![[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}
|
|
|
|
|
||
Vol. 290, Issue 2, 569-577, August 1999
Department of Pharmacology, Columbia University, New York, New York
 |
Abstract |
|---|
 Top
 Abstract
 Introduction
Materials and Methods
Results
Discussion
References
|
|---|
4-Aminopyridine (4AP) binding to rKv1.4 occurs preferentially in the activated state, whereas binding to rKv4.2 occurs in the rested state. To explore structural basis for the different state dependencies of 4AP binding, regions of rKv1.4 that are likely to form the 4AP-binding site and/or the activation gate were replaced by the corresponding rKv4.2 sequences one at a time, and the resulting effects on channel gating and 4AP binding were examined. Replacing the amino acid sequence of rKv1.4 in the intracellular loop between the fourth and fifth transmembrane segments (S4 and S5) with that of rKv4.2 did not alter channels' gating properties or the state dependence of 4AP binding. On the other hand, replacing the rKv1.4 sequence in the cytoplasmic half of S5 (N-S5) or S6 (C-S6) with that of rKv4.2 markedly altered the voltage dependence and kinetics of activation gate function. Importantly, these mutations transferred the rested-state 4AP-binding preference from the donor to the host channel. These data can be explained by a scheme in which the function of the activation gate determines the state dependence of 4AP binding, although the relationship between the binding site and the gate may be similar between rKv1.4 and rKv4.2. The amino acid sequences in the N-S5 and C-S6 domains contribute to this activation gate function.
| |
Introduction |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
4-Aminopyridine
(4AP) is a nonspecific K channel blocker that has been widely used to
block or characterize native K channels (Kirsch et al., 1986
; Simurda
et al., 1989; Kehl, 1990; Choquet and Korn, 1992; Campbell et al.,
1993; Castle and Slawsky, 1993; Wang et al., 1995) and cloned K
channels (Castle et al., 1994; McCormack et al., 1994; Russell et al.,
1994; Shieh and Kirsch, 1994; Yao and Tseng, 1994; Bouchard and Fedida,
1995; Rasmusson et al., 1995; Tseng et al., 1996). One of the most
interesting aspects of 4AP's actions is the variation in the "state
dependence" of 4AP binding to target channels. For some K channels,
4AP binding is facilitated by channel activation when the "activation
gate" is opened (Choquet and Korn, 1992; Castle et al., 1994;
McCormack et al., 1994; Yao and Tseng, 1994; Bouchard and Fedida,
1995). For others, 4AP binding occurs mainly in the rested state when the activation gate is closed, and channel activation induces 4AP
dissociation (Kehl, 1990; Campbell et al., 1993; Castle and Slawsky,
1993; Tseng et al., 1996). The goal of the present study was to explore
the structural basis for this variation in the state dependence of 4AP
binding. This information may be useful in the design of K
channel-blocking (class III) antiarrhythmic agents, for which the state
dependence of drug binding is an important issue (Tseng, 1994). The
information also may be useful in the structure-function analysis of K
channel gating.
For all K channels tested so far, 4AP binding occurs from the
cytoplasmic side of the cell membrane (Kirsch and Narahashi, 1983;
Wagoner and Oxford, 1990; Yao and Tseng, 1994; Bouchard and Fedida,
1995; Tseng et al., 1996). Mutagenesis studies indicate further that
the 4AP-binding site may be in the inner-mouth region of the pore
(Kirsch et al., 1993; Zhang et al., 1998). Structure-function analysis
of voltage-gated K channels have suggested that channels' activation
gates probably are formed by structures around the inner mouth of the
pore (Shieh et al., 1997; Holmgren et al., 1998; Zhou et al., 1998).
The proximity of the 4AP-binding site and the activation gate leads us
to propose two general hypotheses for why the state dependence of 4AP
binding differs among K channels. The first hypothesis is that the
relationship between the 4AP-binding site and the activation gate may
differ among channels. For example, in K channels to which 4AP binding
occurs in the activated state the binding site may be behind (on the
pore side of) the activation gate and thus is protected by the gate. In
other K channels to which 4AP binds preferentially in the rested state,
the 4AP-binding site may be in front (on the cytoplasmic side) of the
activation gate. 4AP binding is not restricted by the gate, and gate
opening reduces the 4AP binding affinity, causing 4AP to dissociate.
