Human 5-HT1A receptor expressed in insect cells activates endogenous G(o)-like G protein(s)

THEJOURNALOF BIOLOGICAL CHEMISTRY

Vol. 269, No. 17, Issue of April 29, pp. 12954-12962, 1994 Printed in U.S.A.

Human 5-HTIAReceptor Expressed in Insect Cells Activates Endogenous Go-likeG Protein(s)* (Received forpublication, April 26, 1993, and in revised form, February 22, 1994)

Janet G. MulheronSB, Shirley J. CasaiiasSB, John M. Arthur, Maria N. GarnovskayaS, Thomas W. GettysSBlI, and JohnR. Raymond$** From the Departments of Medicine (Wephrology and IIGastroenterology) and llCell Biology, Duke University Medical Center and the Medical Service, Wephrology Section, Veterans Affairs Medical Center, Durham, North Carolina 27710

Insect cell expression systems are usedto character- analogous to mammalian G, in Sf9’ cells, which in turn leads to ize signaling components such as G protein-coupledre- an accumulation of cellular CAMP(2,6). Another group showed ceptors. As such, one must know whether endogenous G that the substance P receptor, whichcouples to G proteins from proteins coupleto non-native receptors.We examined G the Gdll class of G proteins, can couple to an undefined G protein linkages after infection of Sporodoptera fiugi- protein population in Sf9 cells (7). However, coupling of mamperdu (Sf9)cells with a baculovirus encoding the 5-HTIA malian receptors to pertussis toxin-sensitive G proteins in inreceptor. Receptor expression was confirmed by immu-sect cells has been much more confusing. The gene for a putanoblot. Someof the receptors were functional, showing tive insect homologue to mammalian G,, has been cloned from guanine nucleotide-sensitivebinding to the specific ago- a Drosophila melanogaster library (12, 13) and a potential innist ligand [SH18-hydroxy-2-(di-n-propylamino)-1,2,3,4sect G,, has been reported in Sf9 cells (2, 81, so there is a tetranaphthalene). Peak expression (=150 fmoYmgof possibility that mammalian receptorscould also couple to this membrane protein) was attained =72-96 h post-infecclass of G protein in insect host cells. Evidence for an intact tion. 5-HT-increased covalent binding of [s2P]GTPazidoanilide to a 40 kDa band, which was identified as a signaling pathway including mammalian receptor, inhibitory G protein by nucleotide blocking, M e dependence, and (pertussis toxin-sensitive) G protein, and effector is scant (14, immunoblot and immunoprecipitation studies. The 15). Oker-Bloom et al. have reported that the q C 4 adrenergic bandcomigratedwith 1) pertussis toxin substrate(s), receptor inhibits forskolin-stimulated CAMP accumulation in and 2) a band recognized by two Go, antisera and one Sf9 cells, although no attempt was made t o characterize the common to heterotrimeric G protein a-subunits,but not responsible G protein(s). Thevery presence of pertussis toxincells has been controversial. Queby sera specific for G,, or Gi,. Labeled species could be sensitive G proteins in insect precipitated with a Go, antiserum. 5-HT-increased label- henberger et al.(5)recently showed that chemoattractant Metcells did ing of the band was prevented by preincubation with Leu-Phe receptors expressed in high density in insect pertussis toxin. These studies suggest that the 5-HTIA not activate GTPaseactivity, nor could they be demonstrated to receptor coupleseffectively to native insect cell Go-like possess GTP-sensitive high affinityagonist binding sites.They proteins. were also unable to show any pertussis toxin substrates in membranes derived from insect cells using a n ADP-ribosylation assay. Those findings were proposed as evidence that the Sf9 The baculovirus expression system has proved to be a con- cell lacks a major category of G proteins (GJ. In contrast, Richardson and Hosey ( 8 ) recently demonstratedthat hm, venient and powerful means of attaining high level expression muscarinic receptors activated GTPase activity and increased of many signal transductioncomponents, including nuclear(1) GTP+ binding to membranes derived from Sf9 cells. These and cell surface receptors (2-7) and G proteins (9-11). One linkages were completely sensitive to small doses of pertussis especially important application of this method is for overextoxin (7). However, distal signaling events were not reported. pression and purification of plasma membrane receptors for Vasudevan et al. (15)showed that therat m3 muscarinic recepbiochemical and reconstitution studies (2-8). In designing such toractivatedpotassiumcurrentswhen expressed ininsect studies, one must take into account the possibility that the cells, and that this activation wasabolished by preincubation heterologously expressed receptors may couple to endogenous with extremely highdoses of pertussis toxin (2 pg/ml for up to G proteins. This is particularly important if the purified recep72 h). Finally, two groups have previously suggested in foottors are tobe used for functional reconstitution studies withG in their Sf9 cell prepanotes thatG,,-like proteins were present proteins. Two groups recently showed that @adrenergic recepnot presented in those rations (2,B), but supporting data were tors couple to and activate an endogenous insect cell G protein manuscripts. Thus, the presence within Sf9 cells of pertussis toxin-sensi* This work was supported in part by National Institutes of Health tive Guoproteins has been controversial. The current studies Grants NS30927 (to J. R. R.) and DK42486 (to T. W. G.), an American The abbreviationsused are: Sf9 cells, cells derived from SporodoptHeart Association Grant-in-Aid,and a Veterans Administration Merit Award (to J. R. R.).The costsof publication of this article were defrayed era frugzperda; [3H]8-OH-DPAT, [3H]8-hydroxy-2-(di-n-propylamino)1,2,3,4-tetranaphthalene); CAMP, adenosine 3’:5’-cyclic monophosin part by the payment of page charges. This article must therefore be hereby marked “advertisement”in accordance with 18 U.S.C. Section phate; G,, stimulatory G protein; Gi, inhibitory G protein; Go, other G protein that has been shown to inhibit adenylylcyclase and o-conotoxin1734 solely to indicate this fact. 6 Indicates that those three authors contributed in a substantially sensitive calcium channels; AA-GTP,[32P]GTP-azidoanilide;PAGE, polyacrylamide gel electrophoresis; Dm, dithiothreitol;GTP, guanosine equal manner to this work. ** To whom correspondence should be addressed: Box 3459, Duke 5”triphosphate; GTPyS,guanosine 5’-0-(34hiotriphosphate); GDP, University MedicalCenter, Durham, NC 27710. Tel.:919-286-0411(ext. guanosine 5”diphosphate; App(NH)p, adenyl-5“yl imidodiphosphate; CHO, Chinese hamster ovary. 7270 or 7345); Fax: 919-286-6879.

