Inhibition and activation of fat cell adenylate cyclase by GTP is mediated by structures of different size

Correlation of transducin photoaffinity labeling with the specific formation of intermolecular disulfide linkages in its α-subunit

Transducin (T) is a heterotrimer of Tα, Tβ, and Tγ subunits. In the presence of light-activated rhodopsin, 8-azidoguanosine triphosphate (8-N3GTP) was covalently incorporated into T in a UV-light photodependent manner, with a low stoichiometry of 0.02 mol of 8-N3GTP per mol of T. Although Tα was preferentially labeled by 8-N3GTP, Tβ and Tγ were also modified. Photolabeling of T was specifically inhibited by GDP and GTP, but not by β,γ-imido-guanosine 5'-triphosphate (GMP-PNP), indicating that 8-N3GTP was modifying the GDP binding site of the holoenzyme. This was consistent with the observation that the photoaffinity probe was completely hydrolyzed to 8-N3GDP by T activated by illuminated rhodopsin. The formation of intermolecular disulfide associations in T was also determined because photolabeling of T was performed under non-reducing conditions. We established that Cys-347 of Tα was the major residue involved in the formation of disulfide-linked T oligomers. Other cysteines of Tα, such as Cys-321, also participated in the formation of disulfide bonds, revealing a complex pattern of intermolecular disulfide cross-links that led to the polymerization of T. The spontaneous generation of these cystines in Tα inhibited the light-dependent GTPase and GMP-PNP binding activities of T. A model was constructed illustrating that when two heterotrimers dimerize through the formation of disulfide bridges between the Cys-347 of their Tα subunits, the guanine ring of the 8-N3GDP bound to one T molecule might approach to the Tβγ-complex of the other heterotrimer. This model provides an explanation for the additional photolabeling of Tβ and Tγ by 8-N3GTP.

Structure-function relationships in adenylate cyclase systems

Hormone-sensitive adenylate cyclase systems are composed of hormone-recognition units (R), a nucleotide-regulatory unit (N) for reaction with GTP and divalent cations, and the catalytic unit (C). From the reported sizes of purified R and N subunits and target analysis of functional sizes of these units, the functions of the components for the binding and actions of hormones and GTP require minimally dimers, homologous or heterologous. It is proposed that the catalytic unit exists in the membrane also as a dimer and that its transition to the active state with MgATP as substrate involves corresponding transitions in linked dimers of the hormone-recognition and nucleotide-regulatory units. It is postulated that hormones trigger the activation process by inducing in concert with GTP and divalent cations the appropriate dimer structure of the holoenzyme. In large aggregates of such structures, realignment of only a few occupied holoenzyme units may be sufficient to induce activation of the total aggregate enzyme. This theory serves to explain the synergistic actions of hormones, and how several hormones can activate a common enzyme. It also provides an explanation for 'spare' receptors, and for the efficacy of hormone action.

Dimerization of mammalian adenylate cyclases

Mammalian adenylate cyclases are predicted to possess complex topologies, comprising two cassettes of six transmembrane-spanning motifs followed by a cytosolic, catalytic ATP-binding domain. Recent studies have begun to provide insights on the tertiary assembly of these proteins; crystallographic analysis has revealed that the two cytosolic domains dimerize to form a catalytic core, while more recent biochemical and cell biological analysis shows that the two transmembrane cassettes also associate to facilitate the functional assembly and trafficking of the enzyme. The older literature had suggested that adenylate cyclases might form higher order aggregates, although the methods used did not necessarily provide convincing evidence of biologically relevant events. In the present study, we have pursued this question by a variety of approaches, including rescue or suppression of function by variously modified molecules, coimmunoprecipitation and fluorescence resonance energy transfer (FRET) analysis between molecules in living cells. The results strongly suggest that adenylate cyclases dimerize (or oligomerize) via their hydrophobic domains. It is speculated that this divalent property may allow adenylate cyclases to participate in multimeric signaling assemblies.

