Identification of three pertussis toxin substrates (41, 40 and 39 kDa proteins) in mammalian brain. Comparison of predicted amino acid sequences from G-protein alpha-subunit genes and cDNAs with partial amino acid sequences from purified proteins

We have determined the partial amino acid sequences of the 40 kDa protein, one of the three pertussis toxin substrates in porcine brain. Purified 40 kDa protein from porcine brain was completely digested with TPCK-trypsin. Digested peptides were separated by reverse-phase HPLC and subjected to analysis by gas-phase protein sequencing. Several sequences of porcine brain 40 kDa protein completely matched with those which were deduced from the nucleotide sequences of the human Gi2 alpha gene and rat Gi2 alpha cDNA. On the other hand, the previously determined sequences of the rat brain 41 and 39 kDa proteins were in complete agreement with the predicted amino acid sequences of rat Gi1 alpha and Go alpha cDNAs, respectively.     

Antibodies directed against synthetic peptides distinguish between GTP-binding proteins in neutrophil and brain

Antisera AS/6 and 7, raised against a synthetic peptide KENLKDCGLF corresponding to the carboxyl-terminal decapeptide of transducin-alpha, react on immunoblots with purified transducin-alpha and with proteins of 40-41 kDa in all tissues tested. The latter represent one or more forms of Gi alpha but not Go alpha, since a synthetic peptide, KNNLKDCGLF, corresponding to the carboxyl-terminal decapeptide of two forms of Gi alpha blocks AS/6 and 7 reactivity with transducin-alpha and Gi alpha on immunoblots, whereas the corresponding Go-related peptide, ANNLRGCGLY, does not. Antisera LE/2 and 3, raised against the synthetic peptide LERIAQSDYI, corresponding to an internal sequence predicted by one form of Gi alpha cDNA (Gi alpha-2) and differing by 3 residues from the sequence of another form, Gi alpha-1, react strongly with a 40-kDa protein abundant in neutrophil membranes and with the major pertussis toxin substrate purified from bovine neutrophils. LE/2 and 3 reveal a relatively faint 40-kDa band on immunoblots of crude brain membranes or of purified brain Gi/Go. LE/2 and 3 do not react with transducin-alpha or Go alpha nor with the 41-kDa form of pertussis toxin substrate in brain, Gi alpha-1. These antisera distinguish between the major pertussis toxin substrates of brain and neutrophil and tentatively identify the latter as Gi alpha-2.     

Inhibition of voltage-dependent Ca2+ currents and activation of pertussis toxin-sensitive G-proteins via muscarinic receptors in GH3 cells.

In the rat pituitary cell line GH3, carbachol inhibits PRL secretion in a pertussis toxin-sensitive manner. For elucidation of the underlying mechanisms, we studied the effect of carbachol on voltage-dependent Ca2+ currents. Under voltage-clamp conditions, carbachol inhibited whole-cell Ca2+ currents by about 25%. This inhibitory action of carbachol was not observed in cells treated with pertussis toxin, indicating the involvement of a pertussis toxin-sensitive G-protein. In membranes of GH3 cells, carbachol stimulated a pertussis toxin-sensitive high-affinity GTPase. In immunoblot experiments with peptide antisera, we identified two forms of the Gi alpha-subunit (41 and 40 kDa) and two forms of the Go alpha-subunit (40 and 39 kDa). The 40-kDa Gi alpha-subunit was recognized by an antibody specific for the Gi2 alpha-subunit, and the 39-kDa Go alpha-subunit was detected by an antibody specific for the Go2 alpha-subunit. Incubation of membranes with the photoreactive GTP analog [alpha-32P]GTP azidoanilide resulted in photo-labelling of 40- and 39-kDa pertussis toxin substrates comigrating with G-protein alpha-subunits of the corresponding molecular masses. Carbachol dose-dependently stimulated incorporation of the photoreactive GTP analog into the 39-kDa pertussis toxin substrate and, to a lesser extent, into 40-kDa pertussis toxin substrates. The data indicate that muscarinic receptors of GH3 cells couple preferentially to Go, which is likely to be involved in the inhibition of secretion, possibly by conferring an inhibitory effect to voltage-dependent Ca2+ channels.     