12954

Sf9 Cell Go-likeProtein

12955

visualized by a chemoluminescencetechnique (ECL,Amersham Corp.). G protein immunoblots were performed after membranes were solubilized on ice for 1h in 20 m~ Tris, 1 nm EDTA, 1m~ Dl", 100 l ~ NaCl, l ~ and 0.3% sodium cholate, pH 8.0. The suspension was then centrifuged at 13,000 x g for 3 min, and protein was assayed in the s u p e m a ~ n t . Soluble proteins were then resolved by SDS-PAGE (12.5% acrylamide, 0.051% N,N"diallyltartardiamide), and electrophoreticallytransferred to Immobilon-P polyvinylsulfidemembranes (Millipore Corp, Bedford, EXPERIMENTALPROCEDURES MA). Following the procedure of Mumbyet al. (26, 27)withminor Materials-Most materials and reagents were obtained from Sigma. modifications (28), polyvinylsulfide membranes were blocked for 1h at (Sf9)cells were obtained from Invitrogen. Grace's insect cell medium room temperature with gentle shaking with 5% dried skim milk in and sf-900 I1 serum free medium wereobtained from Life Technologies, buffer, pH 8.0, containing 50 mM Tris, 2 m CaCb, 80 mM NaCl, 0.02% Inc. Serotonergic ligands were obtained from Research Biochemicals sodium azide, and 0.2% Nonidet P-40 (blotto).The membrane was then Incorporated (Natick, MA). [3H]-8-OH-DPATwas purchased from Re- washed twice in the same low detergent blotto. The primary antisera search Products International (Mt. Prospect, IL). I3'P1-NAD+ (30 Ci/ were diluted in low detergent blotto and added for each type-specific antibody. The blots were incubated for 1 h at room temperature with mmol) was obtained from DuPont NEN.132P1GTP-azidoanilide(AAGTP) was synthesized de novo from O~[~~PI-GTP described as below. gentle rocking. After three washes in low detergent blotto, lz5I-1abeled goat anti-rabbit IgG (1x lo6 countdmin/ml) was incubated with the blot Unlabeled nucleotides were obtained from Boehringer Mannheim. Anfor 1 h a t room temperature. The membranes were subsequently tisera were obtained from sources named below. Cell Culture-Insect cells were grown in monolayer or insuspension washed three times with low detergent blotto and twice in Tris-buffered saline without skim milk, blotted dry, and exposed to KodakXAR film culture with serum-supplemented Grace's insect cell medium or with a defined non-serum supplemented medium (sf-900 11) at 26-28"C in with intensifjhg screens. Ligand Binding-Membranes were prepared by hypotonic lysis in unsupplemented ambient atmosphere. ice-cold lysis buffer (50m Tris, 5 m EDTA, pH 7.4) with 10 strokes of Expression of the Human 5-HT,, Receptor in SP Cells-The construct for the expression of the human 5-HT1, receptor was created by a glass-on-glass Dounce, and collected by centrifugation at 37,000 x g polymerase chain reaction. Oligonucleotides weresynthesized by Oligos for 20 min at 4 "C. Membranes were resuspended in 50 m Tris, pH 7.4, Etc. Inc. (Wilsonville, OR). Sense and antisense primers were synthe- supplemented with 13H18-OH-DPATwith or without 10 p-x 5-HT as sized accordingto the human 5-HTIAreceptor sequence (18)and encom- competingligand, and incubated for 60 min a t room temperature. Mempassed the entire receptor (bases 1-20 (sense) and 1247-1266 (anti- branes were isolated by vacuum filtration through Whatman glass fiber sense). In addition both primers contained an NheI restriction site and filters, then subjected to scintillation counting. Protein Determination-Proteins were measured by the method o f the sense primer contained an additional 27 nucleotides encoding the FLAG epitope (IBI FLAG Biosystem).The polymerase chain reactions Bradford (29) using bovine serum albumin as the standard. Photoaffinity Labeling of Insect Cell G Proteins-AA-GTP was synwere donein 100-plvolumes with 1pg of plasmid containing the human 5-HT1, receptor DNA, taq DNA polymerase buffer (Promega, Madison, thesized as previously describedby Offermans and colleagues (30,31), WI) (50 m KCl, 10 m~ Tris-HCl, pH 8.8, 2.0 rn MgCl,, 0.1% Triton except that AA-GTP was purified using thin layer chromatography, X-100) 1m~ each of &TP, dCW, dGTP, and d T P , 2.5 units of taq DNA Photoaffinity labeling of G proteins was performed using 25 pg of crude polymerase (Promegal, and 0.2 p~ sense and antisense primers. The membrane protein (unless otherwise noted) exactly as described by reactions were carried out in a thermocycler accordingto the following Offermans et al. (30), except that 10 w GDPwasincluded in the protocol. The plasmid was initially denatured by incubating at 95 "C for incubation mixture to suppress basal GTP binding, thus enhancing 1min, amplified by cycling between 95 "C for 1min and 55 "C for 2 min agonist-induced AA-GTP binding. Membranes were preincubated at for 35 cycles, and then incubated at 55 "C for 7 min to ensure complete room temperature for 10min with GDP in the presence or absence of 10 extension. The polymerase chain reaction products were ligated into the w 5-HT in 50-p1 volumes of 30 m HEPES, pH 7.5, 100mM NaC1, 100 pCRlOOO vector fromthe Invitrogen TA cloning kit, and thesequence of nm EDTA, 1m benzamidine, 50 leupeptin. MgC1, (5mfand 10 phi the amplified 5-HT,, receptor was checked by chain termination se- GDP wereincluded in themixture unless otherwise indicated. Then, O X quencing (19). The amplified segment was then subcloned into the pCi ofAA-GTP13160CVmmol) was added for an additional 10-min pBlueBae transfer vector (Invitrogen, San Diego, CA) at the NheI site. incubation. The membranes were collected by centrifugation at 13,000 Plasmid from a large scale plasmid prep was purified by CsC1-ethidium x g for 10 min and resuspended in theidentical buffer lacking GDP but supplemented with 2 m Dl". After exposure t o ultraviolet light for 3 bromide centrifugation and then used for transfection of Sf9 cells. Insect cells were cotransfected with the pBlueBac containing the min at 4 "C the reactions were stopped by addition of Laemmli buffer 5-HT, receptor construct and linear AcMNPV viral DNA using cationic followed by 10% SDS-PAGE and autoradiography. liposomes according to the Invitrogen MaxBac manual. The recombiIm~unoprecipitation-The putative G protein band was imrnunoprenant virus was purified by several rounds of plaque purification follow- cipitated after labeling witn AA-GTP. Membranes were solubilizedin 25 pl of 50 m sodium phosphate, pH 7.4,l mM Dl" supplemented with 5 ing the protocols found in the Invitrogen MaxBac manual. P1-P5 viruses were collected, and the P5 virus was titered by the end point pg/ml of soybean trypsin inhibitor and 0.125% SDS(immunoprecipitation buffer 1).The suspension was heated to 60"C for 5 min, then cooled dilution method (20). Characteristics of Antisera Used for This Study-Antisera used for to room temperature, and 100 pl of immunoprecipitation buffer 2 (50 these stuhes were as follows. Antiserum 982 was raised against the II~Msodium phosphate, pH 7.4, 1 m Dl", 1.25% Nonidet P-40, 1.25% carboxyl terminus peptide K3*1ENLKDCGLF350 ofGi,, and G,,, both sodium deoxycholate, 187.5 nm NaCl, 5 pg/ml of soybean trypsin inof which it recognizes. Antiserum 978 recognizes an internal sequence hibitor, and 0.125% SDS) was added. The mixture was incubated with of Giel t L i 5 ~ D R ~ Q P ~ for' ~which ) , it is highly specific. Antiserum prewashed Pansorbin (25 pl of10% solution in 50 m sodium phos977 was raised against the carboxyl-terminal segment of Gie3 phate, pH 7.4) for 30 min on ice,after which the Pansorbin was pelleted (K34SNNLKECGLYJ") and interacts specifically with Gia3.Antiserum by centrifugation. The antisera were added to the supernatant at 1:40 976 wasraised against the same sequence, but recognizes both Giasand or 1:lOO dilution, then were incubated overnight at 4 "C. Pansorbin (50 Go,,. Antiserum 951 was raised against the carboxyl-terminal fragment pl) was added as before, incubated for 30 min on ice, and layered onto of G,, (R3ssMHLRQYELL394), for which it is highly specific. Those sera a cushion of immunoprecipitation buffer 1 supplemented with 20% suwere raised in one of the authors' laboratories (T. W. G.); specificity crose. The Pansorbin was isolated by centrifugation at 12,000 x g for 5 was evaluated by immunoblot and immunoprecipitation assays using min, then washed once by resuspension in immunoprecipitation buffer purified or partially purified G proteins expressed in bacteria (21). 1followed bycentrifugation. Each pellet was resuspended in 50 pl of 2 Antiserum P960 was raised against the peptide sequence x sample buffer, freeze-thawed,heated to 60 "C for 5 min, cooled to room G4TSNSGKSTNKWMKMcommon to most mammalian a-subunits and temperature, and subjected to SDS-PAGE and immunoblot. recognizes at least G,,, G,,, G,, Go,,, and Gze.P960 (32) was a giR from Pertussis Toxin-facilitated ADP-Ribosylation ofInsect Cell Membrane Dr. Pat Casey (Duke University). Antiserum U46 (gift of Dr. S. Mumby, Proteins-Pertussis toxin (2 pg/15 pVassay) was preactivated by incuUniversity of Texas Southwestern Medical Center, Dallas) was raised bation a t 30 "C for 30 min in buffer consisting of 50 m~ HEPES, pH 8.0, against an internalsequence of G , ( ~ z L ~ D G I S for ~ ~which 3 5 )1 mgiml of bovine serum albumin, 20 nm Dl", and 0.1% SDS as deit is highly specific. For labeling experiments, the actiscribed by Kopf and W ~ l k a l i(33). s Western Blot-Receptor immunoblots were performed as described vated toxin (15 plt was added t o 60-pl aliquots of membranes. The final previously (22, 23) using affinity-purified 5-HTIA/TIL,antipeptide IgG concentration of the components of the assay mix were 20 nm HEPES, (23-25). The protocolwas modified so that immunoreactive bands were pH 8.0, 1 m EDTA, 4 nm Dm, 10 n m thymidine, 200 pg/ml of bovine

