(Received for publication, January 27, 1997, and in revised form, April 29, 1997)
From the D-myo-Inositol
1,4,5-trisphosphate (InsP3) 5-phosphatase and 3-kinase are
thought to be critical regulatory enzymes in the control of
InsP3 and Ca2+ signaling. In brain and many
other cells, type I InsP3 5-phosphatase is the major
phosphatase that dephosphorylates InsP3 and
D-myo-inositol 1,3,4,5-tetrakisphosphate. The
type I 5-phosphatase appears to be associated with the particulate
fraction of cell homogenates. Molecular cloning of the human brain
enzyme identifies a C-terminal farnesylation site CVVQ.
Post-translational modification of this enzyme promotes membrane
interactions and changes in specific activity. We have now compared the
cytosolic Ca2+ ([Ca2+]i) responses
induced by ATP, thapsigargin, and ionomycin in Chinese hamster ovary
(CHO-K1) cells transfected with the intact InsP3
5-phosphatase and with a mutant in which the C-terminal cysteine cannot
be farnesylated. [Ca2+]i was also measured in
cells transfected with an InsP3 3-kinase construct encoding
the A isoform. The Ca2+ oscillations detected in the
presence of 1 µM ATP in control cells were totally lost
in 87.5% of intact (farnesylated) InsP3 5-phosphatase-transfected cells, while such a loss occurred in only
1.1% of the mutant InsP3 5-phosphatase-transfected cells. All cells overexpressing the InsP3 3-kinase also responded
with an oscillatory pattern. However, in contrast to control cells, the
[Ca2+]i returned to base-line levels in between a
couple of oscillations. The [Ca2+]i responses to
thapsigargin and ionomycin were identical for all cells. The four cell
clones compared in this study also behaved similarly with respect to
capacitative Ca2+ entry. In permeabilized cells, no
differences in extent of InsP3-induced Ca2+
release nor in the threshold for InsP3 action were observed
among the four clones and no differences in the expression levels of the various InsP3 receptor isoforms could be shown between
the clones. Our data support the contention that the ATP-induced
increase in InsP3 concentration in transfected CHO-K1 cells
is essentially restricted to the site of its production near the plasma
membrane, where it can be metabolized by the type I InsP3
5-phosphatase. This enzyme directly controls the
[Ca2+]i response and the Ca2+
oscillations in intact cells.
Hydrolysis of phosphatidylinositol 4,5-bisphosphate by
phospholipase C produces the second messengers
D-myo-inositol 1,4,5-trisphosphate (InsP3)1 and diacylglycerol,
which function in mobilization of intracellular Ca2+ and
activation of protein kinase C, respectively (1). InsP3 is
metabolized by an InsP3 5-phosphatase to produce inositol
1,4-bisphosphate, which is inactive in Ca2+ mobilization
(2). In addition, InsP3 can be phosphorylated to
D-myo-inositol 1,3,4,5-tetrakisphosphate
(InsP4) by an InsP3 3-kinase (3). Depending on
the cell type, different effects of InsP4 have been
reported, some of which are directly or indirectly related to
Ca2+ homeostasis (reviewed in Refs. 4 and 5). In sea urchin eggs, in perfused lacrimal cells, and in permeabilized mouse lymphoma cells, synergistic effects of InsP4 on
InsP3-stimulated Ca2+ release have been
reported (6-8). However, in mouse lacrimal acinar cells, high
concentrations of InsP4 were reported to inhibit the
inositol 2,4,5-trisphosphate-induced Ca2+ entry (9). In
mouse exocrine pancreatic acinar cells, InsP4 is effective
in mobilizing Ca2+ and promoting Ca2+ influx in
conditions where phospholipase A2 activity, and thus arachidonic acid production, is inhibited (10). Recently, a high
affinity InsP4-binding protein has been purified from pig platelets (11). The cDNA encoding this putative receptor encoded a
GTPase-activating protein. In vitro, it shows
GTPase-activating protein activity against Ras that is inhibited by
phospholipids and specifically stimulated by InsP4
(12).
Both the InsP3 3-kinase and the 5-phosphatase consist of a
family of multiple isoenzymes. In particular, the various
5-phosphatases dephosphorylate InsP3, InsP4,
phosphatidylinositol 4,5-bisphosphate, and phosphatidylinositol
3,4,5-trisphosphate with different specificities, Km, and Vmax (13-15).
In vitro, type I InsP3 5-phosphatase (also
designated as the 43-kDa 5-phosphatase) can use either
InsP3 or InsP4 as substrates but not the
phosphoinositides (16-20). Based upon kinetic analysis
(Km for InsP3 and
Vmax values compared with other 5-phosphatases),
this enzyme is generally considered as a signal-terminating enzyme for
InsP3, analogous to the cyclic nucleotide
phosphodiesterases for cyclic AMP (21).
