(Received for publication, June 7, 1996, and in revised form, September 11, 1996)
From the To determine whether functional receptor-G
protein coupling or signaling are required for internalization of the
thyrotropin-releasing hormone receptor (TRHR), we compared the
endocytosis of Gq-coupled and uncoupled receptors. A
hemagglutinin epitope-tagged TRHR (HA-TRHR) was in the
Gq-coupled state when bound to the agonist, MeTRH, and in a
nonsignaling state when bound to the HA antibody (12CA5). 12CA5 did not
induce an increase in [Ca2+]i or inositol
phosphates and did not inhibit [3H]MeTRH binding or
MeTRH-induced production of second messengers. Both agonist- and
antibody-bound HA-TRHRs were rapidly internalized via the same pathway;
internalization was sensitive to hypertonic shock, and both types of
internalized receptors were sorted into lysosomes. In addition, the
amino acid sequence CNC (positions 335-337) in the C-terminal tail
of the TRHR, which is important in ligand-induced receptor
internalization as determined by deletion mutagenesis (Nussenzveig, D. R., Heinflink, M., and Gershengorn, M. C. (1993) J. Biol. Chem. 268, 2389-2392), was also important for
12CA5-induced internalization. We expressed two truncated receptors,
HA-K338STOP and HA-C335STOP, in GH12C1
pituitary cells. Both HA-TRHR and HA-K338STOP were localized at the
plasma membrane of untreated cells and were translocated to
intracellular vesicles after MeTRH or 12CA5 binding; however,
HA-C335STOP was internalized and recycled constitutively. The
intracellular localization of HA-C335STOP was not altered by MeTRH;
however, 12CA5 binding induced the disappearance of internalized
HA-C335STOP and caused its localization at the plasma membrane,
indicating that constitutively cycling HA-C335STOP cannot be
reinternalized after antibody binding. Thus, amino acids 335-337,
which are important for the internalization of Gq-coupled
TRHRs, are also required for the sequestration of functionally
uncoupled TRHRs, and in addition, they act as an inhibitory signal that
prevents constitutive receptor internalization. Specifically, the Cys
residues at positions 335 and 337 are important for preventing
constitutive TRHR internalization, because a mutant HA-C335S/C337S
receptor was sequestered constitutively. We conclude that release from
a negative regulatory internalization sequence or domain is important
for HA-TRHR internalization and that the role of the CNC sequence in
internalization is independent of functional TRHR-Gq
coupling.
In the unoccupied state, cell surface 7 transmembrane domain
receptors do not stimulate signal transduction pathways, and they are
not usually internalized. After agonist binding, these receptors become
activated, couple to a G protein, and are internalized (1). It is not
understood how unoccupied receptors are prevented from becoming
sequestered constitutively or how agonist binding induces receptor
internalization. Because antagonist-bound gonadotropin-releasing hormone receptors and Gq coupling-impaired mutant type 1 muscarinic acetylcholine receptors were not internalized (2, 3), it has
been proposed that agonist-induced receptor-G protein coupling is
required for sequestration. However, this mechanism does not apply to
all 7 transmembrane domain receptors. For example, agonist-bound yeast
The thyrotropin-releasing hormone receptor (TRHR), a member of the
Gq-coupled receptor family (9-11), regulates prolactin release from anterior pituitary cells (12). It has been proposed that
functional receptor-Gq coupling is required for TRHR
internalization. For example, whereas agonist-bound TRHRs were
sequestered to intracellular vesicles at 37 °C (13,
14),2 antagonist-bound receptors remained
at the plasma membrane (15). In addition, two mutant TRHRs, one with a
deletion of the third intracellular loop ( We examined this possibility directly by studying the internalization
of another type of Gq-uncoupled TRHR. When bound to antibody, epitope-tagged TRHRs did not activate Gq.
However, antibody-bound, nonsignaling TRHRs were rapidly internalized
via the same pathway as agonist-bound, Gq-coupled
receptors. Antibody-bound receptors used the same C-terminal amino acid
sequence (CNC at positions 335-337) as agonist-bound receptors for
their sequestration. These results indicate that involvement of the CNC
internalization signal in TRHR sequestration does not require
functional receptor-G protein coupling. In addition, we found that the
same amino acid sequence (positions 335-337) that is required for
agonist- or antibody-induced TRHR internalization appears to be
important for preventing constitutive internalization of the receptor.
Therefore, release from tonic inhibitory signals is important for TRHR
internalization.
[3H]3-methyl His-TRH
([3H]MeTRH) (62.8 Ci/mmol) and
[3H]myo-inositol (10-25 Ci/mmol) were
purchased from DuPont NEN, and 3-methyl-His-TRH (MeTRH) was from
Peninsula Laboratories, Inc. (Belmont, CA). Goat anti-mouse IgG
conjugated to fluorescein or Texas Red were from Pierce, and
FITC-dextran was from Molecular Probes (Eugene, OR). Culture media were
purchased from Life Technologies, Inc.