provide evidence that the 5-HT,, receptor, which exclusively couples to pertussis toxin-sensitive G proteins in mammalian cells (16,17), can couple to an endogenous Go-like protein in Sf9 cells. These observations should be considered for any Gv,,linked receptor that i s expressed in insect cells for the purpose o f purification and reconstitution with G proteins.

12956 WESTERN BLOT "

Stds C

24

48

72

96

C

24

48

72

96

B

3 ~

C

24

96

48

FIG.1. Expression of human 5-FlT,, receptors in insect cell membranes. Membranes were preparedby mechanical disruption of cells in lysis buffera s described under "ExperimentalProcedures." Panel A, membranes were thenrun under reducing conditions on 10% SDS-PAGE gels and subjected to either silver stainingor immunoblot with affinity-purified anti-5-HT,, receptor IgG. The lefthand panel shows the resultsof a silver staining experiment.The standards (Stds)used were (fromtop: myosin, phosphorylase b, bovine serum albumin, ovalbumin, andcarbonic anhydrase). The ovalbumin standard is prestained with yellow dye and runstypically a t -53 kDa when compared with non-dyed ovalbumin. Each lane contained 30 pg (control( C ) ,24, 48, and 72 h) except for the 96 h lane, which contained 50 pg of protein. The righthand gel was treated identically, exceptthat it wassubjected to immunoblot with antiserum5-HT,,PTIL4 (23). Inductionof a new proteinof about 46 kDa(arrow)was noted on boththe silver stain and immunoblot. Similar results were obtained in two other experiments. The core theoretical molecular massof the 5-HT,, receptor is 46 kDa (18).Corresponding agonist ligand binding experiments with r3H18-OH-DPAT(done in triplicateat least three times) demonstrated the following amount of receptor at each time: control (none), 24 h (none detected), 48 h (45* 3 fmoVmg protein), 72 h (106 t 15 fmoVmg protein), and 96 h (150 t 23 fmoVmg protein). Panel B, comparison of relative amounts of ligand binding (left axis, white bars) and densitometrically quantified immunoreactive receptor bands (right axis, gray bars). serumalbumin, 1% lubrol, 2 pg of pertussistoxin,and2pCi of [3zPlNAD+(33) Incubations were terminated with the addition of a n equal volume of 2 x Laemmli buffer, boiled for 3 min, then subjectedto SDS-PAGE and autoradiography. Data Analysis-Binding and CAMPdata were analyzedby non-linear least squares regression analysis using InPlot (Graphpad, SanDiego, CA).

current studies. We expected much higher levels of binding based on the silver stain results (Fig. LA ). I t seems likely that a substantial fraction of the membrane-bound receptors are unable to bind [3H]8-OH-DPAT, possibly due to improper or inefficient processing. We compared the relative amounts of the immunoreactive receptors by scanning densitometry with ligand binding results (Fig. l€?1. The comparison suggests thata RESULTS relatively constant fraction of the expressed immunoreactive receptors are capable of binding to [3H18-OH-DPAT.Although it Expression of human 5-HTIAReceptors in Insect CellsExpression of receptor protein was confirmed by immunoblot is likely that many(or most) of the expressed receptors are not with a specific affinity-purified antipeptide IgG raised against functional, a level of 100-150 fmol of functional receptodmg of a portion of the putative third intracellular loop of the human protein fallswell within the physiological range of expression of 5-HT,, receptor (Fig. LA 1. Immunoblottable protein was detect- the 5-HT,, receptor in brain (17) (see legend of Fig. 1for more able in the membrane fraction at 48-96 h. In fact, a silver- details). We next determined whether the human 5-HT1, recells. stained band of similar mobility was also seen at 48-96 h only ceptor coupled to G proteins in the insect host in cells infected with the construct bearing the 5-HT1, receptor. Presumptive Evidence for Coupling of 5-HT,, Receptors to G The functionality of some of the expressed receptors was con- Proteins: Guanine Nucleotide Sensitivity of Agonist Bindingfirmed by ligand binding experiments. Surprisingly, relatively The ability of guanine nucleotides to reduce the high affinity (.-100-150 fmol of binding of the specific agonist ligand [3H]8-OH-DPATto the little ligand binding was detected receptodmg of protein) compared with expression levels re- 5-HT1, receptor was determined in two different assays. As ported for several other G protein-coupled receptors ( 2 4 , shown in Fig. 2A, 10 GTP shifted the Scatchard plot of the which ranged from 20 to nearly 250-fold higher than in the binding isotherm for [3H]8-OH-DPAT from a predominantly

