Institute of Veterinary Pharmacology, University of Bern, Bern,
Switzerland (H.D., M.M., C.T.G., G.S.); and Institute of Physiology,
University of Zurich, Zurich, Switzerland (F.J.P.K., C.K., N.G.G.)
 |
Introduction |
Voltage-gated
sodium channels are large integral membrane glycoproteins. They consist
of a major glycosylated 
-subunit (260-295 kDa) and up to three
small auxiliary 
-subunits of which the 3 has been identified
recently (Morgan et al., 2000
). The major -subunit is a large
protein with four internal homologous domains (DI-DIV), each
containing multiple potential -helical transmembrane segments
(S1-S6). The positively charged S4 segments in each domain serve as
voltage sensors and their movement serves as link to the
voltage-dependent gating. All isoforms of voltage-gated sodium channels, with the exception of dog nodose ganglion sodium channel (Chen et al., 1997) show a high similarity, which varies between 60 and
80%. Numerous isoforms of sodium channels were identified (for review,
see Denac et al., 2000; Goldin, 2001), and their common function in the
rising phase of the action potential in most excitable cells is well
recognized. Most recently, the first data on the three-dimensional
structure of the entire sodium channel protein were published (Sato et
al., 2001). However, compounds acting on sodium channels will remain a
tool for probing the structure of these large proteins until a higher
resolution can be obtained by NMR or X-ray crystallography.
Sodium channels display differences in tissue localization, kinetics,
or sensitivity to various toxins. According to their particular
properties, Catterall (1980) defined five groups of toxins with
corresponding binding sites. Sodium current enhancers, typified by
(±)-4-[3-(4-diphenylmethyl-1-piperazinyl)-2-hydroxy propoxy]-1H-indole-2-carbonitrile (DPI 201-106) compose a
group of compounds with an yet undefined binding site. DPI 201-106 was the first synthetic compound to prolong the open time of sodium channels by retardation or removal of inactivation with a resulting positive inotropic effect in a cAMP-independent manner (Scholtysik et
al., 1985; Salzmann et al., 1986; Gerard et al., 1989).
The present study sought to gain further insight into the mode of
action of the synthetic sodium channel modulators by associating the
structural variations of the channel proteins to the aberrant drug
response. Previous electrophysiological studies conducted with
multicellular myocardial preparations from different species suggested
that the bovine heart sodium channel differs from known cardiac sodium
channels by possibly lacking the binding site for the respective
synthetic modulators (Mevissen et al., 2001). Such naturally occurring
differences were already successfully used for the detailed
characterization of the tetrodotoxin (TTX) binding site (Fozzard and
Lipkind, 1996) and enabled a better understanding of the sodium channel
structure. We have cloned the full-length cDNA encoding the sodium
channel from bovine myocardium and named it bovine heart sodium channel
(bH1). bH1 was successfully expressed in Xenopus laevis
oocytes at high expression, and the inactivation of the sodium channels
could be examined by analyzing isolated sodium currents.
Concentration-response relationship was investigated using TTX, ATX-II,
and the purine derivate
(
)-(S)-6-amino--[(4-diphenylmethyl-1-piperazinyl)-methyl]-9H-purine-9-ethanol (SDZ 211-939) (Scholtysik et al., 1993), which shows the same electrophysiological properties as DPI 201-106 but is more potent and
more soluble (Mevissen et al., 2001). The results presented herein may
form the basis of a future analysis of the binding site and kinetics of
synthetic sodium current modifiers.
| |
Materials and Methods |
RNA Isolation.
Tissue samples were obtained from a
6-month-old calf and from 4-week-old Wistar rats, and total RNA was
isolated using TRIzol reagent (Invitrogen, Carlsbad, CA).
Northern Blot Analysis.
Northern blot analysis was performed
on 20 µg of each total RNA sample. Hybridization was carried out with
digoxigenin-labeled cRNA probes, and chemiluminescent detection was
performed using anti-digoxigenin-AP antibodies and CDP-Star according
to the manufacturer's instructions (Roche Applied Science, Basel,
Switzerland). The cRNA probes used were specific for the rat heart
sodium channel -subunit (rH1; GenBank accession no. M27902,
nucleotides 3132-3639) and for the rat brain sodium channel
2-subunit (rBIIa; GenBank accession no. X03639, nucleotides
3329-3690).
Molecular Cloning.
The general strategy pursued to obtain a
cDNA encoding the complete bovine cardiac sodium channel coding
sequence is described in more detail below. Bovine heart total RNA was
reverse transcribed with Expand Reverse Transcriptase (Roche Applied
Science) and oligo(dT)18 primer. The central part
of the bovine heart sodium channel cDNA (nucleotides 726-6075) was
subsequently amplified using Expand Long Template PCR System (Roche
Applied Science) and rH1-specific primers M27902:6131
(5'-TCGTGACGCTGTCGTAGGAC-3') and M27902:810
(5'-GCATACACAACTGAATTTGTGGA-3'). To exclude possible PCR-introduced
mutations, two independent PCRs were set. Both amplification products
were cloned into the pCR-XL-TOPO vector (Invitrogen, Groningen, The
Netherlands), resulting in pCRXL-bHM and pCRXL-bHMa plasmids, and both
inserts were sequenced.