Nobel Lecture. Signal transduction: evolution of an idea

“In general there is no set of observations conceivable which can give enough information about the past of a system to give complete information as to its future”: Norbert Wiener. “Think simplicity; then discard it”: Alfred North Whitehead

The disaggregation theory of signal transduction revisited: further evidence that G proteins are multimeric and disaggregate to monomers when activated

We have compared the sedimentation rates on sucrose gradients of the heterotrimeric GTP-binding regulatory (G) proteins Gs, G(o), Gi, and Gq extracted from rat brain synaptoneurosomes with Lubrol and digitonin. The individual alpha and beta subunits were monitored with specific antisera. In all cases, both subunits cosedimented, indicating that the subunits are likely complexed as heterotrimers. When extracted with Lubrol all of the G proteins sedimented with rates of about 4.5 S (consistent with heterotrimers) whereas digitonin extracted 60% of the G proteins with peaks at 11 S; 40% pelleted as larger structures. Digitonin-extracted Gi was cross-linked by p-phenylenedimaleimide, yielding structures too large to enter polyacrylamide gels. No cross-linking of Lubrol-extracted Gi occurred. Treatment of the membranes with guanosine 5'-[gamma-thio]triphosphate and Mg2+ yielded digitonin-extracted structures with peak sedimentation values of 8.5 S--i.e., comparable to that of purified G(o) in digitonin and considerably larger than the Lubrol-extracted 2S structures representing the separated alpha and beta gamma subunits formed by the actions of guanosine 5'-[gamma-thio]triphosphate. It is concluded that the multimeric structures of G proteins in brain membranes are at least partially preserved in digitonin and that activation of these structures in membranes yields monomers of G proteins rather than the disaggregated products (alpha and beta gamma complexes) observed in Lubrol. It is proposed that hormones and GTP affect the dynamic interplay between multimeric G proteins and receptors in a fashion analogous to the actions of ATP on the dynamic interactions between myosin and actin filaments. Signal transduction is mediated by activated monomers released from the multimers during the activation process.

Heterotrimeric G proteins in synaptoneurosome membranes are crosslinked by p-phenylenedimaleimide, yielding structures comparable in size to crosslinked tubulin and F-actin

We have treated rat brain synaptoneurosomes with the crosslinking agent N,N'-1,4-phenylenedimaleimide under conditions that cause extensive crosslinking of tubulin, F-actin, and the alpha and beta subunits of three major types of heterotrimeric GTP-binding regulatory proteins (G(o), Gs, Gi) present in brain membranes. The major crosslinked products are coeluted from Bio-Gel sizing columns as very large structures that do not penetrate stacking gels during SDS/PAGE. The alpha subunits but not the beta subunits of Gs, G(o) and Gi also yield crosslinked products of intermediate sizes. None of the products are as small as the heterotrimeric G proteins extracted from brain by cholate or Lubrol. However, the large and intermediate crosslinked structures are strikingly similar to the large, polydisperse structures of the alpha subunits of Gs, Gi, and G(o) extracted from synaptoneurosomes by the detergent octyl glucoside, which have sedimentation properties of multimeric proteins. Several ways in which multimeric forms of G proteins can explain the dynamic and pleiotropic actions of hormones and GTP on signal-transducing systems are discussed.

Octyl glucoside extracts GTP-binding regulatory proteins from rat brain "synaptoneurosomes" as large, polydisperse structures devoid of beta gamma complexes and sensitive to disaggregation by guanine nucleotides

