(Received for publication, October 22, 1996, and in revised form, March 24, 1997)
From the Department of Pharmaceutical Sciences, Northeastern University, Boston, Massachusetts 02115
We have previously shown that
2-adrenergic receptor-mediated coupling to
phospholipase D (PLD) in vascular tissues requires a tyrosine kinase
activity (Jinsi, A., Paradise, J., and Deth, R. C. (1996) Eur.
J. Pharmacol. 302, 183-190). To further clarify this mode
of regulation we reconstituted
2A/D-adrenergic
receptor-stimulated PLD activity in PC12 cells expressing the cloned
receptor. [3H]Myristic acid-labeled cells were lysed by
nitrogen cavitation, and aliquots of subnuclear fraction were utilized
in the PLD assay. Agonist-stimulated PLD activity was measured in the
presence of 0.4% butanol as [3H]phosphatidylbutanol
formation. Both GTP and its non-hydrolyzable analog guanosine
5
-O-(thiotriphosphate) stimulated PLD activity in a
concentration- and time-dependent manner that required
co-activation of protein kinase C by phorbol dibutyrate. Addition of
epinephrine produced a 3-fold stimulation of PLD activity in the
presence of GTP and GDP. This agonist-stimulated PLD activity was
completely blocked by the
2-adrenergic receptor
antagonist rauwolscine and by Clostridium botulinum toxin
as well as by antibodies directed against either pp60src, RhoA,
or Ras GTPase-activating protein. These results indicate that coupling
of the
2A/D-adrenergic receptor to PLD is complexly regulated by both the tyrosine kinase pp60src and the low
molecular weight G protein RhoA.
In recent years, efforts to elucidate the biochemical and
molecular mechanisms of phospholipase D
(PLD)1 activation have focused interest on
both heterotrimeric and Ras-related low molecular weight G proteins
(LMWGs) that have been linked to receptor-mediated signal transduction.
Evidence for a functional role of G proteins in PLD activation comes
from studies in plasma membranes obtained from mammalian cells (1-4)
and in cell-free systems prepared from leukocytes (5) and neutrophils
(6). A GTPS-dependent PLD activity was observed in these
preparations only when fractions of both membranes and cytosol were
combined. It has been proposed that the cytosol-independent
(membrane-dependent) PLD activity represents only a small
portion of the total response and that full activation of PLD requires
the interplay of cytosolic factors such as PKC, a cytosolic GTP-binding
protein, and calcium (7).
A number of studies have provided evidence indicating a role for cytosolic LMWGs such as the Rho and Arf families in regulating membrane-associated PLD activity (8-11). Localization of LMWGs appears to be controlled by their interaction with regulatory proteins such as GTP/GDP dissociation stimulators, GDP dissociation inhibitors, and GTPase-activating proteins (GAPs) (12). Thus reconstitution of receptor-stimulated PLD activity would also be likely to require protein factors from both plasma membrane and cytosolic fractions.
Despite the lack of information about the molecular and cellular distinctions between various isoforms of PLD, the actions of its hydrolytic product (phosphatidic acid) have been documented in a broad spectrum of physiologic events and disease states including metabolic regulation, inflammation, secretion, cellular trafficking, diabetes, mitogenesis, oncogenesis, and senescence (13). The exact mechanism by which such diverse short and long term effects are mediated by the activation of PLD is not yet clearly understood. Recently Hammond et al. (14) identified a highly conserved human PLD gene family whose enzyme product is membrane-associated, stimulated by phosphatidylinositol 4,5-bisphosphate, activated by the monomeric G protein ADP-ribosylation factor-1, and inhibited by oleate.
The 2A/D-adrenergic receptor has been demonstrated to
transduce a cellular proliferation response mediated through
Gi proteins (15-17). It has become evident that this
mitogenic signal results from activation of the Ras/mitogen-activated
protein (MAP) kinase cascade (18, 19), which appears to involve
activation by G
subunits (20). The intermediate
tyrosine kinase pp60src (a member of the src family
of tyrosine kinases) has been proposed to phosphorylate Shc, an adapter
protein that associates with Grb2-Sos complexes through the SH2 domain
of Grb2, thereby leading to Ras activation (21).
Given these findings we sought to investigate the involvement of
Ras-related proteins in the 2A/D-adrenergic receptor
cellular signaling process. The
2A/D-adrenergic receptor
has been shown to couple to PLD in a tyrosine
kinase-dependent manner (22, 23) and is also involved in
mediating vasoconstrictor responses that are selectively blocked by
tyrosine kinase inhibitors (23, 24). The aim of this study was to
directly examine the possible involvement of a LMWG and a tyrosine
kinase in regulating
2-adrenergic receptor-mediated PLD
activity in PC12/
2D cells by reconstituting agonist-stimulated PLD activity in a crude cell lysate preparation.
