Download Modulation of Jellyfish Potassium Channels by External Potassium

Modulation of Jellyfish Potassium Channels by External
Potassium Ions
NIKITA G. GRIGORIEV,1 J. DAVID SPAFFORD,1 AND ANDREW N. SPENCER2
1
Department of Cell Biology and Anatomy, Faculty of Medicine, University of Calgary, Health Sciences Centre, Calgary,
Alberta T2N 4N1; and 2Department of Biological Sciences, The University of Alberta, Edmonton, Alberta T6G 2E9 and
Bamfield Marine Station, Bamfield, British Columbia V0R 1B0, Canada
INTRODUCTION
The large diversity of potassium channels provides richness to
the functional repertoire of excitable cells. Their associated currents shape action potentials and regulate firing frequency in
addition to maintaining the membrane resting potential (Connor
and Stevens 1971; Hille 1992). The amplitude of the ionic current
passing through the selective pore of these channels is determined
by channel conductance, the transmembrane electrical field, and
the potassium gradient. When the extracellular potassium concentration ([K1]out) is increased, the chemical driving force is reduced, which results in reduced currents. In addition to this
intrinsic property of all ionic channel membranes, it has been
shown that altering [K1]out produces a broad spectrum of modulatory effects on the delayed rectifier current in Xenopus axonal
membrane (Safronov and Vogel 1996), on the fast inactivating
K1 current in rat hippocampal neurons (Pardo et al. 1992), and on
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1728
heterologously expressed cloned representatives of the Shaker
subfamily of rapidly inactivating channels (Baukrowitz and
Yellen 1995; Lopez-Barneo et al. 1993; Tseng and Tseng-Crank
1992). These modulatory effects include the following: regulation
of the number of channels available for activation, alterations in
the rate of C-type inactivation, and changes in frequency-dependent cumulative inactivation resulting from an interaction between
N- and C-type inactivation. It has been suggested that these
modulatory effects might allow excitable cells to compensate for
increases in [K1]out in intercellular space as a result of repetitive
firing. Hounsgaard and Nicholson (1983) clearly demonstrated
that an elevation of [K1]out by as little as 1 mM may alter the
pattern of spontaneous activity in guinea-pig Purkinje neurons.
Synchronous activation of a large population of these cells in
brain slices can raise the level of [K1]out from 6 to 10 mM.
In higher vertebrates, extracellular potassium concentrations
are regulated as a result of homeostatic mechanisms at the
organ, tissue, and cellular levels. In contrast, there are no
known organs or tissues maintaining relatively constant extracellular potassium levels in lower metazoans such as the cnidarians (hydroids, jellyfish, corals, etc). It is likely that neurons
and other excitable cells in jellyfish have to be able to adapt to
unstable extracellular potassium concentrations as a result of
phasic accumulation of potassium during firing, or tonic
changes due to naturally occurring salinity differences.
Despite their phylogenetic position at the base of the metazoans, hydrozoan jellyfish display a variety of potassium channels and currents (Meech and Mackie 1993; Przysiezniak and
Spencer 1994). Several members of Shaker and Shal subfamilies of genes encoding a-subunits of voltage-gated potassium
channels were cloned from the jellyfish Polyorchis penicillatus. Their products, when expressed in the Xenopus oocyte
expression system, demonstrate biophysical properties similar
to other known A-like currents (Jegla et al. 1995; Jegla and
Salkoff 1997). The data presented in this paper demonstrate
that potassium channels in motor neurons from the jellyfish
Polyorchis penicillatus are modulated by [K1]out as are two
heterologously expressed Shaker channels, jShak1 and jShak2,
from Polyorchis. Modulation by potassium is likely to be
essential for nervous systems that experience variable external
concentrations of this ion. Increasing [K1]out increased the
peak current amplitude for all three channels studied. We have
shown that occupation by K1 binding sites in the external
channel mouth in the closed state is crucial for modulation of
current amplitude by extracellular potassium. jShak2 channels
exhibited strong potassium dependence of the inactivation rate
0022-3077/99 $5.00 Copyright © 1999 The American Physiological Society
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on June 15, 2017
Grigoriev, Nikita G., David Spafford, and Andrew N. Spencer.
Modulation of jellyfish potassium channels by external potassium
ions. J. Neurophysiol. 82: 1728 –1739, 1999. The amplitude of an
A-like potassium current (IKfast) in identified cultured motor neurons
isolated from the jellyfish Polyorchis penicillatus was found to be
strongly modulated by extracellular potassium ([K1]out). When expressed in Xenopus oocytes, two jellyfish Shaker-like genes, jShak1
and jShak2, coding for potassium channels, exhibited similar modulation by [K1]out over a range of concentrations from 0 to 100 mM.
jShak2-encoded channels also showed a decreased rate of inactivation
and an increased rate of recovery from inactivation at high [K1]out.
Using site-directed mutagenesis we show that inactivation of jShak2
can be ascribed to an unusual combination of a weak “implicit”
N-type inactivation mechanism and a strong, fast, potassium-sensitive
C-type mechanism. Interaction between the two forms of inactivation
is responsible for the potassium dependence of cumulative inactivation. Inactivation of jShak1 was determined primarily by a strong “ball
and chain” mechanism similar to fruit fly Shaker channels. Experiments using fast perfusion of outside-out patches with jShak2 channels were used to establish that the effects of [K1]out on the peak
current amplitude and inactivation were due to processes occurring at
either different sites located at the external channel mouth with
different retention times for potassium ions, or at the same site(s)
where retention time is determined by state-dependent conformations
of the channel protein. The possible physiological implications of
potassium sensitivity of high-threshold potassium A-like currents is
discussed.
MODULATION OF K1 CHANNELS BY EXTERNAL K1
METHODS
Molecular biology
All jShak2 mutants were constructed using cassette, PCR-based,
site-directed mutagenesis as described previously (Grigoriev et al.
1997). Mutants were verified by sequencing in both directions using
a Perkin-Elmer ABI 373A sequencer and an ABI Prism Dye-Terminator Cycle Sequencing Kit. Construction of the mutant jShak1D2–24
(jShak1T) was described by Jegla et al. (1995). Capped mRNAs were
prepared by run-off transcription using mMessage mMachine kits
(Ambion) for T3 (jShak1 and jShak1D2–24) or T7 (jShak2 and jShak2
mutants). The plasmid containing a rat skeletal muscle sodium channel, the a-subunit gene, rSkM1, was linearized with Sal I and transcribed using the Ambion kit for T7.
Electrophysiological recording from swimming motor
neurons
Primary cultures of swimming motor neurons of the jellyfish
Polyorchis penicillatus were prepared as described previously
(Przysiesniak and Spencer 1994) with the modification that cells were
exposed, with agitation, to collagenase for only 1 h. Swimming motor
neurons were identified by their large size, clear cytoplasm, and a
nucleus surrounded by membranous structures. Whole cell patch
recordings were made using 1–2 MV borosilicate glass pipettes filled
with a solution that contained (in mM) 500 KCl, 2 MgCl2,10 HEPES,
1 CaCl2, and 11 EGTA at pH 7.5 adjusted with N-methyl glucamine
(NMG). The extracellular bathing solution contained (in mM) 450
NMG-Cl, 50 MgCl2, and 10 HEPES, at pH 7.5 adjusted with HCl.