GTP-binding regulatory proteins are generally purified from cholate-extracted membranes in the form of heterotrimers (G proteins) consisting of a GTP-binding subunit (alpha protein) complexed with a tightly interacted heterodimer termed beta gamma. In this study we extracted the proteins from rat brain "synaptoneurosomes" using the neutral detergent 1-octyl beta-D-glucopyranoside (octyl glucoside). Using specific antibodies for detection by immunoblotting and sucrose gradients for analyzing hydrodynamic properties, we found that each species of alpha protein (alpha subunits of stimulatory, inhibitory, and brain GTP-binding proteins) exhibited a broad range (4 S to greater than 12 S) of polydisperse structures with peak values (5 S to 7 S) considerably greater than that of heterotrimeric G proteins. The beta subunit proteins, for example, appeared as a homogeneous peak at 4.4 S within which only a fraction of the total alpha proteins can be associated. Incubation of octyl glucose extracts at 30 degrees C rapidly sedimented the alpha proteins but not the beta proteins. Incubation at 30 degrees C with guanosine 5'[gamma-thio]triphosphate (10-100 microM) prevented rapid sedimentation. Hydrodynamic analysis revealed that all alpha proteins were converted to approximately 4 S structures by the actions of guanosine 5'-[gamma-thio]triphosphate without change in the hydrodynamic properties of the beta proteins. Extraction of the membranes with sodium cholate instead of octyl glucoside resulted in complete loss of the large, polydisperse structures of the alpha proteins; the S values were approximately 4 S, in the range for beta proteins. These findings suggest that the transducing GTP-binding proteins in synaptoneurosomes exist as polydisperse, possibly multimer, structures of various size that are stable in octyl glucoside but destroyed by cholate. The polydisperse structures are not associated with beta gamma complexes and are sensitive to the disaggregating effects of guanosine 5'-[gamma-thio]triphosphate.

Modulation of ischemic-induced damage to cerebral adenylate cyclase in gerbils by calcium channel blockers

It has been previously established that prolonged bilateral carotid occlusion followed by recirculation produces damage to the synaptic enzyme adenylate cyclase in the frontal cortex of the gerbil. Since calcium entrance into the brain may account in part for the deleterious consequences of stroke, the present study examined whether pretreatment with calcium channel blockers would modify the effects of 60 min of bilateral ischemia plus 40 min of reflow on various parameters of cortical adenylate cyclase activation. In this context activation of cerebral homogenates by norepinephrine with or without 5'-guanylyl imidodiphosphate was preserved by pretreatment of ischemic gerbils with verapamil but worsened by flunarizine. In contrast, in particulate fractions (treated with EGTA to reduce metallic ion levels) the damage to the Mn2+-sensitive catalytic site of adenylate cyclase was prevented only by flunarizine. Pretreatment with the two calcium channel blockers resulted in an elevated basal activity of the enzyme, thereby reducing the response in the homogenate preparation to forskolin. Gerbils pretreated with verapamil tended to have an increased ability for survival resulting from the ischemic episode. Under in vitro conditions the enzyme preparations were not markedly influenced by either drug.

Further probes into the molecular sites of damage to cerebral adenylate cyclase following postischemic reperfusion

A variety of pharmacological agents were used as experimental probes to determine with greater precision the site(s) of damage to cerebral adenylate cyclase as a consequence of postischemic reperfusion in the gerbil. A paradigm of 60-min bilateral ischemia followed by 40-min reperfusion results in a decreased sensitivity of the catalytic site of adenylate cyclase to Mn2+. Likewise, the GTP-transducer site (guanine nucleotide regulatory or G protein) revealed depressed responses to GTP in the absence or presence of norepinephrine, dopamine agonists, substance P, yohimbine, and cholera and pertussis toxins. Moreover, a crude preparation of GTPase disclosed that damage elicited by postischemic reperfusion was directed to the higher-affinity form of this enzyme, which is associated with the overall function of the guanine nucleotide regulatory protein. Injury to adenylate cyclase was unrelated either to the ability of adrenergic ligands to bind to associated receptor sites or to the capacity of the brain to generate visual evoked potentials in response to visual stimuli.

