Review |
From the Department of Molecular Physiology, National Cardiovascular Center Research Institute, Suita, Osaka, Japan.
Correspondence to Munekazu Shigekawa, MD, PhD, Department of Molecular Physiology, National Cardiovascular Center Research Institute, Fujishiro-dai 5-7, Suita, Osaka 565-8565, Japan. E-mail shigekaw{at}ri.ncvc.go.jp
| Abstract |
|---|
|
|
|---|
-1 and
-2 repeat structures and
their neighboring membrane-spanning segments. Finally, we discuss new
classes of NCX inhibitors with improved selectivity. One of
these, 2-[2-[4-(4-nitrobenzyloxy)phenyl]ethyl]isothiourea
methanesulfonate (KB-R7943), appears to exhibit unique selectivity for
Ca2+-influxmode NCX activity. Data
obtained with these inhibitors should provide a basis for
designing more selective and clinically useful drugs targeting
NCX.
Key Words: Na+-Ca2+ exchange Ca2+ transport cardiac muscle excitation-contraction coupling
| Introduction |
|---|
|
|
|---|
30%
of the Ca2+ required to activate the
myofilaments in rabbit, guinea pig, and human ventricles but a much
smaller portion (
7%) in rat and mouse ventricles. Sarcoplasmic
reticulum (SR) Ca2+-ATPase removes most of
the remaining Ca2+. In failing rabbit or
human heart, NCX and SR Ca2+-ATPase
contribute nearly equally to Ca2+ removal
from the cytoplasm.2 Such
differences in the NCX contribution mostly reflect differences in the
expression levels of NCX activity in the
sarcolemma,3 4 ,
Ca2+-ATPase activity in the
SR,4 5 and possibly
[Na+]i6
in cardiomyocytes from these animal species in normal and
disease conditions. The physiological significance of Ca2+ influx via NCX in the excitation-contraction coupling of cardiac muscle is controversial. Triggering the release of Ca2+ from the SR via NCX during membrane depolarization has been shown to be much less efficient than via L-type Ca2+ channels.7 8 9 However, when both L-type Ca2+ channels and NCX are at work, Ca2+ entry via NCX appears able to synergistically amplify the effect of triggering SR Ca2+ release via the L-type current.10 In heart failure, on the other hand, enhanced Ca2+ entry due to increased NCX expression may provide inotropic support for failing myocytes, in which the SR function is often defective.4 11 Under other pathological conditions, such as cardiac ischemia/reperfusion12 or digitalis intoxication,13 the NCX-mediated increase in Ca2+ entry or decrease in Ca2+ exit due to a rise in [Na+]i results in Ca2+ overloading of the SR, leading to mechanical and electrical dysfunction of myocytes.
Because the SR Ca2+ load, which is a predominant determinant of cardiac contractility,2 is determined by the competition between SR Ca2+-ATPase and NCX for the cytosolic Ca2+, modulation of NCX activity by physiological regulatory factors as well as by alterations in cytosolic ion concentrations and the action potential duration in disease states exerts profound influences on the overall contractile function of the heart. In the present review, we describe recent advances in molecular and pharmacological studies of the cardiac NCX function. The mechanism and further physiological and pathological aspects of NCX function have been discussed in several recent reviews14 15 16 and short comments.17 18
| Cardiac NCX |
|---|
|
|
|---|
70% amino acid identity in the overall sequences and thus
presumably have a very similar molecular structure. NCX1 is the first
NCX cloned and is highly expressed in cardiac muscle and brain and to a
lesser extent in many other tissues. NCX2 and NCX3 are not expressed in
adult rat heart at the protein level, but an NCX2-specific transcript
can be detected faintly by using reverse transcriptasepolymerase
chain reaction.22 NCX2 and
NCX3 are expressed in a few limited tissues, such as brain, and their
molecular properties and functions remain unclear. Besides these
isoforms, splice variants with variation in a small region of the large
central loop of the exchanger molecule are generated from NCX1 and NCX3
genes in a tissue-specific
manner.22 23
However, the physiological significance of such
diversity is not clear, although some splice variants of NCX1 seem to
exhibit the regulatory property
differences.24 25
In heart, a splice variant of NCX1, designated NCX1.1, is predominantly
expressed.22
NCX1 gene expression occurs early in cardiogenesis in the
mouse embryo, before the onset of ventricular myosin light
chain 2 expression and before the occurrence of the first heart
beat.26 It is expressed in a
heart-restricted pattern through the critical early stages of heart
development until at least 11 days postcoitus (dpc). Recent targeted
disruption of the NCX1 gene revealed that NCX1-deficient
(Ncx1-/-)
embryonic mice died between 9 and 10
dpc.27 In the case of
Ncx1-/-
embryos at 9.5 dpc,
70% of them showed no heartbeat, whereas the
remainder exhibited only very slow and arrhythmic heart contraction.
The ventricular wall was very thin and contained few
myocytes in
Ncx1-/-
mice compared with Ncx1+/+ mice, although
there was apparently no defect in the expression of heart-specific
genes in the ventricles of
Ncx1-/-
mice. Myocytes displayed morphological signs of apoptosis. In
the normal mammalian heart, NCX1 expression reaches a maximum near
birth and then decreases postnatally to a significantly lower level in
the adult stages.28 This is
in contrast to SR Ca2+-ATPase, which is
increasingly expressed postnatally. In rat heart, thyroid hormones play
an important role in the reciprocal control of the expression of these
2 Ca2+
transporters.29 Thus, the
contribution of NCX to the control of
[Ca2+]i is greater
in the immature heart than in the mature heart.
