SYNAPSE 7:207-215 (1991)
MK-801 Inhibition of Nicotinic Acetylcholine Receptor Channels MARIANO AMADOR AND JOHN A. DAN1 Department of Molecular Physiology and Biophysics, Baylor College of Medicine, Houston, Texas 77030
KEY WORDS
NMDA receptor, Noncompetitiveinhibition, Open channel block, Neuromuscular junction
ABSTRACT MK-801 is a potent inhibitor of the NMDA subtype of glutamate receptors. Single-channel and macroscopic currents indicate that MK-801 also inhibits nicotinic acetylcholine receptors (nAChRs).MK-801 does not significantly increase desensitization of the nAChRs or compete for the ACh binding site. Although there is a slight inhibition of the closed nAChR, the main action of MK-801 is to enter and block the open channel. The voltage dependence for block is consistent with a single binding site within the channel that is 50%of the way through the membrane field. The IC,, for block is 3 pM at - 70 mV for currents induced by 0.5 pM ACh. The data from both single-channel and macroscopic currents can be used to estimate a & (0) of 7 pM, which is about 40 times higher than the &(O) for MK-801 binding to the NMDA receptor. The relative potency of tricyclic compounds like MK-801for various neurotransmitter systems points out that the pharmacologic action of these drugs could involve complicated interactions in vivo.
INTRODUCTION Many organic cations block the nicotinic acetylcholine receptor (nAChR) channel (reviewed by Changeux and Revah, 1987). Although other binding sites may exist (Papke and Oswald, 1989), voltage-dependent noncometitive inhibitors often exert their primary action by glocking the open pore. Neher and Steinbach (1978) provided the first sin le-channel records showing open channel block of nA hRs. At low concentrations, the uaternar ammonium derivative of lidocaine, QX-222 Lowed w at are now classic characteristics of open channel block. Single-channel openings were shorter and occurred in bursts containing brief closures. The voltage-dependence of the block indicated that QX-222 entered the pore. Site-directed mutations revealed that X-222 inter2 (Charnet I acts with serine residues in the a-helix,% et al., 1990; Leonard et al., 1988). In addition to the results with QX-222, other studies using site-directed mutations (Imoto et al., 1986, 1988) provided evidence that the M2 a-helix lines the pore (reviewed by Dani, 1989a). Two other noncom etitive blockers, chlorpromazine and tri henylmet ylphosphonium, photoaffinit label the 2 a-helix (Giraudat et al., 1986; Hucho et a ., 1986). Therefore, there is direct evidence that these three noncompetitive inhibitors enter the open pore when they block the nAChR channel. Some noncompetitive inhibitors of the nAChR also inhibit the N-methyl-D-aspartate (NMDA) subset of lutamate receptors (reviewed by Kemp, 1987).The best anown is phencyclidine (PCP), which appears to have a high affinity site within the NMDA receptor channel (Anis et al., 1983; Kloog et al., 1988) and within the nAChR channel (Agua o et al., 1986; Oswald et al., 1984; Papke and Oswa d, 1989). Other noncom etitive inhibitors of both channels include mecamy amine,
8
K
P
K
2
P
0 1991 WILEY-LISS,INC.
P
empidine (O’Dell and Christensen, 1988), tetrahy8roaminoacridine (Nilsson et al., 1988) and chlor romazine (Heidmann et al., 1983; Reynolds and Mi ler, 1988a). The dibenzocyclohepteneimine MK-801 is a otent noncompetitive inhibitor of the NMDA receptor YWong et al., 1986). Evidence indicates that MK-801blocks the open channel. Both binding and unbinding of MK-801 require an open NMDA rece tor channel, and recovery from block is volta e-depen ent (Huettner and Bean, 1988).In addition, &g+ +,which blocks the o en channel (Nowak et al., 1984), relieves MK-801 bloc (Huettner and Bean, 1988) possibl by electrostatically repelling MK-801 from its intra-ciannel site or by an allosteric interaction (Reynolds and Miller, 1988b).PCP and similar compounds compete with MK-801, indicatin that they share the same site within the channel ( ong et al., 1986; Reynolds and Miller, 1988a).Recent preliminar results indicate that MK-801 also inhibits the n A d R (Halliwell et al., 1989; Kavanaugh et al., 1989; Ramoa et al., 1989; Amador et al., 1990). In these ex eriments, we determined the effect of external MK-{Ol on the nAChR by collecting macroscopic and single-channel currents. The main mechanism of inhibition is open channel block. The singlechannel currents indicate that the open time of the nAChR channel decreases, but the single channel conductance is not affected. The results show that MK-801 inhibits nAChR channels, and the results provide evidence for the mechanism of inhibition. The conclusions are valuable for assessin the otential clinical value of MK-801 as a specific bloger ofthe NMDA receptors.
P
!
E
&v
Received June 6,1990; accepted in revised form August 23,1990.
208
M. AMADOR AND J.A. DAN1
MATERIALS AND METHODS Tissue culture Nicotinic AChR channels were studied in clonal BCSH-1 mouse muscle cells. The clonal BCSH-1 cells were maintained usin standard tissue culture techni ues (Dani, 1989b; ine and Steinbach, 1984; Sine an% Taylor, 1979). The cell line was maintained in Dulbecco’s Modified Ea le medium (DMEM; Hazleton, Lenexa, KS; Gibco, Gait ersburg, MD) containin 10% to 20% heat-inactivated fetal bovine serum (FBS; ibco Gaithersburg, MD; Hyclone, Logan, UT), 100 units/mi of penicillin, 100 kg/ml of stre tomycin (Hazleton, Lenexa, KS) and 2 mM glutamine PHazleton, Lenexa, KS) in an incubator at 37°C with 5% CO . For patch-clamp experiments, the cells were late8 onto cover gIass (BellcoGlass Co., Vineland, N J a n d were maintained in DMEM with 0.5% FBS. The cells were fed every other day, and the were used from 5 t o 20 days after being plated onto t e cover glass.
8 5
8
K
1
Y I
1 PM
T
r
Patch-clamp techniques Whole-cell and single channel nAChR currents were measured in BC3H-1 cells usin standard patch-clamp techniques (Dani, 1989b; Hami et al., 1981; Sakmann and Neher, 1983). Patch pipettes were pulled with a two-stage electrode puller (PP-83; Narishige USA, Greenvale, NY) using 100 ~1disposable pipettes (Drummon Scientific, Broomall, PA) or Kovar 7052 glass (Garner Glass, Claremont, CA). To decrease ca acitive noise, the ipettes were coated with S 1 ar silicon elastomer ( ow Corning Corp., Midland, i&fior olystyrene Q-Dope (GC Electronic, Rockford, IL). he tips were polished immediately before the ex eriment using a microforge (MF-83; Narishige USA, reenvale NY). The platinum wire of the microforge was coated with Kovar glass. Currents were amplified and filtered using an Axopatch 1B (Axon Instruments, Burlingame, CA). Currents were digitally sampled with a 1Zbit analog to
fi
a .e
1;
cp
10 pM
T
I
30
100
i -
100 p M
T
l200pA L
20s
I
d
20s
100 t
c Q)
L L
3 0
SO 60
U Q)
.-
40
E L
20
N
0
0
z 0 1
3
10
[ M K - 8 0 1 ] (pM) Fig. 1. Dose-response relationship for inhibition by MK-801. Three macroseopic currents activated by0.5 pM ACh a t a ho ding otential of - 70 mV are shown. The ACh was applied at the downwartfarrow and was washed off at the upward arrow. MK-801 was a lied at a concentration of 1, 10, or 100 pM as indicated by the solic?\orizontal
bar. Dose-res onse curves are shown at -70 mV (W and at -20 mV ( 0 ) . The curves i t to the data are based on a single blocking site as described in the text. Each data point is the mean of 3 to 18 measurementsandisshownwithitsSE.
209
MK-801 INHIBITION OF THE A C h R CHANNEL
A
B +50mV
10.0
1 .o
J
-70mV
/250pA
0.1 -80
10s
-40
0
40
Vh ( m V )
Fig. 2. Voltage dependence of MK-801 block. A Macrosco ic cur rents were induced by 0.5 M ACh. During the current, 10 pM RK-80; was ap lied with the ACK for about 20 seconds. The current a t the end of tge 20 second application of MK-801 (Imk)was compared with the current after recove from block (Ic). The dotted horizontal line indicates the baselineyefore ACh a plication. The inhibition was more complete a t negative potentials. The voltage dependence of the
MK-801 block is plotted according to equation 1,which is based on a single binding site within the membrane field. Each data point is the mean of 9 to 21 measurements and is shown with its SE. They intercept indicates that the IC ,(O) equals 9.7 pM, and the slope of the !ine indicates that the binkng site is halfway through the membrane field (i.e., 6 = 0.5).
