(Received for publication, September 9, 1996, and in revised form, January 29, 1997)
From the Institute of Pharmacology, University of Bern, CH-3010 Bern, Switzerland
To assess the role of G16, a
trimeric G protein exclusively expressed in hematopoietic cells,
G16 antisense RNA was stably expressed in human
erythroleukemia (HEL) cells. Western blot analysis showed that in
transfected cell lines, the expression of endogenous G
16
protein was suppressed, but the expression of G
q/11, G
i2, and G
i3 remained unaffected.
Suppression of G
16 in transfected HEL cells did not
interfere with transient elevations of intracellular free
Ca2+ concentrations induced by prostaglandin E1
(PGE1), platelet-activating factor, or thrombin. In
parental HEL cells, UTP and ATP mobilized Ca2+ from
intracellular stores with half-maximum effective concentrations of
3.6 ± 0.7 and 4.7 ± 1.6 µM, respectively,
apparently by stimulating P2U purinoceptors. By contrast,
Ca2+ mobilization by UTP or ATP was completely abrogated in
G
16-suppressed cells, indicating specific coupling of
G16 to P2U purinoceptors. Pertussis toxin
inhibited the effect of UTP in parental HEL cells by 57.6 ± 4.9%. These data indicate that signaling by the P2U purinoceptor obligatorily requires G16 but may be modulated
further by activation of Gi. Priming of HEL cells with UTP
or ATP prior to stimulation with PGE1 markedly enhanced the
PGE1-induced intracellular Ca2+ release. This
indirect, potentiating effect of UTP and ATP was not impaired in
G
16-suppressed cells but was inhibited by pertussis
toxin, indicating that functional P2U purinoceptors are
present on these cells and that the potentiating effect primarily depends on Gi. The data demonstrate (i) that
G
16 antisense RNA selectively inhibits endogenous
G
16 protein expression in HEL cells; (ii) that
stimulation of endogenous P2U (P2Y2)
purinoceptors leads to the mobilization of intracellular
Ca2+ by a mechanism that strictly depends on
G
16; and (iii) that P2U purinoceptors in HEL
cells can communicate with two distinct signaling pathways diverging at
the G protein level.
G proteins, a group of heterotrimeric membrane-associated
proteins, composed of a GTP-binding -subunit and a
-heterodimer, transduce the signals of hormones and
neurotransmitters that bind to a large family of receptors with seven
transmembrane domains. Effector systems for G protein-coupled receptors
include adenylyl cyclases, phosphoinositide-specific phospholipases C
(PLC),1 or ion channels. Based on amino
acid sequence homology and sensitivity to bacterial toxins, the
GTP-binding
-subunits of G proteins are grouped into four
subfamilies, Gs, Gi, Gq, and
G12/13 (1, 2). Two subfamilies (Gs and
Gi) activate or inhibit adenylyl cyclases, respectively,
and are sensitive to ADP-ribosylation catalyzed by either cholera toxin
or pertussis toxin (PTX). Although many G protein subtypes do not show
any tissue specificities, the
-subunit of G16, a member
of the Gq subfamily, is expressed exclusively in
hematopoietic cells (3). Members of this subfamily, which lack an
ADP-ribosylation site, activate PLCs and thus mobilize Ca2+
from inositol 1,4,5-trisphosphate (InsP3)-sensitive
intracellular Ca2+ stores.
In some cases, the expression pattern of G proteins is subject to
changes during cellular differentiation (4-8). In the hematopoietic cell line HL-60, G16 is down-regulated during terminal
differentiation (3). Analysis of G
16 expression in
patients with acute lymphoid leukemia suggests that G
16
is expressed in progenitor B cell-derived malignant cells but not in
leukemic cells that are derived from more mature differentiation stages
(7). Interestingly, G
16 displays growth inhibitory
effects when overexpressed as a constitutively active
(GTPase-deficient) form in Swiss 3T3 fibroblasts or in small cell lung
carcinoma cells (9, 10). Thus it appears that G
16
expression is regulated during cellular differentiation and that its
expression levels may be critical for controlling proliferation rates
in slowly cycling stem and progenitor cells.
The signaling pathways that engage G16 in hematopoietic
cells are still poorly defined. Recent studies suggest that
G16 can interact indiscriminately with a multitude of G
protein-coupled receptors when receptors and G16 were
overexpressed in COS-7 cells (11, 12). On the other hand, some
receptors for chemotactic factors such as C5a,
formyl-methionyl-leucyl-phenylalanine, or interleukin-8 appear to
interact with G16 but not with Gq or
G11, to stimulate the production of InsP3
(13-16). Thus, some receptors may utilize G16,
preferentially, to mobilize intracellular Ca2+. However,
these studies could not address the question of which receptors may
couple specifically to G16 in native cells. Overexpression of wild type
-subunits in transfected cells potentially shifts the
balance between signal transduction through
- and
-subunits because of a redistribution of
-subunits bound to the
overexpressed versus the endogenous G proteins. In transient
transfection assays, competition between
-subunits for common
-dimers was indeed demonstrated (17). Moreover, altering the
stoichiometric relationships between G protein and effector systems may
result in spillover effects in transfected cells which do not occur in
wild type cells. Other studies employed cells overexpressing
GTPase-deficient mutants as tools to define the role of individual
Gq subfamily members (9, 10, 18). However, uncertainties
remain in the interpretation of results from such studies because these
mutants act independently of
-heterodimers. As a more gentle
alternative, an antisense-based approach with oligodeoxynucleotides or
expression of antisense RNA to reduce the expression of specific
-subunits has proven to be a powerful tool in defining the roles of
specific G protein isoforms of various subfamilies (19).
