Nondepolarizing neuromuscular blocking agents (NMBAs) are classic competitive-inhibitors at the muscle nicotinic acetylcholine receptor (nAChR). Although the fetal subtype muscle nAChR has been extensively studied at a molecular level, less is known about the interaction between nondepolarizing NMBAs and the human adult muscle nAChR. The aim of this study was to investigate the effect of clinically used nondepolarizing NMBAs at human adult muscle nAChRs and the mechanisms behind the inhibition.


Human subunits for the adult alpha(1)beta(1)delta(epsilon) muscle nAChR were cloned and expressed into Xenopus oocytes and thereafter studied with two-electrode voltage clamp. The effect of the clinically used nondepolarizing NMBAs, including atracurium, cis-atracurium, mivacurium, pancuronium, rocuronium, vecuronium, and d-tubocurarine, on acetylcholine-induced and dimethylphenylpiperazinium-induced currents were investigated.


All nondepolarizing NMBAs tested inhibited acetylcholine- and dimethylphenylpiperazinium-induced currents in human adult alpha(1)beta(1)delta(epsilon) muscle nAChRs, and no receptor activation was seen. Interestingly, acetylcholine desensitized the human adult alpha(1)beta(1)delta(epsilon) muscle type receptor and attenuated the inhibition caused by nondepolarizing NMBAs, as evident by lack of increase in IC(50) values for the nondepolarizing NMBAs with increased concentrations of acetylcholine. In contrast, dimethylphenylpiperazinium-induced currents were competitively inhibited by the nondepolarizing NMBAs.


This study demonstrates that nondepolarizing NMBAs inhibit human adult muscle nAChRs expressed in Xenopus oocytes by mixed mechanisms. When using the nondesensitizing agonist dimethylphenylpiperazinium, inhibition by the NMBA is competitive, whereas activation with high concentrations of acetylcholine in combination with NMBA induces a noncompetitive inhibition, which the authors speculate can involve receptor desensitization similar to that observed in the neuromuscular junction.

NONDEPOLARIZING neuromuscular blocking agents (NMBAs) are extensively used in routine practice of anesthesia and intensive care medicine to provide muscle relaxation. It is well known that nondepolarizing NMBAs block the transmission in the neuromuscular junction by inhibition of nicotinic acetylcholine receptors (nAChRs), both presynaptically and postsynaptically.1,2 

The nAChR is the prototype of the ligand-gated ion channel superfamily and is composed by five subunits arranged around a central cation pore.3The nicotinic receptors are further subdivided into muscle and neuronal subtypes, the muscle subtypes being composed of two α1 subunit: one β1, Δ, and γ/ϵ subunit, and the neuronal subtypes are α and β heteromers or α homomers.4In the neuromuscular junction, neuronal α3β2nAChRs are presynaptical autoreceptors, whereas the adult muscle α1β1Δϵ and/or the fetal α1β1Δγ subtype are located postsynaptically. Finally, during denervation/immobilization the neuronal α7nAChR has been demonstrated postsynaptically.5,6The fetal and adult subtypes of nAChRs have individual biophysical and pharmacological properties7,8; historically, however, the fetal muscle nAChR has been more extensively studied, and less is therefore known about the adult muscle subtype. Although affinity and potency data for nondepolarizing NMBAs at human  adult muscle nAChRs are sparse,1rodent data on adult muscle nAChRs expressed in Xenopus  oocytes has been published previously with somewhat divergent result.9–11Garland et al.  found vecuronium to be more potent at mouse adult nAChRs compared to d-tubocurarine and pancuronium, whereas Paul et al.  demonstrated that pancuronium was the most potent NMBA of the three.10,11In addition, a recent study found d-tubocurarine to have the highest affinity of several clinically used NMBAs at the mouse adult nAChR.12Notably, previous studies have shown that nondepolarizing NMBAs can act as both agonists, competitive and noncompetitive antagonists at rodent muscle nAChRs,11–14but we lack information on the mechanisms behind the human receptor inhibition. Although there is a large sequence homology between human and rodent nAChRs, differences in single amino acid sequences can cause relatively large functional affinity and kinetic differences.8,12To our knowledge, only one study has described the effect of clinically used nondepolarizing NMBAs at the human adult muscle nAChR, and no further characterization has been done.1 

Therefore, the aim of this study was to investigate the inhibition at the human adult muscle nAChR produced by clinically used nondepolarizing NMBAs.

