(Received for publication, February 24, 1997, and in revised form, March 27, 1997)
From the Department of Molecular Pharmacology and Biological Chemistry, Northwestern University Medical School, Chicago, Illinois 60611
Short term exposure of m2 muscarinic acetylcholine receptors (m2 mAChRs) to agonist causes a rapid phosphorylation of the activated receptors, followed by a profound loss in the ability of the m2 mAChR to activate its signaling pathways. We have used site-directed mutagenesis to identify two clusters of Ser/Thr residues in the third intracellular loop of the m2 mAChR that can serve as redundant targets for agonist-dependent phosphorylation. Mutation of both clusters of Ser/Thr residues to alanines abolished agonist-dependent phosphorylation, while wild-type levels of m2 mAChR phosphorylation were observed in mutant receptors with only one or the other cluster mutated. However, the functional effects of phosphorylation of these two "redundant" clusters were not equivalent. No receptor desensitization was observed in an m2 mAChR with residues Thr307-Ser311 mutated to alanine residues. In contrast, mutation of the other Ser/Thr cluster, residues Ser286-Ser290, to alanines produced a receptor that continued to desensitize. Internalization of the m2 mAChR was promoted by phosphorylation of either cluster, suggesting that distinct mechanisms with unique structural requirements act downstream of m2 mAChR phosphorylation to mediate receptor desensitization and receptor internalization.
The m2 subtype of muscarinic acetylcholine receptor (m2
mAChR)1 signals through heterotrimeric
G-proteins to inhibit adenylyl cyclase activity (1) and to activate
inwardly rectifying potassium channels (2), among other effectors.
These responses are rapidly attenuated in the continued presence of
agonist (2, 3), a process termed "receptor desensitization."
Desensitization may involve several distinct events. Phosphorylation of
the activated receptor is a critical event leading to receptor
desensitization. The kinase(s) that phosphorylate the agonist-occupied
m2 mAChR include one or more members of a family of G-protein-coupled
receptor kinases (GRKs) (3-5) and exhibit an exquisite specificity for activated G-protein-coupled receptors (GPRs) (6). Phosphorylation of
the m2 mAChR by GRKs enables the m2 mAChR to bind arrestin proteins
(6), which "arrest" or "desensitize" GPR signaling by
uncoupling the receptors from their G-proteins (7, 8). A second event
that may contribute to desensitization is the rapid and reversible
internalization of the m2 mAChR away from the plasma membrane during
the first minutes of agonist exposure (3, 4, 9). Internalization of the
m2 mAChR has been suggested to be facilitated by receptor
phosphorylation by GRKs (4), and studies with the -adrenergic
receptor (
-AR) have suggested a role of arrestin-receptor
interaction in internalization (10). However, internalization is not
sufficient for m2 mAChR desensitization, and the processes mediating
receptor/G-protein uncoupling and internalization can be dissociated
(3). Thus, it is possible that while receptor phosphorylation
contributes to both uncoupling and internalization of the m2 mAChR,
different mechanisms may be involved in the two regulatory events. We
have constructed a series of substitution mutants of the m2 mAChR to
identify the in vivo sites of agonist-dependent
phosphorylation and to investigate the contributions of the
phosphorylated sites to m2 mAChR desensitization and m2 mAChR
internalization. Previous studies have shown that phosphorylation of
the m2 mAChR occurs at a stoichiometry of 3-5 mol of phosphate/mol of
receptor in intact cells (3, 11, 12) and have identified a region in
the third intracellular loop (i3 loop) containing residues 282-323 as
a region required for agonist-dependent phosphorylation
(3). This region contains two separate clusters of five amino acids
with four Ser/Thr residues each; these Ser/Thr-rich clusters are both
homologous to the cluster of residues in the
2A-AR shown
to undergo agonist-dependent phosphorylation (13). The
amino acids 282-323 region also contains seven other Ser/Thr residues
not located in these two clusters. In this study we mutated the Ser/Thr
residues in the 282-323 region, including those in the two
clusters, to alanines and investigated the effects of the mutations on
agonist-dependent phosphorylation, desensitization, and
internalization of the m2 mAChR.