digital converter controlled by a ersonal computer (AxonInstruments, Burlingame, CAP and were saved on the hard disk. The computer system also delivered the holding otentials through a 12-bit digital to analog output. eurrents were sampled every 0.1 ms and filtered at 2 kHz. Single-channel currents were studied using outside/ out patches. The solution bathing the cells was (mM) 150 NaC1, 3 KC1, 2 CaC12, 1 M Clz, 10 HEPES, 10 glucose, pH 7.35. The internal so ution in the i ette was 140 Cs-methanesulfonate, 10 NaCl, 10 E G t l , 10 HEPES, H 7.35. The external control solution was 150 2 CaCl,, 10 HEPES, 10glucose, 10 PM csc1,3 atropine, pH 7.35. The ACh that activated channels was added to the external solution. The hydroxide of the main cation was used to adjust the pH. Solution changes were made with large outflow tubes (40 pm inner diameter) that were positioned in a row next to each other (similar to Johnson and Ascher, 1987). The outflow tubes were mounted on a motorized manipulator (Newport Corp., Fountain Valley, CA); therefore, they could be repositioned for ra id solution changes that were complete in tens of mi iseconds. One of the outflow tubes always contained pure control solution, and one alwa s contained the control solution with agonist. For who e-cell currents, there was at least 1 minute between ACh applications to allow the cell to recover
from desensitization. MK-801 was urchased from Research Biochemicals Inc. (Natick, I&). RESULTS Dose response and voltage dependence of MK-801 block Figure 1shows that externally applied MK-801inhibits ACh-induced currents in a dose-dendent manner. The three macrosco ic currents were induced b 0.5 pM ACh a plied to B&H-1 cells voltage clampeJ to - 70 mV. T e horizontal solid bars indicate when 1, 10, or 100 pM MK-801 was a lied during the ACh-induced currents. The bottom oFFigure 1 shows dose-response curves at -20 mV and at -70 mV. On a Hill plot, the data behave as though there is one binding site for MK-801:the Hill coefficient is 1.Therefore, we fitted the data with a single-site equation (Woodhull, 1973). The MK-801 concentrations for 50% inhibition (IC, ) of the ACh-induced current are 7.6 pM at -20 mV ana3.3 pM at -70 mV. The volta e de endence suggests that the block may result from%K-f;Ol binding to a site within the open channel. Figure 2 presents a more complete look at the voltage dependence of the block b 10 pM MK-801. At all potentials, the onset of the bTock is rapid (Fi voltage dependence of steady-state block comparing the current in the presence of MK-801 wit
g:
5
Rc~,
B
P
~
R
210
M. AMADOR AND J.A. DAN1
and the other constants have their usual meaning. These two equations are combined and linearized to give ln(IJIrnk- 1)= -GzFV/RT
.................................................................................
I
+ ln{[MK-8011/IC50(0)}.(1)
Figure 2B shows that the voltage dependence of steadystate block follows the single-site equation. The slope of the plot gives the fraction of the membrane potential sensed at the binding site (6 = 0.51, and the y interce t gives an IC (0) = 9.7 M. The steady-state block ecreases e-fofj per 49 m depolarization. Recovery from block also is voltage-dependent (Fig. 2A). After washin off MK-801, the recovery follows a single exponentia with a T, of 3.2 & 0.2 (n = 12) s at -70 mV, 2.6 2 0.1 (n = 4) s at -50 mV, 2.0 2 0.2 (n = 6) s at -20 mV, and less than 0.4 (n = 4) s at +50 mV. For a sin le-site model, T, is proportional to the resident time o the blocker on its site. As the membrane potential is depoiarized, MK-801 resides for a shorter time at the site and dissociates more quickly. Mechanism of MK-801 block Since many organic cations have weak agonist effects (Sanchez et al., 19861, MK-801 might com ete at the ACh binding site. Inhibition by 10 JLM M!k -801 was tested at three concentrations of ACh. At +50 mV, intrachannel block would be minimized so competitive inhibition would be easier to see. Figure 3 shows that as the ACh concentration increases so does inhibition. Competitive inhibition would produce the opposite effect. A reasonable interpretation is that MK-801 blocks the open channel. As the ACh concentration is increased, more channels open and MK-801 has access to a greater pro ortion of the channels. Therefore, the steady-state b ock increases. There is evidence that some noncom etitive inhibitors, including MK-801 (Halliwell et a ., 1989; Javitt and Zukin, 1989), may block closed channels (Blanton et al., 1988).Figure 4A shows there is a sli ht inhibition of current caused by preincubating the cel for 1minute in 10 pM MK-801. After preincubation in MK-801, the initial hase of the ACh-induced current is decreased. Since &K-801 is not present in the ACh containing solution, all three currents are the same at the end of the ACh ap lication (Fig. 4A). The MK-801that inhibits the initial Rase of the current is washed away by the end of the A8h a lication. The initial inhibition could be caused by K-801 blocking the closed channel or by MK-801 causing desensitization. The currents labeled 4 through 10 in F i p r e 4B sug est that MK-801 does not strong1 desensitize the nA8hR. Between exposures to ACh $us MK-801, the cell rested in MK-801. After the first round of block, resting in MK-801 did not further inhibit the currents, indicating that desensitization was not adding onto the Also, at the end of the exposure to a brief pause in pure bath was enou h for current (current 10 of Fig. 4B). IfM -801 desensitization, a lon er time might have been needed to recover from the s ow phases of desensitization (Boyd, 1987). MK-801 becomes trapped inside the NMDA receptor ore (Huettner and Bean, 1988), and o en channel Elockers can become trapped in the nACh (Neely and Lingle, 1986). The currents labeled 1, 2, 3 and 9, 10 in
2
f
k
B
J1
I
10s
.................................................................................
2pM A C h
! .......................................................................................
Fig. 3. MK-801 blocks the open nAChR channel. Macroscopic currents were induced at t 50 mVb three different concentrations ofACh as indicated next to each trace. $he dotted horizontal line in each trace indicates the baseline before the outward-going (upward) currents were activated. The solid horizontal bar indicates when 10 FMMK-801 was present. As the concentration of ACh increases, a eater ro ortion of the nAChRs are open and, therefore, are availabgto be hocied by MK-801. At ver low concentrations ofACh, MK-801 does not block, indicating that Mi-801 is not a competitive inhibitor.
the current after recovery from inhibition: I &. Steady-state block is greater at negative potentiars as indicated by Imk/Ic: 0.22 2 0.04 (n = 18) at -70 mv, 0.29 i 0.04 (n = 21) at -50 mV, 0.40 i 0.04 (n = 9) at -20 mV, 0.72 2 0.02 (n = 13) at +50 mV. Based on a single bindin site, the block at any potential can be written as fol ows:
k
Irn&
=
(1 + [MK-8011/IC50(V)}-'.
The volta e de endence for binding within the membrane fie1 can e written explicitly as
Yi
= IC50(0)exp(SzFV/RT), ICBO(V)
where IC50(0)is the concentration for half inhibition at 0 mV, 6 is the fraction of the membrane otential sensed rane l potential, at the binding site, V is the applied mem?
P
P
ir
a
5
K
211
MK-801 INHIBITION OF THE nAChR CHANNEL
Figure 4B indicate that MK-801 is not trap ed in the seen in the underlying single-channel currents. MK-801 nAChR pore. After block by MK-801, the A&-induced alters the kinetics but not the conductance of the chancurrent quickly and completely recovers without re uir- nels. ing that the channel o en to release tra ed MK-811. At negative potentials, where the inhibition by MKAnother feature of igure 4B shoulcfie noted. The 801 is greatest, two factors contribute to the decrease in initial moments of the currents labeled 2 and 4 are current: the single-channel openings are shorter and 0.07 almost as large as the control currents labeled 1,3, and less frequent. The mean open time fell from 3.12 I 10. Those results suggest that the channels open as ms to 1.94 F 0.05 ms. When the channel o ens, MK-801 usual when MK-801 is present. After the channels o en, enters and blocks the channel, thereby ecreasing the MK-801 blocks, most likely by entering the open c an- mean open time. Figure 5B shows difference pulse nel. correlation histograms of the sin le-channel currents. A pulse correlation histo am (I$ eher and Steinbach, The effect of MK-801 on single-channel currents 1978) indicates the pro ability of an opening (which Figure 5A shows sin le-channel currents induced by could be an unblocking) at some delay after an opening. ACh in the presence an%absence of 10 pM MK-801. The The difference ulse correlation histo am shows the qualitative features of the macroscopic currents can be increased probaYJility of opening imme lately following
p.
B
il
r
A 1
r
60s
*
r
7
I'
I
10s
B
2 10s Fig. 4. MK-801 causes a slight inhibition of the closed channel, but MK-801 does not become trapped within the channel. A Three inward currents were induced b 0 5 pM ACh while the holdin potential was -50mV. Before the midr$e current was obtained, thecefi was bathed in bath solution containing 10 M MK-801 (solid horizontal bar). Then, the cell was quickly exposedLto regular bath solution for less than 1 second (indicated by the *) to wash away the MK-801 in the bath. Immediately following the 1-second wash, ACh was applied with no MK-801 present. The initial peak of the middle current is smaller than the peak of the two bracketing control currents. All three currents are
the same at the end of the ACh application because MK-801 has been washed away by the end of the middle record. Because complete solution changes were very fast, the results ofthis ex eriment were the same whether or not the 1-secondwash (*) was used.%: The membrane potential was held at -50 mV while 10 currents were induced by 0.5 pM ACh. Each current is numbered. The solid horizontal bar indicates when 10 M MK-801 was present. The ra id complete recover of the currents fabeled 3 and 10 indicate that dK-801 is not trappelin the closed channel.
