(Received for publication, December 2, 1996, and in revised form, February 3, 1997)
From the Veterans Administration Medical Center,
Department of Internal Medicine, and Howard Hughes Medical Institute,
University of Iowa College of Medicine, Iowa City, Iowa 52240, the
¶ School of Pharmacy and Pharmacology, University of Bath,
Claverton Down, Bath BA2 7AY, United Kingdom, and ** Biological Research
Laboratories, Sankyo Co. Ltd., Tokyo 108, Japan
Adenophostin A possesses the highest known
affinity for the inositol 1,4,5-trisphosphate
(Ins(1,4,5)P3) receptor (InsP3R). The
compound shares with Ins(1,4,5)P3 those structural elements essential for binding to the InsP3R. However, its adenosine
2-phosphate moiety has no counterpart in the Ins(1,4,5)P3
molecule. To determine whether its unique structure conferred a
distinctive biological activity, we characterized the
adenophostin-induced Ca2+ signal in Xenopus
oocytes using the Ca2+-gated Cl
current
assay. In high concentrations, adenophostin A released Ca2+
from Ins(1,4,5)P3-sensitive stores and stimulated a
Cl
current that depended upon the presence of
extracellular Ca2+. We used this Cl
current
as a marker of Ca2+ influx. In low concentrations, however,
adenophostin A stimulated Ca2+ influx exclusively. In
contrast, Ins(1,4,5)P3 and
(2-hydroxyethyl)-
-D-glucopyranoside 2
,3,4-trisphosphate, an adenophostin A mimic lacking most of the
adenosine moiety, always released intracellular Ca2+ before
causing Ca2+ influx. Ins(1,4,5)P3 could still
release Ca2+ during adenophostin A-induced Ca2+
influx, confirming that the Ins(1,4,5)P3-sensitive
intracellular Ca2+ stores had not been emptied.
Adenophostin- and Ins(1,4,5)P3-induced Ca2+
influx were not additive, suggesting that both agonists stimulated a
common Ca2+ entry pathway. Heparin, which blocks binding to
the InsP3R, prevented adenophostin-induced Ca2+
influx. These data indicate that adenophostin A can stimulate the
influx of Ca2+ across the plasma membrane without
inevitably emptying the Ins(1,4,5)P3-sensitive intracellular Ca2+ stores.
Stimulation of many plasma membrane receptors increases the intracellular concentration of the second messenger inositol 1,4,5-trisphosphate (Ins(1,4,5)P3).1 In its best characterized activity, Ins(1,4,5)P3 binds with high affinity to its receptor (InsP3R), a Ca2+ channel that traverses the membranes enclosing intracellular Ca2+ stores (1). As a result of this interaction, the InsP3R Ca2+ channel opens, and Ca2+ flows from the internal stores into the cytosol. Very often, however, Ins(1,4,5)P3 not only releases intracellular Ca2+, it also stimulates the influx of Ca2+ across the plasma membrane. The cellular mechanism by which Ins(1,4,5)P3 stimulates Ca2+ influx remains enigmatic. Although Ins(1,4,5)P3 may directly regulate Ca2+ channels at the plasma membrane in T lymphocytes (2), most experimental data suggest that Ins(1,4,5)P3 stimulates Ca2+ influx indirectly, by depleting the intracellular Ca2+ stores. We do not know how depletion of the Ca2+ store would translate into the opening of putative plasma membrane "Ca2+ release-activated Ca2+" (CRAC) channels (3). Perhaps a diffusible Ca2+ influx factor (CIF), whose chemical structure and mechanism of action remain to be elucidated (4-7), could bridge the gap between the Ca2+ stores and the plasma membrane. Alternatively, depletion of the Ca2+ stores may be associated with, but not required for, the development of Ca2+ influx. In the "conformational coupling" model, for example, the InsP3R itself interacts with CRAC channels to stimulate Ca2+ influx (8). In this work, we used the compound adenophostin A and a novel synthetic analogue thereof to investigate if it is essential to deplete the Ca2+ stores to stimulate Ca2+ influx.