Cell culture reagents were purchased from Life Technologies, Inc. 32Pi and [3H]quinuclidinyl benzilate ([3H]QNB) were from Amersham Corp. 3H-Labeled N-methyl scopolamine ([3H]NMS) was purchased from DuPont NEN. Primers were synthesized by Life Technologies, Inc. or the Northwestern University Biotechnology Facility. Other reagents were from Sigma or previously identified sources (3). Allan Levey, Emory University, provided rat anti-m2 mAChR monoclonal antibody ascites (also purchased from Chemicon) (14).
Construction of the Mutant m2 mAChRsAll studies were
performed using the human m2 mAChR cDNA. The mutant STSVS AAAVA
(amino acids 286-290), which we termed the NAla4 mutant
(Fig. 1), was described previously (15) and was obtained from Wolfgang
Sadee (University of California, San Francisco). A
NheI/XmaI fragment was subcloned from the
pSG-m2ST/A (15) vector into the same site in pCR3-m2KT3 (3) to create mutant NAla4. Mutant TVSTS
AVAAA (amino acids
307-311), termed CAla4, was constructed from the wild type
human m2 mAChR using megaprimer PCR mutagenesis. The first-round
primers were primer KT3
(5
-CGGGATCCAGGCCTATTATGTTTCAGGTTCAGGAGGAGGTGTCCTTGTAGCGCCTATG-3
and the mutagenic primer Csense
(5
-CAGGATGAAACGCAGTTGCCGCTGCCCTGGGCC-3
) using pCR3-m2KT3 as a
template. The second round reaction used primer m2-385/407
(5
-GAATTCGTGTGACCTTTGGCTAGCCCTGG-3
) with the first round reaction
product and pCR3-m2KT3 as a template. Mutant N,CAla8,
containing both the NAla4 and CAla4 mutations,
was constructed by four-primer PCR mutagenesis. The first round
reactions used primers KT3 and Csense or primers
m2-385/407 and Cantisense (the exact compliment
of Csense) with pCR3-m2KT3 NAla4 as a template. The second round reaction combined the first round products with primers KT3 and m2-385/407. All first round
reactions were performed with Pfu polymerase (Stratagene);
all second round reactions used Taq polymerase
(Perkin-Elmer). The second round reaction products were cut with
NheI and XmaI and subcloned into pCR3-m2KT3. All four receptors (wild type (WT), NAla4, CAla4,
and N,CAla8) were epitope-tagged with a His6
tag on the C termini, replacing the modified KT3 tag (3). Vector
pCR3-m2KT3 was used as a template for a PCR reaction using
primers m2-385/407 and m2his6
(5
-AACTGCAGCTACCTGTGGTGGTGGTGGTGGTGTGTAGCGCCTATGTTC-3
). This
product was cut with NheI and PstI and
subcloned into this site in pCR3-m2KT3 to make pCR3-m2His6. A
XmaI-PstI fragment was excised from pCR3-m2His6
and ligated into the same site in each of the three mutant receptors to
create His6-tagged versions of each of these receptors.
Single and double point mutations of the remaining Ser/Thr residues
were constructed by the same four-primer PCR mutagenesis method. All
sequences were confirmed by dideoxy-DNA sequencing. All mutant
receptors were expressed at similar levels in transient or stable
transfection systems (see below) and were analyzed in radioligand
binding assays to compare their abilities to bind agonist and
antagonist. All mutants used here exhibited similar agonist and
antagonist binding properties as the WT m2 mAChR.
Cell Culture and Transfection
Cell culture and transfections in human embryonic kidney (HEK) 293 cells were performed as described (3). Transient transfections were performed using a clone of HEK293 cells stably expressing SV40 large T antigen, HEK-tsA201 (16), to increase expression levels. Transiently transfected cells expressing WT or mutant receptors yielded 1.5-3.0 pmol of m2 mAChR/mg of total cell protein. Stable cell lines were made by transfecting HEK293 cells with the pCR3-m2 mutant plasmids. G418-resistant clones were selected and analyzed for antagonist binding to determine receptor expression levels. The stable cell lines used in this study stably expressed the following amounts of m2 mAChR, as assessed by [3H]QNB binding assays: WT, 1.0 pmol/mg of protein; NAla4, 0.8 pmol/mg of protein; CAla4, 0.9 pmol/mg of protein; N,CAla8, 1.1 pmol/mg of protein.