The type I 5-phosphatase has a higher affinity but lower
Vmax for InsP4 as compared with
InsP3, suggesting that it could be rapidly saturated with
InsP4, particularly in brain, which produces more
InsP4 than any other tissue (22). In most tissues,
particularly in brain, total InsP3 5-phosphatase activity
measured in homogenates is associated with the particulate fraction
(13, 23). This activity is predominantly due to type I, as shown by
Western blot analysis and by immunoprecipitation of the activity in
particulate fractions of various tissues by using specific antibodies
(18, 24). Molecular cloning of the dog and human type I
InsP3 5-phosphatase identifies a 412-amino acid protein and
a C-terminal farnesylation site, CVVQ (25-27). Removal of this
isoprenylation site shifted the protein from largely particulate to
soluble, and immunofluorescence analysis with confocal microscopy
showed a shift from the plasma membrane to the cytosol. This was
evidenced by generating and using two different mutants in the
isoprenylation motif and transfecting them in COS-7 cells (28).
Different cells show a different activity ratio between the
InsP3 5-phosphatase and 3-kinase activities,
e.g. high levels of InsP3 3-kinase A are found
in the rat or human hippocampal CA1 pyramidal cells, while no
detectable InsP3 3-kinase activity (and therefore
intracellular InsP4 formation) could be shown in ram sperm
(29, 30). In rat AR4-2J pancreatoma cells, InsP3 3-kinase
activity could only be seen in high speed supernatants of cellular
homogenates, and these cells may actually overexpress InsP3
5-phosphatase activity as compared with non-tumoral acinar cells (31).
Finally, InsP3 3-kinase activity is modulated by the
Ca2+-calmodulin complex; the extent of this modulation
varies between the various isoenzymes in different cell types
(32-37).
Many cell types respond to agonists acting via the
phosphoinositide signaling pathway by the generation of
Ca2+ oscillations (38). Although InsP3 plays a
key role in the various models that have been proposed to explain this
behavior, the influence of the InsP3-metabolizing
activities on the oscillatory cytosolic calcium
([Ca2+]i) response in single cells is largely
unknown. Does the InsP3 3-kinase or the 5-phosphatase
control the [Ca2+]i oscillations generated in
response to an agonist in intact cells? Could isoprenylation of the
type I InsP3 5-phosphatase and its targeting to the cell
surface influence the [Ca2+]i response? These
questions were addressed in the present study by comparing, for the
first time, the [Ca2+]i responses induced by ATP
in CHO-K1 cells transfected with the intact InsP3
5-phosphatase and with a mutant in which the C-terminal cysteine cannot
be isoprenylated. The data have been compared with CHO-K1 cells
transfected with an InsP3 3-kinase construct encoding the
human brain A isoform.
pcDNA3 vector is an expression vector
developed by Invitrogen. Radioactive products, including
[3H]InsP3 (40-60 Ci/mmol for binding
assays), [3H]InsP4 (15-30 Ci/mmol), and
[32P]InsP4 were obtained or prepared as
described previously (20, 28, 35, 39, 40). Anti-rabbit and anti-mouse
Vistra ECF Reagent Packs were obtained from Amersham International plc.
ATP, UTP, Pefabloc, and leupeptin were from Sigma. Cell culture medium and antibiotics were from Life Technologies, Inc.
Stable CHO-K1 cell
lines overexpressing the InsP3-metabolizing enzymes were
generated as described previously (28). The human InsP3
3-kinase A coding sequence (36) was subcloned using the
BamHI and EcoRI restriction sites of the
pcDNA3 expression vector for transfection into CHO-K1 cells. The
DNA corresponding to the InsP3 3-kinase was transfected
into CHO-K1 cells using the calcium phosphate precipitation method
(41). Cells were transfected in 6-cm diameter culture dishes. Two days
after transfection, selection of transfected cells was started by
addition of fresh complete medium (Ham's F-12 medium supplemented with
10% fetal calf serum, 1% fungizone, and 2% penicillin/streptomycin)
containing 400 µg/ml Geneticin G418. After death of all
non-transfected cells, 16 G418-resistant clones were isolated and
transferred into 9-cm diameter culture dishes. Medium was changed every
48 h. The CHO-K1 cells were maintained in complete medium
containing Geneticin in a 5% CO2 atmosphere at 37 °C.
Out of 16 clones tested, 4 were positive with high InsP3
3-kinase activity. One of these clones was clone A4. A construct in
pcDNA3 encoding a C-terminal mutant InsP3 5-phosphatase
with a serine replacing a cysteine at position 409 of the wild type
enzyme (C409S) has been described (28). The pcDNA3 vector alone, as
well as the intact and the mutant InsP3 5-phosphatases were
transfected in CHO-K1 cells as described above.
CHO-K1 (10 × 106) cells were resuspended in 400 µl of homogenization
buffer A containing 15 mM Tris/HCl, 2 mM
MgCl2, 0.3 mM EDTA, 1 mM EGTA, 2.5 µM leupeptin, and 0.4 mM Pefabloc (pH 7.5). The cells were homogenized by passage through a 26-gauge needle (approximately 10 times). For the subcellular distribution of activities, the cell lysate was centrifuged at 80,000 × g for 30 min at 4 °C and the pellet resuspended in the
same volume as the original homogenate (17, 28). InsP3
5-phosphatase and 3-kinase activities were measured as reported
previously (23). InsP3 3-kinase was determined in the
presence of 0.1 µM calmodulin and either 1 mM
EGTA or 10 µM free Ca2+ (34).
