Pfltigers Arch (1993) 424:343-353
Journal
The molecular mode of action of the Ca agonist (-) BAY K 8644
on the cardiac Ca channel
M. Bechem, H. Hoffmann
Bayer AG, Institute for Cardiovascular and Arteriosclerosis Research, P. O. Box 101709, D-42096 Wuppertal, Germany
Received August 19, t992/Received after revision March 3, t993/Accepted March 19, 1993
Abstract. The primary drug action of ( - ) BAY K 8644
on whole-cell Ca current in atrial myocytes was mea-
sured under conditions where secondary Ca-mediated
changes of Ca channel activity were minimized. The
most direct action of ( - ) BAY K 8644 is the change of
gating kinetics which results in a strictly voltage-depen-
dent increase of the peak current in the voltage range
between - 4 0 and 0 mV. Peak currents were increased
dose dependently in the concentration range from 1 to
30 nM. Analysis of peak current/voltage relations re-
vealed a linear shift of the current activation by approxi-
mately 23 mV to more negative membrane potentials,
without any change in its voltage dependence and in the
current reversal potential or the maximum whole-cell
conductance. Measurement of Ca current activation and
deactivation time constants suggests that ( - ) BAY K
8644 prolongs the single-channel open time without af-
fecting the closed time. From the shift of the open time
function to more negative voltages by about 50 mV the
energy transferred to the gating process is calculated to
be 5.4kJ/mol (1.3 kcal/mol). The drug-induced slow
component of tail current has been used to estimate the
true dose/response relation for ( - ) BAY K 8644. A KD
value of 4.3 nM and a Hill coefficient of 1.25 were de-
termined. Flash-induced competition experiments with
the Ca antagonist nifedipine allowed the measurement
of binding kinetics of ( - ) BAY K 8644. The association
rate constant is estimated to about 5 • 106 mo1-1 - s -1 and
dissociation time constant is approximately 5 0 - 7 0 s;
both are in close agreement with receptor binding stud-
ies. Results are discussed in relation to models for drug
action of dihydropyridine-type compounds and to impli-
cations for the structure of the Ca channel protein.
Key words: Ca channel - ( - ) BAY K 8644 - Ca
agonists - Cardiac cells - Ca current - Mode of ac-
tion - Dihydropyridine receptor
Correspondence to: M.Bechem
of Physiology
9 Springer-Verlag 1993
Introduction
The group of 1,4-dihydropyridines (DHP) consists of
two structurally closely related groups of substances
which are pharmacologically characterized as Ca antag-
onists and Ca agonists. The Ca antagonists are in wide-
spread therapeutical use, whereas the Ca agonists have
so far only attracted scientific interest (for review: [17]).
BAY K 8644 [34], the prototype of the Ca agonistic
DHPs has been widely used as a tool to probe Ca chan-
nel function in various tissues and cells [4]. The Ca
agonists have been found to facilitate greatly the investi-
gation and characterization of single-channel activity of
the L-type Ca channel. As Ca channel measurements,
with single-channel events of only about 1 pA amplitude
and durations in the millisecond time domain, in general
suffer from the limitations of patch-clamp recording, the
prolongation of single-channel open times by the Ca
agonist can help to facilitate the measurement of single-
Ca-channel activity [16].
Since the primary sequences of the al subunits of
several Ca channel types are known [5, 26, 28] DHP Ca
agonists and antagonists have become a basic tool for
the classification of the cloned and expressed Ca chan-
nels [35] and for proving the functional relevance of the
expressed subunits [30, 36]. It is expected that DHPs
may also help to map out the functionally relevant areas
within the al subunit and to construct integrated, func-
tional models if the location of the DHP receptor is
known - which is now under progress [7, 31] - and if
their molecular mode of action is understood.
However, despite this widespread use of DHPs, the
precise mechanism of their molecular interaction with
the Ca channel still remains unclear. KD values from
electrophysiological studies [14, 22, 33] are not in agree-
ment with those from receptor binding studies [23, 37]
raising the possibility of two different binding sites be-
ing involved in the various aspects of the Ca agonist's
action [22]. Also the voltage dependence of the agonist's
action [18, 19] and its dependence on the phosphory-
344

lation status o f the C a c h a n n e l are still u n d e r discussion. A o 1.7 nM ,3.0 nM 1"7 nM 80 nM

I I I I I I I I
A s a c o n s e q u e n c e o f these uncertainties, n o n e o f the p r o -
p o s e d m o d e l s [3, 15, 16, 22, 33] is either g e n e r a l l y ac-
c e p t e d or a c c o u n t s for all o f the e x p e r i m e n t a l findings.
The present paper includes a comprehensive kinetic ~ -lO0
a n a l y s i s o f the w h o l e - c e l l C a currents, as m o d i f i e d b y \
the D H P a g o n i s t ( - ) B A Y K 8644, a n d f o c u s e s on its
s p e c i f i c a n d h i g h affinity i n t e r a c t i o n ( c o n c e n t r a t i o n !
r a n g e 1 - 3 0 n M ) w i t h the C a channel. A d d i t i o n a l l y , n e w g' -200
f l a s h - i n d u c e d c o m p e t i t i o n e x p e r i m e n t s with n i f e d i p i n e
are d e s c r i b e d w h i c h a l l o w the m e a s u r e m e n t o f r a p i d
binding kinetics of DHP compounds under equilibrium
conditions. A s s o c i a t i o n a n d d i s s o c i a t i o n rates are deter- ~ \
m i n e d and d i s c u s s e d w i t h r e s p e c t to the m o l e c u l a r inter- -300 i i i i i i

20 40 60 80 1O0 120
a c t i o n o f the a g o n i s t w i t h the C a channel. time / rain
-30 mV
B J 1-50 mV membrane potential
-70 mV
Materials and methods
0
Ca currents were recorded in cultured, atrial myocytes from adult
guinea-pigs by means of the whole-cell recording configuration of
the patch-clamp technique [13]. The atrial myocytes were isolated
by Langendorff perfusion with collagenase and elastase as de- -100
scribed elsewhere [2]. Collagenases from different sources were
tested prior to their use to ensure activity and to guard against cell "E
damage. We currently use Sigma C7176 type L at a concentration
of 0.5 mg/ml. The freshly isolated, Ca-tolerant cells were taken I
into short-term tissue culture [meditun: DMEM with 25 mM 4- -2oo \ / b) 1.7 nM (-)BAY I<8644
(2-hydroxyethyl)-l-piperazineethanesulphonic acid (HEPES) and \ // c) 3.0 nM (-)BAY K8644
5 % - 1 0 % fetal calf serum, 5 mg/ml gentamicin] and used for the
experiments from days 2 to 7. During this time cells had a rounded,
~ / d) 17.0 rim (-)BAY K8644

ball-like geometry which makes them highly suitable for whole-

-300
cell patch-clamp recording. In contrast to freshly isolated cells the
membrane voltage can be precisely controlled and intracellular di- ; 2'0 ;0 6'0 ;0
time /ms
alysis with a single patch pipette is more reliable.
Cells were grown on specially designed culture dishes, which +10 mV
contain a cover slip bottom to facilitate the UV-light radiation dur-
-70 : V
ing the flash experiments. For measurements the culture dishes
were placed on the stage of an inverted microscope (Zeiss IM 35)
and kept at room temperature (23-26~ Medium was replaced
by a modified Tyrode's solution of the following composition a

