Phenol has various medical applications but can cause convulsions and
cardiac arrhythmia suggestive of K+ channel block. We
examined phenol inhibition of the delayed-rectifier RCK1 (Kv1.1)
K+ channel cloned from rat brain and expressed in
Xenopus laevis oocytes. Phenol (2.5 mM)
caused a 43 ± 5 mV depolarizing shift in the RCK1 half-activation
voltage (Vg) but only a 10 ± 3%
decrease in the peak conductance at 80 mV. The 10-90% rise time was
slightly increased, but this was not simply the result of the
activation shift. By contrast, deactivation kinetics at 
40 mV were
greatly accelerated. The importance of the phenolic hydroxyl group was assessed by comparing the effects of p-cresol (a phenol)
and its structural isomer benzyl alcohol (an aryl alcohol).
p-Cresol (1.5 mM) produced a 53 ± 2 mV
depolarizing shift in Vg, but benzyl alcohol
was much less effective
20 mM caused a depolarizing shift of only 23 ± 1 mV. Both isomers also accelerated channel
deactivation. Phenol and p-cresol are better hydrogen
bond donors than acceptors, whereas benzyl alcohol is a better acceptor
than donor. A hydrogen bond between the phenolic hydroxyl and a
presently unknown acceptor group may therefore underlie some aspects of
K+ channel inhibition. Depolarizing shifts in
Vg and accelerated tail kinetics are
consistent with 1) preferential phenol binding to resting channels,
causing the shift in Vg, and 2) a conducting phenol-bound open state with faster deactivation kinetics than the
unbound open state.
 |
Introduction |
Phenol has a number of clinical
applications [e.g., the induction of chronic nerve block (1) and
chemical face peeling (2)]. Serum phenol levels can be as high as 3.6 mM after chemical peels (3). It has been suggested that
clinical phenol administration may lead to cardiac dysrhythmia (1-4),
which emphasizes the need to identify phenol-induced changes in ion
channel gating. In the laboratory, phenol causes convulsions in mice
(5) and has a facilitatory action at the skeletal neuromuscular
junction (6). These effects imply significant changes in nerve and
muscle excitability, but relatively little is known concerning the
actions of phenol on voltage-gated ion channels. Kaila and Saarikoski
(7) suggest that phenol and its weakly acidic derivatives produce a
fairly specific inhibition of voltage-dependent potassium conductance in crayfish giant axons but did not demonstrate this action under voltage clamp. Zamponi and French (8) reported a much greater block by
phenol of cardiac sodium channels compared with those from skeletal
muscle. However, these were batrachotoxin-activated channels in planar
bilayers, and such channels do not gate normally.
Our first aim was to determine the effects of phenol on rat brain RCK1
(Kv1.1) potassium channels expressed in Xenopus laevis oocytes. When injected without additional 
-subunits, these channels act as delayed rectifiers (9). Phenols and alcohols both contain a
hydroxyl group; in phenols, the hydroxyl oxygen is attached directly to
the benzene ring, whereas in aryl alcohols, one or more intermediate
methylene groups separate the two. This has a major effect on the
acidity of the hydroxyl hydrogen and on the hydrogen bond donor versus
acceptor character of the molecule (10). Our second aim was to compare
the effects of two methods of adding a methylene group to the phenol
molecule (Fig. 1). Addition at the 4 position of the
benzene ring gives p-cresol, a phenol. Insertion between the
ring and the hydroxyl group gives benzyl alcohol, an aryl alcohol. This
comparison confirmed the importance of the phenol hydroxyl group and,
presumably, the phenolic hydrogen bond, because p-cresol was
found to be slightly more potent, and benzyl alcohol much less potent,
than phenol. A preliminary account of some of this work has been
presented to the Physiological Society (11).
| |
Materials and Methods |
cRNA preparation and injection.
Cloned cDNA (a gift from
Prof. O. Pongs, Hamburg, Germany) was transcribed in vitro
using the mMessage mMachine kit from Ambion (Austin, TX). Stage V-VI
oocytes were isolated from the ovarian lobes of X. laevis
frogs and defolliculated by incubation for 1-3 hr at room temperature
in a solution of 0.03% collagenase (Type XI; Sigma, Poole, UK) in
calcium-free OR2 solution (80 mM NaCl, 2 mM
KCl, 1 mM MgCl2, 2.5 mM Na-HEPES,
2.5 mM HEPES, pH 7.5, with NaOH). The frogs
were anesthetized in a solution of 0.15-0.35% tricaine in ice water
and either killed by destruction of the brain or allowed to recover
after the surgery for one further operation. Oocytes were injected with
50 nl of capped RNA (0.1-0.5 µg/µl) using a Nanoliter Injector
(World Precision Instruments, Stevenage, Hertfordshire, UK) and were
incubated in modified Barth's medium [78 mM NaCl, 1 mM KCl, 2.4 mM NaHCO3, 0.82 mM MgSO4, 0.33 mM
Ca(NO3)2, 0.41 mM
CaCl2, 10 mM Na-HEPES, 10 mM
HEPES, pH 7.2, with NaOH] containing 10 units/ml
penicillin and 10 µg/ml streptomycin at 19° for 1-10 days before
recording.