Potassium was introduced at the indicated concentrations by equimolar NMG substitution. Cells were microperfused using a manifold
with a dead volume of ,1 ml. Solutions were completely exchanged
within 2 s. All recordings were carried out at room temperature
(20 –22°C).
Whole cell, two-electrode recording from Xenopus oocytes
Xenopus oocytes were prepared and injected with mRNA as previously described (Grigoriev et al. 1997). mRNA (1–5 ng) was injected in each oocyte using a volume of 50 nl. The amount of injected
RNA was adjusted for each expressed channel type to minimize the
effects introduced by high levels of channel expression (Grigoriev et
al. 1999). Whole cell currents were recorded between 2 and 3 days
after injection using a two-microelectrode voltage clamp (CA-1, Dagan, Minneapolis, MN). Cells were constantly microperfused with a
gravity fed system.
Outside-out macropatch recording from Xenopus oocytes
Outside-out macropatches were obtained and recordings made as
described by Stühmer et al. (1992). Patch recordings, as well as whole
cell patch recordings from swimming motor neurons, were made
using an Axopatch 1D amplifier (Axon Instruments). Fast perfusion
was via two PE tubes (7405, Intramedic) glued to the bottom of a 35
mm plastic Petri dish with their orifices perpendicular and in contact
with one another. The two resulting perfusion streams were deflected
where they met and then ran parallel to one another. Because of the
small dimensions and relatively low flow rates, the perfusion system
was operating at a low enough Reynold’s number as to produce
laminar flow. Because NMG substituted for potassium, it was possible
to visualize an optical density difference between the K mM 5 0 mM
and the K mM 5 100 mM streams under phase contrast optics, and
mixing was not seen at the boundary between the two solutions. Each
of these tubes was connected to a reservoir filled with the relevant
solution. These reservoirs were separately connected to a pair of
Picospritzers (General Valve Corporation, Fairfield, NJ) that were
controlled by a computer. At the beginning of each recording session,
control and test solutions were perfused at 0.5 cm s21 by gravity.
Pipettes containing the macropatch were positioned closer to the
orifice of the tube containing the control solution than the test solution. Pressure pulses (1 bar) applied to the reservoir containing the test
solution by a Picospritzer increased the velocity of the test solution to
5 cm s21, thereby deflecting the control stream and exposing the
macropatch to the test solution. Solutions could be changed in this
way within 1 ms. The extracellular solution without potassium,
[K1]out 5 0, contained 100 mM NMG-Cl, 3 mM MgCl2, 10 mM
HEPES-acid adjusted to pH 7.5; whereas the solution, [K1]out 5 100
mM contained KCl 95 mM, MgCl2 3 mM, 10 mM HEPES/1/2K
adjusted to pH 7.5. Intermediate concentrations of K1 were made by
mixing these two solutions in the required proportion. The intracellular solution contained (in mM) 100 KCl 100, 3 MgCl2, 10 EGTA,
and 10 HEPES adjusted to pH 7.5. Experiments were carried out at
20°C using a temperature controller TC-10 (Dagan, Minneapolis,
MN). Conductance calculations for whole oocyte recordings were
made using intracellular potassium activity of 147 mM as reported by
Kusano et al. (1982).
Data acquisition and experimental control
All data acquisition and experimental control was achieved with a
Digidata 1200 acquisition system (Axon Instruments, Foster City,
CA) running pClamp 6.1 software (Axon Instruments). Analysis and
fitting of experimental data were done using the Clampfit program of
the pClamp 6.1 suite and SigmaPlot 4.00 (SPSS, Chicago, IL). All
results are expressed as means 6 SE.
RESULTS
External potassium modulates A-like currents in identified
neurons and heterologously expressed jellyfish Shaker
channels
Recordings from cultured, identified, motor neurons that
control swimming demonstrate the presence of two potassium
currents; a fast inactivating A-like current, IK-fast, and a delayed
rectifier, IK-slow (Przysiezniak and Spencer 1994). In this study
we were able to show that IK-fast current was strongly modu-
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on June 15, 2017
while recovery from inactivation was also affected by variations in [K1]out. Shaker potassium channels have at least two
inactivation mechanisms: N- and C-type. N-type inactivation is
associated with blockage of ionic current by the cytoplasmic
N-terminus of the a-subunit of the channel protein, often called
the “ball and chain” mechanism of inactivation (Hoshi et al.
1990; Zagotta et al. 1990). The molecular mechanism of Ctype inactivation is less clear, but it appears to be associated
with conformational changes in the external mouth of the
channel (Baukrowitz and Yellen 1995; Liu et al. 1996; Molina
et al. 1997).
Mutational analysis indicated that jShak2 channels show a
strong, potassium-dependent C-type inactivation mechanism and
that N-type inactivation is weak and “implicit.” Enhancement of
N-type inactivation of jShak2 by transplantation of a charged
amino acid cluster from the N-terminus of the jShak1 channel
sequence (jShak1 channels possess strong “explicit” N-type inactivation) rendered the inactivation rate less susceptible to variation
of [K1]out. We suggest that an interplay between weak, “implicit,”
N-type inactivation and strong, and fast, C-type inactivation of
jShak2 is responsible for the potassium-dependent cumulative
inactivation observed in this channel.
1729
1730
N. G. GRIGORIEV, J. D. SPAFFORD, AND A. N. SPENCER
lated by [K1]out (Fig. 1). Eliminating potassium ions from the
external solution completely inhibited this current (Fig. 1,
inset), leaving a “potassium insensitive,” slowly activating
current, IK-slow. Increasing [K1]out from 1 to 100 mM amplified
the peak, whole cell conductance of channels passing IK-fast
more than threefold (Fig. 1). We were able to demonstrate the
potassium insensitivity of IK-slow by using a holding potential
of 240 mV, which completely inactivated IK-fast (Przysiezniak
and Spencer 1994), and then altering [K1]out from 1 to 100
mM (data not shown).
Jellyfish jShak1 and jShak2 genes encode high-threshold, Alike currents (Jegla et al. 1995). When expressed in Xenopus
oocytes, jShak1 and jShak2 currents showed a strong dependence
on the concentration of extracellular potassium ions (Fig. 2, A and
B). When [K1]out was increased from 1 to 100 mM, the conductance for both channels increased by approximately fourfold and
the current by 33%; we call this the amplitude effect. The effect
of increasing [K1]out on the conductances of jShak1 and jShak2
could be fitted by a combination of an equation similar to a
Michaelis-Menten relationship, having an apparent Km of ;1.5
mM, and a linear relationship with a slope of 0.005 mM21 (Fig.
2C). We suggest that this nonhyperbolic component represents a
low-affinity mechanism that appears as a linear slope within the
range of concentrations used.
With increasing concentrations of external K1, jShak2 inactivated more slowly (Fig. 2D, t 5 15.67 6 0.27 ms and
71.81 6 7.75 ms at [K1]out 5 1 mM and 100 mM, respectively, n 5 9). This change in t with alterations in [K1]out was
not seen in either jShak1 (Fig. 2D) or IK-fast (Fig. 1, inset). The
time constant of IK-fast inactivation was not significantly different (P 5 0.45, n 5 5) between 1 and 100 mM [K1]out.