(raM): NaC1 130, KC1 2, MgC12 1, NaHCO3 5, HEPES 20, pH 7.4,

glucose 10, CaC12 1 or 5; containing 10 -5 g/ml tetrodotoxin to
block Na currents. The pipette solution [(in raM): K-citrate 60, c~ - 1 0 0

CsC1 30, HEPES 10, pH 7.2, MgClz 1, Mg-ATP 4, dibutyryl cyclic

d a
adenosine monophosphate (cAMP) 0.1, ethylenebis(oxonitrilo)-
tetraacetate (EGTA) 2] was designed to block K currents, to re-
move inactivation of the Ca current (compare [2]) and to bring the g, -2oo
Ca channels into the fully phosphorylated state. b) 1,7 nM (-)BAY K8644
Patch pipettes were made from pyrex glass (pipette puller: c) 3.0 nM (-)BAY K8644
Zeitz Instrumente, DMZ-universal puller) having a resistance of d) 17.0 nM (-)BAY K8644
about 3 - 7 Ms For measurements of current and voltage control,
-300
a patch-clamp amplifier (List LMfEPC 7) was used. A PDP-11
computer system allowed both the application of voltage steps un- o ;o ;o 60 .'o
der program control and the recording (digitized at 20 kHz), stor- time / ms
age and analysis of whole-cell currents. During experiments the Fig. 1. A Plot of peak Ca current at a clamp potential of - 3 0 mV
linear leak and capacity currents were monitored by the application (11) and + 10 mV (V) versus time of the experiment. The cumu-
of short repetitive voltage steps from a holding potential of lative application of ( - ) BAY K 8644 (1.7, 3.0, 17, 80 nM) results
- 7 0 mV to - 6 0 mV and to - 9 0 mV and were subtracted from in a current increase only at - 3 0 mV. Original current recordings
the current traces during off-line data analysis. Further analysis of from the same experiment are shown in parts B and C. Although
current traces is described in the Results. peak currents at + 10 mV were not affected by the drug tail current
For flash experiments the microscope was equipped with an decay is slowed (C)
epifluorescence setup and an arc-lamp (HBO 100 W) was used as
the light source. By means of a dichroic mirror (Zeiss FT410), the
excitation light is directed to the objective (100• N.A. 1.25 oil
immersion) and focused on the specimen. Between the light source seconds. Light energy of the flashes was adjusted by variation of
and the dichroic mirror a computer-driven electronic shutter has flash duration. Control experiments established that the duration
been mounted to allow illumination times from 2 ms to several of illumination used in the present work (from 4 to 20 ms) neither
 345

Table 1. Mean values +-_ SD from the analysis of the current/volt- membranepotential/ mV
-60 -40 -20 20 40 60
age relations at control and in the presence of different concen- I I i
trations of ( - ) BAY K 8644
A ~
( - ) BAY ParamNer
K 8644
concen- V05 k Vr.~ Gm= n
tr~ion -100 /
(nM)

Control - 9 . 1 + 6.1 9.7 + 1.0 74.5 _+ 5.5 5.03 _+ 0.7 8

0.8 - 1 9 . 2 + 0.6 7.7 _+ 0.4 68.3 + 2.4 3.4 2
1.7 - 1 8 . 9 + 6.6 10.1 _+ 0.7 68.9_+ 4.3 4.35 +_ 0.8 6
3.0 - 2 4 . 0 + 4.6 10.3 +_ 1.1 68.0_+ 7.8 3.5 + 0.4 3
a
17 - 3 1 . 0 + 5.0 9.6 + 1.8 61.5 + 1.5 4.1 _+ 0.8 2 v 1.7 nM (-)BAY K 8 6 4 4
30 - 3 0 . 3 _+ 6.8 7.7 _+ 1.3 67.8 _+ 5.6 4.1 _+ 1.5 3 a 10.0 nM (-)BAY K8644
300 - 3 3 . 1 + 6.6 7.8 +_ 1.0 67.3 + 5.3 4.0 + 1.7 2 = 30.0 nM (-)BAY K 8 6 4 4
-300
Vos, Voltage (in mV) of half-maximal activation; k, steepness of
activation (in mV for e-fold change); V~. reversal potential (in Ca-current / pA
mV); Gin=, maximal whole-cell conductance (in nS); n, number
of experiments G / Gin,x

affects the Ca current itself nor the action of ( - ) BAY K 8644 on B

the Ca current. Due to the very limited area of illumination, which
0.75
was nearly identical to the focus area of the objective, the de-
stroyed nifedipine is rapidly exchanged by intact substance from
outside the focus area so that flashes can be repeated several times.
Nifedipine and ( - ) BAY K 8644 were dissolved in a stock
0.50
solution of 10 .3 g/mi in dimethylsulphoxide (DMSO) and then
subsequently diluted with Tyrode's solution so that the final
DMSO concentration was below 0.1%. This concentration does
not affect the stability of the cells or the Ca current.
0.25 9 control
v 1.7 nM (-)BAY K 8 6 4 4
o 10.0 nM (-)BAY K 8 6 4 4
Results = 30.0 nM (-)BAY K 8 6 4 4
0.00
Effects of ( - ) BAY K 8644 on peak Ca current i i i i i I