NCX is a high-capacity and
lowCa2+-affinity transporter exchanging 3
Na+ and 1 Ca2+
across the plasma
membrane7 14 30
(however, see a recent study by Fujioka et
al31 reporting coupling
ratios >3 to 1). The maximum turnover numbers (up to 5000 per
second)32 33 for
NCX1 and its
Km for
intracellular Ca2+ (
6
µmol/L)34 35 36
are much greater than those for SR
Ca2+-ATPase (100 to 150 per second for
Ca2+ transport calculated from the activity
of purified ATPase at
37°C37 and
0.3 µmol/L
for Ca2+
affinity5 38 ). The
density of NCX1 in the sarcolemma has been estimated to be 250 to 400
NCX/µm2 in the guinea pig
ventricular myocyte by electrical measurement of NCX
currents from whole-cell and excised giant
patches.32 33 The
density of NCX1 in the sarcolemma appears to differ significantly in
some species. In one report,3
the rate of Ca2+ extrusion via NCX1 per unit
membrane capacitance is 2- to 4-fold larger in guinea pig and hamster
myocytes than in rat myocytes. Immunocytochemically, NCX1 is localized
in the T-tubule membrane as well as in the peripheral
sarcolemma and the intercalated disks in rat and guinea pig
ventricular myocytes, which does not appear to support the
predominant proximity of NCX1s to the ryanodine receptors in the dyadic
junction.14 39
Whether NCX1 with a relatively low Ca2+
affinity is capable of reducing
[Ca2+]i to a low
resting level is an interesting question. NCX1 has been shown to be the
predominant Ca2+ extrusion mechanism in
resting rabbit or guinea pig
cardiomyocytes,7 40
although the [Ca2+] underneath the
sarcolemma could be significantly higher than the bulk resting
[Ca2+]i. When
cloned cardiac NCX1 is expressed in heterologous CCL39 cells with
little endogenous NCX, the exchanger is capable of
completely suppressing Ca2+-regulated plasma
membrane processes, such as
Ca2+/calmodulin-dependent
activation of the
Na+-H+ exchanger
NHE1 by a physiological agonist,
-thrombin.41
Figure 1A
shows a simplified model for
Na+-Ca2+ exchange
by NCX1 that is supported by many previous
studies.33 34 42 43 44
The model represents a consecutive or ping-pong mechanism in
which only 1 substrate ion is translocated at a time, with the
indicated values for the apparent transport and regulatory site
affinities for extracellular and intracellular
Na+ or
Ca2+.34 35 36 45
Interaction of NCX1 with these ions is intrinsically asymmetric in that
the apparent affinity for intracellular Ca2+
is several hundred times higher than that for extracellular
Ca2+, although the affinities for
intracellular Na+ and extracellular
Na+ differ little, and these 2 transport
substrates exert regulatory influences from only the inside (see
below). The effects of alkali metal cations and protons on
Na+-Ca2+ exchange
are also highly asymmetric (also see below). In addition to
Na+-Ca2+
exchange, NCX also catalyzes coupled exchanges of internal
Ca2+ for external
Ca2+ and internal
Na+ for external
Na+.14
The voltage dependence of
Na+-Ca2+ exchange
by NCX1 is attributed mostly to a voltage dependence of the
Na+ translocation step or
Na+ binding, which is rate limiting in
overall
reaction.33 34 44
However, the net negative charge is also moved by cardiac NCX1 during
outward Ca2+ translocation from the
cytoplasmic side, suggesting that the ion-binding site in unliganded
NCX protein has a negative charge slightly more than
-2.32 44
|
| Regulation |
|---|
|
|
|---|
|
In neonatal cardiomyocytes and NCX1-transfected
cells, the NCX1 protein is phosphorylated at specific
serine residues under basal and stimulated conditions, but the
phosphorylation is abolished by PKC
inhibitors.46 48
However, phosphorylation of the NCX1 protein is not
required for PKC-dependent NCX activation, because cells expressing an
NCX1 mutant with its phosphorylatable residues mutated to alanines
exhibit a normal response to PKC
activation.48 Interestingly,
the phorbol esterdependent stimulation of NCX activity is abolished
in cells expressing an NCX1 mutant lacking most (amino acids 246 to
672) of the central hydrophilic loop (see
Figure 2A
) or in a loss-of-function mutant of the XIP
segment in the same
loop48 51 (see
below). These findings, together with an absence of effect on cloned
NCX2, suggest that enhanced NCX activity is not secondary to the
PKC-dependent modulation of other ion transporters. One likely
mechanism for the PKC-dependent regulation of NCX1 is involvement of a
phosphorylatable cytosolic ancillary protein(s) that interacts with the
exchanger. In this context, it is important to note that a novel 13-kDa
cytosolic protein recently isolated from the axoplasm and brain of
squid has been shown to be involved in ATP-dependent regulation of
NCX-SQ1.52
|
Protein kinase A activation by forskolin and/or other reagents has been reported to stimulate activities of cardiac NCX1 expressed in BHK cells50 and Xenopus oocytes,24 although the same clone expressed in CCL39 cells is not affected by 8-bromo-cAMP or 8-bromo-cGMP.48 In frog heart, however, ß-receptordependent or cAMP-dependent stimulation results in the inhibition of NCX activity, in which a novel 9amino acid exon of the frog exchanger seems to play a role.53 However, the part played by protein kinase A in this inhibition is not clear. On the other hand, Condrescu et al54 found that protein phosphatase inhibitors, such as calyculin A, caused substantial inhibition of intracellular Na+dependent Ca2+ influx into CHO cells expressing bovine cardiac NCX1. Because use of the PKC inhibitor and deletion of the central hydrophilic loop of NCX1 did not block such inhibition, protein phosphorylation by a kinase(s) other than PKC might be involved.
To date, studies of NCX currents using excised giant cardiac or oocyte membrane patches have failed to provide evidence for the involvement of protein kinases, although high ATP concentrations (K1/2 >4 mmol/L) have been shown to stimulate NCX1 activity in these patches.55 56 Such failure could be due to loss of diffusible factors from excised patches. Using dialyzed squid axons, DiPolo and Beaugé57 58 have accumulated a large body of data supporting the involvement of protein phosphorylation in the ATP-dependent regulation of squid NCX-SQ1, although known protein kinase inhibitors are reportedly unable to abolish such an ATP effect. In the squid axon, ATP markedly increases the affinities of NCX-SQ1 for the transport substrates (intracellular Ca2+ and extracellular Na+) without a change in Vmax and also increases the affinity for the regulatory intracellular Ca2+58 59 (see below). However, comparable data have not yet been obtained for the cardiac NCX1.