Materials and Methods

In Vitro  Transcription

The human nAChR subunits α1, β1, Δ and ϵ were cloned from a human complementary DNA library, and the complementary DNA were subsequently subcloned into an expression vector, pKGem (AstraZeneca, Wilmington, DE).1,15messenger RNA was transcribed in vitro  using the mMessage mMachine® T7 kit (Ambion, Austin, TX) and analyzed using a bioanalyzer (Agilent Technologies, Palo Alto, CA).

Xenopus  Oocyte Injection

The study was approved by the local animal ethics committee at Karolinska Institutet, Stockholm, Sweden. Preparation and injection of oocytes and the electrophysiological recordings were done as previously described.1,15Briefly, Xenopus laevis  oocytes were isolated by partial ovariectomy from frogs anesthetized with 0.2% Tricaine. The ovaries were mechanically dissected to smaller lumps and digested in OR-2 buffer (in mm, NaCl 82.5, KCl 2, MgCl21, HEPES 5, pH adjusted to 7.5 with NaOH) containing 1.5 mg/ml collagenase (type 1A; Sigma, St. Louis, MO) for 60 min to remove the follicular epithelia from the oocytes. After 1–24 h, the oocytes were injected with 0.2–18 ng of messenger RNA in a total volume of 30–40 nl/oocyte. The subunit combinations were injected at a 1:1:1:1 ratio. The oocytes were maintained in Leibovitz L-15 medium (Sigma) diluted 1:1 with Millipore-filtered ddH2O (Billerica, MA) and 80 μg/ml gentamycin, 100 units/ml penicillin, and 100 μg/ml streptomycin added. Oocytes were incubated at 18–19°C for 2–7 days after injection before being studied.

Electrophysiological Recordings

All recordings were performed at room temperature (20–22°C). During recording, the oocytes were continuously perfused with ND-96 (in mm), NaCl 96.0, KCl 2.0, CaCl21.8, MgCl21.0, HEPES 5.0, pH 7.4, adjusted with NaOH. Oocyte recordings were performed using an integrated system that provides automated impalement of up to 8 oocytes in parallel. Two-electrode voltage clamp and current measurements were automatically coordinated with fluid delivery throughout the experiment. The individual recording chambers has a groove for the oocyte less than 2 mm from the drug-dispensing pipette on one side of the oval chamber, with aspiration on the opposite side of the oval. The ring design and the total chamber volume of 150 μl enables sub-second complete fluid exchanges at the oocyte (OpusXpress 6000A; Molecular Devices, Union City, CA). Electrodes were made from 1.5-mm borosilicate tubes (World Precision Instruments Inc, Sarasota, FL) and filled with 3 m KCl (0.5–2.5 MΩ resistance). The oocytes were voltage clamped at –60 mV.


Oocytes were continuously perfused with ND-96 at a rate of 2 ml/min in a 150-μl chamber. Drugs were delivered from a 96-well plate using disposable tips and administered at a rate of 2 ml/min for the first 2 s and at 1 ml/min thereafter. Concentration–response curves for the agonists, acetylcholine and dimethylphenylpiperazinium (DMPP), were constructed. To determine whether the tested nondepolarizing NMBAs activate and furthermore inhibit acetylcholine-induced currents, nondepolarizing NMBAs were applied for 55 s before a 20-s coapplication (preapplication) of both antagonist and agonist or as coapplication with 20 s of simultaneous agonist and antagonist application. The concentration of agonist in these inhibition experiments were chosen to represent concentrations below and above the EC50. Inhibition by nondepolarizing NMBAs of 1 μm acetylcholine has been published as control data in a previous paper.1Between each drug application, there was a 6-min washout period to allow clearance of the drugs and to avoid desensitization of the channels. Before and after each concentration–response experiment, three control responses were recorded at approximately EC50agonist concentration to exclude desensitization. Experiments were rejected if the postcontrol response was less than 80% of the precontrol response. To adjust for the level of channel expression, the responses in acetylcholine or DMPP concentration–response experiments were normalized to the peak response in each individual oocyte. For inhibition experiments, responses in each oocyte were normalized to the mean of the second and third agonist precontrols.