The second hypothesis is that the structure of the activation gate determines the state dependence of 4AP binding, although the
relationship between the 4AP-binding site and the activation gate may
be the same among K channels.
We took advantage of the availability of two K channel clones, rKv1.4
and rKv4.2, that manifest different state dependencies of 4AP binding.
For rKv1.4, 4AP binding requires channel activation (Yao and Tseng,
1994). On the other hand, 4AP binding to rKv4.2 seems to occur
exclusively in the rested state (Tseng et al., 1996). We designed
mutagenesis experiments to test the two hypotheses described above. The
assumption was that the 4AP-binding site and the activation gate could
be functionally assigned to specific domains of a channel. Therefore,
by replacing the amino acid sequences in individual domains of one
channel with the corresponding sequences of the other and measuring how
the activation gating function and the state dependence of 4AP binding
were altered, one could infer whether the state dependence of 4AP
binding was determined by the relationship between the binding site and
the activation gate or by the activation gate structure. As discussed
above, the 4AP-binding site may be formed by domains lining the inner mouth of the pore (Durell and Guy, 1996). These include the
intracellular loop between the fourth and fifth transmembrane segments
(S4-S5; Slesinger et al., 1993), the cytoplasmic half of S5 (N-S5;
Shieh and Kirsch, 1994), and the cytoplasmic half of S6 (C-S6; Choi et
al., 1993; Kirsch et al., 1993; Lopez et al., 1994; Zhang et al.,
1998). Furthermore, the N-S5 and C-S6 domains may participate in the
formation of the activation gate (Shieh et al., 1997; Zhou et al.,
1998; Holmgren et al., 1998). These domains are marked in the schematic
diagram of a K channel subunit in Fig.
1A. We replaced the amino acid sequences
in the S4-S5, N-S5, and C-S6 regions of rKv1.4 by the corresponding
sequences of rKv4.2 one at a time (Fig. 1B), and the resulting effects
on channel activation gating function and 4AP binding were examined.
|
| |
Materials and Methods |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
In Vitro Mutagenesis and Oocyte Preparation.
Oligonucleotide-directed mutagenesis was carried out on the
rKv1.4
3-25 background (Tseng-Crank et al., 1993) by using a
commercial mutagenesis kit according to the manufacturer's
instructions (Altered Sites System; Promega, Madison, WI). Mutations
were confirmed by direct DNA sequencing using the dideoxy-mediated
chain termination method.
3-25 were linearized with appropriate restriction enzymes and
used as templates for cRNA synthesis. In vitro transcription reactions
were carried out by using a commercial kit and T7 RNA polymerase
according to the manufacturer's instruction (mMessage mMachine;
Ambion, Austin, TX). The quality and quantity of cRNA product of each
transcription reaction were checked by denaturing agarose gel electrophoresis.
Preparation of Xenopus oocytes and injection of cRNA have been
described previously (Tseng-Crank et al., 1990). Four to six hours
after isolation, each oocyte was microinjected with 20 to 40 nl of a
cRNA solution. The oocytes then were incubated at 16°C in the
following medium for 2 to 4 days before voltage-clamp studies: 96 mM
NaCl, 2 mM KCl, 1.8 mM CaCl2, 1 mM
MgCl2, 5 mM HEPES, 2.5 mM Na-pyruvate, pH 7.5, with NaOH, supplemented with 50 U/ml penicillin, 50 µg/ml
streptomycin, 10 µg/ml gentamycin, and 4% horse serum.
Electrophysiological Experiments.
Whole-cell currents were
recorded using a modified two-microelectrode voltage-clamp method
(Schreibmayer et al., 1994). The voltage-sensing and the
current-passing pipettes both had a tip resistance of 0.1 to 0.3 M
.
An Oocyte Clamp (OC-725B) amplifier was used (Warner Instruments,
Hamden, CT). During current recordings, the oocytes were superfused
with a Cl-free solution to minimize interference from endogenous Cl
channel currents. The composition of this solution was as follows
unless otherwise stated: 96 mM NaOH, 2 mM KOH, 1.8 mM
Ca(OH)2, 1 mM MgSO4, 5 mM
HEPES, 2.5 mM Na-pyruvate, pH 7.5 with methanesulfonic acid. The bath
temperature was 23-25°C.