Electrophysiological measurements.
Whole-cell currents were
recorded using a two-microelectrode voltage clamp amplifier
specifically designed for recording fast, large-amplitude,
voltage-gated currents from oocytes (TURBO TEC 10CD; NPI Electronic,
Tamm, Germany). Voltage command pulses were generated, and currents
were recorded using a TL-1 interface and pCLAMP 6 software (Axon
Instruments, Foster City, CA). All currents were corrected for
capacitative and linear leakage currents before analysis using a
standard P/4 subtraction protocol. Currents were filtered at 1 kHz, and
the sampling frequency was between 3 and 4 kHz. Data were analyzed
using Clampfit (Axon Instruments, Foster City, CA), PSI-Plot (Poly
Software International, Salt Lake City, UT) and Microsoft Excel
(Redmond, WA) software.
The recording electrodes were filled with 2 M KCl (pH 7.2 with 5 mM HEPES and KOH) and had a resistance of between
0.6 and 0.9 M
. Ag/AgCl2 pellets were used as reference
electrodes. The bath solution was normal frog Ringers solution (110 mM NaCl, 2.5 mM KCl, 1.8 mM
CaCl2, 5 mM Na-HEPES, 5 mM
HEPES, pH 7.2, with NaOH), and solutions were exchanged by
total replacement of the bath medium. All experiments included control,
test, and wash. The chemicals used were Analar grade from BDH Chemicals
or Sigma (both Poole, Dorset, UK). The phenol was 99.5% pure and the
benzyl alcohol 99% pure (Sigma); the p-cresol was 99% pure
(Aldrich Chemical, Gillingham, Dorset, UK).
Experiments were performed at room temperature (23 ± 2°). All
numerical data are quoted as mean ± standard error for a minimum of three experiments unless otherwise stated.
Pulse protocols and data analysis.
Oocytes were
voltage-clamped at a holding potential of 80 mV. The I-V relationship
was determined by applying a 200-msec test pulse to between 70 and 80 mV. In many cases, this was immediately followed by a second 200-msec
test pulse to 40 mV (tail-IV). The stimulating frequency was 0.2 Hz
for an ordinary I-V protocol and 0.1 Hz for a tail-IV. The peak chord
conductance (gK) at each potential was
calculated from the corresponding peak current, assuming a reversal
potential (VK) of 80 mV: gK
= IK/(VVK).
The potential at which gK is half its maximal
value (gmax) is termed
Vg, and the slope factor of the normalized
conductance-voltage relationship is termed kg.
Vg and kg were determined
from a least-squares fit of a simple Boltzmann relationship to the
data: g/gmax = 1/(1 + exp((VgV)/kg)),
where g/gmax is the normalized
gK and V is the test potential.
Tail currents were fit using single or double exponential functions:
amplitude = A*exp(time/
) + C and
amplitude = Aslow*exp(time/slow) + Afast*exp(time/fast) + C, where the amplitude at time 0 is A + C or Aslow + Afast + C, respectively, and
C is the amplitude at infinite time. The relative
contribution of fast was calculated as
Afast/(Afast + Aslow). Fits were started 5 msec after the
beginning of the tail voltage step and extrapolated back to the
beginning. This time was chosen to avoid a clear settling phase of some
3 msec. All functions were fit to data using the nonlinear regression features of either pCLAMP or PSI-Plot. Ten- to ninety-percent rise
times were calculated using pCLAMP.
Numerical models.
Numerical simulations of the
HodgkinHuxley axon and K+ channel models were made
using a fourth order Runge-Kutta algorithm written in QuickBasic
(Microsoft, Redmond, WA) and based on that given by Constantinides
(12). The integration step size was manually adjusted to give a stable
solution.
| |
Results |
Phenol.
Phenol produced a large but extremely
voltage-dependent inhibition of the RCK1 current. Fig.
