(Received for publication, October 8, 1996, and in revised form, November 4, 1996)
From the Departments of Medicine and Biochemistry, Howard Hughes Medical Institute at Duke University Medical Center, Durham, North Carolina 27710
G protein-coupled receptor kinases phosphorylate the agonist occupied conformation of G protein-coupled receptors in the plasma membrane, leading to their desensitization. Receptor resensitization requires receptor dephosphorylation, a process which is mediated by a plasma and vesicular membrane-associated form of PP-2A. We present evidence that, like receptor phosphorylation, receptor dephosphorylation is tightly regulated, requiring a specific receptor conformation induced by vesicular acidification. In vitro, spontaneous dephosphorylation of phosphorylated receptors is observed only at acidic pH. Furthermore, in intact cells upon agonist stimulation, phosphorylated receptors traffic from the plasma membrane to vesicles where they become physically associated with the phosphatase and dephosphorylated. Treatment of cells with NH4Cl, which disrupts the acidic pH found in endosomal vesicles, blocks association of the receptors with the phosphatase and blocks receptor dephosphorylation. These findings suggest that a conformational change in the receptor induced by acidification of the endosomal vesicles is the key determinant regulating receptor dephosphorylation and resensitization.
As with other members of the G protein-coupled receptor
superfamily, the functional status of the
2AR1 is determined by its
phosphorylation state (1). One group of kinases which phosphorylate the
receptor in an agonist-dependent manner consists of members
of the family of G protein-coupled receptor kinases (GRKs) (2, 3).
Phosphorylation of the receptors and the subsequent binding of members
of a family of cytosolic proteins (arrestins or
-arrestins) serves
to uncouple the receptor from its cognate G protein (4, 5). This
results in a decreased responsiveness of the signaling system to
agonist, termed desensitization.
Upon removal of agonist, the attenuated responsiveness is reversed in a process known as resensitization, which involves dephosphorylation of the receptors (6). Much less is known about the resensitization and dephosphorylation processes. We have recently identified a membrane-associated phosphatase which dephosphorylates GRK phosphorylated G-protein-coupled receptors (7). This phosphatase, referred to as GRP (G protein-coupled receptor phosphatase), is a member of the PP-2A family of protein serine/threonine phosphatases. Surprisingly, under in vitro conditions, GRP-mediated dephosphorylation of phosphorylated receptors is entirely latent at neutral pH, with activity being observed only in the presence of activators of PP-2A such as protamine or freeze/thawing (7). This indicates that some form of activation must be necessary for receptor dephosphorylation to occur under in vivo conditions.
Over the same time frame as desensitization and resensitization occur,
the 2AR becomes sequestered into vesicles (1). Although
evidence indicates that sequestration is not required for
desensitization (8, 9, 10), it appears to be required for resensitization.
Agents which block sequestration, as well as sequestration-deficient
mutants of the
2AR, do not resensitize (6, 10, 11, 12).
These data, together with earlier findings which indicate that
receptors in a vesicular fraction are in a less phosphorylated state
than receptors in the plasma membrane (12), led to a model proposing
that dephosphorylation of the receptor occurs upon its sequestration
into a vesicle population (10). Consistent with these results, a recent
paper by Pippig et al. (6) shows that treatment of cells
with concanavalin A in addition to agonist prevents not only
sequestration and resensitization but also dephosphorylation of the
2AR. Since sequestration of the receptors into
internalized vesicles appears to be required for their
dephosphorylation, and since these vesicles were previously identified
as endosomes (13, 14), an acidified vesicle population (15), we tested
the hypothesis that the uniquely low pH of the vesicles is the key
regulator of GRP-mediated receptor dephosphorylation.
Materials
Okadaic acid was purchased from Calbiochem. The monoclonal antibodies 12CA5 and M2 were obtained from Berkeley Antibody Co. and Eastman Kodak Co., respectively. Antibody recognizing the catalytic subunit of PP-2A was obtained from Promega.
