Splice variants of the alpha subunit of the G protein Gs activate both adenylyl cyclase and calcium channels

Signal transducing guanine nucleotide binding (G) proteins are heterotrimers with different alpha subunits that confer specificity for interactions with receptors and effectors. Eight to ten such G proteins couple a large number of receptors for hormones and neurotransmitters to at least eight different effectors. Although one G protein can interact with several receptors, a given G protein was thought to interact with but one effector. The recent finding that voltage-gated calcium channels are stimulated by purified Gs, which stimulates adenylyl cyclase, challenged this concept. However, purified Gs may have four distinct alpha-subunit polypeptides, produced by alternative splicing of messenger RNA. By using recombinant DNA techniques, three of the splice variants were synthesized in Escherichia coli and each variant was shown to stimulate both adenylyl cyclase and calcium channels. Thus, a single G protein alpha subunit may regulate more than one effector function.

Recombinant alpha i-3 subunit of G protein activates Gk-gated K+ channels

G proteins, particularly those sensitive to pertussis toxin, are difficult to separate biochemically, creating uncertainty in functional assignments. For this reason the cDNAs encoding G alpha i-3 and two of the G alpha s splice variants were expressed as fusion proteins in Escherichia coli using a T7 promoter-based expression system. These proteins were denoted r alpha i-3 and r alpha s (short and long) and accumulated in bacteria to as much as 5-10% of total cellular protein, of which 5-10% was soluble in lysates. Soluble r alpha subunits were tested for stimulation of K+ channel activity in inside-out atrial membrane patches and for reconstitution of cyc- adenylyl cyclase activity. r alpha i-3, activated either by guanosine 5'-(3-thio)triphosphate (GTP gamma S) or AlF-4, stimulated in a concentration-dependent manner single channel K+ currents in isolated atrial membrane patches of three species: guinea pigs, neonatal rats, and embryonic chick. In contrast, GTP gamma S-activated r alpha s did not. In agreement with a similar study by Graziano et al. (Graziano, M. P., Casey, P. J. and Gilman, A. G. (1987) J. Biol. Chem. 262, 11375-11381), both r alpha s forms reconstituted GTP gamma S-stimulated cyc- adenylyl cyclase activity, albeit at concentrations 50-100 times higher than those needed with native Gs. The concentrations of r alpha i-3 needed to stimulate the K+ channels were also higher than needed with native human erythrocyte Gk, in this case 30-50 times. Single K+ channel currents stimulated by r alpha i-3 were indistinguishable from those stimulated by the natural effector acetylcholine. Thus, bacterial expression of G alpha subunits provided the means to demonstrate unequivocally that Gi-3 has intrinsic Gk activity.

The G protein-gated atrial K+ channel is stimulated by three distinct Gi alpha-subunits

The guanine nucleotide-binding protein, Gi, which inhibits adenylyl cyclase, has recently been shown to have three subtypes of the alpha-subunit, termed Gi alpha-1, Gi alpha-2 and Gi alpha-3. They share 87-94% amino-acid sequence homology and so are difficult to separate from one another. Among other functions, purified preparations activate K+ channels but there is confusion over which of the subtypes activates the muscarinic K+ channels of the atrial muscle of the heart: Gi alpha-3, also termed Gk, has been shown to activate this channel but it is not clear whether Gi alpha-1 does or does not. To clarify this problem, we expressed the subtypes separately in Escherichia coli to eliminate contamination by other subtypes and tested the recombinant alpha- chains on atrial muscarinic K+ channels. Although we anticipated that only Gi alpha-3 would have Gk activity, to our surprise all three recombinant subtypes were active, from which we deduce that the Gi subtypes are multifunctional.

Newly identified brain potassium channels gated by the guanine nucleotide binding protein Go

Potassium channels in neurons are linked by guanine nucleotide binding (G) proteins to numerous neurotransmitter receptors. The ability of Go, the predominant G protein in the brain, to stimulate potassium channels was tested in cell-free membrane patches of hippocampal pyramidal neurons. Four distinct types of potassium channels, which were otherwise quiescent, were activated by both isolated brain G0 and recombinant Go alpha. Hence brain Go can couple diverse brain potassium channels to neurotransmitter receptors.

