Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

On the kinetics of NMDA-associated ion channels : different agonists and the effects of anaesthetic drugs Sawyer, Dale Christopher 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1992_fall_sawyer_dale.pdf [ 2.7MB ]
JSON: 831-1.0086580.json
JSON-LD: 831-1.0086580-ld.json
RDF/XML (Pretty): 831-1.0086580-rdf.xml
RDF/JSON: 831-1.0086580-rdf.json
Turtle: 831-1.0086580-turtle.txt
N-Triples: 831-1.0086580-rdf-ntriples.txt
Original Record: 831-1.0086580-source.json
Full Text

Full Text

ON THE KINETICS OF NMDA-ASSOCIATED ION CHANNELS: DIFFERENT AGONISTS AND THE EFFECTS OF ANAESTHETIC DRUGS by DALE SAWYER B.Sc. (Biochemistry) A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF PHARMACOLOGY & THERAPEUTICS We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA June 1992 (c) Dale Sawyer, 1992 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my writ ten permission. (Signature) Department of The University of British Columbia Vancouver, Canada Date 92/10/08 DE-6 (2/88) ii ABSTRACT The majority of studies to date dealing with NMDA-activated ion channels have used whole-cell preparations to study macroscopic currents. The whole-cell current response to applied agonists consists of several components including the number of single channels activated by the agonist, the magnitude of the current passed by each ion channel, and the mean open time and frequency of opening of the ion channels. In order to determine the physiological and pharmacological properties of these individual components, it is useful to record single-channel currents through the NMDA ion channel. In this work I have used the cell-attached configuration of the patch clamp technique to study the effects of (a) different agonists and their stereoisomers and (b) intermediate and long-chain alcohols and volatile general anaesthetic agents on the single channel properties of the NMDA channel. An important point in this work is that intact cells, where channel gating from intracellular factors were preserved, were studied whereas the majority of single channel data involving NMDA channels has been obtained using outside-out patches, with little or no attempt to correlate such measurements with ion channel behaviour in the intact cell. I have studied the differences in mean open time and opening frequency (and thus probability of the channel being in the open state at a particular time) produced by different agonists of the NMDA channel; as well, the effect of agonist concentration on the frequency of channel opening was investigated. Cultured rat hippocampal CAl neurones were used in order to assess the single-channel kinetics of the stereoisomers of two different NMDA agonists, N-methylaspartic acid and homocysteic acid, as determined by on-cell patch-clamp recording with the agonists included in the pipette. In the absence of external Mg2+ the mean open time of NMDA channels, with all agonists, were diminished in an exponential manner with increasing patch hyperpolarization. No significant differences in conductance were observed between any of the agonists. However, significant differences in mean open time were found between the enantiomers NMDA and NMLA and between the enantiomers D- and L-homocysteic acid. Increasing the concentrations of the agonists in the patch pipette significantly increased the frequency of channel openings. These data on agonist-induced unitary currents can be correlated with whole-cell electrophysiological studies and autoradiographical binding affinity studies. Further studies were undertaken in order to assess the effects of several anaesthetic drugs, including the normal aliphatic alcohols butanol, pentanol, and octanol, as well as the volatile anaesthetics halothane and isoflurane, on the kinetics of unitary NMDA agonist-activated currents as recorded in the cell-attached configuration. While the channel conductance was not iv affected, mean open times were diminished at all patch potentials by the drugs; other effects, such as reduced frequency of channel opening, were also observed. Additional data, using the fluorescence probe fura-2 to illustrate the depressant effect of the n-alkanols on NMDA-induced increases in intracellular calcium, are provided in support of the results from the electrophysiological experiments. The information obtained in this study provides an explanation for such previous findings as the inhibition of whole-cell NMDA currents by n-alkanols and the attenuation of glutamate- and NMDA-stimulated intracellular calcium signals by volatile anaesthetics. TABLE OF CONTENTS V Abstract ii Table of contents v List of tables x List of figures xi 1. AGONIST STUDIES 1.1. INTRODUCTION 1 1.1.1. The NMDA receptor 1 Early studies 1 Subdivision of the glutamate receptor 2 Further subclassification of glutamate receptors 4 Functional role of the NMDA receptor 5 1.1.2. Relevant studies of NMDA ion-channel behaviour 8 Single-channel recording of NMDA ion-channel behaviour 8 Kinetic differences between agonists 10 1.2. METHODS 13 1.2.1. Tissue culture 13 1.2.2. Patch-clamp recordings 14 Pipette preparation 14 The recording bath 15 Pipette mounting 16 Seal formation 17 Options: patch excision and whole-cell mode 18 Recording at different patch potentials 19 Data acquisition and storage 20 Data analysis 20 RESULTS 23 1.3.1. Single-channel currents 23 1.3.2. Current-voltage relation 2 6 1.3.3. Inside-out patches 29 Patch excision 29 Current-voltage relation 31 1.3.4. Agonist effects on channel kinetics 31 Data from individual cells 31 Limitations in data collection: spontaneous excision and noise 34 Amplitude histograms 35 Open duration histograms 35 1.3.5. Mean open times with different agonists 38 On-cell recordings 38 Channel kinetics in the inside-out mode 41 1.3.6. Closed duration analysis 47 1.3.7. Frequency dependence on agonist concentration 51 1.3.8. Open state probability 54 1.4. DISCUSSION 57 1.4.1. Channel conductance 57 1.4.2. Channel kinetics 58 Modulation with patch potential 58 Kinetics in cell-free patches 61 1.4.3. Agonist-dependent channel kinetics 63 Mean open time 63 Shut time kinetics 65 Concentration effects 67 2. INTERACTION OF ALCOHOLS 2.1. INTRODUCTION 70 2.1.1. Alcohol 70 Prevalence of use 70 Consequences of mass alcohol consumption 71 Pioneering research 72 Current research 74 Ethanol and NMDA activity 75 2.1.2. Volatile anaesthetics 76 Historical background 76 Theories concerning anaesthetic action 80 2.2. METHODS 82 2.2.1. Cell preparation and patch-clamp recordings 82 2.2.2. n-alkanol and anaesthetic perfusion 82 2.2.3. Current recording during drug application 83 2.2.4. Fura-2 measurements. 85 Fura-2 loading 85 Fluorescence measurements 8 6 Calculation of [Ca2+]i 88 Determination of n-alkanol effects on NMDA-induced [Ca2+]i 89 RESULTS 90 2.3.1. The alcohols 90 J7-Alkanol effects on unitary currents 90 Current-voltage relation 93 Distributions 93 Reduction of mean open time by n-alkanols 98 Alcohol and open state probability 101 2.3.2. Spectrof luorimetric analysis of Ji-alkanol effects on NMDA-induced changes in [Ca2+]i 107 2.3.3. Volatile anaesthetics 108 Channel open time 108 Open state probability 113 Open and closed duration histograms 113 DISCUSSION 115 2.4.1. Effects of the n-alkanols 115 Channel conductance 115 Reduction in mean open time 116 Effects on other kinetic parameters 117 Depression of NMDA-induced [Ca2+]i increases 118 Proposed mechanism of action on NMDA currents 119 Functional consequences 122 2.4.2. Volatile anaesthetics 123 Influence on NMDA channel kinetics 123 Functional relevance 125 SUMMARY AND CONCLUSIONS 127 REFERENCES 130 X LIST OF TABLES Table I. Channel conductance and zero-current potential 30 Table II. Effects of the n-alkanols on mean channel open time and [Ca2+]i increases 99 Table III. Effects of the volatile anaesthetics on mean open time and open proability 109 XI LIST OP FIGURES Figure 1. Unitary currents 24 Figure 2. I-V plots 27 Figure 3. I-V plot from excised patches 32 Figure 4. Amplitude histograms 36 Figure 5. Open duration histograms 39 Figure 6. Differences in mean open time between agonists 42 Figure 7. Kinetics of inside-out patches 45 Figure 8. Closed duration histograms 48 Figure 9. Dependence of NMDA channel opening rate on agonist concentration 52 Figure 10. Open state probability 55 Figure 11. Unitary currents in control and in the presence of alcohol 91 Figure 12. I-V curves in control and alcohol solutions 94 Figure 13. Open and closed duration histograms 96 Figure 14. Relationship between patch potential and mean open time 102 Figure 15. Spectrofluorimetric analysis 105 Figure 16. Distributions of open and closed times 111 1 1.1. INTRODUCTION 1.1.1. THE NMDA RECEPTOR Early studies The NMDA receptor has become the most well-studied subtype of the glutamate receptor, and investigations into its nature and behaviour have represented a significant portion of neurophysiological and biopsychological research for well over ten years. The discovery of the excitatory action of L-glutamate on central vertebrate neurones by Hayashi in 1954 led to Curtis and Watkins' classic experiments in the late 1950's and early 1960's in which the depolarizing actions of glutamate and several related substances on various mammalian central neurones were studied and categorized (Curtis and Watkins, 1963). At this point the investigators had some difficulty with the concept of L-glutamate, or L-aspartate, another endogenous dicarboxylic amino acid with similar excitant properties, participating in central neurotransmission, since a seemingly non-specific depolarizing effect occurred in every spinal cord neurone to which they applied the substances. However, with the later development of improved methodologies and drugs for studying the responses of neurones to applied substances, it became clear that 2 glutamate's effects were indeed being exerted through membrane-associated receptors and selective ion channels (Watkins, 1989). The stereospecificity of the neuronal responses in those early studies, as can be seen when the results with D- and L-glutamate and D- and L-homocysteic acid are compared, is in retrospect the earliest evidence of a receptor-mediated excitatory action (Curtis and Watkins, 1963). The strict structural reguirements for excitation among the many other compounds tested also suggested the potential role that substances of this type would have in receptor-mediated excitatory neurotransmission. One of the substances tested in these early studies was N-methyl-D-aspartate; this synthetic compound was discovered to be the most potent excitatory agent out of 30 other glutamate-like substances examined (Curtis and Watkins, 1963). Subdivision of the glutamate receptor The eventual development of antagonists that could specifically block the excitatory actions of certain substances with glutamate-like actions, such as NMDA, while leaving the actions of other neuroexcitants unaffected led to the establishment, by the turn of the 1980's, of glutamate receptor subtypes (Watkins and Evans, 1981; Collingridge and Lester, 1989; Olverman and Watkins, 1989). The first distinction to be made was that between NMDA and non-NMDA receptors: neurones which responded to both L-glutamate and NMDA had their responses blocked by substances such as D-a-aminoadipic acid (DAA) and 2-amino-5-3 phosphonovalerate (APV) while neurones that would only respond to L-glutamate, and not NMDA, were unaffected by DAA and APV and instead were more sensitive to blockade by other agents, such as glutamic acid diethylester (GDEE), that did not affect NMDA activity (Mayer and Westbrook, 1987b). As well, aspects of the conductance properties of neuronal membrane currents elicited by substances acting specifically on either of the two types of receptors were found to be fundamentally different: while membrane depolarization was always observed upon application of any glutamate-like substance to feline motoneurones (Curtis and Watkins, 1963), those depolarizations evoked by such agents as NMDA and D,L-homocysteic acid were accompanied by either no change or a voltage-dependent decrease in membrane conductance, producing non-linear current-voltage relations (Mayer and Westbrook, 1987a). The explanation for this phenomenon was first advanced in 1984 upon the discovery of the voltage-dependent open-channel block of NMDA-associated ion channels by magnesium and the observation of linear current-voltage relations for currents flowing through NMDA-specific ion channels in the absence of magnesium (Nowak et al, 1984; Mayer et al, 1984; MacDonald and Wojtowicz, 1982). Activation of non-NMDA receptors, on the other hand, always yielded linear I-V relations and consistent conductance increases in both the presence and absence of Mg2+ (Mayer and Westbrook, 1987a,b). 4 Further subclassification of glutamate receptors The subdivision of the glutamate receptor has continued through the 1980's (Watkins et al, 1990). Non-NMDA receptors have been further subdivided, mostly on the basis of regional variations within the mammalian CNS in the relative sensitivity to different non-NMDA agonists, especially when responses to the non-NMDA agonists kainate and guisqualic acid are compared (Mayer and Westbrook, 1987b; Collingridge and Lester, 1989). As well, differences between the non-NMDA agonists in their sensitivity to non-NMDA antagonists were observed: for example, responses of neurones to guisqualic acid were discovered to be potently blocked by GDEE, which left responses to another non-NMDA agonist kainic acid relatively unaffected (Mayer and Westbrook, 1987b; McLennan, 1988). Observations such as these led to the original subclassification of non-NMDA receptors into quisqualate-preferring and kainate-preferring receptors (Watkins and Evans, 1981; Ascher and Nowak, 1988a). In recent reports, the nomenclature has been adjusted so that the quisqualate receptor has been renamed in order to refer to the agonist so far observed to be the most specific for its subtype, 3-hydroxy-5-methylisoxazole-4-(2-amino)-propionic acid, or AMPA (Watkins et al, 1990). Attempts to rename the kainate receptor for other compounds that are believed to be more specific for its excitation, most notably domoic acid, have not been met with unanimous approval (Watkins et al, 1990). Two other glutamate 5 receptor subtypes have been even more recently identified. One, whose activation results in such second messenger phenomena as increased intracellular inositol 1,4,5-triphosphate (IP3) production and mobilization of intracellular calcium (Collingridge and Lester, 1989), is not associated with direct ion channel activation and thus has been termed the metabotropic glutamate receptor (Watkins et al, 1990). The other subclassification stemmed from observations of the depressant effect of the NMDA receptor agonist L-2-amino-4-phosphonobutyric acid (APB, or L-AP4) on certain presumed glutamatergic synapses (Watkins et al, 1990). The inability of this agent to antagonize the action of coapplied excitatory amino acid agonists, along with its lack of effect on miniature excitatory postsynaptic potential amplitudes suggests a presynaptic site of action for L-AP4 (Monaghan et al, 1989). Only a few compounds, such L-serine-O-phosphate and glutamate itself, can mimic the effects of L-AP4, with no known antagonists, and this has led investigators to postulate the existence of a presynaptic inhibitory autoreceptor influencing the chain of events leading to excitatory amino acid release; this receptor has been termed the L-AP4 receptor (Watkins et al, 1990; Monaghan et al, 1989). Functional role of the NMDA receptor While the non-NMDA receptors are believed to play a dominant role in fast excitatory neurotransmission in the CNS (Headley and Grillner, 1990; Collingridge and Lester. 6 1989; Mayer and Westbrook, 1987b), the function(s) of NMDA receptor activation appear to be far more complex and studies into this area are expected to answer at least as many questions about the basic nature of the psychophysiological process itself as about the specific role the NMDA receptor plays in CNS functioning. For example, while monosynaptic afferently-evoked EPSP's in the mammalian spinal cord are blocked by non-NMDA excitatory amino acid receptor antagonists (Headley and Grillner, 1990; Lambert et al, 1989), responses most likely involving poly-synaptic pathway are sensitive to NMDA antagonists such as APV (Collingridge and Lester, 1989; Davies and Stanley, 1988). NMDA receptors are believed to be co-participatory with non-NMDA receptors in mediating, and modulating, fast excitatory transmission (Eccles, 1989; Watkins, 1989; Dale, 1989; Lambert et al, 1989) and this aspect of NMDA function may indicate some role in the maintenance of consciousness and arousal. The fact that drugs such as the dissociative anaesthetic ketamine (MacDonald et al, 1989; MacDonald et al, 1987) and a-opiate receptor ligands such as phencyclidine (PCP) (MacDonald et al 1989; Lodge, 1987) exert profound effects on consciousness while at the same time having specific actions to depress NMDA ion channel activity via specific binding sites within the channel pore (Watkins, 1989; MacDonald and Nowak, 1990; Collingridge and Lester, 1989) supports the concept of a significant NMDA-mediated component of consciousness. NMDA receptors in the 7 spinal cord may participate in the perception of pain at the spinal level (Dickenson, 1990; Raigorodsky and Urea, 1990). It is also in the spinal cord that NMDA receptors may exercise influence over locomotion and other forms of movement control by organizing and coordinating complex motor patterns (Dale, 1989). The contribution of the NMDA receptor subtype to the process of long-term potentiation (LTP) has been we 11-documented (Sastry et al, 1990; Errington et al, 1983; Collingridge et al, 1983) and the NMDA receptor may be involved in such LTP-associated processes as learning and memory (Izquierdo, 1991; Morris et al, 1989; Morris, 1988). NMDA receptors have also been implicated in synaptic plasticity and in the growth of developing synapses (Eccles, 1989; Dale, 1989). The self-enhancing nature of NMDA receptor activation, which involves uncovering of NMDA-mediated currents after the voltage-dependent magnesium block has been relieved by depolarization (Ascher and Johnson, 1989; Lambert et al, 1989; Johnson and Ascher, 1988a), has implicated NMDA ion channels in the self-sustaining and chaotic (van Rossum and de Bie, 1991) processes of seizure generation and epileptogenesis (Dingledine et al, 1990) and in synaptic transmission following kindling (Lambert et al, 1989). The neurotoxic effects of passage of high levels of calcium (Ascher and Nowak, 1986; MacDermott et al, 1986; Nowak and Ascher, 1985) into neurones via NMDA receptors (Choi et al, 1988; Choi, 1987) has led to the study of the NMDA system 8 with respect to its role in neurodegenerative processes such as Parkinson's disease (Schmidt et al, 1990; Choi, 1988) and neuronal cell death associated with hypoxia and stroke (Collingridge and Lester, 1989; Choi, 1988). Therefore, drugs that block NMDA receptor-ion channel activity, either inhibiting through competitive or non-competitive means such as ion-channel blockade or enhancing by direct agonism or an allosteric activation, are currently being vigourously investigated in terms of potentially providing useful clinical weapons against such diseases as epilepsy and Parkinson's disease. 