State-Dependent Nickel Block of a High-Voltage–Activated Neuronal
Calcium Channel
MATTHEW B. MC FARLANE 1,2 AND WILLIAM F. GILLY 2
1
Department of Molecular and Cellular Physiology, Stanford University School of Medicine, Stanford 94305; and
2
Department of Biological Sciences, Hopkins Marine Station, Stanford University, Pacific Grove, California 93950
INTRODUCTION
Many divalent cations, such as Mg 2/ and the transition
metals Cd 2/ , Co 2/ , and Mn 2/ block voltage-dependent
Ca 2/ channels (Byerly et al. 1985; Hagiwara and Byerly
1981; Hille 1992). In most cases, divalent cations block by
binding inside the channel pore while the channel is open
and impeding the flux of Ca 2/ ions through the channel
(Hess and Tsien 1984). This results in a rapid ‘‘flickery’’
block at the single channel level (Chesnoy-Marchais 1985;
Lansman et al. 1986; Winegar et al. 1991). Most divalent
ions can also permeate open Ca 2/ channels and reduce the
current flowing through the channel by conducting at a much
slower rate than do Ca 2/ ions (Chow 1991; Hille 1992).
Block by Ni 2/ , however, exhibits very different characteristics. For example, several studies have shown that of all
divalents tested only nickel ions failed to permeate through
open Ca 2/ channels (Jones and Sharpe 1994; Shibuya and
Douglas 1992). Furthermore, nickel block of single Ca 2/
channels exhibits a unique blocking profile characterized
both by flickery events as described above and by much
longer duration blocking events (Chesnoy-Marchais 1985;
Winegar et al. 1991).
Whole cell voltage clamp was used to examine the properties of nickel block of Ca 2/ channels in squid giant fiber
lobe (GFL) neurons. These channels are blocked by v-agatoxin-IVA and are thought to be of a class similar to the
mammalian P/Q-type Ca 2/ channel (McFarlane and Gilly
1678
1996). Nickel ions were found to block in a state-dependent
manner because Ni 2/ interacted with GFL Ca 2/ channels in
three different phases of activation. First, nickel ions impeded channel activation by stabilizing one or more closed
channel states. Second, nickel significantly reduced the maximal whole cell Ca 2/ conductance at steady-state, consistent
with block of open channels. Third, nickel block became
more pronounced during longer pulses to strongly depolarizing potentials, the duration and strength of which were sufficient to cause a majority of channels to enter a second open
state (McFarlane 1997). Nickel ions thus appear to block
this second open state with higher affinity.
The overall characteristics of nickel block suggest that,
in addition to its role as an open channel blocker, nickel
ions can interact with GFL Ca 2/ channels at a site (or sites)
not directly associated with the conduction pore. The close
correlation between increased nickel block during long
pulses and the transition to a second open state could indicate
that changes in channel gating are accompanied by conformational changes in the extrapore regions of the channel
protein. This idea is reinforced by the observation that vagatoxin IVA, a highly charged polypeptide unlikely to venture deeply into the pore, appears to differentiate between
the two open states in a similar fashion (McFarlane 1997).
METHODS
California market squid (Loligo opalescens) GFL neurons were
isolated and cultured (1–5 days) as previously described (McFarlane and Gilly 1996).
Ca 2/ currents (ICa ) of GFL somata without processes were isolated with an internal solution that contained 451 mM tetramethylammonium (TMA) aspartate, 25 mM TMA-fluoride, 25 mM tetraethylammonium (TEA) chloride, 20 mM ethylene glycol-bis( baminoethylether)-N,N,N *,N *-tetraacetic acid (EGTA), 20 mM
N-2-hydroxyethylpiperazine-N *-2-ethanesulfonic acid (HEPES),
and 4 mM MgATP. The external solution consisted of 480 mM
TMA-chloride, 60 mM CaCl2 , 10 mM TEA-chloride, 10 mM
HEPES, and 500 nM tetrodotoxin (Sigma, St. Louis, MO). Where
indicated, external Ca 2/ concentration (Cao ) was lowered to 15
mM without any other adjustment (e.g., addition of MgCl2 ). Ultrapure NiCl2 (Aldrich, St. Louis, MO, purity ú99.9999%) was prepared as a 1-M stock solution in deionized water and added directly
to the external solution before each experiment. All solutions were
adjusted to 990–1,010 mosM and pH 7.6–7.7. Experiments were
performed at 15–177C.
Whole cell currents were acquired with a patch-clamp amplifier
with a 20-MV feedback resistance in the headstage and electronic
series resistance compensation. Signals were low-pass filtered at
10 kHz with an eight-pole Bessel filter. All tail current records
0022-3077/98 $5.00 Copyright q 1998 The American Physiological Society
/ 9k2d$$oc05 J975-7
09-15-98 19:57:20
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 15, 2017
McFarlane, Matthew B. and William F. Gilly. State-dependent
nickel block of a high-voltage–activated neuronal calcium channel.
