Correlations among Hill parameters reflect models of activating ligand-gated ion channels.

Molecular biology has contributed a concept, novel in pharmacology, in which the receptor is an independent variable. Site-directed mutagenesis of ligand-gated ion channels is now commonplace. The mutant receptor is usually characterized by the Hill parameters that describe concentration-response curves from transfected, voltage-clamped cells. In this article, Charles Spivak describes how to convert parameters for realistic models of channel activation into Hill parameters. Correlations among the Hill parameters that the models enforce can be useful in tentatively assigning a physiological function to the mutation site.

The ionic channel of the nicotinic acetylcholine receptor is unable to differentiate between the optical antipodes of perhydrohistrionicotoxin.

The enantiomers of perhydrohistrionicotoxin were studied in their effects on endplate currents recorded at the junctional region of sartorius muscles of Rana pipiens. The two optical antipodes progressively decreased the peak amplitude of the endplate currents and were indistinguishable from each other at all times. The enantiomers shortened equally the time constants for endplate current decay, but did not alter their voltage sensitivities. Although perhydrohistrionicotoxin contains 4 chiral centers, complete steric inversion does not alter its effects on the acetylcholine receptor-ion channel complex. By contrast the recognition site of the AcChR is extremely sensitive to any change in the chirality of agonists.

Isoarecolone can inhibit nicotine binding and produce nicotine-like discriminative stimulus effects in rats.

Isoarecolone methiodide has been reported previously to be a potent agonist at peripheral nicotinic-cholinergic receptors. Both isoarecolone methiodide and isoarecolone HCl can produce contractures of the frog rectus abdominis muscle and can inhibit binding of [3H]-(-)-nicotine to rat brain membranes, although the methiodide is much more potent than the hydrochloride. In rats trained to discriminate the effects of nicotine from saline, there is generalization to isoarecolone HCl at doses that reduce overall rates of responding. This effect and the similar relative potencies of isoarecolone and nicotine in the biochemical and behavioural procedures support the view that the high-affinity binding site for [3H]-(-)-nicotine is the receptor mediating the discriminative effect.

Carbamyl analogues of potent nicotinic agonists: pharmacology and computer-assisted molecular modeling study.

To investigate how the substitution of NH2 for CH3 affects the activity of three, potent, semirigid nicotinic agonists, carbamyl analogues were synthesized. The carbamyl agonists were 1-methyl-4-carbamyl-1,2,3,6-tetrahydropyridine methiodide (1), 1-methyl-4-carbamylpiperidine methiodide (2), and 1-methyl-4-carbamylpiperazine methiodide (3). Their potencies (reciprocals of the equipotent molar ratios) at the frog neuromuscular junction with reference to carbamylcholine were 0.77, 0.052, and 0.15, respectively. The acetyl analogues were more potent by factors of 65, 175, and 17, respectively. Explanations for this variable reduction in activity were sought by using computer-assisted molecular mechanics and calculations of electrostatic potential contours. Bioactive conformations of 1-3 were assigned on the basis of a well-supported pharmacophore and the ground-state conformation of the highly potent (50 times that of carbamylcholine) prototype, isoarecolone methiodide (4). Agonist 3 and its acetyl analogue superimposed closely in their ground-state, bioactive conformations, and the differences in their electrostatic potential contours were the least among the three pairs. Accordingly, their potencies differed the least. Agonists 1 and 2 both showed greater differences (with respect to their acetyl analogues) in their electrostatic potential contours and greater differences in potency. Agonist 2, in addition, could achieve the bioactive conformation only at the expense of 2.8 kcal mol-1, and, correspondingly, its activity relative to its acetyl analogue was lowest of all.

Synthesis, pharmacology, and molecular modeling studies of semirigid, nicotinic agonists.

Eight nicotinic agonists were synthesized, and their potencies were estimated by contracture of the frog rectus abdominis muscle. The most potent, 1-methyl-4-acetyl-1,2,3,6-tetrahydropyridine methiodide (3b), 50 times as potent as carbamylcholine, served as a template for the rest. Although all of the agonists could easily conform to the putative nicotinic pharmacophore, their potencies spanned a nearly 10,000-fold range. This pharmacophore, therefore, may be necessary but deficient. Computer-assisted molecular modeling studies helped to delineate additional factors that may contribute to potency. The factors are (1) the ground-state conformation, (2) superimposability of the hydrogen bond acceptor and the cationic head onto the template, (3) electrostatic potential at the cationic head and at the hydrogen bond acceptor site, and (4) the presence of a methyl group bonded to the carbon atom that bears the hydrogen bond acceptor. A new program, ARCHEM, was used to calculate and to visualize electrostatic potentials at the van der Waals surfaces of the agonists.