1.1.2. RELEVANT STUDIES OF NMDA ION-CHANNEL BEHAVIOUR Single-channel recording of NMDA ion-channel behaviour In order to develop drugs or other strategies that target disease states or drug reactions in which NMDA receptors are involved, as well as to better understand the psychophysiological processes mediated in some way by NMDA receptors, the interaction of NMDA agonists with the receptor, and the relationship between this interaction and the currents produced upon NMDA receptor activation, must be investigated. Electrophysiological studies have been central to NMDA research, and much useful information has been gathered since the early 1980's with such techniques. NMDA agonists preferentially activate a large (40-50 pS) 9 conductance state (Cull-Candy and Usowicz, 1989; Cull-Candy and Usowicz, 1987; Nowak et al, 1984). The relatively large size of NMDA-type currents makes the NMDA receptor-ion channel complex amenable to study with patch-clamp single-channel recording (Gibb and Colquhoun, 1991; McLarnon and Curry, 1990a,b; Ascher et al, 1988), which allows observation of microscopic currents through single ion channels in cell membranes (Hamill et al, 1981). The initial observation of the open-channel blocking effect of magnesium, which greatly reduces the apparent open time of single NMDA-activated ion channels in a voltage-dependent manner (Johnson and Ascher, 1988; Mayer et al, 1984; Nowak et al, 1984), led to the realization that NMDA ion channel kinetics are best studied in Mg -free solutions in order to elucidate the magnesium-independent gating behaviour of the channel species. Initial single-channel studies, using cell-free membrane patches, indicated that removal of Mg2+ from solutions bathing either side of the membrane also removed the voltage-dependent kinetics of the behaviour of the ion-channel (Mayer and Westbrook, 1987a; Ascher et al, 1988; Ascher and Nowak, 1988b). However, in the absence of Mg2+, records from cell-attached patches showed the mean open time of the channel was strongly modulated by potential. These results suggested that unknown intracellular components were modifying channel kinetics (McLarnon and Curry, 1990a,b). The conductance and kinetic mechanisms of the NMDA-receptor ion channel have been observed to be dependent on a variety of extracellular constituents (MacOonald et al, 1989; Reynolds and Miller, 1988). These include glycine (Oliver et al, 1990; Ascher and Johnson, 1989; Johnson and Ascher, 1988; Forsythe et al, 1988; Reynolds et al, 1987; Johnson and Ascher, 1987), zinc (Forsythe et al, 1988), magnesium (Nowak et al, 1984), pH (Traynelis and Cull-Candy, 1990), and the total receptor redox state, especially with respect to disulphide bonds accessible through the membrane exterior (Wright and Nowak, 1990; Reynolds et al, 1990). The recent demonstration of the modulatory effects of intracellular polyamines (Ransom and Deschenes, 1990) and phosphorylation of receptor amino acid residues (MacDonald et al, 1989) indicates that the state of the intracellular milieu guite likely affects the manner in which the NMDA ion-channel behaves in situ. The discrepancy in the kinetics of the unitary currents observed between cell-attached and cell-free patches suggests that intracellular substituents can exert some control over the NMDA ion channel gating mechanism (McLarnon and Curry, 1990a,b). Therefore caution must be used in interpreting data obtained from excised patches, as opposed to intact recording configurations, in studying the kinetics of the NMDA-gated ion channel. In this work we have used single-channel recordings from cell-attached patches, with nominally Mg2+-free solutions in the pipette and extracellular bathing solutions, over a range of patch potentials in order to study the voltage-dependent kinetics of intact NMDA ion channels. Kinetic differences between agonists Numerous studies have demonstrated the ability of a variety of substances to activate the NMDA receptor subtype, including NMDA itself, L-glutamate, L-aspartate (Watkins and Evans, 1981), the Amanita mushroom constituent ibotenic acid (Collingridge and Lester, 1989), the synthetic compound D-cis-l-amino-l,3-cyclopentanedicarboxylic acid (ACPD) (McLarnon and Curry, 1990a), and the endogenous tryptophan metabolite quinolinic acid (McLarnon and Curry, 1990b; Peet et al, 1986; Perkins and Stone, 1983). It was found that the single-channel currents produced by NMDA, ACPD, and guinolinate differed in their mean open time as well as in the concentrations required to produce similar rates of channel openings when recorded from cell-attached patches (McLarnon and Curry, 1990a,b). We have studied more closely the difference in activation kinetics observed when agonists of different structure are used to activate the NMDA ion channel. For agonists, we have used N-methylaspartate, as this agent is still considered to be the most specific for the NMDA receptor subtype (Olverman and Watkins, 1989; Davies and Stanley, 1988; Curtis and Watkins, 1963), and the endogenously-occurring excitatory amino acid homocysteic acid. L-homocysteic acid has excitatory actions on the NMDA receptor (Watkins, 1990; Davies and Stanley, 1988; Baudry et al, 1983; Curtis and Watkins, 1962), and its presence and 12 Ca -dependent release upon both chemical and electrical stimulation of brain slices have been demonstrated using High-Performance Liquid Chromatography (HPLC) (Cuenod et al, 1990; Do et al, 1988; Do et al, 1987; Kilpatrick and Mozley, 1987; Cuenod et al, 1986). The presence of enzymes capable of manufacturing L-homocysteic acid from its precursor methionine in striatal slices (Do et al, 1988; Do et al, 1987) and the presence of specific uptake systems for L-homocysteate (Zeise et al, 1988), have led to the suggestions that this excitatory amino acid could serve as an endogenous ligand for the NMDA receptor subtype. As well as studying L-homocysteic acid and NMDA, the two enantiomers of these compounds were also investigated. Comparisons were made between the pairs of enantiomers in order to demonstrate that chiral orientation, in addition to the basic ligand structure, affects not just the binding affinity of the ligand in question but indeed influences the gating behaviour of the ion channel. The gating kinetics were studied at different pipette concentrations for each agonist in order to provide additional information on the relationship between receptor activation and single-channel behaviour. 13 1.2. METHODS 1.2.1. TISSUE CULTURE Wistar rat hippocampal cultures from embryos in the 18th after conception were prepared according to the method of Banker and Cowan (1977). Briefly, hippocampi of day-18 fetal Wistar rats were dissected out, dissociated into single cells by trituration through Pasteur pipettes, and suspended in HBSS for counting in a hemocytometer. The HBSS consisted of Hank's Balanced Salt Solution (Gibco) modified with 4.5 mM sodium bicarbonate (Fisher Scientific Co.), 27.8 mM glucose, 15 mM HEPES (both from the Sigma Chemical Co.), 100 U/mL penicillin, 100 jLtg/mL streptomycin (both from Gibco), all adjusted to pH 7.14-7.18 with NaOH. Once counted, the cell suspension was diluted to 105 cells/ml with Dulbecco's Modified Eagle's Medium (Gibco), which was in turn modified to contain 45 mM NaHC03, 8.33 mM glucose, 10 mM HEPES buffer, 100 U/mL penicillin, 100 jug/mL streptomycin, 20% horse serum (Gibco), and, if prevention of yeast contamination was necessary, fungizone from Gibco (if added, this would result in 135 nM amphotericin B and 247 nM sodium desoxycholate in the DMEM) and plated onto poly-D-lysine/laminin-coated 18 mm circle coverslips. These were then inserted into 6-well plates and incubated in 5% CO2 at 35.5-36°C until use. Cells survived under these conditions 14 for up to 8 weeks; however, only cultures that were less than days following isolation were used, as experience proved that this age range was optimal for achieving patch seals under the experimental conditions outlined below. As well, NMDA responses were not observed in cells from cultures that were 7 days or less post-isolation; therefore, only cells that had been in culture between 8 and 16 days were used. 1.2.2. PATCH-CLAMP RECORDINGS The details concerning the theory and practice of single-channel patch-clamp recording can be found in Hamill et al, 1981. Pipette preparation The cell-attached patch clamp configuration, was used to record unitary NMDA currents since preliminary studies showed a voltage dependence of mean open time that was not seen with outside-out or inside-out patches (McLarnon and Curry, 1990a,b). The pipette glass, (A-M Systems, Washington), was made from Corning #7052 glass without microfilaments and had an outer diameter of either 1.2 mm, (0.68 mm inner diameter), or 1.65 mm (inner diameter of 1.2 mm) . The tip diameters of the pipettes were approximately 1-2 fim corresponding to a pipette resistance of 4-8 Mn. The pipette were prepared with a standard two-pull technique using a Narishige PP-83 glass microelectrode puller. The 15 pipettes were then fire-polished on a microforge consisting of a glass-coated heating filament, whose temperature and heating area were controlled by a jet of air, positioned on the stage of an ordinary laboratory ocular microscope. The pipettes were then filled with the agonist-containing solution, which consisted of bath solution (Section, below) with the addition of 1 juM glycine (Sigma Chemical Co.) and agonist. The agonists used were 25 and 100 /xM L-homocysteic acid, 50 and 100 /zM D-homocysteic acid, 50 and 200 JUM N-methyl-L-aspartic acid (NMLA) , and 30 fill NMDA, all from Sigma. The choice of concentrations was based on preliminary studies so that event frequencies would generally be in the range of 5-10 Hz. Previous to filling, all pipette solutions were filtered twice using Millipore Type GS circle filters with a 0.22 JUM pore size. The recording bath The bath solution consisted of (in mM) : 140 NaCl, 5 CsCl, 1.8 CaCl2, 10 HEPES, and 1 TTX; the pH was adjusted to 7.20 with NaOH. Cesium was used in place of potassium in order to minimize the effects of potassium currents. No magnesium was added to any of the solutions, as this divalent cation has profound effects on the kinetics of the NMDA channel (Jahr and Stevens, 1990; Mayer et al, 1984; Nowak et al, 1984), but the possibility of contaminating Mg being present in the solutions necessitated an analysis of these solutions for residual magnesium. The residual Mg2+ level in the bath solution, as determined by flame 16 analysis on a Jarrell-Ash 280 atomic absorption spectrometer, was found to be 0.4±0.2 /xM. The NaCl, CsCl, and CaCl2 were from BDH Chemicals, Ltd., the NaOH from the Fisher Scientific Co., and the agonists, glycine, TTX, and HEPES were all purchased from the Sigma Chemical Co. For the patch-clamp recordings, a coverslip with the attached hippocampal cells was removed from its 6-well plate and placed in the Perspex circular recording chamber. The chamber was filled with bath solution and was affixed to the stage of a Nikon TMS phase-contrast microscope; this system allowed viable cells to be easily distinguished from dead cells (by the living cells1 characteristic "glow") and enabled clear visualization, at 300x magnification, of the pipette tip during the patch formation procedure. The bath was rinsed three times with bath solution after the cover slip was inserted so that the culture medium was completely exchanged. An agar-filled bridge was submerged in the bath and a wire connected to ground was inserted into the bridge so that the recording circuit was complete. Experiments were performed at room temperature (22-25°C) and ambient atmospheric pressure. Pipette mounting The pipettes were loaded into the pipette holder which inserts into the current-voltage converter (Axon model CV-3 with 1/100 gain) for the Axopatch. These were mounted onto an Optikon lab jack for manipulation of the pipette position; the microscope and lab jack were situated on a 17 Newport Stable Top nitrogen-suspended floating table to enable the cells in the bath to be in a vibration-free environment. After mounting into the pipette holder a slight positive pressure was applied to the pipette to prevent buildup of debris around the tip area; the pipette was then placed into the bath solution and manoeuvred into position just above the cell chosen for patching. Only smooth, healthy-appearing cells with a bipolar shape and an average length of approximately 10-12 jum along the long axis were chosen for recording. Seal formation The pipette was slowly lowered into physical contact with the soma of the cell using a Newport MPH-1 microdrive with an 860 series motorizer and velocity controller attached to the lab jack. Pipette contact with the cell was viewed through the phase-contrast microscope and was also monitored electronically (using a 0.2 mV test pulse to test for seal resistance); as well, an audio signal was applied during the lowering procedure for further monitoring. Unitary currents were recorded with the Axopatch IB patch clamp amplifier (Axon Instruments, Inc.), and the Axopatch output signal was visualized on a Kikusui 5020A digital storage oscilloscope. Once the pipette was touching the cell, a gigohm (>10 Gil) seal was made by applying negative pressure through a mouth suction tube which was hooked up to the pipette holder. The large increase in pipette tip resistance associated with seal formation was monitored by 18 the loss of the 0.2 mV test pulse on the baseline noise and was confirmed by a change in the audio feedback signal characteristic of the formation of a tight seal. Attempts to form a gigohm seal were not always successful, and if only a partial seal was attainable the pipette was discarded. After formation of the seal channel activity in the patch could be observed as rectangular pulses on the oscilloscope screen. Options: patch excision and whole-cell mode Since the objective of these experiments was to record channel openings in the intact cell-attached configuration, usually no further manipulation of the pipette was performed; however, on occasion, the microdrive was used to pull the pipette upwards, excising the patch to the inside-out configuration. In this case, the solution facing the internal side of the membrane was identical to the external bath solution used in the cell-attached configuration. Recordings were made in this manner in order to compare the channel opening kinetics recorded in the cell-attached configuration and in the more widely studied excised-patch configuration. As well, the inside-out mode was used to help identify the channel under study by measuring the channel conductance in this configuration and checking for agreement with that measured in the cell-attached mode. Alternatively, instead of excising the patch, extra suction could be applied after obtaining the cell-attached configuration, producing a whole-cell patch. This was done 19 with 6 cells with the Axopatch was set in the "1=0" mode (the "V-clamp" mode was used for single-channel recordings) in order to measure the cell resting potential; the mean resting potential was determined to be -54 mV. One drawback of using the whole-cell patch configuration to measure cell resting potential is that dialysis of neurone with the pipette contents occurs during the procedure, introducing bias and variability into the measurements. Alternatively, I-V curves for currents recorded in the cell-attached mode for a given cell usually produced extrapolated single-channel current reversals at pipette potentials close to -60 mV. Previous studies have found the reversal potential for NMDA currents to be around -5 mV (Mayer and Westbrook, 1987a,b; Ascher et al, 1988; Ascher and Nowak, 1988b; Nowak et al, 1984). Thus, a cell resting potential (corresponding to 0 mV pipette potential) of -60 mV was assumed during the experiments and the results analysis. Recording at different patch potentials For cell-attached single-channel recordings, hyperpolarizing command potentials, corresponding to positive pipette potentials, were applied in steps of 20 mV for 10 second segments, with "rest" periods of 5 seconds with no applied potential between voltage commands, in order to record currents over a range of cell membrane voltages; depolarizing command potentials of -20 mV were also used. Several recordings were made at each pipette voltage, and the sequence of pipette potentials applied for recording 20 were constantly varied in order to eliminate any possible interactive effects with previously applied voltages. Patch potentials will hereby be referred to as VJJ, where VJJ = vrest ~ vcommand* Data acquisition and storage The patch-clamp data were recorded onto a Multitech MPF-PC/900 computer loaded with Microsoft MS-DOS 3.2 and analyzed offline using pCLAMP (version 5.0, Axon Instruments). The computer was interfaced with the Axopatch using a Labmaster data acquisition system from Axon Instruments and the computer output was monitored with a Datatrain DC-353S EGA monitor. For backup recording an Instrutech VR-10 digital data recorder was connected to a Panasonic Omnivision Hi-Fi MTS VHS (model PV-4760-K). Patch-clamp recordings were usually sampled at 5 kHz and filtered at 2 kHz (low-pass filter, Bessel configuration with -3 dB rolloff). In some cases when channel open times were short, the sampling rate was increased to 20 kHz and a low pass filter set at 5 kHz was used. When the frequency of channel openings was low, an AI 2020 event detector was also used so that computer recording commenced only once an event had occurred. Data analysis The data were analyzed using the pSTAT 5.00 subprogram, which generated the mean amplitudes, mean open times, and mean closed intervals. Single-channel events in the data records, which consisted of either 10- or 2.5-second 21 segments, were selected manually according to the criteria outlined below, and the amplitude, open time, and duration of the previous closed interval for each selected event were stored in a separate record which could then be used to generate histograms and means. The initiation of an event was defined as the point where the current trace crossed the amplitude level that was 50% of a predetermined threshold; when the trace crossed the same 50% threshold level on its return to the baseline, this was defined as termination of the event and therefore channel closure. Events selected for the idealized record were constrained to conform to minimum amplitude criteria; these criteria were unique for each pipette voltage and were established such that selecting events conforming to these criteria would produce monotonic, normally-distributed amplitude histograms. Finally, with the filter settings used and the seal resistances obtained in this study only openings of 400 j^s or greater were properly resolved; openings briefer than this were not selected for the idealized records. The minimum resolvable closed gap duration was set at 100 us. The amplitude, open time, and closed time histograms, such as those shown in the Results section, were plotted using a Hewlett Packard Color Pro graphics printer; in the Results section, the amplitude, open time, and closed times histograms shown for a particular agonist were all constructed from data produced by the same single cell. The open and closed time histograms were curve fitted by pSTAT, 22 which used a non-linear least-squares curve-fitting routine to fit 1- or 2- exponential functions by the Levenberg-Marquardt (LMQ) technique. Here, the data are weighted by the x2 value and, starting from seed values that are either manually entered or automatically determined by pSTAT, the fit proceeds until convergence, defined either when improvements in x2 are less than 2.5 x 10~7, or if all of the parameters converge when parameter change vectors go to zero. In order to determine if a particular histogram could be fitted with more than two exponentials (as in the case of the closed duration histograms), the two fast time components would first be fit to the left-most portion of the histogram; then, the curve fit would be performed again with the data range mow modified to include the portion of the curve fitted with the second component plus the data to be fit with a third component, and so on. 23 1 . 3 . RESULTS 1 . 3 . 1 . SINGLE-CHANNEL CURRENTS Examples of unitary currents activated by the four agonists studied are shown in Figure 1. The rectangular deflections from the baseline represent current flow through the ion-channels isolated in the cell-attached patches, and their downward orientation indicates flow of positive ions from the pipette medium into the cell. The traces shown were recorded with the higher agonist concentrations studied; decreasing the pipette agonist concentrations had no effect on the amplitude or mean duration of the channel openings although fewer openings per unit time were observed. The top traces, for each agonist, were recorded with the patch potential held at -40 mV. Each successive trace for a given agonist represents an increase of 20 mV in patch hyperpolarization. Thus the second trace for each of A - D was with the patch pipettes held at -60 mV so that the driving force for inward current was the cell resting potential (nominally -60 mV in these neurones). Increasing the driving force across the cell membrane increased the amplitude of the openings, and, as found in previous studies with cell-attached patches, decreased the mean open time 24 Figure 1. Unitary currents. Examples of single-channel currents recorded with 1 /xM glycine and either A) 100 /xM L-homocysteic acid, B) 100 ixM D-homocysteic acid, C) 200 /xM NMLA, or D) 30 juM NMDA in the pipette. TTX was included in both the pipette and the cell bathing medium in order to inhibit activation of voltage-dependent sodium channels; the solutions were "nominally Mg2+-free". The top trace for each agonist represents recordings taken at VM = -40 mV, while going down the traces increases the membrane hyperpolarization, and thus the driving force for the unitary currents, by 20 mV so that the bottom traces were recorded at VM = -120 mV. The channel openings are the rectangular deflections from the baseline; note the increase in amplitude with increasing VM. 25 L-HCA mV -40 V i ' ^ ^ J ^ S / I I^M* ( W%^ -60 WwyV/YS WM rVVMWvM f& •80 >M ^^t^ WV^ M LlA^W^ Vji •120 i p * W | f ^ ^ ^ A NMLA -60 /KWO^VM^/I -80 *\ -100 "wM^ • 1 2 0 V M H M '^r(frT<VH^V^ c D-HCA ty%W\ wNjI ^HAy^vfW flVW ,V>H vw B NMDA ,! 3 ms 26 of the events (McLarnon and Curry, 1990a,b), an effect also observed when the sampling rate was Increased to 20 kHz. A secondary type of opening with an amplitude about 2/3 of the primary conductance state was also evident (not shown). Such openings, which showed no obvious mean open time dependency on patch potential, were relatively infrequent in most patches. Other lower conductance states were also occasionally observed in patches. Currents other than those associated with the primary conductance were not analyzed further. 1.3.2. CURRENT-VOLTAGE RELATION Figure 2 shows the current-voltage plots for each of the agonists: D- and L-homocysteic acid in (a) and N-methyl-D- and L-aspartate in (b) . The straight lines are linear fits to the data and the slopes were not significantly different for any of the agonists. The channel conductances (see Table I) were estimated from the slopes of the fitted lines and were close to the conductance values previously reported for the primary conductance state of the NMDA channel (in the 40-50 pS range when outside-out patches were used) (Ascher et al, 1988; Cull-Candy and Usowicz, 1987; Mayer and Westbrook, 1987b). Table I also shows the zero-current potentials for each agonist, 27 Figure 2. I-V plots. Shown are the current-voltage relations for A) D- and L- homocysteic acid and B) NMDA and NMLA. Each point represents mean current amplitudes (I) taken at the corresponding transmembrane potential (VM) from 12-18 cells. The lines are linear fits to the data using linear regression analysis as performed by Slidewrite 4.1. Typical error bars are shown for one of the points in each graph. 