 |
INTRODUCTION |
ATP-dependent K+ channels
(KATP channels)1
are gated by the intracellular ATP/ADP ratio with ATP inducing channel
closure and MgADP channel opening. Functionally, these channels link
membrane potential and excitability to the metabolic state of the cell (1, 2). Pharmacologically, KATP channels are closed by the hypoglycemic sulfonylureas such as glibenclamide and their benzoic acid
analogs, the glinides; these drugs are used in the treatment of
diabetes type 2. KATP channels are activated by the
K+ channel openers, a chemically heterogeneous group of
compounds including the cyanoguanidines like pinacidil and P1075 and
the benzopyrans such as levcromakalim; these compounds relax smooth muscle and induce hypotension (2-5).
KATP channels are heteromeric complexes composed of
pore-forming inwardly rectifying K+ channels (Kir6.x) and
sulfonylurea receptors (SURs) (6, 7) (for reviews, see Refs. 8-10).
Kir6.x is encoded by either one of two genes giving rise to two
subtypes, Kir6.1 and -6.2, with the latter being present in most
KATP channels. ATP binding to Kir6.2 induces channel
closure, thus accounting for the inhibitory action of the nucleotide
(11). SUR is a member of the ATP-binding cassette protein superfamily
(7, 12). It carries binding sites for nucleotides, the nucleotide
binding folds (13), and for the sulfonylureas and the openers (12, 14,
15). SUR also is encoded by either one of two genes. SUR1,
expressed in the pancreatic 
-cell and predominantly in neurons,
exhibits high affinity for the sulfonylureas and low affinity for the
openers. SUR2 is expressed in the various muscle types and shows high
affinity for the openers but lower affinity for the sulfonylureas (14, 15) (for reviews, see Refs. 5 and 16). Alternative splicing of the SUR2
mRNA leads to two isoforms, SUR2A and SUR2B, which differ only in
the last carboxyl-terminal exon. SUR2A is expressed in skeletal and
heart muscle; SUR2B is expressed in smooth muscle (17, 18).
The nucleotides not only gate the channel, they also profoundly affect
the potency and efficacy of the synthetic KATP channel modulators by allosteric interactions. First, there is a positive allosteric interaction between nucleotide and opener binding, enabling
SUR to bind openers with high affinity (14-16, 19, 20). Conversely,
openers enhance the ATPase activity of SUR2A, thereby promoting channel
activation (21). Second, there is a negative allosteric interaction
between nucleotide and sulfonylurea binding (22, 23). Third, there is a
negative allosteric coupling between opener and sulfonylurea binding,
generally leading to mutually exclusive binding of the two ligands to
SUR (14, 15, 19, 24).
In varying the cyanoguanidine structure, Khan and colleagues (25)
discovered that introduction of a phenyl ring in the side chain (Fig.
1) gave compounds that potently reversed the vasodilation produced by
various KATP channel openers but not that by other vasodilatory maneuvers; in addition, the inhibitory effect resided in
one (i.e. the R-) enantiomer. It was concluded
that the novel compounds acted as KATP channel blockers.
The results of their study may be interpreted in two ways. First, the
compounds may have acted as "neutral antagonists" (i.e.
by displacing the opener from SUR without directly affecting channel
activity); channel closure would then be induced by the high
intracellular ATP concentrations in the vascular smooth muscle cell.
Alternatively, the compounds may have acted as "inverse agonists"
(i.e. they displaced the opener, and, similar to the
sulfonylureas, their binding induced channel block also in the absence
of ATP). In order to decide between these alternatives, the enantiomers
of one such compound, PNU-94750 (Fig. 1),
were examined in electrophysiological experiments using recombinant
KATP channels. Since KATP channel openers and blockers (such as sulfonylureas) differ in their allosteric coupling to
MgATP binding (see above), we also investigated the effect of MgATP on
the binding of the enantiomers in radioligand binding assays. The study
showed that the enantiomers of PNU-94750 differed in their coupling to
MgATP binding and that the R-enantiomer (PNU-96293) was
negatively coupled and inhibited KATP channels, whereas the S-enantiomer (PNU-96179) was positively coupled and was an
opener.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 1.
Structures of P1075, pinacidil, and
PNU-94750. PNU-94750 is the racemic mixture of PNU-96179
(S-enantiomer) and PNU-96293 (R-enantiomer). The
asterisk indicates the chiral center of the compound.
|
|
| |
EXPERIMENTAL PROCEDURES |
Cell Culture and Transfection--
SUR2B(Y1206S) was constructed
from murine SUR2B (GenBankTM D86038 (18)) as described
(26). Human embryonic kidney (HEK) 293 cells were cultured in minimum
essential medium containing glutamine and supplemented with 10% fetal
bovine serum and 20 µg/ml gentamycin (14). Cells were transfected
with the expression vector pcDNA3.1 (Invitrogen) containing the
coding sequence of murine SUR2B, SUR2B(Y1206S), murine SUR2A
(GenBankTM D86037 (18)), or rat SUR1 (GenBankTM
X97279), and cell lines stably expressing these proteins were generated
as described (14). Cotransfection of SUR with murine Kir6.2 (D50581
(17)) was done transiently at a molar plasmid ratio of 1:1 using
LipofectAMINE and Opti-MEM (Invitrogen) as described previously (14).