In the present study, we sought to identify signaling pathways that
obligatorily require G16 in a hematopoietic cell line, human erythroleukemia (HEL) cells. Full-length antisense RNA expression was used to suppress selectively the endogenous biosynthesis of G16. In G
16-suppressed cells, a specific
deficit of the P2U
(P2Y2)2 purinoceptor to
stimulate the release of intracellular Ca2+ was detected,
whereas other functions triggered by the same receptor remained
intact.
Tissue culture media and
media supplements were purchased from Life Technologies, Inc. DOTAP,
ATP, and UTP were from Boehringer Mannheim, G418 sulfate was from
Promega Corp. Fura-2/AM was from Molecular Probes. PGE1,
PGE2, and thrombin were obtained from Sigma. PAF was
purchased from Calbiochem-Novabiochem. UK14304 and thapsigargin were
from Research Biochemicals International (RBI). Antibodies against G
protein - and
-subunits were a generous gift from Dr. K. Spicher
and Dr. B. Nürnberg, both from the laboratory of Prof. G. Schultz, Dept. of Pharmacology, Free University of Berlin, Germany (see
below for details). All other reagents were from Fluka Chemie or from
Merck Darmstadt.
The cDNA for human
G16 including 5
- and 3
-flanking regions (224 and 60 base pairs, respectively) (3) was inserted into the mammalian
expression vector pcDNA3 (Invitrogen Corp.) in antisense
orientation with respect to the direction of transcription from the
cytomegalovirus promoter, thus allowing the expression of
G
16 antisense RNA in transfected cells. The resulting plasmid, pG16AS, also harbors a copy of the aminoglycoside
phosphotransferase gene, allowing selection of transfected cells with
G418 sulfate.
HEL cells (22) and K-562 cells were obtained from the American Type Culture Collection. The cells were maintained at densities between 5 × 104 and 5 × 105 cells/ml in a humidified atmosphere of 95% air and 5% CO2 in RPMI medium that was supplemented with 10% fetal bovine serum, 2 mM Glutamax I (Life Technologies, Inc.), 1 mM sodium pyruvate, 100 units/ml penicillin, and 100 µg/ml streptomycin.
Stable Transfection of HEL CellsHEL cells (2.5 × 107 cells at a density of 2 × 106 cells/ml) were transfected in Opti-MEM (Life Technologies, Inc.) with 50 µg of the plasmid pG16AS, using DOTAP as the transfection reagent according to the manufacturer's instructions. Sixteen h later, the medium was replaced with RPMI medium plus supplements, and the cells were seeded at a density of 106/ml. Twenty-four h later the cells received fresh medium that included 400 µg/ml medium of active G418 sulfate. The following day, the cells (8.6 × 107) were transferred into 96-well dishes (105 cells/well) to select independent G418 sulfate-resistant clones. Fourteen days after transfection, 47 of 864 wells contained viable, G418 sulfate-resistant clones, which were picked and expanded. Six clonal cell lines out of 21 that showed sustained growth were selected randomly and characterized further.
Measurement of Intracellular Free CalciumHEL cells were loaded with fura-2/AM (9 µM) as described previously (6). The loaded cells were washed, and a stock suspension was prepared which contained 5 × 106 cells/ml medium A (phosphate-buffered saline supplemented with 1 mM MgCl2, 5 mM KCl, 22 mM glucose, and 5% fetal bovine serum). The Ca2+ concentration of medium A was around 90 µM as measured by atomic absorption spectrophotometry. Prior to use, cells from the stock suspension were diluted to a density of 2.5 × 105/ml into medium A and preincubated for 5 min at 25 °C. Agonist-induced changes in [Ca2+]i were measured in a thermostatted (25 °C) cuvette in a Perkin Elmer LS-50B dual wavelength fluorescence spectrophotometer, which was equipped with the fast-fura measuring device (Perkin-Elmer). Emission intensities (at 485 nm) at excitation wavelengths of 340 and 380 nm were recorded. At the end of each experiment, maximum and minimum fluorescence ratios were determined by adding Ca2+ (1 mM) plus digitonin (12.5 µM) followed by Tris-buffered EGTA (20 mM, pH 8.5), respectively, to the incubation medium. Intensity ratios were converted to intracellular Ca2+ concentrations using the FL data manager software (version 3.5) and were based on an apparent Kd of 224 nM for the fura-2·Ca2+ complex (23).