212
M. AMADOR AND J.A. DAN1
A
1 0 / . ~ MM K - 8 0 1
Control
-50mV
I
20ms
B +50 mV
1
100 -
0
- 5 0 mV
I5O
I II
50
1
0 a a
aLt
0
100
200
300
400
500
0
D e l a y (ms)
100
200
300
400
500
Delay (ms)
Fig. 5. The effect of MK-801 on sin le channel currents. A Singlechannel currents induced by 0.1 pM & ! h are shown at +50 mV and -50 mV in the absence (control)or in the presence of 10 pM MK-801. At -50 mV there were, 2,433 control events and 2,001 MK-801 events from 4 patches. MK-801 decreases the mean open time and the number of openings per second. At +50 mV there were 385 control events and 711 MK-801 events from 2 patches. At +50 mV, MK-801 has less of an effect. Because the mean open time of the control currents is shorter, MK-801 has less of an opportunity to block. Also, the residence time of MK-801 is short, therefore, multiple openings, interpreted as MK-801 entering and leaving the open channel, are seen. B: Difference pulse correlation histograms were constructed as follows from singlechannel currents in the absence and in the resence of 10 pM MK-801. A pulse correlation histogram was plotted &r the control currents and
separately for the MK-801 currents. Those histo ams are of the delay between an opening and all other openings. %e pulse Correlation histograms were normalized to 100% at 250 to 500 ms, which is well after a constant rate of o ening had been reached. Then, the control histogram was subtracteffrom the MK-801 histogram. The difference histograms show that there is an increased probability of seeing an opening within a 5 to 15 ms delay after an opening when MK-801 is present. The y axis is the percent increase in the openin frequency above the constant rate (the 0 position) seen at longdela s kote that at t 5 0 mV the scale is 6 times greater than a t -50 mV. &en MK-801 is present, there are increased number of openings (unblockings)immediately followingan opening. This finding is especiallytrue a t +50 mV, where MK-801 resides in the channel for a shorter time before leaving (unblocking).
when MK-801 is present. These increased an opening result from MK-801 the channel's gate closes. In spite of the increased openings due to unblocking, at -50 mV, the number of openings per second decreases in a atch from 21 in control solution to 17 in 10 pM M -801. Various factors contribute to the overall decreased o enin frequency. First, MK-801 inhibits some closed c anne s, preventing them from opening (Fi 4A). Second, the residence time of MK-801 at its blocking site is longer at negative potentials, and while channels are blocked, they cannot contribute openings. Finally, some
channels might close while they are blocked; therefore, not all blocked channels contribute reopenings due to unblockin . That blocked channels can close is suggested by t e overall decrease in the number ofopenings and by the time dependence of the difference correlation histogram (Fig. 5B). An increased pro abilit of o ening is seen for only 10 ms after an opening. If bgcke8 channels could not close, the time for increased openin s would reflect the open time and the residence time o the MK-801, which is longer than indicated by the difference pulse correlation histogram (see below). The situation is different at positive potentials, where
2
R f
a
guise
B
213
MK-801 INHIBITION OF THE nAChR CHANNEL
block by MK-801 is less effective. At t 5 0 mV, the decreased mean open time is partially compensated for by an overall increase in the opening fre uency. The mean o en time decreased from 1.48 ? 0.1 ms in control sogtion to 1.20 t 0.08 in 10 pM MK-801. The number of openings er second in a patch increased from 0.7 to 0.8 in 10 p MK-801. These results suggest that MK-801does not block as well as positive potentials for several reasons. At depolarized potentials, the mean o en time of the nAChR in control solutions is shorter. &ere fore, the channels frequently close by normal gating before MK-801 can block. In addition, MK-801 resides at the intrachannel site for a shorter time at ositive potentials (Fig. 2). The shorter blocked time Lads to two factors that cause an increase in the number of openin s. The channel recovers from the blockade quickly; t en, the channel can be reopened b the a onist. In addition, MK-801 quickly goes on and o f roduce flickers durin an opening are seen in the dif erence pulse correlation histogram (Fi . 5B), where there is a dramatic increase in the pro abilit of opening (unblocking) shortly after an opening. Jhese unblockings are about 6 times more likely at t 5 0 mV than at -50 mV because of the difference in the residence time of MK801.
3
h!
a
P
B
The data can be used the intrinsic dissociation constant, Bean, 1988). Because the only a small contribution (Fig. 4A) and because many unblockings occur in MK-801, it is assumed that the Closed +Closed-Blocked transition is either relativel slow or rare. If the closed and opened states equi ibrate quickly, then & = u/b. The blocking rate (b) can be estimated from the single-channel mean open times in the resence (tmk)and absence (t ) of MK-801 from the fol owing equation (Neher and Steinbach, 1978):
P
P
t,
=
(blMK-8011 + 1/t,J1
At -50 mV, the forward blocking rate e uals 2 x lo7 M-Is-', which is close to the blocking r a t e l the NMDA receptor and is the same as the b ocking MK-801rate Of by QX-222 of the nAChR. The unblockin rate (u)can be estimated from the 7 for recovery from b ock:
7
%
DISCUSSION MK-801 primarily blocks the open nAChR channel Our results indicate that MK-801 inhibits ACh-induced currents by entering and blocking the open nAChR channel. MK-801 is not a competitive antagonist, and it does not act as a weak a onist that might increase desensitization (Figs. 3,4). urrents activated by the simultaneous application of ACh with MK-801 initially are the same as uninhibited currents, and then, they fall to a smaller value (Fig. 4B). That result suggests that the channels open before block begins. The voltage dependence of the block is consistent with MK801 reaching a single site in the channel 50% of the way through the membrane field (Fig. 2). The voltage dependence falls in the range found with other open channel blockers of the nAChR (see Table I, Sanchez et al., 1986). The results are not consistent with a completelysequential blocking model: Closed + Opened+ Blocked. The se uential mechanism requires that the total time a channe is open remain constant because the blocked channel must pass through the open state before closin total cially at ne ative potentials, MK-801 decreases t eEspeopen time ecause the mean open time is shorter and because the number of openings decreases. There are fewer o enin s mainly because the closed channel can be blocfed (fig. 4A) and because the blocked channel can close. Although the blocked channel can close, MK801 does not become trapped in the nAChR channel (Fig. 4B). The rimary action of MK-801 is block of the o en channel, ut the overall inhibition of the nAChR c annel is consistent with the following state diagram:
8
7
B
fi
f:
i
P b Closed f Opened+ Opened-Blocked a u Closed-Blocked
*
@
7,
=
i
Mpou),
where pois the robability of a channel being open in the absence of M -801. The p, was estimated to be about 0.008 from our lon single-channel records and from the work of Dilger anc f Brett (1990).The rate of unblocking equals about 48 s-', or the residence time of MK-801 at its blocking site is 21 ms. Therefore, the intrinsic dissociation constant (&) for MK-801 binding to the nAChR is estimated to be 2.4 KMat -50 mV. The & also can be estimated at t 50 mV. Since there are so many reopenings (unblockin s)in MK-801at t 5 0 mV, it is assumed that the unbloc ing rate dominates the time component seen in Figure 5B. In that case, the residence time of MK-801 is about 3 ms or u is about 333 s-I. The blockin rate is estimated as shown above: b is 1.5 X lo7 M - W5 . The Kd at +50 mV is estimated to be 22 pM. Comparison of MK-801 block of NMDA receptor and nAChR channels At -50 mV, MK-801 has about a 100 times affinit for the NMDA receptor than for the nACE,a& at 0 m there is on1 a factor of 40 difference (Huettner and Bean, 1988). he dissociation constants for the NMDA rece tor and nAChr are, res ectively, 22 nM and 2.4 pM at -go mV and 0.15 pM ani6.7 pM at 0 mV. For both rece tors, MK-801 may produce a slight closed channel b ock (see Javitt and Zukin, 19891,but its main action is as an open channel blocker. Althou h there are similarities between the block of the two c annels, there are im ortant differences. The onset of the MK-801 block of N DA receptors is slower than for the nAChR, and the onset is further slowed b magnesium (Huettner and Bean, 1988). Once the bloc has been established, however, it is more complete and lon lastin in part, because MK-801 becomes tra ped in t8e N d A rece tor but not in the nAChR. &ese differences are of p ysiologic importance. Only NMDA rece tors at a ver active glutamate synapse are bloc ed by MK-801 iecause, then, the channels have a higher probability of being open and unoccupied by magnesium. Open NMDA receptors free of magnesium are blocked with high potency by MK-801. Therefore,
2
a
G
.3:
P
a
it
f
E
R
214
M. AMADOR AND J.A. DAN1
MK-801 blocks NMDA receptors at opportune times: during epileptiform activity or during hyperactivity that could lead to excitotoxicity.Nicotinic AChR channels are blocked more quickly but less com letely. Once blocked, nAChRs also recover quickly. T erefore, the activity of nAChRs is deceased, but the receptors can still function during each synaptic event. Since neuronal nAChRs function mainly to modulate the excitability of neurons (Brown et al., 1983,1984),blockade of nAChRs may alter excitability. The blockade of nAChRs also could 801. For instance, MK-801 b preventing calcium (Jinkbeiner and transport calcium well (Decker and Dani, 1990), a secondary block of nAChRs may contribute to the therapeutic effect of MK-801. To ap reciate the potency and action of tricyclic compounds ike MK-801, it is interesting to examine NMDA receptor inhibition by desipramine, an antidepressant. Like MK-801, desi ramine is a tricyclic compound that primarily blocks t e open channel. The &(O) for desiramine is 52 ~J.M(Sernagor et al. 1989). Therefore, {esipramine is a less potent blocker of the NMDA receptor than MK-801 is of the nAChR, indicatin the surprising1 hi h affinity of MK-801 for the nACh5. It seems like y t at the tricyclic compounds discussed here may influence various neurotransmitter systems (Serna or et al. 1989). Their main action or secondary side effects couid involve more than one of these systems which may explain why the clinical action of these antidepressants is so complicated. ACKNOWLEDGMENTS We thank Dr. J.H. Steinbach for he1 ful discussions. We also thank P. Neal, J. Merchant, an8Dr. T. Reuhl for their he1 with cell culture or with reliminary experiments. T is work was supported by IH grant NS21229 and by the Muscular Dystrophy Association. REFERENCES
R
P
R
i 2
R
I5
A~LI~JO, L.G., and Albu uerque, E.X. (1986) Effects of phencyclidine an its analogs on t%e end late current of the neuromuscular junction. J. Pharmacol. Ex g e r . , 239:15-24. Amador, M., Merchant, J., gkuhl, T., and Dani, J.A. (1990) 0 en channel block of the nicotinic acetylcholine receptor and the NhfDA rece tor. Biophys. J.,57:122a. Anis, k . ~ . B , erry, S.C., Burton, N.R., and Lodge, D. (1983) The dissociative anesthetics, ketamine and phencyclidine, selectively reduce excitation of central mammalian neurons by N-methyl-aspartate. Br. J . Pharmacol., 79565-575. Blanton, M., McCardy, E., Gallaher, T., and Wang, H.H. (1988) Noncompetitive inhibitors reach their binding site in the acetylcholine rece tor by two different paths. Mol. Pharmacol., 33:634442. Boyd, %.D. (1987) Two distinct kinetic hases of desensitization of acetylcholine receptors of clonal rat P812 cells. J . Physiol. (Lond.), 398145567, Brown, D.A., Docherty, R.J., and Halliwell, J.V. (1983) Chemical transmission in the rat interpeduncular nucleus in uitro. J. Physiol. (Lond.), 34L655-670. Brown, D.A., Docherty, R.J., and Halliwell, J.V. (1984) The action of cholinomimetic substances on impulse conduction in the habenulointerpeduncular pathway of the rat in uitro. J. Physiol. (Lond.), 3531101-109. Changeux, J.P., and Revah, F. (1987) The acetylcholine receptor molecule: allosteric sites and the ion channel. Trends Neurosci., 10:245-250. Charnet, P., Labarca, C., Leonard, R.J., Vogelaar, N.J., Czyzyk, L., Gouin, A,, Davidson, N., and Lester, H.A. (1990) An o r - c h a n n e l blocker interacts with adjacent turns of a-helices in t e nicotinic acetylcholine receptor. Neuron, 4:87-95.