Several monovalent cations, other than K1, produced qualitatively similar effects on jShak2 current, but with substantially lower efficacy than potassium ions. Their modulatory
effectiveness for both current amplitude and inactivation kinetics could be ranked as follows: K1 . Rb1 . Cs1 . Na1
(data not shown). A similar sequence of sensitivity of current
peak amplitude to extracellular monovalent cations was dem-
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on June 15, 2017
FIG. 1. Rapidly inactivating component, IK-fast, of the outward potassium current in swimming motor neurons is sensitive to
[K1]out. Plots of conductance of IK-fast vs. membrane potential at [K1]out 5 1 mM (●) and 100 mM ( ); results are from 5
experiments. Conductance of IK-fast was calculated after isolation of the rapidly inactivating current by digital subtraction of IK-slow
(using a holding potential of 240 mV) from total potassium current (using a holding potential of 280 mV). Currents were
normalized to the maximum current at [K1]out 5 1 mM. Data were fitted with a Boltzmann equation (g9 5 g9max/(g9max 1
exp[2(V 2 V50)/a], where g9 5 g/gmax [K1]out 5 1mM is normalized conductance at voltage V, g9max is maximal normalized
conductance, V50 is voltage at which g9 5 g9max/2 and a is the slope factor. Conductances were determined by correcting peak
current for driving force, calculated by using the Nernst equation for the extra- and intracellular potassium concentrations used in
the experiments. Inset: typical potassium current traces recorded in the whole cell recording configuration at external [K1] of 0,
1, and 100 mM produced by depolarizing pulses to 160 mV from a holding potential of 280 mV. Note that the current remaining
at [K1]out 5 0 mM is a slowly inactivating current IK-slow (Przysiezniak and Spencer 1994) that is [K1]out insensitive.
MODULATION OF K1 CHANNELS BY EXTERNAL K1
1731
onstrated for the rat neuronal channel, RCK4 or Kv1.4 (Pardo
et al. 1992).
Modulatory effects of [K1]out involve N-type inactivation
It is well established that N-type inactivation in Shaker
potassium channels involves a “ball and chain ” mechanism
(Hoshi et al. 1990; Zagotta et al. 1990). Deletion of the cytoplasmic N-terminus of the Shaker channel protein eliminates
fast (N-type) inactivation revealing a slow (C-type) inactivation that has been associated with conformational changes in
the external mouth of the channel (Baukrowitz and Yellen
1995; Liu et al. 1996; Molina et al. 1997).
Extracellular potassium strongly affects the inactivation rate
of jShak2 but not of jShak1 currents, which may reflect differences in the molecular mechanism of inactivation. It is known
that partial blockade of Shaker channels by extracellular application of TEA is accompanied by a concomitant slowing
down of C-type inactivation. Conversely, rapid N-type inactivation is TEA-insensitive (Choi et al. 1991; Grissmer and
Cahalan 1989). Extracellular administration of TEA slowed
down the rate of inactivation of jShak2 but not fast inactivation
of jShak1 (Fig. 3), indicating that the mechanisms of inactivation of jShak1 and jShak2 differ.
To examine N- and C-type inactivation mechanisms more
closely, we constructed mutants with modified N-termini (Fig.
4). A previous study (Jegla et al. 1995) showed that deletion of
the first 23 residues of jShak1 yielded channels lacking fast
inactivation (Fig. 5A). This mutant, jShak1D2–24, showed
noticeable conductance even in the absence of external potassium. In the absence of [K1]out there were two distinct components of activation of jShak1D2–24, slow and fast. Presum-
FIG. 3. Effects of TEA administration on inactivation rate of jShak1 and jShak2 currents. A and B: currents elicited in jShak1 and
jShak2 by depolarization from a holding potential of 280 to 165 mV before (control), and after 10 mM TEA administration. Dotted
traces are scaled up current traces obtained after TEA administration. The external potassium concentration was 1 mM. The ratio “t for
inactivation in 10 mM TEA/t for inactivation in control” was 1.01 6 0.1 for jShak1 (n 5 4) and 7.1 6 2.1 for jShak2 (n 5 6).
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on June 15, 2017
FIG. 2. Modulation of cloned jellyfish potassium currents by [K1]out. A and B: whole
cell currents obtained from oocytes injected
with cRNA coding for jShak1 and jShak2, respectively. Currents were elicited by depolarization to 165 mV from a holding potential of
280 mV at [K1]out 5 0, 1, 20, 100 mM. C:
plot of normalized conductances of jShak1 (●)
and jShak2 (ƒ) vs. the extracellular potassium
concentration. The experimental data were fitted using the equation g 5 gmax*[K1]out/
(Kgapp 1 [K1]out) 1 [K1]out *Cg where gmax
is maximal conductance at [K1]out 5 100 mM,
Kgapp is the apparent constant of the hyperbolic
component, and Cg is the slope of the linear
component. D: plot of inactivation time constants determined for jShak1 (●) and jShak2 (ƒ)
currents vs. extracellular potassium concentration. Inactivation time constants were determined by fitting current decays with single
exponentials. The curve for jShak2 was obtained by fitting the equation t 5 tmax*[K1]out/
(Ktapp 1 [K1]out) 1 [K1]out *Ct where tmax is
the maximal time constant of inactivation at
[K1]out 5 100 mM, Ktapp is the apparent constant of the hyperbolic component, and Ct is
the slope of the linear component. The data
points for jShak1 were connected by straight
lines. Data for C and D were obtained from 6
to 9 oocytes from different batches.
1732
N. G. GRIGORIEV, J. D. SPAFFORD, AND A. N. SPENCER
FIG. 4. Aligned amino acid sequences of the N-terminal region of Wild jShak 2 and jShak 1; the N-terminal deleted mutants,
jShak1D2–24 and jShak2D2–38 plus its additional mutants
jShak2N1, jShak2N2, jShak2N1/2. Positively charged residues
are shown in bold black and the negatively charged residues in
bold gray. Truncation of jShak1D2–24 and jShak2D2–38 mutants
were made on the basis of sequence alignment up to the beginning
of the sequence homologous to the T1-region (tetramerization
domain) (Li et al. 1992).
proportion of the fast process. Although the jShak2D2–38
mutant also demonstrated potassium dependence, recovery
from inactivation became monoexponential, indicating the
presence of only one mechanism. Disappearance of the slowrecovering component of the jShak2 current after N-terminal
truncation shows that this part of the protein is involved in the
inactivation process. Thus the presence of the second component of recovery in Wild-type jShak2 and its absence in
jShak2D2–38 together with the absence of a pronounced
change in the inactivation rate at low [K1]out leads us to the
assumption that jShak2 experiences both weak, N-type inactivation and strong, fast, C-type inactivation. It is important to
note that, under similar ionic conditions, C-type inactivation in
Drosophila Shaker channels was two orders of magnitude
slower than in jShak2 (Baukrowitz and Yellen 1995).