2A shows the effects of 2.5 mM phenol on
currents elicited by test pulses to voltages between 70 and 80 mV,
with a subsequent step to 40 mV. The I-V relationship in Fig. 2B
shows almost total suppression at low depolarizations and near zero
suppression at high depolarizations. Onset and offset of all drug
effects immediately followed commencement of solution changes. Averaged
data for fractional block by 1, 2.5, 4, and 8 mM
phenol are shown in Fig. 2C. Even 1 mM caused more
than 85% inhibition at 40 mV. The voltage-dependent nature of the
inhibition makes the quotation of a single IC50 value
somewhat arbitrary. One possibility is to consider the degree of block
at the voltage that causes 50% activation of control channels, which
is close to 30 mV. One millimolar phenol caused 78 ± 1% block
at 30 mV (Table 1); therefore, the IC50 on
this scale is somewhat less than 1 mM.
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 2.
Voltage-dependent RCK1 current inhibition by
phenol. Experimental currents (A) and an I-V relationship (B)
demonstrate the effect of 2.5 mM phenol. Scale
bars, 20 µA and 100 msec. C, Averaged data show the reduction
in peak current caused by 1, 2.5, 4, and 8 mM phenol. The
fractional block was calculated as 1 [2 × test amplitude/(control amplitude + wash amplitude)]. D and E, Depolarizing shifts in the conductance-voltage curve caused by 1 and 2.5 mM phenol, respectively. Test and wash data from each
individual experiment were normalized with respect to the maximal
control conductance for that oocyte before constructing the mean curves shown herein. Phenol caused substantial depolarizing shifts in activation voltage dependence with relatively little effect on the
conductance at 80 mV.  , control;  , test;  , wash.
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 1
Summarized effects of phenol, p-cresol and benzyl alcohol
on the RCK1 channel
One hundred percent block indicates total suppression of the current.
 and  indicate decrease and increase respectively. A change from
4 to 8 is a 100% increase, a change from 8 to 4 is a 50% decrease.
The potential is shown where appropriate. gk is
the peak chord conductance,  Vg the change in
potential for 50% activation and kg the slope
of the normalized conductance-voltage curve. fast and
slow are the time constants derived from double exponential fits to tail currents. Drug treatment not only reduces both
time constants but also increases the contribution of
fast. All experiments had a wash phase and effects are
calculated as test relative to the mean of control and wash.
|
|
The voltage-dependent block by phenol is linked to a depolarizing shift
in the activation curve. Averaged data for the effects of 1 and 2.5 mM phenol on the voltage dependence of channel activation are given in Fig. 2, D and E. Even 1 mM phenol caused a
13-mV depolarizing shift in Vg, the potential
for 50% activation, and 2.5 mM gave a massive shift of
more than 40 mV.
The fall in control RCK1 conductance at high potentials and the
plateauing of the I-V relationship result from a voltage-dependent block by internal magnesium (13). This phenomenon does not affect the
main conclusions regarding the effects of phenol and the other compounds studied. It would, however, complicate the comparison of peak
conductances in control and test conditions, as the phenol-treated currents reach full activation at more depolarized voltages than the
control currents (i.e., at potentials in which there is greater magnesium block). We have therefore compared conductances at a constant
potential of 80 mV, where 2.5 mM phenol is seen to cause only a 10% fall (Table 1).
Fig. 3 shows the effects of 2.5 mM
phenol on channel kinetics. Phenol slows the current rising phase
(Fig. 3, A and C) and greatly accelerates the speed of
closure at 40 mV (Figs. 2A and 3B). However, the increase in rise
time induced by 2.5 mM phenol is much less than that which
would result directly from a 43-mV depolarizing shift of the control
curve along the voltage axis (Fig. 3C). This departure from a simple
shift along the voltage axis was also evident with 4 and 8 mM phenol. The tail current decay could be adequately fit
by a first-order exponential function (the visibly separate lines in
Fig. 3B are first-order fits), but a second-order function gave a near
perfect fit (the second-order fits in Fig. 3B are indistinguishable
from the data). Analysis of 16 control experiments gave a
fast of 17 ± 1 msec, a slow of
52 ± 2 msec, and a 30 ± 3% contribution of the fast
process to the control decay. Phenol (2.5 mM) reduced both
the fast and slow time constants (by 65 ± 8% and 61 ± 2%,
respectively) and increased the relative contribution of
fast to 84 ± 2%; i.e., it more than doubled the
contribution of the fast decay.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 3.