Dani, J.A. (1989a) Site-directed mutagenesis and sin le channel currents define the ionic channel of the nicotinic acetylgiine receptor. Trends Neurosci., 12:125-128. Dani, J.A. (1989b) 0 en channel structure and ion binding sites of the nicotinic acetylchoyine rece tor channel. J. Neurosci., 9:882-890. Decker, R.D., and Dani, J.1. (1990) Calcium permeability of the nicotinic acetylcholine receptor: the single-channel calcium influx is significant. J. Neurosci., 10:3413-3420. Dilger, J.P., and Brett, R.S. (1990) Direct measurement of the concentration- and time-dependent o en probability of the nicotinic acetylcholine receptor channel. Biop[ys. J., 57:723-731. Finkbeiner, S., and Stevens, C.F. (1988) Applications of quantitative measurements for assessing glutamate neurotoxicity. Proc. Natl. Acad. Sci. U.S.A., 85:40714074. Giraudat, J., Dennis, M., Heidmann, T., Chanf? J.-Y., and Changeux, J.-P. (1986)Structure ofthe hi h affinity bin ingsite for noncompetitive blockers of the acet lch3ine receptor: serine-262 of the delta subunit is labeled by [3h]chlorpromazine. Proc. Natl. Acad. Sci. U.S.A., 83~2719-2723. Halliwell, R.F., Peters, J.A., and Lambert, J.J. (1989)The mechanism of action and pharmacological specificity of the anticonvulsant NMDA antagonist MK-801: a voltage clamp study on neuronal cells in culture. Br. J. Pharmacol., 96:480494. Hamill, O.P., Marty, A,, Neher, E., Sakmann, B., and Sigworth, F.J. (1981) Improved patch-clamp techniques for high-resolution current recordings from cells and cell-free membrane patches. Pflugers Arch., 391:85-100. Heidmann, T., Oswald, R.E., and Changeux, J.P. (1983) Multiple sites of action for noncompetitive blockers on acetylcholine receptor rich membrane fragments from Torpedo rnarrnoruta. Biochemistry, 22~3112-3127. Hucho, F., Oberthur, W., and Lottspeich, F. (1986) The ion channel of the nicotinic acetylcholine receptor is formed by the homologous helices MI1 of the receptor subunits. FEBS Lett., 205:137-142. Huettner, J.E., and Bean, B.P. (1988) Block of NMDA-activated current by the anticonvulsant MK-801: selective binding to open channels. Proc. Natl. Acad. Sci. U.S.A., 85:1307-1311. Imoto, K., Busch, C., Sakmann, B., Mishina, M., Konno, T., Nakai, J., Bujo, ,H., Mori, Y., Fukuda, K., and, Numa, S. (1988) Rings of negatively charged amino acids determine the acetylcholine receptor channel conductance. Nature, 335:645-648. Imoto, K., Methfessel, C., Sakmann, B., Mishina, M., Mori, Y., Konno, T., Fukuda, K., Kurasaki, M., Bujo, H., Fujita, Y., and Numa, S. (1986) Location of the &subunit re 'on determining ion transport through the acetylcholine receptor cfannel. Nature, 324:67&674. Javitt, D.C., and Zukin, S.R. (1989) Biexponential kinetics of 13H]MK801 binding: evidence for access to closed and o en N methylD-aspartate rece tor channels. Mol. Pharmacol., 35:!87-393. Johnson, J.W., and)Ascher, P. (1987) Glycine otentiates the NMDA response in cultured mouse brain neurons. d t u r e , 325529-531. Kavanaugh, M.P., Tester, B.A.C., and Weber, E. (1989) Interaction of MK-801 with the nicotinic acet lcholine receptor-associated ion channel from electroplax. Eur. J. Jharmacol., 164:397,398. Kemp, J.A., Foster, A.C., and Wong, E.H.F. (1987) Non-competitive antagonists of excitatory amino acid receptors. Trends Neurosci., 10:294-298. Kloo Y., Nadler, V., and Sokolovsky M. (1988) Mode of binding of [3€!~dibenzocycloalkenimine (MK-801) to the N-methyl-D-as artate (NMDA) receptor and its therapeutic implication. F E B J Lett., 230:167-1 70. Leonard, R.J., Labarca, C.G., Charnet, P., Davidson, N., and Lester, H. (1988) Evidence that the M2 membrane-s anning region lines the ion channel ore of the nicotinic receptor. gience, 242:157&1581. Neely, A., and tingle, C.J. (1986) Trappin of an open-channel blocker a t the frog neuromuscular acetylchofine channel. Biophys. J., 50:981-986. Neher, E., and Steinbach, J.H. (1978) Local anesthetics transient1 block currents through single acetylcholine-receptor channels. Physiol. (Lond.), 277:153-176. Nilsson, L., Adem, A., Hardy, J., Winblad, B., and Nordberg, A. (1988) Do tetrahadroaminoacridine (THAI and physostigmine restore acetylcholine release in Alzheimer brains via nicotinic receptors. J. Neural Transm., 70:357-368. ODell, T.J., and Christensen, B.N. (1988) Mecamylamine is a selective non-competitive antagonist of N-methyl-D-aspartate- and aspartate-induced currents in horizontal cells dissociated from the catfish retina. Neurosci. Lett., 94(1-2):93-98. Oswald, R.E., Bamberger, M.J., and McLaughlin, J.T. (1984) Mechanism of phencyclidine binding to the acetylcholine receptor from Torpedo electroplaque. Mol. PharmacoI., 25:360-368. Papke, R.L., and Oswald, R.E. (1989) Mechanisms of noncompetitive
MK-801 INHIBITION OF THE nAChR CHANNEL
-
inhibition of acetvlcholine-inducedsingle-channel currents. J. Gen. Physiol., 93:785-811. Ramoa, AS., Alkondon, M., Aracava, Y., Irons, J., Lunt, G.G., Deshpande, S.S. Wonnacott, S., and Albuquerque, E.X. (1989) Noncompetitive blockade of the peri heral and central nicotinic acetylcholine receptors (AChR) by dK-801 and PCP. Neurosci. Abst.,
-15.876 -. -- -.
Reynolds, I.J., and Miller, R.J. (1988a) [3H]MK-801 Binding to the N-methyl-D-aspartate rece tor reveals dru interactions with the zinc and magnesium binling sites. J. Pfarmacol. Exp. Ther., 247:1025-1031. Reynolds,I.J.,and Miller, R.J. (1988b)Multiple sites for the regulation of the N-methyl-D-aspartate receptor. Mol. Pharmacol., 33581-584. Sakmann, B., and Neher, E., eds. (1983) Single-Channel Recording. Plenum Press, New York, 503 pp. Sanchez, J.A., Dani, J.A., Siemen, D., and Hille, B. (1986) Slow
215
ermeation of organic cations in acetylcholine receptor channels. J. 6 en. Ph siol ,87 985 1001 Sernagor, E., Kuhn, D . , i klicky, L., Jr., and Mayer, M.L. (1989)Open
channel block of NMdA receptor responses evoked by tricyclic antide ressants. Neuron, 2:1221-1227. Sine, S d . , and Steinbach, J.H. (1984)Activation of a nicotinic acetylcholine receptor. Biophys. J., 45:175-185. Sine, S.M., and Taylor, P. (1979) Functional consequences of a onistmediated state transitions in the cholinergic receptor. S t u i e s on cultured muscle cells. J . Biol. Chem., 254:3315-3325. Wong, E.H.F., Kemp, J.A., Priestly, T., Kni ht,A.R., Woodruff, G.N., and Iversen, L.L. (1986) The anticonvufsant MK801 is a potent N-methyl-D-aspartate antagonist. Proc. Natl. Acad. Sci. U.S.A., 83:7104-7108. Woodhull, A.M. (1973) Ionic blockage of sodium channels in nerve. J. Gen. Physiol., 61:687-708.