In this study, we used the following strategies to study the heterologously expressed m2 mAChR. Transient transfections were used to initially characterize mutant m2 mAChRs for their ability to become phosphorylated in an agonist-dependent manner. Use of transiently expressed receptors allowed a rapid screening of mutant receptors in phosphorylation assays. However, for technical reasons, it is easier to study the m2 mAChR adenylyl cyclase coupling in stable cell lines (e.g. to measure coupling to adenylyl cyclase in the transiently transfected cells requires co-transfection of a stimulatory receptor). Thus, once interesting mutants were identified in the phosphorylation studies, we produced stable cell lines expressing these mutants to use for signaling studies. We had previously established that the stoichiometry for agonist-dependent phosphorylation occurred to 3-5 mol of phosphate/mol of receptor in both transient transfections and stable HEK clones and that receptor desensitization occurred equivalently in these two expression systems (3). Therefore, the validity of making comparisons between effects of agonist to induce receptor phosphorylation and uncoupling in studies of transiently and stably transfected cell lines has been shown experimentally (3). In contrast, m2 mAChR internalization occurred at different rates and to different extents in the two expression systems. Thus, both stable clonal lines and transient transfections were used to assay mutant m2 mAChR internalization in this study.
Intact Cell Phosphorylation and m2 mAChR ImmunoprecipitationTransiently transfected HEK-tsA201 cells were labeled for 4 h with 0.25 mCi/ml 32Pi in phosphate-free Dulbecco's modified Eagle's medium as described (3). The cells were treated with or without drugs for 15 min at room temperature and homogenized in 20 mM NaPO4, 20 mM NaF, 5 mM EDTA, 5 mM EGTA, pH 7.4, plus protease inhibitors (3), and the homogenate was centrifuged at 100,000 × g for 30 min at 4 °C. The resulting crude particulate fraction was then solubilized in homogenization buffer plus 0.4% digitonin and 0.08% cholate, and the solubilized proteins were separated from the insoluble fraction by centrifuging the mixture at 100,000 × g for 45 min at 4 °C. The solubilized m2 mAChRs were immunoprecipitated by an overnight incubation with anti-m2 mAChR antibody (14) and protein G-agarose beads. The immunoprecipitates were washed with solubilization buffer until no radioactive counts could be further washed from the pellets, and the proteins were eluted by incubating for 3 min at 95 °C in SDS-PAGE sample buffer (Laemmli) (17). The eluted proteins were separated on 8% acrylamide gels and transferred to nitrocellulose, and the receptors were analyzed by immunoblotting and phosphor imaging.
Immunoblot AnalysisNitrocellulose filters (0.2 µm) were blocked for 30 min in Blotto (5% powdered milk in Tris-buffered saline). For detection of the receptors, the anti-m2 mAChR antibody (14) was diluted 1:25,000 in Blotto and incubated with the filter for 2 h at room temperature. The filter was rinsed with TBS and incubated for 1-2 h at room temperature with sheep anti-rat IgG coupled to horseradish peroxidase, diluted 1:4000 in Blotto. Enhanced chemiluminescence (ECL) was used as the detection method for all immunoblotting experiments.
Analysis of Phosphorylated ReceptorsPhosphorylation of receptors was analyzed by SDS-polyacrylamide gel electrophoresis and phosphor imaging using a Fuji Bas2000 PhosphorImager. The amount of radioactivity in each receptor band was quantified using standards containing known amounts of radioactivity on each phosphor image. Because variable amounts of receptor protein were recovered in the immunoprecipitation, in each experiment the amount of receptor in each lane was quantified by immunoblotting the immunoprecipitates followed by analysis using a Bio-Rad scanning densitometer; by using known amounts of purified receptor, we constructed a standard curve in each analysis to convert the densitometric values to mol of protein, as described previously (3). The immunoblotting was performed under conditions where the staining of the receptor standards was in the linear range. Samples outside the linear range of exposure in the ECL assay, as assessed by multiple exposures of different times, were not included in the analyses. The results of the protein and radioactive analyses were combined to determine the amounts of radioactivity/unit of protein. This allowed a rigorous comparison of the extent of phosphorylation of the WT and mutant receptors. In addition, since the stoichiometry of agonist-dependent phosphorylation of the WT m2 mAChR was previously determined by us to be 3-5 mol of phosphate/mol of receptor in the HEK cell expression systems (3), this analysis allowed us to estimate the stoichiometries of phosphorylation of the mutants receptors.