PC12 cells stably transfected to overexpress
the cloned 2D-adrenergic receptor
(PC12/
2D cells, kindly donated by Dr. Stephen Lanier)
were grown in Dulbecco's modified Eagle's medium containing 10%
fetal calf serum, 5% horse serum, glutamine, and
penicillin/streptomycin/Fungizone and were maintained in 5%
CO2, 95% O2 at 37 °C.
PLD activity was measured using a
transphosphatidylation reaction as described by Halenda and Rehm (25).
PC12/2D cells were labeled overnight with 2-5 µCi/ml
[3H]myristic acid and washed in hypotonic lysis buffer
containing 25 mM HEPES (pH 7.4), 1.5 mM
MgCl2, 0.2 mM EDTA, 1 mM
dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 2 µM leupeptin, 2 µM pepstatin, and 100 nM phorbol dibutyrate (PDBu). Cells were allowed to swell for 10 min on ice and then were lysed by nitrogen cavitation at 650 p.s.i. for 20 min. The lysate was then centrifuged, and a postnuclear supernatant was obtained. Aliquots of the lysate
(approximately 100 µg of total protein/sample) were preincubated for
5 min at 37 °C in buffer containing 100 mM KCl, 3 mM NaCl, 5 mM MgCl2, 1 mM dithiothreitol, 25 mM HEPES (pH 7.4), 10 µM CaCl2, 1 mM ATP, and 1 mM isobutylmethylxanthine. Following this preincubation samples were incubated for the indicated time with the agonist in the
presence of antagonist, tyrosine kinase inhibitors, or other additions.
To exclude any contribution of residual catecholamines in the lysate,
agonist-induced stimulation was compared with a control level that
contained 100 nM rauwolscine. All antibodies were incubated
with the lysate at a concentration of 1:100 for 1 h at 4 °C
before carrying out the assay. In some studies antibodies were
previously incubated with their control peptides to validate their
specificity. In other experiments the lysate was pretreated with 10 µg/ml Clostridium botulinum toxin (C3 toxin) for 1 h
at 37 °C before carrying out the assay.
The transphosphatidylation reaction was terminated by adding ice-cold CH3OH/CHCl3 (2:1). This mixture was left on ice for 1 h or overnight in the freezer. Samples were homogenized in additional CH3OH/CHCl3 (2:1) containing 0.1% HCl and centrifuged at 2000 rpm for 20 min. The pellet was then further extracted with CH3OH/CHCl3/HCl, and the supernatants were combined. Phases were resolved by the addition of water and CHCl3. The lower CHCl3 layer containing phospholipids was separated and dried under N2. Samples were redissolved in CHCl3/CH3OH for spotting on Silica Gel G plates and separated using the following solvent system: benzene/chloroform/pyridine/formic acid (45:38:4:2). Commercial standards were separated in the same lane with the sample or in adjacent lanes and were visualized by iodine vapor staining. Corresponding [3H]phosphatidylbutanol (RF = 0.45) bands were scraped off the plate and counted. The pattern of migration was also confirmed by autoradiography in some samples.
Data AnalysisData are reported as the fraction of
[3H]phosphatidylbutanol formed of the total
lipid-associated radioactivity. Typical control values of
[3H]phosphatidylbutanol formed ranged from 200 to 500 cpm, and total lipid-associated radioactivity was between 75,000 and
100,000 cpm. Data means were analyzed by paired Student's t
tests. A probability of p 0.05 was selected as the
criterion for statistical significance.
[3H]Myristic acid was purchased
from DuPont. Tyrosine kinase inhibitors were kindly provided to us by
Dr. Alan Hudson of the Wellcome Foundation Ltd. (Bechenham, Kent,
U.K.), and antisera against the -subunit of Gi
1/2,
Gi
3, and Gi
s were kindly supplied by Dr.
Christopher Lynch. For phospholipid standards, 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphate and
1,2-dioleoyl-sn-phosphatidylbutanol were purchased from
Avanti Polar Lipids (Alabaster, AL). ADP-ribosyltransferase C3 toxin
was purchased from Wako Bioproducts (Richmond, VA), anti-RhoA and
anti-src were from Santa Cruz Biotechnology, Inc. (Santa
Cruz, CA), and anti-GAP was from Upstate Biotechnology Inc. (New York, NY). All other chemicals were of the highest reagent grade and were
obtained from Sigma.