GH12C1 and
F4C1 cells (both of which lack TRHRs) are rat
pituitary somatomammotrops that have been used extensively in control studies on the actions of TRH (17). The cells were grown at 37 °C in
monolayer culture in Ham's F-10 medium supplemented with 15% horse
serum and 2.5% fetal bovine serum in a humidified atmosphere of 5%
CO2-95% air (18). For confocal microscopy experiments, poly-L-lysine-coated glass coverslips were seeded with
equal numbers of cells from a pool of transiently transfected cultures.
GH12C1 cells were transfected with a
hemagglutinin (HA)-tagged cDNA (HA-TRHR, see below) by the
DEAE-dextran method. 1 µg of DNA was mixed in phosphate-buffered
saline (PBS) in a total volume of 170 µl. 85 µl of DEAE-dextran (2 mg/ml) were mixed with an equal volume of PBS. The DNA and DEAE
solutions were mixed and added to cells plated on 100-mm dishes for 15 min at room temperature. The solution was aspirated and the cells were
treated with 5 ml 10% dimethyl sulfoxide (diluted in PBS) for 2 min at
room temperature. After being washed with PBS, the cells were incubated
with 100 µM chloroquine in F-10 medium at 37 °C for
2 h. They were subsequently washed with PBS, incubated in F-10
medium at 37 °C for 16 h, and then plated on
poly-L-lysine coverslips at 1:3 dilution. Experiments were
performed 72 h after transfection. F4C1
cells were transfected by electroporation. 40 µg/ml of DNA was used
for transfection of 106 cells in a total volume of 0.3 ml
of F-10 medium. The cells were electroporated at 270 volts and 960 microfarads capacitance. They were plated in 100-mm dishes at 37 °C
for 16 h, and they were then split into 35-mm plates at 1:6
dilution. Ligand binding experiments for both types of transfected
cells were performed 48 h after transfection.
We used the Bluescript vector
containing the nucleotide sequence for the 9-amino acid HA epitope
(YPYDVPDYA), a generous gift from Dr. Tomas Kirchhausen (Harvard
Medical School), for tagging the N-terminal 233 amino acids of the
TRHR. The appropriate restriction enzyme sites (StuI and
EcoRI) for subcloning this N-terminal portion of the
receptor in frame to the 3 Mutant TRHRs were constructed by
site-directed mutagenesis using the polymerase chain reaction. Codons
for Lys-338 and Cys-335 were replaced by stop codons in the truncation
mutants. Codons for Cys-335 and Cys-337 were replaced by Ser codons in
the substitution mutants. Oligonucleotide primers containing the
desired mutations were designed and used in one- or two-step polymerase
chain reactions. Amplified polymerase chain reaction fragments were
subcloned into BglII and XmnI sites of TRHR
cDNA. The DNA sequences were verified by sequencing using the
dideoxynucleotide method (19).
F4C1 cells were transiently
transfected with the mutant (C335STOP, K338STOP, C335S/C337S) and
wild-type TRHR cDNAs. Each of the mutant receptors could activate
the inositol lipid signaling pathway to mediate increases in
[Ca2+]i. In the K338STOP mutant, TRH mediated
increases in [Ca2+]i that were
indistinguishable from the response of wild-type TRHRs. In the C335STOP
and C335S/C337S mutants, the [Ca2+]i
responses induced by TRH were 30 and 50%, respectively, of those given
by the wild-type TRHRs. All receptors were expressed at similar levels.
Similar results were obtained from cytosensor microphysiometry
measurements in F4C1 cells transiently
transfected with wild-type and mutant
TRHRs.3
Transiently transfected
GH12C1 cells, plated on
poly-L-lysine glass coverslips, were incubated at 37 or
4 °C with 1 µM MeTRH diluted in Hepes buffered salt
solution containing 0.5 mM Ca2+
(HBSS/Ca2+). The cells were then washed with ice-cold
HBSS/Ca2+ and fixed with 2% paraformaldehyde
(methanol-free) in PBS, pH 7.4, at room temperature for 20 min. The
slides were incubated in 50 mM NH4Cl for 10 min, and the cells were permeabilized with 1% digitonin for 5 min at
room temperature. After incubation with blocking solution (PBS/5%
bovine serum albumin/0.1% Triton X-100), the cells were incubated with
primary 12CA5 monoclonal antibody diluted (1:1000) in PBS/1% bovine
serum albumin/0.1% Triton X-100. They were then incubated with FITC-
or Texas Red-conjugated secondary antibody, diluted 1:100 and 1:200,
respectively, in PBS/1% bovine serum albumin/0.1% Triton X-100.
Incubations with the blocking solution and primary and secondary
antibodies were for 1 h at room temperature. The coverslips were
mounted on slides for confocal laser scanning microscopy (Bio-Rad,
MRC-600). Lysosomes were labeled by incubating cells with 20 mg/ml
FITC-dextran in HBSS/Ca2+ at 37 °C for 24 h,
washing with PBS, and reincubating in HBSS/Ca2+ for an
additional 1.5 h.