Sfs Cell Go-likeProtein

A

2000 4000 6wo

ED00 Im 12ooo

Bound Counts 120

-9

-8

-7

-6

-5

-4

-3

log [Nucleotidel (M)

FIG. 2. Guanine nucleotide sensitivity of ['H]&OH-DPAT binding to human 5-HT,, receptors in insect cell membranes. Membranes were prepared as described under "Experimental Procedures." Ligand bindingexperiments were performedat room temperature for 1 h, followed by termination by vacuumfiltration as previously described (23-25). Panel A shows that 10 GTP shifted the Scatchard plot ofthe binding isotherm from a high affnity (steep slope) to a low affinity (shallow slope)configuration, indicating a loss of the high affinity binding site as expected if most of the 5-HTu receptors are coupled to G proteins. The plot is representative of six of nine assays where there was a clear shift from a single high affinity to a single low affinity site. In three other experiments, the plot was shifted from a biphasic eonfiguration to a monophasic low affinity configuration. Therepresentative plot in panel B shows that therank order of potency for displacing the high affinity binding (1 I" L3H18-OH-DPAT)is as expected for the (0,4124 high affinity state of the receptor-G protein complex: GTPyS n ~ z )GTP (0,320 2 200 n ~ )(*,GDP, 2.6 2 1.7 p a ) . All experiments were performed in duplicate or triplicate at least three times.

high affinity (3.22 1.1 m) to a low afltinity site (23.1-+ 3.4 m), indicating a loss of the high affinity binding site, as expected if a measurable amount of5-HT,, receptors are coupled to G proteins. Those Kd values correlate well with those obtained from binding on brain 5-HT1, receptors (17).Total number of receptors did not change. Those results suggest that the majority of 5-HT1, receptors capable of binding [3H]8-OH-DPAT were in thehigh affinity state. However, in threeof nine assays, there was a clear shift from a biphasic to a monophasic configuration, indicating that in those cases, measurable quantities of receptors were also in the low affinity state. In most cases, however, there was a shift from a monophasic high to a to G proteins is low affinity state. Furtherevidence for coupling shown in Fig. 2 B , where guanine nucleotides are shown to displace >70% of 13H18-OH-DPAT binding. The rank order of potency for displacing the high affinity binding is also as expected forthe high affinity state of the receptor-(; protein complex (GTPyS > GTP =- GDP, 41 4 m, 320 2 200 nM, and 2.6 -c 1.7 w, respectively). These results are consistent with the ability of the 5-HT, receptor to couple to endogenous G proteins in insect cells, but reveal little about the natureof the G proteins involved. Direct Evidence for Coupling to G Proteins: 5-HTinduced Increases in AA-GTP Labeling in Sfs Cell Membranes-In order to establish more directly whether G proteins functionally coupled to the 5-HT, receptor in insect cells, we used the strategy of Offermans et al. (30, 31)to covalently label G proteins in the insect cell membranes with a photoreactive GTP analog, AA-GTP. Their approach takes advantage of the obser-

12957

vation that receptor activation can increase the GTP-binding rate of some G protein a-subunits. As shown in Fig. 3, 5-HT treatment of membranes incubated with AA-GTP caused a dose-dependent 2-3-fold increase in the labeling of a 40 kDa band, which presumably represents G proteinb) activated by the 5-HT, receptor. In 3 of 12 experiments, 5-HT also appeared to increase the labeling of a 65 kDa band, but thisobservation was not consistently reproduced. In fact, in the presence of GDP, the higher band was often not visualized. As previously observed for opioid receptors in NG108-15 membranes, GDP was required for agonist-promoted G protein labeling of the 40 kDa band to beobserved (31).Those authors ascribed this phenomenon to the observations of Florio and Sternweis (38) suggesting that receptor activated G proteins have a high affinity for GTP and its analogs, whereas non-activated G proteins have a higher affinity for GDP. Thus, this effect has been attributed to a selective suppression by GDP of basal AA-GTP binding to G proteins that are not coupled to receptors. For the 5-HTIAreceptor expressed in insect cells, the best results were . half-maximal obtained a t a GDP concentration of 10 p ~ The dose forthe stimulation of labeling as determined by scanning densitometry and curve fitting was 80 * 15 m,indicating a potential physiologic interaction between the 5-HT, receptor and the putative G protein. In order to firmly establish that the labeling of the 40 kDa band was increased through a pathway involving the 5-HT, receptor, several experiments were performed on non-infectedcells and cells infected with a baculovirus construct bearing the human &-adrenergic receptor (provided byDr.Bob Lefkowitz). There was no5-HT-induced increase in labeling of the 40 kDa band in either case (n = 5 and 3,respectively, Fig.3C). Furthermore?in cells infected with the 5-HT,, receptor-bearing baculovirus construct, the labeling couldbeblockedby preincubation with the 5-HT, receptor antagonist spiperone a t 100 p~ (Fig. 3). Characterization of the Putative 40-kDa G Protein Labeled by AA-GTP-In order to establish that the 40 kDa band is a G protein, we performed experiments to determine the sensitivity of the labeling to nucleotides and magnesium. Mammalian G proteins typically are sensitive both to guanine nucleotides and magnesium and are relatively insensitive to adenine nucleotides. Labeling of the 40 kDa band could begreatly attenuated by GTPyS, GTP, and GDP, but not by App(n)hp (n= 4 in duplicate, not shown), supporting the notion that theband is a G protein. We were also able to show a clear magnesium dependence(EC,, 5 m)of the incorporation of AA-GTP into the 40 kDa band (n = 5, not shown). Thosestudies demonstrate that thelabeling of the 40 kDa band conforms well with properties expected of mammalian G proteins. To further characterize the nature of the 40 kDa band, membranes were also subjected to immunoblot with a panel of antibodies directed against various regions of mammalian G, and G, a-subunits (Fig. 4). Those results showed no 40 kDa immunoreactivity with antisera raised or G,. against the carboxyl-terminal regions of Gial,Gia2,Girr3, GTP-binding (Fig. 4A).In contrast, antisera raised against the domain common to mammalian G protein a-subunits, andtwo distinct regions ofG, recognized a prominent 40 kDa band in insect cell membranes (Fig. 4B).Because that band mmigrates with that labeled by AA-GTP, it suggested, but did not prove, that they may be the same protein. Fig. 4C demonstrates that an antiserum against the GTP binding region of Ga (P960) is capable of immunoprecipitating the 40 kDa band, as isan anserum tiserum (976) that reacts against GoJGia8. Non-immune does not immunoprecipitate the 40 kDa band. These findings demonstrate a clear physical association between the G protein labeled by AA-GTP and the G, immunoreactive specie. They confirm that the40 kDa band, theAA-GTP labeling of which is is a G protein with antigenic enhanced by the 5-HTIA receptor,