GTP-binding proteins in human platelet membranes serving as the specific substrate of islet-activating protein, pertussis toxin

Two GTP-binding proteins serving as the specific substrate of islet-activating protein (IAP), pertussis toxin, were purified from human platelet membranes as heterotrimers with an alpha beta gamma-subunit structure. The alpha of the major IAP substrate had a molecular mass of 40 kDa and differed from that of Gi 1 or Go previously purified from brain membranes. The partial amino acid sequences of the 40 kDa alpha completely matched with the sequences which were deduced from the nucleotide sequences of the human Gi 2 alpha gene. On the other hand, the alpha of the minor IAP substrate purified from human platelets was about 41 kDa and cross-reacted with an antibody raised against alpha of brain Gi 1 (Gi 1 alpha). These results indicate that the major IAP substrate present in human platelet membranes is a product of the Gi 2 alpha gene.     

Possible involvement of different GTP-binding proteins in noradrenaline- and thrombin-stimulated release of arachidonic acid in rabbit platelets.

Noradrenaline (NA) stimulated the release of arachidonic acid (AA) from the [3H]AA-labelled rabbit platelets via alpha 2-adrenergic receptors, since the effect of NA was inhibited by yohimbine. The stimulatory effect of NA in digitonin-permeabilized platelets was completely dependent on the simultaneous presence of GTP and Ca2+. The NA- and thrombin-stimulated releases of AA were markedly decreased by the prior ADP-ribosylation of the permeabilized platelets with pertussis toxin. Antiserum directed against the pig brain Go (a GTP-binding protein of unknown function), recognizing both alpha 39 and beta 35,36 subunits, but not alpha 41, of pig brain, reacted with 41 kDa and 40 kDa bands, with not one of 39 kDa, in rabbit platelet membranes. Anti-Go antiserum inhibited guanosine 5'-[gamma-thio]triphosphate-, A1F4(-)-, NA- and thrombin-stimulated AA releases in the membranes. Although the effect of thrombin was inhibited by low concentrations of anti-Go antiserum, high concentrations of the antiserum was needed for inhibition of the NA effect. Antiserum directed against the pig brain G1 (inhibitory G-protein), recognizing both alpha 41 and beta 35,36 subunits, but not alpha 39, of pig brain, reacted with the 41 kDa band in platelets. Anti-G1 antiserum inhibited only the effect of NA. Reconstitution of the platelet membranes ADP-ribosylated by pertussis toxin with Go, not Gi, purified from pig brain restored the thrombin-stimulated release of AA. In contrast, reconstitution of those membranes with Gi, not Go, restored the NA-stimulated release of AA. These results indicate that different GTP-binding proteins, Gi- and Go-like proteins, may be involved in the mechanism of signal transduction from alpha 2-adrenergic receptors and thrombin receptors to phospholipase A2 in rabbit platelets.     

Ontogenesis of α2-Adrenoceptor Coupling with GTP-Binding Proteins in the Rat Telencephalon