Comparisons of the sequences of jShak1, jShak2, and fly
Shaker N-termini indicate some structural conservation (Jegla
et al. 1995); however, there are differences in the number and
distribution of charged amino acids in the first 25 amino acids
(Fig. 4). jShak1 carries two distinct clusters of charge on its
N-terminus: RRKKE with a strong net positive charge, and
KDDE with a net negative charge. Such pronounced clusters
are not present in the N-terminal region of jShak2. It was
shown that reducing the charge on the N-terminus of fly Shaker
channels by incremental shortening of the N-terminus alters the
rate of inactivation (Zagotta et al. 1990). Furthermore, struc-
FIG. 5. Effect of external potassium ion concentration
on conductance and the time constant of inactivation in
the N-terminal deleted mutants jShak1D2–24 and
jShak2D2–38. A and B are current traces obtained under
different [K1]out conditions from oocytes expressing
jShak1D2–24 and jShak2D2–38, respectively. The stimulus protocol used was 100 ms depolarizing pulses of 165
mV from a holding potential of 280 mV. C: plot of the
peak conductance (g) vs. the concentration of extracellular
potassium for the 2 mutants, jShak1D2–24 (●) and
jShak2D2–38 (E). Conductance for jShak1D2–24 at 0 mM
of [K1]out was calculated for the fast-activating part of the
current. D: plot of the time constant of inactivation for
jShak2D2–38 (E) vs. the concentration of extracellular
potassium. Solid line results from fitting with the equations shown in Fig. 2. Dashed line is the result of fitting
experimental data for the Wild-type, which still has the
N-terminal region present.
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on June 15, 2017
ably the slow component resulted from additional activation of
the channels by potassium effluxing through the already open
channels and accumulating in the intermicrovillar space of the
oocyte (Grigoriev et al. 1999). Dose-response curves suggest
an increase in the apparent affinity of this mutant channel to
external potassium: the K value for gapp of jShak1D2–24 was
0.2 mM compared with 1.5 mM for Wild-type jShak1 (Fig. 5,
A and C). A truncated jShak2 mutant, (jShak2D2–38) was not
as sensitive to [K1]out, having a Kgapp of ;2 mM (Fig. 5, B
and C). The effect of N-terminal deletion on the inactivation
rate of jShak2 was not pronounced (Fig. 5, B and D), with
statistically significant differences only being observed for
concentrations of [K1]out . 30 mM. Although the results
obtained for jShak1D2–24 are consistent with a “ball and
chain” mechanism, the data obtained for jShak2D2–38 are
more difficult to associate with such a mechanism and suggest
that inactivation is due to rapid, C-type inactivation. To determine whether any N-type inactivation is present in jShak2, we
compared the rates of recovery from inactivation in both
jShak1 and jShak2 and their N-terminal deleted mutants using
a two pulse protocol (Fig. 6). Recovery from inactivation of
Wild jShak2 could be fitted with the sum of slow (t 5 1 s) and
fast (t 5 0.1 s) exponents, suggesting that there are two distinct
processes involved in inactivation. The relative contribution of
the slow and fast components depended on [K1]out, such that
increasing the external potassium concentration increased the
MODULATION OF K1 CHANNELS BY EXTERNAL K1
1733
tural analysis of the N-termini of RCK4 and Raw3 showed that
negatively and positively charged residues may be distributed
in such a way as to confer dipole properties on the ball (Antz
et al. 1997). Therefore it is possible that the transmembrane
electrical field orients the ball in the cytoplasmic mouth of the
channel during the inactivation process. It is also possible that
the presence of the charged residues in the N-terminus sequence might be required for its effective interaction with the
ball receptor located in the internal channel mouth. The presence of electrostatic and steric interactions has been suggested
for the ball peptide and the ball receptor (Holmgren et al.
1996).
To examine the influence of these clusters of charged residues on the sensitivity of inactivation kinetics to extracellular
potassium, we constructed mutants of jShak2 (Fig. 4) in which
positively (RRKKE) and negatively charged (KDDE) clusters
were transplanted from jShak1 to the jShak2 N-terminal region,
both separately (jShak2N1 and jShak2N2) and together
(jShak2N1/2). Introduction of a positively charged cluster
reduced the influence of [K1]out on the rate of inactivation,
whereas inclusion of a negatively charged cluster of amino
acids enhanced the potassium dependency of this phenomenon
(Fig. 7A). The decrease in potassium sensitivity of inactivation
shown by jShak2N1 was accompanied by a far slower (almost
100-fold) rate of recovery from inactivation (Fig. 7B). Conversely, addition of a negative cluster slightly increased the
rate of recovery (Fig. 7B). Inclusion of both charge clusters
(jShak2 N1/2) had intermediate effects on both the potassium
sensitivity of the inactivation rate and the time for recovery
from inactivation (Fig. 7B).
The effectiveness of the N-terminal region as an inactivation particle can be judged from the rate of recovery from
inactivation. Charge-dependent changes in the recovery
rates seen in N-terminal mutants probably reflect differences
in the strength of interaction between the inactivation particle and presumed receptor sites in the internal channel
mouth. Figure 7C shows the effect of these mutations on the
effectiveness of [K1]out modulation of both peak current
and the inactivation time constant. Altering the effectiveness of the “ball and chain” mechanism did not significantly
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on June 15, 2017
FIG. 6. N-terminal deletion affects recovery from inactivation in jShak2. Recovery from inactivation was tested using a 2-pulse
protocol. A conditioning pulse to 165 mV (300 ms duration) from a holding potential of 2100 mV was followed by a test pulse
of the same amplitude (150 –200 ms duration), after a variable time T. The ratio of the peak current amplitude during the second
pulse (P2) to the peak current amplitude during the 1st pulse (P1) was plotted against time T.  and ●, data obtained for Wild-type
jShak2 at [K1]out 5 1 and 100 mM, respectively; ƒ and E, data obtained for the N-terminal deleted mutant jShak2D2–38 at
[K1]out 5 1 and 100 mM, respectively. Dotted lines result from fitting the means with a single exponent for jShak2 D2–38 and the
solid lines from fitting with 2 exponents for Wild-type jShak2. Fitting yielded the following time constants: for jShak2 Wild-type
at [K1]out 5 1 mM, tslow 5 944 6 57.4 ms, tfast 5 83.6 6 18.5 ms, Aslow/Afast 5 5.4 (Aslow/Afast is the ratio of the amplitudes
of slow and fast exponents); for Wild-type jShak2 at [K1]out 5 100 mM, tslow 5 1,017 6 59.1 ms, tfast 5 70.3 6 2.2 ms,
Aslow/Afast 5 0.82; for jShak2 D2–38 at [K1]out 5 1 mM, t 5 415 6 22.6 ms; for jShak2 D2–38 at [K1]out 5 100 mM, t 5 100 6
8.5 ms. Data were obtained from 8 to 10 independent measurements from different oocytes.
