Previous Article | Next Article 
The Journal of Neuroscience, April 15, 1999, 19(8):2906-2918
Functional Analysis of a Mouse Brain Elk-Type K+
Channel
Matthew C.
Trudeau1,
Steven A.
Titus2,
Janet L.
Branchaw1,
Barry
Ganetzky2, and
Gail A.
Robertson1
1 Department of Physiology, University of
Wisconsin-Madison Medical School, and 2 Laboratory of
Genetics, University of Wisconsin, Madison, Wisconsin 53706
 |
ABSTRACT |
Members of the Ether à go-go (Eag) K+
channel subfamilies Eag, Erg, and Elk are widely expressed in the
nervous system, but their neural functions in vivo
remain largely unknown. The biophysical properties of channels from the
Eag and Erg subfamilies have been described, and based on their
characteristic features and expression patterns, Erg channels have been
associated with native currents in the heart. Little is known about the
properties of channels from the Elk subfamily. We have identified a
mouse gene, Melk2, that encodes a predicted polypeptide
with 48% amino acid identity to Drosophila Elk but only
40 and 36% identity with mouse Erg (Merg) and Eag (Meag),
respectively. Melk2 RNA appears to be expressed at high
levels only in brain tissue. Functional expression of Melk2 in Xenopus oocytes reveals large,
transient peaks of current at the onset of depolarization. Like Meag
currents, Melk2 currents activate relatively quickly, but they lack the
nonsuperimposable Cole-Moore shift characteristic of the Eag
subfamily. Melk2 currents are insensitive to E-4031, a class III
antiarrhythmic compound that blocks the Human
Ether-à-go-go-Related Gene (HERG) channel and its counterpart in
native tissues, IKr. Melk2 channels exhibit inward
rectification because of a fast C-type inactivation mechanism, but the
slower rate of inactivation and the faster rate of activation results
in less inward rectification than that observed in HERG channels. This
characterization of Melk currents should aid in identification of
native counterparts to the Elk subfamily of channels in the nervous system.
Key words:
Melk2; Elk; Eag; brain; channel; C-type inactivation
| |
INTRODUCTION |
Members of the Ether-à-go-go
(Eag) family of K+ channels are expressed across a
broad range of species and tissues where they serve diverse
physiological roles. The observation that mutations in the
Drosophila gene eag cause spontaneous action
potentials in motor neurons and enhanced transmitter release at
the Drosophila neuromuscular junction (Ganetzky and
Wu, 1983
; Wu et al., 1983) was the first indication that these genes
regulate membrane excitability. The subsequent cloning of
eag (Drysdale et al., 1991; Warmke et al., 1991) led to the
identification of many related genes including the Human
Ether-à-go-go-Related Gene (HERG; Warmke and
Ganetzky, 1994), which is critical for maintaining the normal rhythmic
activity of the human heart (Curran et al., 1995).
Several members of the Eag and Erg subfamilies have been expressed in
heterologous systems with the aim of using the biophysical and
pharmacological properties revealed by such studies to identify their
corresponding currents in native tissues. Within the Eag subfamily, all
mammalian representatives identified to date have properties typical of
delayed rectifiers with no measurable inactivation (Ludwig et al.,
1994; Robertson et al., 1996; Frings et al., 1998); only
Drosophila Eag channels exhibit partial inactivation
(Robertson et al., 1996). A shared characteristic of members of the Eag
subfamily is a nonsuperimposable Cole-Moore shift (Ludwig et al.,
1994; Robertson et al., 1996), in which the time course of activation becomes slower and more sigmoidal as the prepulse or holding potential is made more negative and, as a consequence, the currents do not superimpose when aligned by shifting along the time axis (Cole and
Moore, 1960; Young and Moore, 1981).
In contrast to Eag channels, all Erg channels studied to date exhibit
significant inactivation. HERG channels open and then rapidly enter a
highly stable inactivated state during depolarization, effectively
suppressing outward current at positive voltages (Sanguinetti et al.,
1995; Trudeau et al., 1995) (cf. Shibasaki 1987). The predominant current thus occurs during repolarization as channels recover from inactivation and slowly deactivate (Zhou et al., 1998)
(cf. Zeng et al., 1995). These unusual gating properties together with
a sensitivity to the pharmacological agent E-4031 suggested that
HERG subunits underlie cardiac IKr (like that
described in native tissues by Sanguinetti and Jurkewicz,
1990) and led to the conclusion that mutations in
HERG cause long QT syndrome by disrupting this
repolarizing current (Sanguinetti et al., 1995; Trudeau et al., 1995).
As such, HERG is the only member of the Eag family with a clearly
defined native counterpart and physiological role.
We have identified a member of the Elk subfamily from mouse (Melk2)
which, based on sequence analysis, defines a separate subclass within
the Elk subfamily distinct from a rat channel first identified (Relk1;
S. Titus and B. Ganetzky, unpublished data) (Shi et al., 1998). We find
that Melk2 mRNA is expressed at high levels only in the
brain. In contrast to Eag or Erg currents, Melk2 currents expressed in
Xenopus oocytes exhibit a large, transient outward component
during depolarization. If presented with a subsequent repolarizing
ramp, a second component of outward current is evoked as channels
recover from inactivation and revisit the open state before closing.
These properties suggest that Melk2 channels may participate in action
potential repolarization as well as regulation of burst properties. A
detailed characterization is presented here to facilitate the
identification of Melk2 currents in native tissues.
| |
MATERIALS AND METHODS |
Molecular biology. A computer-based search of a mouse
dBEST database identified a partially sequenced (190 bp) mouse brain cDNA encoding a polypeptide segment that had significant similarity with the corresponding region of the Drosophila Elk
polypeptide. This cDNA (IMAGE consortium number 319419) was obtained
and sequenced on both strands using automated fluorescent sequencing
(model 377; Applied Biosystems, Foster City, CA). Sequence analysis
indicated that this cDNA encodes the complete open reading frame of a
member of the Elk subfamily of potassium channels, which we named
Melk2. Sequence alignments were generated and analyzed using
Lasergene software programs (DNAstar, Madison, WI).
For oocyte expression studies, the Melk2 cDNA was subcloned
into a derivative of the pGEMHE expression vector (Liman et al., 1992)
and linearized at the NotI site for cRNA transcription. The
HERG and Meag constructs used here were
previously described (Trudeau et al., 1995, 1996; Robertson et al.,
1996). Point mutations in Melk2 were generated with a
PCR-based approach using mutagenic primers (Higuchi et al., 1990). All
sequences generated by PCR were confirmed by sequencing on both strands
as described above.