The ontogenesis of alpha 2-adrenoceptors and GTP-binding proteins and their coupling activity were investigated in telencephalon membranes of developing rats. The manganese-induced elevation of [3H]clonidine binding was increased in an age-dependent manner but the guanosine 5'-O-(3-thio)triphosphate-induced decrease in binding did not change. The extent of the binding of [3H]clonidine at 15 nM (saturable concentration) increased in an age-dependent manner and reached the adult level at 4 days after birth. Cholera toxin and pertussis toxin catalyzed ADP-ribosylation of proteins of 46 and 41/39 kilodaltons (kDa) in solubilized cholate extracts of the membranes. The 41/39-kDa proteins ADP-ribosylated by pertussis toxin (Gi alpha + Go alpha) were increased with age and reached the adult level at day 12, whereas the 46-kDa protein (Gs alpha) reached its peak on day 12 and then decreased to the fetal level at the adult stage. The immunoblot experiments of the homogenates with antiserum (specific antibody against alpha- and beta-subunit of GTP-binding proteins) demonstrated that the 39-kDa alpha-subunit of (Go alpha) and the 36-kDa beta-subunit of GTP-binding protein (beta 36) increased with postnatal age. In contrast, 35-kDa beta-subunit (beta 35) did not change. From these results, it is suggested that the coupling activity of alpha 2-adrenoceptor with GTP-binding protein gradually develops in a manner parallel with the increase of alpha 2-adrenoceptor and pertussis toxin sensitive GTP-binding proteins, Gi, and that alpha 39 beta 36 gamma may be related to the differentiation and/or growth of nerve cells in rat telencephalon.     

G-protein distribution in canine cardiac sarcoplasmic reticulum and sarcolemma: comparison to rabbit skeletal muscle membranes and to brain and erythrocyte G-proteins

Herein we describe the distribution of G-proteins in canine cardiac sarcolemma (SL) and sarcoplasmic reticulum (SR) and in rabbit skeletal muscle SL, T-tubules, and junctional and longitudinal SR in comparison to G-proteins of human erythrocyte and bovine brain. G-proteins were unequivocally present in cardiac SL and SR and in skeletal T-tubules. Both cardiac fractions had two substrates specifically ADP-ribosylated by cholera toxin migrating on a sodium dodecyl sulfate-polyacrylamide gel at about 42 and 45 kDa. In skeletal muscle membranes, cholera toxi-labeled substrates migrated at about 42 and 62 kDa. Three substrates for pertussis toxin were resolved by sodium dodecyl sulfate/urea-polyacrylamide gel electrophoresis in cardiac SL at about 38, 40, and 43 kDa. Only the two higher molecular weight substrates were detected in cardiac SR and in any of several skeletal muscle membrane fractions. Comparison of G-proteins in muscle membrane fractions with G-proteins isolated from bovine brain and human erythrocyte as well as their reaction with antisera to either a common sequence of alpha subunits of G-proteins (G alpha common antibody) or to a unique sequence of the alpha subunit of Go (G alpha o antibody) indicated that the two lower molecular weight bands in cardiac SL are Go or Go-like, and therefore the upper band is probably Gi. These data demonstrate that pertussis toxin substrates are more heterogeneous than previously described and have implications for studies attempting to attribute physiological functions to G-protein isolates.     

Identification of G Protein Subtypes in Peripheral Nerve and Cultured Schwann Cells

In this study, we investigated the expression of various G proteins in whole sciatic nerves, in myelin and nonmyelin fractions from these nerves, and in membranes of immortalized Schwann cells. In myelin, nonmyelin, and Schwann cell membranes we detected two 39-40-kDa pertussis toxin substrates that were resolved on separation on urea-gradient gels. Two cholera toxin substrates with apparent molecular masses of 42 and 47 kDa were present in nerve and brain myelin and in Schwann cell membranes. In these membranes, a third 45-kDa cholera toxin substrate, which displayed the highest labeling, was also present. Immunoblotting with specific antisera allowed the identification of G(o) alpha, Gi1 alpha, Gi2 alpha, Gi3 alpha, Gq/G11 alpha, and the two isoforms of Gs alpha in nerve homogenates, nerve, and brain myelin fractions. In Schwann cell membranes we identified G(o) alpha, Gi2 alpha, Gi3 alpha, and proteins from the Gq family, but no immunoreactivity toward anti-Gi1 alpha antiserum was detected. In these membranes, anti-Gs alpha antibody recognized the three cholera toxin substrates mentioned above, with the 45-kDa band displaying the highest immunoreactivity. Relative to sciatic nerve myelin, the Schwann cell membranes revealed a significantly higher expression of Gi3 alpha and the absence of Gi1 alpha. The different distribution of G proteins among the different nerve compartments might reflect the very specialized function of Schwann cells and myelin within the nerve.     