Phenol slows the RCK1 current rising phase but
accelerates channel closure. A, The rising phase of control
(c), test (t), and wash
(w) current traces at 50 mV from a single experiment with 2.5 mM phenol. B, Tail currents for the control
(c) and test (t) traces. Currents were
normalized with respect to the control to facilitate comparison of the
rise time and the decay on repolarization to 40 mV. Single and double
exponential fits are superimposed on the current tails. Fits are shown
extrapolated back to the beginning of the tail voltage step, but the
first 5 msec were not included in the fitting process. The single
exponential fit is a reasonable approximation but can be distinguished
from the data; the double fit is a near perfect match. C, Averaged data show the effects of 2.5 mM phenol on the 10-90% rise time
over the full voltage range. The control curve is also shown moved along the voltage axis by an amount equal to the mean shift in half-activation voltage (Vg). Phenol
increases the rise time but by an amount much less than that predicted
by a simple shift. , control; , test; , wash; ×, control
shifted by Vg mV.
|
|
Benzyl alcohol and p-cresol.
The effects of 20 mM benzyl alcohol on RCK1 currents are compared with those
of 1.5 mM p-cresol in Fig. 4.
p-Cresol was more potent than phenol, but even 20 mM benzyl alcohol had less effect on
Vg than did 2.5 mM phenol. Benzyl
alcohol caused more high-depolarization block than either phenol or
p-cresol (Fig. 4C). Both p-cresol and benzyl
alcohol accelerated tail kinetics (Fig. 4, A and B), but benzyl alcohol
had little effect on the 10-90% rise time (Fig. 4D), whereas
p-cresol produced the largest slowing of activation kinetics
(Fig. 4E). Both compounds caused depolarizing shifts in the voltage
dependence of activation (Fig. 4, D and E).
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 4.
Effects of benzyl alcohol and
p-cresol on RCK1 currents. A and B, Currents from single
experiments show the effects of 20 mM benzyl alcohol and
1.5 mM p-cresol, respectively. Note that although both cause a substantial acceleration of the tail currents at
40 mV, only p-cresol noticeably slows the activation
kinetics. Scale bars, 40 µA and 100 msec (A) and 20 µA and 100 msec (B). C, Voltage dependence of block by 8 and 20 mM benzyl alcohol and 1.5 mM
p-cresol. Twenty millimolar benzyl alcohol causes a
greater high-depolarization block than either p-cresol
or phenol (see Fig. 2C). D and E, Averaged data show shifts in
activation voltage dependence and effect (or lack of effect) on the
10-90% rise time. ×, control rise time data translated by
Vg mV. , control; , test; ,
wash.
|
|
| |
Discussion |
Phenol and the firing behavior of excitable cells.
Phenol has
a biphasic effect on the excitability of isolated preparations and
whole organisms. At relatively high concentrations (7-10
mM), phenol blocks nervous conduction (14, 15). By
contrast, 0.25 mM phenol increases the frog end-plate
potential (6), and low doses facilitate neuromuscular transmission in
the cat (16) and produce convulsions in mice (5). We have shown that phenol produces a voltage-dependent inhibition of RCK1 channels and
that this inhibition is associated with substantial depolarizing shifts
in the activation threshold. Concentrations in the range known to exist
in phenol-treated patients [up to 3.6 mM (3)] induce
depolarizing shifts of 13 mV (1 mM) and 43 mV (2.5 mM) in the RCK1 conductance-voltage relationship. Such
shifts, even without K+ channel block at high potentials,
are effective in producing spontaneous firing in a model axon, and that
firing continues when most of the Na+ channels are also
blocked. Fig. 5 compares the effect on the firing
behavior of the Hodgkin-Huxley model squid axon of a 15-mV shift in the
K+ channel activation voltage dependence with that of a
50% voltage-independent K+ channel block. There is no
electrical stimulation; therefore, the control axon does not fire (data
not shown). When a 15-mV bias voltage is added to the K+
channel activation parameter (n), the axon fires spontaneously at 75 Hz. Voltage-independent block of 50% of the K+ channels
also causes spontaneous firing, but at only 50 Hz. The phenol-treated
axon continues to fire spontaneously even when 80% of the
Na+ channels are blocked. A 90% Na+ channel
block is required to damp the phenol-induced spontaneous firing. By
contrast, when the K+ channel block is voltage-independent,
a 30% Na+ channel block damps firing. Phenol (10 mM) causes 50-60% block of rat brain IIa Na+
channels in X. laevis oocytes (17). The biphasic effect of phenol on excitability is thus consistent with the combination of a
substantial influence on K+ channel gating and a less
potent inhibition of Na+ channels.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 5.
A depolarizing shift in the voltage dependence of
the K+ channel activation parameter (n) causes spontaneous
firing of the Hodgkin-Huxley axon. One hundred-twenty-millisecond
duration plots of voltage against time are shown for the electrically
unstimulated Hodgkin-Huxley axon. A 15-mV depolarizing shift in the
voltage dependence of n (15 mV K bias V) causes spontaneous firing that persists even when 80% of the Na+ channels are blocked. By
contrast, the lower frequency firing induced by a 50%
voltage-independent K+ channel block is critically damped
by a loss of only 30% of the Na+ channels. Scale
bars, 20 mV and 30 msec.
|
|
Mechanisms and drug-site interactions.
