responses. Upon activation, G protein-coupled receptors (GPCRs) interact with their cognate heterotrimeric G protein, inducing GDP release with subsequent GTP binding to the ␣ subunit. The exchange of GDP for GTP leads to dissociation of the G␤␥ dimer from the G␣ subunit, and both initiate unique intracellular signaling responses (for review, see Refs. 1 and 2). Molecular cloning has resulted in the identification of 18 distinct G␣ subunits that are commonly divided into four families based on their sequence similarity: G i , G s , G q , and G 12 . Similarly, multiple G␤ (5) and G␥ (11) subunits have been identified.
In all G proteins studied GTP is bound as a complex with Mg 2ϩ , and the GTP-and Mg 2ϩ -binding sites are tightly coupled. Dominant-negative constructs of the ␣ subunit of G proteins have been made in which mutations are made in residues that contact the magnesium ion. Although this approach was quite successful with p21 ras and other small G proteins (3,4), dominant-negative G␣ i , G␣ o , G␣ q , and G␣ 11 have been less effective (5)(6)(7)(8)(9)(10). This is probably because of the degree to which Mg 2ϩ is necessary to support GDP binding. p21 ras forms a tight and nearly irreversible GDP⅐Mg 2ϩ complex, whereas G␣ subunits bind Mg 2ϩ in the GDP⅐Mg 2ϩ complex with lower affinity than in the GTP⅐Mg 2ϩ complex (11)(12)(13)(14).
Specific determinants of receptor-G protein interaction have been under investigation for many years. It is thought that there are multiple sites of contact between the activated receptor and the G protein. Studies using ADP-ribosylation by pertussis toxin, site-directed mutagenesis, peptide-specific antibodies, and chimeric proteins indicate that the carboxyl terminus of the G␣ subunit is not only an essential region for receptor contact, but is also important for determining G protein receptor specificity (reviewed in Refs. 1 and 15). The crystal structures of various G␣ subunits show that the last 4 -7 amino acids of G␣ were not observed (16 -22) indicating that the region is conformationally flexible in the absence of other interactions.
In vitro assays, as well as microinjection studies of intact cells, indicate G␣ carboxyl-terminal peptides can competitively block G protein-coupled downstream events (23)(24)(25)(26). A carboxyl-terminal peptide from G␣ t not only binds, but will also directly stabilize photoactivated rhodopsin (27,28). Using a combinatorial peptide library Martin et al. (29) have shown that specific residues within the carboxyl terminus of G␣ t are critical for high affinity binding of the G␣ t peptide to rhodopsin. Similarly, a carboxyl-terminal peptide from G␣ s (384 -394), but not corresponding peptides from G␣ i1/2 , inhibits the ability of ␤ 2 -adrenergic receptors to activate G␣ s and adenylyl cyclase (30). In addition, a carboxyl-terminal undecapeptide from G␣ i1/2 can bind the adenosine A1 receptor, whereas the corresponding peptide from G␣ t , which differs by only 1 amino acid residue does not (31). Thus, the carboxyl terminus of G protein ␣ subunits is critical in both mediating receptor-G protein interactions and in receptor selectivity (31)(32)(33)(34)(35).
"Minigene" plasmid vectors are constructs designed to express relatively short polypeptide sequences following their transfection into mammalian cells. Minigenes have been used by investigators to look at a variety of responses related to G proteins including (i) binding of pleckstrin homology (PH) domains to G␤␥ (36), (ii) inhibiting GPCRs by expressing the carboxyl terminus of ␤2 adrenergic receptor kinase (37)(38)(39), and (iii) identifying intracellular domains of GPCRs critical for G protein coupling (40 -44). Experiments using minigenes that express the last 55 amino acids of G␣ q to target the receptor-G q interface to achieve class-specific inhibition were recently published by Akhter et al. (45). Transient transfection of COS-7 cells with ␣ 1B -adrenergic receptors or M 1 muscarinic receptors and the G␣ q carboxyl-terminal minigene (residues 305-359) inhibits agonist stimulated inositol phosphate (IP) production, whereas co-expression with the G␣ q amino terminus (residues 1-54) has no effect. Inhibition by G␣ q (305-359) was apparently specific for G q -coupled receptors because neither ␣ 2Aadrenergic receptor-mediated IP production (G i -coupled), nor dopamine D 1A receptor-mediated cAMP production (G s -coupled) were inhibited. In addition, transgenic mice made by targeting the G␣ q carboxyl-terminal minigene to the myocardium resulted in a marked inhibition of ␣ 1B -adrenergic receptor-mediated IP production and blockade of cardiac hypertrophy.