Over a longer time range, upregulation of NCX1 activity
occurs due to increased gene expression at the transcription level in
adult cardiac myocytes in the pressure-overloaded heart, in the
infarcted heart (peri-infarcted area), and in the chronically failed
heart.4 60 61
In isolated adult and neonatal cardiomyocytes, NCX1 gene
expression is increased significantly in response to
-adrenergic
(with
phenylephrine)62
or ß-receptor/cAMPdependent
stimulation.63 Furthermore,
long exposure to high external Ca2+ and
enhanced Na+ influx, procedures that raise
[Ca2+]i or
[Na+]i, result in a
significant increase in the NCX1 mRNA level in adult and neonatal
cardiomyocytes.60 63
In neonatal cardiomyocytes, transforming growth
factor-ß1 and interleukin-1ß
stimulate and inhibit NCX1 gene expression, respectively, and
PKC inhibitors significantly reduce both the basal and
transforming growth factor-ß1stimulated
levels of NCX1 mRNA in these
cells.64 The NCX1 gene
contains 3 tissue-specific promoters and multiple 5'-untranslated
region exons upstream from the coding region that undergo alternative
splicing.65 66 67
From an analysis of the cardiac-specific promoter,
cis-acting elements important
for the expression of NCX1 in neonatal cardiomyocytes as
well as for its induction by
1-adrenergic
stimulation have been
identified.67 However, the
mechanisms for upregulation or downregulation of the NCX1 gene by many
of the above signals remain unclear.
Cardiac NCX activity is intrinsically regulated by
ATP-dependent mechanisms as well as by
physiologically important cations
(Table
).
When ATP levels in adult rat cardiomyocytes are depleted
>90% by treatment with metabolic inhibitors,
NCX activity measured as intracellular
Na+dependent
Ca2+ uptake is inhibited by
80%.68 ATP depletion
also reduces exchange activity in cells expressing cloned cardiac
NCX1.46 69 ATP
depletion affects many aspects of cell metabolism, causing
reduced phosphorylation of proteins and active
metabolites, such as phosphatidylinositol 4,5-bisphosphate
(PIP2), alteration of the cytoskeleton
structure, and induction of various stress
responses.70 In CHO cells
expressing cloned NCX1, the effect of ATP depletion can be partially
mimicked by the action of cytochalasin D, an agent that enhances
depolymerization of actin
microfilaments.69
Furthermore, NCX1 protein has previously been shown to bind to the
cytoskeletal protein
ankyrin.71 Therefore, ATP
depletion may inhibit NCX activity by changing its membrane anchorage
because of disruption of the submembrane cytoskeleton. However, in rat
adult cardiomyocytes, cytochalasin D was reported not to
affect the NCX activity
significantly.72 The effect
of ATP depletion is absent or greatly reduced in cells expressing an
NCX1 mutant lacking a large portion of the central hydrophilic
loop48 69 or a
loss-of-function mutant of the XIP or
Ca2+-regulatory segment in the same
loop51 (see below),
suggesting an important role for this loop in regulation by ATP
depletion.
Collins et al,55
Hilgemann et al,56 and He et
al73 observed that
millimolar ATP markedly activates NCX currents in giant
membrane patches excised from guinea pig ventricular
myocyte blebs or oocytes expressing cloned NCX1. This activating effect
of ATP is mimicked by exogenous PIP2 applied to
the cytoplasmic surface of the patch, but it is attenuated by
procedures that reduce the effective PIP2 level
in the patch, such as treatment with anti-PIP2
antibody.73 74
Thus, the ATP-dependent NCX activation is most likely due to the
generation of PIP2 from its precursors in
excised patches. This situation resembles that of ATP-sensitive
K+ (KATP) channels,
in which ATP usually suppresses channel activity with a
Ki of
1 mmol/L in intact
cells.75 In giant patches
excised from oocytes expressing cloned KATP
channels, the
Ki for
ATP is as low as
10 µmol/L, which increases to a normal value when
PIP2 is added to the cytoplasmic surface of the
patch. Therefore, the PIP2 level is low in
excised giant patches, and the resting endogenous level of
PIP2 is likely to be required to maintain the
physiological ATP sensitivity of
KATP channels in intact
cells.75 It is possible that
endogenous PIP2 plays a similar
constitutive role in cardiomyocytes, thus maintaining NCX
activity at a high level. However, because PIP2
exerts a strong regulatory influence on NCX activity in excised
patches, changes in levels of PIP2 or in other
acidic phospholipids in myocytes in response to stimuli may contribute
to the regulation of NCX activity in intact cells. This requires
further clarification.
Besides being transport substrates, intracellular
Ca2+ and Na+
exert important modulatory effects on NCX
activity.56 58 76 77
Both the Ca2+-influx and
Ca2+-efflux modes of NCX1 are
activated only when regulatory intracellular
Ca2+ binds to a high-affinity site located
in the central hydrophilic loop of
NCX135 78
(Figure 2A
). For this intracellular
Ca2+dependent activation,
K1/2
values of 0.1 to 0.4 µmol/L have been obtained by measuring the peak
of the outward exchange current in excised membrane patches without ATP
(giant patches excised from myocyte
blebs,56 oocytes expressing
cloned NCX1,35 79
and large inside-out "macropatches" excised from intact
myocytes36 ). In contrast, a
much smaller
K1/2 of
22 nmol/L was obtained by Miura and
Kimura45 for activation of
the whole-cell outward exchange current in guinea pig myocytes. It is
possible that the excised patch experiments without ATP could have
produced overestimates of
K1/2,
because high ATP concentrations have been shown to reduce
K1/2
values in excised bleb
patches55 or in dialyzed
squid axons58 (see
Table
).