The distribution of the Melk2 transcript was assessed using
a mouse multiple tissue Northern blot (Clontech, Palo Alto, CA) probed
with a 32P-labeled fragment corresponding to the 3' end of
Melk2 from bp 2410-2880. High-stringency conditions were
used with ExpressHyb solution (Clontech). Bands were visualized after a
3 d exposure to Biomax MS film at 
70°C.
Oocyte handling and RNA preparation. Procedures used for
oocyte preparation were the same as those detailed previously (Herzberg et al., 1998). Oocytes were surgically removed from anesthetized female
frogs (Xenopus laevis, Nasco) and defolliculated by
treatment with 1 mg/ml collagenase B (Boehringer Mannheim,
Indianapolis, IN), followed by an osmotic shock procedure (Pajor and
Wright, 1992). cRNAs were transcribed from the T7 promoter of DNA
templates using the mESSAGE mACHINE kit (Ambion, Austin, TX) and
diluted in sterile water to yield ~20 ng of cRNA per oocyte in an
injection volume of 37 nl. Oocytes were injected using a Drummond
microinjector, after which they were stored for 1-7 d at 18°C in
standard ND-96 solution (in mM: 96 NaCl, 2 KCl, 1 MgCl2, 1.8 CaCl2, and 5 HEPES, pH 7.4) supplemented with 10 µg/ml gentamycin sulfate.
Electrophysiological measurements and analysis. Currents
were recorded with the two-electrode voltage-clamp technique (OC-726C; Warner, Hamden, CT) after 1-7 d incubation. All experiments were conducted at room temperature (21-23°C). Data were sampled at 1 kHz.
Electrode resistances were 0.5-1 M
when filled with 2 M
KCl. The bath solution contained (in mM): 95 NaCl, 5 KCl, 1 MgCl2, 0.3 CaCl2, and 5 HEPES,
adjusted to pH 7.4 with NaOH, unless otherwise noted. Data acquisition
and analysis were performed with pClamp 6.0 software (Axon Instruments,
Foster City, CA). Curve fitting in Clampfit was performed using the
Chebyshev method. Additional analysis and Boltzmann curve fits were
performed with Origin 4.0 software (Microcal, Amherst, MA). In some
cases, small corrections for leak were made using an off-line linear
leak subtraction protocol in Clampfit (based on current evoked at
voltage steps to 100 mV), but correction was often not necessary
because of negligible leak (<1% of total conductance). If leak
exceeded 10% of total conductance, the data were discarded. Sample
numbers (n) refer to the number of individual oocytes recorded.
The permeability ratio for Na+ and
K+ was determined from a simplified form of the
Goldman-Hodgkin-Katz equation, where the reversal potential
Erev = 58 log {[PNa
(Na)out + PK
(K)out]/[PNa (Na)in + PK (K)in]}, and we have assumed
no permeability to Cl
ions.
Erev for Melk2 was determined from the
experiments in Figure 5A; Kout and
Naout were 5 and 95 mM, respectively. Values
for Kin and Nain are 92.5 and 6.2 mM, respectively, as reported by Barish (1983).
| |
RESULTS |
Sequence analysis of Melk channels
A partially sequenced mouse brain cDNA encoding a polypeptide
segment with strong amino acid similarity with the
Drosophila Elk sequence was identified in a computer-based
search of a mouse dBEST database. As shown by the alignment in Figure
1, the predicted polypeptide encoded by
this cDNA shares significant similarity with other members of the Eag
family of potassium channel polypeptides across its entire length.
Because the highest degree of identity is shared with members of the
Elk subfamily, the sequence was named Melk (for mouse Elk). For the
core of the polypeptide, extending from the S1 domain through the
region with homology to previously characterized cyclic nucleotide
binding domains (cNBD-like domain), Melk exhibits 38% amino acid
identity with Meag, 45% identity with HERG, 48% identity with
Drosophila Elk, and 63% identity with a rat Elk (Relk1)
gene cloned from rat sciatic nerve (Titus and Ganetzky, unpublished
data) (Shi et al., 1998). Sequence comparisons among Melk, Relk and a
recently identified human Elk (Helk; Titus and Ganetzky, unpublished
data) indicate that there are at least two Elk subtypes in mammals. One
subtype is defined by Relk1, which is ~63% identical (S1 through the
cNBD-like domain) with either Melk or Helk. The mouse and human Elk
sequences share 97% identity and thus appear to belong to a second
subclass. Because we obtained the Relk sequence before that of Melk or
Helk, we refer to the mouse polypeptide characterized here as
Melk2.
View larger version (94K):
[in this window]
[in a new window]
|
Figure 1.
Amino acid sequence alignment of Melk2. Sequence
comparisons of Melk2, Drosophila Elk, Meag, and HERG
were made using Lasergene's Clustal method (DNAstar). Identical amino
acids are indicated by the black shading, and gaps in
the sequence are indicated by dashes. Amino acids
are numbered on the right. The
membrane-spanning regions (S1-S6), the pore
domain of the hydrophobic core and the region of homology with a
cyclic nucleotide binding domain (cNBD) are
indicated by lines above the sequence.
Asterisks indicate the residues mutated for experiments
described in Figure 10.
|
|
A multiple tissue Northern blot probed with a fragment from the C
terminus of Melk2 (bp 2410-2880) revealed a single transcript of 5 kb
expressed specifically in brain (Fig. 2).
A band of ~2 kb was detected in testis, but it has not yet been
established whether this represents a bona fide Melk2 mRNA isoform.
View larger version (101K):
[in this window]
[in a new window]
|
Figure 2.
Tissue expression of Melk2. A mouse multiple
tissue Northern blot (Clontech) was probed using a
32P-labeled C-terminal fragment of Melk2 corresponding to
bp 2410-2880 under high-stringency conditions. A single band of ~5
kb was observed specifically in brain.
|
|
Functional expression of Melk2 channels in
Xenopus oocytes
To characterize the functional properties of Melk2 channels, we
performed two-electrode voltage-clamp analysis of Melk2 currents expressed heterologously in Xenopus oocytes. With voltage
steps up to ~0 mV the currents are sustained, with little evidence of inactivation (Fig. 3A). At
more positive voltage steps, currents have a transient, early
inactivating component (see expanded view in Fig. 3A,
inset). The peak I-V relation measured within the first 40 msec is relatively linear (Fig. 3B, filled
symbols), whereas the steady-state I-V relation
inwardly rectifies because of inactivation at voltages positive to 20 mV, giving a region of negative slope conductance (Fig. 3B, open
symbols).