To label the phosphoinositides, the CHO-K1 cells were
incubated for 24 h in inositol-depleted medium supplemented with
myo-[3H]inositol (10 µCi/ml). When the
medium was removed, the cells were washed twice in KRH solution (4.97 mM KCl, 1.24 mM
MgSO4.7H2O, 124 mM NaCl, 25 mM Hepes, 8.1 mM glucose, 1.3 mM
KH2PO4, and 1.5 mM
CaCl2.2H2O, pH 7.5). This was followed by the
addition of fresh medium containing the agonist for 10 s to 1 min.
The incubation was stopped by removal of the medium, followed by the
addition of 2 ml of perchloric acid (3%). The cells were scrapped and
the dishes rinsed with 2 ml of perchloric acid (1%). After a 30-min incubation at 4 °C, the solution containing the inositol phosphates was neutralized with 1 ml of Hepes (0.375 mM) and about 1.3 ml of KOH (0.765 M) to a final pH of 7.7-7.8. After 30 min
on ice, the supernatant was collected and added to
Na2B4O7 (5 mM) and EDTA
(0.5 mM) (42). The separation of inositol phosphates was achieved by chromatography on Dowex AG 1-X8 (formate form,
100-200-µm mesh) column. By this method, we do not separate the
various inositol phosphate isomers in each fraction. Inositol,
glycerophosphoinositol, InsP1, InsP2,
InsP3, and InsP4 were eluted sequentially with: 20 ml of water; 8 ml of 60 mM ammonium formate, 5 mM Na2B4O7; 20 ml of
150 mM ammonium formate, 5 mM
Na2B4O7; 18 ml of 400 mM ammonium formate, 0.1 M formic acid; 12 ml
of 700 mM ammonium formate, 0.1 M formic acid;
and 24 ml of 1.5 M ammonium formate, 0.1 M
formic acid. The radioactivity of each fraction was estimated by liquid
scintillation. The cell debris in the bottom of the dishes were
dissolved in 1 M NaOH and counted as phosphoinositides (PI). The radioactivity associated to each fraction was normalized to
the radioactivity present in the PI fraction (43).
Unidirectional
45Ca2+ fluxes on confluent monolayers of
permeabilized CHO-K1 cells were performed on a thermostated plate at
25 °C, basically as described previously for smooth muscle cells
(44, 45). In brief, the culture medium was aspirated and replaced by 1 ml of permeabilization medium containing 120 mM KCl, 30 mM imidazole-HCl (pH 6.8), 2 mM
MgCl2, 1 mM ATP, 1 mM EGTA, and 20 µg ml Wild-type and transfected cells
were cultured up to confluence in Doubletray Units (1200 cm2, Nunc). From each unit, 200-250 × 106 cells were harvested by mild trypsinization. A total
microsomal fraction was prepared as described previously for several
other cell lines (40). This particulate fraction from CHO-K1 cells contained 25 ± 1% (n = 11) of the cellular
proteins. This amounted to 43 ± 2 (n = 14) mg
protein/109 cells. No differences between the clones were
observed.
[3H]InsP3 binding was measured with a rapid
filtration assay using 4 nM
[3H]InsP3 and 200 µg of protein/filter, as
published previously (40). Specific binding amounted for all samples to
more than 95% of the total binding activity.
InsP3R isoform determination was performed after SDS-PAGE
(3-12% gradient gels) and Western blotting. Previously characterized isoform-specific antibodies against InsP3R-I (40),
InsP3R-II (40), and InsP3R-III (47) were used.
Detection was performed using secondary antibodies coupled to either
horseradish peroxidase and using 3,3 Single-cell
[Ca2+]i measurements were performed with a
laser-scanning MRC-1000 system (Bio-Rad, Hertfordshire, UK) attached to
an inverted Nikon Diaphot 300 epifluorescence microscope with a CF
Fluor 40 × (NA = 1.3) oil immersion objective (48). Briefly,
after removing the culture medium and washing the cells, they were
incubated for 30 min with 5 µM Indo-1-AM dissolved in a
modified Krebs solution of the following composition: 135 mM NaCl, 5.9 mM KCl, 1.5 mM
CaCl2, 1.2 mM MgCl2, 11.6 mM Hepes, and 11.5 mM glucose (pH 7.3). The
cells were then further incubated for at least 1 h in the absence
of Indo-1. During the experiment, the cells were continuously
superfused from a pipette placed on top of the cell. In the experiments
where Ca2+ was omitted from the medium, 2 mM
EGTA was added. All experiments were performed at room temperature.
Results were always expressed as ratios of emitted fluorescence.
Scanning was done each 0.25 s, with a scanning box containing
384 × 256 lines. The signals were not averaged.