In a recent study, Weber et
al80 examined the effect of
changing [Ca2+]i on
NCX activity in voltage-clamped intact cardiomyocytes while
monitoring the bulk
[Ca2+]i. They
reported that NCX activity in ferret myocytes is regulated by a
physiological range of
[Ca2+]i with a
K1/2 of
0.125 µmol/L, although similar regulation is not detected in mouse
myocytes at [Ca2+]i
>0.1 µmol/L. It has previously been observed that NCX1 is capable of
extruding Ca2+ from resting
rabbit7 or guinea
pig40
cardiomyocytes. On the other hand, Haworth and
Goknur,81 who
analyzed the beat-dependent activation of
22Na flux in isolated adult rat
cardiomyocytes, have suggested that NCX1 is
activated predominantly at a
[Ca2+]i above the
resting level. Overall, NCX appears to exhibit relatively low activity
in intact myocytes at rest, with further activation occurring at a
higher physiological
[Ca2+]i. In
addition, there might be some species-specific differences in this
regulation.80
In the presence of extracellular
Ca2+ and regulatory intracellular
Ca2+, a high
[Na+] applied to the cytoplasmic surface
of inside-out excised patches rapidly activates the exchanger,
followed by an inactivation process in which the exchange current
slowly decays to a steady
state77 (see
Figure 1B
). This process, called intracellular
Na+dependent inactivation, is suggested to
occur when the transport sites in NCX1 are fully loaded with
Na+ from the cytoplasmic side. An
analysis of exchange current noise suggests that this process
is a manifestation of the relatively slow (t1/2
1 to 5 seconds) conformational transition between the active and
inactive states of the exchanger that occurs during
Na+-Ca2+
exchange.82 Intracellular
Na+dependent inactivation is influenced by
a variety of factors; it is enhanced by high intracellular protons (at
low
pHi)77 83
but attenuated by micromolar intracellular
Ca2+
(K1/2
2 µmol/L), millimolar ATP, or
PIP2.56 73 74
Furthermore, this inactivation process and its modulation by the above
factors are all abolished when the intracellular surface of the NCX
protein is partially digested with a proteolytic
enzyme,84 suggesting that
the large cytoplasmic loop is involved in the exchanger inactivation.
Likewise, treatment of the NCX1 protein with redox reagents
(dithiothreitol and FeSO3) results in marked
stimulation of NCX
activity.85 A recent study
has provided evidence that redox reagents enhance NCX activity mainly
by attenuating the intracellular
Na+dependent
inactivation.86
In whole-cell patch-clamp measurements, NCX activity can be
modulated by an inactivation process similar to the intracellular
Na+dependent inactivation observed in
excised giant patches.87
Therefore, an intriguing question is to what extent the activity of
NCX1 is modulated by regulatory intracellular
Ca2+ and Na+ in
beating cardiomyocytes, in which
[Ca2+]i oscillates
between
0.1 and
1.0 µmol/L with some variation in
[Na+]i around
10
mmol/L.2 6 The
activation and deactivation of NCX activity by regulatory intracellular
Ca2+ per se appear to be very fast, because
the inward NCX currents can be turned on or off by the addition or
removal of intracellular Ca2+ within the
time needed for fast solution
change.35 36 On
the other hand, the intracellular
Ca2+dependent modulation of the
steady-state outward NCX current evoked under a high
[Na+]i (see
Figure 1B
) in giant excised cardiac bleb or oocyte patches
exhibits much slower kinetics (t1/2 5 to 10
seconds).35 56
However, recent experiments using macropatches excised from intact
cardiomyocytes36
have shown that deactivation and activation of the steady-state outward
NCX currents in 50 mmol/L intracellular
Na+ by removal and readdition (to 5
µmol/L) of intracellular Ca2+ occur in
fast and slow phases, with an overall t1/2 of
0.5 and
0.4 seconds, respectively. These kinetics are much faster
than those obtained from giant excised
patches.35 56
Therefore, at the physiological
[Na+]i, NCX
activity is likely to be modulated on a beat-to-beat basis by
intracellular Ca2+dependent modulation.
This interpretation is consistent with a recent demonstration
of very fast induction of outward NCX current in the intact ferret
myocytes when the latter are exposed to
caffeine.80 However, it is
not clear to what extent a normal level of intracellular
Na+ causes intracellular
Na+dependent inactivation and how fast
physiological levels of intracellular
Ca2+ deactivate this intracellular
Na+ effect in beating myocytes.
The steady-state exchange activity of NCX1 exhibits a pronounced pHi dependence, increasing monotonically from a near-zero activity at pHi 6 to a high activity at pHi 9.83 88 In myocardial ischemia/reperfusion, the intracellular Na+dependent inactivation of NCX activity due to a prevailing high intracellular Na+, low ATP, and/or low pHi level may be beneficial, because the resulting decrease of NCX-mediated Ca2+ entry would reduce Ca2+ overloading and myocyte damage.12 Consistent with this idea, cells expressing an NCX1 mutant exhibiting no intracellular Na+dependent inactivation are highly sensitive to cell damage caused by intracellular Na+dependent Ca2+ overloading.51 In contrast, reperfusion after a prolonged period of ischemia is known to be associated with a burst of oxygen-derived free radical production.89 Goldhaber90 has provided evidence that oxygen-derived free radicals enhance Na+-Ca2+ exchange in intact rabbit ventricular myocytes. Because redox reagents activate NCX activity primarily by attenuating the intracellular Na+-dependent inactivation (see above), free radicals generated during reperfusion after prolonged myocardial ischemia may enhance NCX activity, promoting intracellular Ca2+ overload and triggering arrhythmia. The molecular mechanism of free radicalinduced activation of NCX is currently unclear.