The adipocyte Go alpha-immunoreactive polypeptide is different from the alpha subunit of the brain Go protein.

Rat adipose tissue possesses two Bordetella pertussis toxin (PTX) substrates and, in the same 39-41 kDa molecular mass range, positive immunoreactivity has also been reported with antibodies against the alpha subunit of Go, the major brain GTP-binding protein (G-protein). In this study, the presence of the brain Go alpha subunit at 39 kDa in adipocytes was reassessed, since direct correspondence between PTX substrates and Go alpha immunoreactivity has not yet been clearly established. On resolutive SDS/polyacrylamide-gel electrophoresis, the PTX substrates of human adipocytes were compared with the three PTX substrates found in brain. No ADP-ribosylated substrate at the level of the 39 kDa brain Go alpha could be detected in adipocyte membranes. Immunoblotting of human adipocyte membranes stained with our anti-Go alpha antibodies confirmed the presence of a positive immunoreactivity in this tissue, but the apparent molecular mass of the immunoreactive polypeptide in adipocytes was higher than that found in nervous tissues. Taken together, these results indicate that the brain Go alpha subunit is not present in adipose tissue. They also suggest the existence of a G-protein in adipocytes which is immunologically related to Go alpha but having a slightly higher molecular mass.     

A new GTP-binding protein in brain tissues serving as the specific substrate of islet-activating protein, pertussis toxin

A new GTP-binding protein serving as the specific substrate of islet-activating protein (IAP), pertussis toxin, was purified from porcine brain membranes as an alpha beta gamma-heterotrimeric structure. The alpha-subunit of the purified protein (alpha 40 beta gamma) had a molecular mass of 40 kDa and differed from that of Gi (alpha 41 beta gamma) or Go (alpha 39 beta gamma) previously purified from brain tissues. The fragmentation patterns of limited tryptic digestion and immunological cross-reactivities among the three alpha were different from one another. However, the beta gamma-subunit resolved from the three IAP substrates similarly inhibited a membrane-bound adenylate cyclase and their beta-subunits were immunologically indistinguishable from one another. Thus, the alpha 40 beta gamma is a new IAP substrate protein different from Gi or Go, in the alpha-subunit only.     

A Go-like protein in Drosophila melanogaster and its expression in memory mutants.

G proteins couple receptors for extracellular signals to several intracellular effector systems and play a key role in signalling transduction mechanisms. In particulate preparations of Drosophila melanogaster heads, only one substrate for pertussis toxin at 39-40 kd was detected. This substrate, which showed only one isoform when analysed by isoelectric focusing, was recognized by immunoblotting and immunoprecipitation techniques using a polyclonal antibody against the alpha subunit of the Go protein purified from bovine brain and can be thus considered as a Go-like protein. Antibodies obtained against a carboxy-terminal sequence of the alpha subunit of Go (but not of Gi1 or Gi2) and against an internal sequence shared by all the alpha subunits, were also able to cross-react with the alpha subunit of this protein in insects. We have also studied the Go-like protein in several D.melanogaster mutants, primarily in memory and learning mutants. In these mutants there was a sex-dependent enhancement in pertussis toxin-catalysed ADP-ribosylation with respect to the wild-type. This increase could be attributed in part to an increase in the alpha subunit of the Go-like protein, as revealed by immunoblotting with anti-Go alpha polyclonal antibody. This report constitutes the first evidence for the participation of a Go protein in learning and memory.     