-60 -40 -20 0 20 40 60

Under the chosen experimental conditions, in which in-
activation is reduced and Ca channels are fully phos-
phorylated, exposure to the Ca agonist ( - ) BAY K 8644
increases peak Ca currents only at negative clamp poten-
tials between - 4 0 and 0 mV. At more positive mem-
brane voltages peak Ca currents are almost unchanged.
As the biggest increase in peak current was always
found at - 3 0 mV we chose this potential to measure a
dose/activity relation. Figure 1A shows the peak cur-
rents at clamp potentials of - 3 0 mV and 10 mV from
an experiment in which the whole cumulative dose/re-
sponse curve was measured in one cell Concentrations
of only 1 nM result in a doubling of the current ampli-
tude (see also original traces in Fig. 1B). Increasing the
concentration to 3 and 17 nM results in a further increase
in the peak Ca current. As an application of 80 nM did
not increase the peak current further we estimate an EDso
of about 3 nM for peak current increase at - 3 0 mV. For
a more complete analysis see below and Table 1.
The peak currents at a clamp potential of + 10 mV,
recorded at the same time, can be regarded as a control
for the stability of the Ca current measurement. Peak Ca
currents are not affected by ( - ) BAY K 8644 except for
a small decrease, not related to drug exposure (Fig. 1 A,
C). The latter is probably due to a "run down" phenom-
enon during the 2-h experiment.
membranepotential/ mV
Fig. 2. Current/voltage relations (A) and voltage-dependent acti-
vation curves (B) in the presence of increasing concentrations of
( - ) BAY K 8644. Current increase is strictly voltage dependent
Two remarkable observations were made routinely
during these experiments. First, it often takes about 1 0 -
15 min of drug exposure, especially at the very low con-
centrations, to reach equilibrium. This may be indicative
of the reported high partition coefficients between aque-
ous solution and the membrane phase [24]. Second,
when peak Ca currents became very large in the pres-
ence of ( - ) BAY K 8644 (see Fig. 1B) the inactivation
process was enhanced. A simple direct effect of the
agonist on inactivation or a facilitation of Ca-induced
inactivation by the increased Ca current can be excluded,
as an enhanced inactivation was not seen at more posi-
tive potentials (see + 10 mV in Fig. 1C).
Whole-cell current/voltage curves were recorded
(Fig. 2A) to obtain a more complete description of the
voltage dependence at low drug concentrations. Ca cur-
rents are only increased at negative membrane potentials
leading to a shift of the activation potential in the nega-
346
five direction. We analysed the current/voltage relation 8 - +10 mV
J T -70 mV ~ -40 rnV
by fitting the peak Ca current values to the function:
Ic,(V) = Gmax" Po(V) " (V-V~L) =" 8-
with an assumed single-channel open probability, '-~ 5 - 300

Po(V) = {1 + exp[(Vo.5-V)/k]} i/ l\T o

/I 'L o a) co,,trol
where V is the membrane potential, V~ the reversal po- .~ 4. ~./ - IX,, 9 b~ >30 .= ( - ~ ' , K8644
tential for the Ca current, G~,x the maximum cell con-
ductance, Vo.5 the potential of half-maximal activation
and k the steepness of the activation. Single-channel E 2.
open probability [Po(V)] is expected to be represented
by G(V)/G~x under whole-cell recording conditions (see 1

Fig. 2B). From these parameters increasing concen- 0

trations of ( - ) BAY K 8644 only affect the Vo.5 value. -;o-~o 2 . o - ; o - ; o - ; o - ~ o - ; o ; ,o ; ;
This can more easily be depicted from a plot of the volt- membrane potential/ rnV
age dependence of the whole-cell Ca conductance in
Fig. 2B. The continuous line in Fig. 2A was calculated Fig. 3. Plot of the relaxation constants for current activation or
deactivation versus clamp potential under control conditions (open
according to the given function. circles) and in the presence of a saturating concentration of (-)
Mean values from a series of dose/activity relations BAY K 8644 (filled circles; mean _+ SD from 6-8 experiments).
are included in Table 1. Increasing concentrations result Inset shows typical current traces under both conditions
in a progressive shift of the Vo.5 value without significant
effects on the other parameters of the current/voltage
curves. Half-maximal shift is found with concentrations Table 2. Summary of gating parameters
of about 3 nM and saturation is seen at concentrations
above 17 nM, resulting in a maximum shift of approxi- Gating properties Condition
mately 22 mV.
In a few experiments (data not shown), in which the Control (-) BAY
K 8644
Ca agonist wa applied without inclusion of cAMP to
the pipette solution also whole-cell conductance (G) was Peak current analysis: fi 9.7 8.4
increased. However, Ca current recordings were gener- P0.5 -9.1 -31.4
ally less stable, and control peak currents were subject Kinetic analysis: fi~ 23.8 24.4
to greater variations. Vc 17.8 17.8
rio -24.4 -22.7
Vo -49.6 -99.3
Ca current kinetics Po.5 -15.0 -43.0
( - ) BAY K 8644 not only affects the amplitude of the Gating charges: 2.2-3.0
peak currents, but also produces marked changes in the Energy transfer by
voltage-dependent current activation and deactivation (-) BAY K 8644: 5.4 kJ/mol (t.3 kcal/mol)
("tail cmncent") kinetics (e. g. Fig. 1 B). Therefore, time
constants of current activation and deactivation were The voltage constants 6 are given in mV for an e-fold change. Pos,
analysed over a broad membrane potential range by fit- is the voltage where the open probability is 0.5; Vc and Voare the
voltages where the single channel closed or open times are I ms
ting single exponentials to the current traces [4]. Time
constants from a group of experiments under control
conditions and with a high, saturating concentration of
( - ) BAY K 8644 have been plotted in Fig. 3. Values where Vo, Vc mark the voltages where open and closed
from - 3 0 mV to + 30 mV were calculated from the ac- times are 1 ms and Vo, Vc are the voltage constants for an
tivation time constants and values between 0 and e-fold change. Table 2 gives a comparison of the values
- 8 0 mV from the deactivation time constants. As- obtained from the fit in Fig. 3 and from the analysis of
suming first-order kinetics for the voltage-dependent peak currents.
gating process the relaxation time constant (T~) is given In the presence of ( - ) BAY K 8644 closed time con-
as: stants are not affected, but open times are prolonged
without any change in the voltage dependence. The
"c. (V) = [ 1 / T o ( v ) + 1 / r e ( v ) ] -~ function of [To(V)] of the open times is shifted to more
with To and Tc as the single-channel open and closed negative membrane potentials by about 50 mV. There-
time constants. On the basis of an exponential voltage fore, the shift of open probability, which is half as large,
dependence of both time constants the data could be fit- can be entirely attributed to this shift of the open time
ted well (continuous lines in Fig. 3) by: constants.
Voltage constants for both open and closed times
"co(V) = exp [(V-Vo)/Vo] (Table 2) are in good agreement with the k values from
Tc(V) = exp [(V-Vc)/vc] peak current analysis. From voltage constants we calcu-