In cotransfections prepared for electrophysiological experiments, the
pEGFP-C1 vector (CLONTECH, Palo Alto, CA), encoding for green fluorescent protein, was added for easy identification of
transfected cells. 2-4 days after transfection, cells were used for
binding studies and electrophysiological experiments.
Electrophysiological Experiments--
The patch-clamp technique
was used in the inside-out, the cell-attached, and the whole-cell
configuration as described by Hamill et al. (27). Patch
pipettes were drawn from borosilicate glass capillaries (GC 150 or GC
150T; Harvard Apparatus, Edenbridge, UK) and heat-polished using a
horizontal microelectrode puller (Zeitz, Augsburg, Germany). For
experiments using inside-out patches, bath and pipette were filled with
a high K+-Ringer solution containing 142 mM
KCl, 2.8 mM NaCl, 1 mM MgCl2, 1 mM CaCl2, 11 mM
D-(+)-glucose; 10 mM HEPES, titrated to pH 7.4 with NaOH at 22 °C. After filling with buffer, pipettes had a resistance of 1-1.5 megaohms. After excision of the patch, the pipette
was moved in front of a pipe with a high K+-EGTA-Ringer
solution containing 143 mM KCl, 0.85 mM
MgCl2, 1 mM CaCl2, 5 mM
EGTA, 11 mM D-(+)-glucose, 10 mM
HEPES, titrated to pH 7.2 with NaOH at 22 °C. The PNU compounds and
glibenclamide were dissolved in stock solutions as described under
"Materials"; ATP was dissolved in buffer, and an appropriate amount
of MgCl2 was added to keep free Mg2+ constant
at 0.7 mM. Substances were applied to the patch via the
pipe, and patches were clamped to 
50 mV. For evaluation of the
inhibition by PNU-96293, traces were individually corrected for run down.
For experiments in the cell-attached configuration, bath and pipette
were filled with the high K+-Ringer solution described
above. After filling, the pipettes had a resistance of 1-1.5 megaohms.
Patch experiments were performed at 37 °C, and patches were clamped
at 50 mV.
Experiments using the whole-cell configuration were performed as
described by Russ et al. (28). The bath solution was 142 mM NaCl, 2.8 mM KCl, 1 mM
MgCl2, 1 mM CaCl2, 11 mM D-(+)-glucose, 10 mM HEPES,
titrated to pH 7.4 with NaOH at 37 °C. Patch pipettes were filled
with 132 mM potassium glutamate, 10 mM NaCl, 2 mM MgCl2, 10 mM HEPES, 1 mM EGTA, 1 mM Na2ATP, titrated to
pH 7.2 with NaOH, and had a resistance of 3-5 megaohms. Isolated cells were chosen. After establishing the whole cell configuration, cells
were clamped to 60 mV, and every 12.5 s, the following square
pulse protocol was applied. After 1 s at 60 mV, cells were
clamped to the test potential of 110 mV for 0.5 s and then, after 1 s at 60 mV, to the next test potential of 90 mV, etc. (seven steps increasing by 20 mV from 110 to 10 mV). At the end of
the protocol, capacitance and series resistance were measured, and the
latter was compensated by 70%. To show the time-dependent current at a specified voltage, data were averaged over the last 125 ms
spent at that voltage during the cycle and spliced together to give a
continuous curve.
Data were recorded with an EPC 9 amplifier (HEKA, Lambrecht, Germany)
using the "Pulse" software (HEKA). Signals were filtered at 200 Hz
using the four-pole Bessel filter of the EPC9 amplifier and sampled
with 1 kHz.
Membrane Preparation, Equilibrium Competition Experiments, and
Dissociation Kinetics--
For cells stably expressing SUR alone, the
antibiotic was withdrawn from the culture medium 1 week prior to
membrane preparation. Membranes were prepared as described (14).
Protein concentration was determined according to Lowry et
al. (29) using bovine serum albumin as the standard.
In equilibrium binding assays, membranes (final protein concentration
0.2-0.5 mg/ml) were added to the incubation buffer containing 139 mM NaCl, 5 mM KCl, 5 mM HEPES,
2.2/1.2 mM MgCl2, 1/0.003 mM Na2ATP and supplemented with the radioligand
([3H]glibenclamide ~2.5 nM or
[3H]P1075 ~2 nM) and the inhibitor of
interest at 37 °C. In the Mg2+-free experiments,
Mg2+ was omitted from the incubation solution, and EDTA (1 mM) was added. At equilibrium (15 min for
[3H]glibenclamide and 30 min for [3H]P1075
binding), incubation was stopped by diluting 0.3-ml aliquots in
triplicate into 8 ml of ice-cold quench solution (50 mM
Tris-(hydroxymethyl)-aminomethane, 154 mM NaCl, pH 7.4) and
rapid filtration under vacuum over Whatman GF/B filters (Whatman,
Clifton, NJ). Filters were washed twice with 8 ml of ice-cold quench
solution and counted for 3H in the presence of 6 ml of
scintillant (Ultima Gold; Packard Instrument Co., Meriden, CT).
Nonspecific binding of [3H]P1075 was determined in the
presence of 10 µM P1075, and that of
[3H]glibenclamide was determined in the presence of 100 µM P1075 or 1 µM glibenclamide for
SUR2B(Y1206S) and SUR1, respectively.