In this paper, we study the effects of several carboxyl termini G␣ peptides using a minigene approach. To test whether minigene constructs encoding the carboxyl-terminal 11 amino acid residues from G␣ subunits could effectively inhibit G proteincoupled receptor-mediated cellular responses, we chose a system in which 1) the importance of the carboxyl terminus and 2) the downstream effector system had been well established. Numerous studies (46 -49) have shown that the M 2 muscarinic receptor (mAChR) couples exclusively to the G i /G o family. The M 2 mAChR can efficiently couple to mutant G␣ q in which the last 5 amino acids of G␣ q are substituted with the corresponding residues from G␣ i or G␣ o (34), suggesting that this receptor contains domains that are specifically recognized by the carboxyl terminus of G␣ i/o subunits. The effector system that we selected was the M 2 mAChR-activated inwardly rectifying K ϩ channel (I KACh ). In cardiac cells, the I KACh channel is formed as a heterotetramer of G protein-regulated inwardly rectifying K ϩ channels (GIRK), with two GIRK1 and two GIRK4 subunits (50,51). This channel is activated upon stimulation of M 2 mAChR in a manner that is completely pertussis toxin-sensitive and is the prototype for a direct G␤␥-activated channel (52)(53)(54). Our experiments indicate that the G␣ i carboxyl terminus minigene construct can completely block M 2 mAChR-mediated K ϩ channel activation. The inhibition appears specific as constructs producing G␣ s , G␣ q , or a scrambled G␣ i carboxylterminal peptide had no effect.

Construction of G␣ Carboxyl-terminal Minigenes-
The cDNA encoding the last 11 amino acids of human G␣ subunits (G␣ i1/2 , G␣ s , G␣ q ) or the G␣ i1/2 carboxyl terminus in random order (G␣ i R) were synthesized (Great American Gene Co.) with newly engineered 5Ј-and 3Ј-ends ( Fig.  1). The 5Ј-end contained a BamHI site followed by the ribosome binding consensus sequence (5Ј-GCCGCCACC-3Ј), a methionine (ATG) for translation initiation, and a glycine (GGA) to protect the ribosome binding site during translation and the nascent peptide against proteolytic degradation. A HindIII site was synthesized at the 3Ј-end immediately following the translational stop codon (TGA).
The DNA was brought up in sterile ddH 2 O (stock concentration 100 M). Complimentary DNA was annealed in 1ϫ NEBuffer 3 (50 mM Tris-HCl, 10 mM MgCl 2 , 100 mM NaCl, 1 mM dithiothreitol; New England Biolabs) at 85°C for 10 min then allowed to cool slowly to room temperature. The annealed cDNA were ligated for 1 h at room temperature into pcDNA 3.1(Ϫ) plasmid vector (Invitrogen) previously cut with BamHI and HindIII. After digestion with each restriction enzyme, the pcDNA 3.1 plasmid vector was run on an 0.8% agarose gel, the appropriate band cut out, and the DNA purified (GeneClean II Kit, Bio101). For the ligation reaction the ratio of insert to vector was approximately 25 M to 50 ng, respectively. Following ligation, the samples were heated to 65°C for 5 min to deactivate the T4 DNA ligase.