Receptor-mediated Modulation of cAMP ProductionCyclic AMP
levels were determined using [3H]adenine-labeled whole
cells as described previously (3). Cells were incubated for 10 min with
no drug, 10 µM isoproterenol, or 10 µM
isoproterenol plus 0.1 µM to 1 mM carbachol,
all in the presence of 1 mM isobutyl 1-methylxanthine.
Cellular cAMP was recovered by the method of Salomon et al.
(18, 19). Isoproterenol treatment produced a 1.5-10-fold (average
7.4 ± 1.2-fold, n = 16) increase in cAMP levels
compared with basal levels. While the magnitude of the isoproterenol
response varied in different experiments as indicated, no significant
differences were measured within each experiment in the -AR response
or the forskolin (20 µM) response for cells expressing
the different WT or mutant receptors or in cells pretreated with
carbachol versus naive cells. No response to carbachol was observed in parental HEK293 cells expressing no exogenous m2 mAChR, indicating that this cell line does not express any endogenous muscarinic receptors that affect adenylyl cyclase signaling. Since the
[3H]adenine whole cell labeling method does not yield an
estimate of the pmol cAMP/mg protein, to compare results from
experiment to experiment and to take into account the experimental
variations in isoproterenol- or forskolin-stimulated levels of cAMP,
the amounts of cellular cAMP for each experiment were normalized by setting basal levels to 0% and isoproterenol- or forskolin-stimulated levels to 100%. Signaling through the m2 mAChR was expressed as the
percentage decrease in cAMP levels due to carbachol-induced attenuation
of the isoproterenol response. The desensitization of the adenylyl
cyclase response was analyzed using pairs of cells in which we compared
the decrease in isoproterenol-induced cAMP levels produced by carbachol
in naive cells versus cells pretreated with agonist to
induce desensitization. The data were analyzed with paired t
tests.
Agonist-induced decreases in the number of receptors located on the cell surface were measured using whole cell radioligand binding with the hydrophilic mAChR antagonist, [3H]NMS, which cannot cross cellular membranes. Transiently or stably transfected cells were incubated for 0-120 min in the presence or absence of 1 mM carbachol, washed five times with 3 ml of ice-cold phosphate-buffered saline, and subjected to radioligand binding with saturating concentrations of [3H]NMS for 2 h at 4 °C using methods previously described (3). Nonspecific binding was defined in the presence of 10 µM atropine. Protein assays were performed to assess differences in cell density on each treated plate so that radioligand binding data could be analyzed as mol of receptor/mg of protein. The amount of receptor present on the cell surface as a function of time of exposure to carbachol was expressed as a percentage of cell surface receptor present at zero time. For internalization studies, the percentage of mutant receptor remaining on the cell surface after various times of agonist exposure was compared with levels observed for the WT m2 mAChR and analyzed by a student's t test. No loss of total receptor number was observed during the 2-h duration of agonist treatment as assessed by use of the hydrophobic radioligand [3H]QNB.
A deletion of amino acids 282-323 in the central
portion of the i3 loop of the human m2 mAChR was previously shown to
completely abolish agonist-dependent phosphorylation of the
m2 mAChR (3). This domain contains two Ser/Thr-rich clusters that are
potential phosphorylation sites, since they are in acidic domains and
analogous to the phosphorylation domain of the 2A-AR
(14). We mutated each of these Ser/Thr clusters to alanine residues to
test the hypothesis that these Ser/Thr residues are phosphorylated and contribute to the desensitization of the m2 mAChR. The Ser/Thr residues
in the cluster near the N terminus of this region (N cluster, Fig. 1) were mutated to alanine in mutant
m2-NAla4 (S286A, T287A, S288A, S290A), and Ser/Thr residues
in the cluster near the C terminus of this region (C
cluster, Fig. 1) were mutated to alanine in mutant
m2-CAla4 (T307A, S309A, T310A, S311A). These mutant
receptors were expressed in HEK cells, and their properties were
analyzed in ligand binding assays. No significant differences were
detected in their ability to bind the antagonist NMS or the agonist
carbachol compared with the WT m2 mAChR (Table I). The levels of expression of the WT and mutant receptors in the HEK-tsA201 cells were also similar (see "Experimental Procedures").