Reticulocyte lysates synthesize an active alpha subunit of the stimulatory G protein Gs

We placed the cDNAs encoding one of the short types of alpha s (alpha s-1) with Asp-Ser in positions 70 and 71 and one of the long types of alpha s (alpha s-2) in which Asp-Ser are substituted with a string of 16 amino acids, into the pGEM-3 transcription vector downstream from its T7 RNA polymerase promoter, obtained transcripts and translated the mRNAs using a rabbit reticulocyte lysate system, to determine if the molecules would be synthesized and, if so, whether they would be active as assessed in cyc- reconstitution assays. The translation products obtained from both alpha s RNAs were a mixture of primarily three polypeptides of which one (approximately 40-50% of total) represented the complete translation product and the other two appeared to be due to internal translation starts at Met60, before the splice difference between the RNAs, and the other at the first Met after the splice difference. Lysates incubated with short or long alpha s RNA when added to cyc- membranes reconstituted fluoride and GTP[gamma S]-stimulated activities. Thus, in vitro synthesized alpha s subunits are active in interacting both with guanine nucleotides and the adenylyl cyclase enzyme. On incubation without and with the receptor agonist isoproterenol, using GTP as sole added guanine nucleotide, both types of alpha s subunits reconstituted the isoproterenol-stimulated adenylyl cyclase activity. Thus, the synthetic alpha s also interact with receptors, and by inference with beta-gamma dimers, shown previously to be needed for activation by receptor. Quantitative assays in which the activity of the synthetic alpha s-1 was compared to that of native purified human erythrocyte type-1 Gs, indicated that the two products are equipotent within a 2-fold margin of error. Thus, the lysate made fully active alpha s subunits, and alpha s subunits require no post-translational modifications dependent on microsomal processes. This approach may be useful in studying biological functions of other cloned alpha subunits of G proteins.

Alpha i-3 cDNA encodes the alpha subunit of Gk, the stimulatory G protein of receptor-regulated K+ channels

cDNA cloning has identified the presence in the human genome of three genes encoding alpha subunits of pertussis toxin substrates, generically called "Gi." They are named alpha i-1, alpha i-2 and alpha i-3. However, none of these genes has been functionally identified with any of the alpha subunits of several possible G proteins, including pertussis toxin-sensitive Gp's, stimulatory to phospholipase C or A2, Gi, inhibitory to adenylyl cyclase, or Gk, stimulatory to a type of K+ channels. We now report the nucleotide sequence and the complete predicted amino acid sequence of human liver alpha i-3 and the partial amino acid sequence of proteolytic fragments of the alpha subunit of human erythrocyte Gk. The amino acid sequence of the proteolytic fragment is uniquely encoded by the cDNA of alpha i-3, thus identifying it as alpha k. The probable identity of alpha i-1 with alpha p and possible roles for alpha i-2, as well as additional roles for alpha i-1 and alpha i-3 (alpha k) are discussed.

Screening of cDNA libraries with oligonucleotides as applied to signal transducing G proteins, receptors and effectors

Screening of cDNA libraries constructed in phage or plasmids with oligonucleotide probes has become one of the preferred cloning techniques with the least number of false positive failures. In this article we present our current protocols for designing the procedure to detect cDNA inserts and isolate them. We illustrate with primary screens for G protein subunits and membrane receptors.

Guanosine 5'-O-(3-thiotriphosphate) reduces ADP-ribosylation of the inhibitory guanine nucleotide-binding regulatory protein of adenylyl cyclase (Ni) by pertussis toxin without causing dissociation of the subunits of Ni. Evidence of existence of heterotrimeric pt+ and pt- conformations of Ni