1734
N. G. GRIGORIEV, J. D. SPAFFORD, AND A. N. SPENCER
change the amplitude effect of [K1]out, but these mutations
did noticeably alter the sensitivity of the rate of inactivation
to the extracellular potassium concentration. For example,
when N-type inactivation was more effective (jShak2N1),
the external potassium concentration no longer had a strong
influence on the inactivation rate. Decreasing the effectiveness of N-terminal inactivation (jShak2N2) had the opposite
effect. Compared with the Wild-type, there was an additional increase in the sensitivity of the inactivation mechanism to [K1]out. The mutant jShak2N1/2 demonstrated an
intermediate sensitivity of the inactivation rate to [K1]out.
Fast perfusion experiments reveal differences between
amplitude and inactivation effects of [K1]out
When patch recordings of jShak2 current were made in the
outside-out configuration at a high external [K1] (100 mM),
there was a noticeable increase in the inactivation rate (t 5
12.8 6 1.3 ms, n 5 6) compared with whole cell recordings
(t 5 71.8 6 7.7 ms, n 5 6; Fig. 8A). We suggest that this
phenomenon is associated with the accumulation of potassium
ions in intermicrovillar space close to the oocyte membrane
during channel activation when making whole cell recordings.
The original microvillar structure is not preserved in outside-
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on June 15, 2017
FIG. 7. Effect of extracellular potassium concentration on current amplitude and the inactivation kinetics of 3 N-terminal jShak2
mutants where clusters of charged residues were introduced from the sequence found in jShak1. The 3 mutants of jShak2 were as
follows: jShak2N1 (net positively charged cluster RRKKE), jShak2N2 (net negatively charged cluster KDDE), jShak2N1/2 (both
the above clusters inserted). The N-terminal sequences of these mutants are shown in Fig. 4. A: representative current traces for
Wild-type jShak2 and the above mutants obtained using test pulses to 165 mV from a holding potential of 2100 mV. Solid traces
were recorded in the presence of [K1]out 5 1 mM, and the dashed traces at [K1]out 5 100 mM. B: time courses of recovery from
inactivation for Wild-type jShak2 and various N-terminal mutants (●), Wild-type jShak2; E, N-terminal deleted mutant jShak2
D2–38; , jShak2N2; u, jShak2N1; , jShak2N1/2. The protocol used for these experiment was the same as described for Fig. 6
using 1 mM [K1]out. Data presented were obtained from 5 to 7 experiments for each curve. C: bar diagram showing the effect of
[K1]out on peak current amplitude and the time constant of inactivation. Data are presented as the ratio of peak current
recorded at [K1]out 5 100 mM to the peak current recorded at 1 mM [K1]out (■), and as the ratio of inactivation time constants
recorded at [K1]out 5 100 and 1 mM, respectively (o). Each bar represents 7 to 9 independent measurements in different
oocytes.
MODULATION OF K1 CHANNELS BY EXTERNAL K1
1735
out patch recordings, thus eliminating many of the barriers to
diffusion of potassium ions as they efflux through the pore
(Grigoriev et al. 1999).
Using a fast perfusion system (solution exchange in ,1
ms), we were able to detect differences in the number of
channels available on depolarization and in their rates of
inactivation when [K1]out was rapidly altered (Fig. 8A).
Channels that had been activated instantly changed their
time constants of inactivation from 12.8 6 1.3 ms (n 5 6)
to 3.5 6 0.3 ms (n 5 6) following a rapid switch from 100
to 0 mM [K1]out. In Fig. 8A the amplitude effect is seen as
a decrease in the peak amplitude of currents recorded at
increasing intervals after switching from 100 to 0 mM. The
rate of reduction in the peak current reflects the rate of
elimination of potassium from the channel mouth in the
resting state. This removal of [K1] from the channel mouth
by diffusion (dekalification) occurred more slowly in the
resting state (t 5 13 ms) than in the open state when the rate
of dekalification is comparable with the speed of switching
between solution (i.e., ,1 ms). These results indicate that
both the amplitude and inactivation effects of [K1]out are
due to processes occurring at either different sites with
different retention times for potassium ions or at the same
site where retention time is determined by conformation of
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on June 15, 2017
8. Outside-out, macropatch recordings of expressed jShak2 currents experiencing rapid changes of [K1]out. A: current
traces obtained by step depolarizations from a holding potential of 280 to 1100 mV. In this experiment 14 test pulses (only
4 shown for clarity) were applied at different times following a fast change of [K1]out from 100 to 0 mM. During the 30-s
interval between each test pulse the patch was bathed in [K1]out 5 100 mM. Inset shows plots of the normalized peak currents
(■) and normalized time constants of inactivation (E) vs. the time the step pulse application was applied after the change of
[K1]out from 100 to 0 mM. Both the peak currents and time constants of inactivation were normalized to the corresponding
parameters for the current at [K1]out 5 100 mM. The curve for normalized peak currents was fitted by a single exponent with
a time constant of 12.8 ms 6 1.3 ms (n 5 6). Data points for inactivation time constants were joined by straight lines (n 5
6). B: membrane current and channel activity (evaluated using current variance) during rapid switching to low [K1]out in
macropatches containing 10 – 40 channels. Top pair of traces are controls recorded with [K1]out 5 100 mM, and the bottom
pair are traces obtained when [K1]out was rapidly switched to 0 mM for 160 ms during the initial test pulse (shaded bar).
Each pair of traces consists of a current trace (I) and a current variance trace (s), which were obtained by averaging 16 paired
trials (control and experimental). The stimulation protocol consisted of paired test pulses of 165 mV given from a holding
potential of 280 mV (lowest trace). Trace I2 shows that when potassium ions are removed there is immediate channel closure
that is associated with a marked drop in variance. Channels that inactivated as a result of dekalification recovered slowly, as
can be seen from the decreased amplitude of the current peak produced by the 2nd test pulse. This occurred in spite of [K1]out
being returned to its initial level of 100 mM. The ratio of IP2/IP1 is a measure of the rate of recovery of channels from
inactivation (where IP1 and IP2 are the peak currents at the 1st and 2nd test pulses, respectively). The ratio IP2/IP1 (0.3 6 0.04)
following dekalification is significantly different (P 5 0.017) from the control ratio (0.47 6 0.04). The tail current, which
is seen after the 1st pulse during dekalification, is an artifact associated with an undercompensated slow capacitative current
because, unlike the control traces, no corresponding change in variance accompanies the current. The small downward
deflection of the current trace I2 seen at the end of the [K1]out 5 0 mM pulse probably represents the reappearance of current
passing through endogenous inward rectifier channel(s).
FIG.
1736
N. G. GRIGORIEV, J. D. SPAFFORD, AND A. N. SPENCER
the channel protein. Experiments in which dekalification of
channels occurred after opening indicate that, as a result of
losing potassium, they were converted to a state from which
recovery was slow. This happened in spite of a high [K1]out
being restored rapidly after inactivation (Fig. 8B).
Physiological implications of channel sensitivity to [K1]out
Although the data presented above showed that [K1]out can
regulate both the conductance and inactivation rate of jShak2,
we wanted to determine whether such modulation of channel
properties could modify the excitability properties so as to be
physiologically significant. We were able to produce synthetic
action potentials by co-expressing jShak2 and the rat skeletal
muscle sodium channel a-subunit, rSkM1 (Fig. 9A). These
action potentials repolarized with two distinct phases: a rapid,
early phase provided by the A-like properties of jShak2, followed by a slow phase presumably associated with an inward
rectifier current endogenous to Xenopus oocytes (Bauer at al.