-

12958 FIG.3. Agonist-induced AA-GTP covalent binding to proteins in insect cell membranes. Experiments were performed as described under "Experimental Procedures."As shown in this representative autoradiogram, photoincorporation into a 40 kDa band was more efficient in the presence of 5-HT.As noted under "Results," a -65 kDa band was also labeled infrequently. Autoradiograms of the gels were subjected to scanning densitometry. The area under the curve was measured for each condition. Thecontrol wasassigned a value of 1 arbitrary unit in each assay, and all other values were normalized to the control value. Bargraph at the left depicts the means 2 standard error for three assays, and the inset shows a representative autoradiogram (exposed to Kodak X-AFt film for 72 h at -80 "C). The calculated half effective dose of 5-HTwas 80 t 15 nM. In the experiments depicted in the righthandpanel, membranes were pretreated with vehicle or the antagonist spiperone (100 PM) briefly before addition of 5-HT (1 PM). Spiperone alone had no effect on basal labeling (not shown, n = 3). The representative autoradiogram was exposed for48 h at -80 "C. Panel C shows that the increased labeling induced by 5-HT was present only in cells infected with the virus bearing the 5-HT,, receptor. This autoradiogram was exposed for 12 h at room temperature.

Protein Sfs Cell Go-like

A

1: C

Not lnfected

Bz-AR

S-HTIA-R

41 kDa-

5-HT

[

+ - + - +

" " "

-

sites thatconform to domains of mammalian G,. In order to firmly establish a link between the AA-GTPIf the labeled band is truly analogous to mammalian G,, i t labeled 40 kDa band and thatlabeled by pertussis toxin in Sf9 should be a substrate for pertussis toxin-catalyzed ADP-ribo- cells, we pretreated intact cells with pertussis toxin in an atsylation. As shown in Fig. 5A, there is a prominent 40-kDa tempt to eliminate the5-HTIA receptor-augmentedcomponent substrate in Sf9 cell membranes ADP-ribosylated in the pres- of AA-GTP labeling. Unfortunately, we were not able toreduce ence, but not in the absence, of pertussis toxin. Immunoblot the amountof pertussis toxin substrate by more than 25-50% as assessed by subsequent membraneADP-ribosylation studies analysis of CHO-K1 cell membranes in our laboratories (not shown), and intwo others (39,401, have demonstrated that the (not shown). Cells were treated with up to 2 pg/ml of toxin for major pertussis toxin substrates inCHO cells are GiuZand Giu3, up to 48 h. Because of two previous reports of pertussis toxinsensitive signaling pathways inSf9 cells (8, 151, we tested six although another group has suggested that the major substrates in CHO-DG44 cells are Giu3and G, (41). Regardless, differentbatches of pertussis toxin from two distributors the Sf9 cell substrate migrates with a slightly smaller mass (Sigma and List) and two different batches of Sf9 cells (one than substrateslabeled in CHO-K1 cell membranes, a slightly from Invitrogen and one from Bob Lefkowitz) with similar relarger mass than bovine brain G, and appears to be present in sults. We performed experiments in serum-supplemented meabout half the amount inSf9 cells compared with CHO cells. dia (n = 3) and unsupplemented defined media (n = 4) with Pertussis toxin-facilitated ADP-ribosylation of the 40-kDa sub- similar results.Those same batchesof pertussis toxin were able strate from Sf9 cell membranes strongly suggests that the sub- to effectively catalyze the membrane reaction after preactivastrate is a G,-like protein, but does not prove it. Therefore, we tion (Figs. 5 and 6), and the toxin itself appeared to possess a next used antisera P960 and 976 to immunoprecipitate the functional enzymatic capability. To evaluate the unlikely pospertussis toxin-labeled band. As shown in Fig. 5B, these anti- sibility that all six batches of toxin were accidentally preactiserabut notnon-immune serum precipitated theband, vated by denaturation or exposure to D'I"l' (thus, causing sepastrongly supporting the notion that it isa Go-like protein. Be- ration of the enzymatically active A component from the B cause neither serum P960, nor serum 976 completely cleared component important for cellular uptake), we performed conthe AA-GTP-labeled band from the soluble preparations (each trol experiments on intact CHO cells. Those studies showed cleared 30-50%),we performed serial immunoprecipitations that treatmentof intact CHO cells with only 100 ng of the toxin with 976 from the remaining supernatants to determine if the for 4 hcompletely eliminated subsequentlabeling of membrane ADP-ribosylated bands could be completely cleared. As shown pertussis toxin substrates (n = 3, not shown). Our conclusion in Fig. 5C, a second immunoprecipitation cleared a further 50% was that under our experimental conditions, pertussis toxin of the labeled band, whereas a third did not further clear the was eitherpoorly taken up into Sf9 cells or was poorly activated band. The inability to completely clear the bandcould be due to by them. In orderto answer the question of the pertussistoxin sensia relative inefficiency of the serum, or to the presence of another 40-kDa G protein pool. Our data do not allow us to dis- tivity of the AA-GTP labeling, we needed to circumvent the tinguish between those possibilities. However, our resultscon- technical problem described above. We did so by treating memfirm that at least 75% of the AA-GTP-labeled 40-kDa branes derived from Sf9 cells infected with the5-HTIA receptor substrates areGo-like proteins. construct with preactivated pertussis toxin, then subjecting

Sf9 Cell Go-like Protein

12959

5 IO "_

20

67 kDa41 kDa-

c

I

67 kDa-

-

976

"

Nonimmune "

FIG.4. Immunoblots with a panel of antisera raised against specific mammalian G protein a-subunits andimmunoprecipitation of the AA-GTP-labeled band with G protein antisera. Panel A, aliquots of membranes prepared from Sf9 cells and rat adipocytes were subjected to immunoblot and visualization with '''I-goat anti-rabbit I g G a s described under "Experimental Procedures." The amountof membrane IgG fractions were isolatedfrom serum by protein A high protein analyzed under eachcondition is noted in thefigure. For the sera in this panel, performance liquid chromatography prior to immunoblot and resuspended in volume a of phosphate-buffered saline equal tothe amountof serum from which they were derived. Dilutions of the IgG fractions usedfor these studies were1:4000 for 978 and 982,1:20,000 for 951, and1:40,000 for 977. Autoradiograms wereexposed to Kodak X-ARfilm for less than 24 hat -80 "C. Panel B, for these experiments, purifiedG protein standards derived from bovine brain (fromDr. Pat Casey) were used as controls. Dilutions of crude sera used for these blots were 1:2000 for P960, 1:400 for U46, and 1:16,000 for 976. Autoradiograms wereexposed to Kodak X-AR film for 5 h a t -80 "C. Blots were repeated at least three timesfor each serum or IgG fraction. Panel C , aliquots of membranes prepared from Sf9 cells (40 pg) were subjected AA-GTP labeling, solubilization in 100-pl volumes, and immunoprecipitation as described under "Experimental Procedures." P960 was used a t a 1:40 dilution and 976 and non-immune serum at 1:lOO dilution. After preclearing with Pansorbin, the supernatants were subjected to SDS-PAGE or reacted with antisera. The supernatants represent the starting material from which the immunoprecipitates were derived. Autoradiograms are representative of three to five experiments and wereexposed to Kodak X-AR film overnight a t -80 "C.