The effect of the addition of guanosine 5'-O-(3-thiotriphosphate) (GTP gamma S), the GTP analog which activates the inhibitory guanine nucleotide-binding regulatory protein of adenylyl cyclase (Ni), on the pertussis toxin-mediated ADP-ribosylation reaction was studied in detail. Two effects were discerned: a stimulation of the ADP-ribosyltransferase activity of the toxin, akin to what was described for ATP and GDP in a previous report (Mattera, R., Codina, J., Sekura, R., and Birnbaumer, L. (1986) J. Biol. Chem. 261, 11173-11179), and a decrease in the ability of Ni to be a substrate for the activated toxin. Both effects were time-dependent with activation of the toxin being somewhat faster than inactivation of Ni. The effect of the addition of GTP gamma S on Ni was readily reversed by excess GDP and attenuated by increasing EDTA in the medium from 0.35 to 10 mM, suggesting dependence on trace concentrations of a divalent cation. It is suggested that this cation is Mg2+ on the basis that low (5-10 nM) concentrations of Mg2+ are needed for the endogenous GTPase activity of Ni (Sunyer, T., Codina, J., and Birnbaumer, L. (1984) J. Biol. Chem. 259, 15447-15451). Sucrose density gradient analysis of the Ni X GTP gamma S complexes with decreased susceptibility to ADP-ribosylation by pertussis toxin showed the same sedimentation parameters as Ni or Ni X GDP complexes, indicating that the molecule of Ni with GTP gamma S bound is heterotrimetric as opposed to dissociated into alpha i X GTP gamma S plus beta X gamma. Thus, these experiments define two conformations of heterotrimeric Ni: one -pt+, ADP-ribosylated by pertussis toxin, and the other pt-, poorly or not ADP-ribosylated by pertussis toxin. This latter, hitherto unrecognized conformation, is stabilized by the addition of strongly activating guanine nucleotides such as GTP gamma S and guanyl-5'-yl imidodiphosphate and should be important in the train of events that lead from an inactive heterotrimeric Ni to a fully active and dissociated Ni.

The human genome encodes at least three non-allellic G proteins with alpha i-type subunits

The amino acid sequence and composition of alpha-subunits of signal transducing G proteins of the same kind appear to vary by no more than 2% from species to species. Here we isolated a human liver cDNA using an oligonucleotide complementary to the sequences encoding the pertussis toxin (PTX) ADP-ribosylation site of the alpha-subunit of the rat brain G protein called Gi. Its open reading frame characterizes it as an alpha i-type cDNA--as opposed to alpha o-type--but predicts an amino acid composition that differs by 7% and 14%, respectively, from two other human alpha i-type molecules. Together with human brain alpha i (type-1) and human monocyte alpha i (type-2), the new human liver alpha i cDNA (type-3) forms parts of a family of alpha i molecules. Type-3 alpha i cDNA hybridizes to a approximately 3.6 kilobase long mRNA and type-2 alpha i cDNA hybridizes to an mRNA species of approximately 2.7 kilobases. This indicates that the human genome has at least three non-allellic genes encoding non-alpha o-type PTX substrates and provides structural evidence for the hypothesis that distinct effector systems are regulated by similar but nevertheless distinct PTX substrates.

Adenosine diphosphate ribosylation of G proteins by pertussis and cholera toxin in isolated membranes. Different requirements for and effects of guanine nucleotides and Mg2+

ADP ribosylation of membranes by pertussis toxin (PT) and cholera toxin (CT) was studied as a function of addition of ATP, various guanine nucleotides, Mg2+, and inorganic phosphate (Pi). ADP ribosylation of a 40 kilodalton (kDa) band by PT is markedly enhanced by ATP and GTP and is strongly inhibited by Pi or Mg2+. GTP analogs (GTP gamma S and GMP-adenyl-5'-yl imidodiphosphate) were less effective. In contrast, ADP ribosylation of two substrates for CT (of 42 and 50 kDa) is stimulated by Pi, Mg2+, and GTP or GTP analogs such as GTP gamma S, but is unaffected by ATP. These stimulatory conditions correlate well with GTP-mediated activation of stimulated nucleotide-binding regulatory component of adenyl cyclase. Optimal conditions for ADP ribosylation by PT do not correlate simply with conditions thought to lead to stabilization of an inactive form of inhibitory nucleotide-binding regulatory component of adenyl cyclase (Gi) or Gi-like protein; rather, the data suggest the involvement of both a stimulatory nucleotide site on PT (positively affected by either ATP or GTP) and a stabilizing site on the PT substrate (affected by GDP, GDP beta S, or GTP). Treatment of membranes with Lubrol PX increased ADP ribosylation by PT by as much as 25- to 30-fold, but inhibited the action of CT. Using defined conditions for ADP ribosylation by PT and CT, distinct labeling patterns were observed in thyroid, brain, corpus luteum, liver, heart, and erythrocytes membranes. All membranes were more intensely labeled by PT rather than CT.