For measurement of the dissociation kinetics, membranes were incubated
with 2 nM [3H]P1075 in the presence of 2.2 mM MgCl2 and 1 mM
Na2ATP at 37 °C for 30 min. Dissociation was initiated
by the addition of P1075 + PNU-96293 (10 µM each), and
300-µl aliquots were withdrawn at different times for filtration as
described above.
Equilibrium Binding Experiments in Cells--
Experiments were
conducted at 37 °C with an incubation time of 30 min as described
previously (14, 28). Cells were suspended by rinsing with a
HEPES-buffered physiological salt solution containing 139 mM NaCl, 5 mM KCl, 1.2 mM
MgCl2, 1.25 mM CaCl2, 11 mM D-(+)-glucose, 5 mM HEPES;
gassed with 95% O2 and 5% CO2; and titrated
to pH 7.4 with NaOH at 37 °C. Binding experiments were started by
the addition of cells (final concentration 1 × 106
cells/ml, corresponding to 0.25 mg of protein/ml) to the buffer supplemented with ~2 nM [3H]P1075 and the
inhibitor of interest. After 30 min, incubation was stopped as
described above, and aliquots were filtered over Whatman GF/C filters.
Data Analysis--
Concentration dependencies were analyzed by
fitting the logistic function
|
(Eq. 1)
|
to the data. Here, A denotes the extent of the
effect (amplitude); n (= nH) is the Hill
coefficient, x is the concentration of the compound under
study, and IC50 the midpoint of the curve with
px = log x and
pIC50 = log IC50.
In binding experiments, the dependence of the midpoint of an inhibition
curve (IC50 value) on the concentration of the radioligand, L*, was calculated according to the equation by Cheng and
Prusoff (30),
|
(Eq. 2)
|
where Ki is the inhibition constant and
KL is the equilibrium dissociation constant of
the radioligand, L*. In general, this correction did not
exceed a factor of 1.5. In cases when a homologous competition
experiment (e.g. L* L) was
conducted in the presence of a fixed concentration of a further ligand, C, that also competed with the radioligand, Equation 2 was
expanded to read as follows,
|
(Eq. 3)
|
where KC is the equilibrium dissociation
constant of competitor C. The homologous competition curves
(in the absence and presence of competitor) may be transformed
according to Scatchard (31),
|
(Eq. 4)
|
where B and F represent the bound and free
concentration of the unlabeled ligand, L. B (in
percent) is calculated from the radioactive ligand specifically bound,
B*, as B = 100 B*.
Fits of the equations to the data were performed according to the
method of least squares using the program SigmaPlot 6.1 (SPSS Science,
Chicago, IL). Individual binding competition experiments were analyzed
according to Equation 1. Errors in the parameters derived from the fit
to a single curve were estimated using the univariate approximation
(32). Amplitudes and pK values are normally distributed
(33); here, K values with the 95% confidence interval
in parentheses are given. Propagation of errors was taken into account
according to Bevington (34).
Materials--
PNU-96293 and PNU-96179 were synthesized
according to Humphrey et al. (35). [3H]P1075
(specific activity 4.5 TBq (118 Ci) mmol
1) was purchased
from Amersham Biosciences, and [3H]glibenclamide
(specific activity 1.85 TBq (50 Ci) mmol1) was from
PerkinElmer Life Sciences. The reagents and media used for cell culture
and transfection were from Invitrogen (Karlsruhe, Germany).
Na2ATP was from Roche Diagnostics, and glibenclamide was
from Sigma. P1075 was a kind gift from Leo Pharmaceuticals (Ballerup,
Denmark). KATP channel modulators were dissolved in dimethyl sulfoxide/ethanol (1:1) and further diluted with the same
solvent or with incubation buffer. In binding studies, the final
solvent concentration in the assays was always below 0.3% (in
electrophysiological experiments 0.1% or below).
| |
RESULTS |
PNU-96293 as an Inhibitor of Kir6.2/SUR2B
Channels--
Khan and colleagues (25) have shown that PNU-96293, the
R-enantiomer, reversed the vasorelaxation induced by other
K+ channel openers. We therefore examined the effect of
this compound on the recombinant KATP channel in
(nonvascular) smooth muscle, Kir6.2/SUR2B. After application of
PNU-96293 (10, 30, and 100 µM) to an inside-out patch in
the presence of 1 mM MgATP, no current developed (data not
shown). In agreement with Khan et al. (25), this indicated
that the compound was devoid of opener activity. When the channel was
opened by reducing the ATP concentration to 3 µM, the
compound inhibited the channel concentration-dependently (Fig. 2A). At 100 µM, the highest concentration tested, inhibition was
~50%. In comparison, glibenclamide, at the saturating concentration of 1 µM, inhibited the current by 80%. The inhibition
curve of the PNU enantiomer extrapolated to a maximum inhibition of
46% with an IC50 value of 16 µM (Fig.
2B, Table I). In the absence of ATP, the current was inhibited maximally by 30% with an
IC50 value of 12 µM (Table I).
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 2.
PNU-96293 inhibits Kir6.2/SUR2 channels in
inside-out patches. A, original recording of
Kir6.2/SUR2B channels in the presence of 3 µM ATP
([Mg2+]free = 0.7 mM) at 50 mV.