1996). Increasing [K1]out from 1 to 40 mM in the presence of
constant [Na1]out caused an increase in the rate of early repolarization from 73.3 6 2 s21 (n 5 4) at [K1]out 5 1 mM to
86.7 6 4.5 s21 (n 5 4) at [K1]out 5 40 mM (P 5 0.034) and
an associated exaggeration of the plateau phase. Increasing
[K1]out caused late repolarization to become much slower and
the spike broader as a result of the decreased electrochemical
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on June 15, 2017
1
FIG. 9. Physiological effects of [K ]out and 4-aminopyridine (4-AP) on the shape of molecularly synthesized action potentials
and the effect of [K1]out on the cumulative inactivation of Wild jShak2 and jShak2 D2–38 during repetitive stimulation. A: synthetic
action potentials generated by oocytes injected with a mixture of RNAs encoding channels for inward and outward currents (the
K1 channel, jShak2 plus sodium channel, rSkM1). Action potentials were evoked by 3-ms depolarizing current pulses (5 mA) using
an extracellular solution containing 1 mM K1 (black trace), 40 mM of K1 (gray trace), and 40 mM of K1 with 2 mM of 4-AP
(gray dashed trace). In all cases the extracellular solution contained 60 mM Na1. The membrane potential was continuously
adjusted to 280 mV by injection of constant hyperpolarizing current. B: cumulative inactivation of Wild jShak2 (circles) and
jShak2 D2–38 (squares) as peak current (normalized to the amplitude at a stimulation rate 0.05 Hz) at different stimulation
frequencies and different [K1]out. Stimulus pulses were of 60 ms duration applied from 2100 to 165 mV at frequencies of 0.1,
1, and 2 Hz at [K1]out 5 1 mM (black symbols) and 100 mM (gray symbols).
MODULATION OF K1 CHANNELS BY EXTERNAL K1
DISCUSSION
Figure 10 is a kinetic model for both jShak1 and jShak2 and
provides a background for the discussion.
Jellyfish A-like current (IK-Fast) recorded from swimming
motor neurons and currents in oocytes expressing jShak1 and
jShak2 showed modulation of the peak current amplitude by
altering the external potassium concentration. Only jShak2
experienced modulation of its inactivation rate by changes in
[K1]out. An effect of external [K1 ] on current amplitude has
been reported for other potassium channels, such as the delayed rectifier in Xenopus axonal membrane (Safronov and
Vogel 1996), heterologously expressed mammalian neuronal
RCK4 channels (Pardo et al. 1992), and the T449K mutant of
fly Shaker (Lopez-Barneo et al. 1993). RCK4 channels in the
absence of external potassium became nonconducting, although gating currents could still be recorded, indicating that
the voltage-sensing mechanism remains operational. Mutation
of the pore region of the fly Shaker channel protein by substitution of tyrosine residue 449 with lysine made peak current
amplitude of this channel strongly dependent on [K1]out
(Lopez-Barneo et al. 1993). It was suggested that potassium
ions, as well as other monovalent cations, can occupy site(s) in
the external channel mouth preventing development of C-type
inactivation by a “foot in the door” mechanism (Labarca and
MacKinnon 1992; Lopez-Barneo et al. 1993). If depolarization
occurs in the absence of [K1]out, channels proceed rapidly to
the C-type inactivated state after opening (Fig. 10). C-type
inactivation occurs sufficiently rapidly so as to convert many
K1 channels to a nonconducting state before any significant
current can be recorded and before sites in the mouth become
occupied by potassium effluxing from the cytoplasm. Current
can only be detected when C-type inactivation is slowed sufficiently by extracellular potassium ions. An efficient ball and
FIG. 10. Kinetic model for the action of extracellular potassium on inactivation mechanisms in jellyfish Shaker channels, jShak1
and jShak2. Channels can be in 2 resting states, (RK1) when K1 (shown by *) occupies a site(s) in the channel mouth and R when
these sites are unoccupied by K1. In both these states, extracellular, but not intracellular, potassium ions have free access to the
binding site(s). On depolarization channels in both states can be opened. After opening channels without potassium ion(s) in the
channel mouth (O) immediately proceed to the C-type inactivated state (C). Open channels with potassium ion (OK1) in the channel
mouth can stay in the open state longer because this site is replenished by intracellular potassium ions transiting the pore. The
N-type inactivation mechanism (ball and chain) clogs the pore from the inside preventing site(s) in the channel mouth from being
replenished by intracellular potassium (NK1); diffusion of potassium ions away from the site(s) promotes fast conversion to the C
1 N inactivated state. Another possible fate of the OK1 channel is dekalification, which redirects its path to the O state. Asterisks
on the cartoons symbolize potassium ions.
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on June 15, 2017
driving force on K1 through endogenous inward rectifier channels. In these experiments the decreased resting potential resulting from an increase in [K1]out was compensated by adjusting the holding current in current clamp mode. Inhibition of
jShak2 current by application of 2 mM 4-aminopyridine (4-AP)
decreased the rate of early repolarization and positioned the
plateau closer to the level of the sodium reversal potential. It
should be noted that the endogenous inward rectifier current in
oocytes is not sensitive to 4-AP (Bauer et al. 1996). We suggest
that this modulation by [K1]out could stabilize the plateau
phase of action potentials when potassium accumulates in
restricted extracellular spaces during repetitive firing.
Another mechanism that could counteract the effect of potassium accumulation on the driving force and hence action
potential shape is the inhibitory effect of [K1]out on the cumulative inactivation of jShak2 in the course of repetitive stimulation (Fig. 9B). Elevation of extracellular potassium concentration makes cumulative inactivation of channels less
pronounced. The severe reduction of cumulative inactivation
observed for the jShak2 D2–38 mutant suggests that the Nterminal “ball” might play a pivotal role in this process.
1737
1738
N. G. GRIGORIEV, J. D. SPAFFORD, AND A. N. SPENCER
could substitute for K1 in the process of recovery of Shaker B
channel from inactivation. There are at least two possible
mechanisms by which occupation of this site(s) can promote
displacement of the inactivation ball. One explanation involves
electrostatic interaction between ion(s) occupying site(s) in the
external channel mouth and the inactivation ball (GomezLagunas and Armstrong 1994). Another suggested mechanism
is that occupation of the K1 binding site prevents the conformational changes associated with C-type inactivation. The
latter mechanism assumes that such conformational changes
can have a remote influence on more internalized parts of the
protein participating in ball-receptor interaction. The ultimate
effect of the conformational change is to hold the inactivation
particle in place for a longer time. This mechanism can also
explain the absence of visible jShak2 tail currents at low
concentrations of [K1]out, as seen in Fig. 2B, because most of
the inactivated channels are in a C 1 N type inactivated state
and channel reopening after repolarization (Ruppersberg et al.
1991) cannot be observed.
Both jShak2 and fly Shaker exhibit potassium-dependent
cumulative inactivation. This dependence was explained by
Baukrowitz and Yellen (1995) by assuming that there is an
interplay between an “explicit” N-type mechanism and slow
C-type inactivation. However, for jShak2 channels, we suggest
that the potassium dependence of cumulative inactivation is
caused by a combination of “implicit” N-type inactivation and
fast C-type inactivation.