Methods
Phosphatase AssaysDephosphorylation of GRK2 phosphorylated
2AR (phosphorylated to a stoichiometry of 2-4 mol of
Pi/mol of receptor) (16), utilizing salt-washed bovine
brain membranes as the source of GRP (7), was performed by incubating
10 µl of phosphorylated
2AR (100 nM) with
10 µl of GRP and 10 µl of 50 mM acetic acid-acetate buffer. The proportion of acetic acid relative to acetate in the 50 mM buffer was used to adjust the pH of the overall assay
between 4.0 and 6.5. For assays performed at pH 7.0, 50 mM
Tris-HCl (pH 7.0) substituted for the acetic acid-acetate buffer. Where
indicated, 1 mg of protamine sulfate/ml or 5 µM okadaic
acid was added to the reaction mixture. Incubations were performed for
60 min at 30 °C and were terminated by the addition of SDS sample
buffer. The amount of labeled receptor was quantified by PhosphorImager analysis and the percentage of phosphate released was calculated relative to controls containing okadaic acid (7). Dephosphorylation of
phosphorylated casein (17) or phosphorylase a (18) was performed by incubating the respective substrate (at a concentration of
1 mg/ml) with an aliquot of GRP for 15 min at 30 °C, at the indicated pH (7).
HEK293 cells
transfected with DNA encoding FLAG-tagged 2AR were
incubated for 15 min either in medium alone, medium containing 10 µM isoproterenol, or medium containing both 20 mM NH4Cl and isoproterenol. Cells were placed
on ice, washed, and incubated with phosphate-buffered saline containing
250 µg of concanavalin A/ml to minimize vesicularization of the
plasma membrane (19) during Dounce homogenization (in 50 mM
Tris-HCl (pH 8.0) and 5 mM EDTA (TE) containing 5% sucrose
and protease inhibitors). Intact cells and nuclei were removed by
centrifugation at 300 × g for 5 min. Crude membranes
were obtained by centrifuging the resulting supernatant at 300,000 × g for 30 min. Pellets were resuspended in TE containing
5% sucrose, layered on top of a 5-50% continuous, nonlinear sucrose
gradient (19) and centrifuged at 100,000 × g for 100 min at 4 °C. Gradient fractions were collected and assayed for
125I-labeled cyanopindolol binding to determine the
localization of
2ARs. The sucrose density in each
fraction was determined by refractometry.
Fractions from the sucrose gradients containing receptor associated with either vesicles or plasma membrane were pooled, diluted with TE and centrifuged at 300,000 × g for 30 min at 4 °C. Pellets were resuspended in buffer B (10 mM Hepes (pH 7.4), 1 mM EDTA, 1 mM dithiothreitol, 100 mM NaCl, and 1% CHAPS) (20), and receptor complexes were immunoprecipitated with 15 µg of monoclonal antibody M2 in the presence of protein G-coupled agarose beads. Immunoprecipitates were washed in buffer B and resuspended in SDS sample buffer. Samples were subjected to SDS-PAGE, transferred to nitrocellulose, probed with antibody against the PP-2A catalytic subunit, and visualized by incubation with a horseradish peroxidase-conjugated secondary antibody followed by processing of the blot as recommended by the manufacturer (ECL by Amersham Corp.).
Whole Cell Dephosphorylation ofHEK293
cells were transfected with DNA encoding HA-tagged 2AR.
The intracellular ATP pool was 32P-labeled as described
(21). Cells were incubated with 10 µM isoproterenol
either alone or in the presence of 20 mM NH4Cl
for 10 min at 37 °C. Cells were washed with phosphate-buffered
saline and either harvested or incubated an additional 20 min with
medium alone or medium containing 20 mM NH4Cl
prior to harvesting. The
2AR was immunoprecipitated as
described (21), and equivalent amounts of protein were subjected to
SDS-PAGE. The amount of labeled receptor was quantified by
PhosphorImager analysis with the percentage remaining on the receptor
after the 20-min incubation calculated relative to the amount of label
present after the initial agonist stimulation.
The effect of acidic pH on the in vitro
dephosphorylation of GRK2 phosphorylated 2AR catalyzed
by GRP was examined (Fig. 1A). While no
GRP-mediated dephosphorylation of the receptor was observed at pH 7.0, significant dephosphorylation occurred as the pH was decreased below
7.0, with maximal dephosphorylation occurring at pH 4.85. This
pH-dependent dephosphorylation was completely inhibited by
okadaic acid, an inhibitor of PP-2A (Fig. 1A). In contrast,
when assayed in the presence of protamine, a PP-2A activator,
significant dephosphorylation of the receptor occurred at pH 7.0, while
less dephosphorylation occurred as the pH was lowered (Fig.
1A). These data indicate that acidification stimulates
GRP-mediated
2AR dephosphorylation in vitro.