MgATP (1 mM) was applied repetitively for 15 s,
blocking the channel and inducing channel refreshment upon washout with
3 µM MgATP. The current developing during MgATP washout
was inhibited by glibenclamide (1 µM) and PNU-96293 (10, 30, and 100 µM). B,
concentration-dependent inhibition of Kir6.2/SUR2B ( )
and Kir6.2/SUR2A ( ) currents by PNU-96293 in the presence of 3 µM ATP; numbers in parentheses give
the number of patches. Inhibition by 1 µM glibenclamide
is shown for comparison. The parameters determined from the fit of
Equation 1 with Hill coefficient 1 to the data are listed in Table
I.
|
|
View this table:
[in this window]
[in a new window]
|
Table I
Inhibition of recombinant KATP channels by PNU-96293
Experiments were performed as shown in Figs. 2-4. The logistic
function (Equation 1) with Hill coefficient 1 was fitted to the data.
IC50 values are followed by the 95% confidence interval in
parentheses. A, the extent of inhibition (amplitude); i/o, the
inside-out configuration.
|
|
The effect of PNU-96293 was also studied in the whole-cell
configuration at 37 °C (Fig. 3).
Dialysis of the cell with 1 mM MgATP generated a current
that was inhibited by the PNU enantiomer in the nanomolar concentration
range (IC50 = 78 nM) with a maximum inhibition
of 74%; glibenclamide (1 µM) induced almost total
inhibition (Table I). Analogous experiments using the
Kir6.2/SUR2B(Y1206S) channel gave very similar results (Table I),
showing that the mutation did not alter the sensitivity of the channel
to PNU-96293.
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 3.
PNU-96293 inhibits Kir6.2/SUR2B channels in
whole-cell experiments. A, original recording. The cell
was clamped at 60 mV, and voltage was stepped from 110 to 10 mV in
20-mV steps as described under "Experimental Procedures." During
dialysis of the cell with MgATP (1 mM) at 37 °C, a
current developed that was inhibited by PNU-96293 in a
concentration-dependent manner. After washout,
glibenclamide (1 µM) produced total inhibition.
B, concentration-dependent inhibition at 60 mV.
The fit of Equation 1 with Hill coefficient 1 to the data (mean values
from the number of patches given in parentheses) gave the
parameters listed in Table I.
|
|
PNU-96179 as an Opener of the Kir6.2/SUR2B
Channel--
Fig. 4A
presents an original recording from an inside-out patch containing
Kir6.2/SUR2B channels. The channel was kept closed by MgATP (1 mM); the addition of PNU-96179 (100 µM)
generated a current that was completely inhibited by glibenclamide (3 µM) in a reversible manner. Further experiments showed
that the current induced by PNU-96179 (100 µM) was about
10-20% of that produced by P1075 at the maximally effective
concentration of 0.1 µM; lower concentrations of the PNU
compound showed smaller and inconsistent effects.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 4.
PNU-96179 opens Kir6.2/SUR2B channels.
A, original recording from an inside-out membrane patch.
After excision of the patch in the Mg2+-containing
solution, strong run-down was observed. ATP (1 mM) induced
channel closure; the further addition of PNU-96179 (100 µM) induced a current that was reversibly blocked by
glibenclamide (3 µM). Note the "refreshment" of the
channels upon washout of MgATP. B, cell-attached
configuration. PNU-96179 and P1075 were applied as indicated by the
bars. In both configurations, recordings were made in high
[K+] buffer (140 mM) at a holding potential
of 50 mV; temperature was ~22 °C in the inside-out and 37 °C
in the cell-attached configuration.
|
|
The ability of PNU-96179 to open the Kir6.2/SUR2B channel was also
examined in the whole-cell configuration; however, it proved difficult
to keep the channel closed during prolonged dialysis of the cell with
3-10 mM MgATP. Experiments were therefore performed in the
cell-attached configuration. Fig. 4B shows that PNU-96179 (100 µM) activated the channels to 13% (mean 15.6 ± 1.6; n = 6) of the level produced by P1075 (0.1 µM); at 10 µM, the PNU compound was
ineffective. Collectively, the data show that PNU-96179 is an opener of
the Kir6.2/SUR2B channel, albeit with low potency and efficacy. The
opening properties were not investigated further, since channel opening
is an effect expected for a cyanoguanidine.
Interaction of PNU Enantiomers with SUR2B: [3H]P1075
Binding Experiments--
Fig. 5
illustrates the inhibition of [3H]P1075 binding to SUR2B
by the two enantiomers in membranes. In the presence of 1 mM MgATP, the compounds inhibited binding of the
radioligand up to 100% with Hill coefficients of 1 and
Ki values of 2.3 and 0.66 µM for
PNU-96179 and PNU-96293, respectively (Table
II). Hence, stereoselectivity of binding,
measured as the eudismic (= Ki) ratio, was weak
(3.5). Similar experiments were performed also at an ATP concentration
of 3 µM (i.e. the lowest concentration at
which high affinity binding of the radioligand (which requires MgATP)
is detected) (14, 15). Under these conditions, the inhibition curve for
PNU-96179 was shifted to the right by a factor of 2.7 and that of
PNU-96293 to the left by a factor of 10, giving now a eudismic ratio of
100. Hence, increasing ATP increased the potency of PNU-96179 and
strongly decreased that of PNU-96293, indicating positive allosteric
coupling between PNU-96179 and MgATP binding, whereas for PNU-96293,
the coupling to MgATP was negative.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 5.