Extraction of G ProteinCells were harvested and washed once in phosphate-buffered saline. Cells were then disrupted by three freeze/thaw cycles (liquid nitrogen) in buffer 1 (137 mM NaCl, 20 mM Hepes, pH 7.4, 2 mM MgCl2, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 10 µg/ml leupeptin, and 10 µg/ml aprotinin). Membranes were collected by centrifugation (14,000 × g for 2 min), and membrane-bound proteins were solubilized with Triton X-100 lysis buffer (137 mM NaCl, 20 mM Tris, pH 7.4, 1 mM MgCl2, 1% Triton X-100, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 10 µg/ml leupeptin, and 10 µg/ml aprotinin). Insoluble material was removed by centrifugation (14,000 × g for 2 min).
Immunoblotting and AntibodiesSolubilized membrane proteins
were separated by sodium dodecyl sulfate-polyacrylamide gel
electrophoresis (10% acrylamide) and electrophoretically transferred
to nitrocellulose filters (Bio-Rad). The -subunits of
G16, Gi3, Gi2, and
Gq/11 were detected by immunostaining with
affinity-purified antibodies (AS 339, AS 86, AS 269, and AS 368,
respectively) that were raised in rabbits against sequence-specific
peptides (24). G protein
-subunits in cellular extracts and in a
preparation of purified G
from porcine brain (a gift
from Dr. B. Nürnberg and Dr. G. Schultz) were detected with an
antibody (AS 11) that was raised against a peptide corresponding to a
conserved sequence within G
1, G
2, and
G
4 (25). Bound antibodies were visualized by enhanced
chemiluminescence (Amersham Corp.), using horseradish peroxidase-conjugated anti-rabbit IgG (Sigma) as the secondary antibody. In some instances, nitrocellulose filters were incubated twice for 90 min each with stripping buffer (100 mM
glycine, pH 2.2, 20 mM magnesium acetate, 50 mM
KCl) to remove antibodies. Filters were then blocked again and used for
reprobing with a different antibody.
Statistical significance was calculated by unpaired t tests using StatView for the Macintosh, version 4.0 (Abacus Concepts, Inc.). Half-maximum effective concentrations (EC50) were calculated by fitting the Hill equation to the data with the Levenberg-Marquardt fit algorithm of Kaleidagraph, version 3.0 (Abelbeck Software).
Individual G418 sulfate-resistant HEL cell lines that
were isolated following transfection with a plasmid expressing
G16 antisense RNA (pG16AS) were analyzed by protein
immunoblot to assess the extent of suppression of endogenous
G
16 expression. In three G418 sulfate-resistant cell
lines (1E3, 3D4, and 1G3), G
16 expression was greatly
diminished compared with parental HEL cells (Fig.
1A). In a fourth cell line (7H6) and two
others (data not shown), G
16 levels were not decreased,
although these cell lines were resistant to the antibiotic. Thus, the
cell line 7H6 was used further as a nonsuppressed control along with parental HEL cells. The specificity of the antibody (AS 339) for G
16 was assessed using wild type K-562 cells, which do
not express detectable levels of G
16 (3), and two
K-562-derived cell lines (K-562-D5 and K-562-A4) that were stably
transfected with a plasmid containing the full-length clone of
G
16 under the transcriptional control of the
cytomegalovirus promoter. A protein immunoblot with cell extracts from
K562-D5 and -A4, but not from parental cells, showed a unique major
immunoreactive band demonstrating that the antibody (AS 339) is highly
specific for G
16 (Fig. 1A). The
electrophoretic mobility of overexpressed G
16 was
identical to that of the major immunoreactive protein detected in HEL
cells.
To assess the specificity of G16 down-regulation in the
transfected cell lines, the expression levels of other G protein
-subunits were also examined in transfected and parental cells. A
major band was detected with the antibody raised against a common peptide of G
q and G
11, two
phylogenetically close congeners of G
16, which comigrate
on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The two
-subunits (collectively termed G
q/11) were expressed
at comparable levels in G
16-suppressed and nonsuppressed
cells as demonstrated in Fig. 1B. G
i3 and
G
i2 expression also remained unchanged. Since the
expression level of G protein
-subunits and G
may
be coordinately regulated (26), G
16-suppressed and
nonsuppressed HEL cells were analyzed for the level of
-subunit
protein, with an antibody (AS 11) that recognizes G
1,
G
2, and G
4 (25). However, no differences in G
protein expression were observed (Fig.
1C). In conclusion, the data indicate that
G
16 antisense RNA expression inhibits the expression of
endogenous G
16 in HEL cells, but the expression of other
G protein subunits is not affected.