J. Neurophysiol. 80: 1678–1685, 1998. Effects of nickel ions
(Ni 2/ ) on noninactivating calcium channels in squid giant fiber
lobe (GFL) neurons were investigated with whole cell voltage
clamp. Three different effects of Ni 2/ were observed to be associated with distinct Ca 2/ channel activation states. 1) Nickel ions
appear to stabilize closed channel states and, as a result, slow
activation kinetics. 2) Nickel ions block open channels with little
voltage dependence over a wide range of potentials. 3) Block of
open channels by Ni 2/ becomes more effective during an extended
strong depolarization, and this effect is voltage dependent. Recovery from this additional inhibition occurs at intermediate voltages,
consistent with the presence of two distinct types of Ni 2/ block
that we propose correspond to two previously identified open states
of the calcium channel. These results, taken together with earlier
evidence of state-dependent block by v-agatoxin IVA, suggest that
Ni 2/ generates these unique effects in part by interacting differently
with the external surface of the GFL calcium channel complex in
ways that depend on channel activation state.
STATE-DEPENDENT CA 2/ CHANNEL BLOCK BY NICKEL IONS
were sampled at 20 ms/point. Sampling for other pulse patterns
varied from 50 to 500 ms/point as required (collection limit of 750
points/pulse). Linear ionic and capacity currents were subtracted
by a P/ 04 method from the holding potential of 080 mV. All curve
fitting utilized a nonlinear least-squares minimization algorithm
(Origin, Microcal Software, Northampton, MA).
1679
Data are shown expressed as the means { SE.
RESULTS
Voltage-independent block of open channels by nickel
FIG . 1. Nickel block of giant fiber lobe (GFL) neuronal ICa . A: macroscopic ICa was recorded in response to a 25-ms pulse to 0 mV in the absence
( ● ) and presence of 6 mM NiCl2 ( s ), and on return to control solution
(wash). [cell 24e] B: macroscopic Ca 2/ conductance (gCa ) was calculated
from tail current amplitude at 080 mV after 10-ms pulses to various potentials by using the equation gCa Å DI/ DV , where DI is the current difference
resulting from a DV potential change as indicated in A. gCa values were
obtained for 25-ms pulses in 60 mM Cao for control ( j ) and 6 mM Ni 2/
solutions ( h ), and are shown plotted as a function of activation pulse
voltage. Control values are shown with the best fit of a Boltzmann function,
gmax /1 / exp{(V 0 V1 / 2 )/k} (fit parameters: V1 / 2 Å 04.4 mV, k Å 7.6
mV, gmax Å 177.7 nS;
). This fit was multiplied by 0.47 to fit gmax
values obtained in Ni 2/ ( – – – ). C: gCa similarly obtained in 15 mM Cao
in control ( s ) and 1.5 mM Ni 2/ solutions ( ● ). Control values are shown
with the best fit of a Boltzmann function (V1 / 2 Å 015.8 mV, k Å 8.6 mV,
gmax Å 126.9 nS;
), and this fit was multiplied by 0.520 to fit the
values obtained in Ni 2/ ( – – – ). [cell 24e]
/ 9k2d$$oc05 J975-7
09-15-98 19:57:20
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 15, 2017
Macroscopic Ca 2/ currents (ICa ) of GFL neurons were
recorded during 25-ms pulses to 0 mV from a holding potential of 080 mV in the absence and presence of 6 mM NiCl2
and after return to control solution (Fig. 1A). Whole cell
conductance (gCa )-voltage relationships were calculated
from ICa records by using the relationship DI/ DV, which
measures the amplitude difference between tail currents recorded at 080 mV and at the end of a 10-ms voltage pulse
to a family of potentials (Fig. 1B). Control gCa values (Fig.
1B, j ) are shown fit with a single Boltzmann function, and
this fit was multiplied by a scaling factor (0.47) to compare
these data with corresponding values similarly obtained in
the presence of 6 mM NiCl2 (Fig. 1B, h ). Thus the reduction of gCa by Ni 2/ is not voltage dependent. Lowering Cao
from 60 to 15 mM produces a negative shift of Ç10 mV in
the gCa-voltage relationship (Fig. 1C, ●, vs. Fig. 1B, j ).
Nickel ions act more potently in lower external Ca 2/ , and
1.5 mM NiCl2 (Fig 1C, s ) reduces gCa by a constant factor
of 0.52 at all voltages.
Lack of voltage dependence for block of gCa by Ni 2/
during brief pulses was also examined through analysis of
tail currents. Tail currents were measured at several voltages after maximal channel activation by a 10-ms depolarizing pulse to /60 mV ( Fig. 2 A ) . At a particular voltage,
the peak amplitude of tail currents is directly proportional
to the number of open channels at the end of the activating
pulse, and the decay reflects the kinetics of channel deactivation. Figure 2 A displays these tail currents in the absence
( left ) and presence of 3 mM ( middle ) and 15 mM NiCl2
( right ) . NiCl2 block causes the overall reduction of macroscopic ICa .