them to AA-GTP labeling in thepresence and absence of 5-HT. In order to eliminate any effects of D'IT (with which the pertussis toxin was preactivated)on the disulfide bonds within the receptor, the preactivated pertussistoxin was dialyzed against 30 mM HEPES, pH 7.5, 100 mM NaCl, 100 mM EDTA, 1 mM benzamidine, 50 1.1~leupeptin usinga Centricon 3000 MW cutoff filter before addition to the membrane preparation. The results of those experiments(Fig. 6) showed that pretreatment with pertussistoxin eliminated the ability of the 5-HTlAreceptor to increase the AA-GTP-labeling of the 40-kDa substrate. Those findingsprovide further evidence that the5-HT,, receptor in Sf9 cells couples to a pertussis toxin-sensitive Go-like protein and increasesits labeling by AA-GTP. It does not rule out coupling to other G proteins in those cells. Final proof of the coupling of the 5-HT, receptor to the Go-likeprotein requires a clear demonstrationof physical concordance between the immunoreactive Go-like protein and the 40 kDa band, thelabeling of which is increased by the 5-HT, receptor. That proof is provided in Fig. 7, in which membranes were exposed to AA-GTP in the presence or absence of 1 J ~ M 5-HT, solubilized and then immunoprecipitated with serum 976. After immunoprecipitation, the increase in labeling by

5-HT is readily apparent. Thus, we have shown physical concordance betweena 40 kDa band that labeled is by AA-GTP and ADP-ribosylated by pertussis toxin, which has immunoreactivity in immunoblot and imunoprecipitatation assays with Go antisera, andwhich shows increased labelingby GTP whenthe 5-HTlAreceptor is activated. The increased labeling is present only when Sf9 cells are infected with 5-HTIAreceptor (but not P,-receptors) and are treatedwith 5-HT and can be blocked by the 5-HTlAreceptor antagonist spiperone. DISCUSSION

The baculovirus/insect cell system'has become a powerful tool for the expression of various signal transduction components, including G protein-coupled receptors. It has already been demonstrated that P-adrenergic receptors can stimulate CAMPaccumulation through a n insect cell G proteinanalogous to mammalian G, (2,6). Functionally important eventssuch as agonist-induced phosphorylation of hm, muscarinic receptors (8) and various processing events of P,-adrenergic receptors have also been demonstrated tooccur in insect cells (2,3,8,34). However, the nature of the Sf9 cell G proteins that couple to receptors normally coupled to mammalian Gi is unclear. The

Sf9 Cell Go-like Protein

12960

A

B CHO Cells

pg Protein {

" 2.5 25 25 " -

sf9 Cells

Immunoprecipitate

Immuno- Immunoprecipitate precipitate

I

I

"

25 250 250 ~

I

11 11 11

Supernatant

Supernatant

Supern:rtant

-

" " "

Cia (P960)

Goa (976) Non-immune

Irnrnunoprecipitate

Irnmuno- Immunoprecipitate precipitate

41 kDaO

W

C

I

I

I

I

11 11 11

Supernatant " " "

Supernatant

Supernatant

" " "

++++-" " "

IP #2

IP # I

1P #3

FIG.5 . Pertussis toxin-mediatedADP-ribosylation of a40-kDa substrate in Sf!) cells. Experiments were performedas described under "Experimental Procedures." Panel A, aliquots of membranes prepared from CHO cells and Sf9 cells were subjected to ADP ribosylaton in the absence or presence of pertussis toxin. Aliquots were loaded onto 12% SDS-PAGE gels in duplicate amounts as noted in the figure. After electrophoresis, gels were washed extensivelyin 20% methanol, 10%acetic acid, dried, andexposed to KodakX-ARfilm a t room temperature ( n = 5). The gel depicted was exposed for 2 h. Panel B , aliquots of membranes prepared from Sf9 cells (20 pg) were subjected toAA-GTP-labeling, solubilization in 100-pl volumes, and immunoprecipitation as described under "Experimental Procedures." Aliquots were preclearedwith protein or P960 a t 1:lOO dilutions overnight a t 4 "C. Another 20-1.11 aliquot of 50% (w/v) A-Sepharose, then were incubated with non-immune serum, 976 protein A-Sepharose beads in phosphate-buffered saline was added to each sample and incubatedon a rotator for 30 min at room temperature. Immune complex precipitates were separatedfrom the supernatantsby centrifugation (500 x g for 1min).The immunoprecipitates were rapidly washed thrice with ice-cold phosphate-buffered saline followed by centrifugation. The supernatants represent what remains from the starting material from which the immunoprecipitates were derived. Autoradiograms are representative of three experiments and wereexposed to Kodak X-AR film for 2 days a t -80 "C. Panel C , serial immunoprecipitations of ADP-ribosylated proteins were performed with serum 976. After IP 1, supernatant and precipitates were subjected SDS-PAGE. to The lefthand two lanes show that equal amountsof ADP-ribosylated proteins were in the supernatant and immunoprecipitate. An aliquot corresponding to the remaining supernatantfrom I P 1 was added to another aliquotof 976 (1:lOO) and subjected to immunoprecipitation (ZP2),again showing clearing of about 50% of the labeled band (middle two lanes). A third precipitation U P 3 ) from a n aliquot corresponding tothe supernatant remaining after IP 2 was subjected to further treatment with 976. I P 3 did not precipitate any further ADP-ribosylated proteins (right two lanes).

current studies demonstrate a functional coupling of human 5-HTIAreceptors to a pertussis toxin-sensitive G protein pool in Sf9 insect cells, thus demonstrating that this class of receptors can modulate G protein function in these host cells. The coupling of the 5-HTIA receptors to pertussis toxin-sensitive G proteins in insect cells will provide a unique opportunity to examine therelationship between G protein structureand functional coupling to inhibitory receptors such as the 5-HTIA receptor. It will be informative to compare the sequences of mammalian G, proteins and the endogenous insect cell G proteins thatcouple to the 5-HTIAreceptor. These studies underscore the amazingconservation of structure and function among the G proteins and their "cognate" receptors throughout the animal kingdom. This conservation was previously demonstrated both by the ability of P-adrenergic receptors to stimulate CAMPaccumulation through an insect cell G protein analogousto mammalianG, (2,6), andby the ability of Drosophila 5-HTd,, and 5-HTd,, receptors expressed in mammalianNIH-3T3 cells to inhibit murine adenylyl cyclase activation through a pertussis toxin-sensitive G protein pool (35, 36). 5-HT1, receptors have already been shown to be somewhat "promiscuous" in coupling to mammalian Guoproteins (16, 21, 37). The current studies show that human5-HTIA receptors can couple functionally to endogenous G proteins expressed in insect cells. These findings are particularly important because