28 l -V PLOT for all HCA cells Vm CmV) -160 -140 -120 -100 KpA) • L - H C A O D - H C A B Vm frnV) i -V PLOT for ail NMA cells KpA) -6 -7 • M^OA O NM_A 29 determined by extrapolating the I-V curve for the particular agonist to the voltage axis. These values can be considered an estimate of the mean resting potential of the cells, assuming a reversal potential of 0 mV for the NMDA channel which is equally permeant to the monovalent cations Na+, K+, and Cs+ (Mayer and Westbrook, 1987a; Ascher and Nowak, 1988b). 1 . 3 . 3 . INSIDE-OUT PATCHES Patch excision On occasion, after sufficient cell-attached data had been obtained, the patch was excised to the inside-out configuration. In some cases openings were no longer observed following the excision process. However, exposing the patch very briefly to air before reinserting the patch into the bath solution sometimes restored activity at both positive and negative pipette potentials so that inside-out data could be recorded. Presumably this procedure disrupted a uni- or bi-directional current block that had resulted from the formation of a vesicle during the excision process (Hamill et al, 1981). Current-voltage relation Figure 3 depicts current-voltage plots obtained from six inside-out patches recorded with either one of the studied agonists in the pipette. Note the reversal 3 0 Slope conductance fpS) Zero-current potential (mV) L-homocysteic acid D-homocysteic acid NMLA NMDA 17 15 16 7 42.8±0.8 42.6±0.7 42.0±0.7 41.0±1.1 -57.9±0.6 -56.7±0.8 -58.8±0.6 -58.8 ±1.0 Table I. Channel conductance and zero-current potential. Slope conductances were measured from the I-V relations for each agonist, pooled for concentration. Zero-current potentials were obtained from extrapolation of the I-V curves to the current axes, and are estimates of the mean cell resting potentials. 31 potentials were close to 0 mV (as expected for a channel non-selective between the monovalent cations Na+ and Cs+; see Mayer and Westbrook, 1985; Ascher and Nowak, 1988b) and the conductances were similar to those observed during cell-attached recording. 1.3.4. AGONIST EFFECTS ON CHANNEL KINETICS Data from individual cells The mean open time, and mean frequency of single channel openings were studied in 55 patches, for -140 mV < VM < ~ 4 0 m V where possible. There was a large variation in opening frequency between cells, even at a single concentration of a single agonist; which made reliable frequency comparisons between the groups somewhat difficult. This necessitated careful choice of high and low concentrations (determined during pilot experiments) for each agonist and a large sample size (>8) at each concentration; the criteria for adequate sample size in studying the differences in channel open time were easily satisfied by meeting the requirements for frequency analysis. Cells that showed instability or unusually high variance in frequency, amplitude, or mean open time were discarded from the analysis. 32 Figure 3. I-V plot from excised patches. Current (I)-pipette voltage (V) relation for NMDA-agonist induced openings. Curve is a linear fit to the data. While some error bars, which are ± S.E.M., are apparent, most are below the resolution of the graph. 33 I (PA) 5 -r -120 -100 -80 -60 -40 -20 20 40 60 80 100 120 V(mV) -J--3 -4 -L-5 34 Limitations in data collection: spontaneous excision and noise In most experiments channel activity was sustained over a period of time to obtain sufficient data over a range of pipette potentials. However, in some cases, a decline of channel activity with time, resulting in an eventually complete cessation of activity, was observed; this was often accompanied by an apparent movement of the cell away from the patch pipette (due to motion of the cover slip in the bath) and was presumed to reflect a partial, to full, excision of the patch. In other cases, extreme baseline noise, characterized by non-rectangular baseline deflections that did not conform to discrete current levels, would intrude upon the patch signal. This latter problem of noise in the patch would often appear if the integrity of the seal was somehow compromised (Hamill et al, 1981), as indicated by an increased test pulse amplitude. Applying voltages to the pipette, especially negative potentials that depolarized the patch and large positive (hyperpolarizing) potentials (VJJ < -140 mV), would occasionally result in a type of activity that, since it conformed to discrete amplitude levels, was judged to represent activation of small-conductance channels. If the opening frequency decreased markedly during the course of an experiment, the subsequent data was considered unreliable and was not used in the data analysis. 35 Amplitude histograms Histograms for amplitude, open time, and closed time were constructed for all cells at each voltage. Figure 4 shows typical amplitude histograms from cells recorded for each of the agonists; the Vj^  was -80 mV in all cases. The single peak distributions indicate little contribution from channel openings other than those under study. Channel opening amplitudes were not significantly different between agonists at a given voltage, as determined using one-way ANOVA. Open duration histograms Typical open duration histograms are shown in Figure 5, and were well-fit with single-exponential curves. The suitability of a single component was determined by comparing the goodness of fit when one or two components were used, The scheme which resulted in convergence of the iterating program, or which resulted in the lowest error in the time constant(s), was judged to represent the closest fit; the visual appearance of the fit qualities was almost always consistent with these criteria. Evidence of a second fast time component was detected in a few cells; since our data acquisition system was set up to detect openings of 0.4 ms or greater, a system enabling fine resolution of very short open times would be required in order to establish the existence of this very fast time component. Other studies dealing with NMDA channel kinetics have also found single-36 Figure 4. Amplitude histograms. Histograms of current amplitudes were constructed for each agonist and are shown in a typical cell for currents recorded at VM = -80 mV: A) 100 [M L-HCA, B) 100 fiM D-HCA, C) 200 MM NMLA, and D) 30 /xM NMDA. As with the channel conductances and the reversal potentials (see Table 1 and Figure 2) the mean current amplitudes were not significantly different for each agonist. Number of events o > 3 a c a (D z Number of events Number of events > 3 •a c a n> K o > Number of events CD > 3 K IH •a > r D • X o > •J 38 exponential open time curves (Jahr and Stevens, 1990; Ascher et al, 1988). Like the mean open time, the time constants of the fitted curves also were voltage dependent, that is, their values decreased with increased membrane hyperpolarization. To facilitate the general analysis, the arithmetic mean open time was assumed to reflect a valid marker for channel opening kinetics. In fact, the two measures were identical, differing only by the constant factor of system resolution time (Colquhoun and Sigworth, 1983); in the present case, time constants of the fitted single-exponential curves were consistently 0.4 ms less than the mean open times for a particular patch held at a given potential. As well, significant differences in mean open time were found between openings produced by the pairs of enantiomers, and this can be seen by comparing Figures 5a and 5b for L- and D-homocysteic acid and 5c and 5d for NMLA and NMDA. Single-exponential functions adequately fit most of the open time histograms, indicating a single open state of the NMDA channel in Mg2+-free solutions. 1.3.5. MEAN OPEN TIMES WITH DIFFERENT AGONISTS On-cell recordings In Figure 6, semilog plots of mean open time versus VJJ for each of the agonists are shown. The data points have been fitted with straight lines for each agonist; the slopes of these lines were not significantly different. These data 39 Figure 5. Open duration histograms. Histograms of channel open time were constructed for each agonist and are shown in a typical cell for currents recorded at VM = -80 mV: A) 100 /iM L-HCA, B) 100 MM D-HCA, C) 200 MM NMLA, and D) 30 MM NMDA. The histograms were adequately fit with single-exponential curves by a chi-squared curve-fitting routine. The time constants of the fitted curves were, in ms, A) 3.11±0.27, B) 1.85±0.07, C) 1.91±0.09, and D) 2.5310.11. Each histogram represents data comprising a minimum of 200 events. 4 0 L-HCA D-HCA c > 0) at J3 E 3 z 190 r1 Open duration (ms) Open duration (ms) NMLA c u > <0 at .a E 3 z NMDA Open duration (ms) Open duration (ms) 41 indicate that there is a significant difference in mean open time within the two pairs of enantiomers as well as between the voltages studied (two-way ANOVA; for homocysteic acid, P<0.001 for voltage and 0.002<P<0.005 between the D and L forms, while for NMA, P<0.001 both for voltage and between NMDA and NMLA). When post-hoc statistical analysis is applied to these results, significant differences between enantiomers are found only between open times recorded at -40 mV and -60 mV (0.001<P<0.005, Student Newman-Keuls test) and weakly at -80 mV (0.05<P<0.10) for homocysteic acid and at -40 mV and -60 mV (P<0.001) and -80 mV (0.01<P<0.025) for NMA. There was also some indication that this dependence of mean open time on agonist extended to the openings observed in inside-out patches (data not shown). Channel kinetics in the inside-out mode While the conductance properties of the channel were not affected by switching to the inside-out patch configuration, the kinetics of the channel were altered upon patch excision. The voltage-dependence of mean open time was reduced with patches recorded in the inside-out configuration. For a given cell, mean open times were similar, or only slightly dependent on patch potential, for most of the hyperpolarizing (that is, pipette positive) voltages, with the exception of the most extreme voltages (most notably +100 and +120 mV) which produced shorter mean open times than those recorded at the smaller voltages. A 42 Figure 6. Differences in mean open time between agonists. The graphs compare mean open times recorded over the transmembrane voltage range -40 mV to -120 mV for A) NMDA (n = 8) and NMLA (n = 15) and B) L-homocysteic acid (n = 18) and D-homocysteic acid (n=16). Significant differences in mean open time were found between each 20 mV voltage step (P<0.001 for each agonist). Differences in mean open time were also found between stereoisomeric pairs (P<0.001 for NMA and P<0.005 for HCA) . No dependence of mean channel open time on agonist concentration was observed. Statistics were performed using two-way ANOVA with voltage as one factor and stereoconfiguration as another factor. 4 3 £ CD E c CD % C (0 CD £ 2h • f - 4 0 NMDA mo.t.vs.Vm , i. ., - 6 0 i -ao Vm (mV) O N T f - 1 0 0 NM_A m.o.t.vs.Vm \ T o - 1 2 0 B to £ CD £ c CD a o c ID CO £ 2h _i_ - 4 0 L-HCA - 6 0 - 8 0 Vm (mV) - 1 0 0 D-HCA m.o.t.vs.Vm - 1 2 0 44 particularly striking effect of the patch excision in the semi-log plot (Figure 7) was the deviation from linearity of the dependence of the mean channel open times on patch potential for positive patch potentials. A channel that behaved quite differently from the NMDA channel openings under study often became activated when inside-out patches were applied with negative voltages. This channel had a considerably higher conductance, perhaps twice that of the NMDA channel, and usually went into a long-lived open state, in the order of hundreds of milliseconds, which usually appeared as noise in the data records. These long, noisy openings would tend to obscure or distort NMDA agonist-induced currents. When Cl~ ions in the bath and pipette solutions were replaced with S042~ ions, openings of this type did not appear when recording from inside-out patches with NMDA or L-homocysteic acid in the pipette, indicating that these openings were most likely produced by anionic currents flowing through an anion channel selective for Cl~ (Franciolini and Nonner, 1987). However, it was possible to record some data at negative potentials before this channel became activated and the data obtained indicated that openings which occurred at negative pipette voltages were approximately 20% longer on the average than those occurring at positive pipette voltages (Figure 7). 45 Figure 7. Kinetics of inside-out patches. Semi-log plot of mean open time vs. pipette potential for inside-out patches recorded with 100 /xM L-homocysteic acid in the pipette. Error bars, where appropriate, are mean ± S.E.M. 4 6 V (mV) T i T 1 100//M L-HCA 7 • 6 T I 5 4 m.o.t. (ms) T 1 I T 1 I T 1 — i 1 1 1 1 1 — -120 -100 -80 -60 -40 -20 0 T 1 n i i i i i ^20 +40 +60 +80 +100+120 47 1.3.6. CLOSED DURATION ANALYSIS In contrast to the open duration histograms, shut time distributions clearly fit functions with at least two exponentials (see Figure 8). Even in the limited examples presented in Figure 1 the grouping of openings into clusters, or bursts, separated by gaps of longer duration than the gaps within the bursts, is evident; the single-channel records for all agonists showed this type of behaviour. The existence of the two components in the shut-time distributions can be attributable to the burst-like behaviour of the channel, with the gaps within bursts contributing chiefly to the fast closed time component r^, and the gaps between bursts being involved in the generation of the slow closed time component T2« The tendency for NMDA-gated ion channels to open in clusters, producing bi-exponential closed time curves, has been previously observed (Howe et al, 1988; Gibb, 1989; Gibb and Colquhoun, 1991). The variation in either of the two closed components was large between cells and/or groups of cells (voltage, agonist, or concentration groups) and this made any kind of quantitative analysis difficult; however, there appeared to be a tendency towards longer closed time constants (either Tl o r T2) f°r t n e agonists producing shorter mean open times. Between the two concentrations for each agonist, the slower closed time component T2 showed a tendency towards greater values with the decreased agonist concentration, 48 Figure 8. Closed duration histograms. Histograms of durations between channel openings were constructed from a typical cell for each agonist and are shown for currents recorded at VM = -80 mV: A) 100 /xM L-HCA, B) 100 nM D-HCA, C) 200 /xM NMLA, and D) 30 JUM NMDA. The histograms were usually well-fit with two-exponential curves by a chi-squared curve-fitting routine. The fast time constants of the fitted curves were, in ms, A) 2.16±3.17, B) 7.18±1.38, C) 6.25±1.00, and D) 2.63±5.55. The late closed time constants were, in ms, A) 34.78±7.80, B) 74.85±35.37, C) 128.79±127.38, and D) 25.32±6.97. Each histogram represents data comprising a minimum of 200 events. Number of events Number of events o o & SI jr A Y^ w J t P •f 1 H -1 to CJ1 " J CO W - 3 V> III • r H «* II to b> CO 3 (A \ o z 2 D > Number of events o o in n> a. a. c Number of events DO o o U> IB D . Q . c 00 3 (A a • o > 00 ai 3 CA 50 with an increase in the area contained underneath the late portion of the curve as well. The modulation of the closed time distribution with voltage was complex, and no clear trend could be surmised, except that changing the patch potential produced effects only on the fast time component Tl with negligible effects on the slow component X2- The behaviour of the fast time component was related in inverse fashion to the open state probability (Po) for a cell at a given voltage: if Po decreased or did not change when the patch was hyperpolarized, the fast time component was usually observed to increase with voltage; whenever Po increased with voltage, as in those seldom cases where the opening frequency increased with patch hyperpolarization, r^ tended to increase with patch potential. Between the 55 cells studied, values for T^ ranged from about 0.9 ms at depolarized patch potentials to as much as 8 ms for extreme hyperpolarizations. Usually T 2 fell to values of about 45-60 ms with the higher agonist conentrations; at low agonist concentrations T 2 would increase to 80-90 ms, and shifting the data range allowed reasonable fits of a third closed component of about 150 ms, with r2 estimates being around 75-90 ms by this technique. Studies utilizing patch recordings for extended periods of time at a single potential have shown evidence of a closed time component in the same range as the third component observed here (Gibb and Colquhoun, 1991). Values of T 2 were lowest with L-HCA, averaging around 30 ms but never going higher than 50 ms 51 with 100 /xM L-HCA in the pipette; values for NMDA were next, ranging from 40-65 ms. T2 estimates for NMLA and D-HCA were highly variable, and while values of 60-75 ms were the most common, 40 ms and 90 ms time constants were not atypical. 1.3.7. FREQUENCY DEPENDENCE ON AGONIST CONCENTRATION For each agonist, there was a dependence of opening frequency on the concentration of agonist. Figure 9a shows the bar graph for frequency of single-channel events recorded at cell resting potential. Voltage, on the other hand, did not affect opening frequency, as determined by Kruskal-Wallis (non-parametric) two-way ANOVA. This allowed grouping together of data from all potentials as in Figure 9b, which depicts the difference in opening rate produced by the two agonist concentrations with data from all voltages combined. The two "distomeric" agonists, NMLA and D-HCA, produced quantitatively similar profiles of opening frequency versus concentration, while L-homocysteic acid was approximately 4-5 times as potent as the former two agonists in terms of opening frequency. The other "eutomer", NMDA, was even more potent than L-HCA, although it is difficult to claim this with certainty since only one concentration of NMDA was tested. 52 Figure 9. Dependence of NMDA-channel opening rate on agonist concentration. (a) Significant differences in the frequency of NMDA ion channel opening at cell resting potential (corresponding to 0 pipette potential) were observed between high and low pipette concentrations of each agonist (***P<0.02; **P<0.05; *0.05<P<0.10; two-tailed Mann-Whitney test); (b) As no significant differences in the channel opening frequency between membrane potentials were found for either of the agonists, the data from the cells in (a) for each agonist pipette concentration were pooled for all the voltages studied (-40 mV<V<-120 mV) . Significant differences were also found between concentrations for all three agonists (**P<0.001; *P<0.05; two-way Kruskal-Wallis test with voltage as the other factor). Sample sizes: n = 9 for 25 MM L-HCA and 100 /xM L-HCA, n = 8 for 50 juM D-HCA, 100 /zM D-HCA, and 200 AtM NMLA and n = 7 for 50 /zM NMLA. While only one concentration of NMDA was studied, it is included in the bar graphs so that its opening rate at 30 AIM can be compared with the other agonists. -60 mV 53 3 0 . 5 0 20O . 25 10O . 5 0 100 PIPETTE CONCENTRATION (M x 10"*) NMDA NMLA L-HCA l:-'--------\ D-HCA All Potentials 30 . 50 2 0 0 . 25 100 . 5 0 100 PIPETTE CONCENTRATION (M x 10"*) NMDA NMLA L-HCA D-HCA 54 1.3.8. OPEN STATE PROBABILITY The open state probability for a given patch is the ratio of net time spent in the open state to the total time recorded at a certain potential. The net open state time was calculated as the product of the mean open time and the channel opening frequency. Probabilities of opening in the 55 cell-attached patches studied were generally low, varying from between 2.8 x 10~ , with 25 AXM L-homocysteic acid in the pipette, to 1.8 x 10" , recorded with 100 /xM L-homocysteic acid in the pipette (indicating the channel was open 18% of the time). As with mean open time, significant differences between L- and D-homocysteic acid (P<0.001, 3-way ANOVA with voltage and concentration as the other factors) and between NMDA and NMLA (P<0.001, 2-way ANOVA with voltage as the other factor) were found. Increasing the agonist concentration in the pipette also increased the mean probability of opening, as shown in Figure 10 (P«0.001 for 50 and 200 iiM NMLA, P<0.001 for 25 and 100 /iM L-homocysteic acid, and 0.01<P<0.02 for 50 and 100 JUM D-homocysteic acid, all by separate 2-way ANOVA's with voltage as the other factor). Increasing the patch hyperpolarization generally produced a small decrease in open state probability; this change is presumably due to the decrease in mean open time associated with patch hyperpolarization. The modulation of open state probability with voltage was not significant, as determined by ANOVA. 55 Figure 10. Open state probability. Percentages of time channel spent in the open state for each of the agonists at each concentration used. Data were pooled for all patches using a particular agonist concentration, and for all patch potentials, since potential did not significantly affect open state probability for each agonist. Each higher concentration produced significantly higher open probabilities than the lower concentrations for a given agonist (*P<0.02; **P<0.001, 2-way ANOVA). Sample sizes are as given in Figure 9. 