MK-801 Inhibition of Nicotinic Acetylcholine Receptor Channels MARIANO AMADOR AND JOHN A. DAN1 Department of Molecular Physiology and Biophysics, Baylor College of Medicine, Houston, Texas 77030
KEY WORDS
NMDA receptor, Noncompetitiveinhibition, Open channel block, Neuromuscular junction
ABSTRACT MK-801 is a potent inhibitor of the NMDA subtype of glutamate receptors. Single-channel and macroscopic currents indicate that MK-801 also inhibits nicotinic acetylcholine receptors (nAChRs).MK-801 does not significantly increase desensitization of the nAChRs or compete for the ACh binding site. Although there is a slight inhibition of the closed nAChR, the main action of MK-801 is to enter and block the open channel. The voltage dependence for block is consistent with a single binding site within the channel that is 50%of the way through the membrane field. The IC,, for block is 3 pM at - 70 mV for currents induced by 0.5 pM ACh. The data from both single-channel and macroscopic currents can be used to estimate a & (0) of 7 pM, which is about 40 times higher than the &(O) for MK-801 binding to the NMDA receptor. The relative potency of tricyclic compounds like MK-801for various neurotransmitter systems points out that the pharmacologic action of these drugs could involve complicated interactions in vivo.
INTRODUCTION Many organic cations block the nicotinic acetylcholine receptor (nAChR) channel (reviewed by Changeux and Revah, 1987). Although other binding sites may exist (Papke and Oswald, 1989), voltage-dependent noncometitive inhibitors often exert their primary action by glocking the open pore. Neher and Steinbach (1978) provided the first sin le-channel records showing open channel block of nA hRs. At low concentrations, the uaternar ammonium derivative of lidocaine, QX-222 Lowed w at are now classic characteristics of open channel block. Single-channel openings were shorter and occurred in bursts containing brief closures. The voltage-dependence of the block indicated that QX-222 entered the pore. Site-directed mutations revealed that X-222 inter2 (Charnet I acts with serine residues in the a-helix,% et al., 1990; Leonard et al., 1988). In addition to the results with QX-222, other studies using site-directed mutations (Imoto et al., 1986, 1988) provided evidence that the M2 a-helix lines the pore (reviewed by Dani, 1989a). Two other noncom etitive blockers, chlorpromazine and tri henylmet ylphosphonium, photoaffinit label the 2 a-helix (Giraudat et al., 1986; Hucho et a ., 1986). Therefore, there is direct evidence that these three noncompetitive inhibitors enter the open pore when they block the nAChR channel. Some noncompetitive inhibitors of the nAChR also inhibit the N-methyl-D-aspartate (NMDA) subset of lutamate receptors (reviewed by Kemp, 1987).The best anown is phencyclidine (PCP), which appears to have a high affinity site within the NMDA receptor channel (Anis et al., 1983; Kloog et al., 1988) and within the nAChR channel (Agua o et al., 1986; Oswald et al., 1984; Papke and Oswa d, 1989). Other noncom etitive inhibitors of both channels include mecamy amine,
8
K
P
K
2
P
0 1991 WILEY-LISS,INC.
P
empidine (O’Dell and Christensen, 1988), tetrahy8roaminoacridine (Nilsson et al., 1988) and chlor romazine (Heidmann et al., 1983; Reynolds and Mi ler, 1988a). The dibenzocyclohepteneimine MK-801 is a otent noncompetitive inhibitor of the NMDA receptor YWong et al., 1986). Evidence indicates that MK-801blocks the open channel. Both binding and unbinding of MK-801 require an open NMDA rece tor channel, and recovery from block is volta e-depen ent (Huettner and Bean, 1988).In addition, &g+ +,which blocks the o en channel (Nowak et al., 1984), relieves MK-801 bloc (Huettner and Bean, 1988) possibl by electrostatically repelling MK-801 from its intra-ciannel site or by an allosteric interaction (Reynolds and Miller, 1988b).PCP and similar compounds compete with MK-801, indicatin that they share the same site within the channel ( ong et al., 1986; Reynolds and Miller, 1988a).Recent preliminar results indicate that MK-801 also inhibits the n A d R (Halliwell et al., 1989; Kavanaugh et al., 1989; Ramoa et al., 1989; Amador et al., 1990). In these ex eriments, we determined the effect of external MK-{Ol on the nAChR by collecting macroscopic and single-channel currents. The main mechanism of inhibition is open channel block. The singlechannel currents indicate that the open time of the nAChR channel decreases, but the single channel conductance is not affected. The results show that MK-801 inhibits nAChR channels, and the results provide evidence for the mechanism of inhibition. The conclusions are valuable for assessin the otential clinical value of MK-801 as a specific bloger ofthe NMDA receptors.
P
!
E
&v
Received June 6,1990; accepted in revised form August 23,1990.
208
M. AMADOR AND J.A. DAN1
MATERIALS AND METHODS Tissue culture Nicotinic AChR channels were studied in clonal BCSH-1 mouse muscle cells. The clonal BCSH-1 cells were maintained usin standard tissue culture techni ues (Dani, 1989b; ine and Steinbach, 1984; Sine an% Taylor, 1979). The cell line was maintained in Dulbecco’s Modified Ea le medium (DMEM; Hazleton, Lenexa, KS; Gibco, Gait ersburg, MD) containin 10% to 20% heat-inactivated fetal bovine serum (FBS; ibco Gaithersburg, MD; Hyclone, Logan, UT), 100 units/mi of penicillin, 100 kg/ml of stre tomycin (Hazleton, Lenexa, KS) and 2 mM glutamine PHazleton, Lenexa, KS) in an incubator at 37°C with 5% CO . For patch-clamp experiments, the cells were late8 onto cover gIass (BellcoGlass Co., Vineland, N J a n d were maintained in DMEM with 0.5% FBS. The cells were fed every other day, and the were used from 5 t o 20 days after being plated onto t e cover glass.
8 5
8
K
1
Y I
1 PM
T
r
Patch-clamp techniques Whole-cell and single channel nAChR currents were measured in BC3H-1 cells usin standard patch-clamp techniques (Dani, 1989b; Hami et al., 1981; Sakmann and Neher, 1983). Patch pipettes were pulled with a two-stage electrode puller (PP-83; Narishige USA, Greenvale, NY) using 100 ~1disposable pipettes (Drummon Scientific, Broomall, PA) or Kovar 7052 glass (Garner Glass, Claremont, CA). To decrease ca acitive noise, the ipettes were coated with S 1 ar silicon elastomer ( ow Corning Corp., Midland, i&fior olystyrene Q-Dope (GC Electronic, Rockford, IL). he tips were polished immediately before the ex eriment using a microforge (MF-83; Narishige USA, reenvale NY). The platinum wire of the microforge was coated with Kovar glass. Currents were amplified and filtered using an Axopatch 1B (Axon Instruments, Burlingame, CA). Currents were digitally sampled with a 1Zbit analog to
fi
a .e
1;
cp
10 pM
T
I
30
100
i -
100 p M
T
l200pA L
20s
I
d
20s
100 t
c Q)
L L
3 0
SO 60
U Q)
.-
40
E L
20
N
0
0
z 0 1
3
10
[ M K - 8 0 1 ] (pM) Fig. 1. Dose-response relationship for inhibition by MK-801. Three macroseopic currents activated by0.5 pM ACh a t a ho ding otential of - 70 mV are shown. The ACh was applied at the downwartfarrow and was washed off at the upward arrow. MK-801 was a lied at a concentration of 1, 10, or 100 pM as indicated by the solic?\orizontal
bar. Dose-res onse curves are shown at -70 mV (W and at -20 mV ( 0 ) . The curves i t to the data are based on a single blocking site as described in the text. Each data point is the mean of 3 to 18 measurementsandisshownwithitsSE.
209
MK-801 INHIBITION OF THE A C h R CHANNEL
A
B +50mV
10.0
1 .o
J
-70mV
/250pA
0.1 -80
10s
-40
0
40
Vh ( m V )
Fig. 2. Voltage dependence of MK-801 block. A Macrosco ic cur rents were induced by 0.5 M ACh. During the current, 10 pM RK-80; was ap lied with the ACK for about 20 seconds. The current a t the end of tge 20 second application of MK-801 (Imk)was compared with the current after recove from block (Ic). The dotted horizontal line indicates the baselineyefore ACh a plication. The inhibition was more complete a t negative potentials. The voltage dependence of the
MK-801 block is plotted according to equation 1,which is based on a single binding site within the membrane field. Each data point is the mean of 9 to 21 measurements and is shown with its SE. They intercept indicates that the IC ,(O) equals 9.7 pM, and the slope of the !ine indicates that the binkng site is halfway through the membrane field (i.e., 6 = 0.5).