Three 100-mm dishes of
transiently transfected GH12C1 cells were used
for each measurement. Cells were detached from the dishes by incubation
with HBSS/0.2% EDTA and harvested by gentle centrifugation in
HBSS/Ca2+ buffer. Cells were loaded with 1 µM
fura 2/AM at 37 °C for 60 min. After being washed, cells were
resuspended in 3 ml of HBSS/Ca2+, and fluorescence
measurements were performed at 37 °C in a Spex Fluorolog F111A
spectrofluorometer (Spex Industries, Edison, NJ) at an excitation
wavelength of 342 nm and an emission wavelength of 492 nm (20).
GH12C1 cells transiently
transfected with the HA-TRHR cDNA were incubated with
[3H]myo-inositol in F-10 medium at 37 °C
for 16 h. They were then washed and reincubated in F-10 medium
without [3H]myo-inositol at 37 °C for
4 h. The cells were treated with 5 mM LiCl in
HBSS/Ca2+ at 37 °C for 10 min and incubated with MeTRH
or 12CA5 at 37 °C for 60 min. After extraction with 1 ml of ice-cold
trichloroacetic acid (final concentration, 5%), the radioactive
inositol polyphosphates were isolated by anion exchange chromatography
using 1 ml of Dowex-1 resin and a 0.1 M formic acid/1
M ammonium formate solution for elution (21). The eluates
were scintillation counted.
F4C1 cells transiently
transfected with the HA-TRHR were incubated with 1 µM
MeTRH or 12CA5 (diluted 1:250) at 37 °C for different periods of
time (see "Results"). The cells were subsequently washed with
ice-cold PBS and incubated with ice-cold 0.2 M
CH3COOH/0.5 M NaCl, pH 2.5 (acid/salt wash),
for 5 min. After cells were washed twice with cold PBS, radiolabeled
ligand binding was performed at 4 °C. The cells were incubated with
4 nM [3H]MeTRH in HBSS/Ca2+ (pH
7.2) at 4 °C for 2 h. Nonspecific binding was determined by the
addition of a 250-fold molar excess of unlabeled MeTRH. The binding
medium was aspirated and the cells were washed twice with 2 ml PBS.
They were then solubilized with 0.1 M NaOH, and the
radioactivity was measured by scintillation counting.
To
determine the role of functional TRHR-Gq coupling in
receptor internalization, we perturbed the receptor in a way that did
not activate its signaling pathway. We fused the hemagglutinin 9-amino
acid epitope tag to the extracellular N terminus of the TRHR. This
region of the TRHR is not essential for ligand binding or
agonist-induced signaling (22). The HA-TRHR was transiently expressed
in TRHR-negative GH12C1 cells. Epitope tagging
did not inhibit agonist-induced Gq coupling of the TRHR
because the HA-TRHR mediated an increase in cytosolic free
Ca2+ concentrations ([Ca2+]i) and
inositol polyphosphate production after MeTRH binding (Fig.
1, A and D). However, binding of
the monoclonal anti-HA 12CA5 did not activate HA-TRHR signaling; unlike
MeTRH, saturating concentrations of 12CA5 did not stimulate an increase
in [Ca2+]i or inositol phosphates (Fig. 1,
B and D). The slowly rising "hump" observed
in some of the Ca2+ studies with 12CA5 (Fig. 1B,
lower trace) was a nonspecific effect of the antibody
preparation, because 12CA5 treatment of GH12C1 cells expressing the untagged TRHR also generated a small, slowly rising Ca2+ signal in some experiments (Fig.
1C). Prebinding of 12CA5 did not inhibit subsequent
MeTRH-induced increases in [Ca2+]i (Fig.
1B), and the antibody did not attenuate specific
[3H]MeTRH binding (Fig. 1E). Our results
indicate that whereas MeTRH-bound HA-TRHRs are in the
Gq-coupled, activated state, receptors bound with 12CA5
alone are in a nonsignaling conformation. However, 12CA5 does not act
as an antagonist, because it does not inhibit ligand binding or
agonist-induced HA-TRHR signaling.
Agonist- and antibody-induced HA-TRHR internalization was
studied by immunofluorescence confocal microscopy. The HA-TRHR was localized at the plasma membrane of untreated
GH12C1 cells (Fig. 2A), and cells incubated for 60 min with
agonist or antibody at 4 °C (data not shown). Incubation with either
MeTRH (1 µM) or 12CA5 (1:1000) for 1, 3, 5, or 10 min at 37 °C induced translocation of HA-TRHRs from the plasma
membrane to intracellular vesicles (Fig. 2, B-I). Thus,
both agonist-bound, Gq-coupled and antibody-bound, uncoupled HA-TRHRs were rapidly internalized.
The internalization of both agonist- and antibody-bound receptors was
sensitive to hypertonic shock. GH12C1 cells,
expressing HA-TRHRs, were pretreated with 0.3 M
sucrose-containing medium at 37 °C for 20 min. They were
subsequently incubated with MeTRH or 12CA5 for 20 min at 37 °C.
Control cells were incubated with MeTRH or 12CA5 at 37 °C without
prior exposure to sucrose. Both agonist- and antibody-bound HA-TRHRs
were localized at the plasma membrane of sucrose-treated cells (Fig.