View larger version (25K):
[in this window]
[in a new window]
|
Figure 3.
Functional expression of Melk2 channels in
Xenopus oocytes. A, Family of Melk2
currents generated in response to voltage pulses from 80 to 70 mV
from a holding potential of 80 mV. Small inward tail currents are
visible after repolarization to 80 mV. Inset, Same
current as in A but on a shorter time scale.
B, I-V relations for Melk2 channels
measured at the peak current within the first 40 msec (closed
symbols) or the current at the end of the 1 sec voltage pulse
(open symbols). Both are normalized to the peak current
at 70 mV. Calibration: A, 2 µA, 250 msec;
inset, 2 µA, 25 msec. Points in
B are the mean ± SEM; n = 5.
|
|
Melk2 activation kinetics
To characterize the time course of Melk2 activation, we measured
the amplitude of the tail current evoked at progressively longer times
after a depolarizing pulse to obtain time-dependent increases in
conductance (Fig. 4A,
Table 1). This method, used for
measuring activation in HERG channels (Trudeau et al., 1995), ensures
that temporal overlap with the inactivation process does not confound
measurements of activation. The macroscopic activation time course thus
determined can be described by two exponentials with an early rapid
component (
= 3.71 ± 0.71 msec at 70 mV) followed by a slower
component ( = 36.0 ± 3.6 msec at 70 mV; n = 8). Activation kinetics are faster at more positive voltages (Fig.
4B, Table 1).
View larger version (25K):
[in this window]
[in a new window]
|
Figure 4.
Activation in Melk2 channels. A,
Melk2 activation kinetics measured with an envelope voltage paradigm in
which a series of pulses to 70 mV that increase in either 2 or 10 msec
durations were given from rest at 80 mV. The resulting tail current
at 100 mV was fit with the sum of two exponentials
(y = Afaste(t/ fast) + Aslowe(t/slow))
and extrapolated to the onset of the voltage pulse (t = 0 msec).
(B) These values were plotted versus time and
normalized to the peak tail current at 70. The same experiment in
A was performed at other voltages, as indicated in
B. The points in B represent the
activation kinetics and were best fit with the sum of two exponentials
[y = 1 - (Afaste(t/fast) + Aslowe(t/slow))];
the time constants from these fits are reported in Table 1.
Calibration: 1 µA, 50 msec.
|
|
A defining characteristic of activation in Drosophila and
mammalian Eag channels (Ludwig et al., 1994; Robertson et al., 1996; Terlau et al., 1996) is the presence of a Cole-Moore shift in which
activation becomes slower and more sigmoidal with successively more
hyperpolarizing prepulses (Young and Moore, 1981). To determine whether Melk2 channels exhibit a Cole-Moore shift, we varied the prepulse potential and measured the activation kinetics of currents evoked by a subsequent step to 70 mV using the tail current pulse protocol described above (compare Fig. 4). As shown in Figure 5A, Melk2 activation kinetics
were not measurably altered by prepulses over the range of 120 to
80 mV, in contrast to the effect of prepulse on kinetics of Meag
current activation (Fig. 5B). Like Melk2, HERG activation
kinetics are insensitive to prepulse voltage (Fig. 5C),
suggesting that the presence of the Cole-Moore shift may distinguish
channels in the Eag subfamily from those in the Elk or Erg
subfamilies.
View larger version (15K):
[in this window]
[in a new window]
|
Figure 5.
Effect of prepulse holding potential and external
Mg2+ on channel activation. A-C,
Melk2 (A) Meag (B), and
HERG (C) activating currents at 70 mV elicited
after a 500 msec prepulse holding potential of 80, 100, or 120 mV
as indicated. Activation was measured for Melk2 (A, D)
and HERG (C) by plotting the development of the
tail current at 100 mV after pulses to 70 mV of incremental duration
in steps of 10 msec, as in Figure 5. For Meag
(B), activation is simply represented by the rise
of the outward current. D, Melk2 activation measured in
the presence of 0, 1 (normal Ringers'), or 10 mM external
Mg2+ by the same method as in A and
Figure 4 from a holding potential of 80 mV. Mg2+
reduced Melk2 tail current amplitude (D, inset), but
scaling these currents indicates that the activation time course is not
changed by Mg2+. The peak tail current relative to
that in the absence of Mg2+ was 0.80 ± 0.02 and 0.58 ± 0.01 in 1 and 10 mM
Mg2+, respectively (mean ± SEM;
n = 4 for each point). Calibration:
A-D, 50 msec; A, D, inset, 250 nA;
C, B, 1 µA. The current amplitude calilbration in
A, C, and D is the absolute current value
for comparison with Meag. The results seen in A-D were
also observed for three additional oocytes.
|
|
Consistent with the absence of a Cole-Moore shift is the lack of
sensitivity of Melk2 activation kinetics to increases in external
Mg2+ concentration (Fig. 5D). In
mammalian Eag channels, elevated divalent cation concentrations cause
activation to become slower and more sigmoidal in a manner reminiscent
of the Cole-Moore shift caused by changes in the prepulse potential
(Terlau et al., 1996). Elevation of external Mg2+
concentration causes a reduction of the inward Melk2 tail current, as
represented by the points in Figure 5D, inset.
However, when these data sets are scaled they superimpose,
demonstrating that Melk2 activation kinetics are unaffected by changes
in external Mg2+ concentration (Fig. 5D).
Thus, Melk2 channels can be distinguished from channels in the Eag
subfamily by their lack of a Cole-Moore shift and the insensitivity of
their activation kinetics to external divalent cation concentration.
Melk2 deactivation kinetics
To measure Melk2 deactivation kinetics, the channels were first
activated with a voltage step to 70 mV followed by a series of steps
from 20 to 120 mV to close (deactivate) the channels (Fig.