The instantaneous current-voltage relationship ( ICa – V )
for open Ca 2/ channels is generated by plotting peak tail
current amplitude as a function of repolarization voltage
( Fig. 2 B ) , and the slope of this relationship can be used
to estimate the whole cell Ca2/ conductance at steady state.
Figure 2 B shows linear fits and corresponding conductance
values for control ( Go ) and nickel-treated currents ( G )
generated in Fig. 2 A . Nonlinearity is only apparent at very
negative voltages, consistent with the idea that the block
1680
M. B. MC FARLANE AND W. F. GILLY
of open Ca2/ channels by Ni 2/ is essentially voltage independent.
Steady-state conductance values derived from instantaneous ICa – V curves in the absence and presence of Ni 2/
were used to determine the dose dependence of nickel
block ( Fig. 2C ) . Mean G / Go values are shown plotted as
a function of nickel concentration for experiments performed in 60 mM ( h ) and 15 mM ( s ) external Ca 2/ .
Both data sets are shown fit with the Hill equation ( see
Fig. 2 legend ) . Cao reduction led to an apparent increase
in affinity ( IC50 decreases from 6.6 to 1.4 mM ) . The best
fits for 60 and 15 mM Cao yield Hill coefficients of 1.6
and 1.5, respectively.
Nickel slows activation kinetics
In addition to blocking open Ca 2/ channels as described
above, nickel ions produce slower Ca 2/ channel activation
kinetics during an activating voltage pulse. Scaling control
and nickel-treated current traces from Fig. 1A to the same
peak value reveals a significant slowing of activation kinetics
at this voltage (Fig. 3A). A time constant characterizing
activation ( tact ) was established by fitting a single exponential function to the final approach ( Ç30%) to peak ICa (Fig.
). tact is thus analogous to the Hodgkin-Huxley
3A,
(1952) gating parameter tm for Na / conductance and pro-
/ 9k2d$$oc05 J975-7
vides a convenient means of describing the voltage dependence of activation gating (Hille 1992; Sala 1991). Nickel
application led to an increase in tact (2.10–4.86 ms). Slower
activation kinetics were observed in all neurons tested and
were present over a wide nickel concentration range (0.3–
10 mM), with higher concentrations producing larger effects
(data not illustrated).
tact measurements were made from eight GFL neurons
over a wide range of activating voltages, and mean control
( ● ) and 6-mM NiCl2-treated ( s ) values are plotted as a
function of activation voltage in Fig. 3B. For potentials more
negative than /20 mV, tact values were significantly larger
in the presence of nickel than in controls, and on average
tact increased by 2.62 { 1.26 ms at 020 mV compared with
1.66 { 0.64 ms at 0 mV and 0.25 { 0.20 ms at / 20
mV. The population response can be loosely described as a
positive shift in activation kinetics of Ç14 mV, but this
shift does not appear to be a perfect translation along the
voltage axis.
Measurement of nickel effects on deactivation kinetics
is slightly more complicated because GFL ICa deactivation
kinetics follows a biexponential time course (Chow 1991;
McFarlane 1997), with an 8- to 10-fold difference in time
constants between fast and slow components. The amplitude
of both fast and slow deactivation components was reduced
in the presence of nickel (Fig. 3C).
09-15-98 19:57:20
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 15, 2017
FIG . 2. Voltage-independent nickel block
of open Ca 2/ channels A: GFL Ca 2/ tail
currents were recorded in 60 mM Cao at
various voltages [indicated on control
(left) traces] after a 10-ms step depolarization to /60 mV. This procedure was repeated in the presence of 3 mM (middle)
and 15 mM NiCl2 (right). [cell 24a] B:
values for peak tail current amplitude were
obtained for control ( h ), 3 mM NiCl2 ( s ),
and 15 mM NiCl2 ( n ) and are shown plotted as a function of voltage. These instantaneous ICa-V relationships were then fitted
by a straight line (range 080 to /40 mV),
and the slope of this linear fit is an estimate
of the whole cell Ca 2/ conductance. Values
for conductance for the absence (Go ) and
presence (G) of nickel are indicated next to
the appropriate ICa-V relationship. C: dose
dependence of nickel block was determined
in 60 ( h ) and 15 mM Cao ( s ) by dividing
G values (determined by method in B for
a given nickel concentration) by Go values
in an individual cell. Mean values are
shown fit with the Hill equation [Ni] n /
{[Ni] n / (IC50 ) n } (fit parameters: 60 mM
Cao : IC50 Å 6.6 mM, n Å 1.6; 15 mM Cao :
IC50 Å 1.4 mM, n Å 1.5). Small numbers
near symbols indicate the number of observations.