The cDNA ends were obtained by a rapid amplification of cDNA ends
(RACE) procedure (Frohman et al., 1988) using bovine heart polyA+ RNA (CLONTECH, Palo Alto, CA) as template.
For the 5'-RACE, the Marathon cDNA Amplification kit (CLONTECH) was
used according to the manufacturer's recommendations. The gene
specific (GSP) oligonucleotides were designed based on the consensus
sequence of pCRXL-bHM and pCRXL-bHMa. The primers used in 5'-RACE were
GSP9BOV (5'-AAGACGCTGAGGCAGAAGACCGTGAG-3') and AP1 (supplied in the
kit) and the nested primers GSP10BOV (5'-ATCAGGGCCCCCACGATGGTCTTCAG-3')
and AP2 (supplied in the kit).
The 3'-RACE product was obtained using the 5'/3'-RACE kit (Roche
Applied Science). The reverse transcription reaction was performed
according to the manufacturer's recommendations using the supplied
Oligo(dT)-anchor primer. Hot-start PCR was performed using the PCR
anchor primer (supplied in kit) and GSP8BOV
(5'-CCACCTACATCATCATCTCCTTCCTCA-3'). PCR products from both RACE
reactions were cloned into the pCR-II-TOPO vector (Invitrogen) and sequenced.
Based on the complete sequence of the bovine cardiac sodium channel
-subunit gained from the overlapping clones, additional primers were
synthesized: BH103 (sense,
5'-GCAGGATGAGAAGATGGCAGCCTTCC-3') and BH6006 (antisense,
5'-GGGCTGCGCTCACACGATTGACTC-3'), including the translation
initiation and stop codons, respectively (underlined). These primers
were used to obtain the complete bovine sodium heart channel coding
sequence as a single DNA fragment in a long-template PCR starting from
reverse transcribed bovine heart polyA+ RNA as
template. The amplification product was gel purified and cloned into
the pGEMT-Easy plasmid (Promega, Madison, WI) to obtain pGEM-bH6.6, and
its nucleotide sequence was confirmed by sequencing.
DNA Sequencing and Analysis.
All cloned fragments were
sequenced on both strands on a Li-Cor sequencer using fluorescent
dye-labeled primers and a commercial kit (Amersham Biosciences UK,
Little Chalfont, Buckinghamshire, UK). Raw sequencing data were edited
and assembled in contigs with the Vector NTI Suite software (Informax,
North Bethesda, MD). Consensus pattern for phosphorylation by
protein kinases A and C were analyzed with ScanProsite
(http://www.expasy.ch/tools/scanprosite/).
Expression of bH1 and Rat-Nav1.5 and
Electrophysiological Recordings.
The coding sequence of bH1, which
is the bovine-Nav1.5, was subsequently subcloned
from pGEM-bH6.6 into the pBSTA expression vector (Shih et al., 1998) to
yield pBSTA/bH1. For proper linearization a XbaI restriction
site was introduced immediately downstream of the poly(A) tract by
site-directed mutagenesis (QuickChange; Stratagene, La Jolla, CA).
Before expression in X. laevis oocytes all vectors were
prepared as described previously (Kühn and Greeff, 1999). Two
clones with cardiac sodium channels were used: the newly cloned
pBSTA/bH1 and the pSP64T/rSkM2 clone (provided by Dr. R. Kallen,
University of Pennsylvania, Philadelphia, PA), coding for the rat heart
sodium channel isoform (Kallen et al., 1990; Sheng et al., 1994). It
should be noted that rSkM2 is similar to rH1 except for three
nucleotides that do not lead to changes in the amino acid sequence
(Kallen et al., 1990). Subsequently, the term
rat-Nav1.5 will be used.
Nav1.5 is the term for voltage-gated sodium
channel name for rSkM2, rH1, and hH1 (Goldin, 2001). In addition, one
clone encoding the rat brain isoform rat-Nav1.2 (former name rBIIA), also subcloned into pBSTA expression vector, was
used (Greeff and Kühn, 2000). Oocytes were microinjected with 20 to 40 ng of cRNA (50 nl) and maintained at 18 ± 1°C in modified
Barth's solution (MBS): 88 mM NaCl, 2.4 mM
NaHCO3, 1 mM KCl, 0.82 mM
MgSO4, 0.41 mM CaCl2, 0.33 mM Ca(NO3)2, and 10 mM
HEPES-CsOH, pH 7.5, supplemented with 25 U of penicillin, 25 µg/ml
streptomycin sulfate, and 50 µg/ml gentamycin sulfate. For the
reduction of ionic currents 15 µM TTX (Sigma/RBI, Natick, MA) was added.
Two-electrode voltage clamping was performed 3 to 6 days after cRNA
injection as described previously (Kühn and Greeff, 1999). Further details about compensation of series resistance and the software subtraction of the capacitance transient as done routinely in
the present experiments are given elsewhere (Greeff and Kühn, 2000). After this subtraction, normally only gating currents remain in
the direction of the voltage step. Sometimes at large pulse potentials
the subtraction was not perfect and subtraction artifacts remain (Figs.