6A). The deactivation
time course is reflected in the decay of these currents, which is
generally preceded by a faster, initial rising phase as channels
recover from inactivation via the open state. The decay phase was best
fit with the sum of two exponentials, with a fast deactivation
component ranging from 107.7 ± 22.6 msec at 20 mV to 38.5 ± 7.0 msec at 120 mV and a slow component ranging from 380 ± 127 msec at 20 mV to 223.2 ± 57 msec at 120 mV
(n = 10; Fig. 6B, Table
2). The fast component accounted for
60-80% of the fit over the range of voltages tested.
View larger version (16K):
[in this window]
[in a new window]
|
Figure 6.
Deactivation kinetics in Melk2 channels.
A, Deactivation kinetics were determined by first
activating (and inactivating) channels with a 1 sec pulse to 70 mV.
This was followed by a series of 3 sec pulses from 20 to 120 mV,
during which currents initially increased in amplitude as a result of
recovery from inactivation, followed by current decay attributable to
channel closure. These decaying currents were best fit with the sum of
two exponentials as in Figure 4A, and time
constants derived from these fits (fast and
slow) are plotted in B and listed
in Table 2. Calibration: 1 µA, 1 sec. Points in
B are mean ± SD; n = 10 for
each point.
|
|
Tail currents in Figure 6A were also used to
determine the permeability ratio for Na+ and
K+ ions in Melk2 channels. The reversal potential
determined from these experiments was 70.6 ± 0.4 mV
(n = 4), and solving the Goldman-Hodgkin-Katz
equation (see Materials and Methods) gave a permeability ratio
(PNa/PK) of
0.007, indicating that Melk2 channels allow one Na+
ion to permeate for every 150 K+ ions.
Melk2 inactivation kinetics
Inactivation was measured using a three-pulse protocol similar to
that used in the analysis of HERG channels to isolate inactivation from
the temporally overlapping activation process (Schonherr et al., 1996;
Smith et al., 1996; Spector et al., 1996b; Wang et al., 1996; Herzberg
et al., 1998). First, channels were maximally activated by a 1 sec
pulse to 70 mV and then allowed to recover from inactivation to the
open state by a brief (10 msec) step to 80 mV (Fig.
7A). Before significant
deactivation, the channels were driven from the open state back to the
inactivated state by a third voltage step ranging from 70 to 20 mV. The
currents evoked at this third pulse to 70 mV were twice as large as
those evoked by a single step to 70 mV (peak current shown by brackets) and decayed (inactivated) approximately three times faster (see expanded time scale to the right). Single exponential fits
to these traces had time constants ranging from 6.0 ± 0.68 msec
at 70 mV to 10.8 ± 1.15 msec at 20 mV (n = 4;
Fig. 7B, Table 3). The change
in time constant with voltage reflects an intrinsic voltage dependence
to the Melk2 inactivation process that is distinct from the voltage
dependence of activation.
View larger version (11K):
[in this window]
[in a new window]
|
Figure 7.
Melk2 inactivation kinetics. A,
Inactivation was determined with a three-pulse voltage paradigm.
Channels were maximally inactivated by a 1 sec pulse to 70 mV and then
allowed to recover from inactivation with a 10 msec pulse to 80 mV.
Before substantial channel deactivation a subsequent pulse to a voltage
ranging from 70 to 20 mV was given. The resulting inactivating currents
in A (and in the expanded view in A)
represent channels going from the open to the inactive state.
B, Time constants derived from a single exponential fit
[y = Ae(t/)] to these
currents are shown in Table 3. Calibration: 1 µA, 250 msec. A 1 msec
duration capacitance artifact was removed before the inactivating
currents in A for clarity. Points in
B are mean ± SD.
|
|
Like HERG channels, Melk2 channels recover from inactivation with a
time course much faster than that of deactivation and thus exhibit
relatively large tail currents during repolarization. Because the
recovery from inactivation is rapid, it is not always resolved in each
record, but in many cases it can be seen as a rising phase of the tail
current before deactivation (e.g., Fig. 6A). The time
course of recovery could be measured by stepping from 70 mV to a range
of voltages, as in Figure 6A; at each voltage, the
resulting current was fit with a single exponential function, with time
constants ranging from 16.5 ± 1.1 msec at 20 mV to 3.15 ± 0.2 msec at 80 mV (n = 8; Table
4). As will be discussed below, the rapid
recovery from inactivation together with the rapid activation confer
unique physiological properties to the Melk2 channel.
Steady-state properties of Melk2 channels
The steady-state activation properties of Melk2 channels were
determined from tail current measurements. Inward tail currents were
elicited by repolarizing steps to 100 mV after 3 sec test voltage
steps ranging from 100 to 60 mV (Fig.
8A). The deactivation decay of each tail current was fit with a double exponential function and back-extrapolated to the moment of the voltage change. The resulting values were normalized to the maximum extrapolated values obtained after the step to 60 mV and plotted as relative conductances (g/gmax) as a function
of the preceding voltage step. The values obtained using this procedure
are thus an estimate of the conductances that would be measured from
the instantaneous peak tail currents in the absence of
inactivation, a method previously used for the analysis of HERG
currents (Trudeau et al., 1995; Wang et al., 1997). The g-V
relationship obtained in this manner has a slope factor (k)
of 28.34 ± 1.55 mV per e-fold change in conductance and
half-maximal voltage of activation (V1/2)
of 6.37 ± 1.34 mV (n = 8; Fig. 8C, filled
squares).
View larger version (28K):
[in this window]
[in a new window]
|
Figure 8.
Steady-state properties.
A, Tail current protocol used to generate steady-state
activation curve. Inward tail currents were elicited at 100 mV after
3 sec voltage commands from 100 to 70 mV from rest at 80 mV. Only
the last 10 msec of the 3 sec pulses are shown. The decay
phase was fit with the sum of two exponentials
[y = Afaste(t/fast) + Aslowe(t/slow)]
and extrapolated to the onset of the 100 mV pulse to minimize the
effects of recovery from inactivation (the rising phase of the tail
current). B, Protocol for generating steady-state
inactivation curve. Channels were first activated (and then
inactivated) by a 1 sec pulse to 70 mV (only 10 msec of which is shown
in B) and then were allowed to equilibrate between open
and inactive states during a conditioning pulse ranging from 70 to
100 mV. A second pulse to 70 mV was then given to determine the
proportion of channels in the open state from the instantaneous current
thus evoked. C, Steady-state activation curve
(filled squares) generated from values from
A normalized to the largest value, plotted versus
command voltage and fit with a single power Boltzmann function of the
form y = 1 {1/[1 + e(V V1/2)/k]}.