External alkali metal ions, such as Na+, K+, or Li+, increase the Vmax of NCX activity up to 2- to 3-fold with low affinity91 92 93 94 (K1/2 several tens of mmol/L for NCX1 in the case of Li+). These cations bind to a site(s) that is distinct from the transport sites, and they are not transported by the exchanger.91 94 Under physiological conditions, Na+ functions as both the transport substrate and an activator of NCX activity. In squid axons, intracellular alkali metal cations also stimulate NCX activity, but this seems to require the simultaneous presence of an external alkali metal cation.91 How these cations regulate the transport properties of the exchanger is not clear. On the other hand, Egger and Niggli95 have reported an interesting effect of extracellular protons on the NCX transport property in guinea pig cardiomyocytes; at pHo 5 or 6, the inward NCX current is strongly inhibited, whereas the corresponding rate of extracellular Na+dependent Ca2+ extrusion is decreased only weakly. Thus, extracellular protons appear to modify the electrogenicity or stoichiometry of Na+-Ca2+ exchange.
| Structure and Function |
|---|
|
|
|---|
550 amino acids) forms a large domain exposed on
the cytoplasm. The latter domain does not appear to be required
for the transport function of NCX1, because a mutant lacking most of it
(
240-679) still retains exchange
activity.84 Recent
topological
studies98 99 100
suggest that the mature NCX1 protein comprises 9 transmembrane segments
(TMs) and a large hydrophilic loop between TMs 5 and 6, with N- and
C-termini located on the external and internal sides, respectively (see
the model in
Figure 2A
40 amino
acids, which are designated the
-1 and
-2
repeats.101 These repeat
sequences are conserved in all members of the NCX family as well as in
related cation
exchangers,16 101
suggesting the functional importance of these segments.
In the NCX1 protein, the
-1 repeat consists of portions
of putative TM2 and TM3 and a loop connecting these TMs, whereas the
-2 repeat consists of a portion of putative TM7 and the C-terminal
non-TM segment. Recent substituted cysteine scanning
analyses98 102
have provided evidence suggesting that the loop of the
-1 repeat and
the non-TM segment of the
-2 repeat together with its C-terminal
neighboring region form reentrant membrane loops originating from the
opposite sides of the membrane, respectively
(Figure 2A
). Cysteine substitutions at many residues in these
loop regions render the exchanger sensitive to inhibition by externally
or internally applied membrane-impermeable sulfhydryl reagents,
suggesting that these residues might be exposed on the ion
pathway.98 102
On the other hand, Qiu et
al,103 using cysteine
mutagenesis with disulfide cross-linking, have shown that the same
interface of TM7 is close to TM2 on the extracellular side but is
adjacent to TM3 near the intracellular side of the membrane and that
TM2 adjoins TM8. Thus, the
-1 and
-2 repeats and their loop
regions are most likely in proximity within the membrane
(Figure 2B
).
Site-directed mutagenesis studies have permitted the
identification of a number of amino acid residues in the
repeats
whose mutations significantly alter the transport properties of cardiac
NCX1. When 45Ca2+
uptake activity was measured in
Xenopus oocytes expressing
mutants of carboxyl- or hydroxyl-containing amino acid residues within
putative TMs 2, 3, and 7, many of them exhibited no or only low NCX
activity.104 Furthermore,
mutation of conserved glycines (Gly138 and Gly837) alters the slope of
the current-voltage relationship of NCX1. Mutation of Thr103 at the
cytoplasmic portion of TM2 increases the apparent affinity of NCX1 for
the substrate intracellular Na+ and also
seems to produce Li+ transport capacity,
suggesting alteration in the ionic selectivity of the
exchanger.105 On the other
hand, the putative TMs 4 and 5 contain regions of similarity to the
Na+,K+-ATPase and
SR Ca2+-ATPase, and mutation of Glu199 or
Thr203 in TM5 results in the loss of NCX
activity.104 Glu199 of NCX1
corresponds to Glu309 of SR Ca2+-ATPase,
which, as revealed in the high-resolution 3D structure of SR
Ca2+-ATPase, is one of the residues directly
liganding the transported Ca2+ in the
membrane.106
In the putative loop regions of the
repeats in cardiac
NCX1, mutations of 3 conserved aspartic acids (Asp130, Asp825, and
Asp829) result in up to 6-fold reduction in the apparent affinity for
the substrate, extracellular
Ca2+.102
Furthermore, mutations of other residues in the
repeat loop regions
(Asn125, Thr127, and Val820) render the exchanger up to 8-fold less
sensitive to inhibition by external
Ni2+,107
a competitive inhibitor for the transport substrate,
extracellular
Ca2+.93
Similar studies have resulted in the identification of Val820, Gln826,
and Gly833 in the
-2 repeat loop, whose mutations alter the apparent
affinity for
2-[2-[4-(4-nitrobenzyloxy)phenyl]ethyl]isothiourea
methanesulfonate
(KB-R7943)108 (see below).
Mutation at Gly833 renders NCX1 almost insensitive to inhibition by
KB-R7943. Furthermore, simultaneous mutations of Val820 and
Gln826 alter the extent of stimulation of NCX activity by external
Li+.107
Thus, interactions of NCX1 with the transport substrate (extracellular
Ca2+), inhibitors
(Ni2+ and KB-R7943), and an
activator (Li+) are
significantly influenced by the mutation of residues in the
repeat
loops, although some of the residues identified do not appear to
interact directly with ions and modifiers, because the observed effects
of mutations are mild. Taken together, these data suggest that both the
TMs and loop regions of the
-1 and
-2 repeats participate in the
formation of the ion translocation pathway in NCX1.