+30 mV
-40 mV membrane potential
1000
100
1
10
o) control
b) 3 nM BAY K8644
c) 300 nM BAY K8644
1 0.1
-5 ; ; 1; 1;
time / me
Fig. 4. Comparison of tail current at different concentrations of
(-) BAY K 8644. Currents were normalized with respect to the
current level before repolarisation and plotted logarithmically
against time. At moderate concentrationsthe time course consists
of two exponentialcomponents:the fast control and the slow drug-
induced decay
late an equivalent of 2.2-3.0 elementary charges in-
volved in the gating process. The linear shift of the open
probability (approx. 23 mV) evoked by ( - ) BAY K
8644 can be interpreted as the amount of energy by
which the gating process is modified upon drug binding.
This energy can be estimated by multiplying the number
of elementary charges with the induced voltage shift re-
sulting in about 5.4 kJ/mol (1.3 kcal/mol) based on cur-
rent kinetics. This value is much less than the total bind-
ing energy which amounts to about 46 kJ/mol (11 kcai/
mol) (e. g. [23]).
Dose~activity relation
Dose/response curves of the increase in peak currents in
the presence of ( - ) BAY K 8644 suffer from the voltage
dependence of the drug effect, from changes in inacti-
vation and from run down phenomena during repeated
drug exposures. In order to obtain more precise infor-
mation we analysed the deactivation kinetics at negative
membrane potentials ( - 4 0 to - 8 0 mV) as a function of
drug concentration. The single Ca channel may be either
in the drug-free state having the short control open
times, or in the drug-bound state with the prolonged
open times. Drug concentration should only affect the
relative amount of channels with prolonged open times
but not the time constant itself. As a result tail currents
must consist of two components reflecting the two dif-
ferent open times, and the relative amount of slow com-
ponent should be a function of drug concentration.
Figure 4 shows a semilogarithmic plot of tail cur-
rents from one experiment under control conditions and
in the presence of a moderate (3 nM) and a high concen-
tration (300 nM) of ( - ) BAY K 8644. For a better com-
parison the plots are normalized with respect to pre-volt-
347
100
80
60
~
o 40 • n =1.25
zo
0
i i i i
1 10 100
[(-)BAY K8644]/nM
20
Fig. 5. Dose/response curve for (-) BAY K 8644. Percentage of
the slow, drug-induced component is plotted versus the concen-
tration
age step current. Under control conditions and at the
high drug concentration the time course is mona-ex-
ponential, but at the moderate concentration it is bi-ex-
ponential, containing the time constants of control and
saturating drug concentrations respectively. In Fig. 5
mean values (+_ SD) of the percentage of the slow com-
ponent are plotted against the drug concentration. From
the fit a KD value of 4.3 nM and a Hill coefficient of
1.25 were determined.
Kinetics of drug binding
So far rate constants for the binding reaction have been
measured only be means of receptor binding studies in
membrane preparations [23] or in intact cells [37]. Rate
constants predicted from models based on electrophysio-
logical data [3, 22] are much higher, suggesting that the
real association rate in receptor binding studies might be
obscured by a secondary slower rate-limiting step such
as partitioning into the lipid membrane phase. To mea-
sure the true association and dissociation rates we per-
formed a new type of competition experiment which
makes use of the light sensitivity of nifedipine to pro-
duce a sudden increase of the number of free DHP re-
ceptors. Nifedipine can be destroyed within a fraction of
a millisecond by light flash, resulting in an unblocking
of Ca channels and increase of Ca current [11, 27].
When using a combination of nifedipine and ( - ) BAY
K 8644 the energy of the flash can be adjusted to photo-
lyse only nifedipine without affecting the agonist. After
selective removal of nifedipine ( - ) BAY K 8644 is free
to interact with the unblocked Ca channels and this inter-
action is subsequently monitored by an increasing slow
component of the tail current.
Figure 6 shows the results from control experiments
with nifedipine alone. Photoremoval of nifedipine results
in a biphasic Ca current increase, a rapid increase during
the clamp pulse during which the flash was given
(Fig. 6A top, trace b) and a secondary slower phase
348

A +10 mV
A I 17 .M (-)~u KBe++ I
- 7 0 mV I - 5 0 mV membrone potentiol
I + 300 nM Nifedipine J
0
<= -250 v~vv~. . . . . . . . . . . .
WVW~
-50
,,v
I
6 -100 -500

; 1'o 20 30 4'0 v

-40 g
time / ms T T
I flash flesh
-750
~. -60 V i~~ V

-80
-1000 + i , , ,
-100 0 200 400 1500 800
-'5 ; time / s
-'10
time I s5 10 15
+10 mV
flesh Nifedipine: 100 nM B - 7 0 mV
.~ ] - 3 5 mV membrane o o t e n t i o l

B 1 ,=16s 0 flash
<=
9 eeeeee 9

//
-250
c

uI -500
--~ 9

-750 i i i i = i
"~ -50
0 10 20 30 40 50
a 1 flash time / ms

§ -200 1

I. J 9 d ~ qvv~vvvv~ovvv~
-~oo l ~b vvvv v
-100 **; 9 vv
-400 .. WV
g, ] %~,~v"
-500
-;o ; 2'0 ;o 6'O 8'O 0
i
25
i
50
i
75
i
1O0
time / s
time / s
Fig. 6. Flash photolysis of nifedipine results in an unblocking of
Fig.TA, B. Flash photolysis of nifedipine in the presence of ( - )
Ca channels with a very fast (A, top trace b) and a slower (A,
BAY K 8644. After a flash peak currents are transiently increased
bottom) current increase. Normalized peak currents thereby show
(overview in A). At the same time the percentage of the slow,
a transient increase with an exponential decay to control values
(B) drug-induced tail component rises with time (B, top, traces b - d ) .
When tail currents from traces b and d are compared this change
becomes were obvious

(Fig. 6A bottom, between points b and c). The rapid

phase with a time constant of about 3 ms is comparable
to the voltage-dependent activation of the channels and During flash experiments with nifedipine alone no
obviously arises from channels being blocked in the re- changes in the kinetics of the tail current could be de-
sted state [11]. The slower phase, with a time constant tected. Control experiments with ( - ) BAY K 8644 (data
of about 200 ms, typical for the rate of recovery from not shown) show no effect of the flash on either the
inactivation, indicates that those channels were blocked amplitude of the peak inward current or the kinetics of
in the inactivated state. Under these experimental con- the tail current.
ditions (holding potential - 7 0 mV, nifedipine concen- In Fig. 7A the peak currents from an experiment
tration 300 nM) the amplitude of the rapid phase is about with the combined application of nifedipine and ( - )
3 0 - 4 0 % of total current increase. BAY K 8644 are plotted. Nifedipine given in the con-
The current increase evoked by the flash is transient, tinuous presence of the agonist blocked the peak current
as can be seen in Fig. 6B. This is due to our experimen- by about 75%. As with nifedipine alone, flashes evoke
tal setup where the flash is delivered through the objec- a transient increase of the peak current. However, due
tive of the inverted microscope. Nifedipine is only de- to run down control peak current values could not be
stroyed within the focus area of the objective (diameter established during these transient increases.
approx. 250 gm) leaving most of the drug in the bath Four original current traces from one transient in-
unaffected. Therefore, following the flash, unchanged crease (peak current overview in Fig. 7B bottom) after
nifedipine diffuses back into the focus area and reblocks a flash are shown in Fig. 7B (top). Trace a was recorded
the Ca channels. The time constant for this diffusion- before the flash was given, trace b is the current trace
controlled reblocking was about 16 s. where the application of the flash (5 ms after depolari-
 349
A IO0 ]
sation) results in the sudden current increase, trace c was
0=;sM /~',,~ 17 nM
recorded at the peak of transient and trace d was taken
50 s later. The latter was selected because the current z~ 80
/ =b
/ "r=14s