Inhibition of [3H]P1075 binding
to SUR2B by PNU-96179 and PNU-96293. Experiments
(n = 3) were performed in the presence of MgATP (1 mM () and 3 µM ()) and evaluated
individually. The mean parameters are listed in Table II. The
arrows indicate the shift of the curves induced by
decreasing MgATP concentration. [Mg2+]free
was 1 mM, and [3H]P1075 was ~2
nM; specific binding (BS) (100%)
corresponded to 217 ± 16 and 42 ± 4 fmol/mg protein at 1 mM and 3 µM ATP, respectively.
|
|
View this table:
[in this window]
[in a new window]
|
Table II
Binding of PNU enantiomers to SUR subtypes and coupling to MgATP
binding
Parameters of inhibition curves (Ki, inhibition
constant; A, amplitude (extent) of curve) were determined
from individual binding experiments as described in the legends to
Figs. 5 and 7. Hill coefficients were not different from 1.
|
|
In order to facilitate comparison with the electrophysiological
experiments, binding experiments were also performed in intact HEK
cells cotransfected with Kir6.2 plus SUR2B. Under these conditions, the
enantiomers inhibited binding of the radioligand with
Ki values of 3.6 and 0.2 µM,
respectively. Comparison with the Ki values
determined in membranes with SUR2B alone at 1 mM ATP showed that coexpression with Kir6.2 and/or the whole cell environment increased the Ki value of PNU-96179 slightly by a
factor of 1.6 ± 0.2 but decreased that of PNU-96293 by a factor
of 3.4 ± 0.4 (Table II). The gain in potency of PNU-96293
obtained by these changes resembled that seen earlier with
glibenclamide (26).
Since the PNU compounds and P1075 are structurally related (Fig. 1), we
examined whether these ligands competed for the same site of SUR2B. To
this end, homologous [3H]P1075-P1075 competition
experiments were performed in the absence and presence of PNU-96179 (10 µM) or PNU-96293 (3 µM). The idea behind
this was that the presence of the PNU compounds, if they competed with
P1075 for the same binding site at SUR2B, should shift the P1075
binding curve to the right as predicted by Equation 3. Fig.
6 shows the inhibition curves in the
Scatchard representation (see "Experimental Procedures"). Both PNU
compounds reduced the slope but left the abscissa intercept unchanged.
The observed reductions in P1075 affinity produced by 10 µM PNU-96179 (3.3×) and 3 µM PNU-96293
(4.4×) are in quantitative agreement with those calculated assuming
competition with P1075 (Equation 3) and using the Ki
values determined in Fig. 5.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 6.
Competition between the binding of PNU
enantiomers and P1075 to SUR2B. [3H]P1075/P1075
inhibition curves were performed in the absence () and presence
() of PNU-96179 (10 µM) and PNU-96293 (3 µM). Data are shown in the Scatchard representation
(Equation 4) to highlight competitiveness. MgATP was 1 mM
and specific binding (BS) in the absence of PNU
compounds 202 ± 10 fmol/mg protein. PNU-96179 (10 µM) reduced specific binding to 32 ± 1% and
PNU-96293 (3 µM) to 28 ± 1%.
|
|
If the PNU enantiomers bind to the same site as P1075, they should not
affect the dissociation kinetics of P1075 from SUR2B. In control
experiments, the addition of unlabeled P1075 in large excess (10 µM) to the SUR2B-[3H]P1075 complex induced
dissociation of the complex with a half-time of 11 min. PNU-96293 (10 µM) in the presence of P1075 (10 µM) did
not alter the half-time of dissociation (data not shown).
Interaction of PNU Enantiomers with SUR2B:
[3H]Glibenclamide Binding Experiments--
Glibenclamide
and opener binding sites at SUR2B are linked by strict negative
allosteric coupling (14, 24); in addition, glibenclamide binding and
MgATP binding are negatively allosterically coupled (36). The coupling
between binding of the PNU enantiomers and ATP was therefore examined
also using [3H]glibenclamide as the radioligand. Since
high affinity glibenclamide binding is increased in the absence of
MgATP, this offers the advantage of studying the coupling over a wider
range of MgATP concentrations. However, the affinity of SUR2B for
glibenclamide (KD = 22 nM (36)) is too
low to obtain a good signal; use was therefore made of the
SUR2B(Y1206S) mutant, which has a 5.5 times higher affinity for
glibenclamide (26). We have shown previously that neither opener
binding nor the allosteric coupling between the binding of ATP, opener,
and glibenclamide are perturbed by this mutation (26, 36).
PNU-96179 inhibited [3H]glibenclamide binding only
partially, regardless of the presence or absence of MgATP (Fig.
7A and Table II). In the
presence of 1 mM MgATP, the Ki value was 2.5 µM (i.e. identical with that measured
using [3H]P1075 as the radioligand) (Table II). In the
absence of MgATP (1 mM EDTA, neither Mg2+ nor
ATP added), the Ki value was 83 µM
(i.e. 33 times higher). This indicated positive allosteric
coupling between PNU-96179 and MgATP binding, as observed also with
[3H]P1075 as the radioligand. Surprisingly, PNU-96293
increased binding of [3H]glibenclamide in the
presence of MgATP by 41 ± 3% with an EC50 value of
1.7 µM; in the absence of MgATP, partial inhibition was seen with Ki = 18 nM (Fig.