G
protein-stimulated PLD activity was measured in a
PC12/2D cell lysate using the non-hydrolyzable analog
GTP
S. Addition of GTP
S at 100 µM for 20 min
produced about a 2-fold increase in
[3H]phosphatidylbutanol formation in lysate obtained
after pretreatment of cells with PDBu (100 nM) for 20 min
(Fig. 1), while control lysate failed to produce a
GTP
S-induced stimulation of PLD activity. These results parallel the
PKC activation requirement for agonist-stimulated PLD activity that we
have observed in intact PC12/
2D
cells.2 Additionally, G protein-stimulated
PLD activity was measurable only when the incubation buffer contained
calcium (1 µM; data not shown). Thus direct activation of
G proteins can activate PLD in a reconstituted system but requires the
co-activation of PKC and calcium.
Kinetics and Concentration Dependence of PLD Stimulation by Hydrolyzable and Non-hydrolyzable Guanine Nucleotides
GTP and
GTPS (but not GDP) produced a time- and
concentration-dependent increase in PLD activity measured in
the PC12/
2D cell lysate. Both GTP at 10 µM
and GTP
S at 100 µM each produced a
time-dependent increase in
[3H]phosphatidylbutanol formation (Fig.
2A). However, the increase of PLD activity
stimulated by GTP was significantly higher and faster than that
achieved by GTP
S. When epinephrine (1 µM) was combined
with GTP the rate of activation of PLD was significantly increased at
each time point, reaching 63, 43, and 154% above GTP-only levels at 5, 10, and 20 min.
Fig. 2B shows the concentration-dependent
stimulation of PLD by GTP, GTPS, or GTP combined with 1 µM epinephrine for 20 min. GTP
S produced no
significant increase in PLD activity at concentrations below 10 µM, but at concentrations above 10 µM it
produced an increase in PLD activity of 2-fold. In sharp contrast, GTP
produced about a 2-fold increase at concentrations as low as 10 nM with the concentration curve being more than 2 orders of
magnitude to the left of that obtained with GTP
S. It has been
reported that the two forms of guanine nucleotides differ in their
ability to activate heterotrimeric and monomeric G proteins (27, 28). The extent of PLD activation produced by the two guanine nucleotide analogs might therefore represent their net effect on different pools
of G proteins. Addition of epinephrine (1 µM) with GTP
further moved the curve to the left and increased PLD activity by about 2-fold above the GTP-only level. This implies that the agonist can
increase the efficacy of GTP in stimulating PLD activity.
It has been proposed by Gierschik
et al. (29) that higher concentrations of GDP can the
increase the role of the receptor in promoting activation of G
proteins. In earlier studies with PC12/2D cell membranes
we showed that GDP facilitates agonist response by progressively
suppressing basal [35S]GTP
S binding, thus
proportionately increasing epinephrine stimulation (30). Moreover, this
condition may be more representative of the physiologic environment. We
therefore utilized this strategy to further study agonist-induced PLD
activity, using GTP at 10 nM and GDP at 1 µM.
Under these conditions epinephrine stimulated PLD activity in a
concentration-dependent manner from 10 nM to 1 µM, while at concentrations above 1 µM
there was a trend toward a lower level of stimulated PLD activity (Fig.
3A). However, levels of PLD activity still
remained significantly higher than control at 10 and 100 µM.
To establish its 2D-adrenergic receptor origin,
epinephrine-stimulated PLD activation was measured in the presence of
the selective
2-receptor antagonist rauwolscine (Fig.
3B). GTP/GDP alone produced a 40% increase in PLD activity,
which was not affected by rauwolscine (1 µM) treatment.
Rauwolscine pretreatment not only completely eliminated the
epinephrine-induced increase but reduced PLD levels to 30% below the
unstimulated control group. This confirms that epinephrine-induced
activation of PLD in this lysate preparation occurs through stimulation
of
2-adrenergic receptors.
To
investigate the identity of heterotrimeric G proteins involved in
2D-adrenergic receptor-stimulated PLD activity, the cell
lysate was pretreated with antisera raised against the COOH-terminal region of G protein subtypes. As shown in Fig. 4,
incubation with control serum or anti-G
i3 did not
significantly reduce the epinephrine-stimulated PLD activity in a
PC12/
2D cell lysate; however, both
anti-G
i1/2 and anti-G
s reduced PLD
activity measured in the presence of epinephrine to a level not
different from GTP/GDP alone. Thus
2D-receptors may
activate PLD via coupling to either Gi1/2 and/or
Gs. Alternatively, the coupling to PLD may be indirectly
dependent on these G proteins.