Ligation reaction (1 l) was electroporated into 50-l competent ARI814 cells (Bio-Rad) (Escherichia coli Pulsar; 29) and there were cells immediately placed into 1 ml of SOC (Life Technologies, Inc.). After 1 h at 37°C, 100 l was spread on LB/Amp plates and incubated at 37°C for 12-16 h. To verify that insert was present, several colonies were grown overnight in LB/Amp and their plasmid DNA purifed (Qiagen SpinKit). The plasmid DNA was digested with NcoI (New England Biolabs, Inc.) for 1 h at 37°C and run on a 1.5% agarose gel. Vector alone produced 3 bands (3.4, 1.3, and 0.7 kilobases), whereas vector with insert resulted in 4 bands (3.4, 1.0, 0.7, and 0.3 kilobases). DNA with the correct pattern was sequenced (Northwestern University Biotechnology Center) to confirm the appropriate sequence. The G␣ minigene constructs used for transfection experiments (pcDNA3.1; pcDNA-G␣ i ; pcDNA-G␣ i R; pcDNA-G␣ q , and pcDNA-G␣ s ) were purified from 500-ml cultures using endotoxin-free maxi-prep kits (Qiagen).
To determine minigene RNA expression, transiently transfected cells were washed twice with phosphate-buffered saline, lysed with 350 l of RLT lysis buffer (Qiagen, Rneasy Mini Kit), homogenized using a QIAshredder column (Qiagen), and total RNA was processed according to the manufacturer's protocol. Total RNA was eluted in diethyl pyrocarbonate-treated water, quantified, and stored at Ϫ20°C. cDNA was made from total RNA using a reverse transcribed polymerase chain reaction (RT-PCR) (CLONTECH Advantage RT-for-PCR kit) according to the manufacturer's protocol. To verify the presence of insert in cells transfected with pcDNA-G␣ i or pcDNA-G␣ i R constructs, their cDNA was used as the template for PCR with forward and reverse primers that correspond to G␣ insert and vector, respectively (forward: 5Ј-AT-CCGCCGCCACCATGGGA; reverse: 5Ј-GCGAAAGGAGCGGGCGCT-A). The primers for the G␣ minigenes amplify a 434-bp fragment only if the insert carboxyl termini oligonucleotides are present; no band is observed in cells transfected with empty vector (pcDNA3.1). As controls, PCR was also performed using T7 forward with the vector reverse primer, which amplified a 486-bp fragment in all cDNA tested or G3DPH primers (CLONTECH), which amplified a 983-bp fragment in all cDNA tested.
Additionally, transiently transfected cells were trypsinized, pelleted, washed twice with phosphate-buffered saline, and stored at Ϫ80°C. Cellular extracts were prepared by homogenizing the cell pellets for 15 s (ESGE Bio-homogenizer M133/1281-0) in fractionation buffer (10 mM HEPES, pH 7.3, 11.5% sucrose, 1 mM EDTA, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride). The homogenate was centrifuged at 3,000 ϫ g for 20 min, and the supernatant centrifuged at 100,000 ϫ g for 30 min. The cytosolic fraction from the resulting supernatant was collected; and the fractions stored at Ϫ80°C until needed. For high pressure liquid chromatography analysis, 100 l of cytosolic extract was loaded onto a C4 column (Vydac) equilibrated with 0.1% trifluoroacetic acid in ddH 2 O. Elution of the peptide was performed using 0.1% triflu-oroacetic acid in acetonitrile. The amount of acetonitrile was increased from 0 to 60% over 45 min. Peaks were collected, lyophilized, and analyzed using ion mass spray analysis (University of Illinois-Urbana Champagne).