Data Acquisition and Analysis. The voltage-clamp protocols and methods of data analysis will be described in figure legends. Voltage-clamp protocol generation and data acquisition were controlled by a 486 IBM computer via pClamp software (Axon Instruments, Foster City, CA) and a 12-bit D/A and A/D converter (TL-1 DMA Interface; Axon Instruments). Currents were low-pass filtered with an eight-pole Bessel filter (Frequency Devices, Haverhill, MA) at 2 kHz, digitized, and stored on diskettes for off-line analysis. The sampling interval ranged from 0.1 to 2.5 ms depending on the voltage-clamp protocol. Data analysis was performed using Clampfit (version 6.1 of pClamp), Excel (Microsoft Corporation, Redmond, WA), and PeakFit (Jandel Scientific, Corte Madera, CA) programs. Data are presented as means and S.E. Statistical significance was tested by unpaired t test (SigmaStat 2.0; Jandel Scientific).
| |
Results |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
The mutations were made in a deletion mutant of rKv1.4
(rKv1.43-25) in which the fast inactivation process was disrupted (Tseng-Crank et al., 1993). This avoided the interference from the fast
inactivation process in the measurement of channels' activation
process and 4AP binding. The conclusions obtained from rKv1.43-25
are applicable to rKv1.4 because the deleted region (amino acids 3-25
in the NH2 terminus) is separated from the
domains involved in 4AP binding and activation gating. Indeed, we have shown that rKv1.4 and rKv1.43-25 have the same state dependence of
4AP binding and unbinding. They also have similar 4AP-binding affinity
when determined using a voltage-clamp protocol that avoided the
interference of the fast inactivation process in the WT channel (Yao
and Tseng, 1994).
Comparison of 4AP Actions on rKv1.43-25 and rKv4.2
State dependence of 4AP Binding and Unbinding.
Figure
2A (center) illustrates changes in the
amplitude and kinetics of rKv1.43-25 after the oocyte had been
superfused with a 4AP-containing (0.5 mM) solution for 10 min while the
membrane was held at 
80 mV. This holding voltage
(Vh) was 30 mV negative to the channel's
activation threshold (Tseng and Tseng-Crank, 1992). Therefore, during
superfusion with the 4AP-containing solution, most of the channels were
in the rested state (activation gate closed). After such a prolonged
4AP exposure, the first pulse induced a current with peak amplitude
only modestly reduced relative to the control (by 15%). However, the
current trace during the first pulse decayed at a much faster rate than
the slow inactivation seen under the control conditions. The current
amplitude at the end of this pulse was reduced by 62% relative to
control. These observations indicate that 4AP binding to rKv1.43-25
developed very slowly, if any, at 80 mV before channel activation,
but was accelerated greatly by membrane depolarization and channel activation. Therefore, 4AP binding to rKv1.43-25 displayed an "activated-state preference." The currents induced by pulses 2 and
3 applied 1 and 2 min after the first pulse were superimposable with
each other. Importantly, their peak amplitudes were reduced by 60%
relative to control, comparable to the level of current suppression
seen at the end of the first pulse. This suggests that 4AP blockade of
rKv1.43-25 approached a steady-state by the end of the first pulse.
Furthermore, there was little relief of 4AP blockade during the
interpulse interval between pulses 1 and 2 or between pulses 2 and 3, during which the membrane was held at 80 mV and the activation gate
was closed. This suggests that once bound, 4AP was "trapped" inside
the rKv1.43-25 channel by the closure of the activation gate.
|
80 mV. This prolonged wash period should have
removed 4AP from the bath and from the cytoplasm. After washing out
4AP, the peak current amplitude during the first pulse was increased
only modestly. Therefore, recovery from 4AP blockade was very
incomplete at 80 mV even after a 10-min wash with 4AP-free solution
(i.e., 4AP trapped inside the channel). However, the current trace
during the first pulse after 4AP removal showed a slowed rising phase
and, more importantly, a markedly slowed decay phase relative to the
control trace. This reflected a time-dependent 4AP unbinding from the
channel and relief of blockade during membrane depolarization. The
current traces during pulses 2 and 3 were superimposable with each
other, indicating that the recovery from 4AP blockade approached a
steady state at the end of the first pulse.