3) as compared to internalized receptors in control
cells. Because hypertonicity inhibits formation of clathrin-coated
vesicles (23), these results are consistent with the hypothesis that
both Gq-coupled and uncoupled HA-TRHRs become internalized
via the clathrin-coated pit pathway. We have recently shown directly
that the TRH-TRHR complex does, in fact, become internalized in
clathrin-coated vesicles.2
In general, receptors that are internalized via the clathrin-coated pit
pathway are subsequently transported to a recycling compartment and/or
to lysosomes. Our previous studies demonstrated that the amount of
recycled TRHRs decreased with increasing internalization time,
indicating that after agonist-induced internalization, TRHRs are sorted
into both a recycling and a noncycling pathway (13). We therefore
investigated the intracellular sorting of HA-TRHRs to lysosomes after
either agonist- or antibody-induced internalization by
immunofluorescence microscopy. Lysosomes were labeled by incubating cells with FITC-dextran at 37 °C for 24 h, washing away the
extracellular, unincorporated material, and incubating for an
additional 1.5 h at 37 °C. The FITC-dextran-loaded cells were
then incubated with either MeTRH or 12CA5 at 37 °C for 10 min to
induce agonist- and antibody-induced TRHR internalization. After being
washed, they were incubated at 37 °C for another 10 or 60 min to
allow for sorting of the internalized TRHRs. The HA-TRHRs were detected with Texas Red-conjugated anti-mouse immunoglobulin. There was partial
colocalization (yellow/orange signal) of both the MeTRH- and
12CA5-internalized Texas Red-labeled HA-TRHRs with lysosomal FITC-dextran (Fig. 4, A and B).
Those HA-TRHRs, which were not localized in lysosomes, probably
represent recycling receptors or noncycling receptors that had not yet
reached lysosomes. To eliminate the possibility that colocalization of
internalized HA-TRHRs with the lysosomal marker was an artifact that
resulted from leakage of the FITC-dextran from lysosomes into other
endocytic compartments during cell fixation and permeabilization, we
localized internalized transferrin, which is not sorted to lysosomes,
in cells that had incorporated FITC-dextran into their lysosomes. Internalized transferrin (Fig. 4C, red) was not
colocalized with lysosomal FITC-dextran (green). Thus,
colocalization of internalized HA-TRHRs is due to sorting of the
receptors to lysosomes.
In summary, based on their similarities with respect to sensitivity to
hypertonic shock and to sorting of the sequestered HA-TRHR into
lysosomes, the internalization pathways of both
Gq-coupled and functionally uncoupled HA-TRHRs appear
to be the same and probably represent clathrin-mediated
endocytosis.
We examined the amino acid sequence
involved in the internalization of 12CA5-bound HA-TRHRs. The C-terminal
region between amino acids 335 and 337 (CNC) has been implicated in
agonist-induced TRHR internalization (24). For example, the
internalization of TRHRs that were truncated at Cys-335 and Lys-338 was
20 and 50%, respectively, of the internalization observed for
wild-type receptors. In addition, mutant receptors in which Cys-335 and Cys-337 were replaced by serine or glycine (C335/337S) were
internalized to 50% of the levels of wild-type TRHRs. To determine
whether the same amino acid sequence is also involved in
antibody-induced receptor sequestration, we constructed two HA-epitope
tagged mutant TRHRs: HA-K338STOP and HA-C335STOP, which had the codons
for Lys-338 and Cys-335 mutated to stop codons. The truncated receptors
were expressed transiently in F4C1 cells and
specific [3H]MeTRH binding was measured (Fig.
5). The level of [3H]MeTRH binding was
similar for the HA-K338STOP and wild-type receptors. The HA-C335STOP
receptor also bound agonist, but to a lower extent than the other two
receptors. The lower [3H]MeTRH binding obtained for the
HA-C335STOP receptor was due to its lower expression compared to
wild-type and HA-K338STOP receptors (as determined by
immunofluorescence microscopy). Decreased affinity of the HA-C335STOP
receptor for the agonist is also a possibility that may contribute to
lower [3H]MeTRH binding. However, we did not measure the
affinity of this truncated receptor. Each of the truncated TRHRs was
functional as measured by its ability to mediate increases in
[Ca2+]i upon ligand binding (data not shown;
see "Experimental Procedures"). F4C1 cells
were used instead of GH12C1 in binding
experiments because they expressed higher levels of transfected
HA-TRHRs. F4C1 cells are another clonal strain
of rat anterior pituitary cells that do not express endogenous TRHRs
and exhibit no specific TRH binding (17).