Experimental data from transiently transfected COS cells
showed that heterologously expressed G16 could mediate
the mobilization of intracellular Ca2+, after stimulation
of diverse G protein-coupled receptors that were cotransfected into
these cells (11-13, 27). Being a member of the Gq
subfamily of G protein
-subunits, a role of G
16 in
mediating Ca2+ mobilization is expected. However, it is
unclear whether G
16 is indeed required in hematopoietic
cells to promote Ca2+ mobilization.
G
16-suppressed HEL cell lines thus provided a
possibility of identifying signaling pathways that depend strictly on
the presence of G
16.
A variety of ligands are known to induce a transient increase in
[Ca2+]i in HEL cells in a G
protein-dependent fashion (28). Fig.
2A summarizes the peak increments of
[Ca2+]i over resting levels after
adding different agonists at maximum effective concentrations. In
G16-suppressed and nonsuppressed cell lines,
PGE1 and thrombin gave rise to similar peak increases in
[Ca2+]i, indicating that
G
16 suppression did not inhibit signaling pathways
utilized by these two ligands. Changes in
[Ca2+]i upon stimulation with PAF
appeared to be more heterogeneous between the individual cell lines,
but down-regulation of G
16 certainly did not impair
PAF-dependent signaling.
Agonist-stimulated changes in
[Ca2+]i are transduced by pathways
with different sensitivity to PTX treatment, depending on the extent of
coupling of receptors to Gq type or Gi type G proteins. The downstream effectors, isoforms of PLC, are either activated by GTP-bound G released from Gq or
by G
released from Gi, whereby the
different isoforms of the enzyme show varying sensitivities toward the
G protein subunits (29). The sensitivity to PTX can help to identify
the subclass of G proteins and, consequently, the possible subtypes of
PLC-
which may be involved in intracellular Ca2+
mobilization. Schwaner et al. (28) reported that in HEL
cells, the transient increase of
[Ca2+]i induced by thrombin is
rather insensitive to PTX, whereas the effects of PGE1 and
PGE2 are partially or completely inhibited by the toxin,
respectively. In our experiments in parental HEL and
G
16-suppressed 3D4 cells, PTX inhibited thrombin-induced changes in [Ca2+]i by 23 and 14%,
respectively (Fig. 2B). These data indicate that the
thrombin receptor couples primarily, although not exclusively, to a
PTX-insensitive G protein, probably of the Gq type. Under
the same conditions, the toxin inhibited the effect of PGE1
in HEL and 3D4 cells by 93 and 86%, respectively, suggesting a major
role of Gi type G proteins in PGE1-induced
Ca2+ signaling. To exclude that PTX treatment simply
reduced the filling state of intracellular Ca2+ stores,
PTX-treated cells and nontreated controls were incubated with
thapsigargin, a potent and specific inhibitor of the endoplasmic reticulum Ca2+ pump which blocks reuptake of cytosolic
Ca2+ into the lumen of the endoplasmic reticulum (30).
Parental and G
16-suppressed cells responded with similar
transient, thapsigargin-induced increments in
[Ca2+]i, whether or not they were
treated with PTX (Fig. 2B). PGE2, which acts via
Gi-coupled receptors (28), was completely ineffective after
PTX treatment, demonstrating that PTX treatment was exhaustive (not
shown). The similar relative inhibitions of agonist-induced
Ca2+ changes that were observed after PTX treatment in
parental and G
16-suppressed cells make it unlikely that
the expression pattern of G
- and
G
-sensitive PLC-
isoforms differed considerably
between G
16-suppressed cells and parental HEL cells.
In HEL cells, the concentration of intracellular Ca2+ also
increased transiently when cells were incubated with UTP (100 µM) and reached a maximum between 40 and 60 s after
adding the agonist (Fig. 3A,
inset). The mean peak increase in
[Ca2+]i in parental HEL cells
amounted to 78.8 ± 4.8 nM (mean ± S.E.) (Fig.
3A). In contrast, UTP failed to increase
[Ca2+]i in any of the three
G16-suppressed cell lines (1E3, 3D4, 1G3). Mean changes
in [Ca2+]i were 3.4 ± 0.7 nM for 1G3 cells and even smaller in the other two cell
lines. Similarly, ATP was no longer able to increase
[Ca2+]i in
G
16-suppressed cell lines (data not shown). These
observations indicate a critical role of G
16 in UTP- and
ATP-induced Ca2+ changes. To substantiate further the
requirement of G
16 expression for
UTP-dependent Ca2+ signaling,
[Ca2+]i was also measured in the
G418 sulfate-resistant control cell line 7H6, which expressed wild type
levels of G
16. In this cell line, UTP-induced
Ca2+ changes (74.3 ± 2.2 nM) were not
significantly different from those evoked by the agonist in parental
HEL cells (Fig. 3A). These data strongly suggest that
inhibition of G
16 expression in transfected HEL cells is
accompanied by a severe defect in UTP-dependent Ca2+ signaling.