|
The extent of agonist-dependent phosphorylation of mutant m2 mAChRs was compared with that of the WT m2 mAChR in transiently transfected HEK tsA201 cells. Cells expressing the different forms of the m2 mAChR were labeled with 32Pi, treated with a saturating concentration (1 mM) of the agonist carbachol or vehicle for 15 min, and the receptors were isolated by immunoprecipitation. In previous studies, we demonstrated agonist-dependent receptor phosphorylation in transiently transfected HEK-tsA201 cells and in stably transfected m2 mAChR-expressing HEK cell lines; in both the transient and stable expression systems the stoichiometry of agonist-dependent phosphorylation was measured as 3-5 mol of phosphate/mol of receptor (3). For the present experiments, stoichiometries were not determined; rather, we used the WT m2 mAChR as a standard to which we compared the levels of phosphorylation for the mutant m2 mAChRs. The transient expression system was used to quickly screen for mutations in the m2 mAChR that produced changes in agonist-dependent receptor phosphorylation. For each experiment, the amount of 32Pi incorporated into the various receptors (assessed by phosphor imaging) was expressed as a function of the amount of receptor present in each lane (assessed by densitometric analysis of the immunoblot).
Carbachol treatment of WT m2 mAChR, the m2-NAla4, and the
m2-CAla4 (Fig. 2A) resulted in an
agonist-dependent phosphorylation of each receptor. The
extents of phosphorylation of the m2-NAla4 and the
m2-CAla4 were 97 ± 7% and 107 ± 4%,
respectively, of the levels of agonist-dependent
phosphorylation measured for the WT m2 mAChR (Fig. 2C).
Thus, while precise measurements of stoichiometry were not attempted,
the results suggested that the WT and both mutant receptors were
phosphorylated to similar extents. These results, taken with the
previous results that localized the sites of phosphorylation to the
domain containing amino acids 282-323 (3), suggested that either 1)
the sites of agonist-dependent phosphorylation were located
in the remaining seven Ser/Thr residues in this domain (Fig. 1,
gray residues), or 2) the two clusters of Ser/Thr residues,
which are highly homologous and almost mirror images, were redundant
sites for agonist-dependent phosphorylation. Pairs of the
remaining Ser/Thr residues in this region were mutated to alanines to
test the first possibility. Mutants S282A/S283A, S294A/S315A, and
S320A/T323A all exhibited agonist-dependent phosphorylation like the m2 mAChR (data not shown). To test the second possibility we
constructed a mutant receptor with both Ser/Thr clusters mutated to
alanines (N,CAla8). In marked contrast to all other
mutants, the double cluster mutant did not exhibit
agonist-dependent phosphorylation (Fig. 2, A and
C). Basal phosphorylation of this mutant also was decreased.
These results indicated that either one of the two clusters of Ser/Thr
residues may be phosphorylated in any given receptor, since the extent
of receptor phosphorylation was not reduced when only one Ser/Thr
cluster was present. However, it is not possible to rule out
phosphorylation of some Ser/Thr residues on each cluster in the WT m2
mAChR when both Ser/Thr clusters are intact. These results raised the
questions of which cluster, if either, is the preferred site
phosphorylated in vivo and whether these two potential
phosphorylation sites were equivalent or different in the functional
consequences of their phosphorylation.