The effects of 4AP on rKv4.2 were studied using the same protocols as
described above. Fig. 2A (right) shows that after a 10-min exposure to
4AP (5 mM), while the membrane was held at 80 mV (40 mV negative to
the activation threshold) (Tseng et al., 1996), the first pulse induced
a current with a markedly suppressed peak amplitude (by 64% relative
to control) and a slowed rising phase. Furthermore, this current trace
decayed at a much slower rate than that of control, such that the two
traces crossed over. The current traces induced by pulses 2 and 3 were
superimposable with the first current trace. These changes in current
amplitude and kinetics indicate that 4AP blockade of rKv4.2 had already reached a steady state at 80 mV before the first pulse. 4AP unbound from rKv4.2 upon membrane depolarization, leading to a slower rising
phase than the control. Because channel inactivation and 4AP binding
were mutually exclusive (Tseng et al., 1996), rKv4.2 channels could
inactivate only after 4AP dissociation. This resulted in a much slower
decay (inactivation) phase in the presence of 4AP. Therefore, 4AP
binding to rKv4.2 occurred when the activation gate was closed and
unbound when the gate opened (i.e., rested-state binding preference).
As shown in Fig. 2B (right), washing out 4AP for 10 min while holding
the membrane at 80 mV allowed the recovery of rKv4.2 from 4AP
blockade to approach a steady state, so that the subsequent current
traces induced by three pulses applied 1 min apart were almost or
totally superimposable. Therefore, 4AP could escape from the rKv4.2
channel at 80 mV when the activation gate was closed (although the
rate of 4AP unbinding from the rested state was much slower than that
from the activated state) (Tseng et al., 1996).
4AP Blocking Potency.
Figure 2 also shows that the 4AP
concentration needed to achieve a similar degree of peak current
suppression was about 10 times higher for rKv4.2 than for
rKv1.43-25. This point was examined further in the experiments
shown in Fig. 3. The voltage-clamp protocols used to quantify 4AP's IC50 values
(described in Fig. 3 legend) were designed to reduce the interference
arising from differences in the state dependence of 4AP binding and
unbinding and in the rate of channel inactivation (see
Discussion). The IC50 value for 4AP
blockade of rKv1.43-25 was 0.16 ± 0.04 mM. This was much
lower than that of rKv4.2 (1.05 ± 0.18 mM). Of note, there was a
substantial fraction of rKv4.2 current not suppressed even by high
concentrations of 4AP. This reflects a technical limitation: rKv4.2
current measurement required channel activation that simultaneously
induced 4AP unblock.
|
3-25 and rKv4.2 differ in two
aspects of 4AP actions: the state dependence of 4AP binding and
unbinding, and the 4AP sensitivity. To deduce the structural basis for
these differences, mutagenesis experiments were performed to alter the
putative activation gate and the 4AP-binding site in rKv1.43-25.
The design of mutations was based on current knowledge of the
structure-function relationship of voltage-gated K channels (see
Introduction). The rKv1.43-25 sequence in the S4-S5,
N-S5, and C-S6 regions was replaced by the corresponding sequence of rKv4.2. The mutants are designated by the locations where mutations were made (Fig. 1B).
Effects of rKv1.43-25 Mutations on Channel Gating
Figure 4 compares the voltage
dependence and kinetics of channel activation gating among the WT and
mutants of rKv1.43-25. Replacing the S4-S5 sequence of
rKv1.43-25 with that of rKv4.2 (i.e., changing 6 of 11 residues in
this region; Fig. 1A) had little or no effect on the channel's
activation-gating properties. The activation curve of the S4-S5 mutant
was superimposable with that of the WT channel (Fig. 4B). The rate of
channel deactivation was also similar between the two over a wide
voltage range (Fig. 4, C and D). These data are consistent with the
notion that the S4-S5 loop is not part of the activation gate
[although the S4-S5 loop participates in conformational changes
associated with channel activation (Holmgren et al., 1996) and changes
in a channel's gating function induced by mutations in this region
have been reported (McCormack et al., 1991)].
|
On the other hand, replacing the N-S5 or C-S6 sequence of
rKv1.43-25 with that of rKv4.2 had profound effects on the
channel's activation gate function. Both mutations shifted the voltage
dependence of channel activation in the hyperpolarizing direction,
along with an increase in the steepness of the activation curve (Fig. 4B, legend). These were accompanied by a marked slowing of deactivation over a wide voltage range (Fig. 4, C and D), although there was little
or no change in the rate of channel activation (Fig. 4A). These
observations are consistent with the notion that both the N-S5 and C-S6
domains participate in the formation of the activation gate (Shieh et
al., 1997; Holmgren et al., 1998; Zhou et al., 1998), so that replacing
five residues in the N-S5 region or even one residue in the C-S6 region
(Fig. 1A) could markedly affect the channel's activation gate
function. However, neither N-S5 nor C-S6 recapitulated the voltage
dependence or kinetics of rKv4.2 activation (Fig. 4, B and D, solid
circles). This indicates that both domains, and likely other regions of
the channels, are involved in determining the activation gate function.
