 |
Introduction |
Voltage-gated
sodium channels generate the action potential in nerve and muscle
tissue. Recent advances in molecular techniques, including
site-directed mutagenesis and the formation of channel chimeras, have
provided important clues into the mechanism of channel kinetics and
pharmacology [see Fozzard and Hanck (1996)
and Marbán et al.
(1998) for review). Before the advent of molecular techniques,
naturally occurring neurotoxins [see Strichartz et al. (1987) and
Catterall et al. (1992) for review] were essential tools for studying
sodium channel pharmacology and function because they modify channel
behavior. At least six receptor sites on sodium channels have been
identified by toxin binding studies. Site 1 toxins, such as
tetrodotoxin and saxitoxin, block sodium channels by binding to a
receptor in the pore region near the extracellular face of the channel
(Tomaselli et al., 1995), whereas toxins that bind to sites 2 through 6 modify channel activation and/or open-state inactivation. Of the toxins
that modify channel kinetics, those that bind to sites 2 and 3 have
been studied most extensively. Site 2 toxins, such as batrachotoxin
(Bartels-Bernal et al., 1977), are alkaloids that bind to open
channels. These toxins gain access to the receptor from the
intracellular side of the channel or possibly through the lipid phase
of the channel. Site 3 toxins are peptides from scorpion (Rogers et
al., 1996) and sea anemone (Benzinger et al., 1998) venoms that bind to
a receptor on the extracellular face of the channel.
Neurotoxins that interact with site 2 on the sodium channel are an
interesting set of alkaloids because they can be found in both animal
and plant species. Batrachotoxin, from the skin of South American
Phyllobates sp. frogs, veratridine (lily), grayanotoxin (rhododendron), and aconitine (Aconitum sp.) markedly
affect sodium channel behavior by shifting the voltage dependence of
activation to more hyperpolarized potentials, by removing or reducing
channel inactivation from the open state, and by altering the ionic
selectivity of the channel (Strichartz et al., 1987; Catterall et al.,
1992). Channel modification by site 2 toxins is generally considered to
be irreversible (Hille, 1992) although the effects of veratridine modification (reviewed by Ulbricht, 1998) can be partially removed.
Although the principal ingredient in Aconitum toxin is
aconitine, some species also contain the structurally related
lappaconitine (Fig. 1). Both toxins have
a skeleton consisting of six-, seven-, and five-membered rings.
Lappaconitine has a benzoyl linkage on the six-membered ring at C18,
whereas aconitine has a benzoyl linkage on the five-membered ring at
C14 (Hardick et al., 1996). Despite the structural similarity to
aconitine, lappaconitine seems to be a sodium channel blocker rather
than an agonist. For example, lappaconitine inhibits the population
spike in the rat hippocampal slice preparation (Ameri et al., 1996a;
Ameri and Simmet, 1999) and displays a use- and frequency-dependent
block of the population spike that is reversible upon washing (Ameri et
al., 1996b). Furthermore, lappaconitine raises the stimulation threshold of electrically stimulated guinea pig atria (Heubach and
Schule, 1998) and induces use-dependent arrhythmia in guinea pig heart,
suggesting that lappaconitine has local anesthetic-like properties
(Gutser et al., 1998; Heubach and Schule, 1998).
The present study focused on lappaconitine block of the whole-cell
current of transiently expressed human heart sodium channels (hH1;
Gellens et al., 1992). Several pieces of evidence indicated that
lappaconitine binds primarily to open channels. For example, lappaconitine blocked the channels during repetitive stimulation but
had little effect on either resting or inactivated channels. Use-dependent block by lappaconitine did not reach steady state even
after several hundred pulses, and the rate and extent of use-dependent
block did not depend on the stimulation rate (0.2 to 2 Hz). Unlike
other naturally occurring sodium channel blockers, block by
lappaconitine was irreversible, suggesting that lappaconitine does not
bind to an extracellular receptor. Several studies have indicated that
local anesthetics and site 2 neurotoxins have overlapping but
nonidentical binding domains (Linford et al., 1998; Wang and Wang,
1999). To determine whether the lappaconitine binding site overlaps
with the local anesthetic receptor, channels containing lysine
substitutions within the local anesthetic receptor (F1760K, N1765K)
were assayed for their sensitivity to bupivacaine and lappaconitine.
Both mutants were highly resistant to block by either bupivacaine or
lappaconitine, suggesting that lappaconitine irreversibly blocks hH1
channels by binding to the site 2 receptor.
| |
Materials and Methods |
Site-Directed Mutagenesis and Transient Transfection of HEK293t
Cells.
Site-directed mutagenesis was used to create lysine point
mutations of the hH1-pcDNA I/amp vector (Chahine et al., 1996) at residues hH1-F1760 and hH1-N1765 as described previously for rat skeletal muscle (µ1) sodium channels (Wright et al., 1998). Human embryonic kidney (HEK) 293t cells were transfected with hH1 or mutant
channel plasmid (2-5 µg) and reporter plasmid CD8-pih3m (1 µg) by
the calcium phosphate precipitation method (Cannon and Strittmatter,
1993) as described previously (Wright et al., 1997, 1999). The
transfected cells were replated onto 35-mm culture dishes and used for
experiments up to 3 days. Transfection-positive cells, as identified by
CD8 Dynabeads (Dynal, Inc., Lake Success, NY), were selected for
whole-cell patch recording.
Solutions and Chemicals.