Adenophostin A, a compound isolated from the culture broth of
Penicillium brevicompactum, is the most potent known agonist at the InsP3R. Its affinity for the InsP3R is
~100-fold greater than Ins(1,4,5)P3 itself
(9, 10). The adenophostin A molecule contains a glucose
3,4-bisphosphate and an adenosine 2-phosphate moiety (Fig. 1). The
glucose 3,4-bisphosphate moiety includes the structural elements common
to all of the highly potent inositol phosphate positional isomers (Fig.
1) and thus probably targets the Ins(1,4,5)P3-binding
domain of the InsP3R. The synthetic compound
(2-hydroxyethyl)-
-D-glucopyranoside 2
,3,4-trisphosphate (ADAN 1) (Fig. 1) possesses the glucose 3,4-bisphosphate
but lacks the adenosine 2
-phosphate moiety of adenophostin A. ADAN 1 binds to the InsP3R with a much lower affinity than that of
adenophostin A (11, 12). Thus, the adenosine 2
-phosphate moiety
appears to contribute significantly to the activity of adenophostin A. Because this moiety has no counterpart in the Ins(1,4,5)P3
molecule, we hypothesized that it interacts with a region of the
InsP3R located outside the Ins(1,4,5)P3-binding
domain. If so, then stimulating the InsP3R with
adenophostin A could result in a distinctive Ca2+ signal
that may uncover novel aspects of the InsP3R function. To
begin to test this hypothesis, we microinjected adenophostin A into
Xenopus oocytes and compared the resulting Ca2+
signal with that induced by Ins(1,4,5)P3. Our results show
that in high concentrations, adenophostin A and
Ins(1,4,5)P3 produce a similar Ca2+ signal.
However, in low concentrations, adenophostin A stimulates extracellular
Ca2+-dependent Cl
current
(Ca2+ influx) without first releasing intracellular
Ca2+.
Adenophostin A was isolated (10) and ADAN 1 was synthesized (11) as described before. The purity of adenophostin A was more than 90% by high pressure liquid chromatography analysis, and no contaminating signal was observed by NMR spectrometry. ADAN 1 was purified by ion exchange chromatography and was used as the triethylammonium salt. Ins(1,4,5)P3 was obtained from Calbiochem. BAPTA was from Molecular Probes (Eugene, OR). All other chemicals were from Sigma. Oocyte harvesting, micropipette calibration, and cytosolic microinjections were performed as described previously (13). Injection pipettes were back-filled with adenophostin A, ADAN 1, or Ins(1,4,5)P3. The injection volume was kept at 0.7 nl, as determined by measuring the diameter of a droplet microinjected under oil.
ElectrophysiologyWe assayed changes in free cytosolic
Ca2+ concentration ([Ca2+]i) by
measuring Ca2+-activated Cl currents with the
two-electrode voltage clamp technique as described before (13). This
assay has been extensively validated using Ca2+-sensitive
electrodes (14-16) and fluorescent Ca2+ indicators (13,
17-21). For most experiments, oocytes were initially stimulated in a
bath solution containing 116 mM NaCl, 2 mM KCl, 6 mM CaCl2, 1 mM MgCl2,
5 mM HEPES, pH 7.4. We assessed Ca2+ influx by
measuring the change in Ca2+-gated Cl
current
induced by either lowering the Ca2+ (from 6 to 0.1 mM CaCl2) or increasing the concentration of
the inorganic ions Mn2+ (4 mM),
Ni2+ (5 mM), or La3+ (5 mM) in the bath. Both Ni2+ and La3+
block Ca2+ channels. Although Mn2+ may go
through Ca2+ influx channels, it does not activate the
Ca2+-gated Cl
channel in the oocyte (22). We
preferred using the inorganic ions to reversibly inhibit the
Ca2+-gated Cl
current caused by
Ca2+ influx (22-24) because the plasma membrane electrical
resistance often decreases in low external [Ca2+], making
some of the recordings difficult to interpret (24). The
Ca2+-gated Cl
current reflects
[Ca2+]i just beneath the cytoplasmic membrane;
changes in [Ca2+]i occurring deeper in the cell
are not measured (25, 26). Thus, this assay inherently measures
[Ca2+]i in the cellular area most likely to be
affected by Ca2+ influx. The assay also integrates the
submembranous [Ca2+]i changes across the
oocyte's entire plasma membrane surface, thereby maximizing our
ability to detect Ca2+ influx. Although the extracellular
Ca2+-dependent Cl
current assays
Ca2+ influx indirectly, for clarity, we use the terms
"extracellular Ca2+-dependent
Cl
current" and "Ca2+ influx"
interchangeably.