The half-maximal voltage of activation
(V1/2) is 6.37 ± 1.43 mV per
e-fold change in conductance with a slope factor
(k) of 28.43 ± 1.55 mV;
n = 8. The steady-state inactivation curve
(filled circles) was generated from normalized
instantaneous current values from B plotted as a
conductance versus the conditioning voltage and fit with a single power
Boltzmann function of the form y = 1/[1 + e(V V1/2)/k],
resulting in a V1/2 of
2.45 ± 1.01 mV and k = 21.54 ± 1.01 mV; n = 3. Channel deactivation also affects the
peak taken at the arrow, especially during negative
conditioning pulses from 100 to approximately 50 mV. This was
corrected for by estimating the relative amount of closure during the
40 msec conditioning pulse based on the deactivation time constant,
where % deactivation = 1 (1/e)(40
msec/fast deactivation) (%Afast
deactivation) and adding this correction to the availability
curve. A 1 msec capacitance artifact was removed at the beginning and
end of the conditioning pulse. Open circles represent
steady-state conductance calculated by multiplying the activation and
inactivation curves on the same plot. The error in the calculated
conductance is the sum of the errors in the steady-state curves.
D, The calculated I-V relation
(open symbols) was determined from the conductance in
C, adjusted for the effects of driving force in accord
with Ohm's law, where Icalculated = (gcalculated)(Vm Erev), and scaled to the largest
value (Vm = membrane potential, and
Erev = reversal potential). The experimental
I-V relation is the same as that from Figure
3B, except that it is scaled to its largest value at the
end of the voltage pulse, instead of the beginning of the pulse, for
comparison of nonlinearity.
|
|
The Melk2 steady-state inactivation relation was determined using the
voltage protocol shown in Figure 8B, based on a
similar protocol used to measure steady-state inactivation in HERG
channels (Smith et al., 1996). A 1 sec pulse to 70 mV was used to
maximally activate and then inactivate channels. Conditioning pulses of 40 msec ranging from 70 to 100 mV were subsequently presented to
allow channels to equilibrate between inactivated and open states (see
Fig. 8 legend for the correction procedure for channels entering the
closed state at more negative voltages). The proportion of channels
available to conduct at the end of each conditioning pulse was
determined by measuring the instantaneous current values by a
subsequent step back to 70 mV, as indicated by Figure 8B, large arrow. These values were normalized to the maximum
instantaneous current value obtained and plotted as a conductance
(g/gmax) versus voltage
(Fig. 8C, filled circles). The resulting curve represents the steady-state inactivation relation, which when fit with a Boltzmann
function has a V1/2 of 2.45 ± 1.01 mV and a slope factor of 21.54 ± 1.01 mV per e-fold change
in conductance (n = 3).
The nonlinear I-V relation shown in Figure 3B
should be predicted by the steady-state activation and inactivation
relations if we have correctly identified and characterized the
predominant gating processes. To test this, we multiplied the
steady-state activation and inactivation relations. Their product
yields a bell-shaped steady-state conductance-voltage curve (Fig.
8C, open symbols). An I-V relationship
calculated from this steady-state conductance-voltage relation closely
matches the experimental I-V relationship determined at 1 sec (Fig. 8D; data from Fig. 3B). This
agreement between the predicted and observed results suggests that
inactivation causes rectification of the Melk I-V relationship, as in HERG channels (Smith et al., 1996), and that we
have accounted for the primary determinants of the steady-state Melk2
currents in our analysis.
C-type inactivation in Melk channels
Rapid inactivation causing inward rectification in HERG channels
is attributable to a C-type inactivation mechanism (Schonherr and
Heinemann, 1996; Smith et al., 1996; Herzberg et al., 1998). To test
the possibility that a similar mechanism was responsible for the
inactivation seen in Melk2, we first determined whether external
TEA+ slows inactivation as it does for Shaker (Choi
et al., 1991;) and HERG channels (Smith et al., 1996). We found that
TEA+ slows inactivation (Fig.
9A), with a
K1/2 of 97 ± 16 mM, as
determined from a plot of inactivation rate (1/) versus log [TEA+] (Fig. 9B). Although the
sensitivity to TEA+ is less for Melk2 currents than
for Shaker (K1/2 = 33 mM; Choi et al., 1991) or HERG (K1/2 = 20 mM; Smith et al., 1996), this result supports the
hypothesis of a C-type inactivation mechanism in Melk.
View larger version (16K):
[in this window]
[in a new window]
|
Figure 9.
TEA slows Melk2 inactivation. A,
Inactivating currents after a second step to 70 mV (as in Fig.
7A) in the presence of increasing amounts of external
TEA+ as indicated. Currents were normalized to their
peaks to better compare the effects of TEA+. The
bottom axis represents the time after the second pulse
to 70 mV. B, The inverse of the time constant of
inactivation (1/) plotted versus [TEA+] was fit
with a sigmoidal function of the form y = 1/[1 + [TEA+]/K1/2)],
which yields a half-maximal concentration
(K1/2) for slowing inactivation of
97.09 ± 16.11 mM TEA+;
n = 3 for 200 mM
TEA+; n = 6 for all other
concentrations of TEA+ (mean ± SD).
|
|
As a second test of this hypothesis we determined the effects of two
point mutations in the P region predicted to impair C-type inactivation. One of these mutations, S475A, is at a site equivalent to
a residue involved in C-type inactivation in Shaker (Lopez-Barneo et
al., 1993) and in HERG (Schonherr and Heinemann 1996; Smith et al.,
1996; Ficker et al., 1998; Herzberg et al., 1998; Fig. 1,
asterisk). The second mutation, S464T, is equivalent to a
mutation in HERG channels that eliminates C-type inactivation
(Suessbrich et al., 1997; Ficker et al., 1988; Herzberg et al., 1998).
To characterize inactivation in these channels, the current measured at
the end of the voltage pulse (I1 sec) was
normalized to the instantaneous peak current in the absence of
inactivation determined from the three-pulse protocol
(Imax) and plotted as a function of
voltage in each case (Fig.