4B and 5A). The oocytes were clamped at a holding potential of 100 mV
for at least 5 min to ensure recovery from slow inactivation before
recording started. The experiments were done in MBS at a temperature of
15 ± 1°C. ATX-II and SDZ 211-939 (Novartis, Basel, Switzerland)
were added as aqueous stock solutions directly to the bath and allowed
to equilibrate.
Statistics and Data Analysis.
Concentration-response curves
were calculated from the log concentration-effect curves using a Hill
equation and estimation via least-squares method. The underlying
equation for Hill function is response = Vm · C
· (C + K)
1, where
Vm is the maximal attainable response,
K is the half-effective concentration
(EC50, i.e., the concentration yielding half of the maximum effect), and the exponent describes the shape of the function (Hill coefficient). Statistical significance of any
comparisons made on the basis of this model (e.g., testing to see
whether the Hill coefficient equals 1) was made using a Wald Statistic. Confidence bounds presented for parameters in the Hill model were also
based upon the Wald Statistic (Wald, 1943; Portier et al., 1993). A
further graphical test for a 1:1 binding and Michaelis-Menten kinetics
was made for all experiments using the double reciprocal presentation
of effect versus concentration (Lineweaver-Burk plot; Fig. 5C, inset).
| |
Results |
Molecular Cloning and Sequence Analysis of Bovine Heart Sodium
Channel -Subunit.
Two amplification products of 5.3 kb,
representing the central part (nucleotides 726-6075) of the cDNA
encoding for the -subunit of the bovine heart sodium channel, were
obtained by two independent reverse transcription-PCRs and cloned to
obtain the plasmids pCRXL-bH5M and pCRXL-bH5Ma. They
contained the same sequence and were coding for a novel sodium channel
highly similar to channels of the type Nav1.5
(92% identity).
Using bovine heart polyA+ RNA as a template, the
5' and 3' ends of the cDNA were obtained by a RACE procedure. 5'-RACE
resulted in a product of 845 bp and included a 79-nucleotide-long
5'-untranslated region (UTR) as well as the beginning of the coding
region. 3'-RACE resulted in a product of 1151 bp and included the end
of the coding region as well as a 355-nucleotide-long 3'-UTR.
The alignment of PCR products from 5'- and 3'-RACE with the fragment
representing the central part followed. The analysis of overlapping
clones revealed a sequence of 6503 bp (GenBank/EMBL accession no.
AJ251721) with a predicted single large open reading frame potentially
encoding a 2022-amino acid protein (Fig. 1). The absence of the highly conserved
sequence AATAAA, present in higher eucaryotes, which is usually located
11 to 30 nucleotides upstream of the polyadenylation site, as well as
comparison with nucleotide sequences of other sodium channels such as
rat- and human-Nav1.5, suggest that we did not
reach the very 3' end of the bovine heart sodium channel cDNA.
View larger version (72K):
[in this window]
[in a new window]
|
Fig. 1.
Sequence comparison of cardiac voltage-gated sodium
channels. The amino acid sequence alignment of bH1 (bovine heart), hH1
(human heart), and rH1 (rat heart), which are all
Nav1.5-type channels (Goldin, 2001), is presented; only
differences compared with the bH1 sequence are displayed. Identical
residues are shown as dots and dashes indicate gaps. The putative
transmembrane segments (S1-S6) within each of the four domains
(DI-DIV) are shaded. Serine residues in the cytoplasmic loop between
DI and DII, shown to be phosphorylated by protein kinase C (Murphy et
al., 1996), are indicated by black triangles. The arginine
(R)-to-histidine (H) 529 substitution is boxed.
|
|
The complete bH1 coding sequence, stripped of the UTRs, was
reconstructed by reverse transcription-PCR with primers encompassing the bH1 translation initiation and stop codons. The 6.1-kb
amplification product was cloned into the pGEMT-Easy vector to yield
the pGEM-bH6.6 plasmid. The sequence around the initiation codon was
identical to two described sequences (Rogart et al., 1989; Gellens et
al., 1992; Akopian et al., 1996; Chen et al., 1997). The critical
purine and guanosine residues that influence most translation
initiation (Kozak, 1986) are present at positions 3 (adenosine) and
+4, respectively, relative to the predicted translation initiation codon. As observed in other species (Rogart et al., 1989; Gellens et
al., 1992; T. Zimmer and K. Benndorf, unpublished data; GenBank accession no. AJ271477), an upstream, out of frame ATG triplet in a
weaker context was also found in the bovine heart isoform cDNA at
positions 6 to 8.
The predicted amino acid sequence of bH1 was aligned with those of
other known heart sodium channel -subunits (Fig. 1). All the
relevant landmark amino acid sequences of sodium channels (such as the
positively charged residues of the putative voltage sensor S4 or the
IFM motif within the inactivation gate) are present in each of the four
domains identified. This confirms that the newly cloned bH1 is a
cardiac sodium channel. The homology on the amino acid level of the bH1
segments versus four other isoforms is shown in Table
1. The most variable segment was found in
the intracellular loop between domains II and III. The sequence in this
loop shared about 80% identity with the corresponding fragments of
human, rat, and mouse heart sodium channel.