Internalization of the truncated TRHRs compared to wild-type receptors
was studied in GH12C1 cells transiently
transfected with their respective cDNAs. Both HA-K338STOP and
HA-TRHR were localized at the plasma membrane of untreated
GH12C1 cells (Fig. 6,
A and D) and were translocated to intracellular
sites only after binding MeTRH or 12CA5 at 37 °C (Fig. 6, B,
C, E, and F). However, HA-C335STOP was localized in
intracellular sites in untreated cells (Fig. 6G). Incubation
with 0.3 M sucrose induced localization of HA-C335STOP at
the plasma membrane (data not shown), indicating constitutive
internalization of this receptor, rather than a defect in the
translocation of newly synthesized receptors to the plasma membrane
that could have resulted from this truncation. The intracellular localization of HA-C335STOP was not altered by incubation with MeTRH
binding at 37 °C (Fig. 6H). However, incubation with
12CA5 at 37 °C for 5 min resulted in the disappearance of
internalized HA-C335STOP and caused its localization at the plasma
membrane (Fig. 6I), demonstrating that HA-C335STOP was
cycling constitutively and that after antibody binding to cell surface
receptors, reinternalization was inhibited. Thus, the same C-terminal
CNC sequence (amino acids 335-337) that was previously shown to be
involved in MeTRH-induced TRHR internalization (17) is also important
in the sequestration of 12CA5-bound receptors. Therefore, the endocytic
step that involves the CNC internalization sequence does not require
functional TRHR-Gq coupling or signaling. Furthermore, the
finding that HA-C335STOP, but not HA-K338STOP, was constitutively
internalized suggests that in addition to their role in mediating
agonist- and antibody-induced internalization, amino acids 335-337 are
also important in preventing constitutive TRHR sequestration.
To determine whether specific amino acids within the CNC sequence are
involved in preventing constitutive TRHR internalization, we studied
the internalization of another epitope-tagged mutant receptor, in which
the cystine residues at positions 335 and 337 were replaced by serine
(HA-C335S/C337S). These mutant receptors could signal upon ligand
binding (see "Experimental Procedures"). Whereas wild-type HA-TRHRs
were localized at the plasma membrane of untreated cells (Fig.
7D) and became internalized after agonist or
antibody binding (Fig. 7, E and F),
HA-C335S/C337S mutants were found in intracellular sites of untreated
cells (Fig. 7A), indicating constitutive sequestration of
these mutant receptors. MeTRH or 12CA5 binding did not alter the
intracellular localization of HA-C335S/C337S (Fig. 7, B and
C). We conclude that Cys-335 and/or Cys-337 prevent
constitutive sequestration of the HA-TRHR.
Although the internalization pathways involved in
rapid agonist- and antibody-induced HA-TRHR sequestration were similar, it was possible that the efficiency of the two modes of internalization was different. We quantitated the extent of MeTRH-, 12CA5-, or MeTRH
plus 12CA5-induced HA-TRHR internalization by using radiolabeled ligand
as the tracer. F4C1 cells expressing the
HA-TRHR were incubated with saturating concentrations of MeTRH, 12CA5,
or MeTRH plus 12CA5 at 37 °C for the indicated times. Cells were
then washed with high salt/low pH buffer to remove cell surface
receptor-bound MeTRH or 12CA5. The amount of HA-TRHRs remaining at the
plasma membrane after internalization was determined by incubation with 4 nM [3H]MeTRH at 4 °C. Loss of cell
surface receptors was greater in cells treated with MeTRH or with MeTRH
plus 12CA5 than in cells incubated with 12CA5 alone (Fig.
8A), indicating that a greater number of
agonist-bound than antibody-bound receptors were internalized. Although
we could not detect significant antibody-induced loss of plasma
membrane HA-TRHRs in cells incubated with 12CA5 for up to 60 min using
radiolabeled ligand binding (Fig. 8A), translocation of
12CA5-bound receptors to intracellular sites by the more sensitive immunofluorescence confocal microscopy technique was clearly evident (Fig. 8, C versus D and E). In addition, longer
incubation with antibody at 37 °C for 2 h resulted in a
significant loss of plasma membrane receptors as measured by ligand
binding (Fig. 8B). We conclude that the extent of
sequestration was higher for agonist-bound, Gq-coupled
receptors than for antibody-bound, functionally uncoupled receptors.
The reason that a low number of antibody-bound, rapidly internalized
HA-TRHRs were not detectable by radiolabeled ligand binding and that
the extent of ligand-induced internalization was lower in transfected
F4C1 cells as compared to
GH4C1 cells, which express endogenous TRHRs
(13), may be the higher expression level of HA-TRHRs in transiently
transfected cells as compared to the level of endogenous TRHR
expression. A 50-fold higher receptor expression in transfected
F4C1 cells as compared to endogenous receptor
expression in GH4C1 cells was estimated from
[3H]MeTRH binding and transfection efficiency. Therefore,
the relatively small number of rapidly internalized antibody-bound
HA-TRHRs would not result in a significant decrease in the total number
of cell surface receptors that would be detectable by radiolabeled
ligand binding.