Involvement of Low Molecular Weight G Proteins and Src in Epinephrine-stimulated PLD Activity
In recent years substantial
evidence has accumulated to suggest the involvement of LMWGs in
regulating PLD activity, which may underlie the requirement for
cytosolic factors that facilitate G protein-mediated activation of PLD
(8-11). To probe for the involvement of a Rho-like low molecular
weight G protein in epinephrine-stimulated PLD activity we utilized
both C3 toxin and specific polyclonal antibodies raised against RhoA.
C3 toxin is an exoenzyme produced by C. botulinum that
ADP-ribosylates the Rho family of G proteins at an asparagine residue
(Asn-41) located within or close to the putative effector domain of the
molecule and blocks its downstream signaling (31). Anti-RhoA was an
affinity-purified rabbit polyclonal antibody raised against a synthetic
peptide corresponding to the GTP-binding domain (residues 119-132) of
RhoA, which is the most abundant form of Rho. Incubation of the cell
lysate with C3 toxin (10 µg/ml) not only blocked the
GTP/GDP-stimulated activation of PLD but reduced activity to
significantly below the untreated control level (Fig.
5A). Epinephrine had no effect after C3 toxin pretreatment. Similarly, anti-RhoA pretreatment completely eliminated both the GTP/GDP- and epinephrine-induced increase. Pretreatment of
lysate with anti-RhoA that had been preabsorbed with its control peptide had no significant effect on PLD activity. This provides evidence for an important role of a Rho-like LMWG protein, possibly of
the RhoA subtype, in regulating both basal and receptor-induced PLD
activity.
Polyclonal antibodies against Src and Ras GAP were also used to identify their possible involvement in regulation of PLD. Anti-Src is an affinity-purified rabbit antibody raised against a peptide corresponding to residues 3-18 within the amino-terminal region of pp60src. Anti-Ras GAP is a polyclonal antibody raised against a TrpE fusion protein containing amino acid residues 171-448 corresponding to the SH2/SH3 region of p120 Ras GAP. Incubation of the cell lysate with anti-Src or anti-Ras GAP blocked the PLD activity stimulated by GTP/GDP and also completely blocked epinephrine-induced PLD activity (Fig. 5B). Pretreatment with anti-src preabsorbed with its control peptide had no effect on either basal or epinephrine-stimulated PLD activity. These results indicate a critical role for pp60src and Ras GAP or an associated protein in receptor-mediated PLD activation.
We have developed a method to reconstitute agonist-stimulated PLD
activity in a crude cell lysate preparation obtained from PC12/2D cells. Conditions required for activating the
receptor-mediated PLD pathway included the presence of calcium and ATP
as well as pretreatment with a PKC-activating phorbol ester. Each of
these requirements is consistent with prior studies of PLD activity in
cell homogenates (8-11) as well as our own studies in intact PC12/
2D cells.2 A requirement of
phosphatidylinositol 4,5-bisphosphate has also been reported (14), but
we did not supplement endogenous phosphatidylinositol 4,5-bisphosphate
in our studies. Thus stimulation of PLD by
2D-receptors in this broken cell preparation is characterized by many of the same
regulatory features found in intact cells but is amenable to study with
reagents such as antibodies that otherwise could not be used with
intact cells.
Epinephrine stimulation measured either in the presence of GTP or a
combination of GTP and GDP produced an increase in PLD activity (Figs.
2 and 3A), and its sensitivity to the antagonist rauwolscine
confirmed its 2-adrenergic receptor origin. As reported for Gi-mediated inhibition of adenylate cyclase in intact
PC12/
2D cells (32), PLD stimulation was diminished at
higher agonist concentrations (Fig. 3A), raising the
possibility that the receptor was increasingly involved with other
coupling pathways at higher levels of occupancy or suggesting the
development of a negative influence on the efficiency of PLD coupling.
Studies of adenylate cyclase coupling in intact PC12/
2D
cells found that epinephrine caused inhibition at levels up to 100 nM but stimulation at higher concentrations (32),
reflecting the now well recognized ability of
2-receptors to activate Gi and
Gs pathways with differing efficacy (33). Involvement of a
Gi protein in
2-receptor-mediated PLD
stimulation was also indicated in earlier studies with rat aorta (23)
and in intact PC12/
2D cells2 by the ability
of pertussis toxin pretreatment to block agonist stimulation. In the
current study, antiserum to Gi
1/2 blocked the
epinephrine stimulation of PLD while anti-Gi
3 did not (Fig. 4), suggesting that one or both of the former G proteins is
responsible for agonist-induced activation. The reduction observed with
anti-Gs antibodies is unexpected, but since
2-receptors can activate Gs in these cells,
it is possible that the
subunits common to both types of G
proteins may contribute to PLD activation.