CHO-K1 cell lines overexpressing the
intact type I InsP3 5-phosphatase (clone C1), the
C-terminal cysteine mutant of the type I InsP3
5-phosphatase (clone 61), or the InsP3 3-kinase A (clone A4) were generated as described under "Experimental Procedures." Table I shows the enzymatic activities that were
measured in the different cell lines: the InsP3
5-phosphatase activity was increased by 15-25-fold in
InsP3 5-phosphatase-overexpressing cells as compared with
control (wild type CHO-K1) cells. The InsP3 3-kinase
activity was increased by 9-fold in CHO-K1 cells overexpressing the
InsP3 3-kinase. Since the type I InsP3
5-phosphatase could use either InsP3 or InsP4
as substrates, cells transfected with the 5-phosphatase constructs also
showed increased InsP4 5-phosphatase activity as compared
with wild type cells (Table I). In the two clones transfected with
InsP3 5-phosphatase constructs, the InsP3 3-kinase activity was not detectable because either the substrate or
the product of the InsP3 3-kinase reaction were
dephosphorylated despite the presence of 2,3-bisphosphoglycerate (an
InsP3 5-phosphatase inhibitor; Ref. 35) in our assay
buffer.
Table I.
InsP3, InsP4 5-phosphatase, and InsP3 3-kinase
activities in wild type and transfected CHO-K1 cells
Institute of Interdisciplinary Research,
Laboratory of Physiology,
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
Materials
1 saponin. The saponin-containing solution was
removed after 10 min, and the cells were washed once with a similar
saponin-free solution. 45Ca2+ uptake into the
non-mitochondrial Ca2+ stores was accomplished by
incubation for 45 min in 1 ml of loading medium containing 120 mM KCl, 30 mM imidazole-HCl (pH 6.8), 5 mM MgCl2, 5 mM ATP, 0.44 mM EGTA, 10 mM NaN3, and 100 nM free 45Ca2+. After this phase of
45Ca2+ accumulation, the monolayers were
incubated in 1 ml of efflux medium containing 120 mM KCl,
30 mM imidazole-HCl (pH 6.8), 1 mM ATP, 1 mM EGTA, and 2 µM thapsigargin. The efflux
medium was replaced at fixed time intervals (6 s or 2 min). Replacing
the medium each 6 s has the advantage that it permits rapid
changing of the InsP3 concentration. This allows the
accurate measurement of the threshold for InsP3-induced
Ca2+ release (46). At the end of the experiment, the
45Ca2+ remaining in the stores was released by
incubation in 1 ml of a 2% SDS solution for 30 min.
diaminobenzidine as substrate or
coupled to alkaline phosphatase and using Vistra ECF as substrate. The
latter was used according to the manufacturer's instructions, and the
signal was quantified using the Storm 840 FluorImager equipped with the ImageQuant NT4.2 software (Molecular Dynamics Inc.). Statistical analysis of the results was performed using the paired or unpaired Student's t test, as applicable.
InsP3 5-Phosphatase and 3-Kinase Activities in
Transfected CHO-K1 Cells
1 × (mg of protein)
1 ± S.E. are from one
representative experiment out of a total of three. ND, not detectable.
Cell type
InsP3 5-phosphatase
activity
InsP4 5-phosphatase activity
InsP3
3-kinase activity
CHO-K1 wild type
5.2
± 0.3
1.6 ± 0.1
2.5 ± 0.2
CHO-K1 A4 (InsP3
3-kinase)
4.9 ± 0.5
0.9 ± 0.1
23.6 ± 0.4
CHO-K1 C1 (InsP3 5-phosphatase)
128.4
± 7.2
57.2 ± 3.0
ND
CHO-K1 61 (mutant InsP3 5-phosphatase)
90.8 ± 3.1
45.2
± 1.5
ND
The InsP3 3-kinase activity associated to InsP3 3-kinase A appeared in high speed supernatants of rat or human brain homogenates (35, 36). Western blotting with antibodies to the rat brain 3-kinase A revealed that the enzyme was both in the soluble and particulate fractions of cell homogenates in transfected CHO-K1 cells (49). The 43-kDa 5-phosphatase is targeted to the membrane by farnesylation (28). As expected, 85% of the total phosphatase activity was in the particulate fraction in cells overexpressing the intact InsP3 5-phosphatase, whereas 77% of the total activity was soluble in homogenates of cells transfected with the C-terminal mutant InsP3 5-phosphatase (data from a typical experiment out of two). A similar redistribution from particulate to soluble fractions has been observed before in homogenates of COS-7 cells transfected with similar constructs (28).
Inositol Phosphate Formation in Transfected CHO-K1 CellsATP
stimulated the formation of [3H]inositol phosphates in
transfected CHO-K1 cells prelabeled with
myo-[3H]inositol. All these experiments
were performed in the absence of LiCl to avoid a preferential
accumulation of inositol 1,3,4-trisphosphate (42). In cells transfected
with the pcDNA3 vector alone, [3H]InsP3
was increased in response to 100 µM ATP (Fig.