Agonist-induced translocation of plasma membrane receptors into
clathrin-coated pits depends on the presence of internalization motifs
in the cytoplasmic domains of the receptors. The most common sequestration signals are the NPXY and
YXX-hydrophobic motifs present in nutrient and growth factor
receptors (25). They appear to mediate receptor internalization by
binding to the µ chain of the AP-2 adaptor complex (26). A similar
NPXXY sequence is present in the C-terminal tails of many 7 transmembrane domain receptors (28). Although this sequence is
important for the sequestration of the An NPXXY sequence is present in the TRHR at residues
316-320. No site-directed mutagenesis of these residues has been
performed to determine the role of the NPXXY motif in TRHR
internalization. However, regardless of any role for the
NPXXY sequence in TRHR sequestration, two other C-terminal
regions appear to be important for agonist-induced TRHR
internalization: the CNC sequence (amino acids 335-337) near the
putative seventh transmembrane helix and the more distal SDRFSTEL
sequence (amino acids 360-368) (24). Although mutant TRHRs that were
truncated at residue 368 were internalized like wild-type receptors,
the internalization of agonist-bound TRHRs that were truncated at
residue 360 was decreased by 50%. Receptors that were truncated at
amino acid 338 or receptors with amino acid substitutions at Cys-335
and Cys-337 were also internalized to 50% of the sequestration level
of wild-type TRHRs, whereas the internalization of receptors truncated
at Cys-335 was decreased to 20% of wild-type receptors (24). The
relatively low level of internalization for the agonist-bound
HA-C335STOP receptor could be readily detected by immunofluorescence
confocal microscopy experiments in our studies. However, Ashworth
et al. (14) reported that C335STOP TRHRs were localized
exclusively at the plasma membrane of pituitary GH3 cells
incubated with rhodamine-conjugated TRH at both 4 and 37 °C. The
difference between the results of Ashworth et al. (14) and
our findings may be explained by the low fluorescence signal obtained
with rhodamine-conjugated TRH as compared to the amplified signal
obtained with FITC-conjugated secondary antibodies or by the different
expression levels of the truncated receptors in the two studies.
We investigated the role of functional TRHR-Gq coupling in
TRHR internalization. Functional coupling is defined as an interaction between the receptor and its cognate G protein, which allows activation of the downstream effector enzyme to generate second messenger molecules. Both Gq-coupled and uncoupled TRHRs were
internalized via the same hypertonic shock-sensitive pathway within 1 min of binding to MeTRH and 12CA5, respectively. Like agonist-bound
receptors (24), antibody-bound receptors used the same CNC
internalization signal. For example, whereas HA-K338STOP receptors
became internalized after antibody binding, HA-C335STOP receptors were
trapped at the plasma membrane. We conclude that the TRHR
internalization step that requires the CNC internalization motif is
independent of functional receptor-G protein coupling. Because
internalization sequences mediate receptor association with clathrin by
interacting with adaptor complexes (26), it is possible that the
CNC-dependent sequestration step involves the association
of TRHRs with clathrin-coated pits. However, this step is not
sufficient for TRHR sequestration, because the extent of endocytosis
was higher for agonist-bound HA-TRHRs than for antibody-bound
receptors. This finding could not be explained by a difference in
receptor expression because the cells that were used for MeTRH- and
12CA5-induced receptor internalization were from a common pool. The
exposure of secondary internalization motifs in the agonist-bound but
not in the antibody-bound receptors (reflecting differences in the
conformation of the two types of internalization-competent receptors)
could increase the extent of MeTRH-induced TRHR endocytosis. This
possibility is consistent with our finding that HA-C335STOP receptors,
lacking the CNC internalization sequence, were internalized
constitutively and when bound to agonist but not when bound to
antibody. Thus, TRHR-Gq coupling may increase the number of
internalized TRHRs via mechanisms that do not involve the CNC
sequence.
Gq coupling-dependent and -independent
mechanisms may also regulate In addition to its positive regulatory role in agonist- and
antibody-induced TRHR internalization, the CNC sequence may also be a
negative regulatory motif that prevents constitutive receptor internalization in the absence of ligand. For example, deletion of this
sequence or mutagenesis of Cys-335 and Cys-337 resulted in constitutive
receptor internalization (Figs. 6G and 7B). This C-terminal region may prevent constitutive TRHR sequestration by
masking the accessibility of an additional sequestration signal in the
unoccupied receptor or by binding to internalization-inhibitory factors. Alternatively, removal of the putative isoprenylation site
(Cys-335) in the HA-C335STOP and HA-C335S/C337S receptors may induce
constitutive receptor internalization. Also, because the C335STOP
mutant TRHR was reported to signal constitutively (31), it is possible
that the active, Gq-coupled conformation of the HA-C335STOP
mutant receptor is responsible for the constitutive internalization of
this receptor. No data are yet available on constitutive activation of
the C335S/C337S receptors. We propose that negative regulatory
mechanisms involving Cys-335 and/or Cys-337 inhibit constitutive TRHR
internalization and are overcome upon agonist or antibody binding to
allow receptor sequestration.
In summary, our studies show that when unoccupied, the CNC sequence in
the C terminus of the TRHR directly or indirectly blocks constitutive
receptor internalization. The same sequence is also important for
mediating internalization of agonist- and antibody-bound receptors,
independently of functional TRHR-Gq coupling. However, TRHR-Gq coupling may be important for additional
unidentified steps that increase the number of receptors recruited for
endocytosis. Future studies, using 12CA5 Fab fragments, will determine
whether antibody-induced, Gq coupling-independent HA-TRHR
internalization requires receptor dimerization or whether monovalent
antibody-induced changes in receptor conformation are sufficient for
triggering internalization.