Identification by molecular cloning of two forms of the alpha-subunit of the human liver stimulatory (GS) regulatory component of adenylyl cyclase

Two DNA molecules complementary to human liver mRNA coding for the alpha-subunit of the stimulatory regulatory component Gs of adenylyl cyclase were cloned. One of the two forms is a full-length cDNA of 1614 nucleotides plus a poly(A) tail of 59 nucleotides. The deduced sequence of 394 amino acids encoded by its open reading frame is essentially identical to that of the alpha-subunits of Gs identified by molecular cloning from bovine adrenals, bovine brain and rat brain. Two independent clones of the other type of cDNA were isolated. Both were incomplete, beginning within the open reading frame coding for the alpha s polypeptide. One codes for amino acids 5 through 394 and the other for amino acids 48 through 394 of the above described cDNA of 1614 nucleotides, and both have the identical 3'-untranslated sequence. They differ from the first cDNA, however, in that they lack a stretch of 42 nucleotides (numbers 214 through 255) and have nucleotides 213 (G) and 256 (G) replaced with C and A, respectively. This results in a predicted amino acid composition of another alpha-subunit of Gs that is shorter by 14 amino acids and contains two substitutions (Asp for Glu and Ser for Gly) at the interface between the deletion and the unchanged sequence. We call the smaller subunit alpha s1 and the larger alpha s2. This is the first demonstration of a structural heterogeneity in alpha s subunits that is due to a difference in amino acid sequence.

The interaction of nucleotides with pertussis toxin. Direct evidence for a nucleotide binding site on the toxin regulating the rate of ADP-ribosylation of Ni, the inhibitory regulatory component of adenylyl cyclase

The interaction of nucleotides with pertussis toxin (PT), and their effects on the ability of the toxin to ADP-ribosylate pure Ni, were evaluated. [32P]ATP (10 nM) bound directly to dithiothreitol-activated PT. This binding was competitively inhibited by nucleotides and anions with the following IC50 concentrations in order of decreasing potency: ATP = ATP gamma S (adenosine-5'-O-(3-thiotriphosphate)) = 0.2-0.3 microM, GDP beta S (guanosine-5'-O-(2-thiodiphosphate)) = 2-3 microM, GTP gamma S (guanosine-5'-O-(3-thiotriphosphate)) = 10-15 microM, ADP = 20-25 microM, GTP = 30-40 microM, GMP-P(NH)P (guanyl-5'-yl imidodiphosphate) = 100-150 microM, GDP = 150-200 microM, Pi = SO4(2-) = 20 mM and Cl- = acetate = 30-35 mM. Treatment of PT with ATP, AMP-P(NH)P, GTP, GDP, or GDP beta S, resulted in a stimulated state of NAD+-Ni ADP-ribosyltransferase activity. Addition of ATP, AMP-P(NH)P (adenyl-5'-yl imidodiphosphate), GTP, GDP, and GDP beta S to the ADP-ribosylation reactions resulted in increased rates of ADP-ribosyl-Ni formation. It is concluded that these effects on the nucleotides are due to their action to stimulate the activity of PT. At concentrations of PT between 0.04 and 0.4 microgram/ml, the stimulation of ADP-ribosylation of Ni effected by nucleotides was hysteretic in nature, exhibiting an approximately 25-min long lag when GDP was used as the activating nucleotide. These lags decreased with increasing concentrations of PT, and were abolished by pretreatment of the toxin with GDP or ATP. Preliminary incubation of Ni with GDP had no effect on the lag in its ADP-ribosylation by non-nucleotide treated PT. Addition of divalent cations (Mg2+, Mn2+, and Ca2+) inhibited formation of ADP-ribosyl-Ni, possibly by causing aggregation and denaturation of Ni. This is the first demonstration that both adenine and guanine nucleotides interact directly with PT and act to stimulate its activity to ADP-ribosylate Ni, and that guanine nucleotides do so regardless of whether they are nucleoside di- or triphosphates.