digital converter controlled by a ersonal computer (AxonInstruments, Burlingame, CAP and were saved on the hard disk. The computer system also delivered the holding otentials through a 12-bit digital to analog output. eurrents were sampled every 0.1 ms and filtered at 2 kHz. Single-channel currents were studied using outside/ out patches. The solution bathing the cells was (mM) 150 NaC1, 3 KC1, 2 CaC12, 1 M Clz, 10 HEPES, 10 glucose, pH 7.35. The internal so ution in the i ette was 140 Cs-methanesulfonate, 10 NaCl, 10 E G t l , 10 HEPES, H 7.35. The external control solution was 150 2 CaCl,, 10 HEPES, 10glucose, 10 PM csc1,3 atropine, pH 7.35. The ACh that activated channels was added to the external solution. The hydroxide of the main cation was used to adjust the pH. Solution changes were made with large outflow tubes (40 pm inner diameter) that were positioned in a row next to each other (similar to Johnson and Ascher, 1987). The outflow tubes were mounted on a motorized manipulator (Newport Corp., Fountain Valley, CA); therefore, they could be repositioned for ra id solution changes that were complete in tens of mi iseconds. One of the outflow tubes always contained pure control solution, and one alwa s contained the control solution with agonist. For who e-cell currents, there was at least 1 minute between ACh applications to allow the cell to recover
from desensitization. MK-801 was urchased from Research Biochemicals Inc. (Natick, I&). RESULTS Dose response and voltage dependence of MK-801 block Figure 1shows that externally applied MK-801inhibits ACh-induced currents in a dose-dendent manner. The three macrosco ic currents were induced b 0.5 pM ACh a plied to B&H-1 cells voltage clampeJ to - 70 mV. T e horizontal solid bars indicate when 1, 10, or 100 pM MK-801 was a lied during the ACh-induced currents. The bottom oFFigure 1 shows dose-response curves at -20 mV and at -70 mV. On a Hill plot, the data behave as though there is one binding site for MK-801:the Hill coefficient is 1.Therefore, we fitted the data with a single-site equation (Woodhull, 1973). The MK-801 concentrations for 50% inhibition (IC, ) of the ACh-induced current are 7.6 pM at -20 mV ana3.3 pM at -70 mV. The volta e de endence suggests that the block may result from%K-f;Ol binding to a site within the open channel. Figure 2 presents a more complete look at the voltage dependence of the block b 10 pM MK-801. At all potentials, the onset of the bTock is rapid (Fi voltage dependence of steady-state block comparing the current in the presence of MK-801 wit
g:
5
Rc~,
B
P
~
R
210
M. AMADOR AND J.A. DAN1
and the other constants have their usual meaning. These two equations are combined and linearized to give ln(IJIrnk- 1)= -GzFV/RT
.................................................................................
I
+ ln{[MK-8011/IC50(0)}.(1)
Figure 2B shows that the voltage dependence of steadystate block follows the single-site equation. The slope of the plot gives the fraction of the membrane potential sensed at the binding site (6 = 0.51, and the y interce t gives an IC (0) = 9.7 M. The steady-state block ecreases e-fofj per 49 m depolarization. Recovery from block also is voltage-dependent (Fig. 2A). After washin off MK-801, the recovery follows a single exponentia with a T, of 3.2 & 0.2 (n = 12) s at -70 mV, 2.6 2 0.1 (n = 4) s at -50 mV, 2.0 2 0.2 (n = 6) s at -20 mV, and less than 0.4 (n = 4) s at +50 mV. For a sin le-site model, T, is proportional to the resident time o the blocker on its site. As the membrane potential is depoiarized, MK-801 resides for a shorter time at the site and dissociates more quickly. Mechanism of MK-801 block Since many organic cations have weak agonist effects (Sanchez et al., 19861, MK-801 might com ete at the ACh binding site. Inhibition by 10 JLM M!k -801 was tested at three concentrations of ACh. At +50 mV, intrachannel block would be minimized so competitive inhibition would be easier to see. Figure 3 shows that as the ACh concentration increases so does inhibition. Competitive inhibition would produce the opposite effect. A reasonable interpretation is that MK-801 blocks the open channel. As the ACh concentration is increased, more channels open and MK-801 has access to a greater pro ortion of the channels. Therefore, the steady-state b ock increases. There is evidence that some noncom etitive inhibitors, including MK-801 (Halliwell et a ., 1989; Javitt and Zukin, 1989), may block closed channels (Blanton et al., 1988).Figure 4A shows there is a sli ht inhibition of current caused by preincubating the cel for 1minute in 10 pM MK-801. After preincubation in MK-801, the initial hase of the ACh-induced current is decreased. Since &K-801 is not present in the ACh containing solution, all three currents are the same at the end of the ACh ap lication (Fig. 4A). The MK-801that inhibits the initial Rase of the current is washed away by the end of the A8h a lication. The initial inhibition could be caused by K-801 blocking the closed channel or by MK-801 causing desensitization. The currents labeled 4 through 10 in F i p r e 4B sug est that MK-801 does not strong1 desensitize the nA8hR. Between exposures to ACh $us MK-801, the cell rested in MK-801. After the first round of block, resting in MK-801 did not further inhibit the currents, indicating that desensitization was not adding onto the Also, at the end of the exposure to a brief pause in pure bath was enou h for current (current 10 of Fig. 4B). IfM -801 desensitization, a lon er time might have been needed to recover from the s ow phases of desensitization (Boyd, 1987). MK-801 becomes trapped inside the NMDA receptor ore (Huettner and Bean, 1988), and o en channel Elockers can become trapped in the nACh (Neely and Lingle, 1986). The currents labeled 1, 2, 3 and 9, 10 in
2
f
k
B
J1
I
10s
.................................................................................
2pM A C h
! .......................................................................................
Fig. 3. MK-801 blocks the open nAChR channel. Macroscopic currents were induced at t 50 mVb three different concentrations ofACh as indicated next to each trace. $he dotted horizontal line in each trace indicates the baseline before the outward-going (upward) currents were activated. The solid horizontal bar indicates when 10 FMMK-801 was present. As the concentration of ACh increases, a eater ro ortion of the nAChRs are open and, therefore, are availabgto be hocied by MK-801. At ver low concentrations ofACh, MK-801 does not block, indicating that Mi-801 is not a competitive inhibitor.
the current after recovery from inhibition: I &. Steady-state block is greater at negative potentiars as indicated by Imk/Ic: 0.22 2 0.04 (n = 18) at -70 mv, 0.29 i 0.04 (n = 21) at -50 mV, 0.40 i 0.04 (n = 9) at -20 mV, 0.72 2 0.02 (n = 13) at +50 mV. Based on a single bindin site, the block at any potential can be written as fol ows:
k
Irn&
=
(1 + [MK-8011/IC50(V)}-'.
The volta e de endence for binding within the membrane fie1 can e written explicitly as
Yi
= IC50(0)exp(SzFV/RT), ICBO(V)
where IC50(0)is the concentration for half inhibition at 0 mV, 6 is the fraction of the membrane otential sensed rane l potential, at the binding site, V is the applied mem?
P
P
ir
a
5
K
211
MK-801 INHIBITION OF THE nAChR CHANNEL
Figure 4B indicate that MK-801 is not trap ed in the seen in the underlying single-channel currents. MK-801 nAChR pore. After block by MK-801, the A&-induced alters the kinetics but not the conductance of the chancurrent quickly and completely recovers without re uir- nels. ing that the channel o en to release tra ed MK-811. At negative potentials, where the inhibition by MKAnother feature of igure 4B shoulcfie noted. The 801 is greatest, two factors contribute to the decrease in initial moments of the currents labeled 2 and 4 are current: the single-channel openings are shorter and 0.07 almost as large as the control currents labeled 1,3, and less frequent. The mean open time fell from 3.12 I 10. Those results suggest that the channels open as ms to 1.94 F 0.05 ms. When the channel o ens, MK-801 usual when MK-801 is present. After the channels o en, enters and blocks the channel, thereby ecreasing the MK-801 blocks, most likely by entering the open c an- mean open time. Figure 5B shows difference pulse nel. correlation histograms of the sin le-channel currents. A pulse correlation histo am (I$ eher and Steinbach, The effect of MK-801 on single-channel currents 1978) indicates the pro ability of an opening (which Figure 5A shows sin le-channel currents induced by could be an unblocking) at some delay after an opening. ACh in the presence an%absence of 10 pM MK-801. The The difference ulse correlation histo am shows the qualitative features of the macroscopic currents can be increased probaYJility of opening imme lately following
p.
B
il
r
A 1
r
60s
*
r
7
I'
I
10s
B
2 10s Fig. 4. MK-801 causes a slight inhibition of the closed channel, but MK-801 does not become trapped within the channel. A Three inward currents were induced b 0 5 pM ACh while the holdin potential was -50mV. Before the midr$e current was obtained, thecefi was bathed in bath solution containing 10 M MK-801 (solid horizontal bar). Then, the cell was quickly exposedLto regular bath solution for less than 1 second (indicated by the *) to wash away the MK-801 in the bath. Immediately following the 1-second wash, ACh was applied with no MK-801 present. The initial peak of the middle current is smaller than the peak of the two bracketing control currents. All three currents are
the same at the end of the ACh application because MK-801 has been washed away by the end of the middle record. Because complete solution changes were very fast, the results ofthis ex eriment were the same whether or not the 1-secondwash (*) was used.%: The membrane potential was held at -50 mV while 10 currents were induced by 0.5 pM ACh. Each current is numbered. The solid horizontal bar indicates when 10 M MK-801 was present. The ra id complete recover of the currents fabeled 3 and 10 indicate that dK-801 is not trappelin the closed channel.