Department of Molecular and Cellular
Toxicology,
-factor receptors and mammalian
2-adrenergic receptors (
2ARs)1 became internalized in cells
that did not express the G protein to which these receptors couple (4,
5). Mutant
2ARs and type 1 angiotensin II receptors that did not
effectively couple to Gs
and Gq
, respectively, were
sequestered to the same extent as wild-type receptors (6, 7). In
addition, antagonist-bound angiotensin type 1a receptors were
internalized about 50% as efficiently as agonist-bound receptors
(8).
218-263) and another with
replacement of Asp-71 by Ala (D71A), were impaired in both
Gq-mediated signaling and agonist-induced internalization
(16). However, although the sequestration of these mutant receptors was
greatly decreased, it was not completely abolished (15% of the mutant
receptors still became internalized), raising the possibility that TRHR
internalization is partially independent of functional receptor-G
protein coupling or signaling.
Materials
end of the HA sequence in the Bluescript
vector were generated by the polymerase chain reaction. The sequence
for the 5
primer was GGCACTAGGCCTGGAGAATGAAACCG and the sequence
for the 3
primer was GCCCAAGGTATTCTTGAGGATCCTCATAT. The
epitope-tagged Nterminal portion of the TRHR was then sequenced by
the dideoxynucleotide method (19) and ligated to either the remainder
of the wild-type TRHR, the C335STOP, the K338STOP, or the C335S/C337S
TRHR sequence in the pCDM8 vector. The HA-TRHR, HA-K338STOP,
HA-C335STOP, and HA-C335S/C337S cDNAs were used for transient
transfections (see above).
Antibody-bound HA-TRHRs Are in a Nonsignaling Conformation
Fig. 1.
Binding of monoclonal 12CA5 antibody
did not stimulate signaling by the HA-TRHR and did not inhibit
[3H]MeTRH binding or MeTRH-induced signaling.
Transiently transfected GH12C1 cells
expressing the HA-TRHR were loaded with 1 µM fura 2/AM
and incubated with 1 µM MeTRH (A) or with
12CA5 diluted at 1:250 (B). Cells expressing the untagged
TRHR were incubated with 12CA5 1:250 (C). Alternatively,
cells were prelabeled with [3H]inositol and stimulated
with 1 µM MeTRH or 12CA5 (1:250) at 37 °C for 60 min or remained untreated (D). E,
GH12C1 cells transiently expressing the
HA-TRHR were incubated with 4 nM [3H]MeTRH at
37 °C for 1 h in the presence or absence (Control) of 12CA5 (1:500), and specific binding was measured. There was no
specific binding in nontransfected GH12C1
cells. The values given in D and E are the
average of duplicate samples and the bars represent the
ranges. The small size of the [Ca2+]i responses
to MeTRH is due to the relatively low (<5%) transfection efficiency
of GH12C1 cells with the HA-TRHR.
[View Larger Version of this Image (44K GIF file)]
Fig. 2.
Time course for MeTRH- and 12CA5-induced
HA-TRHR internalization. Cells were incubated with 1 µM MeTRH (B-E) or 12CA5 antibody
(F-I) diluted 1:1000 at 37 °C for 1 (B and
F), 3 (C and G), 5 (D and
H) or 10 (E and I) min. Cells in
panel A were untreated. After incubation, the cells were
fixed, permeabilized and incubated with 12CA5 antibody (1:1000). The
cells were stained by incubation with FITC-conjugated goat anti-mouse
IgG. The bars (lower left) indicate 5 µm.
[View Larger Version of this Image (158K GIF file)]
Fig. 3.
Inhibition of MeTRH- and 12CA5-induced
HA-TRHR internalization by hypertonic sucrose. Cells were
incubated with F-10 medium containing (B and D)
or lacking (A and C) 0.3 M sucrose at
37 °C for 20 min. Subsequently they were incubated with 1 µM MeTRH (A and B) or 12CA5
(1:1000) (C and D) at 37 °C for 20 min. The
cells were fixed, permeabilized, and incubated with 12CA5 (1:1000).
Fluorescence staining was achieved by incubation with FITC-conjugated
goat anti-mouse IgG. Internalized HA-TRHRs were observed in the absence
of high sucrose (A and C), and internalization was blocked completely in the presence of 0.3 M sucrose
(B and D). The bar (lower
right) indicates 10 µm.
[View Larger Version of this Image (116K GIF file)]
Fig. 4.