Cyclic nucleotides in stroke and related cerebrovascular disorders

Evidence has steadily accumulated to indicate that the rapid fluctuations in cyclic nucleotides during primary and secondary stroke are more than epiphenomena of the disease. During acute phases of ischemia, anoxia or hypoxia cyclic AMP rapidly accumulates in cerebral tissue, cerebrospinal fluid (CSF) and venous plasma, while cyclic GMP either remains unchanged or declines. The massive release of transmitters (catecholamines and adenosine) or ionic fluxes (Na+ and K+) may account for these observations. If reflow is established through a previously occluded vessel cyclic AMP content rises even higher in conjunction with a sharp rise in cyclic GMP. It is during this reflow period subsequent to longer term stroke (30-60 min) that the synaptic membrane enzyme, adenylate cyclase, is especially vulnerable. Presumably the cause of injury to cell membrane systems results from excess lactic acid accumulation and/or Ca++ entry through the damaged blood-brain barrier. The latter initiates breakdown of membrane phospholipids with resultant synthesis of vasoactive prostaglandins and formation of free radicals causing further insult to membrane phospholipids. Thus drugs acting to inhibit formation of prostaglandins, scavenge free radicals, reduce lactate formation, inhibit Ca++ entry or stabilize cell membranes have been shown to possess varying degrees of protective action toward adenylate cyclase. Moreover, cyclic AMP has been found to reverse stroke-induced vasospasm in central vessels. Reduced cyclic AMP content in CSF has been used to monitor the severity of coma, whereas clinical improvement was associated with predictable increases in the cyclic nucleotide. Therefore, cyclic nucleotides and related membrane enzyme systems might be used as target molecules in which to develop future therapeutic strategies for prevention or treatment of stroke.

Protective action by methylprednisolone, allopurinol and indomethacin against stroke-induced damage to adenylate cyclase in gerbil cerebral cortex

Adenylate cyclase activity was investigated in either homogenate or particulate fractions from the frontal cerebral cortex of the gerbil following five experimental conditions of bilateral ischemia. After periods of 15 min ischemia, 15 min ischemia plus 15 min of recirculation or 60 min ischemia the enzyme generally displayed enhanced responses to GTP, norepinephrine (NE), dopamine (DA), NE + GTP and DA + GTP. Pretreatment of the gerbils with methylprednisolone, allopurinol or indomethacin did not significantly influence the outcome of these findings. When the animals were subjected to 60 min ischemia plus 15 min of reflow, enzyme responses to the stimulatory agents including forskolin and NaF were all reduced. Pretreatment with methylprednisolone, allopurinol or indomethacin prevented the damage to adenylate cyclase in the 60 min ischemia plus 15 min reflow animals. When animals were made ischemic for 15 min followed by one week of recovery, enzyme sensitivity to GTP, calmodulin-Ca++, NE, combinations thereof and forskolin were reduced in only the particulate fractions. Enzyme damage was reversed following methylprednisolone. Enzyme damage may result from generation of free radicals during reflow and drugs that either inhibit synthesis pathways generating free radicals, stabilize cell membranes or act as free radical scavengers may be therapeutically beneficial under specific conditions of stroke.

Kinetics of GTP-modulation of adenylate cyclase in gerbil cerebral cortex after bilateral ischemia

Changes in the sensitivity of adenylate cyclase and steady-state levels of cyclic AMP (adenosine 3',5'-monophosphate) occur in mammalian brain during ischemic episodes. In our previous investigation with the gerbil model of bilateral ischemia there was an indication that ischemic conditions produced an enhancement of GTP sensitivity of adenylate cyclase within the cerebral cortex. The present study employed a kinetic analysis to evaluate further the role of this GTP modulation of adenylate cyclase in the gerbil frontal cortex during periods of bilateral ischemia and recirculation. In general, after either 15-min (with or without 15-min reflow) or 60-min ischemia the Vmax to GTP (alone or with dopamine and norepinephrine) was increased. Under these conditions the ED50 for half-maximal enzyme activation was decreased, indicating a greater affinity of the transducer site for GTP during ischemia. However, if irreversible 60-min ischemia was followed by 15-min reflow the enzyme responses to GTP were now absent. An unexpected observation showed that the ED50 for GTP activation of cortical adenylate cyclase was likewise attenuated when sham-operated animals were compared to normal gerbils.