If UTP-dependent Ca2+ signaling in HEL cells
requires G16, then no effect of PTX on UTP-induced
Ca2+ changes would be expected. However, PTX inhibited the
UTP-induced increase in [Ca2+]i by
57.6 ± 4.9% (n = 6) compared with nontreated HEL cells (Fig. 3B). Inhibition by the toxin was exhaustive
because the effect of UK14304, a selective agonist for the
Gi-coupled
2-adrenoreceptor, was inhibited
completely under these conditions. Thus, a Gi type G
protein may also be involved in the propagation of
UTP-dependent Ca2+ signaling. However, the role
of Gi seems to be modulatory rather than obligatory because
its action absolutely depends on the costimulation of
G
16.
The UTP-dependent
increase in [Ca2+]i was mainly due
to the mobilization of intracellular Ca2+ as demonstrated
in Fig. 4, A-C. In the presence as well as
in the absence of free extracellular Ca2+, incubation of
HEL cells with UTP led to an increase in
[Ca2+]i (Fig. 4A,
left and right panel, respectively). This indicates that the UTP-dependent increase in
[Ca2+]i is mainly due to the
release of Ca2+ from intracellular stores. A slightly
larger increase in [Ca2+]i was
observed when Ca2+ was present in the extracellular medium
compared with Ca2+-free conditions (mean Ca2+
changes of 55.5 ± 0.6 nM versus 48.8 ± 2.0 nM (n = 4), respectively). The
small, but statistically significant difference
(p0 < 0.02) was consistently observed and may
be due to the contribution of a capacitative Ca2+ influx
that follows the mobilization of intracellular Ca2+ in HEL
cells (31). Additional experiments with thapsigargin confirmed these
results. Incubation of HEL cells with thapsigargin prior to stimulation
with UTP rendered the cells unresponsive to UTP (Fig. 4B).
Furthermore, incubation of HEL cells and G16-suppressed cell lines (3D4 and 1E3) with thapsigargin led to similar
transient increases in [Ca2+]i,
thus reflecting similar filling states of intracellular Ca2+ stores (Fig. 4C).
UTP may activate at least three different receptor subtypes of the
P2 purinoceptor class, which are stimulated by UTP, ATP, and ADP with different potencies (20, 32). The P2U
(P2Y2) subtype shows a potency order of UTP ATP > ADP (33), whereas the potency order for the "novel
P2" (P2Y6) subtype is UTP > ADP > ATP (34). At the uridine nucleotide (P2Y4) receptor, UTP is a full agonist, whereas ATP and ADP are only partial agonists (35, 36).
All of these receptor subtypes are capable of stimulating PLC-
and
hence are able to mobilize intracellular Ca2+ via an
InsP3-dependent pathway. Therefore, UTP-, ATP-,
and ADP-dependent changes in
[Ca2+]i of HEL cells were
monitored to determine the subtype of P2 purinoceptors
through which UTP acted in HEL cells. UTP and ATP were equally potent
in stimulating a rise in [Ca2+]i
with EC50 values of 3.6 ± 0.7 and 4.7 ± 1.6 µM, respectively (mean values of three independent
experiments ± S.E.). When cells were incubated sequentially with
the two agonists, either nucleotide showed cross-desensitization with
respect to its capacity to increase [Ca2+]i (Fig. 4D). This
suggests that both nucleotides act through a common nucleotide
receptor, most likely of the P2U purinoceptor subtype.
Consistently, the presence of receptors in HEL cells for UTP and ATP
has been shown in earlier studies (28). As shown above, UTP did not
depend on extracellular Ca2+ to evoke its effects,
suggesting that the receptor for UTP is not ionotropic (such as
P2 receptors of the X subtype). HEL cells also responded to
the P2Y (P2Y1) receptor agonist ADP, with a transient rise in [Ca2+]i, but the
maximum response to ADP was not affected by prior incubation with UTP
(Fig. 4E). These data rule out a contribution of this
receptor in the UTP-stimulated increase in [Ca2+]i. Taken together, these
results demonstrate that in HEL cells, UTP and ATP stimulate a
P2U purinoceptor, which leads to the mobilization of
Ca2+ from intracellular, thapsigargin-sensitive stores. The
inability of UTP (and ATP) to mobilize Ca2+ in
G
16-suppressed cell lines appears to be a consequence of
the interruption of P2U purinoceptor-dependent
cellular signaling.
In addition to activating an
effector system directly, receptor-mediated stimulation of
hematopoietic cells may also result in potentiating the action of a
second ligand that is administered simultaneously with, or subsequently
to, the primary agonist (31, 37, 38). Thus, the stimulation of HEL
cells with UTP (100 µM) prior to challenging with
PGE1 substantially enhanced the mobilization of
intracellular Ca2+ by the second agonist (Fig.