At present, however, little is known about the detailed
structure of the NCX1 molecule, in particular, the ion-binding sites
and the shape and dimensions of the ion transport pathway, the
requirement of the oligomeric protein structure for the function, and
changes in the conformation of the exchanger associated with ion
transport. Recently, the low-resolution 3D structure of the
Na+-H+ exchanger
NhaA from Escherichia coli has
been solved, which reveals that it has a highly asymmetric molecular
organization comprising 12 TMs and exists as
dimer.109 Therefore, the
structure of NCX1 does not seem to be similar to NhaA. The unique
topology and functional importance of the
-1 and
-2 repeats of
NCX1 are rather reminiscent of a somewhat similar structural feature
reported for the water channel aquaporin-1, in which 2 prominent loop
regions with functionally important residues penetrate the membrane
from opposite sides, thereby forming part of the
pore.110
As noted above, mild proteolysis of the internal surface of
the exchanger greatly stimulates its activity and eliminates regulation
by intracellular Ca2+, intracellular
Na+, protons, PIP2,
and ATP. Similarly, an NCX1 mutant with deletion of a large portion of
the central cytoplasmic loop (
240-679) is not regulated by
intracellular Ca2+, intracellular
Na+,84
or PKC.48 Thus, the large
cytoplasmic loop is most likely to be involved in the regulation of NCX
activity. At the N-terminal end of this loop near the membrane-lipid
interface, there is a 20amino acid segment, designated the XIP
region, whose sequence is rich in both basic and hydrophobic
residues,19 as in the
calmodulin-binding domain (see
Figure 2A
). This region, to which calmodulin
does not seem to bind
strongly,111 is considered
to play a pivotal role in the regulation of NCX activity (see below).
On the other hand, C-terminal to the XIP region, there is a region of
135 amino acids (amino acids 371 to 508) containing 2 conserved
clusters of acidic amino acids. This 135amino acid region, when
expressed as a fusion protein and assayed directly, binds
45Ca2+ with high
affinity.78 NCX1 mutants
carrying mutations within the acidic clusters exhibit markedly lowered
affinity for regulatory intracellular Ca2+,
suggesting that Ca2+ binding to this region
is responsible for intracellular
Ca2+dependent regulation of NCX
activity.35
In the large cytoplasmic loop of the NCX1 protein, there are
two
70amino acid internal repeat motifs, designated the ß
repeats, which are conserved in the NCX
family.101 Although the
functions of these sequences are not clear, the ß-1 repeat almost
overlaps the N-terminal portion of the
Ca2+-regulatory site, which is reportedly
required for high-affinity
45Ca2+
binding.78 The ß-2 repeat
is located on the C-terminal side of the
Ca2+-regulatory site. A recent study of
tryptic digestion of scallop membranes has shown that limited
proteolysis occurs at both ends of the
Ca2+-regulatory site, with the C-terminal
end cleaved only in the absence of
Ca2+.112
Thus, the Ca2+-regulatory site and its
C-terminal neighboring region containing the ß-2 repeat may form a
folded structure, and Ca2+ removal from the
regulatory site appears to induce a large conformational change in
these regions. A substantial conformational change associated with
Ca2+ binding to a fusion protein containing
the Ca2+-regulatory site was also detected
as a large mobility shift during
SDS-PAGE.78 Finally, there
is a region near the C-terminal end of the cytoplasmic domain in which
alternative splicing involving 6 small exons occurs in a
tissue-specific
manner.22 23
As described above, NCX1 is inactivated when
regulatory intracellular Ca2+ is removed.
Similarly, intracellular Na+dependent
inactivation at a high
[Na+]i generates an
almost fully inactive exchanger state in the absence of ATP or at low
pHi. Although the underlying mechanism(s) that
gives rise to such an inactivated state is not clear, the
endogenous XIP region may play a critical role in the
process. First, a synthetic peptide having the same sequence as the
endogenous XIP region completely inactivates
NCX activity
(Ki
0.1 µmol/L) when applied from the intracellular side of an excised
giant patch.111 This
peptide is also an effective inhibitor of the whole-cell
outward exchange current in ventricular
myocytes.113 Second,
mutations of the XIP region eliminate or accelerate the intracellular
Na+dependent inactivation of the exchange
activity as measured by using excised oocyte giant
patches.79 Mutations in the
XIP region that abolish intracellular
Na+dependent inactivation cause loss of
responsiveness to modulation by ATP, PIP2, or
PIP2
antibody.114 Furthermore,
cells expressing an XIP mutant exhibiting no intracellular
Na+dependent inactivation do not respond
to inhibition by ATP depletion or to activation of PKC by phorbol
ester.51 All of these
results are consistent with the hypothesis that the XIP region
functions as an autoinhibitory domain that plays a central
role in the activation and inactivation of NCX activity. The receptor
site interacting with XIP in NCX1 has not yet been
identified.
Modulations of NCX activity by intracellular
Ca2+ and Na+
appear to be complex processes involving multiple regions of the
exchanger molecule. The regulatory effects of both ions are abolished
by deletion of a small segment (
680-685) of NCX1 near the C-terminal
end of the large cytoplasmic
loop.115 Furthermore,
wild-type NCX1s with different splicing patterns in the region located
upstream from the above-deleted residues
(Figure 2A
) exhibit altered kinetics of regulation by
intracellular Na+ or
Ca2+.24 25
In addition, some mutations within the XIP region, which primarily
affect intracellular Na+dependent
inactivation, significantly reduce the apparent affinity for regulatory
intracellular
Ca2+.79
Conversely, intracellular Ca2+-regulatory
site mutations attenuate intracellular
Na+dependent
inactivation.35 Therefore,
it appears that structural integrity of the large cytosolic loop is
required for the transduction of ion-binding signals. On the other
hand, cysteine substitution of Asn101, modeled to be localized near the
cytosolic interface between TM2 and the first intracellular loop,
renders the exchanger insensitive to regulation by intracellular
Ca2+ or
Na+.105
Thus, a region(s) of the NCX1 protein other than the large cytoplasmic
loop may also be involved in the transduction of ion-binding signals,
which transmit regulatory information on transport by influencing the
TMs that catalyze ion translocation. In this sense, CALX1, an NCX from
Drosophila melanogaster, is
interesting in that activity of the wild-type exchanger is inhibited by
intracellular Ca2+, which is the opposite
pattern of intracellular Ca2+dependent
regulation.116 In CALX1,
Ca2+ binding probably occurs normally, but
the binding signal appears to be transduced
differently.