Acetylcholine, DMPP, and d-tubocurarine were purchased from Sigma. Atracurium and cis-atracrium were kindly provided by GlaxoSmithKline (Barnard Castle Durham, United Kingdom). Mivacurium (Mivacron®; GlaxoSmithKline, Mölndal, Sweden) was purchased. Org NC 97 (pancuronium), Org NC 45 (vecuronium), and rocuronium were provided by Organon, a part of Schering-Plough (Roseland, NJ). Chemicals used in buffers were purchased from Sigma unless otherwise stated. Stock solution of 1 m acetylcholine and 100 mm DMPP in ND-96 was prepared and frozen. Nondepolarizing NMBAs were prepared fresh each day and stored at +4°C. All drugs were then diluted in ND-96 immediately before use.

Data Analysis and Statistics

Offline analyses were made using Clampfit 9.2 (Molecular Devices). The baseline current immediately before drug application was subtracted from the response, and the analysis region was 20 s, i.e.  during the time of agonist application. Concentration–response relationships for acetylcholine were fitted by nonlinear regression (Prism 4.0; GraphPad, San Diego, CA) to the 4-parameter logistic equation:

where Y is the normalized response, x is the logarithm of concentration, and EC50is the logarithm of the concentration of agonist eliciting half-maximal response. This equation is based on the assumption that the ligand-receptor interaction yields a measurable response, in this case a current response. When NMBA-induced inhibition was studied, the same equation was used, and EC50was replaced by IC50, which is the concentration of antagonist eliciting half maximal inhibition: Bottom = 0, Top = 1. Unless otherwise stated, data are given as mean ± SEM or 95% confidence interval (95% CI). Differences in IC50values were compared by using paired or unpaired two-tailed Student t  test as appropriate. A P  value less than 0.05 was considered significant. GraphPad Prism (Prism 4.0, GraphPad) was used for statistics and plotting of graphs.


Acetylcholine and DMPP Concentration–response Relationships

Acetylcholine and DMPP produced concentration-dependent inward currents in voltage clamped oocytes injected with messenger RNA encoding the adult α1β1Δϵ muscle-type nAChR, whereas uninjected oocytes did not respond to acetylcholine or DMPP (data not shown). The responses to acetylcholine and DMPP in terms of kinetics and EC50values were consistent with previous reports from our group and others (fig. 1, table 1),1,9,15–17thus confirming functional expression of the receptor in this expression model. Interestingly, repeated concentration-response experiments on the same oocyte demonstrate that acetylcholine in contrast to DMPP seems to desensitize the receptor, as illustrated by a reduction in maximum current, see figure 1and table 1. Notably, there were no changes in the EC50values for acetylcholine (table 1), indicating that those receptors not desensitized had the same affinity for acetylcholine and thus a preserved pharmacology.

Nondepolarizing NMBAs Do Not Activate the Muscle-type Human nAChR

In this study and as previously demonstrated,1application of 1 nm to 100 μm of atracurium, cis-atracurium, d-tubocurarine, mivacurium, pancuronium, rocuronium, or vecuronium to oocytes expressing the human muscle (α1β1Δϵ) nAChR did not elicit any current (data not shown).

Inhibition Acetylcholine-induced Currents by Nondepolarizing NMBAs at the Muscle-type nAChRs

Atracurium, cis-atracurium, d-tubocurarine, mivacurium, pancuronium, rocuronium, and vecuronium concentration-dependently and reversibly inhibited 1 and 10 μm acetylcholine-induced currents in oocytes expressing the human α1β1Δϵ nAChR (fig. 2Aand table 2). As shown in table 2, the IC50for cis-atracurium, rocuronium, and vecuronium decreased with increased concentration of acetylcholine, whereas the IC50values were unchanged for d-tubocurarine, atracurium, mivacurium, and pancuronium. To rule out desensitization at the 10 μm acetylcholine concentration, we applied 1 and 10 μm acetylcholine for 20 s repeatedly with 6-min intervals and demonstrated that the 10 μm acetylcholine-induced response was associated with desensitization at repeated application in this system (fig. 2, B, C, and D).