212
M. AMADOR AND J.A. DAN1
A
1 0 / . ~ MM K - 8 0 1
Control
-50mV
I
20ms
B +50 mV
1
100 -
0
- 5 0 mV
I5O
I II
50
1
0 a a
aLt
0
100
200
300
400
500
0
D e l a y (ms)
100
200
300
400
500
Delay (ms)
Fig. 5. The effect of MK-801 on sin le channel currents. A Singlechannel currents induced by 0.1 pM & ! h are shown at +50 mV and -50 mV in the absence (control)or in the presence of 10 pM MK-801. At -50 mV there were, 2,433 control events and 2,001 MK-801 events from 4 patches. MK-801 decreases the mean open time and the number of openings per second. At +50 mV there were 385 control events and 711 MK-801 events from 2 patches. At +50 mV, MK-801 has less of an effect. Because the mean open time of the control currents is shorter, MK-801 has less of an opportunity to block. Also, the residence time of MK-801 is short, therefore, multiple openings, interpreted as MK-801 entering and leaving the open channel, are seen. B: Difference pulse correlation histograms were constructed as follows from singlechannel currents in the absence and in the resence of 10 pM MK-801. A pulse correlation histogram was plotted &r the control currents and
separately for the MK-801 currents. Those histo ams are of the delay between an opening and all other openings. %e pulse Correlation histograms were normalized to 100% at 250 to 500 ms, which is well after a constant rate of o ening had been reached. Then, the control histogram was subtracteffrom the MK-801 histogram. The difference histograms show that there is an increased probability of seeing an opening within a 5 to 15 ms delay after an opening when MK-801 is present. The y axis is the percent increase in the openin frequency above the constant rate (the 0 position) seen at longdela s kote that at t 5 0 mV the scale is 6 times greater than a t -50 mV. &en MK-801 is present, there are increased number of openings (unblockings)immediately followingan opening. This finding is especiallytrue a t +50 mV, where MK-801 resides in the channel for a shorter time before leaving (unblocking).
when MK-801 is present. These increased an opening result from MK-801 the channel's gate closes. In spite of the increased openings due to unblocking, at -50 mV, the number of openings per second decreases in a atch from 21 in control solution to 17 in 10 pM M -801. Various factors contribute to the overall decreased o enin frequency. First, MK-801 inhibits some closed c anne s, preventing them from opening (Fi 4A). Second, the residence time of MK-801 at its blocking site is longer at negative potentials, and while channels are blocked, they cannot contribute openings. Finally, some
channels might close while they are blocked; therefore, not all blocked channels contribute reopenings due to unblockin . That blocked channels can close is suggested by t e overall decrease in the number ofopenings and by the time dependence of the difference correlation histogram (Fig. 5B). An increased pro abilit of o ening is seen for only 10 ms after an opening. If bgcke8 channels could not close, the time for increased openin s would reflect the open time and the residence time o the MK-801, which is longer than indicated by the difference pulse correlation histogram (see below). The situation is different at positive potentials, where
2
R f
a
guise
B
213
MK-801 INHIBITION OF THE nAChR CHANNEL
block by MK-801 is less effective. At t 5 0 mV, the decreased mean open time is partially compensated for by an overall increase in the opening fre uency. The mean o en time decreased from 1.48 ? 0.1 ms in control sogtion to 1.20 t 0.08 in 10 pM MK-801. The number of openings er second in a patch increased from 0.7 to 0.8 in 10 p MK-801. These results suggest that MK-801does not block as well as positive potentials for several reasons. At depolarized potentials, the mean o en time of the nAChR in control solutions is shorter. &ere fore, the channels frequently close by normal gating before MK-801 can block. In addition, MK-801 resides at the intrachannel site for a shorter time at ositive potentials (Fig. 2). The shorter blocked time Lads to two factors that cause an increase in the number of openin s. The channel recovers from the blockade quickly; t en, the channel can be reopened b the a onist. In addition, MK-801 quickly goes on and o f roduce flickers durin an opening are seen in the dif erence pulse correlation histogram (Fi . 5B), where there is a dramatic increase in the pro abilit of opening (unblocking) shortly after an opening. Jhese unblockings are about 6 times more likely at t 5 0 mV than at -50 mV because of the difference in the residence time of MK801.
3
h!
a
P
B
The data can be used the intrinsic dissociation constant, Bean, 1988). Because the only a small contribution (Fig. 4A) and because many unblockings occur in MK-801, it is assumed that the Closed +Closed-Blocked transition is either relativel slow or rare. If the closed and opened states equi ibrate quickly, then & = u/b. The blocking rate (b) can be estimated from the single-channel mean open times in the resence (tmk)and absence (t ) of MK-801 from the fol owing equation (Neher and Steinbach, 1978):
P
P
t,
=
(blMK-8011 + 1/t,J1
At -50 mV, the forward blocking rate e uals 2 x lo7 M-Is-', which is close to the blocking r a t e l the NMDA receptor and is the same as the b ocking MK-801rate Of by QX-222 of the nAChR. The unblockin rate (u)can be estimated from the 7 for recovery from b ock:
7
%
DISCUSSION MK-801 primarily blocks the open nAChR channel Our results indicate that MK-801 inhibits ACh-induced currents by entering and blocking the open nAChR channel. MK-801 is not a competitive antagonist, and it does not act as a weak a onist that might increase desensitization (Figs. 3,4). urrents activated by the simultaneous application of ACh with MK-801 initially are the same as uninhibited currents, and then, they fall to a smaller value (Fig. 4B). That result suggests that the channels open before block begins. The voltage dependence of the block is consistent with MK801 reaching a single site in the channel 50% of the way through the membrane field (Fig. 2). The voltage dependence falls in the range found with other open channel blockers of the nAChR (see Table I, Sanchez et al., 1986). The results are not consistent with a completelysequential blocking model: Closed + Opened+ Blocked. The se uential mechanism requires that the total time a channe is open remain constant because the blocked channel must pass through the open state before closin total cially at ne ative potentials, MK-801 decreases t eEspeopen time ecause the mean open time is shorter and because the number of openings decreases. There are fewer o enin s mainly because the closed channel can be blocfed (fig. 4A) and because the blocked channel can close. Although the blocked channel can close, MK801 does not become trapped in the nAChR channel (Fig. 4B). The rimary action of MK-801 is block of the o en channel, ut the overall inhibition of the nAChR c annel is consistent with the following state diagram:
8
7
B
fi
f:
i
P b Closed f Opened+ Opened-Blocked a u Closed-Blocked
*
@
7,
=
i
Mpou),
where pois the robability of a channel being open in the absence of M -801. The p, was estimated to be about 0.008 from our lon single-channel records and from the work of Dilger anc f Brett (1990).The rate of unblocking equals about 48 s-', or the residence time of MK-801 at its blocking site is 21 ms. Therefore, the intrinsic dissociation constant (&) for MK-801 binding to the nAChR is estimated to be 2.4 KMat -50 mV. The & also can be estimated at t 50 mV. Since there are so many reopenings (unblockin s)in MK-801at t 5 0 mV, it is assumed that the unbloc ing rate dominates the time component seen in Figure 5B. In that case, the residence time of MK-801 is about 3 ms or u is about 333 s-I. The blockin rate is estimated as shown above: b is 1.5 X lo7 M - W5 . The Kd at +50 mV is estimated to be 22 pM. Comparison of MK-801 block of NMDA receptor and nAChR channels At -50 mV, MK-801 has about a 100 times affinit for the NMDA receptor than for the nACE,a& at 0 m there is on1 a factor of 40 difference (Huettner and Bean, 1988). he dissociation constants for the NMDA rece tor and nAChr are, res ectively, 22 nM and 2.4 pM at -go mV and 0.15 pM ani6.7 pM at 0 mV. For both rece tors, MK-801 may produce a slight closed channel b ock (see Javitt and Zukin, 19891,but its main action is as an open channel blocker. Althou h there are similarities between the block of the two c annels, there are im ortant differences. The onset of the MK-801 block of N DA receptors is slower than for the nAChR, and the onset is further slowed b magnesium (Huettner and Bean, 1988). Once the bloc has been established, however, it is more complete and lon lastin in part, because MK-801 becomes tra ped in t8e N d A rece tor but not in the nAChR. &ese differences are of p ysiologic importance. Only NMDA rece tors at a ver active glutamate synapse are bloc ed by MK-801 iecause, then, the channels have a higher probability of being open and unoccupied by magnesium. Open NMDA receptors free of magnesium are blocked with high potency by MK-801. Therefore,
2
a
G
.3:
P
a
it
f
E
R
214
M. AMADOR AND J.A. DAN1
MK-801 blocks NMDA receptors at opportune times: during epileptiform activity or during hyperactivity that could lead to excitotoxicity.Nicotinic AChR channels are blocked more quickly but less com letely. Once blocked, nAChRs also recover quickly. T erefore, the activity of nAChRs is deceased, but the receptors can still function during each synaptic event. Since neuronal nAChRs function mainly to modulate the excitability of neurons (Brown et al., 1983,1984),blockade of nAChRs may alter excitability. The blockade of nAChRs also could 801. For instance, MK-801 b preventing calcium (Jinkbeiner and transport calcium well (Decker and Dani, 1990), a secondary block of nAChRs may contribute to the therapeutic effect of MK-801. To ap reciate the potency and action of tricyclic compounds ike MK-801, it is interesting to examine NMDA receptor inhibition by desipramine, an antidepressant. Like MK-801, desi ramine is a tricyclic compound that primarily blocks t e open channel. The &(O) for desiramine is 52 ~J.M(Sernagor et al. 1989). Therefore, {esipramine is a less potent blocker of the NMDA receptor than MK-801 is of the nAChR, indicatin the surprising1 hi h affinity of MK-801 for the nACh5. It seems like y t at the tricyclic compounds discussed here may influence various neurotransmitter systems (Serna or et al. 1989). Their main action or secondary side effects couid involve more than one of these systems which may explain why the clinical action of these antidepressants is so complicated. ACKNOWLEDGMENTS We thank Dr. J.H. Steinbach for he1 ful discussions. We also thank P. Neal, J. Merchant, an8Dr. T. Reuhl for their he1 with cell culture or with reliminary experiments. T is work was supported by IH grant NS21229 and by the Muscular Dystrophy Association. REFERENCES
R
P
R
i 2
R
I5
A~LI~JO, L.G., and Albu uerque, E.X. (1986) Effects of phencyclidine an its analogs on t%e end late current of the neuromuscular junction. J. Pharmacol. Ex g e r . , 239:15-24. Amador, M., Merchant, J., gkuhl, T., and Dani, J.A. (1990) 0 en channel block of the nicotinic acetylcholine receptor and the NhfDA rece tor. Biophys. J.,57:122a. Anis, k . ~ . B , erry, S.C., Burton, N.R., and Lodge, D. (1983) The dissociative anesthetics, ketamine and phencyclidine, selectively reduce excitation of central mammalian neurons by N-methyl-aspartate. Br. J . Pharmacol., 79565-575. Blanton, M., McCardy, E., Gallaher, T., and Wang, H.H. (1988) Noncompetitive inhibitors reach their binding site in the acetylcholine rece tor by two different paths. Mol. Pharmacol., 33:634442. Boyd, %.D. (1987) Two distinct kinetic hases of desensitization of acetylcholine receptors of clonal rat P812 cells. J . Physiol. (Lond.), 398145567, Brown, D.A., Docherty, R.J., and Halliwell, J.V. (1983) Chemical transmission in the rat interpeduncular nucleus in uitro. J. Physiol. (Lond.), 34L655-670. Brown, D.A., Docherty, R.J., and Halliwell, J.V. (1984) The action of cholinomimetic substances on impulse conduction in the habenulointerpeduncular pathway of the rat in uitro. J. Physiol. (Lond.), 3531101-109. Changeux, J.P., and Revah, F. (1987) The acetylcholine receptor molecule: allosteric sites and the ion channel. Trends Neurosci., 10:245-250. Charnet, P., Labarca, C., Leonard, R.J., Vogelaar, N.J., Czyzyk, L., Gouin, A,, Davidson, N., and Lester, H.A. (1990) An o r - c h a n n e l blocker interacts with adjacent turns of a-helices in t e nicotinic acetylcholine receptor. Neuron, 4:87-95.