Six and 19 canonical consensus phosphorylation sites for protein
kinases A ([RK](2)-x-[ST]) and C ([ST]-x-[RK]), respectively, were predicted by ScanProsite in cytoplasmic regions of the bovine sodium channel. Most of these sites are conserved in other species. Interestingly, a histidine residue at position 529 in the bovine heart
sequence is observed in place of an arginine residue conserved in the
intracellular loop between domains I and II of other species as marked
in Fig. 1. This amino acid substitution is located in the immediate
vicinity of the two serine residues that have been shown to be
selectively phosphorylated by protein kinase A and proposed to
participate to the cAMP-dependent regulation of cardiac sodium channel
activity (Murphy et al., 1996).
Tissue Distribution of Heart- and Brain-Specific Isoforms.
A
heart-specific cRNA probe recognized selectively the cardiac isoform
transcript as a single signal at approximately 9 kb (Fig.
2B). No signal was detected after
hybridization of this probe with total RNA from rat or bovine brain or
liver tissue. Northern blot analysis demonstrated that the newly cloned
bovine cDNA heart isoform is not detectable in the brain or liver of calf (Fig. 2A).
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 2.
Northern blot analysis. Total RNA samples (20 µg)
purified from rat (heart, lane 1; brain, lane 2; and liver, lane 3) or
bovine (heart, lane 4; brain, lane 5; and liver, lane 6) tissues were
used in a Northern blot experiment with digoxigenin-labeled cRNA probes
derived from the rat brain (A) or bovine heart (B) isoforms. Molecular
weight standards in kilobases are shown to the right of the blots.
|
|
The hybridization of the heart-specific cRNA probe did not show
variations in the quantity of expressed isoforms when different parts
of the heart such as left and right ventricles and atria were examined
(data not shown).
Electrophysiological Characterization of Expressed bH1
Channel.
Two-electrode voltage-clamp recordings confirmed that the
cRNA of bH1 produced normally functioning sodium channels with the typical properties known for heart muscle also in X. laevis oocytes when cloned in the high expression vector
pBSTA. Large currents up to about 30 µA were obtained (Fig.
3A), comparable in size to the currents
recorded in parallel from rat brain sodium channels (rBIIA) in the same
oocyte batch (data not shown). The higher expression level was desired
to investigate for small plateau currents in the presence of SDZ
211-939 (see below). With respect to the voltage dependence of bH1
and rBIIA, we compared currents of about equal size of the two clones
to minimize effects due to residual series-resistance errors. Typical
for heart sodium channels in comparison to brain sodium channels are
the left shift of the I/V curve by approximately 20 mV (Fig. 3C), the
low sensitivity to TTX (Fig. 3B), and the slower inactivation phase
(visible in the 0 µM ATX-II traces in Fig.
4, A and C). Note that the reversal potential for sodium may change in different experiments as found to be
caused by variations in cytoplasmic Na+
concentration (Greeff and Kühn, 2000); this as seen in Fig. 3C
would occur independently of a shift in the voltage dependence. Even in
the presence of 15 µM TTX, about 94% of the current was blocked and
6% of the current persisted, whereas rBIIA would be almost totally
blocked already at 2 µM, e.g., only 0.8% persistent sodium ionic
current (Fozzard and Hanck, 1996; Greeff and Kühn, 2000). A
further characteristic of heart sodium channels is a slower recovery
from fast inactivation compared with brain channels as shown in Fig.
3D.
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 3.
Sodium currents of bH1 and rBIIA channels at high
expression. A, voltage-clamp pulses from 60 to 80 mV in steps of 10 mV elicit maximal inward currents up to 32 µA at 30 mV in bH1. B,
after adding 15 µM TTX still 2 µA at 20 mV were recorded, pulses
40 to 40 in steps of 10 mV. C, comparison of I/V mean curves (from
experiments as in A) of bH1 ( , n = 5) with rBIIA
( , n = 6). Note comment in text about observed
shifts of reversal potential and the total curve. D, recovery from fast
inactivation determined for rat-rBIIA (rat-Nav1.2) (open
symbols) and bH1 (closed symbols) at different recovery potentials.
Test pulses to 0 mV were applied at increasing recovery intervals
(dt-rec.) after a conditioning, inactivating prepulse to 0 mV for 100 ms. The potentials during the recovery interval were either 80 mV
(  ), 100 mV ( ), or 120 mV (  ). Current
amplitudes shown are normalized to the current obtained without
inactivating prepulse. The corresponding recovery time constants are in
milliseconds (mean ± S.E.M.; n = 4 each): for
rBIIA (rat-Nav1.2) 31 ± 5.2 (80 mV), 9.2 ± 0.9 (100 mV), and 3.8 ± 0.5 (120 mV); and for bH1 234 ± 46 (80 mV), 90 ± 2.5 (100 mV), and 30 ± 0.8 (120
mV).
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 4.
Effect of ATX-II at bath concentrations of 0, 0.1, 0.5, and 2.5 µM, pulses to 0 mV, 80 ms for bH1 (A) compared with
rat-Nav1.5 (B) and rBIIA (rat-Nav1.2) (C).