Role of calmodulin in the effect of guanyl nucleotides on rat hippocampal adenylate cyclase: involvement of adenosine and opiates

The adenylate cyclase activity of rat hippocampal plasma membranes can be stimulated by vasoactive intestinal polypeptide (VIP). Low concentrations (10(-9) to 19(-7) M) of 5'-guanylyl-imido diphosphate (GppNHp) evoke a transient inhibition of the enzyme, which is followed by stimulation with increasing GppNHp concentrations (10(-6) to 10(-4) M). Inclusion of Inclusion of ethyleneglycol-bis-(beta-aminoethylether)-N,N'-tetraacetic acid (EGTA) during incubation abolishes the GppNHp inhibition while preserving GppNHp activation. The stimulation induced by GppNHp is amplified by VIP, but the inhibition is unaffected. Adenosine analogs and opiates are inhibitory ligands in the presence of GTP, and their effects can be reversed by the appropriate receptor antagonists, 3-isobutyl-1-methylxanthine and naloxone. Treatment of membranes with trypsin abolishes the GppNHp-induced inhibition without affecting the GppNHp stimulation. The inhibition induced by GppNHp is also abolished by EGTA treatment followed by washing, which coincides with a reduction in the adenosine- and opiate-mediated, GTP-dependent inhibition. The GppNHp inhibition can be restored in EGTA-treated but not in trypsin-treated membranes by addition of calcium-calmodulin but not by Ca2+ or Mg2+. Calcium-calmodulin-depleted membranes lack calcium stimulation as well as GppNHp-induced inhibition, whereas untreated membranes and calcium-calmodulin-depleted membranes plus exogenous calcium-calmodulin showed calcium stimulation and GppNHp inhibition. These results suggest that calmodulin is involved in both Ca2+ stimulation and guanine nucleotide-mediated inhibition of rat hippocampal adenylate cyclase.

Structure-function relationships for hormone receptors and adenylate cyclase: the contribution of target size analysis

Target size analysis of radiation inactivation has been used to determine structural features of hormone receptors and adenylate cyclase. This method allows the determination of molecular sizes of components of enzyme- or receptor-systems in impure preparations and in intact membranes. Principles of application of target size analysis and basic concepts for interpretation are discussed. Reviewing both biochemical and target size data on insulin receptors and adenylate cyclase it is attempted to outline the potential and the limitations of this biophysical approach.

Structure of the turkey erythrocyte adenylate cyclase system

Target analysis of the turkey erythrocyte adenylate cyclase [ATP pyrophosphate-lyase (cyclizing), EC 4.6.1.1] system showed that the molecular weight of the ground state enzyme increases from 92,000 with MnATP as substrate and no stimulatory ligands to 226,000 when activated by fluoride ion or by 5'-guanyl imidodiphosphate (p[NH]ppG) subsequent to clearance of previously bound GDP. The identical increment in size (130,000) suggests that the same regulatory unit is involved in the activation by both effectors. When assayed with isoproterenol and p[NH]ppG, the enzyme system displayed a further increment in size of 90,000 daltons. Based on binding of the antagonist 125I-labeled hydroxybenzylpindolol, the beta-adrenergic receptor is about 90,000 daltons or the same as that seen for activation of the enzyme by isoproterenol through the beta-adrenergic-receptor. Because single targets were seen for the ground state enzyme system under all conditions, it would appear that the various regulatory and catalytic components are structurally linked prior to activation by hormone, guanine nucleotides, and fluoride ion. Furthermore, based on reported subunit sizes of the nucleotide regulatory and receptor components are composed of multiple subunits, either homologous or heterologous in structure.