7B). The opposite sign of the PNU-96293 effect depending on
the presence of MgATP made a straightforward interpretation difficult.
However, the 100-fold greater potency of the compound in the absence of
MgATP suggested strong negative allosteric coupling between binding of
this enantiomer and MgATP binding, again in agreement with the results
obtained in the [3H]P1075 experiments.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 7.
Modulation of [3H]glibenclamide
binding to SUR2B(Y1206S) by the PNU enantiomers in the presence ()
and absence () of MgATP. Individual experiments
(n = 3 each) were normalized to 100% specific binding
(BS) in the absence of PNU compounds. Specific
binding in the absence of MgATP (1 mM EDTA, no extra
Mg2+ or ATP) was 352 ± 24 fmol/mg protein at 2.5 nM [3H]glibenclamide; the presence of
MgATP reduced specific binding to 21 ± 1%. Fitting parameters
are listed in Table II.
|
|
Experiments with Other SUR Subtypes/Recombinant
KATP Channels--
In inside-out patches at 3 µM MgATP, PNU-96293 inhibited the Kir6.2/SUR2A current
with IC50 = 21 µM and a maximum inhibition of
54% (Fig. 2B, Table I) (i.e. with parameters
similar to those found with the Kir6.2/SUR2B channel). In the whole
cell mode, a large scatter of the data was noted. Data were therefore
divided into two groups according to their glibenclamide sensitivity. Analyzing the group with high sensitivity to glibenclamide (at 1 µM glibenclamide: 79 ± 6% block; n = 3), the inhibition curve extrapolated to 68 ± 4% block at
saturation with an IC50 value of 14 (8, 23) nM
(Table I). The inhibition curve of the cells with poor sensitivity to
glibenclamide (block at 1 µM: 47 ± 15%; n = 4) gave a maximum extent of inhibition of 39 ± 11% and an IC50 value 0.5 (0.2, 1.3) µM.
This showed that the channels with poor sensitivity to glibenclamide
were also quite insensitive to PNU-96293. In [3H]P1075
binding studies with SUR2A, binding of the PNU-96179 was positively
allosterically coupled to MgATP binding, whereas PNU-96293 showed
negative coupling (Table II). These results were qualitatively similar
to the observations made with SUR2B; however, the ATP-induced shifts
were smaller (Table II).
In inside-out patches in the absence of ATP, PNU-96293 (100 µM) inhibited the Kir6.2/SUR1 current by 40 ± 5%
(n = 13); similar inhibition was obtained with
tolbutamide at 100 µM (Table I). In binding experiments,
PNU-96179 proved very weak in inhibiting [3H]glibenclamide binding to SUR1: at 1 mM,
inhibition amounted to ~55% independent of the presence of MgATP.
Extrapolating the curves to 100% inhibition at saturation gave
IC50 values of ~550 µM (i.e. an
~100-fold lower potency than found for the SUR2 subtypes) (Table II).
PNU-96293 exhibited a behavior qualitatively similar to that observed
at SUR2B(Y1206S) (Table II). In the presence of MgATP, the compound
increased [3H]glibenclamide binding to 121% with
EC50 = 31 µM; in the absence of MgATP,
partial inhibition was observed (maximum 22% and IC50 = 1.4 µM). Hence, the R-enantiomer was again
more potent in the absence of MgATP, indicating negative allosteric
coupling to MgATP binding as observed also in the binding experiments
with SUR2.
| |
DISCUSSION |
Here we report that the enantiomers of a racemic compound derived
from the KATP channel opening cyanoguanidines interact
specifically with the channel, the R-enantiomer (PNU-96293)
inducing channel inhibition and the S-enantiomer (PNU-96179) opening.
Inhibition of Recombinant KATP Channels by
PNU-96293--
In whole cell experiments and in the presence of ~1
mM MgATP, PNU-96293 inhibited the Kir6.2/SUR2 channels with
IC50 values in the low nanomolar range and a slight
selectivity for the Kir6.2/SUR2A channel. This is a potency similar to
that of glibenclamide under the same conditions (26). The efficacy of
PNU-96293, however, was lower than that of glibenclamide at both
Kir6.2/SUR2 channels (Table I). This may be related to differences in
the negative allosteric coupling to MgATP binding at SUR. Here we have
shown that MgATP decreases the affinity of PNU-96293 for SUR2B; in the case of glibenclamide binding, however, the number of binding sites is
reduced (23, 26).
Another point deserves comment. For the Kir6.2/SUR2B channel, the
binding and the electrophysiological experiments were performed under
similar conditions (whole cells, high MgATP, 37 °C); however, the
channel inhibition curve (IC50 = 78 nM) was
shifted leftward from the binding curve (Ki = KD = 200 nM) by a factor of ~3. A
leftward shift of roughly this magnitude is expected if the channel has
four equal and independent sites for the ligand and if occupation of a
single site is sufficient for channel inhibition (37). Alternatively,
if binding of four molecules is required for channel block, the shift
is by the same factor ((21/4 1)1 = 5.3 (28))
but in the opposite direction.