56 2 W CO +1 o T -X > _l m < m 0 cr a z in a 0 6 0 5 0 4 0 3 0 2 0 10 0 3 0 50 200 25 100 50 100 PIPETTE CONCENTRATION (M x 10"*) NMDA ^ M NMLA L-HCA l-<:^ D-HCA 1.4. DISCUSSION 1.4.1. CHANNEL CONDUCTANCE Measurements of the single-channel conductance, determined from the slope of the I-V relations, suggest that all four agonists studied activated the same conductance state of the NMDA ion channel. The values for the conductance, about 40 pS, are at the low end of the range for the NMDA-activated ion channel as measured in outside-out patches (40-50 pS) (Cull-Candy and Usowicz, 1989; Ascher et al, 1988; Ascher and Nowak, 1988b; Cull-Candy and Usowicz, 1987; Nowak et al, 1984). One possibility to account for the lower conductance value obtained is that a slightly higher calcium concentration, 1.8 mM (this aided in obtaining gigohm seals), was used in this study, while the earlier work used 1.0 mM calcium in the external bath solution. Calcium, possibly through a concentration-dependent charge-screening effect induced by the surface potential at the mouth of the channel, has an inhibitory effect on the channel conductance (Ascher and Johnson, 1990; Ascher and Nowak, 1988; Mayer and Westbrook, 1987a), and the nearly two-fold difference in Ca2+ concentration could very likely have produced the discrepancy in the conductance measurements. As currents were recorded in the absence of external magnesium, the current-voltage plots showed a linear relationship (Ascher et al, 1988). Non-linear macroscopic current-voltage relationships have been shown to be due to the voltage-dependent block of the NMDA channel by Mg2+ (Ascher and Nowak, 1988; Mayer and Westbrook, 1987a,b; Mayer et al, 1984; Nowak et al, 1984). 1.4.2. CHANNEL KINETICS Modulation with patch potential The mean channel open time of unitary currents induced by NMDA agonists consistently decreased exponentially with increasing patch potential for all agonists. Previous studies with NMDA (McLarnon and Curry, 1990a,b), the endogenous excitotoxin and NMDA agonist quinolinate (McLarnon and Curry, 1990b; Collingridge and Lester, 1989), and the potent NMDA agonist (Curry et al, 1987) cis-aminocyclopentanedicarboxylic acid (cis-ACPD) (McLarnon and Curry, 1990a), a geometric isomer of the metabotropic glutamate receptor agonist trans-ACPD (Watkins et al, 1990), have shown the same dependence of mean open time on patch potential when investigated with the same techniques. The possibility should be considered that this voltage dependence was conferred through contaminating levels of magnesium in the pipette solution; atomic absorption spectrometry was used to analyse the levels of magnesium in samples of bath and pipette solution used in the experiments and concentrations of Mg were determined to be less than 59 0.5 /iM. This concentration should have little (<1%) effect on the NMDA agonist-induced single channel currents since the KQ for Mg* -block of NMDA ion channels has been determined to be around 72 /xM (Jahr and Stevens, 1990; Ascher and Nowak, 1988b). The slope of the I-V plot for each agonist in both cell-attached and excised patches was linear over the entire ranges of patch potentials used, a feature which would not be observed if Mg2+ was present since it acts to produce non-linear current-voltage plots (Mayer and Westbrook, 1987a,b; Nowak et al, 1984) through its voltage-dependent channel pore-blocking action (MacDonald and Ascher, 1990; Ascher and Johnson, 1990; Johnson and Ascher, 1988). The appearance of the single channel currents conforms to channel openings not subjected to "fast channel block", that is, currents remained at a single open level for a certain length of time and were not chopped in bursts as were channel openings recorded in 30 JUM Mg2+-containing medium (Ascher and Nowak, 1988b; Mayer et al, 1984). Observed gaps between channel openings tended to be much longer than those expected for magnesium block (0.8-0.9 ms at cell resting potential) (Nowak et al, 1984). Voltage-dependent Mg block of NMDA-activated channel currents was associated with a reduction in burst length (Jahr and Stevens, 1990; Nowak et al, 1984), and thus attenuation of opening frequency, with increased patch hyperpolarization, while in the present studies opening frequency was independent of patch potential. Theoretical 60 surfaces constructed by Jahr and Stevens (1990) showed non-log-linear modulation of open time with voltage due to magnesium block, while the relationship between the two in cell-attached patches held at -60 mV was log-linear. Dependence of mean open time with voltage varied between individual cells more than would be expected if the dependence was due to a relatively constant rate-limiting factor, such as external magnesium. Finally, the process of patch excision altered the voltage-dependent kinetics, without any change in the ionic composition of the extracellular membrane side, the face from which cationic current (including any contaminating magnesium) flows when the patch is hyperpolarized. In the present studies, a single open state of the channel, as reflected in the single-exponential function fits to the open duration histograms, was assumed. Very brief openings, in a range below the resolution of the recording setup used here, were detected in a few patches, and these may be representative of a second, short-lived open state. Since openings in this duration range were not resolvable in this study, it is unknown whether or not patch potential affects both open states uniformly or in fact modifies the proportion of time spent in a particular open state. Recording data using a higher low-pass filter frequency, which would allow proper estimation of the amplitude and dwell time of very brief events (Howe et al, 1991; Hamill et al, 1981), would help in clearing up this point. Channel closed times, on the other 61 hand, were altered less consistently by patch potential. The fast shut time components did show some sensitivity to voltage; however, the relationship between the T^ values measured and patch potential was complex and varied considerably between cells, which made the analysis difficult. The tighter correlation of potential and mean open time when compared with that of potential and the fast shut time component suggests that the voltage-dependent factor governing the gating of the channel most likely affects 6, the inactivation rate constant for channel opening (see Section, such that 6 increases with patch hyperpolarization. The slower closed time components were usually unaffected by voltage, which suggests a lack of effect of transmembrane voltage on agonist binding to the receptor (Gibb and Colquhoun, 1991). The voltage dependence of mean open time observed in these studies conflicts with early studies of NMDA channel kinetics using outside-out patches, which failed to show any effect of patch potential on mean open time (Ascher et al, 1988; Ascher and Nowak, 1988b). The possible reasons for this discrepancy are discussed below. Kinetics in cell-free patches The alteration of channel kinetics observed following patch excision reflect the possible control that unknown intracellular factors may have over NMDA ion channel gating. The NMDA receptor is under multi-level control by the extracellular substituents, and some intracellular factors 62 controlling NMDA activity, such as intracellular calcium levels (MacDonald et al, 1989) and phosphorylation of the receptor-ion channel complex's cytoplasmic face by cellular kinases (MacDonald et al, 1989), have been identified. Previous investigations in which responses of outside-out patches to NMDA agonists were studied indicated that in the absence of external magnesium the channel kinetics were voltage-independent, and this finding was most likely due to the removal of a possible intracellular influence, such as those proposed above over the channel gating. More recent reports of voltage-dependent open times and open state probabilities recorded using outside-out patches in Mg^+-free solutions (Wright and Nowak, 1990), as well as isolation of Mg2+-dependent and Mg -independent processes contributing to the voltage-dependency of NMDA receptor-mediated EPSC's as recorded from granule cells in thin hippocampal slices (Konnerth et al, 1990), have appeared since those original reports. Differences in recording conditions or tissue and species type (earlier experiments were usually performed on mouse cortex or spinal cord while newer studies tend to utilize rat hippocampal neurones) may interact with the effect of cell excision on channel gating. Nevertheless, these results indicate that NMDA-channel kinetics, as well as non-competitive drug actions on those kinetics, observed depend on the patch configuration used. 63 1.4.3. AGONIST-DEPENDENT CHANNEL KINETICS Mean open time The difference («20%) in mean open time between the enantiomers of homocysteic acid and N-methylaspartate, suggests a relationship between the agonist structure and the degree of receptor activation. Therefore differences in agonist potency in inducing whole-cell currents may reflect not only the affinity of the ligand for the NMDA receptor but also its ability to induce current flow through the channel once bound. The data presented in this study indicate a close relationship between the conducting properties of the NMDA-associated ion-channel and the structure of the bound ligand-receptor complex. The fact that the nature of the bound agonist-receptor complex is communicated so rapidly to the ion-channel gating mechanism, modulating events occurring in the low millisecond range, is evidence that the agonist-recognizing receptor and the current-passing ion channel are part of the same macromolecular complex. As well, open time sensitivity to agonist structure was also observed in cell-free patches, suggesting that intracellular factors do not play a dominant role in this particular effect. The results also take on some significance when the controversy over the identity of an endogenous NMDA receptor ligand is considered. L-homocysteic acid was one of the compounds used as an NMDA agonist in this study, and its candidacy as an endogenous 64 neurotransmitter at NMDA sites has been previously proposed (Cuenod et al, 1990; Do et al, 1988; Do et al, 1987; Kilpatrick and Mozley, 1987; Cuenod et al, 1986). Analysis of the microscopic nature of synaptic currents evoked in pathways known to involve NMDA receptors (for instance, by including fractions of perfusate collected following stimulation of these pathways in patch pipettes) may provide more insight into the nature of the endogenous neurotransmitter when compared with kinetic data such as that obtained in the present study. Excitotoxic potential of a compound like L-homocysteic acid could possibly be indexed according to its efficacy in holding the NMDA ion channel open, and thus allowing toxic levels of calcium to enter the cell (Choi et al, 1988; Choi, 1987) either directly through the NMDA ion channel itself or through voltage-sensitive calcium channels activated by the depolarizing action of the agonist. A measure of their pathophysiological potential that is independent of binding affinity and thus concentrations required to produce excitotoxic effects is thus given. Identification of the endogenous ligand for NMDA receptors would be most useful in this regard as enzymatic pathways or transport and compartmentalization of precursors, methionine in the case of L-homocysteic acid, could be subsequently examined, either as sources of the excitotoxic pathogen or as targets for putative drugs treating neurodegenerative disorders. It should be noted that the rank order of mean open time 65 determined in this study, NMDA>L-HCA«NMLA>D-HCA, differs from the rank order of affinity, for the NMDA receptor as determined from radioligand displacement binding studies, L-HCA>D-HCA«NMDA»NMLA (Watkins and Olverman, 1988), and from the potency as measured in in vivo electrophysiological studies, NMDA>D-HCA>L-HCA*NMLA (Curtis and Watkins, 1962). Studies directly comparing agonist efficacies, affinities, and whole-cell effects would be useful in reconciling the anomalies presented by the above comparison, particularly with regard to the weak whole-cell effect produced by L-HCA even though it has been shown to have (a) a high affinity for displacement of NMDA receptor radioligands, and (b) in our hands it proved to be a moderate to highly efficacious agonist in terms of mean open time. Shut time kinetics Comparisons of two-exponential fits to shut-time histograms showed that both early and late closed time constants differed between agonists. Dependence of the fast closed time component on agonist, but not concentration, suggests an influence of the nature of the bound agonist-receptor complex on the activation kinetics. This, along with the dependence of the slow shut time components on agonist concentration indicates the following scheme underlies NMDA channel activation (Gibb and Colquhoun, 1991; Gibb, 1989; Howe et al, 1987): C2g;Ci<fo, where cc= the activation rate constant for channel opening (= l/t^), 6 = the inactivation rate constant for channel opening (= 1/mean 66 open time), and k+i = the rate constant for agonist binding to receptor; C^ and C2 are channel closed states and 0 is the channel open state. The large variability in the estimates of T^ between cells recorded with the same agonist at the same concentration makes any kind of proposal for an effect on a, the on-activation rate constant for the bound agonist-receptor complex, somewhat premature. The source of this variation could quite likely be the complex gating scheme of the NMDA ion channel. In this regard recent studies using low agonist concentrations applied to outside-out patches have shown distributions fitted by curves with as much as three components in the early portion of the closed-time histograms (Gibb and Colquhoun, 1991). The factors governing the activity of this multiple short-lived closed state scheme are unknown and variation between cells in the parameters describing the probability of the ion channel being in a certain shut state may have introduced some uncertainty in estimations of T± when a two-component scheme is used for curve-fitting. It would seem that future studies should focus on recording large numbers of recorded events (in excess of thousands) for accurate determinations of closed time constants. Further experiments with stable cell-attached patches, concentrating on collecting more data at a single voltage, would aid in clearly establishing the exact locus in the activation kinetic scheme of the agonist structure effect. Experiments resolving extremely fast closed times, in the microsecond range, would also be useful 67 adjuncts to the material presented in this study; comparing the values for the fast time components that can be resolved with higher low-pass filter frequency settings would be useful in developing a satisfactory gating scheme for the NMDA ion channel and for further determining the effects of voltage and agonist structure on the components of the receptor activation-ion channel gating scheme and their relationship to each other. The difference between agonists in the slow time component 12 further complicates the analysis; as discussed below, this parameter is a measure of the agonist binding rate during a particular experiment (Gibb and Colquhoun, 1991; Howe et al, 1987), and, since radioligand studies have shown that the binding affinities of the four agonists are not the same (Watkins and Olverman, 1988), different T2 values are not unexpected. The longer T2 values for L-HCA and NMDA suggests a greater tendency for these agonists to produce clusters of opens. While for a particular agonist the two closed time parameters appear, to a first approximation, to be independent of each other, it cannot be stated for certain whether or not differences in binding affinity due to agonist structure can interact with the status of the activation parameters once the agonist is bound. Concentration effects Increasing agonist concentration increased the frequency of opening (and thus Po) ; mean channel open time was unaffected by agonist concentration. Comparison of 68 closed time distributions between cells recorded with low and high agonist concentrations revealed the most salient change was in the slow time component. Data from cells recorded with low agonist concentrations produced distributions with higher T2's than those recorded with high concentrations of agonist in the pipettes. A third, slower component was more prominent in some patches studied with low agonist concentrations. The fact that the slower component(s) were modulated by agonist concentration gives an estimation of the dissociation rate constant for agonist binding to the receptor, by allowing the assumption that gaps falling in that portion of the histogram are due to agonist dissociation and not inactivation of the excited receptor-ion channel state (Gibb and Colquhoun, 1991). Values of T2 recorded at the higher agonist concentrations bottomed out at approximately 30 ms, which corresponds to an association rate constant of approximately 33 (I/T2) s""1. This i2 value agrees with closed time constants attributed to dissociation of agonist molecules to the receptor obtained in other studies using outside-out patches that were stable for extended recording periods (Gibb and Colquhoun, 1991). However, the possibility still exists that raising the agonist concentration may have affected the ion channel's ability to enter into long-lived closed states, and that T2 may represent an agonist-concentration dependent closed but bound state of the ion channel. Studies of evoked synaptic currents in rat hippocampal 69 neurones have suggested the existence of extremely long-lived closed states in the order of 10_1 seconds (Lester et al, 1990). However, an interaction of concentration with the activation kinetics of bound receptor has never been shown, with the possible exception of desensitization processes, which are usually time-dependent and were not observed in this experiment; the 1991 study by Gibb and Colguhoun supports the conclusion that the longest-lived closed state of the NMDA ion channel is described by a time constant of no more than 12 ms, with longer gaps most likely being the result of agonist dissociation from the receptor. The third closed time component observed chiefly with low agonist concentrations (T3«150 ms) may correspond to a low-affinity state of the receptor, wherein the agonist is unbound. It is possible that the ratio of the two late components could represent the mean number of "unsuccessful" agonist bindings, where binding of the agonist molecule(s) to the receptor site was not accompanied by favourable conditions supporting NMDA activity, for example allosteric modulation through strychnine-insensitive glycine binding. Between agonists, the trend in the late closed time component resembled closely the rank order of binding affinity determined in biochemical studies (Olverman et al, 1988a,b); this agreement lends additional weight to the notion that the slow shut time components correspond to agonist unbinding from the NMDA receptor. 