Analysis of Homologous Desensitization of the m2 mAChR Mutants
The ability of the m2 mAChR to attenuate adenylyl cyclase
activity is substantially desensitized following acute exposure of the
receptor to agonist (3). As shown previously, this desensitization required full agonist-dependent phosphorylation of the m2
mAChR; inhibition of receptor phosphorylation by 50% due to expression of a dominant-negative allele of GRK2, or full inhibition of receptor phosphorylation by deletion of residues 282-323 prevented all m2 mAChR
desensitization (3). Because m2 mAChR signaling is technically
difficult to measure in transient expression systems (because it
requires co-transfection and expression of a stimulatory receptor in
the same cells expressing the m2 mAChR) and since we previously
demonstrated that receptor phosphorylation and desensitization occurs
to the same extent in transiently transfected and stably transfected
HEK cells (3), we created HEK293 cell lines stably expressing the WT,
NAla4, CAla4, or N,CAla8 m2 mAChR
mutants and tested whether carbachol treatment caused these receptors
to desensitize in their ability to attenuate the stimulation of
adenylyl cyclase activity by endogenous HEK cell -ARs. Cell lines
expressed 0.8-1.1 pmol of m2 mAChR/mg of protein (see "Experimental
Procedures"). Receptor desensitization was assessed by comparing the
inhibition of isoproterenol-stimulated cAMP production observed before
and after pretreatment of the cells with the agonist carbachol. In naive (untreated) cells, treatment of the WT receptor with 1 mM carbachol reduced cAMP production by 40-45%, while
activation of all three mutant receptors reduced the adenylyl cyclase
response by 50-60% (Fig. 3, open bars).
Pretreatment of wild-type m2 mAChR with 1 mM carbachol for
20 min resulted in desensitization, as seen by the reduction of the
carbachol-mediated attenuation of cyclase activity in this system (Fig.
3A, gray bars). When the CAla4 mutant
was pretreated with agonist, it did not desensitize, even though it was
fully phosphorylated (Fig. 3B, gray bars). In
contrast, agonist pretreatment of the NAla4 mutant did
cause desensitization similar to that observed for the WT m2 mAChR
(Fig. 3C, gray bars). Finally, the double mutant
m2-N,CAla8, like the CAla4 mutant, did not
desensitize (Fig. 3D).
To further investigate the desensitization of the m2-NAla4
receptor, the desensitization observed in the WT and NAla4
m2 mAChRs was measured using different concentrations of carbachol
during the cAMP assay. Cells were pretreated with buffer alone or with 1 mM carbachol as described in the previous experiment,
washed, and then treated with isoproterenol alone or with different
concentrations of carbachol for 10 min. A similar extent of
desensitization of both the WT m2 mAChR and the m2-NAla4
receptor was also observed at these lower doses of carbachol (Fig.
4, A and B). The signaling and
desensitization experiments were repeated using a shorter time for the
carbachol pretreatment. For both the WT and the NAla4 m2
mAChR, significant receptor/G-protein uncoupling was observed following
only 5 min of carbachol pretreatment (Fig. 5). These results are in agreement with the rapid desensitization of m2 mAChR
responses in cardiac myocytes (2), where loss of m2 mAChR signaling was
observed within a few minutes of agonist exposure.
Thus, it is clear that the m2-NAla4 mutant desensitizes similarly to the WT receptor and is markedly different from the m2-CAla4 and m2-N,CAla8 mutants. These results demonstrate a distinct functional role for phosphorylation of the Ser/Thr cluster in residues 307-311 (mutated in the CAla4 receptor) in the process of receptor/G-protein uncoupling, while a role for phosphorylation of residues 286-290 remains unclear. These results lead to several conclusions. First, desensitization of the m2 mAChR required phosphorylation of Ser/Thr residues 307-311. Mutation of these residues in either m2-CAla4 or m2-N,CAla8 eliminated receptor desensitization, strongly suggesting that the wild-type receptor, which desensitized fully, is phosphorylated on residues 307-311. However, the m2-NAla4 mutant, which lacks the Ser/Thr residues in cluster 282-290 but is fully phosphorylated (presumably on residues 307-311), exhibited desensitization. Thus, the two redundant sites for m2 mAChR phosphorylation are not functionally redundant. This further suggests that other local structural elements must also be involved in mediating desensitization events downstream of receptor phosphorylation.