Our aims in the next
two sections are to discuss possible mechanisms of action while
recognizing that benzyl alcohol is a less, and p-cresol a
slightly more, potent inhibitor of RCK1 channel than phenol.
Reyes and Latorre (18) found that 19 mM benzyl alcohol
reduces the surface dipole potential of bacterial
phosphatidylethanolamine monolayers by 20 mV, whereas 51 mM
caused a 60-mV reduction. Our observation of a 23-mV shift in the
voltage dependence of RCK1 channel activation for 20 mM
benzyl alcohol is thus in the same range as the change in
phosphatidylethanolamine monolayer surface potential. However, for that
effect to be fully translated into a gating shift, one would require
either an extremely asymmetrical adsorption of benzyl alcohol [such
that the reduction in dipole potential took place only at the inner
surface of the membrane (19)] or that the peak of the energy barrier
for the voltage-dependent activation transition be located almost
entirely within one membrane dipole layer (20). Furthermore, a simple
surface potential effect would induce a uniform shift in all gating
parameters; i.e., the changes in rise times, for example, would be
accounted for by a shift in the control curve. We did not find this to
be the case in our study.
Phloretin is a phenolic compound with a dipole moment of 5.6 Debye
(21). Phloretin also causes large (30 to 40 mV) depolarizing shifts in
the voltage dependence of axonal K+ channel activation (20,
22). Strichartz et al. (20) suggest that these shifts are
caused by phloretin-induced changes in the axonal membrane surface
dipole potential. Thus, large shifts are linked to a large dipole
moment, and it seems logical to consider a similar explanation for the
effect of phenol on RCK1 channels. However, phenol has a dipole moment
of only 1.2 D, whereas benzyl alcohol has a dipole moment of 1.7 D
(23). This would not account for the greater effectiveness of phenol
without some additional factor that substantially increased phenol
adsorption.
Kitagawa and Hirata (24) and Kitagawa et al. (25) report the
effects of benzyl alcohol, phenol, and p-cresol on the
membrane fluidity of bovine platelets (determined from changes in
the fluorescence anisotropy of diphenylhexatriene) and on platelet
aggregation. Twenty millimolar benzyl alcohol, 10 mM
phenol, and 5 mM p-cresol all caused a 6%
reduction in diphenylhexatriene fluorescence anisotropy (indicating an
increase in bilayer fluidity) and a 92-98% fall in platelet
aggregation. Phenol (1.9 mM) and 1.4 mM
p-cresol gave a 50% inhibition of aggregation, but the
IC50 for benzyl alcohol was not reported. It is possible
that alterations in channel behavior secondary to membrane fluidity
changes account for some of the effects we saw, but the variation in
potency for fluidity changes, a factor of 4 between benzyl alcohol and
p-cresol, is not substantial when compared with the factor
of approximately 30 that was found for RCK1 channel gating.
The water-soluble enzyme firefly luciferase is a popular model for the
interaction of more or less hydrophobic compounds with purely protein
sites of action. Abraham et al. (26) give IC50 values for phenol and benzyl alcohol of 1.1 and 1.7 mM,
respectively. Thus, although the luciferase model has roughly the
correct phenol potency, it does not show the required degree of
selectivity between phenol and benzyl alcohol.
We suggest that an interaction more specific than fluidity or dipole
potential effects is at least partly responsible for K+
current inhibition by phenol. The comparison of RCK1 channel perturbation by phenol, p-cresol, and benzyl alcohol
identifies a possible feature of that interaction. Abraham et
al. (10) describe the partitioning of organic solutes between
water and model phases, such as octanol or hexadecane, in terms of six
solute properties. Table 2 gives these data for phenol,
p-cresol, and benzyl alcohol. The striking difference is
that both phenol and p-cresol are better hydrogen bond
donors than acceptors, whereas benzyl alcohol is a better acceptor than
donor. This suggests that the partition between phenol and its site of
action on, in, or around RCK1 channels involves a strong hydrogen bond
in which the phenol hydroxyl group is the donor and some presently
unidentified group is the acceptor.