Both Arf and Rho LMWGs have been shown to independently and
synergistically regulate PLD activity in an isoform-specific manner (9,
11, 34). The loss of epinephrine-mediated PLD stimulation upon
pretreatment with either C3 toxin or anti-RhoA antibodies (Fig.
5A) suggests that activation of PLD in the
PC12/2D cell lysate was under the control of one of the
Rho family proteins (presumably RhoA). These findings are in agreement
with similar observations implicating involvement of RhoA in the
activation of membrane-associated PLD from several tissues (10, 34,
35). Since C3 toxin, which has a broader specificity than anti-RhoA, significantly reduced basal PLD activity whereas anti-RhoA did not, it
is possible that two or more types of Rho proteins may regulate PLD
activity in PC12/
2D cells. While a non-RhoA LMWG might
regulate basal activity, RhoA is apparently essential for receptor-stimulated PLD activation.
Inhibition of agonist-induced PLD stimulation by anti-Ras GAP antibody treatment (Fig. 5B) provides evidence for the involvement of multiple LMWGs (i.e. Ras and Rho) in this pathway. p120 Ras GAP not only stimulates Ras GTPase activity but contains SH2 and SH3 domains that allow it to complex with other proteins in a signaling cascade, and thus it may function as a Ras-regulated effector. The antibody we used is directed toward the SH2/SH3 binding domain of Ras GAP, which indicates that these domains may be involved in signal transmission from Ras to Rho. For example, it has been reported that Ras GAP forms a complex with a Rho-specific p190 GAP (36), implying that Rho activity may be subject to Ras regulation as has been previously suggested (37, 38).
Stimulation of 2A-adrenergic receptors has been
demonstrated to increase the levels of active Ras, leading to increased
MAP kinase activity via a pertussis toxin-sensitive mechanism (18). This signaling pathway has now been shown to involve the pleckstrin homology domains of
subunits (possibly via their binding to phosphatidylinositol 4,5-bisphosphate), tyrosine phosphorylation of
Shc, and translocation of the Grb2-Sos complex to the plasma membrane
where Sos promotes guanine nucleotide exchange on Ras (39). By analogy,
Rho-dependent activation of PLD by
2D-receptors may be initiated by a
subunit
mechanism through Ras, which can simultaneously provide a mitogenic
stimulus via Raf and the MAP kinase pathway. This implies a bifurcation
of the signaling pathway at the level of Ras GAP or beyond, as
illustrated in Fig. 6.
Elimination of 2D-receptor-stimulated PLD activity
by anti-pp60src antibodies provides evidence for involvement of
this non-receptor tyrosine kinase in the signaling pathway. Jiang
et al. (40) earlier reported that overexpression of v-Src
leads to increased PLD activity, and the GTP dependence of this
increase suggested the involvement of a low molecular weight G protein.
Subsequently, it was shown that PLD activation by v-Src did indeed
involve Ras, although an additional role for Rho was not investigated
(41). Tyrosine phosphorylation of Shc is increased in v-Src-transformed cells, and recently it has been shown that pp60src is
responsible for the
2-adrenergic receptor-mediated
increase in Shc phosphorylation (21). It is also of interest that
wortmannin, an inhibitor of phosphatidylinositol 3-kinase, blocks
2-receptor-induced phosphorylation of Shc at
concentrations that also inhibit
2-adrenergic receptor-mediated vasoconstriction (42) and receptor-induced PLD
activation (43).
RhoA has recently been shown to reduce the activity of myosin
phosphatase (44), which would promote contraction of vascular smooth
muscle. The effects of 2-receptor-dependent
PLD activation in blood vessels (e.g. sustained PKC
activation) could therefore be complemented by inhibition of myosin
dephosphorylation as well as the phosphorylation of caldesmon by MAP
kinase (45) to provide the overall level of contraction. Studies of
genetically transmitted hypertension in rats have identified a
polymorphism within the pp60src locus that is associated with
elevated blood pressure (26). This suggests the further possibility
that enhanced signaling via the RhoA/PLD pathway could play a role in
the elevated vascular resistance that is characteristic of this model
of hypertension. Hyperactivity of the shared Ras/MAP kinase pathway
might also lead to a proliferative component of the hypertension.
In summary, we have for the first time demonstrated the coupling of
2-adrenergic receptors to PLD activation in a broken cell preparation. PLD coupling is GTP-dependent and
requires the participation of RhoA and pp60src as well as the
co-activation of PKC. Stimulation of this pathway plays a significant
role in
2-receptor responses such as the contraction of
vascular smooth muscle and may contribute to some forms of
hypertension.