STATE-DEPENDENT CA 2/ CHANNEL BLOCK BY NICKEL IONS
1681
For the neuron in Fig. 3C, both fast and slow deactivation
was slightly faster in the presence of nickel when measured
at 080 mV (see Fig. 3 legend). Unlike the case for activation
kinetics, however, no clear pattern of kinetic changes
emerged from analysis of several cells. Mean tdeact values
are plotted as a function of voltage in Fig. 3D for the same
experiments as analyzed in Fig. 3B. It is therefore apparent
that nickel affects deactivation kinetics much less than activation kinetics.
Enhancement of nickel block during extended strong
depolarization
During long pulses to strongly depolarizing potentials, we
identified a third effect of nickel that is evident as an apparent
FIG . 4. Nickel blocking affinity increases with duration of strong depolarization. Tail currents were recorded at 080
mV after a voltage step to /60 mV for the
indicated duration ( Dt) in the absence and
presence of 3 mM NiCl2 (arrowhead). Dotted lines indicate the level of Ni 2/ block
for Dt Å 10 ms. [cell 15a]
/ 9k2d$$oc05 J975-7
09-15-98 19:57:20
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 15, 2017
2/
FIG . 3. Nickel slows Ca
channel activation kinetics. A:
traces from Fig. 1A were scaled to the same peak value and
are shown on an expanded timescale. Values for the time constant of activation ( tact ) were obtained by fitting a single exponential to the final approach ( Ç30%) to peak ICa (——, t Å
2.10 ms and 4.86 ms for control and 6 mM NiCl2 , respectively). B: mean (n Å 8) tact values in the absence ( ● ) and
presence ( s ) of 6 mM NiCl2 are shown plotted as a function
of activation voltage. C: traces from Fig. 1A are shown on an
expanded timescale fit with double exponential functions. The
fast and slow components of channel deactivation are indicated,
and the time constant for deactivation ( tdeact ) is determined
from such fits (control vs. nickel tdeact values: fast, 221 vs. 190
ms; slow, 1,453 vs. 1,221 ms). D: mean (n Å 8) tdeact values
were obtained for a variety of voltages after a 10-ms pulse to
/60 mV (see Fig. 1). Values for the fast ( h, j ) and slow
( n, m ) components of deactivation in the absence ( j, m ) and
presence ( h, n ) of 6 mM NiCl2 . Statistical significance between control and nickel populations for each voltage ( B and
D) was determined by one-way analysis of variance (* P õ
0.001).
1682
M. B. MC FARLANE AND W. F. GILLY
2/
FIG . 5. Properties of nickel extra-block. A: additional Ni
block for
increasing duration pulses to /60 mV was quantified by an extra-block
index, EB. EB values were obtained by dividing total tail current amplitude
in the presence of Ni 2/ for a 10-ms depolarization by the amplitude recorded
for pulses of Dt duration. Mean values for this measurement (6 mM NiCl2 ,
n Å 9) are shown plotted as a function of Dt and fit with a single exponential
function ( t Å 140.4 ms) B: tail current amplitudes (Itail ) were recorded at
various voltages immediately after depolarization to /60 mV for Dt Å 5
ms ( s, ● ) and 500 ms ( h, j ). This procedure was performed in the absence
( ●, j ) and presence of 6 mM NiCl2 ( s, h ). [cell 24e] C: long pulses to
0 mV in the presence of 4 mM NiCl2 ( s ) demonstrates the lack of nickel
extra-block at this activation voltage. [cell 19a]
increase in the efficiency of open channel block. The lack
of sizeable currents at these voltages requires the analysis
of tail currents to detect this effect. The amplitude of tail
currents recorded at 080 mV after depolarization to /60
/ 9k2d$$oc05 J975-7
FIG . 6. Recovery from enhancement of nickel block at 0 mV. Tail currents were recorded at 0 mV after 5- and 500-ms pulses to /60 mV; the
last 0.4 ms of the /60 mV depolarization is shown. Traces were obtained
both in the absence (control; 500-ms duration is the lower amplitude trace)
and presence of 6 mM NiCl2 (duration indicated on traces). [ cell 24e]
09-15-98 19:57:20
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 15, 2017
mV in the presence of 3 mM NiCl2 progressively decreased
as pulse duration ( Dt) increased from 25 to 1,000 ms (Fig.
4). Control tail current amplitude did not change in this
manner, but a slow phase became more prominent because
of complexities caused by the existence of a second open
state (McFarlane 1997). With nickel present, peak tail current amplitude (r ) decreased significantly as longer duration
pulses were delivered, leading to a roughly 50% increase in
block for long ( Dt Å 500–1,000 ms) versus short ( Dt Å
10 ms) pulses to /60 mV.