Effects of rKv1.43-25 Mutations on 4AP Actions
State Dependence of 4AP Binding.
As shown in Fig. 2A, changes
in the current amplitude and kinetics during the first pulse after a
prolonged 4AP exposure can tell us a lot about the state dependence of
4AP binding. In particular, the hallmark of a rested-state 4AP-binding
preference includes: 1) a marked suppression of current amplitude in
the beginning of the first pulse after 4AP application (because 4AP
blockade occurs before channel activation) and 2) signs of 4AP unblock during depolarization (e.g., slowed rising phase, followed by a slowed
decaying phase and a crossover with the control current trace if 4AP
binding and channel inactivation are mutually exclusive). Therefore,
the same protocol as described for Fig. 2A was used to study the state
dependence of 4AP binding to the WT and mutants of rKv1.43-25.
80 mV, the peak current during the first pulse was reduced
modestly. The rate of current decay during this pulse was markedly
accelerated. The current traces during pulses 2 and 3 were
superimposable with each other, and the degree of peak current
suppression was similar to that seen at the end of the first pulse.
|
80 mV]. This is consistent
with a very slow rate of 4AP blockade of N-S5 (at 10 mM and 80 mV,
time constant was 62.3 ± 2.2 s, n = 5, Fig.
6). The phenotype of C-S6 was
similar to that of N-S5, indicating a similar state dependence of 4AP binding. However, the current traces recorded after 4AP application did
not cross over the control trace during the 5-s depolarization pulses.
Furthermore, the C-S6 current traces during pulses 2 and 3 were
superimposable with that during pulse 1, indicating that 4AP blockade
of C-S6 was largely reestablished during the interpulse interval. This
is consistent with the observation that 4AP blockade of C-S6 (at 80
mV half-time was 12.5 ± 2.95, n = 4, data not shown) was faster than that of N-S5.
|
Effects of Changing Vh on 4AP Blockade.
The data
presented in Fig. 5 suggest that the N-S5 and C-S6 mutations of
rKv1.43-25 transferred the state dependence of 4AP binding from the
donor (rKv4.2) to the host. However, before accepting this proposal, we
needed to test the effects of changing Vh on 4AP
binding. This was important because one could create an apparent rested-state 4AP-binding phenotype in rKv1.43-25 simply by shifting the Vh from 80 mV to a more depolarized voltage
above the activation threshold. This is illustrated in Fig.
7A. In this experiment, the oocyte was
exposed to 4AP twice. The first exposure was applied at a
Vh of 80 mV, and the resulting changes in
current amplitude and kinetics were similar to those shown in Fig. 2A.
Then, 4AP was washed out and the current recovered from suppression.
The Vh was switched to 45 mV (5 mV above the
activation threshold; Fig. 4B) during the second 4AP exposure. At this
Vh, 4AP could bind before the first
depolarization pulse. Indeed, early during the first pulse after 4AP
application, the current was suppressed severely (by 85%).
Furthermore, current traces during pulses 2 and 3 were superimposable
with that during pulse 1, indicating that 4AP blockade had reached a
steady state before pulse 1. At Vh = 45 mV,
current traces recorded in the presence of 4AP showed a gradual
increase during depolarization. This is because strong membrane
depolarization could reduce 4AP-binding affinity to rKv1.43-25 (Yao
and Tseng, 1994; Zhang et al., 1998). Therefore, in rKv1.43-25, shifting Vh from 80 mV to slightly above the
activation threshold changed the apparent phenotype of 4AP binding to
that of rKv4.2. This sensitivity to Vh was in
sharp contrast to the situation of rKv4.2. One example is illustrated
in Fig. 7B. Indeed, we have shown that an important test of 4AP binding
to the rested state of a channel is to show that the state dependence
and kinetics of 4AP binding are not affected by shifting
Vh (Tseng et al., 1996).
|
80, 100, and 120
mV. If the rested state 4AP-binding preference suggested by Fig. 5 was
actually due to 4AP binding to channels that activated sporadically at
80 mV, one expects to see two things. First, shifting the
Vh from 80 to 100 or 120 mV should markedly
reduce the rate of 4AP reblock by decreasing the probability of channel
opening. Second, a sufficiently negative Vh
should unmask an activated-state 4AP-binding behavior by preventing 4AP binding before membrane depolarization.