Since under these conditions both the receptor and the phosphatase are exposed to the acidic pH, the in vitro assays do not reflect
the cellular situation. In the cell, the sites that are phosphorylated, and which therefore must be dephosphorylated on the
2AR,
lie on the cytoplasmic face of the plasma membrane (1). When
sequestration of the
2AR into endosomes occurs, these
phosphorylated sites presumably face the cytosol. Since it is the
intravesicular side that undergoes acidification, the sites of
phosphorylation as well as the catalytic site of the phosphatase would
not be directly exposed to the decreased pH. This suggests a pH induced
conformational change in the receptor which regulates sensitivity to
phosphatase actions.
To examine the role receptor conformation plays in facilitating the
pH-dependent dephosphorylation of the 2AR,
the effect of acidic pH on the ability of GRP to dephosphorylate native
phosphorylated
2AR versus heat-denatured
phosphorylated
2AR was examined (Fig. 1B).
The stimulatory effect of low pH on dephosphorylation was markedly
blunted for the denatured receptor. When the effect of pH on
dephosphorylation of nonreceptor substrates such as phosphorylase a and phosphorylated casein by the GRP was examined, more
dephosphorylation was observed at pH 7.0 than 4.85 (all substrates were
soluble over the pH range explored, Fig. 1B). These data
indicate that the receptor conformation plays a critical role in
determining the phosphorylation status of the receptor.
If vesicular acidification plays an important role in regulating
2AR conformation and GRP-mediated
2AR
dephosphorylation in a cellular system, then phosphorylated
2AR and GRP should be associated within endosomes
containing sequestered
2AR (13, 14). Continuous,
nonlinear sucrose gradients were utilized to separate
2AR localized in the plasma membrane from that
sequestered within vesicles (Fig. 2A).
Peak 1, localized near the top of the gradient, represented
receptors localized to vesicles since these fractions were composed of
clathrin coated vesicles, as determined by electron microscopy and
immunoblotting (data not shown). Peak 2 represented plasma
membrane-derived material, as determined by the presence of adenylate
cyclase activity (data not shown). Incubating the cells with a
2AR agonist (10 µM isoproterenol), conditions which promote GRK-mediated
2AR
phosphorylation (1, 10, 12), increases the quantity of
2AR present in the vesicular pool (Peak 1)
from 5-10% to about 30% of the total
2AR present (Fig. 2A,
versus
), consistent with the
occurrence of sequestration.
The association of plasma membrane or vesicular localized
2AR with phosphatase was subsequently examined (Fig.
2B). FLAG-tagged
2AR was immunoprecipitated
from Peak 1 and Peak 2, corresponding to
2AR localized to vesicles or plasma membrane,
respectively. Phosphatase associated with the immunoprecipitated
receptor was visualized by immunoblotting, utilizing antibody that
recognizes the catalytic subunit of PP-2A, a component of GRP (7). In the absence of isoproterenol treatment, no phosphatase was found associated with the
2AR either in the plasma membrane or
in the vesicles (Fig. 2B, left panel), despite its presence
in both fractions (Fig. 2B, right panel). The PP-2A present
in these fractions represented GRP, as defined by its ability to
dephosphorylate GRK2 phosphorylated
2AR (data not
shown). When the cells were incubated with isoproterenol to
phosphorylate transfected
2AR prior to harvesting,
phosphatase was found associated with the
2AR
specifically in the vesicles but not in the plasma membrane (Fig.
2B, left panel). Together, these data emphasize that the
vesicular environment is critical for the association of receptor and
phosphatase.
To determine specifically whether a decrease in pH, as occurs in
endosomal vesicles containing sequestered 2AR, is a
critical regulator of receptor dephosphorylation, we examined the
effects of incubation of HEK293 cells with NH4Cl on
receptor dephosphorylation and on the association of receptor with
phosphatase (Fig. 3). NH4Cl is a weak base
used to raise the pH of acidic cellular compartments such as endosomes
and lysosomes (22, 23). The effect of NH4Cl on
dephosphorylation of the
2AR is shown in Fig.
3A. In the absence of agonist, little phosphorylation of the
receptor was observed. Agonist stimulated a 3.5-fold increase in
2AR phosphorylation, which was unaffected by the
presence of NH4Cl (Fig. 3A, time 0). Dephosphorylation of 42% of the phosphorylated
2AR
occurred when agonist was removed, and the cells were incubated an
additional 20 min in medium alone, corresponding to the time during
which receptor resensitization occurs (10). In contrast, only 8% of the phosphorylated
2AR was dephosphorylated when
NH4Cl was included in this 20-min incubation (Fig.