Receptor binding and electrophysiological effects of dehydroepiandrosterone sulfate, an antagonist of the GABAA receptor.

Recently we demonstrated that [3H]dehydroepiandrosterone sulfate binds specifically to two populations of sites in rat brain membranes [Majewska et al. (1990) Eur. J. Pharmac. 189, 307-315]. As an extension of this work, we studied the biochemical and pharmacological properties of [3H]dehydroepiandrosterone sulfate binding to brain membranes and the effects of dehydroepiandrosterone sulfate on GABA-induced currents in cultured neurons. [3H]Dehydroepiandrosterone sulfate binding depended upon incubation time, pH, protein concentration, and incubation temperature. Thermal denaturation or pretreatment of the membranes with protease or phospholipase A2 reduced the binding by 54-85%. The higher affinity [3H]dehydroepiandrosterone sulfate binding sites appeared to be associated with protein and with the GABAA receptor complex. Among substances known to interact with the GABAA receptor complex, pregnenolone sulfate, pentobarbital, and phenobarbital inhibited the binding of [3H]dehydroepiandrosterone sulfate. High micromolar concentrations of dehydroepiandrosterone sulfate inhibited [3H]muscimol and [3H]flunitrazepam binding to rat brain membranes, primarily by reducing the binding affinities. Dehydroepiandrosterone sulfate also produced a concentration-dependent block of GABA-induced currents in cultured neurons from ventral mesencephalon (IC50 = 13 +/- 3 microM). The results of this study are consistent with an action of dehydroepiandrosterone sulfate as a negative noncompetitive modulator of the GABAA receptor. Because concentrations of dehydroepiandrosterone sulfate in the brain undergo physiological variations, this neurosteroid may play a vital role in regulation of neuronal excitability in the central nervous system.

Pharmacology of GABA rho 1 and GABA alpha/beta receptors expressed in Xenopus oocytes and COS cells.

1. The rho 1 protein, which we previously cloned from retina, assembles as a homooligomer that transduces the binding of gamma-aminobutyric acid (GABA) into robust chloride currents. However, its insensitivity to bicuculline, pentobarbitone and benzodiazepines, all potent agents at typical GABAA receptors, suggested that it may react atypically to other GABA agonists and antagonists. 2. cDNAs for the rho 1 and the alpha 5 beta 1 receptors for GABA were expressed as homo- and heterooligomers, respectively, in Xenopus oocytes. The selectivities of the respective receptors for various agonists were investigated using concentration-response experiments in voltage clamped cells. 3. The most potent agonists at the rho 1 receptor were trans-4-aminocrotonic acid (TACA) > GABA > muscimol; at the alpha 5 beta 1 receptor the rank order was muscimol > GABA > 4,5,6,7-tetrahydroisoxazole[4,5-c]pyridine-3-ol (THIP). The most specific agonists were cis-(2-(aminomethyl)-cyclopropyl-carboxylic acid (CAMP) and THIP for the rho 1 and the alpha 5 beta 1 receptors, respectively. 4. Comparing GABA, TACA and cis-aminocrotonic acid (CACA) at rho 1 receptors expressed in COS cells gave results almost indistinguishable from those found at oocytes; the pharmacology of rho 1 seems independent of the expression system. 5. Agonists THIP, piperidine-4-sulphonic acid (P4S), and isoguvacine, whose C-C-C-N chains are constrained by rings into a folded conformation and were potent at the alpha 5 beta 1 receptor, were among the weakest at the rho 1 receptor. However CACA and CAMP, which align better with the extended than the folded conformation, were weakest at the alpha 5 beta 1 receptor but moderately potent at the pl receptor. These findings suggest that the rho l receptor recognizes agonists in the extended conformation, in contrast to GABAA receptors, which are believed to recognize agonists in the partially folded conformation.6. In contrast to the alpha 5 beta 1 receptor, gradations in maximum responses were apparent in the rho l receptor,suggesting various degrees of partial agonism. In particular, imidazole-4-acetic acid (I4AA), whose maximum response was only 3% of GABA's maximum, had an apparent Kd for activating the rho l receptor of 16 microM; but it had an apparent Kd for competitively blocking the receptor of 0.64 microM. This difference suggests that steric constraints in the activated (open channel) receptor are tighter than in the resting receptor.7. Hill coefficients approached 2 at the rho l receptor, but were closer to unity at the alpha 5 beta 1 receptor. Thus,the rho l receptor displayed higher cooperativity.8. Unlike typical GABAA receptors, the rho l receptor was insensitive to the competitive antagonists bicuculline, SR95531, securinine, and (+)-tubocurarine.