Internalization of the Mutant m2 mAChRsWe previously
observed that deletion of amino acids 282-323 decreased the rate and
extent of m2 mAChR internalization in stable cell lines (3). Subsequent
experiments showed that this deletion had even more pronounced effects
in transient expression systems (data not shown). Here we analyzed the
behavior of the WT and cluster mutants in both stably and transiently
transfected cells. Since the transient expression system allowed a
greater range of responses for the various mutants, we first used cells
transiently transfected with the WT m2 mAChR or one of the cluster
mutant receptors to investigate the contribution of the Ser/Thr
clusters to the internalization process. Analysis of these cells with
the hydrophilic antagonist, [3H]NMS, indicated that these
cells expressed 0.7-1.5 pmol of receptor/mg of total cell protein on
the cell surface. This corresponded to half of the total receptor
number in the cells, as determined by analysis with the hydrophobic
antagonist [3H]QNB, which labels both surface and
internal receptors (data not shown). Cells were treated with 1 mM carbachol for 0-120 min, and subsequent changes in the
levels of cell surface receptors were measured with
[3H]NMS. Approximately 30% of the wild-type m2 mAChR
internalized with a half-time of ~30 min (Fig.
6A). Interestingly, the
desensitization-defective CAla4 mAChR exhibited a pattern
of receptor internalization indistinguishable from the WT m2 mAChR
(Fig. 6A), demonstrating that desensitization and
internalization are differentially affected by mutations in this
domain. The NAla4 receptor internalized at a faster rate and to a slightly greater extent than did the wild-type m2 mAChR. In
agreement with earlier data from a m2 mAChR lacking amino acids 282-323 in the i3 loop (3), the N,CAla8 mutant did not
internalize at all in the transient expression system. Since this
mutant failed to undergo agonist dependent phosphorylation, this
suggests that phosphorylation of the receptor may play a role in
internalization as well as in m2 mAChR desensitization.
When the stable HEK lines were used to measure internalization of the m2 mutants, all forms of the m2 mAChR internalized more rapidly and to a greater extent. Nevertheless, the relationships between receptor internalization patterns that were observed in the transient expression system were mostly conserved. The NAla4 and CAla4 mutants internalized to a similar extent and rate to the WT m2 mAChR (Fig. 6B). In these stable lines, the double cluster mutant N,CAla8 internalized, but at a much slower rate and to a markedly reduced extent compared with the WT and single cluster mutants. The internalization of the N,CAla8 mutant was indistinguishable from the internalization previously described for a stably expressed deletion mutant of the m2 mAChR amino acids 282-323 of the i3 loop (3). Taken together, the internalization experiments suggest that phosphorylation of the m2 mAChR promotes receptor internalization. However, in contrast to desensitization, it appears that phosphorylation at either the N or C cluster of Ser/Thr residues can fully support internalization, suggesting that distinct mechanisms are operative in the processes of desensitization and internalization.
A comparison of the results from the transient and stable expression systems suggests that m2 mAChR internalization proceeds by a pathway that can be easily overwhelmed by the number of receptors expressed heterologously in HEK cells. Although the stably and transiently transfected cells have similar numbers of ligand binding sites, the actual receptor number in the transiently transfected cell is higher, since not all cells are transfected. The lower levels of receptor per cell in the stable cell lines may have allowed for faster and more extensive internalization of the WT and single cluster mutants as well as observation of low levels of internalization of the N,CAla8 mutant. It is likely that the differences observed in stable versus transient expression systems reflect a limiting amount of some protein that is involved in recognizing the activated m2 mAChR or linking the activated m2 mAChR with the internalization vesicles. Because the stoichiometry of phosphorylation is identical in stable and transient expression systems (3), it is unlikely that the limiting factor is a GRK, although it has been reported that overexpression of GRKs can enhance the rate of receptor internalization, albeit only at low concentrations of agonist (4). It is not evident why the NAla4 receptor internalizes more than the WT or CAla4 receptors in the transient expression system but not in stably transfected cells. Perhaps this mutant interacts better than the other receptors with the limiting factor in the transient expression system, but when this factor is no longer limiting, as in the stable clones, internalization of the three receptor forms is equivalent. The present data also suggest that while receptor phosphorylation promotes m2 mAChR internalization, it is not absolutely necessary for this process to occur, since the N,CAla8 mutant does internalize in the stable cell lines but does not undergo agonist-dependent phosphorylation. Furthermore, the results point out additional differences between receptor internalization and desensitization in that the N,CAla8 mutant does not desensitize (Fig. 3D) but does slowly internalize (Fig. 6B).