5A, left panel). This potentiating effect of UTP on PGE1-induced Ca2+ release was
not impaired in 3D4 cells (Fig. 5A, right panel), nor in any other of the G16-suppressed cell lines, even
though the mean absolute PGE1-dependent
increase in [Ca2+]i following
preincubation with UTP was higher in nonsuppressed cells than in
G
16-suppressed cells (Fig. 5B). Similar to UTP, ATP also enhanced the response to PGE1 in parental and
G
16-suppressed cell lines (data not shown).
The effect of PGE1 on
[Ca2+]i as well as the
potentiating effect of UTP (and ATP) could be caused either by an
influx of extracellular Ca2+ or by a Ca2+
release from internal stores or by both mechanisms. Therefore, PGE1-dependent Ca2+ signaling was
examined in the presence of different concentrations of extracellular
Ca2+ ([Ca2+]o). When
[Ca2+]o was lowered from 90 µM to < 107 M
(i.e. after chelating extracellular Ca2+ with
EGTA), the PGE1-dependent rise in
[Ca2+]i was reduced by 41.0 ± 7.7% (Fig. 5C). Similarly, when UTP-treated cells were
stimulated with PGE1, the subsequent rise in
[Ca2+]i was reduced by 39.2 ± 3.6% under Ca2+-free conditions. Irrespective of the
level of [Ca2+]o, UTP treatment of
HEL cells always potentiated the PGE1 response compared
with the respective controls (1.7 ± 0.1-fold versus
1.8 ± 0.1-fold, in Ca2+-containing versus
Ca2+-free medium, Fig. 5C). Comparable data were
also obtained with the G
16-suppressed cell line 3D4 (not
shown). We conclude that PGE1 primarily induces
Ca2+ release from internal stores, which subsequently
triggers a capacitative Ca2+ influx that depends on
[Ca2+]o. The relative
contributions of Ca2+ release and influx to the
PGE1-induced overall change in
[Ca2+]i were similar with or
without UTP treatment. UTP seems to enhance both the
PGE1-dependent cellular Ca2+
mobilization and the Ca2+ influx component.
The potentiating effect of UTP on intracellular Ca2+ release was not restricted to cells stimulated with PGE1, nor did it appear to be a general phenomenon elicited by UTP. Potentiation was also observed when UTP-treated cells were stimulated subsequently with adenosine or neuropeptide Y, rather than with PGE1. On the other hand, ADP- or PAF- induced rises in [Ca2+]i were not enhanced by pretreatment with UTP (Fig. 4E and data not shown, respectively). PTX inhibited the effect of UTP on the PGE1-induced rise in [Ca2+]i by 78.8 ± 8.1% (mean ± S.E., n = 3), suggesting a major, although not exclusive, role of Gi in potentiation.
The Potentiating Effect of UTP Is Also Mediated by the P2U PurinoceptorTo assess whether the receptor
subtype that triggered the UTP-dependent mobilization of
intracellular Ca2+ was identical to that provoking the
UTP-mediated potentiating effect on PGE1-induced
Ca2+ mobilization, the potencies of UTP to induce these two
effects were compared with those of ATP. ATP also potentiated
PGE1-induced Ca2+ mobilization (Fig.
6A). Exposing cells to maximum effective
concentrations of ATP and UTP together had no additive effect on
PGE1-induced Ca2+ mobilization, as would be
expected for a P2U purinoceptor-mediated effect (Fig.
6A). As stated above, ATP stimulated the mobilization of
intracellular Ca2+ with a mean EC50 value that
was similar to the one for UTP. If the receptors inducing the direct
and indirect (potentiating) effects were indeed identical, the
EC50 value for UTP for enhancing the PGE1
signal should also be similar to that for ATP. To test this hypothesis
experimentally, HEL cells were stimulated with different concentrations
of UTP or ATP to determine their potencies for direct Ca2+
mobilization. Subsequently, the same cells were challenged with PGE1, and increments of PGE1-induced changes in
[Ca2+]i over control values (from
cells that were not previously treated with nucleotides) were
determined to estimate EC50 values for this potentiating
effect. UTP mobilized Ca2+ with an EC50 value
of 2.4 ± 0.2 µM, whereas the EC50 value
for ATP was 4.0 ± 0.6 µM (Fig. 6, B and
C). The EC50 values for the potentiating effect
were lower by more than 1 order of magnitude for both nucleotides, but
their relative potencies (UTP versus ATP) again were similar
(EC50 = 0.10 ± 0.02 and 0.38 ± 0.07 µM, respectively) (Fig. 6, B and
C). Furthermore, UDP, an agonist with a potency similar to
that of UTP at the uridine nucleotide receptor (P2Y4),
neither stimulated Ca2+ mobilization nor enhanced the
PGE1-induced Ca2+ signal at concentrations that
would have been fully effective for UTP (not shown).