Nondepolarizing NMBA (10 nm) was coapplied to the acetylcholine concentration-response relationships in each oocyte expressing the α1β1Δϵ nAChR subtype. As shown in figure 3, all nondepolarizing NMBAs except pancuronium shifted the acetylcholine concentration-response curve to the right, with an increased EC50value for atracurium, cis-atracurium, and d-tubocurarine (table 3). All nondepolarizing NMBAs except pancuronium and mivacurium reduced the peak acetylcholine response (fig. 3). Interestingly, pancuronium and mivacurium were also the only NMBAs to increase the IC50values with increased acetylcholine concentrations, more clearly showing a competitive inhibition in contrast to the other NMBAs. In these concentration-response experiments, the postcontrol responses were not preserved in a similar way as for repeated acetylcholine concentration-response curves. However, as discussed above (acetylcholine and DMPP concentration-response relationship), repeated acetylcholine concentration-response curves showed similar EC50values despite reduced current amplitude. Similar results for nondepolarizing NMBAs have been demonstrated using the rodent counterpart receptors.9–11 

On the basis of the above results and with the aim to investigate the inhibition of nondepolarizing NMBAs at the human muscle α1β1Δϵ nAChR in more detail, we continued to study the inhibition by rocuronium. Preapplication of rocuronium at 0.3, 1, 3, and 10 μm acetylcholine demonstrated that the IC50value increased with increased acetylcholine concentration in the lower range (0.3 to 1 μm), whereas there was a decrease in IC50values with 3 and 10 μm acetylcholine (table 4, fig. 4). To rule out open-channel block as a possible mechanism of inhibition, we coapplied rocuronium together with 0.3, 1, 3, and 10 μm acetylcholine. The IC50values for coapplication with rocuronium were higher or unchanged compared to the preapplication experiments (table 4, fig. 4), which excludes an open channel block as a mechanism for the inhibition.

Inhibition of DMPP-induced Currents by Nondepolarizing NMBAs at the Muscle-type nAChRs

DMPP is a specific nicotinic agonist without intrinsic channel blocking properties and produces minimal receptor desensitization. Atracurium and rocuronium both inhibited 10 and 100 μm DMPP-induced currents in oocytes expressing the human α1β1Δϵ nAChR in a dose-dependent and reversible fashion (fig. 5). When increasing the DMPP concentration from 10 to 100 μm, the IC50value of rocuronium increases slightly, whereas that of atracurium almost doubles (table 5). The increase is, however, not significant but in direct contrast to the observed decreased IC50values, using increasing acetylcholine concentrations (table 2 and 4). Taken together, this suggests that the NMBA inhibition of DMPP-induced human α1β1Δϵ nAChR currents is competitive rather than dependent on development of a desensitized receptor state.


In the current study, we can for the first time demonstrate that some of the clinically used nondepolarizing NMBAs in combination with acetylcholine inhibit acetylcholine-induced currents in the human adult muscle nAChR (α1β1Δϵ) in a noncompetitive manner. This is in contrast to the competitive inhibition seen with nondepolarizing NMBAs in combination with low concentrations of acetylcholine or DMPP.

Surprisingly, three of the seven nondepolarizing NMBAs tested displayed a decreased IC50value, and the others did not increase the IC50value with increased acetylcholine concentration (1 to 10 μm), a finding that indicates a noncompetitive component of inhibition.

We also found that repeated applications of 10 μm acetylcholine desensitized the receptor, indicating that the decrease in IC50value could at least partly be explained by receptor desensitization. To further probe this hypothesis, we tested the inhibitory mechanism of rocuronium against concentrations of acetylcholine below and above the EC50. We found that the rocuronium inhibition was competitive at low acetylcholine concentrations (0.3–1 μm) when receptor desensitization is limited, and it only became noncompetitive at moderate concentrations (greater than 1 μm). This suggests that the noncompetitive inhibition of the human adult muscle nAChR by rocuronium is dependent on accumulation of desensitized receptor. Coapplication of rocuronium in contrast to preapplication did not increase the affinity, strongly arguing against open-channel block as a mechanism of inhibition.