| Pharmacology |
|---|
|
|
|---|
As blockers of NCX activity, many divalent and trivalent
cations, such as La3+,
Ni2+, and Cd2+
(Table
),
and a variety of organic compounds, including amiloride derivatives
(eg, dichlorobenzamil) or the substituted pyrrolidine ethanamine
(eg, bepridil), have long been known (see previous
reviews14 120
and references therein). These cations and organic agents also exert
blocking effects on other cell systems sometimes showing
cardiovascular activities, which limits their use as
specific blockers of NCX activity. However,
Ni2+ at up to 5 mmol/L is being used as
an NCX blocker under conditions in which other membrane currents
sensitive to Ni2+ are already suppressed by
different agents. More recently, a few new NCX inhibitors,
such as an isothiourea derivative (KB-R7943, formerly No. 7943) and
peptides and their analogues with much improved selectivity for NCX,
have been developed.
KB-R7943 has been reported to be a potent and selective inhibitor of NCX at low micromolar concentrations.121 122 Its effect on NCX1 or NCX2 is 3-fold less potent than its effect on NCX3.93 It is an amphiphilic molecule with a positively charged isothiourea group at neutral pH and is soluble at up to 100 µmol/L in an aqueous buffer. Its inhibitory action is relatively fast and washable. However, when cells are preincubated with KB-R7943 for a longer period (>2 to 3 minutes), more time is needed for its washout. Similarly, the inhibitory potency of the drug seems to increase as the preincubation time is lengthened.
Intriguingly, KB-R7943 exerts a preferential effect on
reverse-mode (Ca2+-influx mode) NCX
activity,121 122 123
although such an effect disappeared under a certain
condition.124 It inhibits
intracellular Na+dependent
Ca2+ influx into rat
cardiomyocytes or some other NCX1-expressing cells with an
IC50 of 1.2 to 2.4 µmol/L, whereas it inhibits
extracellular Na+dependent
Ca2+ efflux from these cells only weakly
(IC50
30
µmol/L).121 On the other
hand, KB-R7943 inhibits whole-cell outward NCX currents from
ventricular myocytes or NCX1-transfected cells with an
IC50 of 0.3 to 0.9
µmol/L,108 122 123
but it is much less potently inhibitory to whole-cell
inward NCX current (IC50 17
µmol/L).122 Inhibition by
KB-R7943 is reportedly noncompetitive with extracellular
Ca2+ or
Na+121
of a mixed-type (competitive and noncompetitive with extracellular
Ca2+)93
or competitive with extracellular
Ca2+,122
suggesting that the mode selectivity may be partly due to competition
with extracellular Ca2+. However, the
selective blocking of Ca2+ influx via NCX is
observed irrespective of the presence or absence of extracellular
Ca2+.121 123
Interestingly, the same mode selectivity as seen in KB-R7943 is seen in
the inhibition of NCX activity by
Ni2+107
or some protein phosphatase
inhibitors54
(see above), whereas the opposite selectivity has been reported for
3',4'-dichlorobenzamil.122
Similarly, in cardiac sarcolemmal vesicles, the XIP
peptide111 and a
Ca2+ channel blocker, nicaldipine (0.1 to 10
µmol/L),125 depressed the
rate of Na+-dependent
Ca2+ uptake much more potently than that of
Na+-dependent
Ca2+ efflux. Although the data suggest a
significant difference in the properties of NCX1 in the forward and
reverse modes, the underlying molecular mechanism for this peculiar
directional specificity is not clear.
Some progress has been made in elucidating the mechanism by
which KB-R7943 acts. First, current evidence strongly suggests that
KB-R7943 inhibits NCX activity from the external side in intact
cells.108 122
When whole-cell outward NCX currents are measured, inhibition is not
observed if the drug is applied internally through a pipette solution.
Furthermore, rapid (<5-second) inhibition of the outward NCX current
is induced when the latter is evoked by external application of both
Ca2+ and
KB-R7943.108 However,
KB-R7943 may be capable of inhibiting NCX activity from the cytoplasmic
side also, inasmuch as it inhibits NCX activity in inside-out giant
patches excised from oocytes expressing
NCX1.126 Second, structural
determinants of KB-R7943 sensitivity in the exchanger have been sought
by analyzing the functions of chimeras between NCX1 and NCX3 and
site-directed mutants of
NCX1108 (see above). The
results suggest that the highly conserved
-2 repeat of the exchanger
is exclusively responsible for the drug response and that mutation at
Val820, Gln826, or Gly833 in the portion of the
-2 repeat forming
the putative reentrant membrane loop
(Figure 2B
) alters drug sensitivity. Mutation at Gly833
causes a particularly large (
30-fold) reduction in KB-R7943
sensitivity. Considering the external side of the drug action in the
intact cell, the simplest interpretation of these data would be that
mutations of above residues alter the conformation of the external
drug-binding site, thereby influencing its affinity for the drug.