As shown in Fig. 6, the current traces during V1
(recorded after the membrane had been held at Vh
for 3 min to ensure that 4AP reblock had reached a steady state) were
similar at Vh values 80, 100, and 120 mV,
i.e., they all showed a severely suppressed initial current followed by
a gradual increase (4AP unblock). Therefore, a hyperpolarizing shift of
Vh by 40 mV did not unmask any activated-state
4AP-binding behavior. Furthermore, when an identical depolarization
pulse (V2) was applied shortly after V1 (e.g., 10 s), the current had a much
higher initial amplitude than during V1.
Therefore, such a short
V1-V2 interval was not long
enough to allow a sizable 4AP reblock. As the
V1-V2 interval was
prolonged, the initial current amplitude during
V2 gradually declined and was superimposable with
that during V1 at a
V1-V2 interval of 180 s. The time courses of 4AP reblock of the N-S5 mutant channel from the
same experiment are plotted in Fig. 6B. The time courses at the three
Vh values were superimposable. After an initial
lag, the rising phases could be reasonably described by a single
exponential time course. The 
values of 4AP reblock from this and
four other experiments are plotted in Fig. 6C. There was no difference
in the values determined at the three Vh
values (p > .05, one-way ANOVA).
The same type of analysis was carried out for the C-S6 mutant (data not
shown). Similar to the N-S5 data, shifting the Vh from 80 to 100 or 120 mV did not unmask any activated state 4AP-binding behavior. For C-S6, the time courses of 4AP reblock manifested a significant initial lag. The rising phase that followed required a double-exponential function for a reasonable fit. Therefore, the T1/2 of 4AP reblock (i.e., the
V1-V2 interval when 50% of the channels unblocked during V1 were reblocked
by 4AP) was used for comparison of data obtained at different
Vh values. In the presence of 5 mM 4AP, the
T1/2 values determined at
Vh of 80, 100, and 120 mV were 12.5 ± 2.9, 13.0 ± 3.4, and 16.8 ± 2.4 s (n = 4 each). There was no statistically significant difference in these
T1/2 values (p = 0.131 by one-way ANOVA).
Although mutating the N-S5 or C-S6 sequence of rKv1.43-25 to that
of rKv4.2 transferred the state dependence of 4AP binding, the kinetics
of 4AP binding and unbinding in these two mutants were not the same as
those of the donor channel. The rates of 4AP binding to the N-S5 and
C-S6 mutants were much slower than that of 4AP binding to rKv4.2 [at
80 mV and 5 mM 4AP, of 4AP reblock of rKv4.2 was 7.7 ± 1.2 s, (Tseng et al., 1996)]. The rates of 4AP unbinding from the
N-S5 and C-S6 mutant channels were also slower than the rate of 4AP
unbinding from rKv4.2. The values of 4AP unbinding from N-S5 and
C-S6 were estimated by fitting a single-exponential function to the
rising phase of current traces induced by the first pulse after 4AP
application (e.g., those shown in Fig. 5). At +60 mV, of 4AP
unbinding from N-S5 was 415.3 ± 35.9 ms (n = 5, 10 mM 4AP), and that of 4AP unbinding from C-S6 was 1480 ± 130 ms
(n = 5, 5 mM 4AP). These values are much longer
than the of 4AP unbinding from rKv4.2 (42 ± 6 ms at +60 mV
and 5 mM 4AP; Tseng et al., 1996).
4AP Blocking Potency.
Figure 8
shows the IC50 values for 4AP blockade of the WT
and mutant channels of rKv1.43-25 determined using the
voltage-clamp protocol described for Fig. 3. The
IC50 value for the S4-S5 mutant appeared lower
than that of the WT channel (0.05 ± 0.003 versus 0.16 ± 0.04 mM). However, the difference was not statistically significant
(p = .121). For the N-S5 mutant, the
IC50 value (0.68 ± 0.15 mM) was
significantly higher than that of the WT channel (p = .04) but was similar to the IC50 value for rKv4.2
(1.05 ± 0.18 mM; p > .05). Therefore, mutating
the N-S5 sequence transferred the phenotype of 4AP sensitivity from the
donor to the host channel. This supports the proposal that the N-S5
domain is important for the formation of the 4AP-binding site in
voltage-gated K channels (Shieh and Kirsch, 1994). For the C-S6
mutant, the IC50 value of 4AP blockade (0.20 ± 0.04 mM) was similar to that of the WT channel. However, it has been
shown that mutating the same residue (threonine at position 529; Fig.