Guanine nucleotide regulation of a mammalian myocardial muscarinic receptor system. Evidence for homo- and heterotropic cooperativity in ligand binding analyzed by computer-assisted curve fitting

Highly purified dog heart sarcolemmal membranes, with a content of approximately 5 pmol of muscarinic acetylcholine receptor (mAChR)/mg of protein, were analyzed for mAChR-mediated inhibition of adenylyl cyclase and ligand binding in the absence and the presence of guanine nucleotides. Adenylyl cyclase was found to be coupled to the mAChR, being attenuated approximately 30% in a GTP-dependent manner. Direct binding studies, using 3H-labeled oxotremorine M, showed high affinity binding (apparent KD = 10 nM) that was reduced on nucleotide addition. Dose-response curves for GDP, GTP, and guanyl-5'-yl imidodiphosphate showed them to be equipotent. On the basis of pirenzepine binding, only one type of mAChR, commonly referred to as M2, was detected. Direct binding of [3H]quinuclidinyl benzilate [( 3H]QNB) uncovered 50% more binding sites than 150 nM 3H-labeled oxotremorine M; addition of guanine nucleotides uncovered the existence of positive cooperativity in the binding of [3H]QNB. Agonist displacement curves of [3H]QNB binding, without and with guanine nucleotides, extended over several orders of magnitude, which is inconsistent with single site competitive kinetics. The results and their analysis by computer-assisted curve fitting indicated that the data are well fitted by a model in which a receptor is at least bivalent and exists in two states: one with and the other without cooperativity between its sites, with guanine nucleotides decreasing both the degree of cooperativity between the sites and the proportion of the receptor that is in the cooperative form. Since the guanine nucleotide effect is mediated by the Ni coupling protein, it is suggested that direct binding detects R'Ni complexes (cooperative), R"NiG complexes (cooperative but distinct from R'Ni), and R0 complexes (non-cooperative and unaffected by Ni or NiG), where R = mAChR, Ni = the inhibitory regulatory component of adenylyl cyclase unaffected by guanine nucleotide, and NiG = Ni affected by guanine nucleotide (G).

Biological and immunological characterization of iodinated bovine growth hormone

Iodinated bovine growth hormone, containing no more than 1 g-atom of iodine per mole of hormone is generally used as a tracer in studies related to the action and metabolism of the hormone. This derivative was tested in different biological and immunological systems in which the hormone is known to be active. The iodinated derivative was almost indistinguishable in its properties from the native hormone when it was examined by the following criteria: body growth promoting activity, rat liver uptake in vivo, binding to rabbit liver microsomes and primary antigen-antibody interactions. Micro-complement fixation experiments suggested that the iodination produces minor alterations in the affinity of some antigenic determinants.

Structural characterization of iodinated bovine growth hormone

Bovine growth hormone (bGH) was submitted to iodination using limited amounts of oxidizing reagent, yielding a derivative with no more than 1-g-atom of iodine per mole of hormone. Analysis of the hydrolysis products indicated that monoiodotyrosine was almost the only product of substitution. Isolation and identification of the tryptic fragments showed that half of the 125I-labeled bGH molecules were iodinated in Tyr 174, followed by Tyr 158 (16%) and Tyr 42 (14%). Frontal gel chromatography indicated that the preparation did not contain significant amounts of unreacted bGH. Circular dichroism evidenced structural similarity between the native and the iodinated bGH. The iodinated hormone, like the native protein, undergoes self-association. The dissociation constant of the iodo-labeled bGH self-association equilibrium showed a two-fold increase when compared to that corresponding to the unlabeled hormone. At pH 8.5, where the equilibrium constant was estimated, one tenth of the molecules bear a charged iodotyrosyl residue (average pKapp = 9.3), which could account for part, if not all, of the observed difference regarding self-association. By this criterion, the presence of the iodine atom does not disturb the area engaged in dimer formation.