70 2.1. INTRODUCTION 2.1.1. ALCOHOL Prevalence of use Alcohol is undoubtedly among the most widely and commonly used drugs in Western culture and its effects and consequences of its usage have appeared in the writings of nearly every civilization in written history; indeed most civilizations had discovered alcohol before they even developed writing. Today, in Canada, alcohol consumption and production plays a vital role in the economy: $3.78 billion in provincial and federal revenues were generated through the alcohol industry in 1986 alone. According to a Health and Welfare Canada survey conducted that same year (the last year on which alcohol statistics are available) 81% of the total Canadian populace of age 15 years or older consumed alcohol in the twelve months preceding the survey. Two-thirds of this group were "current drinkers" - that is, they consume alcoholic beverages on a regular basis - and males in this group consume over 8 drinks per week on the average. Contrary to popular conceptions, the greater the education and/or income of the survey subjects, the greater the likelihood that they would admit to drinking in the past month. Students as a group contain a high proportion of "current drinkers": 39% of students polled would consume 71 greater than 5 drinks at a single party, while only 8% would refrain from drinking any alcohol at all at a party. Amazingly, 53% of students between grades 7 and 13 reported being drunk at least once in the previous year. Canadians over 15 years of age consumed 10.2 litres of alcohol per capita in 1986; if that figure sounds a bit excessive, it should be noted that Canada ranks a mere 21st out of 32 industrialized countries on the scale of alcohol consumption (c.f. Eliany, 1989). Consequences of mass alcohol consumption Despite its widespread acceptance and apparently ubiquitous utilization in Canada, alcohol consumption is not without its problems, particularly social and medical problems. The 1986 survey revealed that 7 out of 10 Canadians feel the government should be involved in controlling alcohol abuse, while only 53% believed that the government should act to cut down on the general population's smoking habits. In total $5,252 billion, or $215.75 per capita, was spent in 1986 alone for costs related to alcohol abuse, including health care, social welfare, law enforcement, and reduced productivity. All through the 1980's, more than 30000 mental, psychiatric, and general hospital separations were attributed to alcohol use, mainly for such conditions as liver disease (cirrhosis, hepatitis), pancreatitis, various gastrointestinal illnesses, anaemias, cardiomyopathies, short-term memory deficits, and Wernicke-Korsakoff and chronic brain syndromes 72 (c.f. Eliany, 1989). It seems somewhat unusual, given all of the above statistics, that the basic mechanism of action of alcohol, despite over 80 years of research into the matter, still eludes the scientific community. Pioneering research The earliest attempts at systematic research into the question of alcohol's mechanism of action were performed by two researchers, E. Overton and H. H. Meyer, working independently at the turn of the century. They found, using aquatic animals such as tadpoles, that the potency of a variety of anaesthetic agents in producing narcosis was related directly to the agent's lipid solubility, measured as the olive oil:water partition coefficient; this concept is now known as the Meyer-Overton rule and the normal alcohols, up to a critical chain length (n=13) at which point the addition of a further methylene group abolished the alcohol's anaesthetic properties, were included as conforming to this rule (Goldstein, 1983). The subsequent debate opened by the Meyer-Overton studies was whether increasing the lipid solubility merely increases access of the agent to sites of action that require the crossing of a lipid barrier or if the lipoidal compartment is in fact the direct site of action. In 1954, L.J. Mullins refined the Meyer-Overton hypothesis by using, instead of simple partition coefficients, ratios of the activities of the anaesthetic agents in olive oil and water and thus achieving a closer lipid solubility-narcotic potency relationship by 73 taking molecular volumes into account (Mullins, 1954). On the other hand, studies were ongoing over the next two decades concerning neuronal responses in terms of electrical and chemical functioning upon alcohol administration, and theories dealing with the observed inhibition of Mg2+-activated Na+-K+-ATPase (Goldstein, 1983; Sun, 1979; Seeman, 1972) and activation of intracellular calcium by ethanol (Krnjevic, 1974) came to be regarded as attractive alternatives (or necessary corollaries) to the idea of membrane phospholipids as the primary site of alkanol action. In the late 1960's, Philip Seeman, working out of the University of Toronto, found that the potency of various anaesthetic agents, including the normal alcohols, was proportional to their antihaemolytic effects via their ability to expand erythrocyte membranes (Seeman, 1972). In 1974, Seeman used a high-precision densitometer to compare membrane expansions by ethanol between model membrane systems and actual biological membranes. What he found was that the membrane expansion was much greater in biological membranes than in the simple lipid mixtures that composed the model membranes; this was attributed to additional swelling of protein structures contained in the membrane (Seeman, 1974). The results of this study were the first indications that alcohols do not merely produce their effects by altering the density of lipid cell membranes but can have a direct influence on the function of enzymes, receptors, ion channels, and pumps contained in the 74 membrane. Eventually, through the work of several investigators working in the field of alcohol and anaesthetic research, it became clear that changes in the physicochemical parameters of lipid bilayer structure, such as bulk fluidity and gel transition temperatures, induced by anaesthetic molecules were best attained at suprapharmacological concentrations, and that at surgical concentrations the effects of most anaesthetics, including the normal alcohols, on membrane lipid parameters was in the fraction of a percent range; moreover, these effects could be mimicked by slight changes in temperature which are well within the range experienced by any organism as part of its usual diurnal fluctuations (Franks and Lieb, 1987). That the lipid-free protein firefly luciferase was found to be inhibited by anaesthetic agents in a manner correlated with their lipid solubility further strengthens the idea that alcohols inhibit protein function as a possible mechanism of action (Forman and Miller, 1989). Thus began the search for the actual intramembrane system(s), probably involving one or more excitable proteins, that contribute to the action of alcohols and anaesthetics. Current research Research investigating the neurochemical target for ethanol's action contained within the cell membrane have been ongoing all through the 1980's. Work using acetylcholine acting on nicotinic receptors in the neuromuscular junction as a model system for alcohol effects 75 on receptor systems determined that ethanol acted to increase the frequency of miniature endplate potentials (mEPP's) and to increase the duration of both the early and late phase of the decay of miniature endplate currents (mEPC's) by both prolonging the duration of the ion channel openings and by extending the lifetime of the bound ACh-nicotinic receptor complex (Linder et al, 1984). The CNS receptor system that has received the most attention in recent years as the primary mediator of the central effects of ethanol has been the inhibitory GABAergic system, particularly since the demonstration of increased GABA^ anion-channel activity and the enhancement of GABA-stimulated 36C1" influx in the presence of ethanol (Allan et al, 1986; Suzdak et al, 1986); however, the effect of enhanced GABA^ activity at low ethanol concentrations has been difficult to reproduce in electrophysiological studies (White et al, 1990), and more work dealing with the apparent heterogeneity of GABA^ receptors and/or cell responsiveness to GABAergic input and modulation must be done before a mechanism of ethanol action pertaining to GABAergic systems can be postulated (Gonzales and Hoffman, 1991). Ethanol and NMDA activity Ethanol has been found to inhibit whole-cell NMDA currents (White et al, 1990; Lovinger et al, 1989) and NMDA-induced intracellular calcium accumulation (Dildy-Mayfield and Leslie, 1991) with a potency and consistency greater than can be observed with any other ethanol effects on 76 nerve-cell functioning (Gonzales and Hoffman, 1991). In the electrophysiological studies, the potency of the NMDA current inhibition was proportional to the chain-length of the normal alcohol applied to the cell preparation. This observation is consistent with the Meyer-Overton rule in that the lipid solubility of the normal aliphatic alcohols is proportional to the number of carbon atoms in the chain. We have studied the effects of two intermediate-chain alcohols, butanol (n=4) and pentanol (n=5), plus a longer-chain agent, octanol (n=8), on the kinetics of NMDA currents as recorded in the cell-attached mode; as indicated in section 1.1., this patch-clamp configuration preserves cellular modulation of NMOA channel kinetics, while excised patch configurations do not. These studies were augmented by microspectrofluorometric experiments using the fluorescence probe fura-2, which measures intracellular Ca^ levels independent of the internal dye concentration (Grynkiewicz et al, 1985), in order to assess the effect of the n-alkanols on NMDA-induced increases in intracellular calcium. 2.1.2. VOLATILE ANAESTHETICS Historical background The use of anaesthetic agents in medicinal practice is nearly as old as the medical tradition itself. While reports of unanaesthetized invasive procedures litter 77 accounts of surgical practices performed anything more than 150 years ago, these situations occurred, or were recorded to occur, in cases where either poverty of the hapless patient was a determining factor towards the withholding of pharmacological comfort, or where the watchful eye of the church or the (very) conservative medical establishment may have been present to strongly discourage the use of consciousness-altering anaesthetic drugs. Agents promoting narcosis to a level sufficient for surgery have been available for use (and were quite often used) since at least Egyptian times, and many were often known to the general populace, especially women, as part of the lore associated with home remedies. Alcohol, cannabis, and opium were well known for their anaesthetic properties, and the use of the mandrake flower, containing quantities of the centrally-acting antimuscarinic L-hyoscyamine or scopolamine, was the most commonly referenced method of surgical anaesthesia from the 2nd millenium B.C. until just a few centuries ago. The era of modern anaesthesia began in the 1840 's with the discoveries of the anaesthetic properties of diethyl ether and chloroform - the former agent being first prepared in 1540 by Valerius Cordus and begun use as an anaesthetic in New York in 1842 by Clarke and in Georgia that same year by Crawford Long, and the latter agent being produced in New York in 1830 by Samuel Guthrie and demonstrated in 1847 by James William Simpson on himself in Boston (c.f. Atkinson and Boulton, 1987). Nitrous oxide, originally isolated in 78 1772 by oxygen discoverer Joseph Priestley, initially fell into disfavour as a useful anaesthetic after a failed public demonstration in Boston by dentist Horace Wells in 184 4; Edmund Andrews1 reintroduction of N2O over twenty years later in Chicago, this time with added oxygen to minimize the possibility of asphyxiation by the patient, regained the credibility of the gas as a general anaesthetic (Eger, 1985). It should be noted that the above three substances would enjoy heavy popularity after, if not before, the discoveries of their anaesthetic properties as recreational inhalants, and the development of modern-day vapourizers and masks for anaesthetic administration has in its roots the devices originally designed for recreational purposes; John Snow's 1846 ether vapourizer, which had as a component the first successfully designed mask for anaesthetic administration, was an improvement upon the novelty contraptions originally designed for recreational inhalation. By the late nineteenth century Theodoor Hammes ("the municipal stupifier"), considered by many to be the first professional anaesthetist, had in his medicine cabinet chloroform, ethyl chloride, nitrous oxide, ether, and carbon dioxide. Around this time chloroform fell into disfavour, mainly in North America, because of the high mortality (1:3000), usually due to cardiac complications such as syncope, associated with its use (c.f. Atkinson and Boulton, 1987). However, new anaesthetics were being introduced, and the early twentieth century brought trichloroethylene 79 (discovered as an anaesthetic in 1917) and cyclopropane, introduced in the 193O's. Even pure oxygen found use as a make-do anaesthetic during World War I. By 1940 nitrous oxide, ether, procaine, cyclopropane, trichloroethylene, and the newly developed thiopentone were the most common first-line anaesthetics (c.f. Smith and Aitkenhead, 1985). However, the problems associated with these agents (cardiac and hepatic complications with cyclopropane and trichloroethylene, incomplete anaesthesia with N2O, explosions with ether) necessitated the development of more favourable agents with respect to volatility, nonflammability, high potency, high safety margin, safety against damage to vital organs, rapid, quiet induction, and cardiovascular toxicity. The fluorinated hydrocarbon halothane, tagged as a potential anaesthetic by Raventos in 1946 after 2 years of testing 46 different halogenated hydrocarbons, satisfied the above criteria and was first manufactured for use in clinical practice by Suckling in 1956 in the U.K. and in 1958 in the U.S.A. Enflurane followed in 1971 after its development and initial clinical studies, in 1963 and 1966 respectively, by Ohio Medical Products of New Jersey. An isomer of enflurane, isoflurane, was developed in 1965 by Ohio Medical Products, but was not introduced into clinical practice until ten years after clinical studies began in 1970 due to early reports of its potential carcinogenicity in laboratory animals (c.f. Smith and Aitkenhead, 1985). These three agents, along with the 80 related methoxyflurane, comprise the volatile anaesthetics whose prevalent use has greatly reduced the worries of anaesthetists and surgery patients alike in the past few decades. While researchers have come a long way in developing safe anesthetics that are not particularly unpleasant to the patient, there is one significant point about the nature of these agents that eludes the medical community at large: their basic mechanism of action (Gonzales, 1990). Theories concerning anaesthetic action For the most part, theories concerning the mechanism of action of the volatile general anaesthetics have centred around the Meyer-Overton rule. The interchangeable use of the terms "alcohol" and "anaesthetic" in Section 2.1.1. of this thesis is a reflection of the state of understanding of the mechanism of action of the two types of drug within the scientific community. Ethanol has anaesthetic properties, and there is a basic stuctural homology between ethanol and the volatile anaesthetics: for example, ethanol is a monohydroxylated ethane, halothane is a halogenated ethane, and isoflurane is a difluoromethane ether of halothane save for the bromide substitution. Both classes of drugs lack specific antagonists for their effects. General anaesthetics, like the normal alcohols, expand biomembranes such as erythrocyte ghosts; like alcohols, the expansion is about 10-fold greater than to be expected by simple insertion of the molecular volume of anaesthetic into the 81 membrane (Seeman, 1972). Similarly, the expansion of biomembranes by general anaesthetics is much greater than the expansion of simple lipid model membranes by the same drugs (Seeman, 1974). While many other studies have been published that confirm the fluidizing effects of volatile anaesthetics in biomembranes (Nandini-Kashore, 1977; Trudell et al, 1973), interactions with membrane proteins have also been demonstrated (Wyrwicz et al, 1983; Godin and del Vicario, 1981; Katz and Simon, 1977), and at more clinically relevant concentrations. As mentioned in Section, theories attributing the mechanism of anaesthetic action to alteration in membrane fluidity do not take into account the high concentrations required to demonstrate a significant fluidizing effect (Forman and Miller, 1989; Franks and Lieb, 1987). If interactions with lipophilic intramembrane sites, such as hydrophobic domains of ion channel proteins (Godin and del Vicario, 1981; Katz and Simon, 1977), are required for anaesthetic action, the Meyer-Overton rule, that is, increased potency with increased lipid solubility, would still be expected to hold, especially if transport through the membrane lipids is necessary to gain access to these sites. We have studied the effects of two inhalational anaesthetics, halothane and isoflurane, on the activation kinetics of the NMDA ion channel, so that analogy with the aliphatic alcohol actions on the same system can be drawn. 82 2.2. METHODS 2.2.1. CELL PREPARATION AND PATCH-CLAMP RECORDINGS Hippocampal CA1 neurons were isolated and cultured exactly as in section 1.2. The protocol for patching the cells was also the same as described in that section, using either NMDA, L- or D-homocysteic acid in the patch pipette. The entire current recording and data acquisition/storage system used in this experiment, as well as the methods used for data analysis, was no different from that described in section 1.2. 2 . 2 . 2 . Jl-ALKANOL AND ANAESTHETIC PERFUSION Drugs were supplied through inverted syringe tubes suspended above the level of the bath so drug solutions would flow out of the syringes by gravity. Polyethylene connecting tubes fitted with a flow-rate control device delivered solutions to the bath chamber. For the alcohols, a certain volume of a concentrated stock solution was added to the standard bath solution in order to achieve the desired concentration, which was either 3 mM for 1-butanol, 1 mM for 1-pentanol, or 0.02 mM for 1-octanol. For the anaesthetics, a Cyprane Fluotec-3 vapourizer was filled with the desired agent and the anaesthetic vapour (set at 4% in 83 95% O2 and 5% CO2, flowing at about 250-300 mL/min) was bubbled into one of the syringe supply tubes for a minimum of 20 minutes to allow the anaesthetic to equilibrate with the bath solution in the reservoir; the bubbling of the solution with anaesthetic was continued during bath perfusion. In both cases, the syringe tubes containing either the alcohols or the volatile anaesthetics were sealed off with Parafilm in order to reduce evapouration of the volatile agent out of the solution. In addition, a special sealed-bath setup was designed for the volatile anaesthetic experiments so that evapouration of the volatile anaesthetic from the bath itself during recordings made while the drug was on board could be minimized. The butanol and octanol were from the Fisher Scientific Co., the pentanol was from BDH Chemicals, Ltd., the halothane came from Ayerst Laboratories, and the isoflurane was supplied by Anaquest, Inc. 2.2.3. CURRENT RECORDING DURING DRUG APPLICATION Initially, control recordings were taken over a range of voltages. The tubes connecting the reservoirs containing the drugs were then opened and solution was allowed to flow into the bath at a rate of 2-3 ml/min. As the bath had a total volume of about 2 ml, a suction tube was used to remove excess solution to prevent overflow; in this way the bath medium could be exchanged completely. Data were 84 recorded after the bath solution had been exchanged with the solution containing the drug at least two or three times. Occasionally, before the reservoir containing the drug could be opened, channel activity as observed in the oscilloscope would cease completely; at this point the experiment was stopped. While a definite explanation for this phenomenon is not possible, checking the position of the pipette on the cell through the microscope upon occurrence of this situation most often revealed that the pipette had moved quite significantly across the body of the cell, or even completely off the cell, indicating the integrity of the patch had been destroyed. As well, the currents would sometimes drop in amplitude when observed at hyperpolarizing voltages, and would not be present at all when no voltage was applied to the pipette. This was considered to represent a spontaneous excision of patch to the inside-out mode, as confirmed by the appearance of the cell and pipette through the microscope. After the full range of voltages had been recorded with the drug on board, the tube connecting the reservoir containing the drug was closed off and a second reservoir, containing the control bath solution with no added drug, was opened in order to take re-control data demonstrating the recovery of the channel kinetics from the effects of the drug. 85 2 . 2 . 4 . FURA-2 MEASUREMENTS. The calcium-sensitive fluorescent dye fura-2 was used to ascertain the effects of the n-alkanols on NMDA-induced increases in intracellular calcium. Rat CA1 hippocampal neurones were cultured in exactly the same manner as described in the first Methods section, and were used for the fluorescence experiments at approximately the same age as were the cells used in the patch-clamp experiments. For details concerning the theory and methodology of intracellular Ca2+ indicators like fura-2, see Grynkiewicz et al, 1985. Fura-2 loading The coverslips with the hippocampal neurones attached were first rinsed twice with the fura-2 loading medium, which contained, in mM: 112.9 NaCl, 26.2 NaHC03, 1.1 Na2HP04-7H20, 3.4 KC1, 6.7 HEPES, 3.3 Na-HEPES, 5.6 d-glucose, 0.8 MgSC^^^O, 1.8 CaCl2, and 5.0 pyruvate, with pH adjusted to 7.25±0.02 with 1.0 N NaOH and 0.0005% phenol red added as an indicator. The NaCl, KC1, HEPES, Na-HEPES, and d-glucose were obtained from the Sigma Chemical Co., the sodium bicarbonate and the sodium hydroxide were from the Fisher Scientific Co., the disodium hydrogen orthophosphate and the magnesium phosphate were from BDH Chemicals, Ltd., the calcium chloride was from Anachemia, Ltd., and the pyruvate was from Calbiochem. Then the coverslips were placed in 2 ml of loading medium that was identical to the 86 loading medium described above except for the addition of 0.1% bovine serum albumin, 7.5 juM (from a stock of 1 mM fura-2/AM in DMSO) fura-2/AM, the permeant acetoxymethyl ester form of the probe, and 6.75 pi of Pluronic F-127, a nonionic dispersing agent for solubilizing large dye molecules (Poenie et al, 1986), from a 20% in DMSO stock (total DMSO 250 nm) ; the above three agents were purchased from Molecular Probes, Inc. The cells were incubated with this fura-2/AM/PF-127/BSA medium for 1.5 hours at 37°C with 5% CO2. Between the end of this incubation time and the actual spectrofluorimetry procedure, another hour or two, the cells were placed in Hank's balanced salt solution (138.6 mM NaCl, 1.0 mM NaHC03, 1.1 mM Na2HP04-H20, 3.4 mM KCl, 6.7 mM HEPES, 3.3 mM Na-HEPES, 5.6 mM d-glucose, 0.8 mM MgS04*7H20, 1.8 CaCl2, and 5.0 mM pyruvate) at room temperature. This constituted an "equilibration period" during which unhydrolyzed fura-2/AM was washed out of the cells so that subsequent leakage of fura-2 from the cells during the recording period and also compartmentalization of the dye into intracellular organelles were minimized (Puil et al, 1990); therefore most, if not all, of the fluorescence signal measured during the experiment represented reaction of the fluorescent probe with cytosolic Ca2+. Fluorescence measurements For the spectrofluorimetric measurements, the coverslips were placed cell-side down in a Perspex laminar 87 flow-through chamber, of approximate volume 350 /xl, and, after sealing the coverslip to the chamber with silicon gel, the unit was placed in a stainless steel bath. A cassette pump supplied, at 2-3 ml/min, the bathing medium to the bath, via polyethylene tubes, from reservoirs containing either control bath solution ("NMDA medium"), or "NMDA medium" plus an n-alkanol; the supply to the bath could be switched between reservoirs at will. The bath medium contained, in mM: 137 NaCl, 0.34 Na2HP04-7H20, 2.8 KC1, 2.8 KH2P04, 5.6 d-glucose, 5.0 HEPES, 0.0005% phenol red, 1.8 mM CaCl2, 1.0 /iM glycine, and 500 nM TTX, all adjusted to pH 7.25 with NaOH. The drug reservoir contained the exact same perfusing solution as above except for the addition of n-alkanol. The perfusing solution was maintained at 23°C by heating elements, located near the bath, which were