Using the inside-out configuration, the potency and efficacy of
PNU-96293 in blocking the Kir6.2/SUR2 channels were reduced. The loss
in potency was dramatic (IC50 values in the 20 µM range); it was also unexpected for two reasons. First,
the contrary is observed with sulfonylureas (e.g.
glibenclamide) and analogs (e.g. repaglinide); in these
cases, the sensitivity for channel block is increased in the inside-out
patch (glibenclamide (26, 38) and repaglinide (39)). Second, the
IC50 values for channel inhibition are 190-250 times
higher than the Ki values of the respective binding
curves. Despite some differences in the experimental conditions (binding experiments were done at 37 °C and to SUR alone), these observations suggest that the transduction of PNU binding into to
channel inhibition is more complex than in the case of the sulfonylureas. One may speculate that the more physiological
environment of the (dialyzed) whole cell offers a component
(e.g. another protein, a metabolic product) not present in
the inside-out patch. Further work is needed to resolve this question.
The other point is the lower efficacy of block observed in the
inside-out configuration. This is less surprising, since it has
generally been observed with all KATP channel inhibitors
acting via SUR (e.g. see Refs. 38, 40, and 41). The reason
for this behavior remains unknown.
Binding Sites of the Enantiomers--
Information on the binding
site(s) of SUR for the PNU enantiomers was obtained from the binding
experiments. First we recall that SUR accommodates the binding sites
for the openers, the sulfonylureas, and nucleotides, and that these
sites are linked by allosteric interactions (Fig.
8A; see also the
Introduction). The partial inhibition of or even the increase in
[3H]glibenclamide binding to SUR2B(Y1206S) and SUR1
excluded the possibility that the enantiomers bound to the
glibenclamide site of SUR. Instead, these results showed that ternary
complexes of SUR with glibenclamide and a PNU enantiomer existed at
equilibrium and that the site(s) of the PNU enantiomers and the
glibenclamide site were linked by allosteric interactions. Similarly,
the modulation of PNU binding by MgATP showed allosteric linkage
between the PNU site(s) and nucleotide binding to the nucleotide
binding folds of SUR. The equilibrium binding experiments with
[3H]P1075 in Fig. 6 and the kinetic experiments, however,
were in quantitative agreement with a model in which the PNU
enantiomers and P1075 competed for the same binding site of SUR2B. In
view of the structural similarity between these three ligands (Fig. 1),
competition for the same binding site seems plausible; however, more
direct biochemical evidence (e.g. by photoaffinity labeling) is desirable. A first consequence of the competitive mechanism was that
the Ki values determined in [3H]P1075
inhibition assays (Table II, Fig. 5) can be equated with KD values for binding of these compounds to SUR2B. A second consequence was that the two enantiomers bind to the same site
of SUR2B, which is of importance in view of their opposite coupling to
MgATP binding and their opposite effects on channel activity.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 8.
Allosteric interactions between binding sites
of SUR. A, SUR has binding sites for nucleotides
(N), openers (O), and sulfonylureas
(SU). These sites are linked by positive and negative
allosteric interactions as indicated. For reasons of simplicity, the
two nucleotide binding sites of SUR were condensed into one.
B, in the presence of MgATP (N), PNU-96293
increases [3H]glibenclamide (G) binding.
Top, SUR with [3H]glibenclamide and MgATP
occupying their sites linked by negative allosteric interaction.
Bottom, by strong negative coupling to the MgATP site,
binding of the blocker PNU-96293 to the opener site induces
dissociation of MgATP from the MgATP site, leading to a
SUR-[3H]glibenclamide-PNU complex with increased
occupancy of the [3H]glibenclamide site. The latter
suggests that the negative allosteric interaction between
[3H]glibenclamide and PNU is weaker than that between
[3H]glibenclamide and MgATP. The strength of the
different allosteric interactions is indicated by the
thickness of the arrows. C, in the
absence of MgATP, PNU-96293 partially displaces glibenclamide by the
negative allosteric coupling between the respective sites.
|
|
Allosteric Coupling of PNU-96293 and MgATP Binding--
MgATP
inhibited binding of PNU-96293 to SUR via a negative allosteric
interaction, regardless of the radioligand used. Two points deserve
comment. First, MgATP induced a large reduction in the affinity of
PNU-96293 binding to SUR2B(Y1206S) (rightward shift ~100-fold),
whereas for glibenclamide the number of binding sites was reduced (23,
36). Regarding binding to SUR1, MgATP induced a rightward shift in both
cases, and the shift for PNU (22, Table II) was larger than that for
glibenclamide (~7 (38)). Assuming that the negative allosteric
coupling between PNU-96293 and MgATP binding is stronger than that
between glibenclamide and MgATP, one can explain that in the presence
of MgATP, PNU-96293 actually increased the binding of
[3H]glibenclamide to SUR2B(Y1206S) (Fig. 7B)
and to SUR1 (Table II). Including MgATP, one is faced with a system of
three binding sites, linked by negative allosteric interactions of
differing strength (Fig. 8B). As PNU-96293 binds to the
SUR-[3H]glibenclamide-MgATP complex, the strong negative
coupling between the PNU enantiomer and MgATP induces dissociation of
MgATP from SUR, and one is left with a
SUR-[3H]glibenclamide-PNU complex. Since more
[3H]glibenclamide is bound, one must conclude that the
negative allosteric interaction between glibenclamide and PNU-96293
binding is weaker than that between glibenclamide and MgATP binding
(Fig. 8B). In the absence of MgATP (Fig. 8C),
only the weak negative allosteric coupling between PNU-96293 and
glibenclamide comes into play, and partial inhibition of glibenclamide
binding results.