View this table:
[in this window]
[in a new window]
|
TABLE 2
Physical properties of phenol, p-cresol, and benzyl alcohol
(from ref. 10)
Both phenols are better hydrogen bond donors (0.60 and 0.57) than
acceptors (0.30 and 0.31) whereas benzyl alcohol is a better acceptor
(0.56) than donor (0.33).
|
|
Urban and Haydon (27) propose that the relatively high selectivity of
certain halogenated ethers for the squid giant axon delayed rectifier
K+ current over the axonal Na+ current may also
result from channel-specific hydrogen bonds in which acidic hydrogens
on the halogenated ether are the donor groups. By contrast, benzyl
alcohol, with its weaker donor ability, is much more effective in
reducing the squid axon Na+ current than the K+
current (28). Rat brain IIa Na+ channels expressed in
X. laevis oocytes require 10 mM phenol or 5 mM p-cresol for 50-60% inhibition, and those
effects are accompanied by a negligible (<3 mV) shift in the voltage
dependence of channel activation (17). This supports the suggestion of a specific interaction between phenols and K+ channels.
Modeling phenol inhibition of RCK1 currents.
We now present a
simple (but not necessarily unique) model of drug action that
reproduces the two most consistent features of phenol,
p-cresol, and benzyl alcohol perturbation of RCK1 channels. These are a depolarizing shift in the voltage dependence of channel activation and an acceleration of channel closure on repolarizing to
40 mV. The model and comparisons between simulated and experimental data are shown in Fig. 6. The control channel has four
resting states (shown as one group for convenience), a preopen state, and an open state (see, e.g., 29). Drug binding can have two possible effects. One is to simply block the channel, as indicated by the bound
blocked state. This accounts for a high-depolarization block but would
not cause the depolarizing activation shift or the accelerated tail
kinetics. These are produced by including a set of drug-bound resting,
preopen, and open states in which the bound open state conducts but is
destabilized. In the example given, drug binding destabilizes the open
state by causing a 1000-fold increase in p, the rate
constant to the preopen state. To maintain microscopic reversibility
(30), this change in p must be matched by making the
affinity of the resting and preopen states for the drug 1000 times
greater than that of the open state. This was done by reducing the
unbinding rate constants (0.003 compared with 3). The rate constants
among the various resting states and between the preopen and last
resting state were not affected by the drug in this simulation.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 6.
Phenol and the RCK1 channela simple model. A, The
control channel is modeled by a linear series of four resting states, a pre-open state, and an open state. The resting states (shown as one
group) are linked by the standard 4:3:2:1 arrangement of
 m [0.5754 × exp(V/48)] and
m [0.1691 × exp(V/48)] rate
constants (see, e.g., 29). The preopen and open states are linked by
the rate constants p [0.24 × exp(0.0001 × V/24.56)] and p [0.004 × exp(0.999 × V/24.56)]. V,
simulation voltage. The depolarizing activation shift and accelerated
tail kinetics were simulated by adding a parallel set of bound states
in which the bound open state still conducts but is destabilized by a
1000-fold increase in p. This is matched by a 1000-fold
increase in the binding affinity of resting and preopen states. A bound
blocked state is included to account for current reduction additional
to that induced by a depolarizing activation shift. The numbers
indicate rate constants for binding or unbinding. B, A comparison of
simulated and experimental currents. The experimental currents are
averaged normalized data from four 2.5-mM phenol
experiments. Scale bar, 100 msec. C, Fraction conducting
versus voltage curves for () experimental control,
( ) simulation control, ()
experimental test, and ( ) simulation test. D, Tail kinetics. Single
exponential fits are shown superimposed on experimental and simulated
tails.
|
|
Fig. 6B compares simulated currents with averaged control and test data
from experiments with 2.5 mM phenol. Fig. 6C shows the
match between simulation and experiment for channel activation, and
Fig. 6D compares deactivation kinetics. The model does not exactly
reproduce the effects of phenol, p-cresol, or benzyl
alcohol, but that is not surprising given its simple nature. For
example, the simulated current tails are all well fit by single
exponential functions, and the changes in rise time do not exactly
match the experimental findings. However, it does reproduce the
depolarizing shift and the reduction in the slope of the activation
voltage dependence, and the deactivation kinetics are accelerated. The inclusion of a conducting bound open state was critical to the production of the faster tails.
A 1000-fold decrease in open-state stability is associated with a free
energy change of 17 kJ/mol. That is equivalent to roughly twice the
combined energy of a hydrogen bond (31) and the interaction between a
cation and the benzene ring 
electrons (32). It is possible that two
phenol molecules must bind to the site of action, and it is reassuring
that the calculated energy change is neither impossibly large nor
ludicrously small.