1A) to leucine or isoleucine could dramatically reduce the 4AP's
blocking potency, whereas replacing this threonine with phenylalanine
had the opposite effect (Zhang et al., 1998). Therefore, the effects of
a single point mutation in the C-S6 domain on 4AP binding depend on the
side chain properties here.
|
| |
Discussion |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
Results from the present study can be summarized as the following:
1) rKv1.43-25 and rKv4.2 were different in the state dependence of
4AP binding and in their sensitivity to 4AP, 2) replacing the amino
acid sequence of the intracellular loop between S4 and S5 (S4-S5) of
rKv1.43-25 with that of rKv4.2 did not affect the channel's
activation gating properties or the state dependence of 4AP binding, 3)
replacing the amino acid sequence in the cytoplasmic half of S5 (N-S5)
or S6 (C-S6) of rKv1.43-25 with the corresponding sequence of
rKv4.2 had profound effects on channel's activation gating properties
(importantly, both mutations transferred the phenotype of rested state
4AP-binding preference from the donor to the host channel), and 4) the
N-S5 mutation also transferred the donor channel's 4AP sensitivity.
Technical Considerations.
In this study we used a deletion
mutant of rKv1.4 in which the fast inactivation process was totally
disrupted (Tseng-Crank et al., 1993). This facilitated the studies of
4AP binding and the channel's activation gate function. Ideally, we
would like to do the same on rKv4.2. Unfortunately, it has not been
possible to identify a domain or domains in rKv4.2 that can abolish the channel's fast inactivation process (Baldwin et al., 1991; Pak et al.,
1991; Jerng and Covarrubias, 1997). The important issue here is: can
the difference in the inactivation rate between rKv1.43-25 and
rKv4.2 account for, or contribute to, the observed difference in the
state dependence of 4AP binding? This is not the case for the following
reasons. First, in the WT rKv1.4 channel, which inactivates as rapidly
as rKv4.2, 4AP binding clearly occurs in the activated state (Yao and
Tseng, 1994). Second, in the N-S5 and C-S6 mutants that inactivate very
slowly, 4AP binds in the rested state.
80
mV, and short depolarization pulses (40 ms for the WT and mutants of
rKv1.43-25 and 15 ms for rKv4.2) to a low voltage (20 mV) were
applied once every 30 s to monitor 4AP's effects on the current
amplitude. For rKv4.2, N-S5, and C-S6, 4AP bound to the channels at
Vh and unbound when channels opened. Because 4AP
unbinding was a time- and voltage-dependent process (Tseng et al.,
1996), the effect of 4AP unbinding on the measurement of its
IC50 values was minimized by using a short-pulse duration and a low-depolarization voltage. For WT rKv1.43-25 and
S4-S5, 4AP binding accumulated gradually during repetitive pulses to
20 mV, with blocker trapped within the channel during the interpulse
interval at 80 mV. The interference from the fast inactivation
process of rKv4.2 [that would hinder 4AP binding (Tseng et al.,
1996)] also was minimized by the short duration and low voltage of the
depolarization pulses (15 ms to 20 mV was too short to induce
significant inactivation; Fig. 3B).
We used a chimera approach to study the structural basis
underlying the differences in 4AP actions between rKv1.43-25 and rKv4.2. The assumption underlying this approach was that a channel function could be determined by a specific domain(s) so that
transplanting individual domain(s) from the donor to the host can
transfer the phenotype of the donor channel. This strategy works well
in certain cases [most notably in identifying the channel domain that
lines the ion-selectivity filter (Hartmann et al., 1991)]. However, if
a channel function requires the simultaneous presence of multiple domains or is affected by global conformational changes of the channel,
transplanting specific domain(s) may alter the channel's function but
the phenotype may be very different from that of the donor channel.
This is probably why in the N-S5 and C-S6 mutants of rKv1.43-25,
the phenotypes of activation gating function were different from that
of the donor channel, rKv4.2 (Fig. 4, B and D). The data support an
important role of the N-S5 and C-S6 domains in the activation gate
function, but suggest further that other domains of the channel are
also involved (Kirsch et al., 1993; Shieh et al., 1997; Holmgren et
al., 1998).