AA-GTP Photolabeling

v o T$

A R C D E F G H I J

-_

" " " "

IpMS-HT(

- -

++

Pertussis ( - - - Toxin

++ ++++ - -

-

+

FIG.6. Effect of pertussis toxin-facilitatedADP-ribosylation of Sf3 cell membranes on the ability of the 5-HT,, receptor to. increase AA-GTP labeling. Membranes from Sf9 cells (50 pgkondition) infected with the baculovirus construct bearing the 5-HT,, receptor were preincubated with vehicle (lanes A-D and I ) or pertussis toxin (2 pgkondition) ( l a m s E-H and J ) for 1 h a t room temperature as described in the protocol for ADP-ribosylation, except that no ["PINAD' was includedin the incubationmix. Those membranes were then either subjected to ADP-ribosylation in the presence of ['*PINAD+and pertussis toxin (2 pgkondition)for 1 h a t room temperature (lanes Z and J ) or to photolabeling with AA-GTP in theabsence (lanes A, B , E , and F )or presence (lanes C , D,G , and H ) of 1 p~ 5-HT. Samples were then subjected to 12% SDS-PAGE and autoradiography for 24 h a t -80 "C. The entire 50-pg sample was loaded for all photolabeling conditions (lanes A - H ) ,but only 10 pg for each ADP-ribosylation(lanes Z and J ) . The autoradiogram is representativeof three experiments with nearly identical results.

Sf3 Cell Go-like Protein Serum 976 Immunoprecipitate

12961

analyses, andimmunoprecipitation of both pertussis toxin- and AA-GTP-labeled 40-kDa substrates. On the surface, these results directly conflict with thoseof Quehenberger et al. (51, who showed a lack pertussis toxin substrates andlack of coupling of 41 kDa chemoattractant Net-Leu-Phe receptor to endogenous G proteins in Sf9 cells. There is no obvious explanation for the dis5-HT crepancies between their findings and ours, although we did FIG. 7. Ability of the 6-HT,,receptor to increase AA-GTPlabel- find that inclusion of lubrol in the ADP-ribosylation assays ing is present in immunoprecipitates derived from serum 976. dramatically increased the pertussis-toxin facilitated labeling Methods were as described in Figs.3 and 4. Membranes were treated or of the 40 kDa band. It should also be noted that two groups not with 1 1.1~5-HT prior to AA-GTP labeling. This autoradiogram have previously suggested in footnotes that G,-like proteins (representative of four experiments) was exposed to Kodak X-AR film were present in theirSf9 cell preparations (2, 8), but supportfor 72 h at -80 “C. ing data was not presented in those manuscripts. Our studies point out the need for special care to eliminate the baculovirushnsect cell expressionsystem has become a the effects of endogenous G proteins when designing experipopular method of expressing G protein-coupled receptors for ments inwhich such coupling might complicate interpretation. purification and reconstitution with mammalian G proteins. The fact that G protein coupling was not demonstrated for The coupling of 5-HT,, receptors to G proteins in insectcells is Net-Leu-Phe receptors expressed at very high density in Sf9 exquisitely sensitive to guanine nucleotides. In the current cells (more than 200-fold greater than in the current experistudies, -70% of receptor bindingto 1nM [3H]-8-OH-DPATwas ments) illustrates thata careful examination is warranted for guanine nucleotide-sensitive. Moreover, 5-HT-induced photoaf- each new receptor heterologously expressed in insect cells (5). finity labeling of 40-kDa putative Go protein(s) was demon- Moreover, an effect should notbe deemed insensitive to pertusstrated in membranes derived from insect cells infected with sis toxin unless care is taken to document that all pertussis baculovirus bearing the DNA of the 5-HT1, receptor. The 40 toxin substrate hasbeen inactivated as assessed by membrane kDa band(s) was shown by a variety of techniques to possess ADP-ribosylation studies. biochemical features of G proteins, Go-like immunoreactivity, The current studiesalso underscore the high degree of conand pertussis toxin sensitivity in membranes. servation of structure andfunction among G proteinsand their It was also perplexing to note that the receptor-modified G “cognate” receptors throughout the animal kingdom, and furprotein(s) appeared to be insensitive to pertussis toxin treat- ther validate the insect cell expression system as a relevant ment of intact cells, yet were susceptible to pertussis toxin physiological model. They also present thefirst direct evidence treatment of membranes. Our experimentsseem to support the that the5-HT1, receptor cancouple to anendogenous insect cell possibility that insect cells either arepoorly able to internalize Go-like protein. or activate the pertussis toxin complex under our experimental Acknowledgments-We are very grateful to Drs. Pat Casey and Tim conditions. The Sf9 cells we used were commercially obtained from a common supplier and also from a colleague, and it is not Fields of the Section on Cell Regulation and Oncogenesis, Duke University Medical Center, for helpful discussions concerning the synthesis easy to directly compare our results with thoseof Richardson of AA-GTP, and for providing the antiserum P960 to the GTP-binding and Hosey, who previously showed that low doses of pertussis site common to mammalian G proteins. Dr. Bob Lefiowitz of the toxin abolished coupling of the hm, muscarinic cholinergic re- Howard Hughes Medical Institute at Duke University generously proceptor to G proteins in insect cells (8). We used six different vided the baculovirus construct containing the DNA of the & receptor. batches of pertussis toxin at very high doses with prolonged Addendum-Ng (Ng, G. Y. K., George, S. R., Zastamy, R. L., Caron, incubations without a discernible effect on intact Sf9 cells. M., Bouvier, M., Dennis, M., and ODowd,B. F. (1993) Biochemistry 32, Those same batches were used to treat intactCHO cells bear- 11727-11733) recently reported the presence of a 41-kDa pertussis ing the 5-HT1, receptor. Preincubation with 100 ng/ml for 4 h toxin substrate and GTP-sensitive agonist binding to 5-HT,, receptors completely eliminated pertussis toxin substrates and blocked in Sf9 cells. the ability of the 5-HT1, receptor to inhibit adenylylcyclase in REFERENCES CHO cells (not shown). Therefore, it is highly unlikely that our multiple batches of toxin were defective. Another possibility is 1. Elliston, J. F., Beekman, J. M., Tsai, S. Y., OMalley, B. W., and %ai, M . J . (1992)J. B i d . Chem. 267,5193-5198 that the5-HT1, receptor is tightlyprecoupled to theG proteins 2. Parker, E. M., Kameyama, K., Higashijima, T.,and Ross, E. M. (1991)J. Eiol. in insectcells, thus blocking access of the toxin to thecarboxyl Chem. 266,519-527 3. Reilander, H.. Boege,F.,Vasudevan, S., Maul, G., Hekman, M., Dees, C., terminus of the G proteins. Although we did not directly test Hampe, W.,Helmreich, E. J. M., and Michel, H.(1991) FEES Lett. 282, that hypothesis, we think it unlikely because of the prolonged 441444 incubations of intact cells with pertussis toxin used in our 4. Berstein. G., Blank, J. L., Jhon, D.-Y., Exton, J. H., Rhee, S. G., and Ross, E. M. (1992) Cell 70,411-418 experiments, and because such would imply a stoichiometric 5. Quehenberger, 0..Prossnitz. E., Cochrane, C. G., and Ye, R. D. (1992)J. Eiol. excess of functional receptor over G protein. On the one hand, Chem. 267, 19757-19760 6. Mouillac, B., Caron, M., Bonin, H., Dennis, M., and Bouvier, M.(1992)J. Biol. such a n excess of functional receptors over G proteins would Chem. 267,22249-22255 seem to be unlikely in the face of the abundant 40-kDa G 7. Kwatra, M.,Schwinn, D., Schreurs,J., Blank, J. L., Kim, C.M., Benovic,J. L., Krause, J. E., Caron, M. G., and Leflcowitz, R. J. (1993)J. Biol. Chem. 268, protein present in Sf9 cells and the low level of receptors ca9161-9164 pable of binding ligand in these studies. On the other hand, 8. Richardson, R. M., and Hosey, M. M. (1992) J. Biol. Chem. 267,22249-22255 because a large amountof receptors incapableof binding ligand 9. Graber, S. G., Figler, R. A,, andGarrison, J. C. (1992) J. Eiol. Chem. 267, 1271-1278 appear to be expressed in the membranesof Sf9 cells (Fig. 11, 10. Graber, S. G., Figler, R. A,, Kalman-Maltese,V., Robishaw,J., and Garrison,J. there remains the remotepossibility that agonist-independent C. (1992) J. Eiol. Chem. 267, 13123-13126 tight precoupling of those 5-HT, receptors to G proteins could 11. Labrecque, J., Caron, M., Torossian, K., Plamondon, J., and Dennis, M. (1992) FEES Lett. 304,157-162 indeed block access of pertussis toxin to the carboxyl terminus 12. Buchner, E. (1991)J. Neurogenet. 7,7153-7192 cysteine acceptor site of the Go-like proteins in Sf9 cells. 13. Provost, N. M., Somers, D. E., and Hurley, J. B.(1988) J. Biol. Chem. 267, 12071-12076 The current studiesclearly documentthe presence of Go-like 14. Oker-Bloom, C., Jansson, C., Karp, M., Lindqvist, C., Savola, J.-M., Vlak, J., proteins in Sf9 insect cells based on pertussis toxin labeling, and Akerman, K. (1993) Eiochem. Biophys. Acta 1176,269-275 nucleotide and Me-sensitive AA-GTP labeling, immunoblot 15. Vasudevan, S.. Premkumar, L., Stowe, S., Gage, P. W., Reilander, H., and