Taken together, these data strongly suggest that direct
Ca2+ mobilization by UTP and its indirect, potentiating
effect on the Ca2+ release by other agonists are both
mediated through P2U purinoceptors. We conclude that in HEL
cells, P2U purinoceptors are coupled to two different
signaling pathways. One pathway leads to the direct mobilization of
intracellular Ca2+ and is strictly dependent on the
presence of G16, the second pathway is independent of
G
16, mediates the potentiation of Ca2+
signals induced by some but not all agonists that raise
[Ca2+]i, and involves
Gi as a major signal transducer.
In the present study, we have used G16 antisense
RNA expression to create G
16-suppressed HEL cell
sublines, which allowed us to assess the role of this protein by
analyzing the resulting defects in receptor-effector coupling. We
conclude from our results that endogenous purinoceptors of the
P2U subtype interact specifically with the
G
16 protein to trigger intracellular Ca2+
release. It seemed obvious from the data in Fig. 1 that the antisense technique, although causing a major reduction, did not result in a
complete elimination of G
16 immunoreactivity. Since the
residual activity could not prevent the loss of a partial function of
the P2U purinoceptor, this function appears to require a
threshold level of G
16. Alternatively, the amount of
immunoreactive material may significantly exceed the amount of
functionally intact G protein. Similar discrepancies between residual G
protein levels after antisense RNA expression or antisense
oligodeoxynucleotide treatment and the associated functional defects
have also been noted by others (39-42).
A number of recent studies indicated that G16 has the
potential to couple to a broad spectrum of seven transmembrane domain
receptors (11-14, 27). In all cases, PLC-
isoforms seem to
represent the primary effector system. These receptor-G protein interactions were detected when different receptors together with the
-subunits of various G proteins were overexpressed in COS-7 cells.
However, whether all of these potential interactions indeed exist, and
more importantly, are functionally relevant in hematopoietic cells,
remains elusive. In HEL cells, the endogenous P2U
purinoceptor productively interacts with G16 but apparently
is unable to trigger Ca2+ release via Gq/11.
Hematopoietic cell lines, including HEL cells, contain high levels of
the latter G proteins, which share the same G protein family with
G
16 and are also linked to intracellular Ca2+ release (43-45). The selective functional defect
resulting from the reduction in G
16 expression clearly
indicates that Gq and G16 are functionally not
equivalent. In HEL cells, receptors for thrombin and PAF appear to
engage Gq/11 (Ref. 28 and this study). Unlike
P2U purinoceptor responses, the cellular response in HEL
cells to the activation of these latter receptors is not affected
significantly by the suppression of G
16 (present data).
Hence, in HEL cells G
16 plays a distinct role
specifically for the transduction of signals triggered by the
P2U purinoceptor. Interestingly, signal transduction
through the thrombin receptor was not affected by suppression of
G
16 in HEL cells, although thrombin receptors may couple
to G
16 in COS cells (11).
The inability of G16-suppressed cells to respond to UTP
by increasing [Ca2+]i could have
been the result of unintended clonal selection, which might have led to
cell lines that no longer expressed P2U purinoceptors. We
have not formally established the presence of P2U
purinoceptors on the various cell lines in binding studies because
there is still a lack of selective ligands with sufficiently high
affinity. The analysis using the potentiating effect of the P2U purinoceptor agonist UTP to stimulate the
Ca2+ response induced by PGE1 was used instead
to demonstrate the presence of P2U purinoceptors in
G
16-suppressed cell lines. Since UTP may also stimulate
the uridine nucleotide receptor and the novel P2 receptor, these
receptors have also to be considered for explaining the potentiating
effect. However, ATP is approximately 1,000-fold less potent than UTP
at the novel P2 receptor (34). At the uridine receptor, ATP is at best
a partial agonist (35, 36). In our study in HEL cells, ATP always
showed the same efficacy as UTP. Thus, the distinct properties of the
uridine and the novel P2 nucleotide receptors regarding UTP and ATP
rule out a contribution of these receptors under our experimental
conditions. As an alternate possibility, the lack of response to UTP
might have been caused by altered expression of G
- and
G
-stimulated isoforms of PLC-
, favoring signaling
by receptors that stimulate G
-sensitive isoforms such
as PLC-
2 or -
3. G
-mediated
stimulation of PLC requires the activation of Gq type G
proteins; G
-mediated activation of the enzyme is
initiated by the stimulation of Gi-coupled receptors and
the subsequent liberation of G
subunits from
Gi (29). A potential shift in the expression of PLC
isoforms in G
16-suppressed cells would therefore have
resulted in enhanced sensitivity of these cells toward PTX when
challenged with agonists that stimulate Gq- and
Gi-dependent pathways. However, the relative
inhibition of thrombin- or PGE1-stimulated changes in
[Ca2+]i in HEL and in
G
16-suppressed cells was similar after PTX treatment.
This indicates that the contributions of Gi-
versus Gq-dependent pathways to the
overall Ca2+ signal are similar in parental and transfected
cell lines when challenged with agonists that stimulate both
pathways.