Behavioural and ligand-binding studies in rats with 1-acetyl-4-methylpiperazine, a novel nicotinic agonist.

The novel nicotinic agonist 1-acetyl-4-methylpiperazine (AMP) has been studied in ligand-binding and behavioural studies. AMP methiodide potently inhibited [3H]-(-)-nicotine and [125I]-alpha-bungarotoxin binding to P2 membranes from rat brain and [125I]-alpha-bungarotoxin binding to rat skeletal muscles. AMP HCl also inhibited nicotinic binding, but it was 100 times less potent than AMP methiodide. In behavioural studies, AMP HCl reduced locomotor activity of experimentally naive rats and mecamylamine blocked this effect. In rats receiving (-)-nicotine chronically, AMP HCl did not increase locomotor activity consistently or to the same extent as (-)-nicotine. In rats trained to discriminate (-)-nicotine from saline in a two-bar operant conditioning procedure with food reinforcement, there was generalization to AMP HCl, but only at doses that reduced the overall rate of responding. The potency and effectiveness of AMP relative to (-)-nicotine varied across the different behavioural procedures. The results suggest that the pharmacodynamic action of AMP differs from that of (-)-nicotine and that it usefully extends the range of agonists that can be used as probes for central nicotinic mechanisms.

Mutagenesis of the GABA rho 1 receptor alters agonist affinity and channel gating.

Seventeen site-directed mutations were constructed in the GABA rho 1 receptor with the aim of finding agonist binding domains common to rho 1 and rho 2 receptors but distinct from those identified in members of the family of homologous, ligand gated ion channels. Mutated cDNAs were expressed in Xenopus oocytes and tested by voltage clamp experiments. Five of the mutations abolished responsiveness to GABA. Mutation Q189H, in the conserved cysteine loop, diminished apparent GABA affinity to about 1/10 of wild type values in a manner consistent with decreased allosteric cooperativity among agonist recognition sites. Mutation R316A, located in the extracellular loop between transmembrane domains II and III, increased the Hill coefficient to 3.9 in a fashion consistent with enhanced open probability of a receptor multimer.

Fluorescein-labeled naloxone binding to mu opioid receptors on live Chinese hamster ovary cells using confocal fluorescent microscopy.

A general method of confocal laser scanning microscopy was used to demonstrate specific binding of fluorescein-labeled naloxone (FNAL, 10-50 nM) to stably transfected mu opioid receptors on live Chinese hamster ovary cells. Nonspecific binding was visually indistinguishable from autofluorescence in cells with intact cell membranes. Fluorescent labeling of cell perimeters, not present in control nontransfected cells, reversed in transfected cells upon washout of FNAL or following the addition of either unlabeled naloxone (25 microM) or the mu specific antagonist CTOP (1 microM). The addition of the delta and kappa specific agonists DPDPE (1 microM) and U50488 (1 microM), respectively, failed to reverse the labeling. Further evidence of specific binding was obtained from kinetic experiments, where it was observed that only transfected cells showed a time-dependent exponential change in fluorescence that permitted estimation of association and dissociation binding rate constants of (5.8+/-0.5, mean+/-S.E.M.)x10(5) M(-1) s(-1) and (3.3+/-0.6)x10(-3) s(-1), respectively and a kinetically derived dissociation constant of 5.7+/-1.4 nM. These estimates were comparable to those obtained under similar conditions in radioligand binding experiments using [3H]-naloxone.

The neurosteroid dehydroepiandrosterone sulfate is an allosteric antagonist of the GABAA receptor.

Binding of the neurosteroid dehydroepiandrosterone sulfate (DHEAS) to rat brain synaptosomal membranes was studied in vitro, and the interaction of DHEAS with the GABAA receptor was tested using biochemical and electrophysiological assays. DHEAS bound to two populations of sites, and its binding was inhibited by barbiturates. DHEAS interfered with barbiturate-mediated enhancement of benzodiazepine binding. In cultured neurons from ventral mesencephalon, DHEAS reversibly blocked GABA-induced currents, behaving as an allosteric antagonist of the GABAA receptor.

Structural and electronic requirements for potent agonists at a nicotinic receptor.