However, the possibility remains that Gly833 is part of the KB-R7943
receptor.108
KB-R7943 at up to 30 µmol/L has little effect on the
Na+-Ca2+-K+
exchanger,108 the
Na+-H+ exchanger,
sarcolemmal Ca2+-ATPase, SR
Ca2+-ATPase, or
Na+,K+-ATPase.121
On the other hand, the drug has been reported to inhibit
voltage-sensitive Na+ currents, L-type
Ca2+ currents, and inward rectifier
K+ currents with
IC50s of 14, 8, and
7 µmol/L, respectively,
in guinea pig
cardiomyocytes.122
However, the ramp pulse protocol used in these latter measurements
appears to have overestimated the effect of KB-R7943 (see pertinent
study122 ). In rat
ventricular myocytes, 5 µmol/L KB-R7943 does not alter
steady-state twitches, Ca2+ transients,
Ca2+ load in the SR, or rest potentiation,
but it prolongs the late low plateau of the action potential,
suggesting modest inhibition of K+
currents.123 In guinea pig
papillary muscle, however, KB-R7943 at up to 10 µmol/L does not
significantly affect the resting membrane potential or various action
potential
parameters.121 127
Similarly, the spontaneous beating rate and developed tension are not
affected by 10 or 30 µmol/L KB-R7943 in isolated guinea pig
atria.128 Thus, KB-R7943 at
<5 µmol/L could be used as a fairly selective blocker for
reverse-mode NCX activity in isolated cardiomyocytes.
However, in other cell types, such as bovine adrenal chromaffin
cells129 and rat
hippocampal neurons,130
KB-R7943 has recently been reported to inhibit neuronal nicotinic
acetylcholine receptors (IC50 0.3 to 6.5
µmol/L) and
N-methyl-D-aspartate
receptor channels (2 IC50s 0.8 and 11 µmol/L),
respectively, although the latter is in contradiction with another
report.131 Furthermore,
store-operated Ca2+ entry into cultured
neurons and astrocytes is significantly inhibited by 10 µmol/L
KB-R7943.132 All these
possible side effects need to be taken into account if KB-R7943 is to
be used as an NCX blocker.
Despite its side effects, several pharmacological actions of KB-R7943 provide information on the role of reverse-mode NCX activity. First, the fact that, as noted above, 5 µmol/L KB-R7943 has no effect on normal Ca2+ transients and contractions123 suggests that Ca2+ influx via NCX is not important for physiological excitation-contraction coupling, at least in rat cardiomyocytes. Second, blocking the Na+-K+ pump by cardiac glycosides increases [Na+]i, thereby inducing positive inotropy as well as toxic myocyte Ca2+ overload.13 In rat ventricular myocytes treated with 50 µmol/L strophanthidin, 5 µmol/L KB-R7943 suppressed glycoside toxicity but preserved positive inotropy.123 Similar effects by 30 µmol/L KB-R7943 were observed in spontaneously beating isolated guinea pig atria pretreated with 3 µmol/L ouabain.128 These data are interpreted by Satoh et al123 as suggesting that the reduction of Ca2+ efflux via NCX due to competition with elevated [Na+]i causes the inotropic effect, whereas the net Ca2+ entry via NCX is responsible for the generation of glycoside toxicity. Third, reverse-mode NCX activity has been implicated as the cause of Ca2+ overload associated with cardiac ischemia/reperfusion.12 Recent studies have provided evidence that KB-R7943 at 3 to 20 µmol/L is effective in reducing cytosolic Ca2+ and Na+ overload, cell injury, and arrhythmias that are associated with ischemia/reperfusion, the Ca2+ paradox, and substrate-free hypoxia/reoxygenation in different types of cardiac preparations (rat cardiomyocytes,121 133 guinea pig papillary muscle,127 and Langendorff-perfused rat hearts133 134 ). KB-R7943 has also been reported to be significantly protective against anoxia/reoxygenation or ischemia/reperfusion damage in brain135 and kidney.136
Some synthetic peptides are also effective NCX inhibitors. The XIP peptide derived from the primary sequence of cardiac NCX1 (see above) decreases the Vmax of NCX activity.111 As noted earlier, it is significantly less potent in inhibiting Na+-dependent Ca2+ efflux from sarcolemmal vesicles than in inhibiting the reverse reaction. XIP has little effect on Na+,K+-ATPase, SR Ca2+-ATPase, or L-type Ca2+ currents, and it does not increase membrane conductance when applied to the intracellular surface by use of the excised-patch technique.111 113 However, it may bind calmodulin and could thus interfere with the function of calmodulin-binding proteins. Furthermore, its usefulness as an NCX inhibitor is limited because it acts only from the cytoplasmic side.
Other peptides, such as the molluscan cardioexcitatory tetrapeptide Phe-Met-Arg-Phe-NH2 (FMRFa) and its analogues and the cyclic hexapeptide Phe-Arg-Cys-Arg-Cys-Phe-CONH2 (FRCRCFa), which are much smaller than XIP, have been reported to inhibit NCX activity.137 138 FMRFa and its related peptides inhibit the NCX activity of cardiac sarcolemmal vesicles, with IC50s ranging from 1 to 1000 µmol/L. On the basis of structure/activity studies of these peptides, a new cyclic peptide, FRCRCFa, with an intramolecular disulfide bond has been synthesized. FRCRCFa exhibits improved inhibitory potency and resistance to proteolytic degradation, and in sarcolemmal vesicles it inhibits NCX activity completely, with an IC50 of 2 to 10 µmol/L without competing with extravesicular Ca2+ and Na+.138 In the rabbit ventricular myocyte, FRCRCFa inhibits whole-cell NCX currents much more potently, with an IC50 of 0.023 µmol/L, and exhibits a rapid onset of action.139 Furthermore, it reportedly has no effect on L-type Ca2+ channels or delayed rectifier and inward rectifier K+ channels. Thus, FRCRCFa appears to have several advantages over XIP. This peptide acts from the intracellular side, but its binding site has not been identified.
| Concluding Remarks |
|---|
|
|
|---|
200%)
or of an NCX1 mutant devoid of intracellular
Ca2+ and intracellular
Na+dependent regulation in transgenic mice
has been reported to produce a relatively normal myocyte function,
except that Ca2+ fluxes via NCX and the SR
Ca2+ content are increased in the former
mice, and rest potentiation is substantially enhanced in papillary
muscle from the
latter.16 115
Finally, highly potent and selective new drugs targeting NCXs are being
developed. The extent to which these can be of clinical benefit remains
to be seen.
| Acknowledgments |
|---|
| Footnotes |
|---|
| References |
|---|
|
|
|---|