Measurement of I KACh Currents-For the measurement of inwardly rectifying K ϩ current, whole cell currents were recorded as described previously (57,58). The extracellular solution contained 120 mM NaCl, 20 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , and 10 mM Hepes-NaOH, pH 7.4. The solution for filling the patch pipettes was composed of 100 mM potassium glutamate, 40 mM KCl, 5 mM MgATP, 10 mM Hepes-KOH, pH 7.4, 5 mM NaCl, 2 mM EGTA, 1 mM MgCl 2 , and 0.01 mM GTP. All standard salts as well as acetylcholine were from Sigma.
To minimize variations caused by different transfections or culture conditions, control experiments (transfection with pcDNA-G␣ i R) were done in parallel. Membrane currents were recorded under voltage clamp, using conventional whole cell patch techniques (59). Patch pipettes were fabricated from borosilicate glass capillaries, (GF-150 -10, Warner Instrument Corp.) using a horizontal puller (P-95 Fleming & Poulsen) and were filled with the solutions listed above. The DC resistance of the filled pipettes ranged from 3 to 6 mega-ohms. Membrane currents were recorded using a patch-clamp amplifier (Axopatch 200, Axon Instruments). Signals were analog filtered using a low pass Bessel filter (1-3 kHz corner frequency). Data were digitally stored using an IBM compatible PC equipped with a hardware/software package (ISO2 by MFK, Frankfurt/Main, Germany) for voltage control, data acquisition, and data evaluation.
To measure K ϩ currents in the inward direction, the potassium equilibrium potential was set to about Ϫ50 mV and the holding potential was Ϫ90 mV as described (57,58). Agonist-induced currents were evoked by application of acetylcholine (ACh; 1 M) using a solenoidoperated superfusion device, which allowed for solution exchange within 300 ms. Linear voltage ramps (from Ϫ120 mV to ϩ 60 mV within 500 ms) were applied every 10 s. By subtracting nonagonist-dependent currents we were able to resolve the current voltage properties of the agonist-induced currents. For analysis of the data the maximal current density (peak amplitude) of ACh-induced inwardly rectifying K ϩ currents were measured at Ϫ90 mV and compared.
Data Analysis-Data are presented as mean Ϯ S.E. The statistical differences were determined using the Student's t test (GraphPad Prism; version 2.0).

RESULTS AND DISCUSSION
Dominant-negative constructs of the ␣ subunit of G proteins have been made in which mutations are made in regions that contact the magnesium ion. For the ␣ subunit of G proteins, this includes mutations of the Gly residue within the invariant sequence (G203T,G204A), as well as mutations of a Ser residue in the effector loop, switch I region (S47C) in either G␣ o or G␣ i (5,6). However, neither of these mutations has resulted in effective dominant-negatives probably because the GDP complexes of G␣ i , G␣ o , and G␣ s have low affinity for Mg 2ϩ . Thus, we looked to other regions on G protein ␣ subunits that could serve to block receptor-G protein interactions, and consequently serve as dominant-negatives. A variety of studies have implicated the carboxyl terminus of G protein ␣ subunits in mediating receptor-G protein interaction and selectivity (for review, see Refs. 14 and 15). We have shown that carboxyl termini from G protein ␣ subunits are important sites of receptor binding, and peptides corresponding to the carboxyl terminus can be used as competitive inhibitors of receptor-G protein interactions (27,29,30). This interaction is quite specific as we found that a difference in 1 amino acid can annul the ability of the G␣ i1/2 peptide to bind the A1 adenosine receptor-G protein interface (31).
To determine whether we could selectively antagonize G protein signal transduction events in vivo by expressing peptides that block the receptor-G protein interface, we generated minigene plasmid constructs that encode carboxyl-terminal peptide sequences from G␣ i1/2 , G␣ q , or G␣ s ( Table I). As a control, we also made a minigene that expressed the carboxyl terminus of G␣ i1/2 in random order (G␣ i R, Table I). The minigene insert DNA were made by synthesizing short complimentary oligonucleotides corresponding to the peptide sequences from the carboxyl terminus of each G␣ with BamHI and Hin-dIII restriction sites at the 5Ј and 3Ј ends, respectively. Complementary oligonucleotides were annealed and ligated into the mammalian expression vector pcDNA 3.1(Ϫ). The DNA was cut with NcoI, and separated on a 1.5% agarose gel to determine whether the insert was present. As shown in Fig. 1, when insert is present there is a new NcoI site resulting in a shift in the band pattern, such that the digest pattern goes from three bands (3345, 1352, and 735 bp) to four bands (3345, 1011, 735, and 380 bp).