3A, time 20). Agonist-induced sequestration of receptors was
not altered by the addition of NH4Cl (data not shown). This
excludes the possibility that an inhibition of receptor sequestration
was responsible for the decreased dephosphorylation observed in the
presence of NH4Cl. Thus NH4Cl, an agent which
inhibits vesicular acidification, significantly impairs
2AR dephosphorylation in HEK293 cells. In A431 cells, a
recent paper by Pippig et al. (6) showed that monensin, a carboxylic ionophore that has also been used to increase vesicular pH
(22, 23), inhibited resensitization but did not affect dephosphorylation of the
2AR in A431 cells. We used the
weak base NH4Cl rather than monensin to avoid the
possibility that previously reported effects of monensin, such as
effects on Golgi morphology and vesicular trafficking (24, 25, 26), might influence the dephosphorylation observed.
Fig. 3B (left panel) shows that NH4Cl
inhibits receptor dephosphorylation by inhibiting the interaction of
GRP with the receptor. While the catalytic subunit of PP-2A was found
associated with vesicular (sequestered) 2AR following
incubation of cells with agonist alone (Fig. 3B, left panel,
NH4Cl), almost no interaction of
GRP with
2AR was detected when sequestered receptor from
cells treated with agonist and NH4Cl was subjected to
immunoprecipitation (Fig. 3B, left panel,
+NH4Cl). Expression and localization
of the phosphatase appeared to be unaffected by NH4Cl
treatment (Fig. 3B, right panel). These studies indicate
that acidification is an important cellular mechanism for regulating
the association of phosphatase with phosphorylated receptor and thus
dephosphorylation.
These data suggest the following model for 2AR
resensitization (Fig. 4). Upon agonist binding, the
receptor becomes desensitized by GRK-mediated phosphorylation and the
subsequent binding of
-arrestin. Recent data indicate that
-arrestin, in addition to its role in desensitization, is also
involved in receptor sequestration (27). The receptor becomes
sequestered into endosomes (13, 14) (Fig. 2A) which become
acidified (15). Acidification facilitates receptor and GRP association
in the vesicles (Fig. 2B) and the subsequent
dephosphorylation of the
2AR (Fig. 3). Since the sites of phosphorylation on the receptor are not directly exposed to the
decrease in pH within the endosome, how might this acidification stimulate dephosphorylation? Residues that are exposed to this decrease
in pH include those that face the extracellular space when the
receptors are localized in the plasma membrane. Protonation of these
residues appears to produce a conformational change in the receptor
that allows the receptor to be dephosphorylated. Residues on this face
include those involved in agonist binding. It has been established that
ligands of certain transmembrane receptors are dissociated in endosomal
compartments upon acidification (23, 28). Perhaps a similar event
occurs for the
2AR, producing a conformational change
that causes the GRP to associate with the receptor and facilitate its
dephosphorylation. Indicative of just such a specific conformational
change within native
2AR is the markedly diminished
effect of low pH on dephosphorylation of the denatured receptor and its
inhibitory effect on GRP-mediated dephosphorylation of nonreceptor
substrates (Fig. 1B). A change in receptor conformation
facilitating GRP-mediated dephosphorylation of the receptor is
analogous to the conformational requirement for
agonist-dependent GRK-mediated phosphorylation of the
2AR. Following dephosphorylation, receptor is recycled
back to the plasma membrane where agonist binding can initiate a new
round of signal transduction (Fig. 4). This model may be applicable to
a larger population of the G protein-coupled receptor superfamily since
the Neurokinin1 receptor was recently shown to require
vesicular acidification for receptor resensitization (29, 30).
The identification of acidification as a novel mechanism regulating
2AR dephosphorylation explains the requirement of
sequestration for
2AR resensitization (6, 8, 9, 10).
Sequestration facilitates the localization of the
2AR to
an acidified compartment, where a conformationally altered receptor
associates with the GRP and is dephosphorylated. These findings
underscore the importance of receptor conformation in the regulation of
both receptor phosphorylation as well as dephosphorylation. Whereas
binding of agonist produces a receptor conformation that is readily
phosphorylated by GRKs, an acidic pH environment produces a receptor
conformation capable of being dephosphorylated by GRP.
We thank Dr. Neil Freedman for providing the
HA-tagged 2AR DNA, Carl Stone and Sturgis Payne for the
preparation of reconstituted
2AR, Dr. Robert Stoffel for
helpful discussions, and Grace Irons for excellent tissue culture
assistance and advice.