The position of the bound blocked state shown in Fig. 6A implies that
phenol has an action equivalent to classical open channel block and
that there are two separate sites of action. An alternative is to have
the bound blocked state coming off the conducting bound open state,
which implies a single site of action with a drug-bound transition
between two states. Each position mimics the main experimental findings, but they are not kinetically equivalent. Our reasons for
choosing the open channel block position relate to experiments on the
effects of local anesthetics and n-alkanols performed in collaboration with J. P. Newman and J. A. Harrold (University of
Dundee, Dundee, Scotland). Lidocaine and bupivacaine cause open-channel
block-like increases in the tail decay time constant. We do not know
the interaction between phenol and lidocaine in RCK1 channels, but
Zamponi and French (8) suggest that for Na+ channels,
phenol mimics one of lidocaine's blocking modes. Our model is
generally compatible with that suggestion; one of phenol's actions
mimics the main effect of lidocaine (open channel block). The effects
of n-pentanol on RCK1 channels can be explained by a
mechanism very similar to the phenol model. n-Nonanol,
however, has a much more obvious open-channel block behavior.
Therefore, our model is formulated to produce a range of activity with
classical open-channel block behavior at one end of the spectrum.
Concluding remarks.
We have shown that phenol, at
concentrations that have been measured in human serum after clinical
administration, causes very large depolarizing shifts in the voltage
dependence of RCK1 channel activation and a substantial acceleration of
channel deactivation. The magnitude of these changes and the relative
potencies of benzyl alcohol and p-cresol suggest a specific
interaction with the channel itself, perhaps involving a strong
hydrogen bond between the phenolic hydroxyl and an acceptor group on
the channel protein. The depolarizing shift in activation threshold and
the marked increase in deactivation kinetics are suggestive of a
mechanism in which 1) phenol binding destabilizes but does not
necessarily block open channels and 2) both resting and pre-open
channels have a higher affinity for phenol than the open channel.
| 1.
|
Morrison, J. E., Jr.,
D. Matthews,
R. Washington,
P. V. Fennessey, and
M. Harrison.
Phenol motor point blocks in children: plasma concentrations and cardiac dysrhythmias.
Anesthesiology
75:359-362 (1991)[Medline].
|
| 2.
|
Warner, M. A. and
J. V. Harper.
Cardiac dysrhythmias associated with chemical peeling with phenol.
Anesthesiology
62:366-367 (1985)[Medline].
|
| 3.
|
Gross, B. G.
Cardiac arrhythmias during phenol face peeling.
Plast. Reconstr. Surg.
73:590-594 (1984)[Medline].
|
| 4.
|
Forrest, T. and
D. T. O. Ramage.
Cardiac dysrhythmia after subtrigonal phenol.
Anaesthesia
42:777-778 (1987)[Medline].
|
| 5.
|
Angel, A. and
K. J. Rodgers.
An analysis of the convulsant activity of substituted benzenes in the mouse.
Toxicol. Appl. Pharmacol.
21:214-229 (1972)[Medline].
|
| 6.
|
Otsuka, M. and
Y. Nonomura.
The action of phenolic substances on motor nerve endings.
J. Pharmacol. Exp. Ther.
140:41-45 (1962).
|
| 7.
|
Kaila, K. and
J. Saarikoski.
Inhibition of voltage-dependent potassium conductance by convulsant phenols in the medial giant axon of the crayfish.
Comp. Biochem. Physiol.
63C:17-24 (1980).
|
| 8.
|
Zamponi, G. W. and
R. J. French.
Dissecting lidocaine action: diethylamide and phenol mimic separate modes of lidocaine block of sodium channels from heart and skeletal muscle.
Biophys. J.
65:2335-2347 (1993)[Abstract/Free Full Text].
|
| 9.
|
Stühmer, W.,
M. Stocker,
B. Sakmann,
P. Seeburg,
A. Baumann,
A. Grupe, and
O. Pongs.
Potassium channels expressed from rat brain cDNA have delayed rectifier properties.
FEBS Lett.
242:199-206 (1988)[Medline].
|
| 10.
|
Abraham, M. H.,
H. S. Chadha,
G. S. Whiting, and
R. C. Mitchell.
Hydrogen bonding. 32. An analysis of water-octanol and water-alkane partitioning and the log P parameter of Seiler.
J. Pharm. Sci.
83:1085-1100 (1994)[Medline].
|
| 11.
|
Elliott, A. A. and
J. R. Elliott.
Phenol inhibition of rat brain RCK1 (Kv1.1) K+ channels expressed in Xenopus oocytes: implications for action potential firing in excitable cells.
J. Physiol. (Lond.)
493:81-82 (1996).
|
| 12.
|
Constantinides, A.
Applied Numerical Methods with Personal Computers. McGraw-Hill, New York (1987).
|
| 13.
|
Ludewig, U.,
C. Lorra,
O. Pongs, and
S. H. Heinemann.
A site accessible to extracellular TEA+ and K+ influences intracellular Mg2+ block of cloned potassium channels.
Eur. Biophys. J.
22:237-247 (1993)[Medline].
|
| 14.
|
Seeman, P.