As our minigene approach depends on competitive inhibition, a key element for success is the expression of adequate amounts of peptides to block intracellular signaling pathways. To confirm the presence of the minigene constructs in transfected cells, total RNA was isolated 48 h posttransfection, cDNA made with RT-PCR, and PCR analysis was performed using the cDNA as template with primers specific for the G␣ carboxyl-terminal peptide insert. Separation of the PCR products on 1.5% agarose gels ( Fig. 2A) indicates the presence of the G␣ carboxyl terminus peptide minigene RNA by a single 434-bp band. Control experiments were done using a T7 forward primer with the vector reverse primer to verify the presence of the pcDNA3.1 vector, and G3DPH primers (CLON-TECH) to approximate the amount of total RNA (data not shown).
To verify that the peptide was being produced in the transfected cells, 48 h posttransfection, cells were lysed and homogenized. Cytosolic extracts were analyzed by high pressure liquid chromatography, and peaks (Fig. 2B) were analyzed by ion mass spray analysis. The mass spectrometer analysis for peak 1 from the pcDNA-G␣ i transfected cells, and peak 1 from cells transfected with a vector expressing the carboxyl terminus in random order (pcDNA-G␣ i R) indicate that a 1450 molecular weight peptide was found in both cytosolic extracts. This is the expected molecular weight for both 13 amino acid peptide sequences. The fact that they were the major peptides found in the cytosol from cells transiently transfected with the pcDNA-G␣ i or pcDNA-G␣ i R vectors strongly suggests that the vectors are producing the appropriate peptide sequences. Therefore, analysis of the transiently transfected HEK 293 cells indicates (1) minigene vectors are present, and (2) the corresponding peptides are being expressed.
We examined whether the presence of the G␣ i carboxylterminal peptide minigene would result in a significant inhibition of a downstream functional response following agonist stimulation of the transiently transfected cells. G protein-regulated inwardly rectifying K ϩ channels modulate electrical activity in many excitable cells (for review, see Refs. 60 -62). Because the channel opens as a consequence of a direct interaction with G␤␥, whole cell patch clamp recording of inwardly rectifying K ϩ currents can be used as a readout of G protein activity in single intact cells. Thus, we tested whether the G␣ carboxyl-terminal peptide minigenes could inhibit M 2 mAChR activation of inwardly rectifying K ϩ currents. Superfusion of HEK 293 cells transiently transfected with GIRK1/GIRK4 and either pcDNA-G␣ i or pcDNA-G␣ i R DNA with 1 M ACh re-TABLE I Carboxyl termini sequences Alignment of the last 11 amino acid residues from human G␣ i1/2 , G␣ q , and G␣ s subunits. Also shown is the peptide sequence of G␣ i R, the G␣ i1/2 sequence in random order, used to construct the control minigene.