The extracellular solution used to
perfuse HEK cells contained 65 mM NaCl, 85 mM choline Cl, 2 mM
CaCl2, and 10 mM HEPES (titrated with
tetramethylammonium hydroxide to pH 7.4). For most experiments in which
a reversed sodium gradient was used, the pipette solution contained 100 mM NaF, 30 mM NaCl, 10 mM EGTA, and 10 mM HEPES. For experiments in
which a more conventional sodium gradient was used, the pipette
solution contained 10 mM NaCl, 120 mM CsF, 10 mM EGTA, and 10 mM HEPES
(titrated with cesium hydroxide to pH 7.2). Lappaconitine hydrobromide
was purchased from CalBiochem (San Diego, CA), and racemic bupivacaine
was purchased from Sigma Chemical Co. (St. Louis, MO). Lappaconitine
was dissolved in 50% ethanol and was stored in 10 mM aliquots at
80°C; bupivacaine was dissolved in deionized water and was stored
in 100 mM aliquots at 20°C. The stock solutions were diluted in
extracellular saline at the appropriate concentration and were applied
to the cell surface from a series of small-bore glass capillary tubes.
Whole-Cell Voltage Clamp and Data Analysis.
Whole-cell
voltage clamp (Hamill et al., 1981) of transfected HEK cells was used
to study macroscopic hH1 sodium currents at room temperature (21 ± 2°C). Electrode resistances ranged from 0.5 to 1.0 M
. Command
voltages were programmed by pCLAMP 7.0 software (Axon Instruments,
Burlingame, CA) and were delivered by a Warner PC501A voltage clamp
(Warner Instrument Corporation, Hamden, CT). Data were sampled at 50 kHz and filtered at 5 kHz. The holding potential for all experiments
was 140 mV. A previous study examined the time-dependent negative
shift in the steady state inactivation curve of heterologously
expressed voltage-gated sodium channels (Wang et al., 1996). As
described in previous papers (Wright et al., 1997, 1999), after
establishment of whole-cell voltage clamp the cells were dialyzed for
25 to 30 min before acquiring data. According to estimates by Wang et
al. (1996), the steady-state inactivation curve would have shifted by
about 5 to 7 mV during the time course (20-30 min) of an experiment. Most of the capacitative current was cancelled by the voltage clamp
circuitry, and the remaining capacitative artifact and the leakage
current were subtracted by the P/4 method. Leakage and capacitance
current subtraction (P/4) was not employed in studies of 2-Hz
use-dependent block. Voltage errors of 
5 mV during test pulses to +30
mV were considered acceptable for the drug study (Bean, 1992).
Least-squares curve fitting was performed with Origin software
(Microcal, Northampton, MA). Statistical analyses (Student's t test) were performed using SigmaStat (Jandel Scientific
Software, San Rafael, CA) to determine the significance of changes in
mean values; p values of < 0.05 were considered
statistically significant. Unless noted otherwise, data are presented
as mean ± S.E.M.
| |
Results |
Lappaconitine
Irreversibly Blocks hH1 Channels.
Application of 10 to 100 µM
lappaconitine produced an initial reduction in the whole-cell current
(Fig. 2A, wash on), but the percentage of
available channels slowly decreased during the 5-min external perfusion
and never reached steady state. This phenomenon was particularly
noticeable at the 60 µM (Fig. 2A, 
) and 100 µM (Fig. 2A, 
)
concentrations. To determine whether the whole-cell current could be
further reduced, the cells were given a 2 Hz train of 100 pulses (5 ms
duration) to +30 mV. The pulse protocol reduced the percentage of
available current and the reduction was concentration dependent. After
2 Hz stimulation, the first test pulse was delivered in the presence of
the drug (at 7 min in Fig. 2A) and during the next 5 min, subsequent
pulses were delivered as the drug was washed off by control external solution. At each lappaconitine concentration, the 5-min wash recovered
only 5 to 7% of the blocked channels.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 2.
Irreversible block of whole-cell hH1 currents by
lappaconitine. A, lappaconitine was applied to the bath (arrow, wash
on) at 10 µM ( , n = 5), 30 µM ( ,
n = 5), 60 µM (, n = 6),
or 100 µM (, n = 9), and the available current
was measured during test pulses to +30 mV. Block of the hH1 current
increased slowly during the initial 5-min perfusion but was markedly
enhanced by delivery of a 2 Hz train of 100 pulses to +30 mV (arrow,
100 pulses at 2 Hz). The first pulse after the 2-Hz train was delivered
while the cell was perfused with lappaconitine. The cells were then
washed by control external saline for 5 min (arrow, wash
off) but the blocked current did not recover. B, the traces on the left
are whole-cell hH1 currents recorded in control external saline and
after 5 min ofexposure to 100 µM lappaconitine. The cell was then given a
2-Hz train of 100 pulses to +30 mV (broken arrow). The traces on the
right show the first pulse recorded after the 2 Hz train (100 µM) and
three superimposed traces recorded in control saline. The traces in
control saline were recorded after 5 and 10 min of washing in control
saline and after delivery of 200 additional pulses to +30 mV in control
saline. Scale, 1 nA, 1 ms. C, normalized membrane conductance (circles)
and normalized steady-state availability (squares) were measured before
(,) and after ( , ) use-dependent block by 30 µM
lappaconitine. Membrane conductance (gm) was
measured during 10-ms voltage steps to the voltages indicated on the
abscissa; gm was determined from the
equation gm = INa/(Em ENa), and the plot
was fitted with an empirical Boltzmann function. The midpoint voltage
(V0.5) and slope (k) of the Boltzmann
function used to fit the data were 48.3 ± 2.4 mV and 11.8 ± 0.6 mV, respectively, for the control data (n = 4, ). After 2 Hz stimulation in 30 µM lappaconitine
(n = 4, ) these values were 50.5 ± 2.4 mV and 13.3 ± 0.4 mV, respectively. The differences between the
V0.5 and k values of the function before and
after lappaconitine block were not significant (p > 0.05). The currents in the inset were recorded before (top traces)
and after (bottom traces) 2-Hz use-dependent block by 30 µM
lappaconitine. Scale, 1 nA, 1 ms. To determine the steady-state
availability of the channels, the cells were given 100 ms
conditioning pulses to the voltage indicated on the abscissa and
current availability was measured during a test pulse to +30 mV. The
V0.5 and k values of the Boltzmann function
used to fit the data were 103.6 ± 0.8 mV and 9.6 ± 0.6 mV, respectively, for the control data (n = 5, ). After 2 Hz stimulation in 30 µM lappaconitine
(n = 5, ) these values were 108.3 ± 1.6 mV and 9.4 ± 0.3 mV, respectively. Although the change in the
midpoint voltage of inactivation was significant (p = 0.03), the negative shift in channel availability after use-dependent
block by lappaconitine was most probably caused by the time-dependent
shift in steady-state inactivation (Wang et al., 1996).