10A-D). The relative
inactivation rates of the mutants were obtained from the relaxation
from Imax, which is replotted on a faster time
scale in Figure 10E. Compared with wild-type
channels, the S475A mutation produces less inward rectification and
slower inactivation, whereas S464T eliminates rectification as well as
inactivation. These effects parallel those of the equivalent mutations
in HERG channels (Suessbrich et al., 1997; Ficker et al., 1998;
Herzberg et al., 1998) and in Shaker channels (Lopez-Barneo et al.,
1993), providing further support for the hypothesis that a C-type
inactivation mechanism is responsible for the inactivation of Melk2
channels.
View larger version (26K):
[in this window]
[in a new window]
|
Figure 10.
C-type inactivation in Melk2 channels.
A-C, Current families of wild-type Melk2 and point
mutants Melk2 S475A and Melk2 S464T, respectively, elicited by voltage
steps from 70 to 70 mV. The final pulse to 70 mV was followed by
second pulse to 70 mV after a 10 msec pulse to 80 mV to estimate the
magnitude of the current in the absence of inactivation and the
inactivation rate. D, Normalized I-V
relations for each construct, as indicated, were compared by
normalizing the current level at the end of the 1 sec pulse
(arrows, I1 sec) to
the current in the absence of inactivation at the peak of the second
pulse to 70 mV (arrows, Imax);
n = 5 for each construct. E, Direct
comparison of the inactivating currents from the second pulse to 70 mV
from traces in A-C, normalized to the peak current for
each trace. The bottom axis is the time in milliseconds
after the second pulse to 70 mV. Calibration: A-C, 1 µA, 250 msec.
|
|
Comparison of currents from Eag family members
To further distinguish Melk2 channels from other members of the
Eag family, we compared the rectification and underlying inactivation properties of Meag, HERG, and Melk, representatives of each Eag subfamily (Fig. 11). Scaling the
currents evoked by the three-pulse protocol (as described for Fig. 10)
to the maximum instantaneous current obtained during the third pulse
(Imax) provides a visual impression of
the differences in rectification among the three channel types (compare
the current amplitudes at I1 sec and
Imax; Fig. 11A-C).
Comparison of the steady-state I-V relations reflects these
differences and shows that Melk is intermediate to Meag, which has
little or no inward rectification, and HERG, which is strongly
rectifying (Fig. 11D). Correspondingly, the
inactivation rate of Melk2 is intermediate to that of Meag, which is
noninactivating, and HERG, which inactivates rapidly (Fig.
11E). Thus, the rate and extent of C-type
inactivation is a distinguishing feature for different subfamilies of
potassium channels within the Eag family.
View larger version (23K):
[in this window]
[in a new window]
|
Figure 11.
C-type inactivation in the Eag channel family.
A-C, Families of currents from Meag, Melk2, and HERG,
respectively. The voltage paradigm (bottom) is identical
to the one in Figure 10. D, The current-voltage
relation for each channel was determined at the end of a 1 sec voltage
pulse (arrows, I1 sec) and normalized
to the current at the peak of the second pulse to 70 mV (arrows,
Imax); n = 5 for Melk2;
n = 7 for Meag; n = 9 for HERG.
E, Comparison of inactivation among Meag, Melk2, and
HERG, as indicated. Inactivating currents from the second pulse to 70 mV are normalized for comparison; the time scale represents the time
after the second pulse to 70 mV. Calibration: A-C, 1 µA, 250 msec.
|
|
Sensitivity to E-4031
A defining characteristic of channels in the Erg subfamily is a
nanomolar sensitivity to class III antiarrhythmic drugs such as
dofetilide and E-4031 (Trudeau et al., 1995; Snyders and Chaudhary, 1996; Spector et al., 1996a; London et al., 1997; Shi et al., 1997;
Zhou et al., 1998). In contrast, Meag (Herzberg et al., 1998) and
bovine Eag (Ficker et al., 1998) channels have 100-fold less
sensitivity to these drugs. We determined the effects of E-4031 on Melk
channels and found that Melk currents elicited by a pulse to 20 mV
after equilibration at 5 mV in 5 µM E-4031 for 10 min
were no different from currents in the absence of drug (data not
shown). This is a somewhat surprising result, because HERG and Melk
channels share all the residues in the P loop that have previously been
identified as determinants of drug binding in HERG (Ficker et al.,
1998).
Physiological roles of Melk2 channels
To begin to address the physiological roles played by Melk2
channels in the brain, we presented a ramp voltage-clamp command to 60 mV with a rise time of 3 msec and repolarizing ramps of varying
durations (Fig. 12). A sharp peak of
outward current is evoked by the initial depolarizing ramp as channels
activate and subsequently inactivate, indicating that Melk2 could
potentially contribute to the repolarization of a single action
potential. As the duration of the repolarizing ramp increases, a second
peak of current emerges as a result of the recovery from inactivation, suggesting that during a long burst of action potentials Melk2 currents
might contribute to the timing of burst duration or to spike frequency
adaptation. Thus, their particular gating properties render Melk2
channels well suited to a dual role in membrane excitability
on a
short (millisecond) time scale to the control of membrane potential and
on a longer (hundreds of milliseconds) time scale to the regulation of
firing frequency or burst duration.
View larger version (14K):
[in this window]
[in a new window]
|
Figure 12.
Melk2 currents in response to voltage ramps.
Voltage was changed from rest (80 mV) to 60 mV with a 3 msec ramp for
all traces and then returned to rest with a ramp of the following
durations: 100, 200, 400, 800, 1200, and 1600 msec. The same results
were obtained in three additional experiments. Calibration: 2 µA, 250 msec.
|
|
| |
DISCUSSION |
The novel properties of Melk2 channels described in this study
expand our understanding of the structural and functional diversity within the Eag family and provide physiological and pharmacological criteria with which the corresponding currents in vivo may
be distinguished. Melk2 mRNA is highly expressed in brain
tissue and not in heart, spleen, lung, liver, skeletal muscle, or
kidney; a smaller band appears in testis, but it is not yet clear
whether this represents a true Melk2 transcript. In
contrast, members of the Erg family are broadly distributed in tissues
such as the brain, heart, skeletal muscle, smooth muscle, and testis
(Curran et al., 1995; London et al., 1997; Shi et al., 1997; Wymore et al., 1997), and Eag members are found in the brain, skeletal muscle, and the retina (Ludwig et al., 1994; Frings et al., 1998).
A distinctive feature of Melk2 currents is a large, early inactivating
component and a corresponding inward rectification that is intermediate
in degree to the strongly rectifying HERG and the noninactivating Meag.