When injected in high concentration (105
M in the pipette), adenophostin A causes a biphasic
response; there is an initial, short lived increase in
[Ca2+]i followed by a slow increase (Fig.
2A, n = 16). This response
pattern is similar to that generated by high concentrations of
Ins(1,4,5)P3 (10
4 M in the
pipette, n = 8; and see Ref. 22). Both components of
the response to adenophostin A are due to an increase in
[Ca2+]i because they can be completely blocked by
the Ca2+ chelator BAPTA (1 nmol, n = 6, flat tracings not shown). The initial transient increase of
[Ca2+]i caused by both agonists persists in the
presence of extracellular Mn2+ (4 mM) (Fig. 2,
C and D) or in low extracellular
[Ca2+] (n = 5 for adenophostin A,
n = 6 for Ins(1,4,5)P3; see also Ref. 23)
and represents the release of Ca2+ from the intracellular
stores. In contrast, the slow increase in [Ca2+]i
can be blocked by adding either extracellular Mn2+ (4 mM, Fig. 2, A and B,
n = 15 for adenophostin A, n = 3 and see Ref. 22 for Ins(1,4,5)P3), extracellular
Ni2+ (5 mM, n = 17 for
adenophostin A, n = 5 for Ins(1,4,5)P3) or extracellular La3+ (5 mM, n = 4 for adenophostin A, n = 6 for
Ins(1,4,5)P3). This slow increase in
[Ca2+]i can also be blocked by decreasing the
extracellular [Ca2+] from 6 to 0.1 mM
(n = 10 for adenophostin A; see Refs. 22 and 23 for
Ins(1,4,5)P3). These results indicate that the slow component of the response results from the influx of extracellular Ca2+. The current-voltage relationship of the
Ins(1,4,5)P3-induced (n = 7) and of the
adenophostin-induced (n = 9) extracellular Ca2+-dependent currents was linear within the
90 to
10 mV range. Reversal potentials were similar for both
agonists (
23 ± 4 for Ins(1,4,5)P3 and
27 ± 4 mV for adenophostin) and consistent with the extracellular
Ca2+-dependent current being carried mainly by
Cl
ions (27, 28) (Fig. 3). While
Ins(1,4,5)P3-induced Ca2+ influx ends after
30-35 min. (22), with adenophostin A, the extracellular
Ca2+-dependent Cl
current
continues for at least 1 h (n = 6). This prolonged
Ca2+ influx probably reflects the poor metabolism of
adenophostin A, since 1) adenophostin A resists inactivation by the
Ins(1,4,5)P3 3-kinase and 5-phosphatase enzymes in
vitro (9) and 2) poorly metabolized inositol phosphate cause
prolonged Ca2+ influx in the oocyte (22, 23). Aside from
the duration of Ca2+ influx, high concentrations of
adenophostin A and Ins(1,4,5)P3 produce similar
Ca2+ signals.
To determine if adenophostin A and Ins(1,4,5)P3 released
Ca2+ from the same intracellular pool, we injected
Ins(1,4,5)P3 and adenophostin A sequentially. As seen in
Fig. 2C, Ins(1,4,5)P3 (104
M in the pipette) did not release Ca2+ when
injected after adenophostin A (10
4 M in the
pipette, n = 8). Conversely, adenophostin A did not release Ca2+ when injected after Ins(1,4,5)P3
(Fig. 2D, n = 3). This cross-desensitization suggests that adenophostin A releases Ca2+ from the
oocyte's Ins(1,4,5)P3-sensitive Ca2+ pools.