Traces for representative experiments normalized to peak current. Note
that the control traces in MBS/0 ATX-II show the inactivation decay to
be slower for the heart clones compared with brain (for the traces
shown,  h,fast/h,slow estimated from
double-exponential fits to the decay are for bH1 1.34/10.7 ms, for
rat-Nav1.5 1.56/13.4 ms, and for rBIIA
(rat-Nav1.2) 0.6/5 ms). In saturating ATX-II, rBIIA shows
besides a slowed decay, a prominent plateau current. In bH1 the decay
is more slowed by ATX-II (h,slow of 24.5 ms) than in
rat-Nav1.5 (h,slow of 9.4 ms) and both show
no or little plateau current.
|
|
To exclude a general insensitivity of bH1 to certain inactivation
modulators, the inactivation kinetics of bH1,
rat-Nav1.5, and rat-Nav1.2
(rBIIA) was determined using increasing concentrations of ATX-II (Fig.
4). ATX-II, a site-3 toxin known to act at the external side of sodium
channels (Narahashi et al., 1969; Romey et al., 1976), induced typical
changes of the inactivation process in all three channels tested.
Depending on toxin concentration, a growing fraction of channels was
affected. As expected, the bH1 channels acquired a much slower
inactivation decay in the presence of ATX-II, similar to
rat-Nav1.5-ATX-II channel-toxin complex. In
contrast to the two heart isoforms, in the brain isoform rat-Nav1.2 (rBIIA), ATX-II not only caused a
slower decay but also a persistent plateau consistent with previously
published results (Benzinger et al., 1999).
Effects of SDZ 211-939 on Sodium Currents of bH1 and
Rat-Nav1.5.
The sensitivity of bH1 and
rat-Nav1.5 to SDZ 211-939 was compared using
concentrations from 0 to 50 µM SDZ 211-939 (Fig.
5, A and B). The
rat-Nav1.5 taken as a positive control (Kallen et al., 1990), showed the effect known for DPI 201-106: a slowed decay and
incomplete inactivation visible as a prominent plateau current during
sustained depolarization (Romey et al., 1987; Krafte et al., 1994).
These effects were obtained in presence of SDZ 211-939 in our
experiments (Fig. 5A). In contrast, inactivation of the cloned heart
isoform bH1 appeared mostly unaffected even at 50 µM. However, a
marginal slowing and a small increase of the plateau current was
visible as illustrated in Fig. 5B. A 5-min incubation was found
sufficient for distribution of SDZ 211-939 in the bath and possibly
through the oocyte membrane, because no further increase of the effect
was observed after 10 and 15 min. The increase in concentration was
achieved by addition of drug dissolved in MBS every 5 min. The total
peak currents showed an average decay of 20% over 25 min, which
corresponds to a normal run-down; this is about 10 times less than the
observed increase in the plateau current. This was also seen in solvent
controls (Figs. 7 and 8), which are discussed below. The pulses shown
were to 0 mV for 80 ms to simulate the conditions that sodium channels are exposed to during the long plateau of an action potential at around
0 mV. Subsequently, the concentration-response relationship has been
obtained from data shown for single oocyte experiments as shown in Fig.
5, A and B. The measurements were repeated in several batches of
oocytes [bH1 (n = 5), rat-Nav1.5
(n = 2)]. For this analysis, the ratios of the
amplitudes of plateau and peak currents were calculated and plotted
versus the SDZ 211-939 concentrations (Fig. 5C). Calculation of the
EC50 (micromolar) shows that the potency for both
heart channels is nearly the same as quantified by the
EC50 values being equal
(EC50 = 10.3; 95% confidence interval of
4.5-23.3 µM for bH1 and EC50 = 10.6;
6.59-17.16, 95% confidence interval for
rat-Nav1.5). However, the maximal effect in both
clones differs as shown by Vm, which
is 0.15 (0.10-0.21, 95% confidence interval) and 0.69 (0.53-0.88,
95% confidence interval) for bH1 and rat-Nav1.5,
respectively (Fig. 5C). The reciprocal representation (Lineweaver-Burk
plot; Fig. 5C, inset) shows that the data fit a straight line
indicating a 1:1 binding of SDZ 211-939 to the channel protein and
follow a Michaelis-Menten kinetics.
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 5.
Effect of SDZ 211-939 (applied concentrations 0, 2.5, 7.5, 12.5, 25, and 50 µM achieved by added stock solution of SDZ
211-939 in MBS every 5 min) on rat-Nav1.5 (A) and bH1 (B)
and representative analysis of these two data sets (C). Sodium currents
for pulses to 0 mV, 80-ms duration; traces normalized to peak current
of each SDZ concentration; peak currents show typically some recovery
initially and then a slight run-down over time, but all within 20% on
average (see Discussion; Fig. 7). The absolute size of
the peak currents for test pulses to 0 mV were 0.5 to 1.5 µA for
rat-Nav1.5 and 8 to 14 µA for bH1. C, response (ratio of
plateau to peak current) plotted versus concentration of SDZ 211-939 for data in A and B. For rat Nav1.5 n = 2 ( ), for bH1 n = 5 (). EC50 was
calculated as described under Materials and Methods. The
double reciprocal plot for the data obtained from the experiments shown
in Fig. 5, A and B (Lineweaver-Burk representation) is given in the
inset (y-intercept = 1/Vm). Note that in this representation the
small plateau values of bH1 by reciprocity appear larger and that the
values follow a straight line.
|
|
Effect of SDZ 211-939 on Voltage Dependence of Activation and
Inactivation.