Inactivation occurs by a voltage-dependent, C-type inactivation
mechanism such as that mediating inward rectification in HERG channels.
Melk2 currents lack the Cole-Moore shift and sensitivity of activation
kinetics to Mg2+ observed in Eag subfamily members
as well as the E-4031 sensitivity characteristic of HERG.
The inactivation mechanism in Melk2 appears to be highly conserved with
the C-type inactivation mechanism described for HERG channels. For
example, the mutation S464T in Melk essentially removes inactivation,
as does the equivalent mutation S620T in HERG, whereas Melk2 S475A
slows inactivation like S631A in HERG (Suessbrich et al., 1997;
Herzberg et al., 1998). This suggests a similar dependence on the side
chains of the amino acid residues present at these two sites in the P
region for control of inactivation in Melk2 and HERG channels. The
essentially null inactivation phenotype of Melk2 S464T also suggests
that additional mechanisms such as N-type inactivation do not
contribute to the inactivation of Melk2 channels.
Despite these similarities, Melk2 inactivation is approximately
threefold to fourfold slower, and the
V1/2 of the steady-state inactivation
relation is shifted by ~80 mV compared with the corresponding properties of HERG (cf. Smith et al., 1996). These differences may
arise from other residues involved in C-type inactivation that are not
conserved between Melk2 and HERG. For example, the rate of C-type
inactivation appears to be proportional to the volume of the amino acid
side chain at equivalent sites in S6 in ShakerB (463; Hoshi et al.,
1991) and HERG (644; Herzberg et al., 1998). In each case an alanine,
with the smaller volume, is associated with slow inactivation, whereas
when the larger valine is present, inactivation is fast. In Melk2
channels the inactivation rate is intermediate, as predicted by the
intermediate volume of the threonine at the corresponding position
(T488) and consistent with the hypothesis that this site may contribute
to the differences in C-type inactivation between HERG and Melk2 channels.
Melk2 activation properties are distinct from those in either the Eag
or Erg channel subfamilies. Macroscopic Melk2 activation is well
described by the sum of two exponentials with time constants similar to
those of rat Eag channels (Terlau et al., 1996); the V1/2 and the slope of Melk2
g-V relation are similar to those of Meag
channels (Robertson et al., 1996). These properties distinguish Melk2
from HERG, which has very slow and sigmoidal activation kinetics and a
g-V relation that is twofold to threefold steeper and
shifted by approximately 25 mV (Sanguinetti et al., 1995; Trudeau et
al., 1995; Wang et al., 1997). However, unlike Eag, but similar to
HERG, Melk2 activation does not exhibit a Cole-Moore shift, which is
attributed to residency in a remote closed state during a
hyperpolarizing prepulse; during a subsequent depolarization, the
transition from that closed state is rate limiting (Young and Moore
1981; Bezanilla et al., 1994). Melk2 channels correspondingly lack a
sensitivity to external divalent cations, which, like hyperpolarizing prepulses, slow activation in Eag channels (Terlau et al., 1996). This
property has recently been used as a criterion to link Eag-type channels with IKx, a current involved in
sensory transduction in the retina (Frings et al., 1998). It is
interesting that the voltage dependence of activation in Melk2 is
similar to that of Meag despite the presence of a proline residue in S4
(P347) instead of the conserved arginine residue present in Eag and
HERG channels. Whether the presence or absence of this charged residue
in S4 contributes to differences in voltage sensitivity among the Eag family members remains to be determined.
Melk deactivation gating is similar to that of certain isoforms of
Merg1 channels that have shortened or novel N-terminal domains
(Lees-Miller et al., 1997; London et al., 1997) and HERG channels with
engineered deletions of the N terminus (Schonherr and Heinemann 1996;
Spector et al., 1996b; Wang et al., 1998). In wild-type HERG and Merg1a
channels (both of which have long N termini) the N terminus modulates
deactivation kinetics, but in the absence of this domain, deactivation
kinetics are ~100-fold faster. The S4-S5 linker is also critically
involved in this process and may form a receptor site for the
interaction of the N terminus with the pore region (Wang et al., 1998).
The relatively rapid deactivation kinetics in Melk2 channels suggest
that such modulation may not occur in these channels; perhaps the
shorter N terminus of Melk2 or amino acid differences in the S4-S5
linker fail to support the modulatory mechanism. Further experiments
will be required to test these hypotheses.
One of the hallmark features of the Erg family of channels is their
sensitivity to inhibition by class III antiarrhythmic drugs (Trudeau et
al., 1995; Snyders and Chaudhary 1996; Spector et al., 1996a; London et
al., 1997; Shi et al., 1997; Zhou et al., 1998). Efforts at localizing
drug binding sites reveal residues in the P loop and part of the S6
domain of HERG necessary for drug block, including S620 and S631
(Ficker et al., 1998). It is interesting that, despite the presence of
identical residues at the equivalent positions (S464 and S475
respectively), Melk channels are insensitive to E-4031. Thus other, as
yet unidentified residues must also be involved in drug binding in
HERG, or alternatively, other Melk residues may disrupt the ability of
drug to bind these channels.
A recent report of the cloning and expression of elk1 from
rat (Relk1; Shi et al., 1998) reveals heterogeneity within
the Elk subfamily. In contrast to Melk2, which is expressed
robustly in brain, mRNA levels of Relk1 were barely
detectable in brain but were much higher in sympathetic ganglia and
sciatic nerve (Shi et al., 1998). The amino acid sequences of the
corresponding polypeptides are ~68% identical and form channels with
distinct biophysical properties. For example, Melk2 exhibits relatively rapid activation and inactivation, whereas Relk1 is slow to activate and shows no apparent inactivation (cf. Shi et al., 1998).