vealed that cells transfected with pcDNA-G␣ i DNA have a dramatically impaired response to the M 2 mAchR agonist (Fig.  3). Fig. 3, A and B shows representative recordings of whole cell membrane currents at Ϫ90 mV. Superfusion of the cells with ACh activates inward currents in cells transfected with pcDNA-G␣ i R (Fig. 3A) but not in cells transfected with pcDNA-G␣ i (Fig. 3B). The inwardly rectifying IV-curve for the ACh-induced current from the experiment shown in Fig. 3A is illustrated in Fig. 3D. The strong inwardly rectifying properties of this current is characteristic of I KACh channels. Summarized data for the maximum amplitude of ACh-evoked currents are shown for three different transfection conditions as indicated by the black bars. The maximum current evoked by ACh was 3.7 Ϯ 1.5 pA/pF (n ϭ 14) in cells transfected with the pcDNA-G␣ i compared with 24.1 Ϯ 8.8 pA/pF (n ϭ 11) in cells transfected with pcDNA-G␣ i R. As a control we transfected cells with empty vector (pcDNA3.1). The ACh responses in these cells (16.5 Ϯ 7.7 pA/pF (n ϭ 5) was not significantly different from responses measured in cells transfected with pcDNA-G␣ i R (Fig. 3C). Basal levels for all three conditions were equivalent (pcDNA 3.2 Ϯ 1.8 pA/pF (n ϭ 5); G␣ i 6.1 Ϯ 0.9 pA/pF (n ϭ 14); G␣ i R 5.6 Ϯ 2.0 pA/pF (n ϭ 10)). To exclude experiments in which we recorded currents from cells that may not have expressed the functional channel, only those cells that exhibited a basal nonagonist-dependent Ba 2ϩ (200 M) sensitive inwardly rectifying current were used for analysis. Thus, it appears that the G␣ i minigene construct completely blocks the agonist-mediated M 2 mAChR GIRK1/4 response, whereas the control minigene constructs (empty vector, pcDNA3.1, and the G␣ i1/2 carboxyl peptide in random order, pcDNA-G␣ i R) had no effect on the agonist-mediated M 2 mAChR GIRK1/4 response.
The cardiac I KACh channel is activated upon stimulation of M 2 mAChR via G proteins of the G i family. The carboxylterminal region of G␣ has also been shown to be critical in determining the specificity of GPCR-G protein interactions (34,63). Substitution of 3-5 carboxyl-terminal amino acids from G␣ q with corresponding residues from G␣ i allowed receptors that signal exclusively through G␣ i subunits to activate the chimeric ␣ subunits and stimulate the G␣ q effector, phospholipase C-␤. To determine whether carboxyl-terminal peptides from other classes of G proteins could inhibit the agonistmediated M 2 mAChR GIRK1/4 response, we transiently transfected HEK 293 cells stably expressing the M 2 mAChR with GIRK1/GIRK4 and with minigene constructs encoding G␣ carboxyl termini for G␣ q or G␣ s . ACh-stimulated I KACh currents from cells transfected with pcDNA-G␣ q (Fig. 4B; 19.5 Ϯ 5.5 pA/pF (n ϭ 6)) or pCDNA-G␣ s (Fig. 4C; 35.5 Ϯ 9.7 pA/pF (n ϭ 5)) were not significantly different from those of cells transfected with the control minigene vector, pCDNA-G␣ i R (23.7 Ϯ 10.5 pA/pF (n ϭ 6) and 26.0 Ϯ 7.9 pA/pF (n ϭ 5), respectively). This is very different from cells transfected with pcDNA-G␣ i whose ACh-stimulated I KACh currents were significantly de-FIG. 1. The cDNA minigene constructs. Insert DNA, all G␣ carboxyl-terminal peptide minigenes contain a BamHI restriction enzyme site at the 5Ј-end followed by a ribosomal binding site sequence, a methionine for translation initiation, a glycine for stabilization of the peptide, the peptide sequence, a stop codon, and a HindIII restriction enzyme site at the 3Ј-end. The G␣ i R contains the G␣ i1/2 carboxyl peptide sequence in random order. Vector, following annealing, complimentary oligonucleotides were ligated into BamHI/HindIII cut pcDNA 3.1 plasmid vector, and the ligated insert/vector DNA was electroporated into competent cells. NcoI digest, plasmid DNA was purified, digested with NcoI, and separated on