This time-dependent increase in the nickel blocking level
was quantified by an ‘‘extra-block’’ index (EB; see Fig. 5
legend), which compares peak tail current amplitude in the
presence of nickel for various duration pulses with /60
mV between individual cells (Fig. 5A). EB values were
computed for nine neurons in the presence of 6 mM NiCl2 ,
and mean values are shown plotted as a function of pulse
length ( , ), and fitted with a single exponential function
( t Å 140.4 ms).
Voltage dependence of Ni 2/ extra block was also investigated in these experiments. Instantaneous ICa-V curves obtained after 5-ms (Fig. 5B, ● ) or 500-ms pulses ( j ) to /60
mV under control conditions were not significantly different.
In the presence of 4 mM NiCl2 , the ICa-V curve for Dt Å 5
ms ( s ) was diminished by a constant factor over the entire
voltage range. For Dt Å 500 ms ( h ), however, an additional
decrease in slope of the ICa-V relationship was observed with
no obvious alteration in shape. These results indicate that,
once Ni 2/ extra-block was established in a time-dependent
manner, Ca 2/ channel block itself is not voltage dependent.
Establishment of extra-block, however, does display an
apparent voltage dependence because it only occurs at posi-
STATE-DEPENDENT CA 2/ CHANNEL BLOCK BY NICKEL IONS
DISCUSSION
This study investigated the effects of nickel on Ca 2/ channels in squid GFL neurons. Previous patch-clamp experiments have shown that these channels exhibit two voltagedependent gating modes that are well described by two open
states connected through a closed (or inactive) state in the
following scheme
C2 ã C1 ã O1 ã CI ã O2
(McFarlane 1997). Brief depolarization results in opening
to the first open state, O1 , but longer pulses to positive
potentials allow channels to pass through state CI and enter
the second open state, O2 . This transition is observable because channels in O2 exhibit slower deactivation kinetics
(McFarlane 1997). Furthermore, relief from block by vagatoxin IVA appears to occur concurrently with O2 occupancy. Like block caused by v-agatoxin IVA, that associated
with nickel was observed to proceed in a state-dependent
fashion. For the gating scheme described previously, this
discussion refers to nickel interaction with channels in three
basic states: closed channels (C2 and C1 ) and open channels
in either O1 or O2 .
Nickel ions interact with closed Ca 2/ channels
One effect of nickel appears to be the result of interaction
with closed states of GFL Ca 2/ channels because activation
/ 9k2d$$oc05 J975-7
kinetics are slower in the presence of Ni 2/ . This causes an
apparent positive shift in the voltage dependence of activation kinetics (Fig. 2). Additionally, this kinetic effect is
selective for channel activation because deactivation rates at
negative voltages are affected little if at all by nickel. The
idea that some divalent cation species interact with closed
channels and selectively slow activation kinetics is largely
based on detailed studies of the effects of Zn 2/ on Na /
(Gilly and Armstrong 1982a) and K / channels (Gilly and
Armstrong 1982b) in squid giant axons. The actions of Ni 2/
reported conform well with the ideas developed in these
previous studies. Although we did not carry out a detailed
modeling analysis of the effect of Ni 2/ on activation kinetics
of Ca 2/ channels in this study, we regard the qualitative
nature of the effects as sufficient evidence to propose a
nickel-induced stabilization of squid Ca 2/ channels.
In contrast to other blocking agents that stabilize closed
states of other types of Ca 2/ channels (Boland and Bean
1993; Patil et al. 1996), the voltage dependence of peak gCa
was seemingly unaffected by nickel (Fig. 1). Given the shift
of activation (but not deactivation) kinetics (Fig. 3), a small
shift in peak gCa is expected for a model in which forward
(but not reverse) rate constants are lower in the presence of
the blocking agent (Gilly and Armstrong 1982a). It is likely
that any shift of gCa is too small to resolve in our experiments,
but other explanations including alternative mechanisms
cannot be ruled out.
Both positive shifting of activation gating and slower activation kinetics in the presence of nickel were previously
noted for cloned neuronal Ca 2/ channels expressed in Xenopus oocytes, but a specific cause-and-effect relationship between slower opening kinetics and nickel stabilization of
closed states was not proposed (Zamponi et al. 1996). Although the degree of shifting of steady-state activation by
nickel may depend on the permeant ion species (Zamponi
et al. 1996), this question was not examined in the present
study.
Nickel blocks open channels: O1
Macroscopic Ca 2/ conductance assayed with a brief activating pulse is significantly smaller in the presence of nickel,
and the instantaneous ICa-V curve is reduced in slope without
a change in shape (Fig. 2B). These effects are almost certainly caused by a decrease in single channel conductance
as a result of Ni 2/ ions blocking open Ca 2/ channels. Single
channel blocking experiments revealed that Ni 2/ ions exhibit
very high frequency block (Chesnoy-Marchais 1985; Winegar et al. 1991), which leads to an apparent reduction
in the amplitude of single channel currents. Whereas open
channel blocking agents with appropriate rates for block and
unblock can cause the appearance of current inactivation
(see Armstrong 1971; Chow 1991), the very fast interactions
between Ni 2/ ions and open Ca 2/ channels (Winegar et al.