1, A, D, and G). Wild
type CHO-K1 cells behaved as cells transfected with the pcDNA3
vector alone (data not shown). The increase in
[3H]InsP3 was markedly reduced in the
5-phosphatase (both clones C1 and 61) and in the 3-kinase (clone
A4)-transfected cells (Fig. 1, A, D, and
G, respectively). After a short time (10 s), the level of
[3H]InsP4, in 3-kinase-transfected cells, was
about 200% of the control value (Fig. 1I). No increase in
[3H]InsP4 could be detected in the
5-phosphatase-transfected cells (Fig. 1, C and F,
respectively). In contrast, a small increase in
[3H]InsP2 could be measured in the same
cells; at 10 s, the levels of [3H]InsP2
were 150% and 135% of the control value (Fig. 1, B and E, respectively). We thus conclude that transfected cells
metabolize InsP3 in a manner that is consistent with the
increase in InsP3-metabolizing activities detected in crude
homogenates (Table I).
InsP3-induced Ca2+ Release and InsP3R Expression in Wild Type and Transfected CHO-K1 Cells
To analyze the properties of InsP3-induced
Ca2+ release in CHO-K1 cells, unidirectional
45Ca2+ efflux experiments on permeabilized
cells were performed. The non-mitochondrial Ca2+ stores in
saponin-permeabilized CHO-K1 cells were loaded to steady state with
Ca2+ and then challenged with 10 µM
InsP3 or 10 µM A23187 to induce Ca2+ release, as indicated by the bars in the
left panels of Fig. 2. The maximal increase
in the rate of Ca2+ release induced by this maximal
[InsP3] relative to that induced by the
Ca2+-ionophore was similar for the four clones. The
threshold for InsP3 action was determined by a technique in
which the [InsP3] was gradually increased from 10 nM to 3.2 µM (right panels of Fig.
2). The lowest [InsP3], at which the rate of
Ca2+ release significantly increased above base-line
levels, gives an indication of the threshold. A threshold of
approximately 30 nM was determined for the various clones,
indicating a similar sensitivity of the Ca2+ stores for
InsP3 and thus presumably a comparable sensitivity of the
InsP3Rs.
The expression level of the InsP3R in wild type and transfected CHO-K1 cells was further investigated by two independent methods. First, an estimation of the density of the InsP3Rs was performed with a classical ligand-binding assay (Table II, A). Using [3H]InsP3 as specific ligand, the microsomal InsP3 binding sites were detected. Since the amount of microsomes obtained per cell was identical in the various clones investigated (25% of the cellular proteins), the measured number of binding sites is roughly equivalent to their density. This value, when measured at a concentration of 4 nM [3H]InsP3, amounted in average to 65 fmol of bound InsP3/mg of microsomal protein in wild type cells. Binding was not statistically different between the four cell clones (Table II, A). It should be emphasized that those values are not equivalent to the Bmax values. Indeed, the three different InsP3R isoforms have different affinities for their ligand (40, 50, 51), and it was recently demonstrated that CHO-K1 cells not only express all three isoforms but also allow their co-assembly in heterotetramers (52), making a correlation between InsP3 binding levels and the concentrations of the various InsP3R isoforms difficult. A second technique was therefore used to determine the expression level of the various isoforms. This method involved the use of isoform-specific antibodies on Western blots and the quantification of the resulting fluorescent signal detected by fluorimaging (Table II, B). The levels of the various InsP3R isoforms were not significantly different between the clones. Most importantly, no differences in expression levels were observed either between the C1 and the A4 cells (overexpressing InsP3 5-phosphatase and InsP3 3-kinase, respectively) or between the C1 and the 61 cells (overexpressing intact and mutant InsP3 5-phosphatase, respectively; Table II, B).
|
Addition of ATP or UTP resulted in a rapid increase in
[Ca2+]i in CHO-K1 cells, which was mediated by
P2y2 receptors (53). Fig. 3 illustrates the
effect of 100 µM ATP in a medium containing 1.5 mM external Ca2+ on the four different cell
types. The responses of the wild type cells were very consistent,
because 99.4% of the cells showed an initial rapid
[Ca2+]i rise, which then declined to a plateau
that was maintained as long as the agonist was present. The
[Ca2+]i returned to base line as soon as ATP was
washed away. This pattern is typical for an initial Ca2+
release from internal stores, followed by a subsequent phase of influx
of external Ca2+. Only 0.6% of the cells presented a
slightly different pattern, with a [Ca2+]i rise
that came back to base line and then showed a single Ca2+
peak. Cells transfected with the pcDNA3 vector alone responded in a
similar way to 100 µM ATP as the wild type CHO-K1 cells
(data not shown).
Cells overexpressing the InsP3 3-kinase activity responded in a much similar way as the wild type cells, i.e. with a plateau that was maintained as long as the ATP was present (93.0% of the cells). However, 7.0% of the cells had a [Ca2+]i rise that was less well maintained. This percentage was slightly higher than in control cells (0.6%).