|
|
Additional washing and delivery of repetitive pulses in control
external solution failed to recover additional amounts of the blocked
current. In Fig. 2B, the traces on the left show the reduction of the
current after a 5-min perfusion of 100 µM lappaconitine, and the
traces on the right show the remaining current after 2 Hz stimulation.
The smaller trace on the right was the first pulse recorded in 100 µM
lappaconitine after the 2 Hz stimulation, and the three larger traces
(superimposed) were recorded in control external solution. Two of the
superimposed traces were recorded after washing the cell with control
external solution for 5 and 10 min. The third trace was recorded after
delivering 200 additional pulses to +30 mV in control solution. In
three other cells, delivery of 100 additional pulses to 20 mV, which
elicited a small inward current, did not increase the available
fraction of channels (not shown). These data indicated that once bound
to the channel, lappaconitine unbinds very slowly if at all.
Site 2 neurotoxins commonly induce a hyperpolarizing
shift in channel activation. Figure 2C shows that lappaconitine had
little affect on the kinetics of available channels. The
conductance-voltage relationship and steady-state availability were
determined in control external solution and after delivery of 100 pulses to +30 mV in 30 µM lappaconitine. To determine the
conductance-voltage relationship, the cells were depolarized to the
voltage indicated on the abscissa from a holding potential of 140 mV.
Fig. 2C, inset, shows the currents from a representative cell before
(Fig. 2C, inset, top) and after (Fig. 2C, inset, bottom) use-dependent block by 30 µM lappaconitine. In contrast to aconitine, which markedly slows the inactivation rate of the current, lappaconitine had
no obvious effect on the time-constant of current decay during channel
inactivation. The steady-state availability of the current (Fig. 2C,
squares) also was little affected by lappaconitine. To determine
channel availability, 100-ms conditioning pulses to the voltage
indicated on the abscissa were delivered to the cells and a subsequent
test pulse to +30 mV was used to measure the available current.
Although the midpoint voltage of steady-state availability was 4 mV
more negative after lappaconitine block, the hyperpolarization was most
likely caused by the time-dependent shift in steady-state inactivation
that is characteristic of these channels (Wang et al., 1996). Thus, in
contrast to the actions of other site 2 neurotoxins, lappaconitine
blocks hH1 sodium channels and does not alter channel kinetics.
Use-Dependent Block by Lappaconitine is Dose-Dependent but Not
Frequency Dependent.
To examine the kinetics of use-dependent
block by lappaconitine, the relative amplitude of the current during
2-Hz stimulation was plotted versus the pulse number (Fig.
3A). The extent of block by lappaconitine
was dose-dependent, and the time course of the reduction in available
current developed very slowly, particularly at the lower
concentrations. The block did not reach steady state even at 100 µM,
and delivery of an additional 200 pulses completely blocked the current
(n = 3; data not shown).
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 3.
Use-dependent block of hH1 current by lappaconitine.
A, use-dependent block was monitored during a 2-Hz train of 100 pulses
to +30 mV. In control saline, the percentage of available channels did
not change during the 100-pulse train. At the 100th pulse of the 2-Hz
stimulation, the current amplitudes in 10 µM, 30 µM, 60 µM, and
100 µM lappaconitine were 89 ± 2%, (n = 5), 71 ± 2% (n = 5), 46 ± 1%
(n = 5), and 34 ± 4% (n = 4), respectively, of the first pulse. B, block of hH1 current at +30
mV by 100 µM lappaconitine during 2-Hz (left) and 0.2-Hz (right)
stimulation. For clarity, only the first pulse and every 25th pulse of
the 100-pulse protocol are shown. Block of the current was very similar
despite the large difference in stimulation frequency. C, use-dependent
block by 30 µM bupivacaine (left) and 30 µM lappaconitine (right).
For bupivacaine block, the 1st, 3rd, and every 10th pulse of the
60-pulse protocol are shown; for lappaconitine block, the 1st and every
100th pulse of the 500-pulse protocol are shown. Note that
use-dependent block by bupivacaine reached steady state by the 20th
pulse, whereas block by lappaconitine did not reach steady state even
after 500 pulses. Scale in B and C, 1 nA, 1 ms.
|
|
Interestingly, the stimulation rate did not affect the extent of
use-dependent block. Figure 3B shows currents recorded at 25 pulse
intervals during 2-Hz (Fig. 3B, left) and 0.2 Hz (Fig. 3B, right)
stimulation. The development and extent of use-dependent block by 100 µM lappaconitine was very similar despite the 10-fold difference in
stimulation rate. At the 100th pulse, the fraction of available
channels was 34 ± 4% (n = 4) at 2 Hz and was
26 ± 4% (n = 4, p > 0.05) at
0.2 Hz. Note that the cells were exposed to lappaconitine for more than
8 min at the 0.2 Hz stimulation rate compared with only 50 s at 2 Hz, suggesting that the drug has limited access to the receptor when
the channel is in the closed resting state.