Identification of a new GTP-binding protein. A Mr = 43,000 substrate for pertussis toxin

In purified preparations of human erythrocyte GTP-binding proteins, we have identified a new substrate for pertussis toxin, which has an apparent molecular mass of 43 kDa by silver and Coomassie Blue staining. Pertussis toxin-catalyzed ADP-ribosylation of the 43-kDa protein is inhibited by Mg2+ ion and this inhibition is relieved by the co-addition of micromolar amounts of guanine nucleotides. GTP affects the ADP-ribosylation with a K value of 0.8 microM. Addition of a 10-fold molar excess of purified beta gamma subunits (Mr = 35,000 beta; and Mr = 7,000 gamma) of other GTP-binding proteins results in a significant decrease in the pertussis toxin-mediated ADP-ribosylation of the 43-kDa protein. Treatment of the GTP-binding proteins with guanosine 5'-O-(thiotriphosphate) and 50 mM MgCl2 resulted in shifting of the 43-kDa protein from 4 S to 2 S on sucrose density gradients. Immunoblotting analysis of the 43-kDa protein with the antiserum A-569, raised against a peptide whose sequence is found in the alpha subunits of all of the known GTP-binding, signal-transducing proteins (Mumby, S. M., Kahn, R. A., Manning, D. R., and Gilman, A. G. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, 265-259) showed that the 43-kDa protein is specifically recognized by the common peptide antiserum. A pertussis toxin substrate of similar molecular weight was observed in human erythrocyte membranes, bovine brain membranes, membranes made from the pituitary cell line GH4C1, in partially purified GTP-binding protein preparations of rat liver, and in human neutrophil membranes. Treatment of neutrophils with pertussis toxin prior to preparation of the membranes resulted in abolishment of the radiolabeling of this protein. From these data, we conclude that we have found a new pertussis toxin substrate that is a likely GTP-binding protein.     

GTP-binding proteins in luminal and basolateral membranes from pars convoluta and pars recta of rabbit kidney proximal tubule.

The GTP-binding proteins on luminal and basolateral membrane vesicles from outer cortex (pars convoluta) and outer medulla (pars recta) of rabbit proximal tubule have been examined. The membrane vesicles were highly purified, as ascertained by electron microscopy, by measurements of marker enzymes, and by investigating segmental-specific transport systems. The [35S]GTP gamma S binding to vesicles, and to sodium cholate-extracted proteins from vesicles, indicated that the total content of GTP-binding proteins were equally distributed on pars convoluta, pars recta luminal and basolateral membranes. The membranes were ADP-ribosylated with [32P]NAD+ in the presence of pertussis toxin and cholera toxin. Gel electrophoresis revealed, for all preparations, the presence of cholera toxin [32P]ADP-ribosylated 42 and 45 kDa G alpha s proteins, and pertussis toxin [32P]ADP-ribosylated 41 kDa G alpha i1, 40 kDa G alpha i2 and 41 kDa G alpha i3 proteins. The 2D electrophoresis indicated that Go's were not present in luminal nor in basolateral membranes of pars convoluta or pars recta of rabbit proximal tubule.     

Purification of heterotrimeric GTP-binding proteins from brain: identification of a novel form of Go

Using high-resolution Mono-Q anion-exchange chromatography, we purified four distinct GTP-binding proteins from bovine brain. Each consists of alpha and associated beta/gamma subunits, and each is a substrate for pertussis toxin catalyzed ADP-ribosylation. We defined the relationship between the alpha subunits of the purified proteins and cloned cDNAs encoding putative alpha subunits (1) by performing immunoblots with peptide antisera with defined specificity and (2) by comparing the migration on two-dimensional gel electrophoresis of the purified proteins, and of the in vitro translated products of cDNAs encoding alpha subunits. Purified G proteins with alpha subunits of 39, 41, and 40 kDa (G39, G41, and G40 in order of abundance) correspond to the products of Go, Gi1, and Gi2 cDNAs. We purified a novel G protein with an alpha subunit slightly above 39 kDa (G39*). G39* is less abundant than G39, elutes earlier than G39 on Mono-Q chromatography, and has a more basic pI (6.0 vs 5.6) than G39. G39 and G39*, however, are indistinguishable on immunoblots with a large number of specific antisera. The data suggest that G39* may represent a novel form of Go, differing in posttranslational modification rather than primary sequence.     