The present study has provided new insights into the molecular events associated with the regulation of the m2 mAChR. A review of all the data, which are summarized in Table II, suggests several interpretations of these results. Overall, the clearest message is that desensitization is tightly linked to receptor phosphorylation at a specific domain. In addition, the data strongly suggest that desensitization and internalization are two structurally and functionally separate processes. While mutation of the C cluster caused a complete loss of desensitization, it had no effect on internalization. Desensitization required phosphorylation of the cluster 307-311; these results provided a convincing role for the process of receptor phosphorylation in receptor/G-protein uncoupling (Table II). Phosphorylation of the N cluster in the wild type receptor may also occur, as the measures of phosphorylation stoichiometry are not exact (3-5 mol of phosphate/mol of receptor is a broad range of values), but it is clear that phosphorylation of the N cluster of Ser/Thr residues is not sufficient to support m2 mAChR desensitization. The results with internalization are more complex. Perhaps the simplest interpretation is that phosphorylation is an important signal for internalization, but the downstream structural requirements are more relaxed than for desensitization. While mutation of the C cluster ablated desensitization but did not alter internalization, in the C mutant full phosphorylation occurred at the N cluster (Fig. 2, Table II), and this "alternative" phosphorylation may be an acceptable signal for internalization although it cannot support desensitization. When both sites of phosphorylation are ablated by mutating both clusters, both desensitization and internalization are severely impaired. This interpretation supports a key role for the phosphorylation of the C cluster for both regulatory processes.
|
The timing of these two regulatory processes also supports the hypothesis that receptor/G-protein uncoupling is functionally distinct from receptor internalization. Receptor desensitization appeared maximal after 5 min of carbachol exposure (Fig. 5), while receptor internalization took somewhat longer to reach its maximum. While it could be argued that the internalization of m2 mAChR following 5 min of carbachol exposure was sufficient to desensitize the receptor, the data from the CAla4-m2 mAChR, which internalizes normally but does not desensitize, do not support the premise that receptor internalization is the primary cause for receptor/G-protein uncoupling. Indeed, these data suggest that in this expression system, loss of 75% of the surface receptor has no effect on m2 mAChR signaling. Rather, it is events initiated by the phosphorylation of the Ser/Thr cluster 307-311 that causes receptor desensitization.
It is interesting that the two Ser/Thr-rich N and C clusters are very
homologous and almost mirror images of each other. Homologous clusters
of serines and/or threonines are also found on other mAChRs (20, 21),
as well as on other GPRs, including the 2A-AR (13), the
angiotensin II-1 receptor (22), the 5HT-1B receptor (23), the adenosine
A1 receptor (24), and several bombesin receptor subtypes (25-27).
Clusters of Ser/Thr residues have already been implicated in regulation
of m1 and m3 mAChR internalization (15), although phosphorylation of
these residues has not yet been investigated. Mutation of analogous
clusters of Ser/Thr residues in other GPRs may allow dissection of the
events that regulate GPR signaling at the level of the receptor from
events which act downstream of receptor activation to prolong or
desensitize second messenger pathways. There is a striking difference
in primary sequence of the potential or proven phosphorylation sites of
receptors with Ser/Thr-rich clusters from the sites identified in the
2-AR (28) and rhodopsin (29-32), leaving open the
possibility that different kinases are involved in regulating these two
groups of GPRs or that a conserved tertiary structure that is not
evident from the primary sequence is important in the ability of these receptors to be agonist-dependent substrates for the
GRK.
In summary, the m2 mAChR contains two clusters of Ser/Thr residues, which are both potential sites for agonist dependent phosphorylation. However, desensitization of the m2 mAChR absolutely requires phosphorylation of the 307-311 Ser/Thr cluster; phosphorylation of the 286-290 Ser/Thr cluster has no effect on m2 mAChR signaling through adenylyl cyclase. In contrast, phosphorylation of either Ser/Thr cluster can support m2 mAChR internalization. The phosphorylation and internalization data suggest that while the kinase(s) and possibly the internalization machinery may recognize these domains in either orientation, the requirements for desensitization are more rigid. Identification of other proteins that may participate in the two events will yield greater understanding of the rapid regulation of m2 mAChR signaling and desensitization.
We thank Katharine Lee for technical assistance and Wolfgang Sadee for the generous contribution of the cDNA encoding m2-NAla4.