The traces in Fig. 3C compare the 2 Hz use-dependent block of hH1
channels by 30 µM bupivacaine (Fig. 3C, left) to the block by 30 µM
lappaconitine (Fig. 3C, right). Bupivacaine block increased rapidly and
reached steady state by the 20th pulse of the 60 pulse protocol (n = 3). For tertiary amine local anesthetics, the steady-state phase of use-dependent block represents an equilibrium between drug binding at each pulse and unbinding of the
drug during the interval between the pulses (Chernoff and Strichartz,
1989). In contrast, use-dependent block of the current by 30 µM
lappaconitine did not reach steady-state even after 500 pulses to +30
mV. In fact, the amount of available current was reduced after each set
of 100 pulses by increments of about 28% (n = 3).
Apparently, lappaconitine irreversibly blocks a small percentage of
open channels during each pulse, thereby eliminating them from the
population of available channels at subsequent pulses. Thus, the extent
of block at a given lappaconitine concentration depended on the number
of times the channels opened.
Lappaconitine Binding to Inactivated Channels Is Minimal.
The
fact that use-dependent block by lappaconitine did not depend on the
stimulation rate suggested initially that channel opening and perhaps
channel activation were important for drug binding and block. However,
these data did not distinguish whether or not lappaconitine binding to
inactivated channels during 5 ms pulses to +30 mV contributed to
channel blockade. To determine whether or not inactivated channels
became blocked by lappaconitine, the cells were given three 10-s pulses
to 70 mV. The conditioning pulses to 70 mV, which inactivated
almost all of the channels (see Fig. 2C), were separated by 30 s
at the holding potential to allow the channels to recover from small
amounts of slow inactivation. Figure 4A
shows that 100 µM lappaconitine blocked hH1 channels to a much lesser
degree during 10-s conditioning pulses to 70 mV () than during 2 Hz stimulation to +30 mV (). The available current after washout of
the drug was about the same as the available current before delivery of
the inactivating pulses to 70 mV. Furthermore, the conditioning
pulses to 70 mV would almost certainly have caused a large fraction
of the channels to enter preactivated states, yet lappaconitine had
little effect on the available current. These data suggested that
lappaconitine is essentially incapable of binding to inactivated
channels and like aconitine and other site 2 neurotoxins, the drug
gains access to the receptor primarily when the channel is open.
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 4.
Insensitivity of inactivated channels to
lappaconitine and the effect of the sodium gradient on the irreversible
block. A, the same general procedures described in Fig. 2A were
followed to examine block of inactivated channels by 100 µM
lappaconitine. The inset in the figure shows the pulse protocol
delivered after the 5-min perfusion of drug. Block of inactivated
channels () was measured by delivering three 10-s conditioning
pulses to 70 mV and then measuring the available current during a
test pulse to +30 mV. The conditioning pulses were separated by 30 s to allow the channels to recover from small amounts of slow
inactivation. The amount of available current was then monitored during
test pulses to +30 mV in control saline. Note that 100 µM
lappaconitine blocked a much larger percentage of channels during a
100-pulse train to +30 mV (, same data as Fig. 2A) than during 10-s
depolarizations to 70 mV. Furthermore, 100 µM lappaconitine blocked
a smaller percentage of channels when the driving force on sodium was
inward at +30 mV (, 65 mM [Na]o/10 mM
[Na]i) compared with when the driving force on sodium was
outward at +30 mV (, 65 mM [Na]o/130 mM
[Na]i). Nevertheless, the extent of block under either
condition was irreversible. B, comparison of use-dependent block when
the driving force on sodium was outward (left) or inward (right) at +30
mV. During a 2-Hz train of 100 pulses, 100 µM lappaconitine blocked
about 75% of the current when the driving force was outward compared
with about 45% when the driving force was inward. The current
amplitudes were measured from the steady-state current levels (dashed
lines) following open channel inactivation. Scale, 500 pA, 1 ms.
|
|
To determine whether the reversed sodium gradient (65 mM
[Na]o/130 mM [Na]i)
contributed to the irreversible block, use-dependent block of the
channels was examined under conditions in which the driving force on
sodium was inward at +30 mV. The squares in Fig. 4A represent the
averaged data from three experiments in which 100 µM lappaconitine
block was measured using a 65 mM [Na]o/10 mM
[Na]i gradient. As in the experiments with the
reversed sodium gradient, channel availability was measured during test
pulses to +30 mV. Under these conditions block of the resting channels during the initial 5-min perfusion of lappaconitine did not steadily increase, as was consistently observed in the reversed sodium gradient,
and the fraction of available channels at the conclusion of the 5 min
perfusion was greater at 88 ± 3%. In addition, a larger fraction
of channels was available after 2 Hz stimulation to +30 mV and
subsequent 5-min wash in control external solution (42 ± 3%,
n = 3) than were available after use-dependent block in
the reversed sodium gradient (28 ± 3%, n = 9).
Furthermore, an inward driving force on sodium at +30 mV reduced the
extent of use-dependent block by lappaconitine (Fig. 4B). Note that
although the extent of use-dependent block was reduced in the 65 mM
[Na]o/10 mM [Na]i
gradient, blocked channels did not recover after several minutes of
washing in control saline, indicating that the reversed sodium gradient
was not responsible for the irreversible block. As demonstrated for
quaternary derivatives of lidocaine (Cahalan and Almers, 1979), these
data showed that external sodium ions influence lappaconitine binding.
Bupivacaine Inhibits Lappaconitine Block.