1991) can explain the lack of macroscopic current decay
during long pulses (Fig. 5C).
Other divalent cations such as Cd 2/ and Co 2/ exhibit only
flickery block of single Ca 2/ channels (Lansman et al. 1986;
Winegar et al. 1991). During block of GFL ICa these ions
do not lead to slowing of activation kinetics (Chow 1991;
unpublished observations). The interpretation of this result,
09-15-98 19:57:20
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 15, 2017
tive voltages. This is evident by direct examination of the
effects of Ni 2/ on inward ICa during a long pulse to 0 mV
(Fig. 5C). 4 mM NiCl2 ( s ) reduced ICa but did not induce
a time-dependent decay of ICa , indicating that nickel extrablock does not occur at 0 mV or at more negative voltages
(data not illustrated). Extra-block becomes progressively
greater as voltage increases from 0 to /60 mV (not illustrated).
One prediction arising from the two distinct ‘‘levels’’ of
nickel block as described is that channels should be able
to recover from extra-block at 0 mV. This recovery was
demonstrated directly by recording tail currents at 0 mV after
short ( Dt Å 5 ms) or long ( Dt Å 500 ms) pulses to /60
mV in the absence and presence of 6 mM Ni 2/ (Fig. 6). At
0 mV, the time course of the tail current reflects the closing
of some channels that were opened by depolarization to /60
mV. In the presence of nickel, the tail current recorded after
the short pulse was roughly a scaled-down version of the
control tail. As expected after development of extra-block
during the long pulse, the tail current amplitude was proportionally smaller, but rather than decaying this tail current
slowly increased in amplitude and converged with the current recorded after the short pulse.
This recovery process was slower at /20 mV, absent at
/40 mV, and markedly faster at 020 mV (data not shown).
For more negative voltages, at which channel closing is
favored, recovery was faster than channel deactivation. Both
of these observations are consistent with the overall voltage
dependence of an open state transition determined by measurement of the slowing of channel deactivation rate
(McFarlane 1997). This point will be considered in greater
detail.
1683
1684
M. B. MC FARLANE AND W. F. GILLY
based on the evidence from single channel experiments, is
that Cd 2/ and Co 2/ ions only block open GFL Ca 2/ channels. Moreover, it seems that the mechanism of Cd 2/ and
Co 2/ block does not discriminate between the two open
states, O1 and O2 . Supporting this idea is the fact that tail
currents exhibit similar ‘‘hooks’’ in the presence of Cd 2/
(Chow 1991) that do not depend on the length and strength
of prior depolarization (unpublished observations). Open
channel block was the main conclusion drawn from a more
detailed kinetic analysis of Cd 2/ block of GFL ICa (Chow
1991), and open channel block is further suggested by the
apparent one-to-one interaction between both Cd 2/ (Chow
1991) and Co 2/ and GFL Ca 2/ channels (unpublished observations).
Nickel blocks open channels: O2
Location of nickel ion binding sites
Similarities between Ca 2/ channel extra-block by Ni 2/
ions and relief of block by v-agatoxin IVA, a highly charged
polypeptide, suggest that the site of nickel action is likely
to be associated with the external surface of the channel
protein. Interactions between Ni 2/ ions and GFL Ca 2/ channels is more complex than a strict one-to-one association;
this is reflected in dose-response relationships that were well
fit with Hill coefficients of 1.5 and 1.6 (measured in 15 and
60 mM external Ca 2/ , respectively). Of all the divalent
metals tested for their ability to block GFL ICa (Cd 2/ , Co 2/ ,
Mn 2/ , Ni 2/ , and Pb 2/ ), Ni 2/ was the only divalent ion
that produced a Hill coefficient ú1.1 under these recording
/ 9k2d$$oc05 J975-7
Why nickel?
Although nickel was long considered a useful pharmacological tool for blocking low-voltage–activated Ca 2/ channels (Soong et al. 1993; Tsien et al. 1988), its relatively
low affinity block of high-voltage–activated Ca 2/ channels
revealed several unusual properties in this and other studies
that are apparently specific for nickel (Chesnoy-Marchais
1985; Shibuya and Douglas 1992; Winegar et al. 1991; Zamponi et al. 1996). But why nickel? The geometric flexibility
possessed by Ni 2/ ions in aqueous solution gives rise to
stable association with several amino acid residues (Hausinger 1993). This ability may confer the capacity for nickel
to substitute for Ca 2/ ions at regulatory binding sites on the
external protein surface (Kostyuk and Mironov 1986) where
other divalents may fail. Alternatively, individual Ni 2/ ions
carry up to six water molecules in aqueous solution (Hausinger 1993); hydration may thus enable weak interactions
(e.g., hydrogen bonding) with reactive residues over large
molecular distances, possibly spanning multiple channel subunits. Indeed, nickel binding to specific sites on multiple
channel subunits was demonstrated for cyclic nucleotidegated channels (Gordon and Zagotta 1995). Given these
observations and possibilities, nickel may prove to be a useful tool for probing important regulatory sites on the external
surface of Ca 2/ channels, which may in turn lead to further
comprehension of the nature of Ca 2/ channel gating.