Cells overexpressing the InsP3 5-phosphatase showed a totally different response to 100 µM ATP. 50.1% of the cells responded with only one short-lasting Ca2+ transient with no subsequent plateau. 32.3%, 11.6%, and 4.3% of the cells showed respectively two, three and four Ca2+ transients. A short-lasting plateau was only seen in 0.8% of the cells. This plateau was, however, made up of fused Ca2+ oscillations and therefore differed from the plateau observed in the wild type cells and in those transfected with the InsP3 3-kinase. 0.9% of the cells failed to respond. The Ca2+ signals in the cells transfected with the mutant InsP3 5-phosphatase were totally different from those in the cells transfected with the intact InsP3 5-phosphatase. 91.3% of the cells responded with an initial Ca2+ peak and a long-lasting plateau. Their responses therefore resembled those observed in the wild type and InsP3 3-kinase-overexpressing cells. 5.8% of the cells showed a plateau that developed into a Ca2+ oscillation, which is a behavior that was also occasionally observed in the wild type cells. Only 2.9% of the mutant InsP3 5-phosphatase-transfected cells responded with a more transient plateau.
The temporal patterns of the Ca2+ signals differed between the various cell types. To analyze possible differences in magnitude of the responses, we converted the fluorescence ratios to [Ca2+]i in some of the experiments. The initial Ca2+ peak in each of the cell clones exceeded 1 µM (data not shown). This value is far above the Kd for Indo-1, thus making exact estimates of the [Ca2+]i less accurate. For this reason, we choose not to convert the ratiometric values to concentration units.
Effect of a Maximal ATP Concentration on the [Ca2+]i in Ca2+-free SolutionSince the [Ca2+]i rises in response to ATP in Fig. 3 were obtained in the presence of 1.5 mM external Ca2+, they represented the effect of both Ca2+ entry and internal Ca2+ release. To assess the effect of overexpression of the InsP3 3-kinase and InsP3 5-phosphatase on the Ca2+ release in the absence of Ca2+ entry, the experiments of Fig. 3 were repeated in the absence of extracellular Ca2+ in a medium containing 2 mM EGTA.
Fig. 4 illustrates that 100% of the wild type cells
responded to 100 µM ATP with an initial Ca2+
peak, after which the [Ca2+]i spontaneously
declined to its resting level. A more-or-less long-lasting plateau was
always observed. A similar response occurred in 100% of the
InsP3 3-kinase-overexpressing cells. Again, the [Ca2+]i responses in the InsP3
5-phosphatase-overexpressing cells were very different, since they
consisted of one (63.4%) or more (28.8%) sharp Ca2+
spikes. An abortive Ca2+ spike was occasionally observed.
Only 0.5% of the cells showed a more sustained response, while 7.3%
of the cells failed to respond. 82.2% of the cells overexpressing the
mutant InsP3 5-phosphatase responded like the wild type and
InsP3 3-kinase-overexpressing cells, i.e. with
an initial Ca2+ peak, a more-or-less long-lasting plateau,
and a subsequent decrease of the [Ca2+]i to its
resting level. In some cases, this plateau was replaced by a series of
Ca2+ oscillations (14.1%). Only a minority of the cells
(0.7% and 3.0%) showed more transient responses.
Effect of Agonist-independent Stimuli for Ca2+ Release in Intact Cells
To prove that the more transient responses in the
intact InsP3 5-phosphatase-overexpressing cells were due to
InsP3 metabolism and not to differences in the filling
state of the stores, we have studied the Ca2+ content of
the latter using stimuli that mobilized Ca2+ independently
of the InsP3 signaling pathway. The left panels of Fig. 5 show the effects of 2 µM
thapsigargin in Ca2+-free solution. The InsP3
5-phosphatase-transfected cells behaved similarly as the other cell
types. The same phenomenon was observed for the responses to 10 µM ionomycin (right panels in Fig. 5). These
findings exclude the possibility that the different responses to 100 µM ATP were caused by different levels of
Ca2+ content of the stores.
Effect on Capacitative Ca2+ Entry
The differences
in agonist-induced [Ca2+]i responses between the
four different cell types occurred both in the presence and absence of
extracellular Ca2+ (Figs. 3 and 4), suggesting that they
represented different patterns of internal Ca2+ release and
not possible effects on Ca2+ entry. To more directly assess
Ca2+ entry, thapsigargin (2 µM) was added for
1 h in the presence of 1.5 mM Ca2+ to
activate the capacitative Ca2+ entry pathway independently
of receptor activation (54). The contribution of Ca2+ entry
was then assessed by removing extracellular Ca2+ and then
adding Ca2+ back (48). Fig. 6 indicates
that, with this procedure, Ca2+ entry could be demonstrated
in the four different cell types. No differences were, however,
observed between these cell lines.
Effect of a Submaximal ATP Concentration on the [Ca2+]i
Fig. 7
illustrates the effect of 1 µM ATP in a medium containing
1.5 mM external Ca2+. 100% of the wild type
cells responded with an oscillatory response. The
[Ca2+]i never dropped to the resting level during
the falling phase of a Ca2+ spike, but only returned to
resting levels after removing the ATP. Cells transfected with the
pcDNA3 vector alone responded in a similar way to 1 µM ATP as the wild type CHO-K1 cells (data not shown).