Having shown that SDZ 211-939 binds to bH1 as
demonstrated by the small but reproducible effect, the next aim was to
investigate whether SDZ 211-939 exerts its effect at this channel via
structures in the gating machinery of activation or inactivation.
Identical standard protocols were repeated with the same cells in the
presence and absence of SDZ 211-939 at the nearly saturating
concentration of 50 µM. The data revealed no difference in the I/V
relationship (Fig. 6A), indicating that
activation was not affected. With respect to the inactivation process,
the steady-state inactivation was assessed as well as the recovery from
fast inactivation. Figure 6B shows the currents evoked by equal pulses
to +20 mV but after different prepotentials to 120, 100, and 60
mV. The resulting peak currents are plotted for a complete series of
prepotentials in Fig. 6D to obtain the steady-state inactivation curve.
By comparison of Fig. 6, B and D, it becomes clear that in SDZ there is
always a plateau of about 10% of the maximal attainable peak. The
effect of the prepotential consists in reducing the peak current only but this effect was seen in presence and in absence of SDZ 211-939. This will be discussed later under Fig. 8, but it is obvious that the
effect on the plateau current is independent of the prepotential. The
recovery from inactivation at three recovery potentials is shown in
Fig. 6C. The fitted time-constants are nearly identical at 120 and
they are slightly smaller at 80 mV in SDZ.
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 6.
Voltage dependence of bH1 gating in absence (filled
symbols) and presence of 50 µM (open symbols) SDZ 211-939. A, peak
current versus voltage (n = 5 each; S.E.M. too
small to be visible). B, steady-state inactivation shown by original
traces for a test pulse to +20 mV after prepotentials during 100 ms to
120, 100, and 60 mV (largest to smallest peak; see analysis in D
and Fig. 8). C, recovery from fast inactivation produced by a
conditioning pulse to 0 mV for 100 ms followed by test pulses to 0 mV
after recovery for the indicated time (dt-rec.) at potentials of 80
mV ( ), 100 mV ( ), and 120 mV ( ). The recovery
time constants in MBS or 50 µM SDZ 211-939 are, respectively
(n = 3 for bH1, n = 4 for
rBIIA), at 120 mV, 30 or 28 ms; at 100 mV, 90 or 74 ms; at 80 mV,
234 or 152 ms. D, steady-state inactivation as obtained from traces as
shown in B for prepotentials ranging from 120 to 0 mV in steps of 10 mV (fitted curves to mean ± S.E.M., n = 3 for
MBS, n = 4 for SDZ 211-939 each). Note that the
figures are normalized to the largest currents at 120 mV. In bH1, the
plateau currents were observed for all prepotentials to be of equal
size which is reflected in the pedestal of the values above 40 mV,
where the test pulse to +20 mV elicited no additional transient sodium
current but only the sustained plateau.
|
|
The absence of an important inward current (slow phase) on the return
to the holding potential indicated that deactivation of the slowly
inactivating current is very rapid.
| |
Discussion |
We have cloned and characterized the first cardiac sodium channel
(bH1), which displays a remarkable low sensitivity to synthetic modifiers. Expression of bH1 and rat-Nav1.5 in
X. laevis oocytes and subsequent
electrophysiological and pharmacological investigation allowed us to
analyze isolated sodium currents. In the mammalian heart, as so far
investigated, SDZ 211-939 and related synthetic modifiers such as DPI
201-106 prolong the action potential associated with a positive
inotropic effect. The equivalent effect was obtained at the cellular
level by whole-cell voltage clamp on neuroblastoma cells and cultured
rat cardiac cells (Romey et al., 1987). In these cells, sodium channel
modulators slow down the kinetics of sodium channel inactivation and
cause a sodium current plateau. This key feature remained when using
heterologously expressed rat-Nav1.5 protein (Fig.
5A).
Our attention was drawn to the fact that in bovine multicellular
myocardial preparations no change or even a shortening of the action
potential duration was observed, accompanied, however, by a small
inotropic effect (Mevissen et al., 2001). This unique insensitivity of
the bovine heart sodium channel to SDZ 211-939, compared with the
effects seen in preparations of other species, was now confirmed in the
recombinant bH1 channel (Fig. 5B). In addition, the interaction of SDZ
211-939 with bH1 has been analyzed in more detail with respect to
possible binding and efficacy, and further insight was gained on how
this class of drugs exert their pharmacological effect at the molecular level.
Electrophysiological Analysis of Small Plateau Current Induced by
SDZ 211-939 in bH1.
Initially, the interaction between SDZ 211-939 and bH1 indicated a lack of affinity of this drug to this channel,
confirming previous observations (Mevissen et al., 2001). However, the
quantitative analysis of the small increase of the plateau current in
response to SDZ 211-939 suggested a similar and simple 1:1 binding
stoichiometry to bH1 as for the control
rat-Nav1.5 (EC50 of around
10 µM). Straight lines, as shown in inset of Fig. 5C, were typical
for the SDZ 211-939 concentration experiments and could already be
taken as an indication for a true effect of SDZ 211-939. However, at
this point it seems necessary to discuss critically whether the
observed small plateau in bH1 might be an artifact. The
concentration-response data of Fig. 5, B and C, showed the plateau/peak
ratio for each SDZ 211-939 concentration. However, effects on the peak
Na+ current could be responsible for this effect.