Previous studies
have demonstrated that local anesthetics allosterically inhibit
batrachotoxin binding to voltage-gated sodium channels (Postma and
Catterall, 1984; Wang and Wang, 1999); that is, channels blocked by
local anesthetic cannot open and become modified by batrachotoxin. As
shown in Figs. 2-4, lappaconitine binds weakly to resting or
inactivated channels, and the kinetics of use-dependent block are very
slow. In contrast, the onset of use-dependent block by bupivacaine is
relatively rapid because local anesthetics have a high affinity for
open channels as well as inactivated channels (Hille, 1992). To examine
whether use-dependent block by bupivacaine could reduce lappaconitine
binding, the two drugs were applied together before delivery of the
2-Hz pulse protocol. In these experiments, the bupivacaine
concentration was varied and the lappaconitine concentration was always
100 µM. The traces in Fig. 5A show the
irreversible block produced by application of 100 µM lappaconitine
alone (Fig 5A, top), and the reduction in irreversible block when the
channels were exposed to both 100 µM lappaconitine and 8 µM
bupivacaine (Fig 5A, middle) or to 100 µM lappaconitine and 300 µM
bupivacaine (Fig 5A, bottom). Delivery of 100 pulses at 2 Hz again
reduced the current, but a much larger fraction of channels recovered
at the higher bupivacaine concentration (Fig. 5A, right traces). The
plot in Fig. 5B shows the time course of these experiments for
bupivacaine concentrations of 1, 8, 60, and 300 µM. These data
clearly demonstrated that block of the channels by bupivacaine
inhibited lappaconitine binding as measured by the reduction in
irreversible block by lappaconitine.
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 5.
Bupivacaine inhibits the irreversible block by
lappaconitine. A, the traces on the left of each panel show the tonic
block after 3-min perfusion of 100 µM lappaconitine alone (top), 100 µM lappaconitine and 8 µM bupivacaine (middle), and 100 µM
lappaconitine and 300 µM bupivacaine (bottom). At the conclusion of
the 3-min perfusion, the cells were given a 2-Hz train of 100 pulses.
The traces on the right of each panel show the first pulse recorded
immediately after the 2 Hz train (unlabeled) and after a 5-min wash in
control saline. Increases in the concentration of bupivacaine increased
the fraction of channels that were protected from irreversible block by
lappaconitine. Scale in all panels, 1 nA, 2 ms. B, average data for
experiments like those shown in A. In addition to perfusion of 100 µM
lappaconitine alone (), bupivacaine (open symbols) was applied
(along with 100 µM lappaconitine) at four concentrations [1 µM,
(n = 4); 8 µM, (n = 4); 60 µM,  (n = 5); 300 µM, (n = 5)]. Increases in the concentration of
bupivacaine reduced the fraction of channels that were irreversibly
blocked by 100 µM lappaconitine during 2 Hz stimulation.
|
|
Fig. 6 shows the dose dependence of the
interaction between bupivacaine and lappaconitine. Each data set was
best fitted by the Hill equation and had a Hill coefficient of about 1, indicating that one drug molecule binds to one channel. The
IC50 values of resting block (at 140 mV) by
bupivacaine alone (Fig. 6, ) and resting block by bupivacaine plus
100 µM lappaconitine (Fig. 6, 
) were similar at 65 and 73 µM,
respectively. To determine the extent of use-dependent block by
bupivacaine alone, the cells were given a 2 Hz train of 100 pulses (5 ms in duration). The available current during the steady-state phase of
2 Hz block by bupivacaine alone (Fig. 6, ) was normalized to the
current amplitude measured in control external solution. The
IC50 value for bupivacaine during this summed
block of resting channels and channels blocked during the 2 Hz train
was 9 µM. As shown in Fig. 5B, the fraction of channels recovered
after the use-dependent protocol increased with increasing bupivacaine
concentration. For each bupivacaine concentration, the fraction of
available channels at the conclusion of the 5-min wash in control
external solution was renormalized according to the fraction of
available channels after use-dependent block by 100 µM lappaconitine
alone. These data were plotted as the fraction of recovered channels in
Fig. 6 (right ordinate, ). The IC50 value for
the fraction of recovered channels after use-dependent block by
bupivacaine and lappaconitine was 12 µM, very similar to the
IC50 value of the summed block by bupivacaine
alone. As one would expect, the accumulation of bupivacaine block
(resting block and use-dependent block) at each concentration reduced
the fraction of available channels to which lappaconitine could bind.
Although tonic block of the channels by bupivacaine undoubtedly
contributed to the fraction of recovered channels, the more rapid
binding of bupivacaine during the use-dependent protocol made the
largest contribution to the increased recovery from block.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 6.
Bupivacaine dose-response curves. The solid symbols
on the plot correspond to the normalized block of the current (left
ordinate) at each bupivacaine concentration. Resting block (140 mV)
by bupivacaine () and by bupivacaine plus 100 µM lappaconitine
() had similar IC50 values of 65 ± 6 µM (Hill
coefficient, h = 1.1 ± 0.1) and 73 ± 14 µM (h = 1.1 ± 0.2), respectively. The plot also contains the
dose-response data for steady-state use-dependent block by bupivacaine
alone (). The available current during the steady-state phase of
2-Hz use-dependent block was normalized to the available current in
control saline. The IC50 value for this sum of resting and
phasic block under these conditions was 9.3 ± 1.2 µM (h = 0.8 ± 0.1). The triangles represent the fraction of channels
recovered (right ordinate) at each concentration of bupivacaine
following use-dependent block (in the presence of 100 µM
lappaconitine). The data at each concentration of bupivacaine are the
data points shown at the conclusion of the 5-min wash in Fig. 5B (i.e.,
10-min time point). The normalized fraction of available current at
each bupivacaine concentration was renormalized according to the
fraction of channels available after block by 100 µM lappaconitine
alone and subsequent 5 min wash (Fig. 5B, , 10-min time point). The
IC50 value for the fractional recovery of channels was
11.9 ± 0.9 µM (h = 0.9 ± 0.1), and was thus similar
to the IC50 value for the summed block (resting plus
use-dependent block) by bupivacaine alone. For all curves, the errors
for the IC50 values and Hill coefficients are the errors of
the best fit to the data.
|
|
Lysine Point Mutations in the Local Anesthetic Receptor Region
Render hH1 Channels Resistant to Lappaconitine.