monitored by thermistors and controlled by a regulation device. The stainless steel holder with the chamber containing the cells was mounted onto the stage of a Zeiss Jenalumar microscope equipped for epifluorescence. Light from a DC-powered 200 W mercury arc light source passed through one of two 10 nm bandpass differential interference filters, 350 or 380 nm (determination of bound [Ca2+]i and free [Ca2+]j_, respectively), which were mounted on a turret driven by a computer-controlled stepping motor. The light then passed through a 410 nm dichroic mirror and a lOOx apochromat oil immersion lens with a numerical aperture of 1.4; light intensity was controlled here with an adjustable diaphragm. A circular diaphragm was located above the lens in order to isolate the soma of individual pyramidal-appearing neurones for study. The selected neuron was excited sequentially with 350 nm and 380 nm light on a 1.8 second time base, and the fluorescent light passed back through the dichroic mirror and through a 490 nm bandpass interference filter to cut out background fluorescence. The light could then be deflected either to the eyepieces of the microscope or to a photomultiplier tube which could convert the emitted fluorescence to a DC voltage, which in turn was converted to digital form with an analog-to-digital converter of a data acquisition system for storage on computer. The DC output of the photomultiplier tube was also connected to a Tektronix T912 storage oscilloscope for monitoring of the fluorescence signal during the experiment. Calculation of [Ca2+]i The concentration of cytosolic free calcium was calculated from the ratio of the emitted fluorescence at 350 nm to 380 nm (i.e. bound to free calcium) according to the method of Grynkiewicz et al, (1985). Briefly, [Ca2+]i = K<j x 8 x (R - Rmin) / (Rmax ~ R) ' where K<j is the equilibrium dissociation constant for the association of fura-2 with cytosolic free calcium, assumed here to be 220 nM, 6 is the ratio of 380 nm light with zero [Ca2+]i to 380 nm light with infinite [Ca2+]i, R is the experimentally determined ratio of 350 nm to 380 nm light, Rmin is the ratio of 350 nm to 380 nm light with zero extracellular Ca2+, and Rmax is the 89 ratio of 350 nm to 380 nm light with infinite external calcium. These values were determined from calibration procedures using the calcium ionophore Br-A23187 either in the absence of external calcium with 10 mM EGTA or with excess (1.8 mM) external calcium. Determination of n-alkanol effects on NMDA-induced [Ca2+]i After taking background readings for R with the NMDA medium and without the light source, baseline measurements were taken with the ultraviolet light and the flow of NMDA medium continuing at 2-3 ml/min. NMDA (200 /xM) was injected as a bolus of 20 /ul into the laminar flow of NMDA medium via a syringe that could empty into the flow line through a 3-way connector; this resulted in an approximate 2- to 3-fold dilution in the bath itself. After at least two control trials with NMDA, the perfusing solution was switched to that containing the ii-alkanol, which was either 3 mM 1-butanol, 1 mM 1-pentanol, or 20 /zM 1-octanol; after a brief period of time (usually 1-2 minutes) to allow for some degree of equilibration between the alcohol and the cells in the recording chamber, one or more trials with injections of NMDA were recorded. Finally, the bath medium was changed back to the NMDA medium without drug, and some more trials with NMDA injections were recorded in order to produce recovery data. 90 2.3. RESULTS 2.3.1. THE ALCOHOLS n-Alkanol effects on unitary currents Figure 11 shows currents recorded during perfusion of (a) 1 mM pentanol (n=5), (b) 3 mM butanol (n=4), and (c) 20 /zM octanol (n=8). The essential primary effect of each of the n-alkanols was to diminish the channel activity produced by the NMDA agonists. Note the apparent decrease in open time as well as a decrease in the frequency of single-channel events when each of the n-alkanols were perfused into the recording bath. The concentrations of butanol and pentanol chosen for this study correspond to concentrations producing maximal effect in the respective dose-response curves for inhibition of whole-cell current responses to bath-applied NMDA (Lovinger et al, 1989), while the octanol concentration is close to the EC50 determined for that agent (Lovinger et al, 1989); however, the inhibitory effects of the two shorter-chain alkanols were also observed at their respective ECSQ'S, 1 mM and 0.3 mM, respectively (data not shown), although the magnitude of the inhibition was quite small. The appearance of the single-channel currents was otherwise unchanged by the n-alkanols. 91 Figure 11. Unitary currents in control and in the presence of alcohol. Single-channel on-cell patch clamp records using 40 /zM NMDA and 1 ^M glycine in the pipette; TTX was also included in the bath and pipette media. The top traces in (a-c) were recorded in different patches under control conditions while the lower traces were obtained from the same cell as the control trace above it in the presence of (a) 1 mM pentanol, (b) 3 mM butanol, and (c) 20 fMM octanol. Pipette potentials were 0 mV in (a), 20 mV in (b), and 40 mV in (c) . 92 1 mM pentanol TinT^/TlP" I , ; ' I Mr *| 3 mM butanol "'I'M'*; yf i^ '^nrvi n flr*T-| ww i/ir r-v* v j^srtrv^x r+trtr-20 //M octanol I 'I r T r*rt r+ftf**tr~C*^ P"HM 3pA 10 ms Current-voltage relation Figure 12 shows the I-V curves constructed from the amplitudes of the single-channel currents recorded in control solution and during perfusion of 1 mM pentanol (n=7). Current amplitudes were not significantly altered by n-alkanol perfusion, as shown in the curves; however, a few of the cells responded to the alcohol perfusion with a small increase in channel amplitude at each voltage. Measurements of cell resting potential during alcohol perfusion using whole-cell patch-clamp revealed a tendency for the n-alkanols to slightly hyperpolarize the neurons by 2-8 mV, and the magnitude of this effect on the resting potential appeared sufficient to explain the size of the single-channel current increase. The slope conductance of the channel during perfusion of pentanol was determined to be 40 pS; similar conductance values were found for 3 mM butanol and 20 jitM butanol. The zero-current potential was determined from extrapolation of the I-V curves to the ordinate and was -58 mV for 1 mM pentanol, with similar values being found for the other two alkanols. Distributions Histograms for open (a, c) and closed (b, d) times for a single selected cell measured in control solution (a, b) and in the presence of 1 mM pentanol (c, d) are shown in Figure 13. Open time distributions were adequately fit with single-exponential functions; note the apparent decrease in the fitted curve time constant (which can be approximately 94 Figure 12. I-V curves in control and alcohol solutions. Current-voltage relations for cell-attached patches in control solution (filled circles, n=7) and for the same cells in the presence of 1 mM pentanol (open circles). The agonists in the pipette were 25 or 40 juM L-homocysteic acid, or 40 /iM NMDA; data were pooled for agonist and concentration. The solid straight line is a visual fit to the control data; S.E.M. values were all below 0.2 pA. Abbreviations: I = unitary current amplitude in pA; V = pipette voltage in mV. 95 -60 -40 -20 96 Figure 13. Open and closed duration histograms. Distributions for open time and closed intervals are depicted in (a) for control and (b) in 1 mM pentanol, using 40 iM L-homocysteic acid in the patch pipette as agonist. The open time distribution in (a) consisted of 515 events and the fitted curve had a time constant of 1.9 ms, while the closed time histogram represented 570 events and was fit with two components of 1.8 and 13.5 ms. In (b), the open duration histogram contained 445 events and was fitted with a single component of 1.3 ms, while the closed duration histogram had 500 events and was fitted with 2.4 and 20.5 ms time constants. Curves were least-squares one- and two-exponential fits for the open and closed durations, respectively, performed using pSTAT 5.0. 100//M L-HCA 150 A (control) 150 r x = 1.9 ms B (1 mM pentanol) o 75 • j T = 1.3 ms 2 0 0 TT = 1.8 ms 100 so CLOSED TIMC I M J 100 150 ti = 2 . 4 ms o 75 J x2 = 20.5 ms visualized as an increase in the rate of fall of the function as open time increases along the ordinate) when going from control solution to pentanol. This effect was also observed in cells where either 3 mM butanol or 20 /iM octanol were added to the bath. Closed time distributions were also affected by addition of alcohol to the bath, with a 2-3x increase in the fast time component T^ being the most obvious effect of the drug treatment on the channel kinetics. On the other hand, the slow time constant 12 w a s unaffected by alcohol perfusion. Reduction of mean open time by n-alkanols The percentage effects on mean open time were averaged for each of the alcohols and these are summarized in part (a) of Table II. As there was no significant difference between NMDA and homocysteic acid in the effectiveness of either of the alkanols in attenuating the currents produced by those agonists, the values in the table have been pooled for the agonists used to induce the single-channel currents. The reductions in mean open time for each n-alkanol at each patch potential were analyzed using the paired student's t-test, and the results were found to be significant at the a=0.05 level for each case except for the effect of 0.02 mM octanol at VM=-100 mV. In experiments where the patch formation was mechanically stable enough to withstand a second bath perfusion, control solution was reperfused into the recording chamber following collection of data with 99 Table II. Effects of the n-alkanols on mean channel open time and [Ca2+]± increases. (a) Percentage reductions in mean NMDA ion channel open time, relative to control, produced by 3 mM butanol, 1 mM pentanol, and 20 /zM octanol, using 30 /xM NMDA as agonist in the pipette; (b) percent reductions in 200 JUM NMDA-induced intracellular calcium increases, as determined with fura-2 microspectrofluorimetry, produced by the same alcohols, at the same concentrations, as in (a) . In both (a) and (b) , all values were statistically significant (P<0.05; paired student's t-test) except for Vp=40 mV and the [Ca2+]i increase, both in 20 /xM octanol. 100 A Mean Open Time Vp « 0 mV Vp = 20 mV Vp = 40 mV Butanol (n = 6) 32 + 3 % 34 + 4 % 27 ± 4 % Pentanol (n » 7) 32 ± 5 % 31 ± 3 % 27 + 4 % Octanol (n = 4) 28 + 4 I 25 ± 4 X 22 ± 7 % B [Ca2+] Butanol (n = 3) Pentanol (n = 6) Octanol (n = 3) 23 ± 6 % 24 ± 5 % 11 i 6 t alcohol in the bath. In the majority of those cases, mean open times showed partial recovery to control values (data not shown); however, recoveries were observed to occur very slowly and for the most part it was impossible to maintain a cell-attached patch for a sufficient length of time in order to demonstrate complete return to control values. Increasing the concentrations of the intermediate chain length alkanols butanol and pentanol, from 3 to 10 mM and 1 to 3 mM, repectively, did not produce any further decrease in mean open time (data not shown), indicating the effect was near-maximal at 3 mM butanol and 1 mM pentanol. Figure 14 shows the dependence of mean open time on applied voltage for currents recorded in control solution and those recorded during perfusion of 1 mM pentanol. One feature of the graph is that the pentanol addition did not alter the modulation of open time with patch potential, which is evidenced by the near-equivalent slopes of the linear fits to the two sets of points. Alcohol and open state probability The effects of n-alkanols on Po were inconsistent and varied considerably; in some cases the open state probability decreased to as low as 13% of the probability recorded in control solution while in other cells the open state probability actually increased to as high as triple the probability observed under control conditions. This high variability can be attributed to the inconsistent 102 Figure 14. Relationship between patch potential and mean open time. Semi-log plot of pipette potential (V) against mean open time (m.o.t.) in control (filled circles; n=7) and in the same patches in the presence of 1 mM pentanol. Error bars are ± S.E.M. r 3 < o o OPEN TIME |ms| to 1/1 1 1 HCH l -O-t H#H f—O—• H O effects of the n-alkanols on the channel opening rate. While 10 of the 17 cells underwent a decrease in opening frequency (a decrease of 20% was typical), 7 out of 17 cells responded to alkanol addition by increasing their opening frequency, usually by about 50% but in one case the frequency was observed to increase by a factor of four. Other investigators have demonstrated biphasic effects on NMDA open probability: specifically, low ethanol concentrations were found to increase open state probability while higher concentrations produced an decrease in probability (Lima-Landman and Albuquerque, 1989). On a few occasions, after addition of alkanol-containing solution to the bath, the opening frequency in the cell being recorded from steadily decreased until openings ceased altogether or dropped in frequency to less than 0.05 s-1; subsequent addition of control solution invariably failed to restore channel activity. This situation was attributed to a mechanical disturbance of the cell-attached patch, which could often be observed through the microscope as a visible displacement of the cell from the pipette; the data from cells in this category were discarded. The average effect of alcohol addition on frequency was a non-significant decrease of 11%. 105 Figure 15. Spectrofluorimetric analysis. Measurements of intracellular calcium increases, with the fluorescent dye fura-2, in a typical cell produced by bolus injections of 200 JUM NMDA into the recording chamber. The bar represents switching to medium containing 1 mM pentanol. Abbreviations: [Ca2+]i = intracellular calcium measurements with fura-2; T = time during experiment. 600r 1 mM pentanol 1 0 6 ^ 400|-5 c r<o1 200 10 "t _L 20 T (mini 30 40 2 . 3 . 2 . SPECTROFLUORIMETRIC ANALYSIS OF 7I-ALKANOL EFFECTS ON NMDA-INDUCED CHANGES IN [ C A 2 + ] i Bolus injections of 200 JUM NMDA into the recording chamber produced transient increases in the level of intracellular free calcium, as measured by increases in the fura-2 fluorescence signals, beginning from resting levels of between 50 to 100 nM Ca2+ to peak levels of 500-600 nM Ca2+. NMDA injections were usually spaced about 10 minutes apart, and the rises in [Ca2+]i produced by a series of NMDA injections seldom deviated from each other by more than 10%, indicating that neither depletion of intracellular calcium stores nor NMDA receptor desensitization were affecting sequential [Ca ]± measurements. Addition of either 3 mM butanol, 1 mM pentanol, or 20 /xM octanol to the recording bath significantly attenuated the responses to the NMDA injections. Results from a typical experiment are shown in Figure 15, where the NMDA-induced [Ca2+]-£ increase was attenuated by about 24% in the presence of 1 mM pentanol. The signal on the extreme right illustrates the fact that, as in the electrophysiological experiments, only partial recovery to pre-alcohol values could be obtained; in the case depicted in Figure 15, the NMDA-induced [Ca2+]i increase returned to 83% of control after washout of the alcohol. The mean percentage effects of the n-alkanols are summarized in part (b) of Table II. The alkanol effects as determined by the microspectrofluorometry are close, although slightly smaller, than those effects observed during the patch-clamp experiments, and indicate reasonable agreement between the two techniques as measures of NMDA ion channel function and its susceptibility to the presence of the n-alkanols. The only difference between the two sets of experiments was, curiously, the ineffectiveness of the long-chain (n=8) agent octanol in diminishing NMDA-induced [Ca2+]i increases while on the other hand producing significant decreases in mean NMDA unitary current mean open time at cell resting or slightly hyperpolarized potentials; this cause of this aberration remains unclear. While the possibility of octanol quenching the fura-2 signal was not investigated, the fact that the signal was only 11% smaller in octanol than in control (NMDA) whereas the shorter-chain alkanols decreased the signal more markedly suggests that octanol affected little the fura-2 quantum yield. 2.3.3. VOLATILE ANAESTHETICS Channel open time Table III shows the effects of the anaesthetics on NMDA channel open times. . Perfusion of the bath chamber with external bath solution equilibrated with 4% halothane or isoflurane during on-cell recording of unitary currents with either 40 /iM NMDA or 100 /zM D-homocysteic acid in the patch pipettes reduced the mean open time of the single-channel currents. Most cells responded to the anaesthetic 109 -60 mV 1M -80 mV -100 mV agonist («M) 40 NMDA 100 D-HCA 40 NMDA 100 D-HCA anaesthetic 4% isoflurane 4% isoflurane 4% halothane 4% halothane iop 76±7 69±6 74+24 83±9 Po 53±9 39±22 34+20 36±6 fe 84+9 91 + 16 80+28 95+10 £Q 47±11 53 ±25 49±22 80±18 ^ 86±5 106±1 1 83±11 94±17 Po 63 ±9 22±3 36+13 72 + 11 Table III. Effects of the volatile anaesthetics on mean open time and open probability. Each datum in the table represents the mean of three patches, except for those cells with 4% isoflurane and 40 NMDA as agonist (n=4). Means are percentages of control expressed ± S.E.M. Abbreviations: t0« = mean open time; Po = mean open channel probability. application in similar fashion as those cells studied with the n-alkanols, displaying a mean open time reduction of about 30%. A few cells, however, did not respond to the treatment, and it is unknown whether the anaesthetic concentration in the vicinity of the recorded cells was subeffective, due to evapouration of the volatile agent and/or sequestering of anaesthetic molecules into hydrophobic areas in the rest of the culture, or if the drugs were exerting variable effects on the different cells. We favour the former explanation: it was difficult to control evapouration of the anaesthetic from the bath during recording, even with the bath cover designed specifically for these experiments in place, while still allowing access to the cells by the patch pipette and avoiding physical contact or vibration of the mechanically fragile on-cell recording configuration. The incidence of cells not responding to anaesthetic application was higher in the earlier experiments where proper design of the bath cover and its placement had not yet been perfected; indeed, no effect of anaesthetic perfusion on the unitary currents could be demonstrated in pilot experiments where the bath had not been covered. The fact that an effect was achieved in many of the cells indicates that volatile anaesthetics are able to affect NMDA ion channel function and the inconsistency of the anaesthetic effects was most likely due to physicochemically-related conditions, such as level of bath solution and thus amount of surface area available for Ill Figure 16. Distributions of open and closed times. Histograms of open and closed times from typical cells in (a) control solution and (b) solution saturated with 4% isoflurane. The open time distribution in (a) consisted of 580 events and the fitted curve had a time constant of 1.8 ms, while the closed time histogram represented 540 events and was fit with two components of 2.8 and 59.0 ms. In (b), the open duration histogram contained 500 events and was fitted with a single component of 1.0 ms, while the closed duration histogram had 535 events and was fitted with 15.6 and 74.6 ms time constants. Curves, one-exponential and two-exponential for open and closed times, repectively, were least-squares fit using pSTAT 5.0. Number of events o ro o O •D ID 3 a. 3 ID 3 (0 CD o D) 3 CD Number of events Number of events 3-f H II 00 3 (A 1 1 o o Number of events I ! f ro 03 3 (A N) evapouration relative to volume of anaesthetic-containing bath solution, or total mass of hydrophobic tissue in the culture. Open state probability The volatile anaesthetics consistently produced decreases in the open state probability Po. Both components of Po, the mean open time and the rate of opening, were diminished by perfusion of the patches with anaesthetic-containing solution. Opening frequency usually decreased from between 50-75% of that recorded in control solution, although on occasion the opening frequency dropped to and maintained at a level of 30% of control. The mean reduction in Po was 49±7% for halothane and 53±6% for isoflurane and values ranged from 10% to 75% that recorded in control solution. The ability of the anaesthetics to reduce Po was unlike the alcohols in that it was observed in all cells recorded; another distinction between the actions of the volatile anaesthetics and the alkanols is that neither isoflurane nor halothane affected the amplitude of the unitary currents whereas, as noted above, in some cases alkanol treatment increased the size of the currents at each pipette potential. Open and closed duration histograms Figure 16 illustrates the effect of anaesthetic treatment with 4% isoflurane on the distributions of channel open times and closed intervals. Time constants of the single-exponential curves fitted to the open duration histograms were reduced by perfusion of volatile anaesthetic solution. The fast shut time components were usually increased by introduction of the anaesthetic-saturated solution into the recording bath, with little effect being observed on the late portion of the closed time curve. 