A noncompetitive mechanism of inhibition for d-tubocurarine and pancuronium at mouse adult muscle nAChR has been reported previously,11but it has not been investigated in more detail thereafter. To our knowledge, the data presented here on the other nondepolarizing NMBAs have not been demonstrated before. Much of the early work on interactions between nondepolarizing NMBAs and muscle nAChRs were made at the fetal muscle nAChR and/or using radioligand binding techniques,14,18thus limited to binding and with no ability to distinguish between different functional receptor states (i.e. , active, closed, desensitized). Recently, a preopening state (flipping) has been identified in single-channel recordings at muscle nAChRs using partial agonists,19demonstrating the importance of functional studies. Also, we speculate that this flipping state might be involved in the reduction of IC50values with increased acetylcholine concentrations. In this voltage-clamp setup, we measure currents upon application of agonist and/or antagonist in real time; we thus investigate the functional affinity of the NMBAs. Furthermore, the fetal and adult muscle nAChRs have divergent biophysical and pharmacological properties7,8that will have implications on the distinct pharmacological profiles as argued above.

Using the nicotinic agonist DMPP devoid of desensitization and open-channel blocking properties,20we demonstrated that atracurium and rocuronium inhibited DMPP-induced currents by a different mechanism than the noncompetitive inhibition displayed against acetylcholine. In fact, increasing concentrations of DMPP yielded increasing IC50values, suggesting competitive inhibition. In this isolated receptor model, some of the nondepolarizing NMBAs tested inhibit the acetylcholine current in human adult muscle nAChR by a mechanism other than competitive inhibition. We speculate that acetylcholine induced desensitization of the receptor in combination with NMBA is a possible mechanism for this inhibition.

Interestingly, repeated acetylcholine concentration-response curves yielded the same EC50, although the peak response declined. This suggests that a population of receptors has been desensitized, and the remaining receptor population available for activation has the same properties because the pharmacology (i.e. , EC50) was unchanged.

Although we have studied human adult muscle nAChRs, they are expressed in a Xenopus  oocyte expression model, and to what extent these data apply for the intact human neuromuscular junction remains to be elucidated. However, we speculate that one possible mechanism behind the increased block seen with high and repeated doses of neostigmine in patients might be the result of an elevation of acetylcholine in the synaptic cleft and a desensitization of the postsynaptic receptor, resulting in an increased degree of blockade. It is not possible to determine if the degree of desensitization seen in this study by 10 μm of acetylcholine adequately reflects the postsynaptic receptor desensitization in the intact neuromuscular junction. The acetylcholine concentration reached at the receptor in vivo  is very difficult to measure because of rapid degradation by acetylcholine esterase, but it is likely to locally achieve much higher concentrations than used in this study, inducing a substantial degree of receptor desensitization.

In this study, the affinity range at 1 μm acetylcholine was mivacurium > pancuronium > d-tubocurarine = vecuronium > cis-atracrium > rocuronium > atracurium. The relations between pancuronium and d-tubocurarine were the same as in the previous mouse receptor study by Garland et al. .11However, comparing with the most extensive study so far on mouse adult muscle nAChR, mivacurium and d-tubocurarine displayed a higher affinity in the human adult muscle nAChR, whereas rocuronium had a lower affinity.10However, cis-atracurium was less potent at the isolated human receptor, and d-tubocurarine more potent than expected from the corrected ED95dose. This probably reflects that the neuromuscular block seen in the clinic is not only dependent on interaction with the postsynaptic muscle nAChR, but also with other structures in the neuromuscular junction as well as protein binding, distribution, and elimination of the nondepolarizing NMBA.

In summary, we demonstrate that nondepolarizing NMBAs inhibit human adult muscle nAChRs expressed in Xenopus  oocytes by mixed mechanisms dependent on the receptor activation mode. When using the nondesensitizing agonist, DMPP inhibition by the NMBA is competitive, whereas activation with high concentrations of acetylcholine induces a noncompetitive inhibition by the NMBA. We speculate that this observation can involve receptor desensitization by the NMBA in combination with high concentrations of acetylcholine similar to that observed in the neuromuscular junction.

The authors thank GlaxoSmithKline (Barnard Castle Durham, United Kingdom) for kindly providing atracurium and cis-atracurium, and Organon, a part of Schering-Plough (Roseland, New Jersey) for providing pancuronium, rocuronium, and vecuronium. The authors also thank AstraZeneca Pharmaceuticals (Wilmington, Delaware) for providing messenger RNA for the nicotinic acetylcholine receptor subunits.