We thank Dr. Reid J. Leonard (Merck & Co.) for generous computer
related support and Dr. Stuart H. Thompson for reviewing the manuscript.
09-15-98 19:57:20
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 15, 2017
Long pulses to /60 mV caused a progressive increase in
the amount of nickel block of GFL Ca 2/ channels, suggesting that an additional amount of nickel block (extrablock) is specifically associated with O2 (Fig. 5). The further
reduction of the slope of the instantaneous current-voltage
relationship for long versus short pulses implies two distinct
levels of block. More support for this inference comes from
the fact that the recovery from the extra-block induced by
long pulses to /60 mV can proceed at voltages (e.g., 0 mV)
favoring O1 occupancy.
Because neither an increase in the blocking efficacy of
nickel nor the relief of v-agatoxin IVA block nor the apparent entry to O2 can be observed at 0 mV (Fig. 5C) (McFarlane 1997), the correlation between nickel extra-block and
the entry to O2 is reinforced. The nickel-sensitive states responsible for extra-block during long pulses to /60 mV
cannot be limited to O1 because increased depolarization
leads to accumulation of channels in O2 (McFarlane 1997).
If nickel block was restricted to O1 , long pulses would cause
an apparent increase in current amplitude as channels entered
O2 during a long pulse to /60 mV. That the time course of
nickel extra-block ( t Ç140 ms at /60 mV) is substantially
faster than the time course of entry to O2 ( t Ç250 ms at
/60 mV) (McFarlane 1997), however, suggests that the
mechanism underlying extra-block might be more complex
than a simple flickery block of channels that entered O2 .
More specifically, Ni 2/ might alter rate constants leading
into or out of O2 or CI, the inactive state connecting O1
and O2 .
conditions (unpublished observations). Complex cooperativity was also observed for nickel block of cloned mammalian Ca 2/ channels expressed in Xenopus oocytes (Zamponi
et al. 1996). This result may not necessarily indicate that
more than one Ni 2/ ion must bind to cause block. Another
explanation, which may be more likely in this case, is that
complexity stems from the changing properties of the substrate, i.e., nickel-binding strength depends on channel state.
v-Agatoxin IVA block of GFL ICa exhibits a similar degree
of apparent cooperativity (n Å 1.4) (McFarlane and Gilly
1996), suggesting that these two agents interact with channels in a comparably complex fashion. Although one cannot
conclude that v-agatoxin IVA and Ni 2/ ions bind at the
same site, results of this paper are consistent with the hypothesis that open state transitions and state-dependent block
result from a common underlying mechanism.
Recent experiments suggested that the bulk of the effects
accompanying nickel block occurs as a result of Ni 2/ binding
to an important divalent binding site not directly associated
with the conduction pore (Zamponi et al. 1996). Such an
external site was postulated to influence gating voltage dependence and/or ion permeation characteristics (Kostyuk
and Mironov 1986), but its specific structural identity remains unknown (see Zamponi and Snutch 1996). The possibility exists that nickel block of some high-voltage–activated Ca 2/ channels may not exclusively involve deep pore
penetration or intrapore binding. This is indeed the case for
GFL Ca 2/ channels because the results of this paper clearly
demonstrate that nickel block is not voltage dependent, suggesting that very little interaction with the membrane electric
field occurs.
STATE-DEPENDENT CA 2/ CHANNEL BLOCK BY NICKEL IONS
This work was supported by National Institute of Neurological Disorders
and Stroke Grant NS-17510 and a predoctoral fellowship from the Ford
Foundation.
Address for reprint requests: M. B. McFarlane, Hopkins Marine Station,
Pacific Grove, CA 93950.
Received 1 December 1997; accepted in final form 15 June 1998.
REFERENCES
/ 9k2d$$oc05 J975-7
current and its applications to conduction and excitation in nerve. J.
Physiol. (Lond.) 117: 500–514, 1952.
JONES, K. T. AND SHARPE, G. R. Ni 2/ blocks the Ca influx in human keratinocytes following a rise in extracellular Ca. Exp. Cell. Res. 212: 409–
413, 1994.
KOSTYUK, P. G. AND MIRONOV, S. L. Some predictions concerning the calcium channel model with different conformational states. Gen. Physiol.
Biophys. 6: 649–654, 1986.