amplitude during the clamp pulse is identical with that

in trace b, so tail currents can be compared directly. The
tail current in trace a follows a single exponential time i
=
D
6o J
course with a time constant typical for ( - ) BAY-K-
8644-modified channels (see above). Ca channels which ~ 40
were not blocked by nifedipine had obviously bound ( - )
BAY K 8644, because of the nearly saturating concen-
tration (17nM) used in this experiment (compare 20
Fig. 1B).
In trace b the tail current, which is recorded 10 ms
0
after the flash, consists of two components as described i i i i
0 " 10 20 30
earlier for conditions in which submaximal concen- time / s
trations of the agonist were studied. The fast component
has the same time constant as that under control con- B ~ flash
ditions and the slow component is identical to that in 9 9 $ Nifedipine ~'=16 s
trace a. Therefore, channels which were unblocked dur- O, 9 9 9
ing the flash did not bind a measurable amount of ( - )
BAY K 8644 during this short clamp step. However, at .E
a very late point during the transient increase, when _E
trace d was recorded, the unblocked channels have -50
bound the agonist, as can be inferred from the kinetics
of the tail current which now shows a single, slowly I
decaying component. o
In a series of experiments we measured the associ- -100 -
ation of ( - ) BAY K 8644 with the Ca channels by ana- 8/~ ~176(-)BAY K8644 + Nifedipine: "r=70 s
lysing the time-dependent increase of the slow compo-
nent of tail current after a flash. This analysis was re- i i i i i i
-40 -20 0 20 40 60 80
stricted to the flash-induced current by subtracting the time / a
Ca current traces recorded prior to the flash (e. g. trace a
in Fig. 7B). The results for three different concentrations Fig. 8. A Plot of the slow drug-induced component versus time
after the flash in the presence of different concentrations of ( - )
of ( - ) BAY K 8644 are plotted against the time after BAY K 8644. Time constants were estimated from exponential
the flash (Fig. 8A). The determined association time fits. B Comparison of the normalized transient peak current in-
constants show a linear concentration dependence as ex- crease from experiments with nifedipine alone (filled squares) and
pected for such a process. From the time constants given in combination with ( - ) BAY K 8644 (open circles). Continous
in the figure a mean association constant (ka) of 5 lines represent exponential fits (data from several different flashes)
X10 -6 mol-ts -1 was calculated. Taking the K , value and
the association constants together allows the estimation
of the dissociation time constant: 50 s. Discussion
The dissociation time constant can also be measured
directly from the reblocking rate during a flash-induced Ca agonists action on peak currents
transient peak current increase. When both nifedipine
and ( - ) BAY K 8644 are present the reblocking by back The results of our study present a framework for under-
diffusion of nifedipine should be hindered by the Ca standing the molecular interaction of the DHP Ca ago-
agonist occupying the receptor. Under these conditions nist with the Ca channel. As some of the previous re-
the dissociation of ( - ) BAY K 8644 becomes the rate- ports on the electrophysiological actions of these drugs
limiting process. Figure 8B compares the peak current were controversal, our aim was to investigate the pri-
increase for both nifedipine alone and the combined ap- mary and most direct action of ( - ) BAY K 8644, na-
plication of the agonist and the antagonist. Therefore mely the changes in gating behaviour, without any inter-
changes in the peak current were normalized with re- ference from secondary effects brought about by the in-
spect to the maximum increase in each experiment and creased Ca entry into the cells. Those secondary effects
plotted versus the time after a flash. Data points were might be a Ca-induced inactivation [2], which may dim-
taken from 3 - 4 experiments under each condition. The inish the peak current increase, and a Ca-induced in-
reblocking rate is clearly slowed in the presence of ( - ) crease in the number of available channels [8, 12], which
BAY K 8644 resulting in a time constant of 70 s. Taking may lead to an overestimation of the current increase
into account that this reblocking time constant tends to and to a "false" voltage dependence.
overestimate the dissociation time constant, because of In general, two distinct changes of the peak current/
a repeated binding of the agonist, there is a good agree- voltage relation have been reported for Ca agonistic
ment with the above given calculated time constant. DHPs: (1) a shift of the current activation to more nega-
350
tive membrane voltages and the concurrent shift of the
peak voltage of the current/voltage curve [19, 33] and,
(2) a voltage-independent increase of current amplitudes
over the whole voltage range resulting in an increased
steepness of the linear part of the current/voltage curve
[14, 33].
We have never observed the latter finding under our
chosen experimental conditions; therefore the voltage-
independent increase of current is probably a secondary
action of the agonists, mediated by increased intracellu-
lar Ca levels. This is in line with observations that pho-
tochemically released intracellular Ca is able to increase
Ca currents in a voltage-independent manner, an effect
which can be blocked by prior application of isopren-
aline [12] and which is most likely produced by an en-
hanced phosphorylation of Ca channels through the Ca-
dependent kinases.
From the complete analysis of the current/voltage re-
lations at different concentrations we conclude that the
only direct action of the Ca agonist on the Ca channel
is a shift of the voltage-dependent activation. This re-
sults in an increased single-channel open probability;
single-channel conductance, number of available chan-
nels and the current reversal potential are unchanged.
This view is well supported by various single-channel
studies [6, 16, 29].
The observation that Ca current increase at very
positive membrane potentials might be due to a second-
ary effect governed by an enhanced phosphorylation is
of great value for recently published studies on the ex-
pression of cloned Ca channels. As the cells used in such
expression studies may have very different activities of
kinases and phosphatases, the Ca-induced increase in
single-channel availability may be absent. So the ex-
pected increase in the Ca current in the presence of BAY
K 8644, which is taken as a proof for DHP-sensitive
channel types, may be lower than anticipated [30, 36,
38], or might be overlooked because of a too positive
clamp potential [28].
Most reports on the dose/response relationship of
( - ) BAY K 8644 failed to show the high-affinity action
expected from receptor binding studies [14, 33]. As peak
current increase is strikingly voltage dependent a reliable
dose/response relation can only be measured at the most
sensitive clamp potential (at the start point of activation
e.g. at - 3 0 mV), or as the drug-induced shift of the
activation curve. Both procedures yield, if equilibration
times for the drug are long enough, and if inactivation
and run down is negligible, reasonable values for half-
maximal activity of about 3 - 5 nM. The values are also
in good agreement with the dose/response relation de-
rived from the independent analysis of the current kin-
etics.
With non-saturating concentrations the activation
curve should be the sum of the activation curves for the
drug-bound and drug-free channel populations. Because
of an enhanced inactivation process in the presence of
the agonist which was predominantly abvious from - 5 0
to - 1 0 mV (compare Fig. 1B, 30 nM trace) we have not
succeeded in resolving these two components. However,
there was always a tendency for a decreased steepness
of the curve compared to control or high drug conditions
which can be taken as an indication for the predicted
"mixed-type" of activation curve.
Changes in current kinetics
Assessment of the activation and deactivation kinetics
of the Ca current is one of the most direct ways to ana-
lyse the gating behaviour of single Ca channels under
physiological ion concentrations, where a single-channel
study is impossible [10]. When secondary processes of
the channel regulation like inactivation or high rates of
phosphorylation, which may obscure the true voltage-
dependent gating properties, are reduced as far as possi-
ble, voltage-dependent activation and deactivation can
be reasonably well described by first order kinetics.
As there are deviations from a single exponential
time course - especially at the beginning of activation
[3, 10] because of the existence of several closed states
[10] - the resulting time constants should be taken as
an approximation. However, several points indicate that
this approach yields very reliable values:
1. Time constants in the potential range from - 4 0
to - 1 0 m V can be checked independently from acti-
vation and deactivation.
2. Independent fits of the relaxation time constants
under control and high-drug conditions result in the
same functions for both open and closed times, except
for the linear shift in open times.
3. Voltage dependence and half-maximal activation
voltage of peak currents and exponential voltage depen-
dence of time constants are in good agreement, support-
ing a first-order process.
4. Effects of ( - ) BAY K 8644 on peak currents and
on gating properties of the channel are in complete
agreement.
Most striking is the latter point: the complete
congruence of peak current analysis with kinetic changes
of current activation and deactivation. Voltage constants
calculated from both independent measurements are in
agreement with an involvement of 2 . 2 - 3 . 0 elementary
charges in the gating process. The observed shift of acti-
vation curves by about 23 mV upon binding of ( - ) BAY
K 8644 can be attributed to the approximately twice as
large shift of the single-channel open times while single-
channel closed times remain unaffected. Due to this ap-
proximate 50 mV shift of the open times function along
the voltage axis without any change in its voltage depen-
dency, open times at every membrane voltage are in-
creased by a constant factor. However, this prolongation
of open times will only result in an increased Ca current
when open probability is much less than 1. If, on the
other hand, open probability is close to 1, as in the range
of voltages positive from about + 10 mV, no change in
the amplitude of Ca current will be seen. Therefore, the
drug action on the current amplitude must be strictly
voltage dependent.
The shift of open times as a function of voltage can
be considered to be the result of a constant energy trans-
fer by the drug, which stabilizes the open channel state,