2.4. DISCUSSION 2.4.1. EFFECTS OF THE n-ALKANOLS Channel conductance Application of the n-alkanols to cell-attached patches while recording single-channel currents activated by NMDA agonists included in the pipette did not alter the basic conductance properties of the NMDA ion channel. The increase in current amplitudes observed in some cells following alkanol perfusion was present at all voltages studied and was consistent with an effect of the alcohols to slightly hyperpolarize the neurones. The action of alcohols to occasionally produce small hyperpolarizations in central mammalian neurones has been previously observed (Gruol, 1982) and was also observed in some of the cells studied with whole-cell patch-clamp measurements of the cell resting potential. The mechanism for this action is poorly understood, but may be attributable to an increase in intracellular calcium (Daniell and Harris, 1988), through inhibition of Ca* binding to endoplasmic reticulum and depression of mitochondrial uptake (Quastel, 1952), which in turn would stimulate Ca-activated potassium conductances (Krnjevic, 1974). The magnitude of the hyperpolarizations and the amplitude increases were small and would be expected to have little effect on whole-cell NMDA currents. Reduction in mean open time The primary effect of the n-alkanols was to decrease the mean open time of the NMDA- and L-homocysteic acid-activated currents studied with the on-cell patch clamp configuration. The magnitude of the inhibition, which, with the n-alkanol conentrations used, varied from 25-35% of control, is comparable to that obtained in whole-cell studies examining the effects of alcohol inhibition of NMDA responses in hippocampal slices (Lovinger et al, 1990). It is a somewhat lesser inhibition than that observed in whole-cell current-clamp experiments using mouse hippocampal neurones (inhibition of 50-60% for similar concentrations) (Lovinger et al, 1989). The latter experiment may have overestimated the inhibitory effect of the alcohols on NMDA currents, as some of the inhibition may have been on currents mediated through kainate and guisgualate receptors which in this study were inhibited by 20-30% at the alkanol concentrations used. The reduction in mean open time of NMDA-induced currents by ethanol has been observed in outside-out patches (Lima-Landman and Albuquerque, 1989), although higher ethanol concentrations (greater than 90 mM) than used in the whole-cell studies were required in this study to achieve a similar level of open time inhibition as observed in our laboratory. It may be argued that some of the decrease in mean open time observed in our experiments was due to the small hyperpolarizing effect produced in some cells; however, this is unlikely since the magnitude of the hyperpolarization would be expected to result in at most a 7% decrease in mean open time, an effect much smaller than that observed and one that is in the range of the standard error associated with measurements of the mean open time (see Figures 6 and 13). Effects on other kinetic parameters Application of the n-alkanols produced variable effects on the channel opening frequency. While a small, non-significant decrease in frequency (11%) was the net effect between all the cells studied, a given cell responded with a definite change in opening rate, that is, the frequency behaviour in alkanol solution was quite distinct from that observed in control. Since some cells responded to alcohol perfusion with large increases in frequency, while still showing a decrease in mean open time, it is doubtful that the main effect of the fl-alkanols on NMDA channel kinetics is through encouragement of desensitization, as may be the case if proposals of ethanol actions on the desensitization-controlling strychnine-insensitive glycine site are correct (Rabe and Tabakoff, 1990). The effects of the alkanols on the opening rate can be compared to the work of Lima-Landman and Albuquerque (1989), where high ethanol concentrations reliably produced decreases in opening frequency while lower concentrations showed more variability but with a tendency to increase the opening rate. The observation of inconsistent frequency effects upon ethanol perfusion of outside-out patches at low concentrations, with decreases in frequency being always seen when ethanol concentrations were raised above 87 mM, suggests that more reliable effects on frequency, and thus open state probability, may have been obtained in the present study had higher concentrations of the J2-alkanols been used. Comparison of the closed time constants obtained in control and in the presence of alcohol revealed a tendency towards increased values for fast closed time components upon alkanol treatment. Just as the open state probability and channel opening frequency were observed to respond inconsistently to alcohol application, the fast components of fits to closed duration histograms responded somewhat variably as well, with an increase (usually accompanied by a net decrease in Po) being the most common observation. The fact that the slow closed time constants changed relatively little following alcohol perfusion indicates that the principal effects of the n-alkanols are on the activation kinetics of the NMDA ion channel and have little to do with agonist binding to receptor. Depression of NMDA-mduced [Ca* ]j_ increases The results obtained using the fluorescent dye fura-2 to determine alterations in NMDA-induced increases in internal calcium indicate an inhibitory response to the three .n-alkanols studied. The magnitude of the inhibition was close to that obtained in the patch-clamp experiments; however, whereas the fura studies measured changes in intracellular calcium (Dildy-Mayfield and Leslie, 1991), much of the signal recorded in the patch clamp experiments represents the passage of ions other than calcium through the patch (Asher and Nowak, 1988b). While the NMDA channel is unique among ligand gated ion channels for its high calcium permeability, with Pca/pNa estimations being about 10.6 (Mayer and Westbrook, 1987a), because of the non-selectivity of the NMDA-gated channel with respect to monovalent cations and because of the ionic gradients involved, the total charge passing through NMDA ion channels consists of about 10% Ca2+ in cells clamped at -60 mV (Mayer and Westbrook, 1987a); the rest is due to sodium, and potassium or cesium ions (Ascher and Nowak, 1988b; Mayer and Westbrook, 1987a,b; Nowak et al, 1984). As well, in the fura measurements a significant portion of the NMDA-induced rise in [Ca2+]i is due to release of the divalent cation from intracellular stores, as illustrated by the sensitivity of NMDA-activated [Ca2+]-[ to dantrolene, which only affects intracellular calcium stores (MacDonald et al, 1989). The slow washout in the present fluorescence experiments, as in the electrophysiological studies, is further evidence to suggest that relevant to the n-alkanols there is a hydrophobic site of action, storage of drug in lipophilic areas, or both. Proposed mechanism of action on NMDA currents The demonstration of effects on unitary currents recorded from cell-attached patches following application of the lipophilic n-alkanols to the bath indicate the effects on mean open time were mediated via equilibration of drug across the membrane patch, since the extracellular face of the receptor-ion channel complex was physically blocked by the patch pipette. The slow recovery of the currents following washout of alcohol from the bath also suggests mediation of the effects through a hydrophobic site, with sequestration of drug in hydrophobic plasma membranes also contributing to the prolonged open time attenuation following exchange of the bath with alcohol-free control solution. Saturable interaction of n-alkanols with hydrophobic membrane sites, as evidenced from the sigmoidal dose-response curves obtained in studies of the effects of the n-alkanol series on voltage-clamped NMDA currents (White et al, 1990; Lovinger et al, 1989), has been proposed, and the lack of further decreases in mean open time observed in our experiments when the concentration of butanol and pentanol were increased to 10 and 3 mM, respectively, replicated this saturation behaviour. As suggested above, effects of the n-alkanols on agonist binding to receptor (Gibb and Colquhoun, 1991) are minor, since slow shut time constants were relatively unaffected by alcohol addition. No fast transitions between open and closed states were introduced by the alkanol perfusion, and therefore the action of the alcohols was probably not due to rapid drug blockade of the open NMDA ion channel. The attenuation of mean open time was observed at all voltages studied and did not affect the slope of semi-log plots of mean open time versus patch potential, indicating that the inhibitory effects of the alcohols on the activation kinetics of the ion channel were not interacting with the intracellularly-mediated voltage modulation of channel behaviour. The correlation of the potency of n-alkanol action on NMDA-induced currents with chain length (Lovinger et al, 1989), and thus lipophilicity, strongly suggests a lipophilic site of action. While this site could be the membrane lipids themselves, since disturbance of their fluid properties would alter the environment of proteins embedded in the lipids (Goldstein, 1981), it should be recalled that intramembrane regions of membrane proteins consist chiefly of hydrophobic domains and thus are also potential sites for lipophilic drug action. Distinguishing between the two possible loci of action is, at this point in neurophysiological research technology, nearly impossible, since the structure of the lipid matrix near membrane proteins seems to be profoundly affected by interactions with components of the protein structure (Jost et al, 1973), and indeed may physically intercalate with protein domains such as a-helices. The closeness in spatial intramembrane location and chemical properties of membrane phospholipids and hydrophobic intramembrane protein domains means that further advances in research technology, such as the forecasted cloning of the NMDA receptor (Gonzales and Hoffman, 1991), are required in order to determine the 122 nature of the NMDA current response to the n-alkanols with more exactitude. Functional consequences Reduction of the mean open time of the NMDA ion channel supports the observations of alcohol-induced depression of whole-cell currents (Lovinger et al, 1989; White et al, 1989; Lovinger et al, 1990) in voltage-clamped neurones. The effect on single channel mean open time was observed with L-homocysteic acid, a substance which could play a role in endogenous NMDA-mediated neurotransmission (Do et al, 1988; Kilpatrick and Mozley, 1987; Cuenod et al, 1986); Similar findings were obtained with NMDA, which produces slightly longer mean open times than L-homocysteic acid in cell-attached control conditions as agonist. Studies employing other endogenous NMDA ligand candidates, particularly L-glutamate and aspartate, would support the hypothesis of alkanol effects on single-channel NMDA receptor-mediated currents in vivo. The inhibition of mean channel open time was independent of patch potential, indicating the effect on unitary NMDA currents would be manifested at various resting potentials and states of cell excitability. Unfortunately, the difficulty in studying NMDA currents at depolarized cell potentials due to low signal-to-noise ratios rendered us unable to study the effects of the alkanols on single-channel currents at the relatively depolarized cell potentials at which the NMDA system has been proposed to exert its greatest influence (Ascher and Johnson, 1989; Johnson and Ascher, 1988; Collingridge and Lester, 1989). However, the nonpolar nature of the alkanols makes any voltage-dependent drug action unlikely. The NMDA response to the longer chain alkanols may be extrapolated to ethanol (Lovinger et al, 1989). Inhibition of NMDA activity by alcohol provides a reasonable explanation for some of the effects experienced during alcohol intoxication, since these effects deal with systems in which NMDA neurotransmission plays a key role. The learning and memory deficits, ataxia, and analgesia produced by alcohol intoxication, and the developmental problems prevalent in children of alcoholic mothers (Ritchie, 1985), all suggest dysfunctioning of central functions in which NMDA receptors participate: long-term potentiation (Sastry et al, 1990; Collingridge et al, 1983), coordination of complex motor patterns (Dale, 1989), spinal nociception (Dickenson, 1990; Raigorodsky and Urea, 1990), and developmental synaptic plasticity (Eccles, 1989; Dale 1989), respectively. 2.4.2. VOLATILE ANAESTHETICS Influence on NMDA channel kinetics The most salient feature of the effects of perfusing solution equilibrated with halothane and isoflurane into the bath containing the cell-attached patches was to diminish the mean open time of NMDA agonist-activated currents. The drug actions were not dependent on potential across the patch and the channel conductance was unchanged from control. The effect of the anaesthetic on NMDA channel kinetics was not likely due to possible changes in the cell resting potential. Hyperpolarizations, accompanied by an increase in resting K+ conductance which may be activated by raised intracellular Ca2+ activity, due to isoflurane have been observed (Krnjevic, 1991), but this effect is mainly observed at higher anaesthetic doses. Similar to what was observed in the fura experiments with the n-alkanols, halothane depressed the Ca2+ response to glutamate in hippocampal neurones (Puil et al, 1990). The volatile anaesthetics tended to decrease the open state probability of the NMDA channel, and this observation may be the most relevant to synaptic transmission mediated through NMDA receptors. The NMDA response in the guinea-pig neocortical slice is depressed by isoflurane at 1.0 MAC (minimum alveolar concentration) (Puil and El-Beheiry, 1990). EPSP's recorded from guinea pig slices have been found to be sensitive to low concentrations of halothane (Krnjevic, 1991) and isoflurane can attenuate bicuculline-induced convulsant activity in neocortical neurones (El-Beheiry and Puil, 1989). The processes affected by the volatiles in these last two experiments are quite likely mediated by excitatory amino acid transmitters, and NMDA-receptor mediated currents may be involved; however, the depression of EPSP's and spiking activity reflects effects on presynaptic mechanisms, most notably transmitter release, as well as on postsynaptic membrane functions such as those studied in the present experiments. Functional relevance The attenuation of opening dwell time and frequency of unitary NMDA currents by the general anaesthetics halothane and isoflurane could suggest a role for the NMDA receptor in the maintenance of consciousness and arousal. However, many aspects of neurophysiological and whole brain function are depressed, or affected in some way, during general anaesthetics, and indeed, there are few areas in the central nervous system that are not depressed at some dose of volatile anaesthetic (Puil et al, 1991). Actual surgical anaesthesia quite likely involves a combination of effects on a number of neural systems. Nevertheless, effects on open time and frequency in this study were obtained at anaesthetic concentrations that are relatively lower than those required to produce effects on other aspects of neuronal function, such as membrane fluidity and responsiveness to the inhibitory neurotransmitter GABA (Puil and El-Beheiry, 1990). Concentration estimates of solutions saturated with 4% v/v volatile anaesthetic using vapourizers have appeared in the literature, and vary from a conservative estimate of 0.25 mM for isoflurane (obtained with similar equipment to that used in the present study and . t 1 Q determined using X^F nuclear magnetic resonance; c.f. Miu, 1988) to higher reports of 1.2 mM as determined by gas chromatography employing a different sampling technique (Yoshimura et al, 1985). These concentrations are at the low end of halothane concentrations required to produce brain tissue levels approximating those attained during surgical anaesthesia (Bazil et al, 1987) which in turn are close to the tissue concentrations obtained in in vitro preparations where techniques nearly identical to those used in the present study were utilized to introduce anaesthetic to the tissue (Bazil et al, 1987). This suggests that during surgical anaesthesia a similar depression of NMDA unitary currents may occur, and this effect may be a substantial component of the production of anaesthetic unconsciousness. The NMDA channel pore blocker, MK-801, has been found to reduce anaesthetic requirements for halothane in rats (Scheller et al, 1989), and the use of ketamine, an open-channel non-competitive NMDA blocker (MacDonald and Nowak, 1990), as an effective general anaesthetic in a variety of surgical and research situations (Smith and Aitkenhead, 1985) can be taken as evidence for the role of NMDA activity and its suppression in general anaesthesia. 3. SUMMARY AND CONCLUSIONS Unitary currents in cell-attached patches activated by NMDA agonists in nominally Mg2+-free solution were sensitive to the potential across the patch. Specifically, the mean channel open time showed exponential dependence on patch potential, such that increasing potential across the patch decreased the mean open time; therefore, some form of contact with intracellular elements is indicated for the expression of the open time modulation with potential. These results would suggest that the use of the excised patch configuration to study NMDA ion channel kinetics may produce misleading results. Mean open times obtained from including either of the two enantiomers for N-methylaspartate and homocysteic acid in the pipette were significantly different at most patch potentials studied in the cell-attached configuration. Analysis of closed time distributions corresponding to data gathered for each of the agonists revealed the difference in kinetics to most likely be through alterations in ion channel activation/deactivation, and not through agonist binding to the receptor. These observations serve as the putative identification of one of the factors contributing to distinctions between the relative potencies, and especially efficacies, of these agonists in producing whole-cell currents mediated through NMDA receptors. As well, the sensitivity of the ion channel activation kinetics to the structure of the ligand inducing channel activity suggests a tight association between the receptor recognition structure and the mechanism governing gating of the ion channel within the NMDA receptor-ion channel macromolecular complex. The observed dependence of channel opening frequency on agonist concentration in the mid- to high-micromolar range indicates that at these concentrations the level of receptor binding is an important determinant of NMDA ion channel activity. Late closed time constants were usually increased or separated into two larger components by decreasing the agonist concentration, while the early time components were unaffected by changes in concentration; a correlation between agonist binding and effects on the late, and not the early, closed time constants is thus proposed. The normal aliphatic alcohols butanol, pentanol, and octanol acted to decrease the mean open time of the NMDA ion channel independent of patch potential; the effect was primarily attributed to alterations in the activation/deactivation kinetics of the channel and not to attenuation of the lifetime of the bound agonist-receptor complex. The increase in inhibitory potency with n-alkanol chain length, as well as the observation of effects on mean open time even though the drugs were required to traverse the plasma membrane in order to gain access to the channel contained in the patch, both indicate that the alcohols' target of action is a lipophilic area associated with, and influencing the behaviour of, the NMDA ion channel. Measurements of intracellular calcium responses to NMDA application using the fluorescence indicator fura-2 confirmed the inhibitory effect of the n-alkanols on NMDA ion channel function. The volatile anaesthetics halothane and isoflurane were found to diminish the mean open time of unitary currents activated by NMDA agonists in a manner similar to that observed with the n-alkanols. The locus of action of the volatiles on the NMDA system in terms of one or more hydrophobic target sites was also found to be similar to the alcohols since the agents could produce an effect even when direct access to the ion channel was blocked by the patch pipette, leaving the plasma membrane surrounding the patch as the only pathway for the drug to the channel. While inhibitory effects on channel opening rate were somewhat more pronounced for the volatile anaesthetics than the inconsistently-acting J3-alkanols, the postulation of a qualitatively different mechanism of action on channel functioning would be overly premature. REFERENCES Allan, A.M., and Harris, R.A. Gamma-aminobutyric acid and alcohol actions: neurochemical studies of long sleep and short sleep mice. Life Sciences 39: 2005-15, 1986. Ascher, P., Bregestovski, P., and Nowak, L. tf-methyl-D-aspartate-activated channels of mouse central neurones in magnesium-free solutions. Journal of Physiology 399: 207-26, 1988. Ascher, P., and Johnson, J.W. The NMDA receptor, its channel, and its modulation by glycine, in The NMDA Receptor, Watkins, J.C., and Collingridge, G.L., eds.: pp. 109-21, IRL Press, Oxford, 1989. Ascher, P., and Nowak, L. Calcium permeability of the channels activated by 2V-methyl-D-aspartate (NMDA) in isolated mouse central neurones. Journal of Physiology 377: 35P, 1986. Ascher, P., and Nowak, L. Quisqualate- and kainate-activated channels in mouse central neurones