Second, the rightward shift in the Ki values of
PNU-96293 upon increasing ATP depended on the SUR subtype (SUR2B > SUR2A; SUR2B(Y1206S) > SUR1; Table II); this suggested
differences in the strength of the negative allosteric coupling between
PNU-96293 and MgATP binding with the SUR subtype. Variable
MgATP-induced rightward shifts were observed previously for the binding
of glibenclamide and of the sulfonylthiourea, HMR 1883, to SUR
subtypes; in this case, the maximum shift (measured as the ratio of the
Ki values) was ~8 for glibenclamide binding to
SUR1 (38). In addition, the glinide, AZ-DF265 (42), also displayed a
rightward shift of 2.8-fold in its interaction with
SUR2B.2 Collectively, these
data show that compounds of different chemical structures, which
inhibit KATP channels by binding to the SUR subunit, are
negatively allosterically coupled to MgATP binding to SUR.
Allosteric Coupling of PNU-96179 and MgATP Binding--
Increased
MgATP induced a leftward shift in binding of the
S-enantiomer to SUR2, indicating positive allosteric
coupling between the respective binding sites. With
[3H]P1075 binding to SUR2B and an increase in MgATP from
0.003 to 1 mM, the leftward shift measured for PNU-96179
was ~3-fold (Table II). This is in the range typical for openers; for
P1075 it was 1.7-fold (14), and for the benzopyran 217-774 (43) it was 3.8-fold.2 Much higher shifts were found with
[3H]glibenclamide as the radioligand, probably because
one can start from MgATP = 0. In this case, and using the mutant
SUR2B(Y1206S), the shift for PNU-96179 amounted to a ratio of 33; for
typical openers, it ranged from 100 to 300 (44). These data show that openers of different chemical structures are positively allosterically coupled to MgATP binding. In view of the fact that openers increase the
ATPase activity of SUR (21), this is not surprising.
Eudismic Ratio and Affinity of the PNU Enantiomers--
At 1 mM MgATP, the PNU enantiomers bound to the different SUR
subtypes in the low (SUR2) or high (SUR1) micromolar range
(i.e. binding was relatively weak). The eudismic (=
KD) ratio of the enantiomers was also weak, ranging
from ~1 (SUR2B(Y1206S)) to 17 (SUR1). Hence, the largest ratio
was found for the SUR with the weakest affinity for the PNU compounds
(i.e. SUR1). This was also true for the
Ki values at low or 0 MgATP (Table I). These
observations are not in accordance with Pfeiffer's rule, which states
that the weaker the racemate, the smaller the potency difference
between the enantiomers (45). One also notes that for all SUR subtypes,
PNU-96293 was more potent than PNU-96179, which agrees with the
electrophysiological experiments.
Significance of Results--
The major result of this study was
that the R-enantiomer of the racemic cyanoguanidine,
PNU-94750, exhibited the higher potency (25), was negatively coupled to
MgATP binding, and inhibited KATP channels; for the
S-enantiomer, the converse was true in all points.
Stereoselective gating of KATP channels has been observed before, both for openers (e.g. for cromakalim
(e.g. see Ref. 46) and for cyanoguanidines like pinacidil
(47)) and for blockers (e.g. the benzoic acid sulfonylurea
analogue AZ-DF 265 (42, 48)). In all of these cases, the distomer
affected the channel in the same way as the eutomer but with reduced
potency. Within the cyanoguanidine class, the PNU eutomer exhibited the
R configuration as in the case of pinacidil. However, the
S-enantiomer can also be the eutomer, indicating that SUR
has considerable flexibility in accommodating the ligand (47) (for a
review, see Ref. 4). More importantly, the PNU eutomer was a
KATP channel blocker, and the two enantiomers had opposite
molecular effects.
These observations are reminiscent of some dihydropyridines where,
depending on membrane voltage and channel kinetics, one enantiomer can
open and the other close the L-type Ca2+ channel (49, 50).
These opposite actions are apparently mediated by binding to the same
binding site of the channel (51). The enantiomers shift the channel
inactivation curve by different degrees to more negative potential,
which, in a certain voltage window, results in an increase or decrease
in the number of activable channels (50). Regarding the PNU
enantiomers, the available evidence suggests that they also bind to the
same binding site of their receptor (SUR); however, they show opposite
effects in their allosteric coupling to ATP binding (and possibly
ATPase activity) and in their modulation of KATP channel
activity. Within the framework of the biochemical mechanism of
KATP channel openers (21, 52) and blockers (13), one may
hypothesize that it is not the binding of a ligand to the opener site
of SUR per se but the modulation by this ligand of ATP
binding and ATPase activity of SUR that determines whether the compound
acts as an opener or a blocker of the channel.
,
TOP
ABSTRACT
INTRODUCTION