Taken together, our data are consistent with adenophostin A stimulating
the InsP3R (10).
When we compared the functional potency of many inositol phosphates, we
noticed that in low concentrations, they only released intracellular
Ca2+; they did not cause Ca2+ influx (29, 30).
Based on these results, we anticipated that when injected in threshold
concentrations, adenophostin A would strictly release intracellular
Ca2+. As the tracing in Fig. 4A
shows, a low concentration of adenophostin A (107
M in the pipette) caused a slow increase in
Cl
current. Contrary to what we had predicted, this
current is caused by Ca2+ influx, since it can be blocked
by either adding Mn2+ (n = 9) or
Ni2+ (n = 7) to the extracellular bath or
by decreasing bath [Ca2+] (n = 4). The
absence of a fast initial component of the response suggested that
threshold concentrations of adenophostin A did not release
intracellular Ca2+. To test this possibility further, we
injected adenophostin A in the continued presence of Mn2+
or Ni2+. As the example in Fig. 4B shows,
adenophostin A did not elicit a response in the presence of
Mn2+ in the bath, whereas subsequent removal of the
Mn2+ confirmed that adenophostin had stimulated
Ca2+ influx (n = 3). We also considered the
possibility that we failed to observe intracellular Ca2+
release because it did not yield a sufficiently high
[Ca2+]i to stimulate the Ca2+-gated
Cl
channels. However, the smallest possible increase in
[Ca2+]i caused by an inositol phosphate is higher
than the threshold [Ca2+]i required to open the
Cl
channels (29, 31). Moreover, the absence of an initial
release of intracellular Ca2+ implied that adenophostin A
had not depleted the Ins(1,4,5)P3-sensitive Ca2+ stores. To verify this prediction, we injected a low
concentration of adenophostin A, blocked the resulting Ca2+
influx with Mn2+, and then injected
Ins(1,4,5)P3. As shown in Fig. 4C, injection of
Ins(1,4,5)P3 during adenophostin-stimulated
Ca2+ influx causes a transient release in intracellular
Ca2+ (n = 10). These results indicate that
adenophostin A can stimulate Ca2+ influx without emptying
the oocyte's Ins(1,4,5)P3-sensitive Ca2+
stores. These results also argue against adenophostin A acting as an
ATP inhibitor to cause Ca2+ influx: if adenophostin A
inhibited the Ca2+ ATPases, then we would have expected the
Ins(1,4,5)P3-sensitive intracellular Ca2+
stores to be empty, and they were not. Furthermore, we could not cause
Ca2+ influx with other purine compounds expected to inhibit
ATPases, such as ADP (10 mM in the pipette,
n = 5) or ATP
S (10 mM in the pipette,
n = 6).
Given that adenophostin A has an extremely high affinity for the
InsP3R and that it cross-desensitizes with
Ins(1,4,5)P3 for the release of intracellular
Ca2+, our working hypothesis was that adenophostin A acted
through the InsP3R to cause Ca2+ influx. This
hypothesis predicted that adenophostin A and Ins(1,4,5)P3 should stimulate a common Ca2+ entry pathway. To test this
prediction, we first injected a high concentration of
Ins(1,4,5)P3 (104 M in the
pipette). When Ca2+ influx reached its maximum, we injected
adenophostin A (10
7 M). As shown in Fig.
5A, adenophostin A caused Ca2+
influx to return to the maximal value that had been reached with Ins(1,4,5)P3 alone but not to exceed this value
(n = 4). This lack of additivity is consistent with
Ins(1,4,5)P3 and adenophostin A ultimately stimulating a
common Ca2+ entry pathway. Contrary to the lack of
additivity for Ca2+ influx, the working hypothesis
predicted that adenophostin A and Ins(1,4,5)P3 would act
additively to release intracellular Ca2+. Microinjections
of adenophostin A in concentrations sufficient to cause
Ca2+ influx, but below the threshold for Ca2+
release, lowered the threshold for Ins(1,4,5)P3-induced
Ca2+ release (Fig. 5B, n = 6).