LANSMAN, J. B., HESS, P., AND TSIEN, R. W. Blockade of current through
single calcium channels by Cd, Mg, and Ca. Voltage and concentration
dependence of calcium entry into the pore. J. Gen. Physiol. 88: 321–
347, 1986.
MC FARLANE, M. B. Depolarization-induced slowing of Ca 2/ channel deactivation kinetics in squid neurons. Biophys. J. 72: 1607–1621, 1997.
MC FARLANE, M. B. AND GILLY, W. F. Spatial localization of calcium channels in giant fiber lobe neurons of the squid ( Loligo opalescens). Proc.
Natl. Acad. Sci. USA 93: 5067–5071, 1996.
PATIL, P., DE LEON, M., REED, R., DUBEL, S., SNUTCH, T. P., AND YUE, D. T.
Elementary events underlying voltage-dependent G-protein inhibition of
N-type Ca channels (Abstract). Biophys. J. 70: 28, 1996.
SALA, F. Activation kinetics of calcium currents in bull-frog sympathetic
neurones. J. Physiol. (Lond.) 437: 221–238, 1991.
SHIBUYA, I. AND DOUGLAS, W. W. Calcium channels in rat melanotrophs
are permeable to manganese, cobalt, cadmium, and lanthanum, but not
to nickel: evidence provided by fluorescence changes in fura-2-loaded
cells. Endocrinology 131: 1936–1941, 1992.
SOONG, T. W., STEA, A., HODSON, C. D., DUBEL, S. J., VINCENT, S. R., AND
SNUTCH, T. P. Structure and functional expression of a member of the
low voltage-activated calcium channel family. Science 260: 1133–1136,
1993.
TSIEN, R. W., LIPSCOMBE, D., MADISON, D. V., BLEY, K. R., AND FOX,
A. P. Multiple types of neuronal calcium channels and their selective
modulation. Trends Neurosci. 11: 431–438, 1988.
WINEGAR, B. D., KELLY, R., AND LANSMAN, J. B. Block of current through
single calcium channels by Fe, Co and Ni. Location of the transition
metal binding site in the pore. J. Gen. Physiol. 97: 351–367, 1991.
ZAMPONI, G. W., BOURINET, E., AND SNUTCH, T. P. Nickel block of a family
of neuronal calcium channels: Subtype- and subunit-dependent action at
multiple sites. J. Membr. Biol. 151: 77–90, 1996.
ZAMPONI, G. W. AND SNUTCH, T. P. Evidence for a specific site for modulation of calcium channel activation by external ions. Pflügers Arch. 431:
470–472, 1996.
09-15-98 19:57:20
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.1 on June 15, 2017
ARMSTRONG, C. M. Interaction of tetraethylammonium ion derivatives with
the potassium channels of giant axons. J. Gen. Physiol. 58: 413–437,
1971.
BOLAND, L. M. AND BEAN, B. P. Modulation of N-type calcium channels in
bullfrog sympathetic neurons by luteinizing hormone-releasing hormone:
kinetics and voltage-dependence. J. Neurosci. 13: 516–533, 1993.
BYERLY, L., CHASE, P. B., AND STIMERS, J. R. Permeation and interaction
of divalent cations in calcium channels of snail neurons. J. Gen. Physiol.
85: 491–518, 1985.
CHESNOY-MARCHAIS, D. Kinetic properties and selectivity of calcium-permeable single channels in Aplysia neurones. J. Physiol. (Lond.) 367:
457–488, 1985.
CHOW, R. H. Cadmium block of squid calcium currents. Macroscopic data
and a kinetic model. J. Gen. Physiol. 98: 751–770, 1991.
GILLY, W. F. AND ARMSTRONG, C. M. Slowing of sodium channel opening
kinetics in squid axon by extracellular zinc. J. Gen. Physiol. 79: 935–
964, 1982.
GILLY, W. F. AND ARMSTRONG, C. M. Divalent cations and the activation
kinetics of potassium channels in squid giant axons. J. Gen. Physiol. 79:
965–996, 1982.
GORDON, S. E. AND ZAGOTTA, W. N. Subunit interactions in coordination
of Ni 2/ in cyclic nucleotide-gated channels. Proc. Natl. Acad. Sci. USA
92: 10222–10226, 1995.
HAGIWARA, S. AND BYERLY, L. Calcium channel. Annu. Rev. Neurosci. 4:
69–125, 1981.
HAUSINGER, R. P. Biochemistry of Nickel. New York: Plenum, 1993.
HESS, P. AND TSIEN, R. W. Mechanism of ion permeation through calcium
channels. Nature 309: 453–456, 1984.
HILLE, B. Ionic Channels of Excitable Membranes. Sunderland, MA: Sinauer, 1992.
HODGKIN, A. L. AND HUXLEY, A. F. A quantitative description of membrane
1685