Distinguishing between Two Possible Mechanisms for the Different
State Dependencies of 4AP Binding.
We propose two possibilities to
explain the difference in the state dependence of 4AP binding between
rKv1.43-25 and rKv4.2. The first possibility is that in
rKv1.43-25 the 4AP binding site is behind, and protected by, the
activation gate, whereas in rKv4.2 the 4AP binding site is in front of
the activation gate. The second possibility is that the structure of
the activation gate determines the state dependence of 4AP binding,
although the relationship between the 4AP binding site and the gate may
be the same in these two channels. Data from the C-S6 mutant argue
against the first possibility. In this mutant, only one residue in the
S6 domain was altered (from threonine to valine; Fig. 1A). This single
point mutation changed the state dependence of 4AP binding (Fig. 5) without altering the IC50 value (Fig. 8). It is
less likely that such a single mutation could shift the 4AP-binding
site from behind the activation gate to in front of the gate or create
an "rKv4.2-like"-binding site in front of the gate. Instead, in
view of the marked alterations in the activation gating process
associated with this point mutation, it is more likely that C-S6
changed the state dependence of 4AP binding by altering the activation
gate function. It is possible that the 4AP-binding site and the
activation gate coincide with each other in K channels, and subtle
changes in the amino acid sequences here can profoundly alter the
activation gate function as well as the phenotype of 4AP binding. Our
data support an important role of the cytoplasmic halves of both S5 and
S6 (N-S5 and C-S6) in forming these structures.
| |
Acknowledgments |
|---|
I thank Bing Zhu for making the mutations and assisting in the oocyte experiments.
| |
Footnotes |
|---|
Accepted for publication March 30, 1999.
Received for publication February 5, 1999.
1 This study was supported National Heart, Lung, and Blood Institute Grant HL 46451, National Institutes of Health, Bethesda, MD.
Send reprint requests to: Dr. Gea-Ny Tseng, Ph.D., Department of Pharmacology, Columbia University, 630 West 168th St., New York, NY 10032. E-mail: gt10{at}columbia.edu
| |
Abbreviations |
|---|
4AP, 4-aminopyridine; Vh, holding voltage; Vt, test voltage; Vr, repolarization voltage; S4, the fourth transmembrane segment of a subunit or a homologous domain of voltage-gated ion channels; S5, the fifth transmembrane segment; S6, the sixth transmembrane segment; WT, wild-type.
| |
References |
|---|
|
Top
Abstract
Introduction
Materials and Methods
Results
Discussion
References
|
|---|
This article has been cited by other articles:
|
|
 |
|
  L. K. Bekar, M. E. Loewen, K. Cao, X. Sun, J. Leis, R. Wang, G. W. Forsyth, and W. Walz Complex Expression and Localization of Inactivating Kv Channels in Cultured Hippocampal Astrocytes J Neurophysiol, March 1, 2005; 93(3): 1699 - 1709. [Abstract] [Full Text] [PDF]
|
|
|
|
![<UP>Fraction of control</UP>=<UP>A/</UP>(<UP>1</UP>+[4<UP>AP</UP>]<UP>/IC</UP><SUB><UP>50</UP></SUB>)+(1−<UP>A</UP>)](/content/vol290/issue2/fulltext/569/img001.gif)
![<UP>Fraction activated</UP>=1/(1+<UP>exp</UP>[(<UP>V<SUB>0.5</SUB></UP>−<UP>V</UP><SUB><UP>t</UP></SUB>)<UP>/k</UP>)])](/content/vol290/issue2/fulltext/569/img002.gif)
![<UP>Fraction activated</UP>=<UP>A<SUB>1</SUB>/</UP>(<UP>1</UP>+<UP>exp</UP>[(<UP>V<SUB>1</SUB></UP>−<UP>V</UP><SUB><UP>t</UP></SUB>)<UP>/k</UP><SUB><UP>1</UP></SUB>])+(1−<UP>A</UP><SUB><UP>1</UP></SUB>)<UP>/</UP>(<UP>1</UP>+<UP>exp</UP>[(<UP>V<SUB>2</SUB></UP>−<UP>V</UP><SUB><UP>t</UP></SUB>)<UP>/k</UP><SUB><UP>2</UP></SUB>])](/content/vol290/issue2/fulltext/569/img003.gif)