The working hypothesis also predicted that heparin, which prevents
Ins(1,4,5)P3 (32) and adenophostin A (9) from binding to
the InsP3R, should prevent adenophostin A from inducing
Ca2+ influx. When we preinjected oocytes with heparin (10 mg/ml in pipette, 30-nl injection volume), adenophostin A
(10
7 M) no longer stimulated Ca2+
influx (n = 5, Fig. 5C). Although heparin
could prevent Ca2+ influx through other mechanisms, our
aggregate data nevertheless suggest that adenophostin A binds to the
InsP3R to stimulate Ca2+ influx.
Despite evaluating 47 of the 64 possible positional isomers, we have
not encountered an inositol phosphate that stimulates Ca2+
influx exclusively (30). Instead, all of the inositol phosphates, including Ins(1,4,5)P3, released intracellular
Ca2+ before causing Ca2+ influx (22). The most
obvious structural difference between adenophostin A and
Ins(1,4,5)P3 is that the former possesses a large adenosine
2-phosphate moiety. We therefore asked if this moiety was responsible
for adenophostin A's unique ability to preferentially stimulate
Ca2+ influx. After establishing that D-glucose
(n = 3) and glucose 1,6-diphosphate (n = 3) were inactive, we injected the polyphosphorylated D-glucose derivative ADAN 1 (11), which also possesses the
glucose 3,4-bisphosphate moiety of adenophostin A (Fig. 1C).
When injected in high concentrations, ADAN 1 (10
4
M in the pipette, n = 6) caused a biphasic
Ca2+ signal similar to that of adenophostin A (Fig.
6A). However, when injected in threshold
concentrations (10
6 M in the pipette), ADAN 1 did not behave like adenophostin A; it did not cause Ca2+
influx. Instead, ADAN 1 behaved like Ins(1,4,5)P3, causing
an oscillatory release of intracellular Ca2+ (Fig.
6B). The results with ADAN 1 suggest that the glucose
3,4-bisphosphate moiety of adenophostin A is sufficient to release
intracellular Ca2+ but that the adenosine 2
-phosphate
moiety is required for adenophostin A to preferentially stimulate
Ca2+ influx. However, we do not think that the adenosine
2
-phosphate moiety is sufficient to stimulate Ca2+ influx,
because the following compounds, which form an increasing part of the
moiety, neither stimulated Ca2+ release nor elicited
Ca2+ influx: D-ribose (n = 3),
adenine (n = 3), adenosine (n = 4), and
adenosine 2
-phosphate (n = 6).
Being a small phosphorylated polar compound, adenophostin A discloses some of the key properties attributed to the CIF partially purified from Jurkat cells (6, 7, 33). Like adenophostin A, CIF stimulates Ca2+ influx (7) and can potentiate the Ins(1,4,5)P3-induced release of intracellular Ca2+ (34) in the oocyte. Unlike adenophostin A however, CIF's Ca2+ influx-stimulating activity is not inhibited by heparin, and CIF does not release intracellular Ca2+ (7). Thus, if CIF and adenophostin A both interact with the InsP3R, they may do so through different mechanisms. Also, CIF was found to be active when added extracellularly (6), and adenophostin A is not (n = 6, Fig. 5D). Although there are potential explanations for each of the observed functional differences between the two compounds, only the purification of CIF will determine if adenophostin A is a prototypical Ca2+ influx factor.
In summary, our results suggest that adenophostin A can stimulate Ca2+ influx without depleting the Ins(1,4,5)P3-sensitive intracellular Ca2+ stores. Although definitive proof must await a specific InsP3R binding inhibitor, our results also suggest that adenophostin A stimulates Ca2+ influx by binding to an InsP3R. When considered along with the recent finding that overexpression of type 3 InsP3R markedly increases the magnitude of Ca2+ influx without affecting the release of intracellular Ca2+ (35), our results raise the possibility that the InsP3R influences Ca2+ influx through mechanisms that extend beyond its ability to release intracellular Ca2+.