Figure 7 gives these data as mean ± S.D. for four experiments. In each of the four oocytes the peak
currents were measured at 0 mV every 5 min after adding the necessary
additional amount of SDZ 211-939. As known for
Na+ channels, there is some recovery from
inactivation after the oocyte is clamped to 100 mV from its resting
potential of about 30 mV. Therefore, the first measurement was taken
after 5 min in the absence of SDZ 211-939. An additional increase of
about 20% was observed 5 min after 2.5 µM SDZ 211-939, which was
then followed by a small run-down of some 20% over 30 min. This was observed in all four experiments. To calculate the average of the
currents for peak and plateau, the data were first normalized for each
single oocyte to the maximal peak current after 5 min (being 1.0 in the
plot of Fig. 7). No differences in the peak currents of more than about
20% were seen and therefore, the increase of the plateau-currents by a
factor of almost 10 is not due to changes in the peak current.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 7.
Effect of increasing concentrations of SDZ on bH1
Na+ currents shown separately for the peak current ()
and the plateau current () for experiments as in Fig. 5B. Test
pulses to 0 mV, holding potential 100 mV. After an initial
equilibration period of 5 min at 100 mV (voltage-clamp switched on),
the first control measurement at 0 SDZ was taken and then SDZ dissolved
in MBS was added for the next concentration, where the measurement was
made after 5 min. Thus, the data at 50 µM SDZ were obtained 30 min
after switching to 100 mV holding potential. To compare the data of
the different sets, for each experiment the currents for peak and
plateau were normalized to the maximal value of the peak at 10 min
(being 1 in the graph). Data are shown as mean ± S.D.
(n = 4). Note the logarithmic abscissa to visualize
the plateau currents better; for the peak, the SD is too small to be
visible. See Discussion.
|
|
Furthermore, solvent controls (MBS) over long periods did not show any
effect in the plateau also when different prepotentials were used as
demonstrated in Fig. 8 (compare also
traces in Fig. 6B).
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 8.
Analysis of the effect of the prepotential to 120,
100, and 60 mV (Fig. 6B) and also the effect of time on the plateau
current in 0 µM (open symbols) and 50 µM SDZ (closed symbols). All
data are normalized to the peak current for 100 mV prepotential for
each experiment (data from four experiments shown). Although the data
for the peak currents nearly superimpose as also seen in Fig. 6D, the
plateau currents differ for 0 and 50 µM SDZ independent of incubation
time: The data in 0 SDZ were obtained after 30 (open squares) or 45 min
(open circles) and show that the plateau remains small at about 1%.
This is taken as solvent control over time. The data in 50 µM SDZ
211-939 are both close together whether SDZ acted 5 (filled circles) or
30 min (filled squares) at about 10% of the peak.
|
|
Interaction between SDZ 211-939 and Sodium Channel Heart
Isoforms.
The similarity in EC50 values
between the isoforms suggests similar binding sites but possibly
different positioning of SDZ 211-939 relative to the gating structures.
In our experiments we observed no change caused by SDZ 211-939 in the
voltage dependence of the activation (I/V curve) and only minimal
changes in the steady-state inactivation or the recovery from fast
inactivation. The prominent effect was the increase of the plateau
current compared with the peak current.
An approximate estimation the effect of SDZ 211-939 on the rates
between open and inactivated state in rat-Nav1.5
and bH1 should be performed. The mechanism of the modulation of
inactivation could concern the voltage-sensor S4DIV (Chahine et al.,
1994; Chen et al., 1996; Kontis and Goldin, 1997), the hinged lid (West et al., 1992), the C-terminal and DI-DII junction (Wei et al., 1999),
and/or its binding site(s) when it closes the pore from the cytoplasmic
side under control of S4DIV (Kühn and Greeff, 1999). As shown
elsewhere (Greeff and Forster, 1991), the inactivation phase and its
time-constant h depend on several rate
constants, especially activation and inactivation and a change of this
parameter would not give information about the possible mechanism.
Nevertheless, h was estimated for six
experiments with and without SDZ 211-939 for pulses to 0 mV and it was
found to be 2.41 ± 0.23 S.D. (n = 6) in MBS and
2.88 ± 0.24 S.D. (n = 6) in 50 µM SDZ 211-939, i.e., an increase of only about 20% at 50 µM SDZ 211-939. As
described in Kühn and Greeff (1999), the inactivation
process most likely includes a first voltage-dependent step to a second
open state due to a movement of sensor S4DIV, and point mutations in
S4DIV did cause strong changes of h. In this
second open state a receptor region is presented for the inactivation
loop (L3 between domains 3 and 4), which then closes the pore. An
interaction of SDZ 211-939 with this step would cause a sustained
plateau, i.e., a flickering between the open (O) and inactivated (I)
states. Therefore, the plateau size allows an estimation of the rates
between (O) and (I) states, especially their change due to SDZ 211-939. During the sustained plateau-current the product of the occupancy of state (O) times the (O)-to-(I) rate equals the occupancy of state (I)
times the (I)-to-(O) rate. Based on this argument one can conclude that
the reopening rate is changed by SDZ 211-939 about 10 times for bH1 and
about 70 times for rat-Nav1.5, because the plateaus grow from about 1 to 10 (bH1) or 70%
(rat-Nav1.5), respectively.