To establish the
proximity of the binding sites for lappaconitine and bupivacaine, I
examined bupivacaine and lappaconitine block using channels containing
point mutations within the local anesthetic receptor region. As
demonstrated for the µ1 isoform (Wright et al., 1998), hH1 channels
containing lysine substitutions at F1760 (µ1-F1579K) or N1765
(µ1-N1584K) expressed well in HEK293t cells (Fig.
7A). Figure 7B shows the normalized
conductance-voltage relationships, and Fig. 7C shows the normalized
steady-state availability relationships of wild-type hH1 channels and
the two lysine mutants. As with the comparable mutations in µ1
channels (Wright et al., 1998), the midpoint voltages of the
conductance-voltage and steady-state availability relationships for the
mutant channels were significantly different (p < 0.05) from those of the wild-type channels. Nevertheless, the channels
otherwise appeared to open and inactivate normally. In contrast to the
comparable lysine mutation in the µ1 isoform (µ1-N1584K),
hH1-N1765K channels did not exhibit an unusually large maintained
current at +30 mV.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 7.
Normalized conductance-voltage and steady-state
inactivation relationships of hH1 channels containing lysine
substitutions. A, channels containing lysine substitutions within the
local anesthetic receptor at residues F1760 or N1765 express well in
HEK cells. The currents were recorded during 10-ms voltage commands
ranging in amplitude from 100 mV to +50 mV. Scale, 1 nA, 2 ms. B,
normalized membrane conductance plotted versus the amplitude of the
10-ms voltage step. As described in the Fig. 2 legend, the data were
fitted with a standard Boltzmann function to obtain the midpoint
voltage (V0.5) and slope (k) of the data.
For hH1 (n = 7), the mean V0.5 and
k values were 49.0 ± 1.4 mV and 10.6 ± 0.8 mV, respectively. For F1760K (n = 7), the mean
V0.5 and k values were 31.2 ± 2.6 mV* and 14.4 ± 0.5 mV*, respectively. For N1765K
(n = 8), these values were 43.2 ± 0.9 mV*
and 9.1 ± 0.8 mV. * indicates p < 0.05 compared with hH1. C, normalized steady-state availability function
(h ) for the channels. Cells were held at 140 mV and
were given 100-ms conditioning pulses ranging in amplitude from 160
mV to 35 mV followed by a test pulse to +30 mV. The mean
V0.5 values (50% availability) and k values
for the fitted Boltzmann functions were: 101.2 ± 1.2 mV and
8.9 ± 0.4 mV, respectively for hH1 channels
(n = 10). The V0.5 and k
values for F1760K channels (n = 9) were
106.2 ± 1.0 mV and 6.1 ± 0.3 mV, respectively, and for
N1765K channels (n = 6) these values were
87.5 ± 1.5 mV and 6.1 ± 0.2 mV, respectively. The mean
V0.5 and k values of both mutants were all
significantly different from the mean values for hH1
(p < 0.05).
|
|
The sensitivity of F1760K and N1765K channels to racemic bupivacaine
was first examined to determine whether the lysine mutations conferred
resistance to local anesthetic block. Bupivacaine (30 µM) blocked
approximately 45% of hH1 channels during the steady-state phase of
2-Hz use-dependent block (Fig. 8A). In
contrast, both F1760K and N1765K channels were extremely resistant to
use-dependent block. The lysine point mutations also reduced
steady-state bupivacaine block of resting channels and inactivated
channels (Fig. 8B). To determine resting channel block, a 10-s
conditioning pulse to 160 mV was followed by 100 ms at the holding
potential and the available current was measured during a subsequent
test pulse to +30 mV. As shown in the traces in Fig. 8B, left, resting
F1760K and N1765K channels were about half as sensitive to 30 µM
bupivacaine compared with hH1 channels. A similar pulse protocol was
used to measure inactivated channel block except that the 10-s
conditioning pulse was to 70 mV. The conditioning pulse to 70
allowed bupivacaine binding to inactivated channels to reach
steady-state, and the 100-ms interval at the holding potential
permitted inactivated but drug-free channels to recover from fast
inactivation (Wright et al., 1997). Compared with hH1 channels,
inactivated F1760K and N1765K channels were > 20-fold less
sensitive to 30 µM bupivacaine (right column in Fig. 8B). These data
were consistent with the >20-fold reduction in cocaine block of
inactivated µ1-F1579K and N1584K channels (Wright et al., 1998).
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 8.
Resistance of lysine mutants to block by 30 µM
bupivacaine. A, the plot shows the use-dependent block of hH1 channels
by 30 µM bupivacaine. The fraction of hH1 channels blocked during the
steady-state phase of the 60 pulse protocol was 45%. In contrast, 30 µM bupivacaine reduced the fraction of available F1760K and N1765K
channels by only 1-2%. B, steady-state block of hH1 and lysine
mutants by 30 µM bupivacaine. The pulse protocol consisted of a 10-s
conditioning pulse to 160 mV or to 70 mV followed by a 100 ms
interval at 140 mV, and a subsequent test pulse to +30 mV. The
conditioning pulse to 160 mV was used to estimate resting channel
block, whereas the conditioning pulse to 70 mV was used to estimate
inactivated channel block. The traces show the available current in
control saline (dashed traces) and in 30 µM bupivacaine (solid
traces). Bupivacaine blocked 26.6 ± 1.8% of resting hH1 channels
(n = 4) and blocked 10.8 ± 3.2% and
13.9 ± 2.5% of resting F1760K (n = 3) and
N1765K channels (n = 5), respectively. Bupivacaine
blocked 87.5 ± 1.0% of inactivated hH1 channels
(n = 4) and blocked 18.5 ± 6.3% and
24.0 ± 4.2% of inactivated F1760K (n = 3)
and N1765K channels (n = 5), respectively. Compared
with hH1 channels, the block of the mutant channels by 30 µM
bupivacaine was significantly less (p < 0.05).