The pronounced differences in gating among channels in the Eag family
likely reflect their distinct physiological roles. HERG currents peak
not during activation but rather as channels recover from inactivation
before returning to the closed or resting state (Sanguinetti et al.,
1995; Trudeau et al., 1995). Thus, HERG is specialized to contribute a
large repolarizing current at the terminal repolarization phase of the
cardiac action potential (Zhou et al., 1998), as previously
demonstrated for the native IKr (Zeng et al.,
1995). In contrast, the faster activation and slower inactivation of
Melk2 channels gives rise to early transient peaks, which occur on a
time scale that may enable them to contribute a repolarizing current
during a neuronal action potential or short spike train. Like HERG,
however, Melk2 recovers from inactivation more rapidly than it
deactivates and thus may contribute a second outward, repolarizing
current near the end of a sustained burst, especially in conjunction
with other K+ currents that initiate the
repolarization process. In neurons with action potentials that have
late plateau phases (cf. McCormick et al., 1997), this second outward
component could be a significant factor in neuronal processing.
| |
FOOTNOTES |
Received Oct. 2, 1998; revised Jan. 29, 1999; accepted Feb. 1, 1999.
This work was supported by National Institutes of Health Grant HL55973
and a National Science Foundation Career Award to G.A.R., a predoctoral
fellowship from the American Heart Association-Wisconsin to M.C.T., and
National Institutes of Health Grant NS15390 to B.G. The Melk2 sequence
has been deposited in GenBank with accession number AF109143. We thank
Cena Meyers for technical assistance, Elon Roti Roti and Nathan Penn
for frog surgery and oocyte preparation, and Eisai (Tokodai, Japan) for
providing E-4031. Thanks to Jinling Wang and Dr. Ed Chapman for helpful
discussions and Jinling Wang for critically reading this manuscript.
Correspondence should be addressed to Dr. Gail Robertson, Department of
Physiology, University of Wisconsin-Madison Medical School, 1300 University Avenue, Madison, WI 53706.
Dr. Trudeau's present address: Department of Physiology and
Biophysics, Howard Hughes Medical Institute, University of Washington, Box 357370, Seattle, WA 98195.
| |
REFERENCES |
-
Barish ME
(1983)
A transient calcium-dependent chloride current in the immature Xenopus oocyte.
J Physiol (Lond)
342:309-325[Abstract/Free Full Text].
-
Bezanilla F,
Perozo E,
Stefani E
(1994)
Gating of Shaker K+ channels: II. The components of gating currents and a model of channel activation.
Biophys J
66:1011-1021[Abstract/Free Full Text].
-
Choi KL,
Aldrich RW,
Yellen G
(1991)
Tetraethylammonium blockade distinguishes two inactivation mechanisms in voltage-activated K+ channels.
Proc Natl Acad Sci USA
88:5092-5095[Abstract/Free Full Text].
-
Cole KS,
Moore JW
(1960)
Ionic current measurements in the squid giant axon membrane.
J Gen Physiol
44:123-167[Abstract/Free Full Text].
-
Curran ME,
Splawski I,
Timothy KW,
Vincent GM,
Green ED,
Keating MT
(1995)
A molecular basis for cardiac arrhythmia: HERG mutations cause long QT syndrome.
Cell
80:795-803[ISI][Medline].
-
Drysdale R,
Warmke J,
Kreber R,
Ganetzky B
(1991)
Molecular characterization of eag: a gene affecting potassium channels in Drosophila melanogaster.
Genetics
127:497-505[Abstract].
-
Ficker E,
Jarolimek W,
Kiehn J,
Baumann A,
Brown AM
(1998)
Molecular determinants of dofetilide block of HERG K+ channels.
Circ Res
82:386-395[Abstract/Free Full Text].
-
Frings S,
Brull N,
Dzeja C,
Angele A,
Hagen V,
Kaupp UB,
Baumann A
(1998)
Characterization of ether-à-go-go channels present in photoreceptors reveals similarity to IKx, a K+ current in rod inner segments.
J Gen Physiol
111:583-599[Abstract/Free Full Text].
-
Ganetzky B,
Wu CF
(1983)
Neurogenetic analysis of potassium currents in Drosophila: synergistic effects on neuromuscular transmission in double mutants.
J Neurogenet
1:17-28[Medline].
-
Herzberg IM,
Trudeau MC,
Robertson GA
(1998)
Transfer of rapid inactivation and E-4031 sensitivity from HERG to M-EAG channels.
J Physiol (Lond)
511:3-14[Abstract/Free Full Text].
-
Higuchi R
(1990)
In: PCR protocols: a guide to methods and applications (Innis MA, Gelfand DH, Sninsky JJ, White TJ, eds), pp 177-183. New York: Academic.
-
Hoshi T,
Zagotta WN,
Aldrich RW
(1991)
Two types of inactivation in Shaker K+ channels: effects of alterations in the carboxy-terminal region.
Neuron
7:547-556[ISI][Medline].
-
Lees-Miller JP,
Kondo C,
Wang L,
Duff HJ
(1997)
Electrophysiological characterization of an alternatively processed ERG K+ channel in mouse and human hearts.
Circ Res
81:719-726[Abstract/Free Full Text].
-
Liman ER,
Tytgat J,
Hess P
(1992)
Subunit stoichiometry of a mammalian K+ channel determined by construction of multimeric cDNAs.
Neuron
9:861-871[ISI][Medline].
-
London B,
Trudeau MC,
Newton KP,
Beyer AK,
Copeland NG,
Gilbert DJ,
Jenkins NA,
Satler CA,
Robertson GA
(1997)
Two isoforms of the mouse ether-a-go-go-related gene coassemble to form channels with properties similar to the rapidly activating component of the cardiac delayed rectifier K+ current.
Circ Res
81:870-878[Abstract/Free Full Text].
-
Lopez-Barneo J,
Hoshi T,
Heinemann SH,
Aldrich RW
(1993)
Effects of external cations and mutations in the pore region on C-type inactivation of Shaker potassium channels.
Receptors Channels
1:61-71[ISI][Medline].
-
Ludwig J,
Terlau H,
Wunder F,
Bruggemann A,
Pardo LA,
Marquardt A,
Stuhmer W,
Pongs O
(1994)
Functional expression of a rat homologue of the voltage gated either a go-go potassium channel reveals differences in selectivity and activation kinetics between the Drosophila channel and its mammalian counterpart.
EMBO J
13:4451-4458[ISI][Medline].
-
McCormick DA,
Bal T
(1997)
Sleep and arousal: thalamocortical mechanisms.
Annu Rev Neurosci
20:185-215[ISI][Medline].
-
Pajor AM,
Wright EM
(1992)
Cloning and functional expression of a mammalian Na+/nucleoside cotransporter. A member of the SGLT family.
J Biol Chem
267:3557-3560[Abstract/Free Full Text].
-
<
TOP
ABSTRACT
INTRODUCTION