and which can be overcome by applying more negative
voltages to obtain the same closing rates. This energy
can be calculated as the product of the gating charges
and the drug-induced voltage shift, which gives 5.4 kJ/
mol (1.3 kcal/mol). The energy needed to modify gating
therefore amounts only to 11% of the estimated total
binding energy of 48 kJ/mol (11.5 kcal/mol) [23].
Binding kinetics of ( - ) BAY K 8644
Although the slowing of deactivation time course ("tail
current prolongation") has been noted in most of the
studies with Ca agonists (e. g. [16, 33]) only two at-
tempts have been undertaken to quantify the changes [9,
22]. The present paper contains the first complete pub-
lished analysis of tail currents including both the volt-
age- and the concentration-dependent changes.
From the distinct changes in the gating kinetics and
from the discrete prolongation of open times upon drug
binding it is clear that non-saturating drug concen-
trations must result in biphasic tail currents. Tail currents
must contain a fast component, characteristic for control
measurements, and a slow component, which reflects the
population of drug-bound channels. The relative amount
of both components should vary with the drug concen-
tration without any change in the time constants of the
components, because binding to the single channel is a
discrete event excluding the occurrence of intermediate
time constants. A concentration dependence of the time
constants, as reported by Lacerda and Brown [22], is
therefore very unlikely.
The dose/response curve for ( - ) BAY K 8644 from
tail current analysis presents unequivocal evidence that
the high-affinity receptor found in binding studies
(2.5 nM in [23]; 0 . 5 - 3 nM in [37], 3.5 nM in [39]) is
also the only functional and pharmacologically relevant
binding site. Differences between KD values from bind-
ing studies and from electrophysiological investigations
reported elsewhere [14, 33] are most probably attributed
to the inaccuracy of the drug-induced peak current
changes.
It has been proposed that Ca agonists preferentially
bind and interact with the open state of the Ca channel
[3, 22]. However, a state-independent, non-modulated
receptor hypothesis has also been suggested [15]. From
receptor binding and functional studies it is known that
Ca antagonists have a strongly voltage-dependent bind-
ing behaviour [1, 21, 39] whereas binding of the Ca
agonists was found to be voltage independent [37, 39].
Photolysis of nifedipine has been used in several
studies to prove the state-dependent interaction of this
antagonist with the Ca channel [11, 27]. Our control
experiments, with the photo-induced removal of nifedi-
pine alone, are in line with those reports. Recovery of
the Ca current is biphasic with a rapid phase (time con-
stant approx. 3 ms) for channels which are blocked in
the rested state and a slower phase (time constant ap-
prox. 150 ms) arising from channels arrested in the inac-
tivated state.
351
In the presence of the Ca agonists the recovery rate
of the current after flash photolysis of nifedipine is un-
changed, but tail current analysis reveals that with time
the fraction of drug-bound channels increases and as-
sociation rates are dependent on the concentration of the
Ca agonist. These experiments clearly show, in func-
tional studies, that DHP agonists and antagonists com-
pete for the same high-affinity receptor site and that
binding of one will completely prohibit the binding and
action of the other. These results present a strong argu-
ment against multiple and/or different binding sites for
Ca agonists and antagonists [21, 22].
The association rate of 5)<106 mol-ls -1 for ( - ) BAY
K 8644 is a factor of 3 - 4 higher than those from recep-
tor binding studies [23, 39]. Both calculated and meas-
ured dissociation time constants of 5 0 - 7 0 s are also
quite similar to those found in binding studies [23]. Not
only receptor affinity, but also binding kinetics, are iden-
tical in both functional electrophysiological and in re-
ceptor binding studies.
In the present experiments the membrane voltage
was maintained at a resting potential of - 7 0 mV and
depolarizing steps last only for about 1% of the time.
Under this condition the Ca channel exists almost exlu-
sively in the resting state, therefore the binding parame-
ters characterize the interaction with the resting state of
the Ca channel. Obviously the interaction with the inac-
tivated state, as monitored in binding studies [23, 39],
is not much different. Taken together, a non-modulated
receptor hypothesis [15] seemed to be appropriate for
a description of the agonists action on the Ca channel.
However, gating analysis has shown that additional en-
ergy - amounting to 5.4 kJ/mol (1.3 kcal/mol) or to a
change of KD value of a factor of 10 -- is transferred
solely to the open channel state. Whether this energy
either stabilizes or destabilizes the binding of the Ca
agonist with the receptor is still unclear. Due to the
general problem of increasing the fraction of open
channels for long times without any inactivation experi-
ments to answer this question cannot easily be obtained.
Models for the agonist's action
Models based on the assumption of an exclusive binding
[3], or a preferential binding [22] to the open state with
proposed associations rates up to 10 l~ mol-ls -1 are not
in agreement with the present results. In addition, this
study, as with most of the receptor binding studies re-
ported so far, provides no indication for multiple binding
sites as proposed in the model of Lacerda and Brown
[22].
Upon binding the agonist may either change only the
closing rate constants as previously reported by Sangui-
netti et al. [33] or both the opening and closing rates.