All cells overexpressing the InsP3 3-kinase also responded with an oscillatory pattern. However, in contrast to the oscillations in the wild type cells, the [Ca2+]i returned to
base-line levels between oscillations.
Cells transfected with the InsP3 5-phosphatase again showed a completely different response. The large majority of the cells (87.5%) did not respond. 5.5% of the cells only showed an abortive Ca2+ spike. Only 7.0% of the cells responded with one or more Ca2+ spikes. In contrast, the cells transfected with the mutant InsP3 5-phosphatase again responded much like the wild type cells. 95.6% of the cells responded with an oscillatory [Ca2+]i response, during which the [Ca2+]i never dropped to the resting level. In 3.3% of the cells the [Ca2+]i did return to resting levels during the application of ATP, while 1.1% of the cells did not respond.
In many cells, the intracellular Ca2+ signal is spatio-temporally organized, and consists of Ca2+ oscillations and Ca2+ waves. Oscillations of [Ca2+]i occur in stimulated cells in response to hormones, neurotransmitters or growth factors, which all induce the formation of InsP3 (1, 38). Various models have been proposed to explain the observed [Ca2+]i oscillations (55). In general, the metabolism of InsP3 is not considered. Only in one report, the InsP3 5-phosphatase and the Ca2+-sensitive InsP3 3-kinase have been introduced in a model that may generate different patterns of Ca2+ oscillations (56). InsP3 5-phosphatase activity is predominantly membrane-associated and can be purified as a 43-kDa protein from either soluble or particulate fractions of tissue homogenates (13, 18, 23, 57). The 43-kDa InsP3 5-phosphatase shows a putative C-terminal isoprenylation site CVVQ, suggesting a mechanism for membrane attachment (25-28). In contrast, previous studies all suggested InsP3 3-kinase A to be cytosolic. InsP3 3-kinase A was initially purified from rat brain cytosolic fraction as a 50-53-kDa protein (6, 34, 35, 58, 59). Antibodies to purified InsP3 3-kinase recognized a 50-53-kDa protein band in a soluble fraction of rat brain homogenates. cDNAs encoding the rat (and human) brain InsP3 3-kinase A have been isolated (39, 60). The encoded protein of 459 amino acids has a calculated molecular weight of about 51,000, comparable with estimation for the native enzyme. This finding indicates that the native protein purified from a soluble fraction is not a proteolytic product from a larger polypeptide form.
The data in the present study indicate that the farnesylated type I InsP3 5-phosphatase, which is targeted to the plasma membrane (28), modifies or abolishes the [Ca2+]i responses to ATP in cells overexpressing this enzyme. First, the Ca2+ oscillations in the presence of 1 µM ATP were totally lost in 87.5% of intact InsP3 5-phosphatase-transfected cells, while such a loss occurred in only 1.1% of the mutant InsP3 5-phosphatase-transfected cells. Second, at 100 µM ATP, the more or less sustained [Ca2+]i responses in the control (wild type) cells and in the cells overexpressing either the InsP3 3-kinase or the mutant InsP3 5-phosphatase were relatively similar, but differed to a large extent from the spiking pattern observed in the cells transfected with the intact InsP3 5-phosphatase. This pattern difference was clearly observed both in the presence and in the absence of extracellular Ca2+.
Differences in InsP3 5-phosphatase activities detected in homogenates and in vitro could not account for the observed divergent [Ca2+]i responses, since both the intact and the mutant InsP3 5-phosphatase-overexpressing cells showed very high enzymatic activities. The two InsP3 5-phosphatase-transfected cells differed, however, by the distribution of activities between soluble and particulate fractions. In transfected COS-7 cells, uniform immunofluorescence labeling was observed throughout the cytosol for the InsP3 5-phosphatase mutant, whereas the intact InsP3 5-phosphatase was found along the plasmalemma (28).
The divergent [Ca2+]i responses were not related to possible effects on the filling state of the stores: first, because the responses to thapsigargin and ionomycin in intact cells were indistinguishable for the various clones; and second, because the effects of a maximal concentration of InsP3 or of Ca2+ ionophore in permeabilized cells were not different.
It is known that prolonged stimulation of the cells by agonists, and the concomitant high InsP3 and Ca2+ levels, can induce down-regulation of the InsP3R (61, 62). Similarly, constitutive activation of phospholipase C in stable NIH-3T3 transfectants led to a decrease in the amount of immunodetectable InsP3R-I (63). Calpain activation by Ca2+ plays a role in the relatively fast down-regulation phenomenon occurring during chronic cell stimulation (62), but the various mechanisms that can regulate the expression levels of the different InsP3R isoforms are not yet elucidated.
Since the InsP3 levels in the transfected cells are shown to differ from those in wild type cells, we investigated the properties of the InsP3-induced Ca2+ release process and the expression levels of the various InsP3R isoforms in the four cell clones. No differences in magnitude or in sensitivity of the InsP3-induced Ca2+ release were detected in permeabilized cells. Most importantly, no differences in threshold for InsP3 action were detected between the different clones. We previously demonstrated that a number of physiological regulators of the InsP3R (46) as well as cell type-specific differences in isoform expression (47) modulate this threshold value.