Antisera against the 3-17 sequence of rat G alpha i recognize only a 40 kDa G-protein in brain

To obtain antisera specific for the GTP-binding protein Gi alpha we immunized rabbits against a synthetic peptide derived from the N-terminal (3-17) sequence predicted from the rat Gi alpha cDNA clone published by Itoh et al. (1986) (Proc. Natl. Acad. Sci. USA 83, 3776-3780). Western-blot analysis of bovine brain G-proteins purified and resolved by hydrophobic chromatography and of rat striatal membranes, indicate that this antiserum does not recognize 41 kDa alpha i or 39 kDa alpha o. However, it reacts with a 40 kDa alpha-subunit. The data suggest that the sequence deduced from the rat G alpha i cDNA corresponds to a G40 alpha protein and that N-terminus directed antisera are useful tools to discriminate between two different G alpha i-like types of G-proteins present in mammalian brain.     

Adenylyl cyclase and G-proteins in Phytomonas

Phytomonas sp. membranes have an adenylyl cyclase activity which is greater in the presence of Mn2+ than with Mg2+. The Mg2+ and Mn2+ activity ratio varies from one membrane preparation to another, suggesting that the adenylyl cyclase has a variable activation state. A[35S]GTP-gamma-S-binding activity with a Kd of 171 nM was detected in Phytomonas membranes. Incubation of these membranes with activated cholera or pertussis toxin and [adenylate 23P]NAD+ led to incorporation of radioactivity into bands of about 40-44 kDa. Crude membranes were electrophoresed on SDS-polyacrylamide gels and analyzed, by Western blotting, with the 9188 anti-alpha[s] antibody and the AS/7 antibody (anti-alpha[i], anti-alpha[i1], and anti-alpha[i2]. These procedures resulted in the identification of polypeptides of approximately 40-44 kDa. Phytomonas adenylyl cyclase could be activated by treatment of membrane preparations with cholera toxin, in the presence of NAD+, while similar treatment with pertussis toxin did not affect this enzyme activity. These studies indicate that in Phytomonas, adenylyl cyclase activity is coupled to an unknown receptor entity through G alpha[s] proteins.     

Presence of three pertussis toxin substrates and Go alpha immunoreactivity in both plasma and granule membranes of chromaffin cells

GTP-binding proteins have been proposed to be involved in some secretory processes. Bordetella pertussis toxin is known to catalyze ADP-ribosylation of several GTP-binding proteins. In this paper, the subcellular localization of B. pertussis toxin substrates has been explored in chromaffin cells of bovine adrenal medulla. With appropriate gel electrophoresis conditions, three ADP-ribosylated substrates of 39, 40 and 41 kDa were detectable in both plasma and granule membranes. The more intense labelling occurred on the 40 kDa component, while the 41 kDa species exhibited electrophoretic mobility similar to that of Gi alpha. Significant immunoreactivity with anti-Go alpha antibodies was detected at the level of the 39 kDa faster component. The association of G-proteins with granule and plasma membranes suggests the involvement of these proteins in the exocytotic process or in its regulation.     

Immunological and biochemical differentiation of guanyl nucleotide binding proteins: interaction of Go alpha with rhodopsin, anti-Go alpha polyclonal antibodies, and a monoclonal antibody against transducin alpha subunit and Gi alpha