We see no evidence for changes of the opening rates as
reported by Hering et al. [15]. Evidence for a change in
opening rates has also come from the analysis of first
opening waiting times and from closed times distri-
butions [22]. However, because of the limited time resol-
ution, short openings, especially under control con-
352
ditions, may have been overlooked and a prolongation
of the open times and an overall increased open prob-
ability in the presence of the Ca agonists must therefore
result in an apparent decrease of waiting and closed
times.
Our analysis of the current kinetics over a broad
range of membrane potentials provides strong evidence
against a change in opening rates. The following scheme
summarizes our results:
KI
C ~K---I O "Nfci
xd ]l r~ I order kinetics which was never observed in our experi-
K1 D /KID
CD ~K-1D OD
where C, O, I, CD and 0D denote the closed, open,
inactivated and drug-bound closed and open states re-
spectively, with
K~(V) = K~D(V) = 10 3 e (V-~8~'24 s -~
K_I(V) = ]03 e -(v+50)/24 s -1
K_ID(V) = ] 0 3 e-(V+99)/24 S-1
where V is given in mV. The association rate KA is about
5Xl06mo1-1 s -~ and dissociation rate Kd = 0.02s ~.
Transition between O and OD can be neglected under
our experimental conditions, because of the rather short
live-time of the open state.
Our experiments yield no information about the inac-
tivation process. We assume that inactivation rates Ki
and KiD are unchanged and expect inactivation to be in-
fluenced only indirectly because of the prolongation of
open-state live-times and increased Ca entry as proposed
by Sanguinetti et al. [33]. However, as inactivation is
also subject to many other regulatory processes, includ-
ing integral intraceHular Ca concentration and channel
phosphorylation, a simple mathematical description is
not feasible.
As transitions from the drug-free to drug-bound
states are rather slow compared to the voltage-dependent
gating, our model predicts a single-channel behaviour
for a given clamp potential which is phenomenologically
similar to the mode model of Hess et al. [16]. But certain
essential characteristics are not fulfilled: 1. We have
found no evidence for a slow- ("mode 2") gating behav-
iour under control conditions. 2. We have evidence that
the prolongation of open times is not governed by an
intrinsic mechanism, but is very substance specific and
may therefore be different for various Ca agonist types.
Implications for the structure of the Ca channel
Recently attempts have been undertaken to localize the
DHP receptor region within the a~ subunit of the Ca
channel [31]. Due to uncertainties in the predicted fold-
ing pattern of the az subunit the location of the DHP
receptor is somewhat controversal. Regulla et al. [31]
favour the cytoplasmic side and Catterall and Striessnig
[7] propose a position on the extracellular surface of the
membrane.
Additional information on the orientation of the DHP
receptor has come from studies with charged DHPs [20]
favouring an extracellular side. The association rate
measured by flash-induced competition experiments
characterizes the binding kinetics under equilibrium con-
ditions, e.g. diffusion, membrane permeation and ac-
cumulation processes having reached steady state. Rates
calculated from this approach are about a factor of 10
higher than those derived from experiments which make
use of rapid perfusion techniques [15, 25] for drug appli-
cation. In addition rapid application of the DHP results
in a delay of up to 1 s and a response following second-
ments. An explanation for these obvious discrepancies
could be due to the fact that in the case of rapid drug
application the Ca agonist has first to equilibrate with
the lipid membrane before getting access to the DHP
receptor.
Although this interpretation fits very well the pre-
viously suggested hypothesis of a "membrane bilayer
pathway" [24], the determined association rate seems to
be very reasonable for a diffusion-limited access from
the aqueous phase to the receptor but too low for the
two dimensional membrane compartment. Taking into
account the high partitioning coefficient of about 104,
the real rate, based on the membrane concentration, is
even slower. Part of this discrepancy may be due to the
fact that the DHPs are not evenly distributed in the mem-
brane phase [24] and that the receptor and the DHP mol-
ecules are not located at the same position.
Structure/activity relations show that the only differ-
ence between agonists and antagonists is the rather small
substituent in the 3' position, which might be either an
ester function for nifedipine or a nitro-group for BAY K
8644. This small modification has a large and distinct
impact on the state transitions of the comparably large
protein molecule.
1. Agonists stabilize the open state without affecting the
other states.
2. Binding of the agonist, however, takes place in a
state- and thereby voltage-independent manner.
3. Open state stabilization by the agonist is due to a
rather low energy transfer in the order of a single
hydrogen bond.
4. The antagonist, on the other hand, stabilize the inacti-
vated and the resting state [32].
5. Binding of the antagonist is strictly voltage and state
dependent.
Taking these results together it becomes clear that the
substituent in the 3' position must be very closely
coupled to one of the key positions of the channel pro-
tein. Identification of the amino acid residues interacting
with this side group will therefore greatly help to en-
lighten the gating and inactivation mechanisms of the
Ca channel.
Acknowledgement. We wouldlike to thank R. Noll and G. Schmidt
for their technical assistance and Prof. R.Grog, Dr. E Rounding
and Dr. M. Schramm for their critical examination of the manu-
script and their helpful advice.