A new agonist, isoarecolone methiodide (1,1-dimethyl-4-acetyl-1,2,3,6-tetrahydropyridinium iodide) was tested at the frog neuromuscular junction. It was 50 times more potent than carbamylcholine, making it one of the most potent nicotinic agonists known. In addition, its cyclic structure and conjugated carbonyl bond endow it with near rigidity. An analogous compound, 1,1-dimethyl-4-acetylpiperazinium iodide, was synthesized because of its similar geometry and rigidity. It was 2.6 times as potent as carbamylcholine but only 0.053 times as potent as isoarecolone methiodide. Computer assisted molecular modeling and molecular orbital calculations revealed steric and electrostatic field differences between these two compounds.

Desensitization and noncompetitive blockade of GABAA receptors in ventral midbrain neurons by a neurosteroid dehydroepiandrosterone sulfate.

Dehydroepiandrosterone sulfate (DHEAS) blocked the GABAA receptor noncompetitively in neurons grown in primary culture from the ventral midbrains of fetal rats. The apparent dissociation constant for this blockade was 4.5 microM, and one molecule of DHEAS was sufficient to block the receptor. The affinity of the blocked receptor for GABA was diminished by about one half. The findings that the DHEAS caused no rectification of chloride currents and that it did not shorten the durations of open ion channels indicated that DHEAS did not act by occluding open ion channels. Neither did it diminish their conductance. DHEAS accelerated desensitization in at least one population of receptors, diminished the amplitudes of inhibitory postsynaptic currents, and shortened their decay time constants in a concentration dependent manner.

The activation of the nicotinic acetylcholine receptor by the transmitter.

Experimental evidence has been published from isolated guinea pig muscle in vitro, and from direct ligand binding to receptors from T. californica, indicating that two agonist ions react with the nicotinic receptor by exchanging for one magnesium ion. It is the basis of the ion exchange receptor pair model, in which two acetylcholine ions exchange for one magnesium ion in contact with and between a pair of negatively charged receptor groups about 4 A apart. In the resting state the electrostatic attraction between the negatively charged receptor groups and the Mg2+ ion exerts a binding force. This binding force is opposed by the quantum mechanical repulsions of the electron clouds of the charged groups and ions in contact, together with the mutual repulsion of the pair of receptor oxyanions. When the Mg2+ ion is replaced by two acetylcholine ions the quaternary heads of the latter are positioned so that they form two mutually repelling ACh+ receptor group dipoles. As the Mg2+ ion leaves, its rehydration energy contributes to the sum of the electron cloud repulsions and the ACh+ receptor group dipole repulsions, causing the receptor groups to be forced apart activating the receptor macromolecule. The subsequent decrease in ACh+ concentration results in the reestablishment of the resting state. The coulombic electrostatic energy, the Born repulsion energy, the London attraction energy and the oxyanion ACh+ dipole repulsion energies have been calculated and shown to be consistent with the model. The displacement of the Mg2+ by two ACh+ ions makes several hundred kcals of energy available for receptor group separation and receptor activation.

Kinetics of recovery from opioids at wild-type and mutant mu opioid receptors expressed in xenopus oocytes.

To investigate a previous observation that classical antagonists behave as agonists at mutant H297N and H297Q mu opioid receptors, we compared the kinetics of recovery from opioids at wild-type and mutant mu receptors expressed in voltage-clamped Xenopus oocytes. The cDNA for the potassium channel GIRK1 was coinjected into the oocytes with that of the mu receptors to transduce agonist binding into a coupled electrophysiological response. The kinetics of recovery were estimated by brief test pulses of the agonist normorphine given at a frequency of 0.67 or 1 per min. After treatment with a variety of agonists, the receptors recovered from desensitization at rates that depended on the agonist, but there was little difference between mutant and wild-type receptors. Antagonists, however, induced agonist-like currents and demonstrated faster recovery at the mutant receptors. These results suggest that His-297 may comprise part of an antagonist subsite. This conclusion, when coupled with the steric theory that intrinsic activity depends on independent binary equilibration of a drug between agonist and antagonist subsites, could unify the paired observations that antagonists become agonists and recover faster at the mutant than at the wild-type receptors. Synapse 38:254-260, 2000. Published 2000 Wiley-Liss, Inc.

Triphenylmethylphosphonium blocks the nicotinic acetylcholine receptor noncompetitively.