in culture. Journal of Physiology 399: 227-45, 1988a. Ascher, P., and Nowak, L. The role of divalent cations in the tf-methyl-D-aspartate reponses of mouse central neurones in culture. Journal of Physiology 399: 247-66, 1988b. Atkinson, R.S., and Boulton, T.B., eds. The History of Anaesthesia, Royal Society of Medicine Services, London, 1987. Banker, G.A., and Cowan, W.M. Rat hippocampal neurons in dispersed cell culture. Brain Research 126: 397-425, 1977. Baudry, M., Kramer, K., Fagni, L., Recasens, M., and Lynch, G. Classification and properties of acidic amino acid receptors in hippocampus II. Biochemical studies using a sodium efflux assay. Molecular Pharmacology 24: 222-8, 1983. Bazil, C.W., Raux, M.E., Yudell, S., and Minneman, K.P. Equilibration of halothane with brain tissue in vitro: comparison to brain concentrations during anaesthesia. Journal of Neurochemistry 49: 952-8, 1987. Choi, D.W. Ionic dependence of glutamate neurotoxicity. Journal of Neuroscience 7(2): 369-79, 1987. Choi, D.W. Glutamate neurotoxicity and diseases of the nervous system. Neuron 1(8): 623-34, 1988. Choi, D.W., Koh, J.Y., and Peters, S. Pharmacology of glutamate neurotoxicity in cortical cell culture: attenuation by NMDA antagonists. Journal of Neuroscience 8(1): 185-96, 1988. Collingridge, G.L., Kehl, S.J., and McLennan, H. Excitatory amino acids in synaptic transmission, in the Schaeffer collateral-commissural pathway of the rat hippocampus. Journal of Physiology 334: 33-46, 1983. Collingridge, G.L., and Lester, R.A. Excitatory amino acid receptors in the vertebrate central nervous system. Pharmacological Reviews 40(2): 143-210, 1989. Colquhoun, D., and Sigworth, F.J. Statistical analysis and fitting of single-channel records, in Single-channel Recording, Sakmann, B., and Neher, E., eds.: pp. 191-263, Plenum Press, New York, 1983. Cuenod, M., Do, K.Q., Matute, C., and Streit, P. Identification of pathways for acidic amino acid transmitters and search for new candidates: sulphur-containing amino acids, in Excitatory Amino Acids, Roberts, P.J., Storm-Mathisen, J., and Bradford, H.F., eds.: pp. 117-29, MacMillan, Toronto, 1986. Cuenod, M., Do, K.Q., and Streit, P. Homocysteic acid as an endogenous excitatory amino acid. Trends in Pharmacological Sciences 11: 477-8, 1990. Cull-Candy, S.G.M., and Usowicz, M.M. Multiple-conductance channels activated by excitatory amino acids in cerebellar neurones. Nature 325: 525-28, 1987. Cull-Candy, S.G.M., and Usowicz, M.M. On the multiple-conductance single channels activated by excitatory amino acids in large cerebellar neurones of the rat. Journal of Physiology 415: 555-82, 1989. Curry, K., Magnuson, D.S.K., McLennan, H., and Peet, M.J. Excitation of rat hippocampal neurones by the stereo-isomers of cis-and tra7js-l-amino-l,3-cyclopentane dicarboxy late. Canadian Journal of Physiology and Pharmacology 65: 2196-2201, 1987. Curtis, D.R., and Watkins, J.C. Acidic amino acids with strong excitatory actions on mammalian neurones. Journal of Physiology 166: 1-14, 1962. Dale, N. The role of NMDA receptors in synaptic integration and the organization of complex neural patterns, in The NMDA Receptor, Watkins, J.C., and Collingridge, G.L., eds.: pp. 93-107, IRL Press, Oxford, 1989. Daniell, L.C., and Harris, R.A. Neuronal intracellular calcium concentration are altered by anesthetics: relationship to membrane fluidization. Journal of Pharmacology and Experimental Therapeutics 245(1): 1-7, 1988. Davies, J., and Stanley, M.D.A. Specificity of excitatory amino acid agonists and antagonists, in Excitatory Amino Acids in Health and Disease, Lodge, D., ed.: pp. 47-62, John Wiley & Sons, Toronto, 1988. Dickenson, A.H. A cure for wind up: NMDA receptor antagonists as potential analgesics. Trends in Pharmacological Sciences 11(8): 307-9, 1990. Dildy-Mayfield, J.E., and Leslie, S.W. Mechanism of inhibition of 2V-methyl-D-aspartate-stimulated increases in free intracellular Ca2 concentration by ethanol. Journal of Neurochemistry 56(5): 1536-43, 1991. Dingledine, R., McBain, C.J., and McNamara, J.O. Excitatory amino acid receptors in epilepsy. Trends in Pharmacological Sciences 11(8): 334-8, 1990. Do, K.Q., Herrling, P., Streit, P. and Cuenod, M. Homocysteate: a rival for aspartate?, in Excitatory Amino Acid Transmission, Hicks, P., Lodge, D., and McLennan, H., eds.: pp. 153-60, Alan R. Liss, Toronto, 1987. Do, K.Q., Herrling, P., Streit, P. and Cuenod, M. Release of neuroactive substances: homocysteic acid as an endogenous agonist of the NMDA receptor. Journal of Neural Transmission 72: 185-90, 1988. Eccles, J. Foreword, in The NMDA Receptor, Watkins, J.c. and Collingridge, G.L., eds.: pp. ix-xi, IRL Press, Oxford, 1989. Eger, E.I., ed. Nitrous Oxide. Elsevier, New York, 1985. El-Beheiry, H., and Puil, E. Postsynaptic depression induced by isoflurane and Althesin in neocortical neurons. Experimental Brain Research 75: 361-8, 1989. Eliany, M. Alcohol in Canada, Health and Welfare Canada, 1989. Errington, M.L., Lynch, M.A., and Bliss, T.V.P. Long-term potentiation in the dentate gyrus: induction and increased glutamate release are blocked by D - ( - ) -aminophosphonovalerate. //euroscience £0(1) : 279-84, 1987. Forman, S.A., and Miller, K.W. Molecular sites of anesthetic action in postsynaptic membranes. Trends in Pharmacological Sciences 10: 447-52, 1989. Forsythe, I.D., Westbrook, G.L., and Mayer, M.L. Modulation of excitatory synaptic transmission by glycine and zinc in cultures of mouse hippocampal neurons. Journal of Neuroscience 8(10): 3733-41, 1988. Franciolini, F., and Nonner, W. Anion and cation permeability of a chloride channel in rat hippocampal neurons. Journal of General Physiology 90: 453-78, 1987. Franks, N.P., and Lieb, W.R. Trends in Pharmacological Sciences 8: 169-74, 1987. Gibb, A.J. Characteristics of channel openings activated by low concentrations of glutamate in cells dissociated from adult rat hippocampus. Journal of Physiology 412: 39P, 1989. Gibb, A.J., and Colquhoun, D. Glutamate activation of a single NMDA receptor-channel produces a cluster of channel openings. Proceeding of the Royal Society of London Series B 243: 39-45, 1991. Godin, D.V., and del Vicario, G. Molecular aspects of inhalational anaesthetic interaction with excitable and non-excitable membranes. Canadian Anaesthetists' Society Journal 28(3): 201-9, 1981. Goldstein, D.B. Biophysical pharmacology: alcohol effects on biomembranes, in Pharmacology of Alcohol, pp. 48-64, Oxford University Press, Oxford, 1983. Gonzales, R.A. NMDA receptors excite alcohol research. Trends in Pharmacological Sciences 11: 137-9, 1990. Gonzales, R.A., and Hoffman, P.L. Receptor-gated ion channels may be selective CNS targets for ethanol. Trends in Pharmacological Sciences 12: 3-5, 1991. Gruol, D.L. Ethanol alters synaptic activity in cultured spinal cord neurons. Brain Research 243: 25-33, 1982. Grynkiewicz, G., Poenie, M., and Tsien, R.Y. A new generation of Ca2+ indicators with greatly improved fluorescence properties. Journal of Biological Chemistry 260: 3440-50, 1985. Headley, P.M., and Grillner, S. Excitatory amino acids and synaptic transmission: the evidence for a physiological function. Trends in Pharmacological Sciences 11(8): 205-11, 1990. Hamill, O.P., Marty, A., Neher, E., Sakmann, B., and Sigworth, F.J. Improved patch-clamp technigues for high-resolution current recording from cells and cell-free membrane patches. Pflugers Archiv 391: 85-100, 1981. Howe, J.R., Colquhoun, D., and Cull-Candy, S.G. On the conductance of large-conductance glutamate-receptor ion channels in rat cerebellar granule neurons. Proceedings of the Royal Society of London Series B 233: 407-22, 1988. Howe, J.R., Cull-Candy, S.G., and Colquhoun, D. Currents through single glutamate-receptor channels in outside-out patches from rat cerebellar granule cells. Journal of Physiology 432: 143-202, 1991. Izquierdo, I. Role of NMDA receptors in memory. Trends in Pharmacological Sciences 12(4): 128-9, 1991. Jahr, C.E., and Stevens, C.F. Glutamate activates multiple single channel conductances in hippocampal neurons. Nature 325: 522-5, 1987. Jahr, C.E., and Stevens, C.F. A quantitative description of NMDA receptor-channel kinetic behavior. Journal of Neuroscience 10(6): 1830-7, 1990. Johnson, J.W., and Ascher, P. The NMDA receptor and its channel. Modulation by magnesium and by glycine, in Excitatory Amino Acids in Health and Disease, Lodge, D., ed.: pp. 143-64, John Wiley & Sons, Toronto, 1988. Johnson, J.W., and Ascher, P. Glycine potentiates the NMDA response in cultured mouse brain neurons. Nature 325: 529-31, 1987. Jost, P.C., Griffith, O.H., Capaldi, R.A., and Vanderkooi, G. Evidence for boundary lipid in membranes. Proceedings of the National Academy of Sciences of the United States of America 70: 480-4, 1973. 135 Katz, Y., and Simon, S.A. Physical parameters of the anesthetic site. Biochimica et Biochimica Acta 471: 1-15, 1977. Kilpatrick, I.e., and Mozley, L.S. Detection of excitatory sulphur-containing amino acids in the rat brain, in Excitatory Amino Acid Transmission, Hicks, P., Lodge, D., and McLennan, H., eds.: pp. 181-4, Alan R. Liss, Toronto, 1987. Konnerth, A., Keller, B.U., Ballanyi, K., and Yaari, Y. Voltage sensitivity of NMDA-receptor mediated postsynaptic currents. Biophysical Journal 57: 128a, 1990. Krnjevic, K. Central actions of general anaesthetics, in Molecular Mechanisms in General Anaesthesia, Halsey, M.J., Millar, R.A., and Sutton, J.A., eds.: pp. 65-89, Churchill Livingstone, Edinburgh, 1974. Krnjevic, K. Cellular mechanisms of anesthesia, in Molecular and Cellular Mechanisms of Alcohol and Anesthetics, Rubin, E., Miller, K.W., and Roth, S.H., eds.: pp. 1-16, New York Academy of Sciences, New York, 1991. Lambert, J.O.C., Jones, R.S.G., Andreasen, M., Jensen, M.S., and Heinemann, U. The role of excitatory amino acids in synaptic transmission in the hippocampus. Comparative Biochemistry and Physiology 93A(1): 195-201, 1989. Lester, R.A.J., Clements, J.D., Westbrook, G.L., and Jahr, C.E. Channel kinetics determine the time course of NMDA receptor-mediated synaptic currents. Nature 346: 565-7, 1990. Lima-Landman, M.T.R., and Albuquerque, E.X. Ethanol potentiates and blocks NMDA-activated single-channel currents in rat hippocampal pyramidal cells. Federation of European Biochemical Societies Letters 24Z(1): 61-7, 1989. Linder, T.M., Pennefather, P., and Quastel, D.M.J. The time course of miniature endplate currents and its modification by receptor blockade and ethanol. Journal of General Physiology 85(5): 435-68, 1984. Lodge, D. Modulating glutamate pharmacology. Trends in Pharmacological Sciences 8(7): 243-6, 1987. 136 Lovinger, D.M., White, G., and Weight, F.F. NMDA receptor-mediated synaptic excitation selectively inhibited by ethanol in hippocampal slice from adult rat. Journal of Neuroscience 10: 1372-9, 1990. Lovinger, D.M., White, G., and Weight, F.F. Ethanol inhibits NMDA-activated ion current in hippocampal neurons. Science 243: 1721-4, 1989. MacDermott, A.B., Mayer, M.L., Westbrook, G.L., Smith, S.J., and Barker, J.L. NMDA-receptor activation increases cytoplasmic calcium concentration in cultured spinal cord neurones. Nature 321: 519-22, 1986. MacDonald, J.F., and Nowak, L. Mechanisms of blockade of excitatory amino acid receptor channels. Trends in Pharmacological Sciences il(4): 167-72, 1990. MacDonald, J.F., and Wojtowicz, J.M. The effects of L-glutamate and its analogues upon the membrane conductance of central murine neurones in culture. Canadian Journal of Physiology and Pharmacology 60(3): 282-96, 1982. MacDonald, J.F., Mody, I., Salter, M.W., Pennefather, P., and Schneiderman, J.H. The regulation of NMDA receptors in the central nervous system. Progress in Neuro-psychopharmacology and Biological Psychiatry 13: 481-8, 1989. MacDonald, J.F., Miljkovic, Z., and Pennefather, P. Use-dependent block of excitatory amino acid currents in cultured neurons by ketamine. Journal of Neurophysiology 58(2): 251-65, 1987. Mayer, M.L., and Westbrook, G.L. Permeation and block of N-methyl-D-aspartic acid receptor channnels by divalent cations in mouse cultured central neurones. Journal of Physiology 399: 501-27, 1987a. Mayer, M.L., and Westbrook, G.L. The physiology of excitatory amino acids in the vertebrate central nervous system. Progress in Neurobiology 28: 197-276, 1987b. Mayer, M.L., Westbrook, G.L., and Guthrie, P.B. Voltage-dependent block by Mg2+ of NMDA responses in spinal cord neurones. Nature 309(5965): 261-3, 1984. McLarnon, J.G., and Curry, K. Quinolinate activation of N-methyl-D-aspartate ion channels in rat hippocampal neurons. Neuroscience Letters 116: 341-46, 1990a. McLarnon, J.G., and Curry, K. Single channel properties of the N-methyl-D-aspartate receptor channel using NMDA and NMDA agonists: on-cell recordings. Experimental Brain Research 82: 82-8, 1990b. McLennan, H. The pharmacological characterization of excitatory amino acid receptors, in Excitatory Amino Acids in Health and Disease, Lodge, D., ed.: pp. 1-13, John Wiley & Sons, Toronto, 1988. Miu, P. Isoflurane induced impairment of synaptic transmission in hippocampal neurons of the guinea pig in vitro. Master's thesis, University of British Columbia, 1988. Monaghan, D.T., Bridges, R.J., and Cotman, C.W. The excitatory amino acid receptors: their classes, pharmacology, and distinct properties in the function of the central nervous system. Annual Review of Pharmacology and Toxicology 29: 365-402, 1989. Morris, R.G.M. Elements of a hypothesis concerning the participation of hippocampal NMDA receptors in learning, in Excitatory Amino Acids in Health and Disease, Lodge, D., ed.: pp. 297-320, John Wiley & Sons, Toronto, 1988. Morris, R.G.M., Davis, S., and Butcher, S.P. The role of NMDA receptors in learning and memory, in The NMDA receptor, Watkins, J.C., and Collingridge, G.L., eds: pp. 137-151, IRL Press, Oxford, 1989. Mullins, L.J. Some physical mechanisms in narcosis. Chemical Reviews 54: 289-323, 1954. Nandini-Kishore, S.G., Kitajima, Y., and Thompson, G.A. Membrane fluidizing effects of the general anesthetic methoxyflurane elicit an acclimation response in Tetrahymena. Biochimica et Biophysica Acta 471: 157-61, 1977. Nowak, L., and Ascher, P. Divalent cation effects on NMDA-activated channels can be described as Mg-like or Ca-like. Society for Neuroscience Abstracts 11: 953, 1985. Nowak, L., Bregestovski, P., Ascher, P., Herbet, A., and Prochiantz, A. Magnesium gates glutamate-activated channels in mouse central neurones. Nature 307: 462-5, 1984. Oliver, M.W., Shacklock, J.A., Kessler, M., Lynch, G., and Baimbridge, K.G. The glycine site modulates NMDA-mediated changes of intracellular free calcium in cultures of hippocampal neurons. Neuroscience Letters 114: 197-202, 1990. Olverman, H.J., Jones, A.W., and Watkins, J.C. [3H]D-2-amino-5-phosphonopentanoate as a ligand for tf-methyl-D-aspartate receptors in the mammalian central nervous system. Neuroscience 2_6(1) : 1-15, 1988a. Olverman, H.J., Jones, A.W., Mewett, K.N., and Watkins, J.C. Structure/activity relations of tf-methyl-D-aspartate receptor ligands as studied by their inhibition of [ H]D-2-amino-5-phosphonopentanoic acid binding in rat brain membranes. Neuroscience 26(1): 17-31, 1988b. Olverman, H.J., and Watkins, J.C. NMDA agonists and competitive antagonists, in The NMDA Receptor, Watkins, J.C, and Collingridge, G.L., eds.: pp. 19-36, IRL Press, Oxford, 1989. Peet, M.,J., Curry, K., Magnuson, D.S., and McLennan, H. Ca -dependent depolarization and burst firing of rat CA1 pyramidal neurones induced by tf-methyl-D-aspartate and quinolinic acid: antagonism by 2-5-phosphonovaleric and kynurenic acids. Canadian Journal of Physiology and Pharmacology 64.: 163-8, 1986. Perkins, M.N., and Stone, T.C. Pharmacology and regional variations of quinolinic acid-evoked excitations in the rat central nervous system. Journal of Pharmacology and Experimental Therapeutics 226: 551-7, 1983. Poenie, M., Alderton, J., Steinhardt, R., and Tsien, R. Calcium rises abruptly and briefly throughout the cell at the onset of anaphase. Science 233: 886-9, 1986. Puil, E., and El-Beheiry, H. Anaesthetic suppression of transmitter actions in neocortex. British Journal of Pharmacology 101: 61-6, 1990. Puil, E., El-Beheiry, H., and Baimbridge, K.G. Anesthetic effects on glutamate-stimulated increase in intraneuronal calcium. Journal of Pharmacology and Experimental Therapeutics 255(3): 955-61, 1990. Quastel, J.H. Biochemical aspects of narcosis. Current Research in Anaesthesia and Analgesia 31: 151, 1952. Rabe, C.S., and Tabakoff, B. Glycine site-directed agonists reverse the actions of ethanol at the tf-methyl-D-aspartate receptor. Molecular Pharmacology 38: 753-7, 1990. Raigorodsky, G., and Urea, G. Involvement of N-methyl-D-aspartate receptors in nociception and motor control in the spinal cord of the mouse: behavioural, pharmacological, and electrophysiological evidence. Neuroscience 36(3): 601-10, 1990. Ransom, R.W., and Oeschenes, N.L. Polyamines regulate glycine interaction with the N-methyl-D-aspartate receptor. Synapse 5: 294-8, 1990. Reynolds, I.J., and Miller, R.J. Multiple sites for the regulation of the tf-methyl-D-aspartate receptor. Molecular Pharmacology 33: 581-4, 1988. Reynolds, I.J., Murphy, S.N., and Miller, R.J. 3H-labeled MK-801 binding to the excitatory amino acid receptor complex from rat brain is enhanced by glycine. Proceedings of the National Academy of Sciences of the United States of America 84: 7744-48, 1987. Reynolds, I.J., Rush, E.A., and Aizenman, E. Reduction of NMDA receptors with dithiothreitol increases [3H]-MK-801 binding and NMDA-induced Ca2+ fluxes. British Journal of Pharmacology 101: 178-82, 1990. Ritchie, J.M. The aliphatic alcohols, in The Pharmacological Basis of Therapeutics, Gilman, A.G., Goodman, L.S., and Gilman, A., eds: pp. 376-90, Collier Macmillan Canada, Toronto, 1985. Sastry, B.R., Maretic, H., Morishita, W., and Xie, Z. Modulation of the induction of long-term potentiation in the hippocampus, in Excitatory Amino Acids and Neuronal Plasticity, Advances in Experimental Medicine and Biology 268, Ben-Ari, Y., ed.: pp. 377-86, Plenum Press, New York, 1990. Scheller, M.S., Zornow, M.H., Fleischer, J.E., Shearman, G.T., Greber, T.F. The noncompetitive tf-methyl-D-aspartate receptor antagonist, MK-801 profoundly reduces volatile anesthetic requirements in rabbits. Neuropharmacology 28(7): 677-81, 1989. Schmidt, W.J., Bubser, M., and Hauber, W. Excitatory amino acids and Parkinson's disease. Trends in Neurosciences 13(2): 48, 1990. Seeman, P. The membrane actions of anesthetics and tranquilizers. Pharmacological Reviews 24 (4) : 583-655, 1972. Seeman, P. The membrane expansion theory of anaesthesia: direct evidence using ethanol and a high-precision density meter. Experientia i5(7): 759-60, 1974. Sigworth, F.J., and Sine, S.M. Data transformations for improved display and fitting of single-channel dwell time histograms. Biophysical Journal 52: 1047-54, 1987. Smith, G., and Aitkenhead, A.P., eds. Textbook of Anaesthesia, Churchill Livingstone, New York, 1985. Sun, A.Y. Biochemical and biophysical approaches in the study of ethanol-membrane interaction, in Biochemistry and Pharmacology of Ethanol, Volume 2, Majchrowicz, E., and Noble, E.P., eds.: pp. 243-67, Plenum Press, New York, 1979. Suzdak, P.D., Schwartz, R.D., Skolnik, P., and Paul, S.M. Ethanol stimulates gamma-aminobutyric acid receptor-mediated CI" transport in rat brain synaptoneurosomes. Proceedings of the National Academy of Sciences of the United States of America 83: 4071-5, 1986. Traynelis, S.F., and Cull-Candy, S.G. Proton inhibition of 2tf-methyl-D-aspartate receptors in cerebellar neurons. Nature 345: 347-50, 1990. Trudell, J.R., Hubbell, W.L., and Cohen, E.N. The effect of two inhalation anesthetics on the order of spin-labeled phospholipid vesicles. Biochimica et Biophysica Acta 291: 321-7, 1973. van Rossum, J.M., and de Bie, J.E. Chaos and illusion. Trends in Pharmacological Sciences 11(10): 379-83, 1991. Watkins, J.C. The NMDA receptor concept: origins and development, in The NMDA Receptor, Watkins, J.C, and Collingridge, G.L., eds.: pp. 1-17, IRL press, Oxford, 1989. Watkins, J.C, and Evans, R.H. Excitatory amino acid transmitters. Annual Review of Pharmacology and Toxicology 21: 165-204, 1981. Watkins, J.C, Krogsgaard-Larsen, P., and Honore, T. Structure-activity relationships in the development of excitatory amino acid receptor agonists and competitive antagonists. Trends in Pharmacological Sciences 11: 25-33, 1990. Watkins, J.C, and Olverman, H.J. Structural requirements for activation and blockade of EAA receptors, in Excitatory Amino Acids in Health and Disease, Lodge, D., ed.: pp. 13-45, John Wiley & Sons, Toronto, 1988. White, G., Lovmger, D.M., and Weight, F.F. Ethanol inhibits NMDA-activated current but does not alter GABA-activated current in an isolated adult mammalian neuron. Brain Research 507: 332-6, 1990. Wright, J.M., and Nowak, L.M. Voltage dependent change in probability of opening (nPo) of tf-methyl-D-aspartate (NMDA) channels in Mg free solutions. Biophysical Journal 5_7: 128a, 1990. Wyrwicz, A.M., Li, Y.-E., Schofield, J.C., and Burt, C.T. Multiple environments of fluorinated anesthetics in intact tissues observed with 19F NMR spectroscopy. Federation of European Biochemical Societies Letters 162(2): 334-8, 1983. Yoshimura, M., Higashi, H., Fujita, S., and Shimoji, K. Selective depression of hippocampal inhibitory postsynaptic potentials and spontaneous firing by volatile anesthetics. Brain Research 340: 363-8, 1985. Zeise, M.L., Knopfel, T., and Zieglgansberger, W. (±)-B-parachlorophenylglutamate selectively enhances the depolarizing response to L-homocysteic acid in neocortical neurons of the rat: evidence for a specific uptake system. Brain Research 443: 373-6, 1988. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items