References
1. Bean BP (1984) Nitrendipine block of cardiac calcium chan-
nels: high affinity binding to the inactivated state. Proc Natl
Acad Sci USA 81:6388-6392
2. Bechem M, Pott L (1985) Removal of Ca-current inactivation
in dialyzed guinea-pig atrial cardioballs by Ca chelators.
Pfltigers Arch 404:1020
3. Bechem M, Schramm M (1987) Calcium-agonists. J Mol Cell
Cardiol 19 [Suppl 11] : 6 3 - 7 5
4. Bechem M, Hebisch S, Schramm M (1988) Ca2+ agonists:
new, sensitive probes for Ca 2+ channels. Trends Pharmacol
Sci 9 : 257-261
5. Biel M, Ruth P, Bosse E, Hullin R, Sttihmer W, Flockerzi V,
Hofmann F (1990) Primary structure and functional ex-
pression of a high voltage activated calcium channel from rab-
bit lung. FEBS Lett 269:409-412
6. Brown AM, Kunze DL, Yatani A (1986) Dual effects of di-
hydropyridines on whole cell and unitary calcium currents in
single ventricular cells of guinea-pig. J Physiol (Lond)
379 : 495-514
7. Catterall WA, Striessnig J (1992) Receptor sites for Ca z+
channel antagonists. Trends Pharmacol Sci 13:256-262
8. Charnet P, Richard S, Gumey AM, Ouadid H, Tiaho F, Nar-
geot J (1991) Modulation of Ca currents in isolated frog atrial
cells studied with photosensitive probes. Regulation by cAMP
and Ca2+: a common pathway? J Mol Cell Cardiol 2 3 : 3 4 3 -
356
9. Droogmans G, Callewaert G (1986) Ca~+-channel current and
its modification by the dihydropyridine agonist BAY K 8644
in isolated smooth muscle cells. Pfiiigers Arch 406 : 259-265
10. Fenwick EM, Marry A, Neher E (1982) Sodium and calcium
channels in bovine chromaffin cells. J Physiol (Lond)
331 : 599-635
11. Guruey AM, Nerbonne JM, Lester HA (1985) Photoinduced
removal of nifedipine reveals mechanisms of calcium antag-
onist action on single heart cells. J Physiol (Paris) 8 6 : 3 5 3 -
379
12. Guruey AM, Charnet P, Pye JM, Nargeot J (1989) Aug-
mentation of cardiac calcium current by flash photolysis of
intracellular caged-Caa+ molecules. Nature 341 : 6 5 - 6 8
13. Hamill OP, Marty A, Neher E, Sakman B, Sigworth FJ (1981)
Improved patch-clamp techniques for high resolution current
recording from cells and cell free membrane patches. Pfl~igers
Arch 391:85-100
14. Hamilton SL, Yatani A, Brush K, Schwartz A, Brown AM
(1987) A comparison between the binding and electrophysio-
logical effects of dihydropyridines on cardiac membranes. Mol
Pharmacol 31 : 221-231
15. Hering S, Kleppisch T, Timin EN, Bodewei R (1989) Charac-
terization of the calcium channel state transitions induced by
the enantiomers of the 1,4-dihydropyridine Sandoz 202 791
in neonatal rat heart cells. Pfltigers Arch 414:690-700
16. Hess P, Lansman JB, Tsien RW (1984) Different modes of
Ca channel gating behaviour favoured by dihydropyridine Ca
agonists and antagonists. Nature 311 : 538-544
17. Janis RA, Triggle DJ (1991) Drugs acting on calcium chan-
nels. In: Hurwitz LH, Partridge LD, Leach JK (eds) Calcium
channels: their properties, functions regulation and clinical
relevance. CRC Press, Boca Raton, pp 195-249
18. Kamp TJ, Sanguinetti MC, Miller RJ (1989) Voltage- and use-
dependent modulation of cardiac calcium channels by the di-
hydropyridine (+)-202-791. Circ Res 64:338-351
19. Kass RS (1987) Voltage-dependent modulation of cardiac cal-
cium channel current by optical isomers of BAY K 8644: im-
plications for channel gating. Circ Res 61 [Suppl I] : 1 1 - 1 5
20. Kass RS, Arena JP, Chin S (1991) Block of L-type calcium
channels by charged dihydropyridines. J Gen Physiol 98 : 6 3 -
75
21. Kokubnn S, Prod'hum B, Becket C, Porzig H, Reuter H
(1986) Studies on Ca channels in intact cardiac cells: voltage-
353
dependent effects and cooperative interactions of dihydropyri-
dine enantiomers. Mol Pharmacol 30:571-584
22. Lacerda AE, Brown AM (1989) Nonmodal gating of cardiac
calcium channels as revealed by dihydropyridines. J Gen Phy-
siol 93 : 1243-1273
23. Maan AC, Hosey MM (1987) Analysis of the properties of
binding of calcium-channel activators and inhibitors to di-
hydropyfidine receptors in chick heart membranes. Circ Res
61 : 3 7 9 - 388
24. Mason RP, Gonye GE, Chester DW, Herbette LG (1989) Par-
titioning and location of BAY K 8644, 1,4-dihydropyridine
calcium channel agonist, in model and biological membranes.
Biophys J 55 : 769-778
25. M6ry PF, Lechene P, Fischmeister R (1992) A loudspeaker-
driven system for rapid and multiple solution exchanges in
patch-clamp experiments. Pfliigers Arch 420:529-535
26. Mikami A, Imoto K, Tanabe T, Niidome T, Mori Y, Takeshima
H, Narumiya S, Numa S (1989) Primary structure and func-
tional expression of the cardiac dihydropyridine-sensitive cal-
cium channel. Nature 340: 230-232
27. Morad M, Goldman YE, Trentham D (1983) Rapid photo-
chemical inactivation of Ca2+-antagonists shows that Ca 2§ en-
try directly activates contraction in frog heart. Nature
304: 635-638
28. Mori Y, Friedrich T, Kim MS, Mikami A, Nakai J, Ruth P,
Bosse E, Hofmann F, Flockerzi V, Furuichi T, Mikoshiba K,
Imoto K, Tanabe T, Numa S (1991) Primary structure and
functional expression from complementary DNA of a brain
calcium channel. Nature 350:398-402
29. Ochi R, Hino N, Niima Y (1984) Prolongation of calcium
channel open time by the dihydropyridine derivative BAY K
8644 in cardiac myocytes. Proc Jpn Acad Ser B Phys Biol Sci
15 : 153-156
30. Perez-Reyes E, Castellano A, Kim HS, Bertrand P, Baggstrom
E, Lacerda AE, Wei X, Birubaumer L (1992) Cloning and
expression of a cardiac/brain fl subunit of the L-type calcium
channel. J Biol Chem 267:1792-1797
31. Regulla S, Schneider T, Nastalnczyk W, Meyer HE, Hofmann
F (1991) Identification of the site of interaction of the dihydro-
pyridine channel blockers nitrendipine and azidopine with the
calcium-channel al subunit. EMBO J 1 0 : 4 5 - 4 9
32. Sangninetti MC, Kass RS (1984) Voltage-dependent block of
calcium channel current in the calf cardiac Purkinje fiber by
dihydropyridine calcium channel antagonists. Circ Res
55 : 336-348
33. Sanguinetti MS, Krafte DS, Kass RS (1986) Voltage-depen-
dent modulation of Ca channel current by BAY K 8644. J Gen
Physiol 88:369-392
34. Schramm M, Thomas G, Towart R, Franckowiak G (1983)
Novel dihydropyfidines with positive inotropic action through
activation of Ca channels. Nature 303 : 535
35. Tsien RW, Ellinor PT, Home WA (1991) Molecular diversity
of voltage dependent Ca +§ channels. Trends Pharmacol Sci
12: 3 4 9 - 354
36. Varadi G, Lory P, Schultz D, Varadi M, Schwartz A (1991)
Acceleration of activation and inactivation by the fl subnnit of
the skeletal muscle calcium channel. Nature 352:159-162
37. Wei XY, Rutledge A, Triggle DJ (1988) Voltage-dependent
binding of 1,4-dihydropyridine Ca 2§ channel antagonists and
activators in cultured neonatal rat ventricular myocytes. Mol
Pharmacol 35 : 541-552
38. Williams ME, Feldman DH, McCue AE Brenner R, Velicelebi
G, Ellis SB, Harpold MM (1992) S~ucture and functional ex-
pression of al, a2, and fl subunits of a novel human neuronal
calcium channel subtype. Neuron 8 : 7 1 - 8 4
39. Zheng W, Stoltefuss J, Goldmann S, Triggle DJ (1992)
Pharmacologic and radioligand binding studies of 1,4-di-
hydropyridines in rat cardiac and vascular preparations:
stereoselectivity and voltage dependence of antagonist and ac-
tivator interactions. Mol Pharmacol 41:535-541