Guanyl nucleotide binding proteins couple agonist interaction with cell-surface receptors to an intracellular enzymatic response. In the adenylate cyclase system, inhibitory and stimulatory effects are mediated through guanyl nucleotide binding proteins, Gi and Gs, respectively. In the visual excitation complex, the photon receptor rhodopsin is linked to its target, cGMP phosphodiesterase, through transducin (Gt). Bovine brain contains another guanyl nucleotide binding protein, Go. The proteins are heterotrimers of alpha, beta, and gamma subunits; the alpha subunits catalyze receptor-stimulated GTP hydrolysis. To examine the interaction of Go alpha with beta gamma subunits and rhodopsin, the proteins were reconstituted in phosphatidylcholine vesicles. The GTPase activity of Go alpha purified from bovine brain was stimulated by photolyzed, but not dark, rhodopsin and was enhanced by bovine retinal Gt beta gamma or by rabbit liver G beta gamma. Go alpha in the presence of G beta gamma is a substrate for pertussis toxin catalyzed ADP-ribosylation; the modification was inhibited by photolyzed rhodopsin and enhanced by guanosine 5'-O-(2-thiodiphosphate). ADP-Ribosylation of Go alpha by pertussis toxin inhibited photolyzed rhodopsin-stimulated, but not basal, GTPase activity. It would appear from this and prior studies that Go alpha is similar to Gt alpha and Gi alpha; all three proteins exhibit photolyzed rhodopsin-stimulated GTPase activity, are pertussis toxin substrates, and functionally couple to Gt beta gamma. Go alpha (39K) can be distinguished from Gi alpha (41K) but not from Gt alpha (39K) by molecular weight.(ABSTRACT TRUNCATED AT 250 WORDS)     

An alpha 40 subunit of a GTP-binding protein immunologically related to Go mediates a dopamine-induced decrease of Ca2+ current in snail neurons

Dopamine induces a decrease in voltage-dependent Ca2+ current in identified neurons of the snail H. aspersa. This effect is blocked by intracellular injection of activated B. pertussis toxin and of an affinity-purified antibody against the alpha subunit of bovine Go protein. The dopamine effect is mimicked by intracellular injection of mammalian alpha o. In snail nervous tissue, pertussis toxin ADP-ribosylates a single protein band on SDS gels, and this band is recognized in immunoblots by the anti-alpha o antibody. We propose that this is a 40 kd alpha subunit of a molluscan G protein immunologically related to alpha o and that it mediates the effect of dopamine on Ca2+ currents in identified snail neurons.     

Major pertussis-toxin-sensitive GTP-binding protein of bovine lung. Purification, characterization and production of specific antibodies

Two GTP-binding proteins which can be ADP-ribosylated by islet-activating protein, pertussis toxin, were purified from the cholate extract of bovine lung membranes. Both proteins had the same heterotrimeric structure (alpha beta gamma), but the alpha subunits were dissociated from the beta gamma when they were purified in the presence of AlCl3, MgCl2 and NaF. The molecular mass of the alpha subunit of the major protein (designated GLu, with beta gamma) was 40 kDa and that of the minor one was 41 kDa. The results of peptide mapping analysis of alpha subunits with a limited proteolysis indicated that GLu alpha was entirely different from the alpha of brain Gi or Go, while the 41-kDa polypeptide was identical with the alpha of bovine brain Gi. The kinetics of guanosine 5'-[3-O-thio]triphosphate (GTP[gamma S]) binding to GLu was similar to that to lung Gi but quite different from that to brain Go. On the other hand, incubation of GLu alpha at 30 degrees C caused a rapid decrease of GTP[gamma S] binding, the inactivation curve being similar to that of Go alpha but different from that of Gi alpha. The alpha subunits of lung Gi and GLu did not react with the antibodies against the alpha subunit of bovine brain Go. The antibodies were raised in rabbits against GLu alpha and were purified with a GLu alpha-Sepharose column. The purified antibodies reacted not only with GLu alpha but also with the 41-kDa protein and purified brain Gi alpha. However, the antibodies adsorbed with brain Gi alpha reacted only with GLu alpha, indicating antisera raised with GLu alpha contained antibodies that recognize both Gi alpha and GLu alpha, and those specific to GLu alpha. These results further indicate that GLu is different from Gi or Go. Anti-GLu alpha antibodies reacted with the 40-kDa proteins in the membranes of bovine brain and human leukemic (HL-60) cells. The beta gamma subunits were also purified from bovine lung. The beta subunit was the doublet of 36-kDa and 35-kDa polypeptides. The lung beta gamma could elicit the ADP-ribosylation of GLu alpha by islet-activating protein, increase the GTP[gamma S] binding to GLu and protect the thermal denaturation of GLu alpha. The antibodies raised against brain beta gamma cross-reacted with lung beta but not with lung gamma.