The membrane actions of anesthetics and tranquilizers.
Pharmacol. Rev.
24:583-655 (1972)[Abstract].
|
| 15.
|
Nathan, P. W. and
T. A. Sears.
Effects of phenol on nervous conduction.
J. Physiol. (Lond.)
150:565-580 (1960).
|
| 16.
|
Blaber, L. C. and
J. P. Gallagher.
The facilitatory effects of catechol and phenol at the neuromuscular junction of the cat.
Neuropharmacology
10:153-159 (1971)[Medline].
|
| 17.
|
Harrold, J. A.,
J. R. Elliott, and
A. A. Elliott.
Inhibition of oocyte-expressed rat brain IIa Na+ channels by phenol. p-cresol and n-pentanol.
J. Physiol. (Lond.)
497:127P (1996).
|
| 18.
|
Reyes, J. and
R. Latorre.
Effect of the anesthetics benzyl alcohol and chloroform on bilayers made from monolayers.
Biophys. J.
28:259-280 (1979)[Abstract/Free Full Text].
|
| 19.
|
Elliott, J. R.,
D. A. Haydon, and
B. M. Hendry.
The asymmetrical effects of some ionized n-octyl derivatives on the sodium current of the giant axon of Loligo forbesi.
J. Physiol. (Lond.)
350:429-445 (1984)[Abstract/Free Full Text].
|
| 20.
|
Strichartz, G. R.,
G. S. Oxford, and
F. Ramon.
Effects of the dipolar form of phloretin on potassium conductance in squid giant axons.
Biophys. J.
31:229-246 (1980)[Abstract/Free Full Text].
|
| 21.
|
Andersen, O. S.,
A. Finkelstein,
I. Katz, and
A. Cass.
Effect of phloretin on the permeability of thin lipid membranes.
J. Gen. Physiol.
67:749-771 (1976)[Abstract/Free Full Text].
|
| 22.
|
Klusemann, J. and
H. Meves.
Phloretin affects the fast potassium channels in frog nerve fibres.
Eur. Biophys. J.
20:79-86 (1991)[Medline].
|
| 23.
|
Lide, D. R. and
H. P. R. Frederikse.
CRC Handbook of Chemistry and Physics. CRC Press, Boca Raton, FL (1995).
|
| 24.
|
Kitagawa, S. and
H. Hirata.
Effects of alcohols on fluorescence anisotropies of diphenylhexatriene and its derivatives in bovine blood platelets: relationships of the depth-dependent change in membrane fluidity by alcohols and their effects on platelet aggregation and adenylate cyclase activity.
Biochim. Biophys. Acta
1112:14-18 (1992)[Medline].
|
| 25.
|
Kitagawa, S.,
M. Orinaka, and
H. Hirata.
Depth-dependent change in membrane fluidity by phenolic compounds in bovine platelets and its relationship with their effects on aggregation and adenylate cyclase activity.
Biochim. Biophys. Acta
1179:277-282 (1993)[Medline].
|
| 26.
|
Abraham, M. H.,
W. R. Lieb, and
N. P. Franks.
Role of hydrogen bonding in general anesthesia.
J. Pharm. Sci.
80:719-724 (1991)[Medline].
|
| 27.
|
Urban, B. W. and
D. A. Haydon.
The actions of halogenated ethers on the ionic currents of the squid giant axon.
Proc. R. Soc. Lond. B Biol. Sci.
231:13-26 (1987)[Medline].
|
| 28.
|
Harper, A. A.,
A. G. Macdonald, and
K. T. Wann.
The effect of temperature on the nerve-blocking action of benzyl alcohol on the squid giant axon.
J. Physiol. (Lond.)
338:51-60 (1983)[Abstract/Free Full Text].
|
| 29.
|
Koren, G.,
E. R. Liman,
D. E. Logothetis,
B. Nadal-Ginard, and
P. Hess.
Gating mechanism of a cloned potassium channel expressed in frog oocytes and mammalian cells.
Neuron
2:39-51 (1990).
|
| 30.
|
Hille, B.
Ionic Channels of Excitable Membranes. 2nd ed. Sinauer Associates Inc., Sunderland, MA (1992).
|
| 31.
|
Jeffrey, G. A. and
W. Sanger.
Hydrogen bonding in biological structures. Springer-Verlag, Berlin (1994).
|
| 32.
|
Schneider, H.-J.,
T. Blatter,
S. Simonova, and
I. Theis.
Large binding constant differences between aromatic and aliphatic substrates in positively charged cavities indicative of higher order electrical effects.
J. Chem. Soc. Chem. Comm.
9:580-581 (1989).
|
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Summary
Introduction
Summary