{

+

12962

Sf9 Celt Go-likeProtein

Chung, S.-H. (1992) FEBS Lett, 311,7-11 16. Fargin, A., Yamamoto, K.,Cotecchia, S., Speigel,A,, Lapetina, E., Caron, M. G., and Lekowitz, R. J. (1991) Cell. Signal. 3,547657 17. Raymond, J. R., El Mestikaw, S., and Fargin, A. (1992) in M o k d a r Biology of ReceptorswhichCoupleto G Proteins (Brann, M., ed) pp.113-121, Birkhauser Boston, Cambridge, MA 18. Kobilka, B. K, Frielle, T.,Collins, S., yang-Feng, T., Kobilka, TS.,Francke, U., Letkowitz, R. J., and Caron, M. G. (1987)Nature 329,75-79 19. Sanger, F.,Nicklen, S., andCoulson,A. R. (1977)proC. Natl. Acad. Sei. U.S. A. 74,5463- 5467 20. Summers, M. D., and Smith,G. E.(1987) Ikx.Agz Erp. Sta.BUIE. 1666, 1-56 21. Raymond, J. R., Olsen, C. L., and Gettys, T.W.(1993)Biochemistry 32,1106411073 22. Raymond, J. R., Albers, F. J., and Middleton, J. P. (1992) Nuunyn-Schmkdeberg's Arch. Pharmncol. 346, 127-137,1992 23. Raymond, J. R., Beach, R. E., Kim, J., and Tisher, C. C. (1993)Am. J. Physiol. 264, F9- F19 24. Fargin, A., Raymond, J. R., Lohse, M. J., Kobilka, B. K., Caron, M. G., and Letkowitz, R. J. (1988)Nature 336, 358460 25. Raymond, J. R., Fargin, A., Lohse, M. J., Regan, J. W.,Senogbs, S., Lefkowitz, R. J., and Caron, M.G. (1989) Mol. Pharmacol. 36, 16-21 26. Mumby, S., Pang, I-K., Gilman, A. G., and Sternweis, P. C. (1988) J. Biol. Chem. 263,202&2026 27. Mumby, S. M., and Gilman, A. G. (1991) Methods Enzymol. 196,215-233

28. Gettys, T.W., Ramkumar, V., Uhing, R. J., Seger, L., and Taylor, I. L. (1991)J. Bid. Chem. 266 15949-15955 29. Bradford, M.M.(1976)Anal. Biochem. 72,248-254 30. Offermans, S., Schultz, G., and Rosenthal, W.(1991) Methods Enzymol. 96, 286-301 31. Offermanns, S., Schultz, G., and Rosenthal, W.(1991) J. Biol. Chem.266, 3365-3368 32. Casey, P.J., Fong, H. K., Simon, W.,and Gilman, A. G.11990)J . Biol. Chem. 266,2383-2390 33. Kopf, G. S., and Woclkalis, M.J. (1991) Methods Enzyml. 196,257-266 Kobilka, T. S., and Koblika, B. K. (1992) J. Bid. Chem. 267, 34. Guan, X."., 21995-21998 35. Hen, R. (1992) %ends Pharmacol. Sei. 13, 160-165 36. Saudou, F., Boschert, U.,Amlaiky, N., Plassat, J.-L., and Hen, R. (1992)EMBO J. 11,7-13 37. Bertin, B., Freissmuth, M., Breyer, R. M., Schutz, W., Strosberg, A. D., and Marullo, S . (1992) J. Bzol. Chem. 287, 82004206 38. Florio, V A,, and Sternweis, P. C. (1989) J. Biol. Chem. 264,3909-3915 39. Gerhardt, M. A., and Neubig, R. R. (1991)Md. P h u ~ o l 40,707-711 . 40. Dell'Acqua, M. L.,Carroll, R. C., and Peralta, E. G. (1993)J. Bid. Chum. 268, 5676-5685 41. Law, S . F., Yasuda, K., Bell, G., and Reisine, T. (1993) J. B i d . Chem. 268, 10721-10727