All Polyorchis potassium-dependent currents, such as
jShak1, jShak2 and IK-fast, show a high activation threshold that
is associated with their roles in shaping the early repolarization
of the action potential. We also observed that when jShak2
contributes outward current in synthetic action potentials expressed in oocytes, this current is capable of both repolarizing
the action potential and forming a plateau. The endogenous
repolarizing outward current in swimming motor neurons of
Polyorchis, IK-fast, also truncates the plateau of action potentials, which has been shown to modulate neuromuscular transmission (Przysiezniak and Spencer 1994; Spencer 1984; Spencer et al. 1989). Jellyfish have no known tissues or cells, such
as glia, that are specialized for K1 homeostasis in the immediate extracellular space surrounding neurons. It is also important to note that nearly all cell types in hydromedusae are
electrically excitable, including epithelial cells (Satterlie and
Spencer 1987), which would drastically increase the number of
potential sources for potassium accumulation in extracellular
space. Therefore it is likely that potassium that accumulates
during repetitive firing could reduce outward current amplitude
and duration, thereby altering action potential shape. Because
IK-fast can be modulated by [K1]out, one can imagine that these
negative influences by accumulating K1 can be compensated
by increased current amplitude. Although the kinetics and
electrical properties of IK-fast are markedly different from those
of heterologously expressed jShak1 and jShak2 currents, we
cannot rule out the possibility that IK-fast is conducted by
channels composed of jShak1 or jShak2 a-subunits because
their properties may be altered by differences in lipid environments (Schetz and Anderson 1993) and/or by the presence of
auxiliary subunits when these channels are expressed in vivo.
In the mammalian CNS, extracellular potassium concentrations
can increase by several millimoles as a result of high-frequency
firing (Hounsgaard and Nicholson 1983; Sykova 1983). It is
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on June 15, 2017
chain mechanism, as was seen for jShak1 in this study, prevents occupation of the proposed potassium binding site(s) by
effluxing potassium ions, which can explain the increased
apparent affinity to [K1]out observed for the N-terminus truncated mutant. The “foot in the door” hypothesis for modulation
by [K1]out also provides an explanation for the effect of
[K1]out, and other monovalent cations, on inactivation of
jShak2. Inactivation of this channel occurred predominantly by
a C-type mechanism, with occupation of the potassium binding
site(s) slowing inactivation. The differences in retention times
of K1 for closed and open channels observed in experiments
involving fast dekalification probably reflects state-dependent
conformational changes, or accessibility, of potassium binding
site(s) located in the external channel vestibule. N-type inactivation of jShak2 channels is “implicit,” but experiments examining recovery from inactivation of Wild-type and
jShak2D2–38 unmasks the presence of an N-type inactivation
mechanism. Its presence can also be detected at [K1]out of .30
mM when C-type inactivation becomes very slow (Fig. 5D).
Conversely, inactivation of jShak1 is mostly by an efficient ball
and chain mechanism (Fig. 10). In the case of jShak2, enhancement of N-type inactivation by transplantation of the jShak1
RRKKE cluster (jShak2N1) makes the potassium dependence
of inactivation less pronounced because the potassium-independent ball and chain mechanism becomes more explicit.
According to the kinetic model suggested by Baukrowitz
and Yellen (1995) for fly Shaker channels, recovery from
N-type inactivation is fast, whereas recovery from C-type
inactivation is much slower. Interaction between slowly recovering, potassium-dependent, C-type inactivation and rapidly
recovering, potassium-independent N-type inactivation explains the potassium dependency of cumulative inactivation
observed for fly Shaker channels. In our experiments with
jShak2 channels, the slower component of recovery appears to
be tightly associated with N-type inactivation. Truncation of
jShak2 channels was accompanied by disappearance of the
slower component of recovery, and introduction of the
RRKKE cluster in the N-terminal sequence made recovery
dramatically slower. It is reasonable to suggest that the increase in effectiveness of interaction between the inactivation
particle and the receptor in the internal channel mouth simultaneously slows down the process of unbinding of this particle
from the receptor during recovery from inactivation. Our data
indicate that N-type inactivation in jShak2 channels is the
major mechanism involved in potassium-dependent cumulative inactivation.
At the end of a depolarizing pulse, jShak2 channels can be in
N, C, and C 1 N inactivated states (Fig. 10). jShak2 channels
with the N-terminus present show potassium-sensitive recovery from inactivation indicating that channels in the C 1 N
inactivated state recover more slowly than those in the N state.
Why do channels in the C 1 N state recover slowly and N-type
inactivated channels recover rapidly? Displacement of the Nparticle from the internal channel mouth will allow the channel
to proceed to the resting state and determine the rate of recovery from inactivation. Considering that impermeable Cs1 ions
(data not shown) can imitate the effect of K1 on the process of
recovery from inactivation in jShak2 then occupancy of the
site(s) at the external channel vestibule, rather than ion flow
through the reopened channels, can explain this effect. Similarly, Gomez-Lagunas and Armstrong (1994) reported that Cs1
MODULATION OF K1 CHANNELS BY EXTERNAL K1
probably significant that the potassium sensitivity shown by various vertebrate (Pardo et al. 1992; Safronov and Vogel 1996) and
fly (Baukrowitz and Yellen 1995) K1 channels can be satisfactorily described by curves with a Kapp close to 2 mM. By comparison, there are additional low-affinity sites that are modulating
potassium currents in Polyorchis. These sites are presumably an
adaptation for the lack of glia cells and the consequently greater
potassium accumulation in extracellular space.
Received 30 March 1999; accepted in final form 16 June 1999.
REFERENCES
ANTZ, C., GEWEYER, M., FAKLER B., SCHOTT, M. K., GUY, R., FRANCK, R.,
RUPPERSBERG, J. P., AND KALBITZER, H. R. NMR-structure of inactivation
gates from mammalian voltage-dependent potassium channels. Nature 385:
272–275, 1997.
BAUER, C. K., FALK, T., AND SCHWARZ, J. R. An endogenous inactivating
inward-rectifying potassium current in oocytes of Xenopus laevis. Pflügers
Arch. 432: 812– 820, 1996.
BAUKROWITZ, T. AND YELLEN, G. Modulation of K1 current by frequency and
external [K1]: A tale of two inactivation mechanisms. Neuron 15: 951–960,
1995.
CHOI, K. L., ALDRICH, R. W., AND YELLEN, G. Tetraethylammonium blockade
distinguishes two inactivation mechanisms in voltage-activated K1 channels. Proc. Natl. Acad. Sci. USA 88: 5092–5095, 1991.
CONNOR, J. A. AND STEVENS, C.F.L. Voltage-clamp studies of a transient
outward current in gastropod neural somata. J. Physiol. (Lond.) 213: 21–30,
1971.
GOMEZ-LAGUNAS, F. AND ARMSTRONG, C. M. The relation between ion permeation and recovery from inactivation of Shaker B K1 channels. Biophys.