Triphenylmethylphosphonium (TPMP) blocked the nerve-elicited twitch tension of frog sartorius muscles by 50% at a concentration of about 20 microM. This neuromuscular blockade by TPMP, which originated at the level of the nicotinic receptor, was due in part to the ability of the drug to block the quiescent receptor by degrees that depended on the TPMP concentration and on membrane potential. In addition, the blockade was markedly enhanced when the receptor was activated by acetylcholine. Finally, TPMP shortened the lifetimes of ACh-activated ion channels. These findings were interpreted as follows. Under resting conditions, TPMP shifted the equilibrium of the receptor channel complex toward the desensitized state. TPMP united with the activated ion channel to shorten channel lifetime and to deepen the blockade. High concentrations (greater than or equal to 50 microM) of TPMP altered muscle action potentials and often increased, by about 30-fold, the frequency of miniature endplate potentials.

Binding of semirigid nicotinic agonists to nicotinic and muscarinic receptors.

Isoarecolone methiodide (1-methyl-4-acetyl-1,2,3,6-tetrahydropyridine methiodide) was previously shown to be among the most potent agonists tested at the frog neuromuscular junction. Because nicotinic receptors from different sources vary in their selectivities, isoarecolone methiodide as well as 19 additional congeners, most of which were also previously tested at the frog neuromuscular junction, were studied in binding assays. Torpedo nobiliana was the tissue source for nicotinic receptors. Two types of experiments were conducted. The first evaluated the affinities of the agonists (including acetylcholine and carbamylcholine) for the recognition site by allowing the agonists to compete for that site with 125I-alpha-bungarotoxin. The inhibition potencies obtained correlated strongly (Spearman's correlation coefficient,-0.91) with the potency obtained at the frog neuromuscular junction. The second type of experiment evaluated the agonists for their ability to activate the receptor. The binding of [3H]perhydrohistrionicotoxin, which was employed as an indicator of the activation of the receptor, was measured in the presence of each of the agonists. Isoarecolone methiodide was the most potent of all. A few of the agonists (partial agonists) were incapable of fully enhancing this binding. For the full agonists, the concentration that produced half of the maximum binding of [3H]perhydrohistrionicotoxin was defined as the EC50. The correlation coefficient (Spearman's) for EC50 versus potency at the frog neuromuscular junction was -0.73, indicating innate differences between Torpedo and frog receptors. In addition, these compounds were tested for their affinity at muscarinic receptors from rat brain. Competition experiments were carried out using [3H]N-methylscopolamine. The affinity of isoarecolone methiodide was only about 7-fold lower than that of acetylcholine and less than 2-fold lower than that of carbamylcholine. In contrast, 1-methyl-4-acetylpiperazine methiodide was much more selective for nicotinic receptors. Its activity was similar to isoarecolone methiodide at the nicotinic receptor, but it was among the weakest compounds in its affinity for the muscarinic receptor.

Potencies and channel properties induced by semirigid agonists at frog nicotinic acetylcholine receptors.

Structure-activity relationships were investigated in a series of semirigid nicotinic agonists. Three of the agonists, (-)-ferruginine methiodide, arecoline methiodide, and its ketonic analogue arecolone methiodide, were cyclic analogues of anatoxin-a, a potent, naturally occurring, bicyclic alkaloid. Two other cyclic agonists, (-)-cytisine and (+/-)-muscarone, and the simplest agonist, the tetramethylammonium ion, were also tested. Arecolone methiodide and (-)-ferruginine methiodide have been tested as nicotinic agonists for the first time. Relative potency was assayed by contracture on the rectus abdominis muscle of the frog Rana pipiens. Natural, (+)-anatoxin-a, the most active of all of the agonists, was more than twice as potent as racemic anatoxin-a. Arecolone methiodide ranked after anatoxin-a in potency, being 8.6 times more potent than carbamycholine. A correlation between nicotinic potency and steric requirements probably involves the position of positively charged groups out of the plane defined by the carbonyl group and its two substituents. Channel properties induced by the agonists were evaluated by Fourier analysis of the end-plate current noise that resulted when the agonists were iontophoretically applied to frog sartorius muscle fibers. Average channel lifetimes were exponential functions of membrane potential, but the voltage sensitivity of channel lifetime seemed to vary among the agonists. Channel conductance, which was independent of membrane potential, also varied significantly among the agonists. The average charge traversing the membrane through each open channel, calculated from the product of average channel lifetime and current, did not correlate with potency. Therefore, the dominant component of potency is the frequency of channel opening. No clear relationship between the structure of the agonist and channel lifetime or conductance was evident.