Dani, J.A. (1989a) Site-directed mutagenesis and sin le channel currents define the ionic channel of the nicotinic acetylgiine receptor. Trends Neurosci., 12:125-128. Dani, J.A. (1989b) 0 en channel structure and ion binding sites of the nicotinic acetylchoyine rece tor channel. J. Neurosci., 9:882-890. Decker, R.D., and Dani, J.1. (1990) Calcium permeability of the nicotinic acetylcholine receptor: the single-channel calcium influx is significant. J. Neurosci., 10:3413-3420. Dilger, J.P., and Brett, R.S. (1990) Direct measurement of the concentration- and time-dependent o en probability of the nicotinic acetylcholine receptor channel. Biop[ys. J., 57:723-731. Finkbeiner, S., and Stevens, C.F. (1988) Applications of quantitative measurements for assessing glutamate neurotoxicity. Proc. Natl. Acad. Sci. U.S.A., 85:40714074. Giraudat, J., Dennis, M., Heidmann, T., Chanf? J.-Y., and Changeux, J.-P. (1986)Structure ofthe hi h affinity bin ingsite for noncompetitive blockers of the acet lch3ine receptor: serine-262 of the delta subunit is labeled by [3h]chlorpromazine. Proc. Natl. Acad. Sci. U.S.A., 83~2719-2723. Halliwell, R.F., Peters, J.A., and Lambert, J.J. (1989)The mechanism of action and pharmacological specificity of the anticonvulsant NMDA antagonist MK-801: a voltage clamp study on neuronal cells in culture. Br. J. Pharmacol., 96:480494. Hamill, O.P., Marty, A,, Neher, E., Sakmann, B., and Sigworth, F.J. (1981) Improved patch-clamp techniques for high-resolution current recordings from cells and cell-free membrane patches. Pflugers Arch., 391:85-100. Heidmann, T., Oswald, R.E., and Changeux, J.P. (1983) Multiple sites of action for noncompetitive blockers on acetylcholine receptor rich membrane fragments from Torpedo rnarrnoruta. Biochemistry, 22~3112-3127. Hucho, F., Oberthur, W., and Lottspeich, F. (1986) The ion channel of the nicotinic acetylcholine receptor is formed by the homologous helices MI1 of the receptor subunits. FEBS Lett., 205:137-142. Huettner, J.E., and Bean, B.P. (1988) Block of NMDA-activated current by the anticonvulsant MK-801: selective binding to open channels. Proc. Natl. Acad. Sci. U.S.A., 85:1307-1311. Imoto, K., Busch, C., Sakmann, B., Mishina, M., Konno, T., Nakai, J., Bujo, ,H., Mori, Y., Fukuda, K., and, Numa, S. (1988) Rings of negatively charged amino acids determine the acetylcholine receptor channel conductance. Nature, 335:645-648. Imoto, K., Methfessel, C., Sakmann, B., Mishina, M., Mori, Y., Konno, T., Fukuda, K., Kurasaki, M., Bujo, H., Fujita, Y., and Numa, S. (1986) Location of the &subunit re 'on determining ion transport through the acetylcholine receptor cfannel. Nature, 324:67&674. Javitt, D.C., and Zukin, S.R. (1989) Biexponential kinetics of 13H]MK801 binding: evidence for access to closed and o en N methylD-aspartate rece tor channels. Mol. Pharmacol., 35:!87-393. Johnson, J.W., and)Ascher, P. (1987) Glycine otentiates the NMDA response in cultured mouse brain neurons. d t u r e , 325529-531. Kavanaugh, M.P., Tester, B.A.C., and Weber, E. (1989) Interaction of MK-801 with the nicotinic acet lcholine receptor-associated ion channel from electroplax. Eur. J. Jharmacol., 164:397,398. Kemp, J.A., Foster, A.C., and Wong, E.H.F. (1987) Non-competitive antagonists of excitatory amino acid receptors. Trends Neurosci., 10:294-298. Kloo Y., Nadler, V., and Sokolovsky M. (1988) Mode of binding of [3€!~dibenzocycloalkenimine (MK-801) to the N-methyl-D-as artate (NMDA) receptor and its therapeutic implication. F E B J Lett., 230:167-1 70. Leonard, R.J., Labarca, C.G., Charnet, P., Davidson, N., and Lester, H. (1988) Evidence that the M2 membrane-s anning region lines the ion channel ore of the nicotinic receptor. gience, 242:157&1581. Neely, A., and tingle, C.J. (1986) Trappin of an open-channel blocker a t the frog neuromuscular acetylchofine channel. Biophys. J., 50:981-986. Neher, E., and Steinbach, J.H. (1978) Local anesthetics transient1 block currents through single acetylcholine-receptor channels. Physiol. (Lond.), 277:153-176. Nilsson, L., Adem, A., Hardy, J., Winblad, B., and Nordberg, A. (1988) Do tetrahadroaminoacridine (THAI and physostigmine restore acetylcholine release in Alzheimer brains via nicotinic receptors. J. Neural Transm., 70:357-368. ODell, T.J., and Christensen, B.N. (1988) Mecamylamine is a selective non-competitive antagonist of N-methyl-D-aspartate- and aspartate-induced currents in horizontal cells dissociated from the catfish retina. Neurosci. Lett., 94(1-2):93-98. Oswald, R.E., Bamberger, M.J., and McLaughlin, J.T. (1984) Mechanism of phencyclidine binding to the acetylcholine receptor from Torpedo electroplaque. Mol. PharmacoI., 25:360-368. Papke, R.L., and Oswald, R.E. (1989) Mechanisms of noncompetitive
MK-801 INHIBITION OF THE nAChR CHANNEL
-
inhibition of acetvlcholine-inducedsingle-channel currents. J. Gen. Physiol., 93:785-811. Ramoa, AS., Alkondon, M., Aracava, Y., Irons, J., Lunt, G.G., Deshpande, S.S. Wonnacott, S., and Albuquerque, E.X. (1989) Noncompetitive blockade of the peri heral and central nicotinic acetylcholine receptors (AChR) by dK-801 and PCP. Neurosci. Abst.,
-15.876 -. -- -.
Reynolds, I.J., and Miller, R.J. (1988a) [3H]MK-801 Binding to the N-methyl-D-aspartate rece tor reveals dru interactions with the zinc and magnesium binling sites. J. Pfarmacol. Exp. Ther., 247:1025-1031. Reynolds,I.J.,and Miller, R.J. (1988b)Multiple sites for the regulation of the N-methyl-D-aspartate receptor. Mol. Pharmacol., 33581-584. Sakmann, B., and Neher, E., eds. (1983) Single-Channel Recording. Plenum Press, New York, 503 pp. Sanchez, J.A., Dani, J.A., Siemen, D., and Hille, B. (1986) Slow
215
ermeation of organic cations in acetylcholine receptor channels. J. 6 en. Ph siol ,87 985 1001 Sernagor, E., Kuhn, D . , i klicky, L., Jr., and Mayer, M.L. (1989)Open
channel block of NMdA receptor responses evoked by tricyclic antide ressants. Neuron, 2:1221-1227. Sine, S d . , and Steinbach, J.H. (1984)Activation of a nicotinic acetylcholine receptor. Biophys. J., 45:175-185. Sine, S.M., and Taylor, P. (1979) Functional consequences of a onistmediated state transitions in the cholinergic receptor. S t u i e s on cultured muscle cells. J . Biol. Chem., 254:3315-3325. Wong, E.H.F., Kemp, J.A., Priestly, T., Kni ht,A.R., Woodruff, G.N., and Iversen, L.L. (1986) The anticonvufsant MK801 is a potent N-methyl-D-aspartate antagonist. Proc. Natl. Acad. Sci. U.S.A., 83:7104-7108. Woodhull, A.M. (1973) Ionic blockage of sodium channels in nerve. J. Gen. Physiol., 61:687-708.