Scale, 500 pA, 1 ms.
|
|
In addition to reducing local anesthetic block, lysine substitutions at
homologous residues in the µ1 isoform eliminate channel modification
by batrachotoxin (Wang and Wang, 1999). To confirm that the lysine
substitutions also conferred resistance to lappaconitine block, the
channels were exposed to 100 µM lappaconitine for 5 min and then
given a 2-Hz train of 100 pulses to +30 mV. Figure 9 shows that both F1760K and N1765K
channels were highly resistant to use-dependent block by 100 µM
lappaconitine. Thus, the substitution of lysine for native residues in
the local anesthetic receptor region reduced block by a local
anesthetic as well as lappaconitine. The relationship between the local
anesthetic receptor region and site 2 neurotoxin binding will be
addressed under Discussion.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 9.
Resistance of lysine mutants to block by 100 µM
lappaconitine. As shown in Fig. 3B (different cell), 100 µM
lappaconitine markedly reduced the available hH1 current during a 2-Hz
train of 100 pulses. In contrast, the same concentration of
lappaconitine had little effect on the available current of F1760K
(n = 5) or N1765K (n = 3)
channels. For clarity, only the 1st and every 25th pulse of the
100-pulse protocol are shown. Scale, 500 pA, 1 ms.
|
|
| |
Discussion |
Site 2 neurotoxins typically induce negative shifts in sodium
channel activation and reduce or eliminate channel inactivation from
the open state. This report focused on the interaction between the
plant alkaloid lappaconitine and hH1 sodium channels. Despite the
structural similarity to aconitine, lappaconitine irreversibly blocked
the channels and did not modify channel kinetics. Lappaconitine did not
interact with the site 1 receptor as do other naturally occurring
channel blockers such as tetrodotoxin. The results in the study
suggested that lappaconitine blocked hH1 channels by binding to the
site 2 receptor. Like most other site 2 toxins, lappaconitine 1) binds
only with open channels, 2) binds irreversibly, and 3) does not bind to
channels containing lysine substitutions within the local anesthetic receptor.
Site 1 neurotoxins such as tetrodotoxin or saxitoxin block sodium
channels by binding to a receptor on the extracellular region of the
domain 1 pore loop between segments 5 and 6 (Tomaselli et al., 1995).
As one would expect for toxins that bind on the extracellular surface
of the channel, block by site 1 toxins is almost always reversible. In
contrast, several minutes of washing in control external solution or
repetitive pulses delivered in control solution failed to alleviate
lappaconitine block. Thus, lappaconitine is a naturally occurring
sodium channel blocker that binds to a receptor that is not on the
extracellular surface of the channel.
Lappaconitine has antiepileptiform activity and exhibits a use- and
frequency-dependent antagonism of the population spike in hippocampal
slice preparations (Ameri et al., 1996a,b; Ameri and Simmet, 1999). In
addition, lappaconitine induces bradycardia in heart preparations
(Gutser et al., 1998; Heubach and Schule, 1998), suggesting that the
drug has class I local anesthetic properties. In retrospect, the idea
that lappaconitine has local anesthetic properties seems reasonable
given that local anesthetics, antiepileptiform drugs, and site 2 neurotoxins have binding domains that seem to overlap (Catterall,
1999). Several pieces of evidence in the present study indicated that
block of hH1 channels by lappaconitine differed from that of tertiary
amine local anesthetics and more closely resembled the block by
quaternary derivatives of lidocaine, such as QX-314 and QX-222.
Lappaconitine almost exclusively preferred open channels and had a low
affinity for resting or inactivated channels. Consistent with an open
channel block, altering the stimulation rate by an order of magnitude
did not markedly affect the time course or extent of use-dependent
block. Similarly, QX-314 and QX-222 bind preferentially to open
channels and have little affect on closed channels. The reversibility
of use-dependent block by QX compounds seems to differ depending on the
preparation. QX-222 block of cardiac sodium channels in dog Purkinje
cells is only partially reversible after several minutes of washing (Hanck et al., 1994). In contrast, frog (Strichartz, 1973) or squid
(Yeh and Tanguy, 1985) sodium channels recover in seconds to minutes
from QX-222 or QX-314 block, and the recovery from use-dependent block
by QX-314 can be accelerated by small repetitive depolarizations
(Strichartz, 1973). In contrast, lappaconitine block was essentially
irreversible during the time course of the experiments and >100 pulses
in control solution did not elicit any use-dependent unblock. These
data suggest that bound channels may become locked in a conformation
that ultimately prohibits reopening and drug unbinding.
Lappaconitine Binding within the Site 2 Domain.
The structures
of aconitine and lappaconitine are remarkably similar, which alone
suggests that the toxins share a binding domain. Evidence obtained from
other preparations supports the notion that the compounds share a
common receptor. For example, lappaconitine competitively inhibits
aconitine-induced increases in synaptosomal intracellular calcium
(Gutser et al., 1998) and aconitine-induced arrhythmia in guinea pig
heart (Heubach and Schule, 1998), suggesting that the toxins bind to
the same receptor or at least that the binding sites overlap.