J. 67: 1806 –1815, 1994.
GRIGORIEV, N. G., SPAFFORD, J. D., GALLIN, W. J., AND SPENCER, A. N. Voltage
sensing in jellyfish Shaker K1 channels. J. Exp. Biol. 200: 2919–2926, 1997.
GRIGORIEV, N. G., SPAFFORD, J. D., AND SPENCER, A. N. The effects of level of
expression of a jellyfish Shaker channel; a positive potassium feed-back
mechanism. J. Physiol. (Lond.) 517: 25–33, 1999.
GRISSMER, S. AND CAHALAN, M. TEA prevents inactivation while blocking
open K1 channels in human T-lymphocytes. Biophys. J. 55: 203–206, 1989.
HILLE, B. Ionic Channels of Excitable Membranes. Sunderland, MA: Sinauer,
1992, p. 607.
HOLMGREN, M., JURMAN, M. E., AND YELLEN, G. N-type inactivation and the
S4-S5 region of the Shaker K1 channel. J. Gen. Physiol. 108: 195–206, 1996.
HOSHI, T., ZAGOTTA, W. N., AND ALDRICH, R. W. Biophysical and molecular
mechanisms of Shaker potassium channel inactivation. Science 250: 533–
538, 1990.
HOUNSGAARD, D. J. AND NICHOLSON, C. Potassium accumulation around individual Purkinje cells in cerebellar slices from the guinea-pig. J. Physiol.
(Lond.) 340: 359 –388, 1983.
JEGLA, T., GRIGORIEV, N., GALLIN, W. J., SALKOFF, L., AND SPENCER, A. N.
Multiple Shaker potassium channels in a primitive metazoan. J. Neurosci.
15: 7889 –7999, 1995.
JEGLA, T. G. AND SALKOFF, L. A novel subunit for Shal K1 channels radically
alters activation and inactivation. J. Neurosci. 17: 32– 44, 1997.
KUSANO, K., MILEDI, R., AND STINNAKRE, J. Cholinergic and catecholaminergic
receptors in the Xenopus oocyte membrane. J. Physiol. (Lond.) 328: 142–
170, 1982.
LABARCA, P. AND MACKINNON, R. Permeant ions influence the rate of C-type
inactivation in Shaker channels. Biophys. J. 61: A378, 1992.
LEE, T. E., PHILIPSON, L. H., AND NELSON, D. J. N-type inactivation in the
mammalian Shaker K1 channel Kv1. 4. J. Membr. Biol. 151: 225–235,
1996.
LI, M., JAN, Y. N., AND JAN, L. Y. Specification of subunit assembly by the
hydrophilic amino-terminal domain of the Shaker potassium channel. Science 257: 1225–1230, 1992.
LIU, Y., JURMAN, M. E., AND YELLEN, G. Dynamic rearrangement of the outer
mouth of a K1 channel during gating. Neuron 16: 859 – 867, 1996.
LOPEZ-BARNEO, J., HOSHI, T., HEINEMANN, S. H., AND ALDRICH, R. Effects of
external cations and mutations in the pore region on C-type inactivation of
Shaker potassium channels. Receptors Channels 1: 61–71, 1993.
MEECH, R. W. AND MACKIE, G. O. Potassium channel family in giant motor
axons of Aglantha digitale. J. Neurophysiol. 69: 894 –901, 1993.
MOLINA, A., CASTELLANO, A. G., AND LOPEZ-BARNEO, J. Pore mutations in
Shaker K1 channels distinguish between the sites of tetraethylammonium blockade and C-type inactivation. J. Physiol. (Lond.) 449: 361–
367, 1997.
PARDO, A., HEINEMANN, S. H., TERLAU, H., LUDEWIG, U., LORRA, C., PONGS,
O., AND STUHMER, W. Extracellular K1 specifically modulates a rat brain K1
channel. Proc. Natl. Acad. Sci. USA 89: 2466 –2470, 1992.
PRZYSIEZNIAK, J. AND SPENCER, A. N. Voltage-activated potassium currents in
isolated motor neurons from the jellyfish Polyorchis penicillatus. J. Neurophysiol. 72: 1010 –1019, 1994.
RUPPERSBERG, J. P., FRANK, R., AND PONGS, O. Cloned neuronal IK(A) channels
reopen during recovery from inactivation. Nature 353: 657– 660, 1991.
SAFRONOV, B. V. AND VOGEL, W. Modulation of delayed rectifier K1 channel
activity by external K1 ions in Xenopus axon. Pflügers Arch. 430: 879 – 886,
1996.
SATTERLIE, R. A. AND SPENCER, A. N. Organization of conducting systems in
‘simple’ invertebrates: Porifera, Cnidaria and Ctenophora. In: Nervous Systems in Invertebrates, edited by M. A. Ali. New York: Plenum, 1987, p.
213–264.
SCHETZ, J. A. AND ANDERSON, P.A.V. Investigations of lipid components of
neuron-enriched membranes of the jellyfish Cyanea capillata. J. Exp. Biol.
177: 23–39, 1993.
SPENCER, A. N. The physiology of a coelenterate neuromuscular synapse.
J. Comp. Physiol. A Sens. Neural Behav. Physiol. 148: 353–363, 1984.
SPENCER, A. N., PRZYSIEZNIAK, J., ACOSTA-URQUIDI, J., AND BASARSKI, T. A.
Presynaptic spike broadening reduces junctional potential amplitude. Nature
340: 636 – 638, 1989.
STUHMER, W., TERLAU, H., AND HEINEMANN, S. H. Xenopus oocytes for
two-electrode and patch clamp recording. In: Practical Electrophysiological
Methods, edited by H. Kettenman and R. Grantyn. New York: Wiley-Liss,
1992, p. 186 –189.
SYKOVA, E. Extracellular K1 accumulation in the central nervous system.
Prog. Biophys. Mol. Biol. 42: 135–189, 1983.
TSENG, G. N. AND TSENG-CRANK, J. Differential effects of elevating [K]o on
three transient outward potassium channels. Circ. Res. 71: 657– 672,
1992.
ZAGOTTA, W. N., HOSHI, T., AND ALDRICH, R. W. Restoration of inactivation
in mutants of Shaker potassium channels by a peptide derived from ShB.
Science 250: 568 –571, 1990.
Downloaded from http://jn.physiology.org/ by 10.220.33.6 on June 15, 2017
We thank Bamfield Marine Station for providing excellent facilities. We are
especially grateful to W. Gallin for advice on molecular techniques and P.
Ruben for providing the plasmid containing the rSkM1 channel gene. We are
also grateful to T. Baukrowitz, P. Ruben, and S. Buckingham for stimulating
discussion and valuable advice. We especially thank Dr. W. Gallin for providing technical guidance for the molecular biological aspects of this study,
which were carried out in his laboratory.
Partial salary support for N. G. Grigoriev was provided by the Western
Canadian Universities Marine Biological Society. J. D. Spafford was supported
by an Alberta Heritage Foundation for Medical Research studentship. This
work was supported by a Natural Sciences and Engineering Research Council
Research grant to A. N. Spencer.
Address for reprint requests: A. N. Spencer, Bamfield Marine Station,
Bamfield, British Columbia V0R 1B0, Canada.
1739