a 1.5% agarose gel to determine whether insert was present. The PCR analysis was completed using the cDNA as template with primers specific for the G␣ carboxyl-terminal peptide insert. Separation of the PCR products on 1.5% agarose gels indicates the presence of the G␣ carboxyl terminus peptide minigene RNA by a single 434-bp band. Lane 1 is a 1-kilobase pair DNA ladder; lane 2 is PCR products from cells transfected with pcDNA-G␣ i R; lane 3 is cells transfected with pcDNA-G␣ i ; lane 4 is cells transfected with pcDNA3.1. B, to verify that the peptide was being produced in the transiently transfected cells, the cells were lysed 48 h posttransfection, homogenized, and cytosolic extracts analyzed by HPLC. Peaks from cells transfected with pcDNA3.1, pcDNA-G␣ i , or pcDNA-G␣ i R were analyzed by ion mass spray analysis. phosphate accumulation. Our G␣ i minigene constructs that encode only the last 11 amino acids of the carboxyl terminus of G␣ i1/2 resulted in an 85% inhibition of the I KACh response. This difference may be caused by variations in the length of the expressed minigene (55 versus 11 residues). The longer peptide may fold in such a way that the critical carboxyl-terminal region is partly buried. We have shown that shorter peptides can effectively bind to receptors (27, 29 -31). Because the extreme carboxyl terminus of G␣ subunits in their GDP-bound conformation is disordered in crystal structures (16,20), the smaller peptide may be able to fit into its binding site more effectively. Alternatively, the difference may be in the amount of peptide being expressed because of differences in methods of transfection and cell type being studied.
Molecular determinants other than the carboxyl terminus are also involved in the recognition between heterotrimeric G proteins and their cognate receptors (27,64). However, a variety of studies have shown that the carboxyl terminus of G protein ␣ subunits is critical in both mediating receptor-G protein interaction and in receptor selectivity (31)(32)(33)(34)(35). Our results confirm that the carboxyl terminus of G␣ i is able to block agonist-mediated responses completely and thus is important in receptor selectivity and specificity. Most importantly, this method appears to be a promising approach for completely turning off G protein-mediated responses in transfected cells and in vivo. Transfection of different G␣ carboxylterminal peptide should allow us to selectively block signal transduction through any G protein and thus provides a novel dominant-negative strategy. We have now made minigene constructs encoding G␣ carboxyl-terminal undecapeptide sequences for each of the G␣ subunits. These minigenes should provide an effective dominant-negative approach that will allow us to define new roles of G proteins in vivo. The approach may also allow us to explore the coupling mechanisms of receptors that interact with multiple G proteins and tease out the downstream responses mediated by each G protein. FIG. 4. Transfection of the G␣ i carboxyl-terminal minigene inhibits the M 2 mAChR activated I KACh response, whereas G␣ s or G␣ q carboxyl-terminal minigenes do not. Stably m2 mAChRexpressing HEK 293 cells were transiently transfected with GIRK1, GIRK4 and pcDNA-G␣ i R, pcDNA-G␣ i , pcDNA-G␣ s , or pcDNA-G␣ q . A, the maximum current evoked by 1 M ACh was 3.7 Ϯ 1.5 pA/pF (n ϭ 14) in cells transfected with pcDNA-G␣ i compared with 24.1 Ϯ 8.8 pA/pF (n ϭ 11) in cells transfected with pcDNA-G␣ i R. B, the maximum current evoked by ACh was 19.5 Ϯ 5.6 pA/pF (n ϭ 6) in cells transfected with the pcDNA-G␣ q compared with 23.8 Ϯ 10.5 pA/pF (n ϭ 6) in cells transfected with pcDNA-G␣ i R. C, the maximum current evoked by ACh was 35.5 Ϯ 9.7 pA/pF (n ϭ 5) in cells transfected with pcDNA-G␣ s compared with 26.0 Ϯ 8.0 pA/pF (n ϭ 5) in cells transfected with pcDNA-G␣ i R.