The binding site and the mode of action of SDZ 211-939 at the molecular
level are at present unknown. Most differences in the primary structure
are found within the large connecting loop between DII and DIII (Fig.
1; Table 1). Whether this causes a different positioning of bound SDZ
211-939 and affecting a voltage-independent part of the
inactivation-related structures will have to be investigated in future
experiments using site-directed mutagenesis. In addition, we observed
that a conserved arginine was substituted for a histidine residue in
the loop between domain I and domain II of bH1 (Fig. 1). Murphy et al.
(1996) have shown that the rat cardiac -subunit is phosphorylated
selectively in vitro and in intact cells by protein kinase A on two
sites in this region (Ser526,
Ser529, rH1 numbering). Because the
arginine-to-histidine substitution in the bovine sequence is located in
between the two serines shown to be phosphorylated, the phosphorylation
and the activity of the bovine heart sodium channel might be affected.
Ser529 in rat-Nav1.5
corresponds to Ser531 in bH1, and this
phosphorylation site is probably lost in bH1. How this substitution in
the bovine heart sodium channel might affect its phosphorylation and
activity is at present unknown and will be investigated in further studies.
In future experiments it would also be interesting to study the degree
of gating-current immobilization. The inactivation lid, besides closing
the open pore (inactivation), also interferes with the voltage-sensors,
"foot-in-the-door-effect" as originally proposed by Bezanilla and
Armstrong (1975). Recent studies on immobilization (Cha et al., 1999;
Kühn and Greeff, 1999; Sheets et al., 2000) seem to confirm this idea.
Correlation of Pharmacological Effect of SDZ 211-939 in Myocardial
Cells and in Oocyte Expression System.
The positive intropy
induced by this class of modulators is convincingly explained by an
increased sodium influx during an action potential. The increased
cytoplasmic sodium reduces the driving force for the
Na+/Ca2+-exchanger, and the
resulting Ca2+ increase enhances the
contractility (Scholtysik, 1989). Based on the observed plateau levels
for clamp pulses to 0 mV (Fig. 5, A and B), which correspond to the
average voltage during an action potential plateau, we can now roughly
assess the relative change in sodium influx during a heart action
potential. Without SDZ 211-939, sodium influx occurs only during the
first fast spike of about 5 ms in response to the clamp pulse. In the
rat clone, SDZ 211-939 resulted in a plateau of about 70% of the peak
current during a sustained voltage-clamp pulse to 0 mV. Thus, SDZ
211-939 resulted in an increase in sodium influx during the plateau of about 300 ms of an action potential compared with the fast initial sodium spike of about 5 ms, and would then be about 300 ms/5 ms times
0.7 (70% plateau), i.e., 42 times larger than normal. In bH1, the
plateau level (probability of open sodium channels) is approximately 5 times smaller. Because the bovine action potential also lasts about 300 ms at about 0 mV, the SDZ 211-939-induced increase in sodium influx and
the positive inotropic effect would be expected to be 5 times smaller.
This fits the earlier observations seen in myocardial cells (Mevissen
et al., 2001).
Conclusions and Outlook.
Availability of two nearly identical
sodium channels with a rather different sensitivity to the same drug of
potential pharmacological and clinical value opens at least two
interesting lines. 1) Experimentally, one can discern two clearly
different phenotypes by simply adding a drug to the bath, and so can
easily screen different clones for their corresponding
EC50 and efficacy of SDZ 211-939. Having characterized the differently responding clones of rat and bovine heart
sodium channels, one can now proceed to vary the interface between drug
and channel. This will be done by construction of chimeras between the
two channels to understand the reasons for the low efficacy in bH1. 2)
Basic recognition of the molecular background of cardiac sodium channel
gating may also enable medicinal chemists to change the structures of
DPI 201-106 or SDZ 211-939 and related substances to investigate the
structure-activity relationship. The therapeutic potential of synthetic
sodium channel modifiers is still not exploited to an adequate extent,
also due to the lack of understanding of their mechanism of action.
Cloning, expression, and characterization of a channel with unique
properties contribute to this understanding.
We are grateful to Dr. R. G. Kallen (University of
Pennsylvania) for generously providing the pSP64T-rSkM2/rH1 clone and
to Dr. A. Goldin (University of California, Irvine, Irvine, CA) for the
rBIIa, the vector pBSTA, and helpful advice. We also thank Dr. C. J. Portier (National Institute of Environmental Health, Research
Triangle Park, NC) for help performing the statistical analysis. We are
indebted to J. Lis (Institute of Veterinary Pharmacology, University of
Bern) for excellent technical help as well as to B. Colomb, Dr. G. Dolf, and Dr. J. Schläpfer (Institute of Animal Breeding,
University of Bern) for help in the course of sequencing. Dr. R. Haltiner (Institute of Veterinary Pharmacology, University of Bern)
contributed substantially to the success of the experimental part
(molecular biology).
This research was supported by a grant from the Schmid-Fonds
(Prof. Tino Hess).
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
)-(S)-6-Amino-
-[(4-diphenylmethyl-1-piperazinyl)-methyl]-9H-purine-9-ethanol
(SDZ 211-939)
Top
Abstract
Introduction