Jonsson M, Gurley D, Dabrowski M, Larsson O, Johnson EC, Eriksson LI: Distinct pharmacological properties of neuromuscular blocking agents on human neuronal nicotinic acetylcholine receptors: A possible mechanism behind the train-of-four fade. Anesthesiology 2006; 105:521–33
Bowman WC: Neuromuscular block. Br J Pharmacol 2006; 147(Suppl 1):S277–86
Karlin A: Emerging structure of the nicotinic acetylcholine receptors. Nat Rev Neurosci 2002; 3:102–14
Lindstrom JM: Nicotinic acetylcholine receptors of muscles and nerves: Comparison of their structures, functional roles, and vulnerability to pathology. Ann N Y Acad Sci 2003; 998:41–52
Tsuneki H, Kimura I, Dezaki K, Kimura M, Sala C, Fumagalli G: Immunohistochemical localization of neuronal nicotinic receptor subtypes at the pre- and postjunctional sites in mouse diaphragm muscle. Neurosci Lett 1995; 196:13–6
Tsuneki H, Salas R, Dani JA: Mouse muscle denervation increases expression of an alpha7 nicotinic receptor with unusual pharmacology. J Physiol 2003; 547:169–79
Mishina M, Takai T, Imoto K, Noda M, Takahashi T, Numa S, Methfessel C, Sakmann B: Molecular distinction between fetal and adult forms of muscle acetylcholine receptor. Nature 1986; 321:406–11
Teichert RW, Rivier J, Torres J, Dykert J, Miller C, Olivera BM: A uniquely selective inhibitor of the mammalian fetal neuromuscular nicotinic acetylcholine receptor. J Neurosci 2005; 25:732–6
Yost CS, Winegar BD: Potency of agonists and competitive antagonists on adult- and fetal-type nicotinic acetylcholine receptors. Cell Mol Neurobiol 1997; 17:35–50
Paul M, Kindler CH, Fokt RM, Dresser MJ, Dipp NC, Yost CS: The potency of new muscle relaxants on recombinant muscle-type acetylcholine receptors. Anesth Analg 2002; 94:597–603
Garland CM, Foreman RC, Chad JE, Holden-Dye L, Walker RJ: The actions of muscle relaxants at nicotinic acetylcholine receptor isoforms. Eur J Pharmacol 1998; 357:83–92
Purohit PG, Tate RJ, Pow E, Hill D, Connolly JG: The role of the amino acid residue at alpha1:189 in the binding of neuromuscular blocking agents to mouse and human muscle nicotinic acetylcholine receptors. Br J Pharmacol 2007; 150:920–31
Lowenick CV, Krampfl K, Schneck H, Kochs E, Bufler J: Open channel and competitive block of nicotinic receptors by pancuronium and atracurium. Eur J Pharmacol 2001; 413:31–5
Fletcher GH, Steinbach JH: Ability of nondepolarizing neuromuscular blocking drugs to act as partial agonists at fetal and adult mouse muscle nicotinic receptors. Mol Pharmacol 1996; 49:938–47
Jonsson M, Dabrowski M, Gurley DA, Larsson O, Johnson EC, Fredholm BB, Eriksson LI: Activation and inhibition of human muscular and neuronal nicotinic acetylcholine receptors by succinylcholine. Anesthesiology 2006; 104:724–33
Chavez-Noriega LE, Crona JH, Washburn MS, Urrutia A, Elliott KJ, Johnson EC: Pharmacological characterization of recombinant human neuronal nicotinic acetylcholine receptors h alpha 2 beta 2, h alpha 2 beta 4, h alpha 3 beta 2, h alpha 3 beta 4, h alpha 4 beta 2, h alpha 4 beta 4 and h alpha 7 expressed in Xenopus  oocytes. J Pharmacol Exp Ther 1997; 280:346–56
Hatton CJ, Shelley C, Brydson M, Beeson D, Colquhoun D: Properties of the human muscle nicotinic receptor, and of the slow-channel myasthenic syndrome mutant epsilonL221F, inferred from maximum likelihood fits. J Physiol 2003; 547:729–60
Sine SM, Taylor P: Relationship between reversible antagonist occupancy and the functional capacity of the acetylcholine receptor. J Biol Chem 1981; 256:6692–9
Lape R, Colquhoun D, Sivilotti LG: On the nature of partial agonism in the nicotinic receptor superfamily. Nature 2008; 454:722–7
Yost CS, Dodson BA: Inhibition of the nicotinic acetylcholine receptor by barbiturates and by procaine: Do they act at different sites? Cell Mol Neurobiol 1993; 13:159–72