The potentiating effects of UTP (or ATP) as outlined in Fig. 5 were also detected when PGE1 was added at the same time as the nucleotides and persisted for at least 15 min after the addition of the nucleotides (data not shown). Thus, potentiation does not appear to be subject to the same rapid desensitization process that was observed for Ca2+ mobilization (Fig. 4D). In rat basophilic leukemia RBL-2H3 cells, a similar potentiating ("priming") effect of thrombin on the chemoattractant-induced release of InsP3 was observed which also lasted several minutes (38). These cross-talk events between two different receptor-effector systems may thus require signaling components distal to receptors and their associated G proteins. Such a model would also predict that the dose-response relationships of agonists that stimulate direct and indirect effects through the same receptor may be different. The striking differences of EC50 values which were observed for the direct, G16-dependent Ca2+-mobilizing effect, and the indirect, Gi-mediated potentiating effect of UTP or ATP thus support the concept that two distinct signaling pathways originate at the P2U purinoceptor.
The observation that PTX inhibited P2U
purinoceptor-stimulated Ca2+ signaling only partially,
while suppression of G16 completely abrogated it,
suggests strongly that G
16 is obligatory for stimulation, whereas stimulation by G
represents a
positively modulatory function. An analogous mechanism has been
reported in the literature for types II and IV adenylyl cyclases
involving G
s as the obligatory and
Gi-derived G
as the modulatory G protein
subunits (46). In HEL cells, the molecular site for integrating
G
- and G
-mediated signaling pathways cannot be deduced from the present studies, but the
2
and
3 isoforms of PLC, which can be activated by both G
protein subunits, may represent possible candidates for such a role.
The priming effect of thrombin on chemoattractant signaling may indeed
critically involve PLC-
2 (38). However, the activation
of PLC-
2 by the two G protein subunits appears to be
independent rather than conditional, as has been concluded from
transient transfections or in vitro reconstitution studies
(29, 47-49). The molecular basis for the conditional,
G
16-dependent priming event observed in this study will need further characterization of the phospholipases and
perhaps other downstream components required in
UTP-dependent Ca2+ mobilization.
The potential of the endogenous P2U purinoceptor to
initiate two distinct signaling pathways that diverge at the level of the engaged G proteins is shared by many other receptors in various tissues (for a review, see Ref. 19). In native cells, multiple G
protein coupling has usually been assumed if PTX inhibited signaling to
some but not all effector systems, or if effector systems were only
partially inhibited by the toxin (28, 50, 51). Such results could best
be explained by alternative coupling of the receptor to PTX-sensitive
(Gi, Go) and PTX-insensitive
(Gq/11, G12/13) G protein subfamilies.
Bifunctional coupling of the P2U purinoceptor has also been
observed in the DDT1 MF-2 smooth muscle cell line (51). In this system,
P2U purinoceptors activate a PLC- via a PTX-resistant
pathway and inhibit adenylyl cyclase activity via a PTX-sensitive
pathway. Since non-hematopoietic cells lack G
16,
PTX-resistant effects of P2U purinoceptor stimulation must
be mediated by other members of the Gq subfamily. Moreover,
in epithelial cells, activation of P2U purinoceptors resulted in a PTX-sensitive release of InsP3, suggesting
interaction with a Gi type protein (52). In hematopoietic
cells, coupling of the P2U purinoceptor to G16
and to Gi was observed (this study). These results indicate
that P2U purinoceptors in native cells (as in the
cotransfection studies mentioned above) are capable of interacting with
G proteins from different subfamilies. The mechanisms determining which
of the G proteins are specifically recruited by individual receptors in
a given cell type are largely unknown. The work of Kleuss et
al. (53, 54) suggests that the expression of specific isoforms of
- and
-subunits of the trimeric G proteins together with a given
-subunit may confer cell-specific receptor-effector coupling.
It is well established that G protein-dependent signals may promote cell proliferation on their own or by modulating growth factor receptor responses (55, 56). Recent reports demonstrated that extracellular UTP and ATP are potent mitogens in rat renal mesangial and aortic smooth muscle cells (57-59). The nucleotides seem to act via stimulation of P2U purinoceptors. In addition, diadenosine tetraphosphate, which is released from neurons, chromaffin cells, and platelets, may also represent a physiological ligand at the P2U purinoceptor (60). The expression of P2U purinoceptors in a variety of tissues (33), including endothelial and hematopoietic cells, makes the P2U purinoceptor an interesting candidate for the hormonal (as opposed to growth factor-dependent) regulation of cell proliferation in the hematopoietic and vascular systems.
We thank Dr. M. I. Simon (California
Institute of Technology, Pasadena, CA) for providing the cDNA for
G16. We are grateful to Dr. K. Spicher and Dr. B. Nürnberg, both from the laboratory of Prof. G. Schultz, Dept. of
Pharmacology, Free University of Berlin, Germany, for antibodies and
for purified G
heterodimer, respectively.