The InsP3 binding activity observed in CHO-K1 microsomes was similar in level to the activity detected under the same conditions in some other cell types like A7r5 smooth muscle cells, C3H10T1/2 fibroblasts, or Jurkat leukemia cells (40, 64). Importantly, no significant differences between the various CHO-K1 cell clones were found. Moreover, there were no differences in InsP3R isoform expression either between the intact InsP3 5-phosphatase and the InsP3 3-kinase-transfected cells or between the intact and the mutant InsP3 5-phosphatase-transfected cells. This contrasted with their [Ca2+]i response to both high and low concentrations of ATP. For all these reasons, it is clear that the differences in Ca2+ signaling observed between the clones are not due to variations in the expression levels of the various InsP3R isoforms.
All these findings make it therefore very likely that the observed differences in [Ca2+]i responses represent effects due to the changes in InsP3 metabolism. We suggest that in CHO-K1 cells, the ATP-induced increase in InsP3 concentration is essentially restricted to the site of its production near the plasma membrane, where it can be metabolized by the type I InsP3 5-phosphatase. In cells transfected with the intact InsP3 5-phosphatase (farnesylated and therefore targeted to the membrane), the catabolism of InsP3 is obviously more efficient and the InsP3 is impaired in reaching its receptor on the Ca2+ stores. This leads to a drastic reduction in Ca2+ signaling capability. InsP3 is metabolized less effectively by the mutant InsP3 5-phosphatase or by the InsP3 3-kinase, both of which are not membrane-associated. This will allow the InsP3 to activate the InsP3R on Ca2+ stores, which may even be located at the cell periphery, and this explains why the [Ca2+]i responses in the latter clones are more in line with the [Ca2+]i responses observed in wild type cells.
Under conditions where InsP3 degradation is inhibited, InsP3 diffuses rapidly and even much faster than Ca2+ (65). It was therefore suggested that InsP3 had a much larger domain of messenger action than Ca2+ (65). In vivo, however, metabolic enzymes may limit the useful range of InsP3 action (66). Ca2+ imaging in polarized epithelia revealed that ATP stimulation induced ipsilateral responses that did not propagate to the other side of the cell (67). InsP3 generation and catabolism were thus colocalized to a cellular domain of the stimulated membrane. The interpretation of Paradiso et al. (67), that InsP3 was generated locally and did not diffuse as a result of local degradation, is entirely consistent with our observations.
Finally, the four cell clones that have been compared in this study behaved similarly with respect to capacitative Ca2+ entry. Cells transfected with the InsP3 3-kinase showed the same [Ca2+]i response to thapsigargin as cells transfected with the InsP3 5-phosphatase. Overexpression of the InsP3 3-kinase had also no effect on thapsigargin-induced Ca2+ entry in Xenopus oocytes (68). These findings argue against a possible role of InsP4 on Ca2+ influx, at least in these cell types. In a previous study, [Ca2+]i responses measured in a cell population of fibroblasts transfected with rat InsP3 3-kinase A did not show any evidence of enhanced Ca2+ mobilization, entry or sequestration (69). In their study, however, Balla et al. (69) did not find any oscillation in [Ca2+]i in response to high or low doses of agonist; therefore, their results could not be directly compared with ours.
Our data indicate that type I InsP3 5-phosphatase directly controls the [Ca2+]i response and the Ca2+ oscillations in intact cells. Thus any change in the activity of this isoenzyme would lead to changed patterns of cell signaling because of an effect on the [Ca2+]i response. For example, the InsP3/InsP4 5-phosphatase activity of human immunodeficiency virus-infected T-helper cells was dramatically reduced during the progression of AIDS but the [Ca2+]i responses were not measured in these reports (70). More recently, underexpression of type I InsP3 5-phosphatase in antisense-transfected rat kidney cells was associated with cellular transformation. In this study, basal Ca2+ levels in cell monolayers increased by underexpressing the 5-phosphatase. The cells also grew faster than control cells (71). Whether the oscillatory [Ca2+]i response is affected in these cells was not determined. Our data suggest that it will be.
In conclusion, the targeting of type I InsP3 5-phosphatase to the membrane is particularly important for the modulation of the intracellular Ca2+ signal. In this context, prenylation of Ras proteins is essential for their biological activity. Mutation of the CAAX box cysteine abrogates cellular transformation by oncogenically activated Ha-Ras (72). Prenylation of Ras proteins promote membrane and protein-protein interactions. The possible interaction of type I InsP3 5-phosphatase with other membrane proteins is currently being studied in our laboratory.
We gratefully acknowledge the technical assistance of A. Florizoone, M. Crabbé, I. Willems, L. Heremans, and Y. Parijs. We also thank S. Swillens for helpful discussions and C. Moreau, L. Drayer, and X. Pesesse for kind assistance in many experiments.