Previous studies have demonstrated that local anesthetics
noncompetitively inhibit the binding of site 2 neurotoxins (Postma and
Catterall, 1984; Wang and Wang, 1999). Likewise, batrachotoxin modification lowers channel affinity for lidocaine (Zamponi et al.,
1993) and for (+) optical isomers of bupivacaine and cocaine (Wang and
Wang, 1992). Studies that have described the allosteric binding
interactions between batrachotoxin and local anesthetics benefited from
the fact that site 2 toxins typically activate the channels and remove
inactivation, whereas local anesthetics block the channels. In the
present report, lappaconitine exhibited use-dependent block of the
channels, albeit with much slower kinetics than use-dependent block by
bupivacaine. As shown in Fig. 6, the inhibition of lappaconitine
binding by bupivacaine was not caused exclusively by tonic bupivacaine
block. Rather, the more rapid kinetics of use-dependent bupivacaine
block protected blocked channels from lappaconitine access by removing
them from the fraction of available channels. Thus, as reflected in
Fig. 6, the dose-response data of use-dependent block by bupivacaine
alone and the fraction of channels protected from irreversible
lappaconitine block were similar. Although the conclusions from these
experiments are simplistic, the data indicated that bupivacaine binding
inhibits lappaconitine binding and further confirmed that lappaconitine
binds only to open channels. Although the precise mechanism for the
allosteric interactions between the binding sites for local anesthetics
and site 2 toxins has not been deduced, one possible explanation for the interaction between bupivacaine and lappaconitine is that bupivacaine binding induces an allosteric change in the channel that
inhibits lappaconitine binding. The inhibition of lappaconitine block
by bupivacaine could be explained if, at closed channels, the
anesthetic has better access to the local anesthetic receptor than
lappaconitine has to the site 2 receptor. More rapid bupivacaine binding during repetitive stimulation could then induce an allosteric change in channel conformation that inhibits toxin binding.
Alternatively, bupivacaine and lappaconitine may compete for binding
sites that are in close proximity. The fact that the mutant channels
were resistant to both bupivacaine and lappaconitine is consistent with
the idea that local anesthetics and site 2 neurotoxins bind to
nonidentical receptors within an overlapping region (Linford et al.,
1998).
Homologous amino acid residues within D4-S6 of rat brain (RBIIA;
Ragsdale et al., 1994) and rat skeletal muscle (µ1; Wright et al.,
1998) sodium channels are critical determinants for local anesthetic
block. In addition, alanine substitution within the local anesthetic
receptor at RBIIA-F1764 inhibits batrachotoxin modification of the
channels (Linford et al., 1998). Lysine substitution within the local
anesthetic receptor region of µ1 channels at residues µ1-F1579 or
N1584 inhibits modification of the channels by batrachotoxin (Wang and
Wang, 1999), and lysine substitution at µ1-F1579 reduces channel
modification by grayanotoxin (Kimura et al., 2000). Consistent with the
results obtained using other channel isoforms, hH1 channels containing
lysine substitution at F1760 or N1765 were resistant to block by
bupivacaine or lappaconitine. With respect to batrachotoxin binding,
the more critical of the two residues is probably hH1-N1765, which
faces away from the pore region according to the Ragsdale et al. (1994)
model of sodium channel D4-S6. Alanine substitution at hH1-N1765 or the
homologous site in µ1 channels (N1584) reduces batrachotoxin
modification to a much greater extent than does alanine substitution at
µ1-F1579 (Wang and Wang, 1999). The importance of sodium channel S6
segments is further supported by more recent studies that implicate a
neighboring residue within D4-S6 (µ1-V1583; Vedantham and Cannon,
2000), as well as residues within D3-S6 (Wang et al., 2000) as critical binding regions for batrachotoxin. In the present study, lysine substitution at hH1-F1760 may disrupt lappaconitine binding simply by
introducing a positive charge within the vicinity of the lappaconitine binding site. On the other hand, the mutation at N1765 may inhibit lappaconitine block because the residue is part of the site 2 receptor.
Although the data in the present study demonstrate that lappaconitine
binds to open channels, the mechanism of block is not entirely clear.
Indeed, no study has determined definitively the access route for site
2 neurotoxins. Like other site 2 toxins, the structure of lappaconitine
indicates that the drug is lipid soluble and can easily cross the
membrane. Once inside the cell, the toxin could access the receptor
using more than one pathway. For example, lappaconitine could simply
enter the pore from the intracellular side while the channel is open,
bind to the receptor, and block the pore. Alternatively, the toxin
could access the receptor through the lipid phase of the channel. The
conformational changes associated with channel opening would then
permit drug binding with the receptor, and the gating mechanism of
channels bound by the toxin could become locked in a configuration that prevents subsequent opening. One other alternative is a combination of
the two possibilities above. That is, the drug may enter the pore
during channel opening, bind to the receptor, and lock the channel in a
nonconducting state. Although I cannot rule out any of these
possibilities to explain lappaconitine binding and block, an allosteric
modification of the gating mechanism would be consistent with the
action of other site 2 neurotoxins.
The most abundant neurotoxin in Aconitum plants is usually
aconitine, with lappaconitine and several other toxins in the minority. The fact that the same plant may synthesize two compounds that bind to
the same receptor, but have opposing and essentially irreversible effects, is difficult to reconcile. The two toxins could target specific isoforms of sodium channel; however, this possibility has not
been fully explored. Although several of the side groups on the
diterpene skeleton may prove to be important in toxin action, the
location of the benzoyl linkage is particularly interesting and may
partially explain the contrasting effects of aconitine and
lappaconitine at voltage-gated sodium channels.
I thank R. Kallen (Department of Biochemistry and Biophysics,
University of Pennsylvania School of Medicine, Philadelphia, PA)
for the hH1 clone, and S.-Y. Wang (Department of Biology, State
University of New York at Albany, Albany, NY) and G. K. Wang
(Department of Anesthesia Research Laboratories, Harvard Medical School
and Brigham & Women's Hospital, Boston, MA) for generously supplying
the HEK293t cell line, CD8-pih3 m plasmid, and lysine mutants. I also
thank G. K. Wang for a critical reading of the manuscript.
This work was supported by grants from the Kentucky Academy of
Science and the National Institutes of Health (R15-GM60927).
![[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}
Top
Abstract
Introduction
-subunits expressed in tsA201 cells.
J Membr Biol
152:
39-48[Medline].
1 subunit.
Biophys J
76:
233-245