Localization of internalized HA-TRHRs in
lysosomes. Cells were incubated with FITC-dextran at 37 °C for
24 h, washed, and reincubated in the absence of FITC-dextran at
37 °C for 1.5 h. Cells were subsequently incubated with 1 µM MeTRH (A) or 12CA5 (1:1000) (B)
for 10 min at 37 °C. They were then washed and reincubated at
37 °C for 10 or 60 min. Control, FITC-dextran-labeled cells were
incubated with 50 nM transferrin at 37 °C for 20 min
(C). Cells were processed for microscopy as described in
previous Figs. HA-TRHRs and transferrin were stained with Texas
Red-conjugated secondary IgG for confocal fluorescence microscopy. The
images indicating lysosomal labeling (green) and HA-TRHR or
transferrin labeling (red) were overlaid, and colocalization
of internalized HA-TRHRs with lysosomal FITC-dextran is indicated by
the yellow/orange punctate signal (arrowheads) (A
and B). There was no colocalization of internalized
transferrin, which is not sorted to lysosomes, with FITC-dextran
(C). The bar (lower right) indicates
10 µm.
[View Larger Version of this Image (92K GIF file)]
Fig. 5.
Mutant TRHRs, HA-K338STOP, and HA-C335STOP
bind [3H]MeTRH. F4C1 cells
transiently expressing HA-K338STOP or HA-C335STOP were incubated with 4 nM [3H]MeTRH at 37 °C for 60 min. The
amount of radiolabeled ligand specifically bound was measured by
scintillation counting. The values presented represent the average of
duplicate samples, and the bars indicate the ranges. The
lower but significant binding observed with the HA-C335STOP mutant was
primarily due to lower expression of this receptor compared to HA-TRHR
and HA-K338STOP (see "Results").
[View Larger Version of this Image (45K GIF file)]
Fig. 6.
Effect of MeTRH and 12CA5 binding on the
localization of HA-K338STOP and HA-C335STOP.
GH12C1 cells expressing wild-type HA-TRHRs
(A-C), mutant HA-K338STOP (D-F), and
HA-C335STOP (G-I) TRHRs were untreated (A, D,
and G) or were incubated with 1 µM MeTRH
(B, E, and H) or 12CA5 (1:250) (C, F,
and I) at 37 °C for 5 min. Green fluorescence
represents HA-TRHRs labeled with 12CA5 primary antibody and
FITC-conjugated secondary antibody after cell fixation and
permeabilization. No internalization was observed in unpermeabilized
cells. The HA-C335STOP mutant was expressed at lower levels than
wild-type and HA-K338STOP receptors (see "Results"). The
bar (lower right) indicates 10 µm. The HA-TRHR and HA-K338STOP receptors were localized at the plasma membrane of
untreated cells and became internalized only after agonist or antibody
binding. The HA-C335STOP receptor was localized in intracellular sites
in untreated cells and in cells treated with agonist. However,
incubation with 12CA5 antibody trapped the HA-C335STOP receptors at the
plasma membrane (I).
[View Larger Version of this Image (89K GIF file)]
Fig. 7.
Effect of MeTRH and 12CA5 binding on the
localization of the mutant HA-C335S/C337S TRHRs.
GH12C1 cells expressing the HA-C335S/C337S
(A-C) or wild-type HA-TRHRs (D-F) were
untreated (A and D) or treated with 1 µm MeTRH
(B and E) or 12CA5 (1:250) (C and
F) at 37 °C for 5 min. Green fluorescence
represents HA-C335S/C337S detected by FITC-conjugated secondary
antibodies. The bar (lower right) indicates 10 µm. HA-C335S/C337S receptors were localized in intracellular sites in
untreated cells (A) and cells incubated with MeTRH or 12CA5
(B and C). Control wild-type HA-TRHRs were localized at the plasma membrane of untreated cells (D) and
became internalized only after MeTRH or 12CA5 binding (E and
F).
[View Larger Version of this Image (75K GIF file)]
Fig. 8.
MeTRH-bound HA-TRHRs were internalized to a
greater extent than 12CA5-bound HA-TRHRs. A and
B, F4C1 cells were incubated with 1 µM MeTRH (open columns), 12CA5 at 1:250
(gray columns), or MeTRH plus 12CA5 (1 µM and
1:250) (filled columns) at 37 °C for the times indicated.
The cells were then washed with a high salt/low pH buffer. Binding of 4 nM [3H]MeTRH was performed at 4 °C. The
amount of [3H]MeTRH bound was measured by scintillation
counting. Each column represents the mean of duplicate
values and the bars show the ranges. Receptors that were
bound to agonist or to agonist plus antibody were internalized to a
greater extent than receptors that were bound to antibody alone.
C-E, confocal immunofluorescence microscopy localizing the
HA-TRHR in F4C1 cells that were untreated (C) or incubated with 1 µM MeTRH
(D) or with 12CA5 (1:1000) (E) at 37 °C for 10 min. Although 12CA5-induced receptor internalization was not detectable
after 10 min of incubation at 37 °C by radiolabeled ligand binding
(A, gray columns), internalized receptors could be observed
by immunofluorescence microscopy (E).
[View Larger Version of this Image (53K GIF file)]
2AR (27), it is not a common
internalization motif among all 7 transmembrane domain receptors (28,
29).
2AR internalization (30). For example,
the initial translocation of
2ARs from microvilli to clathrin-coated
pits at the plasma membrane was dependent on agonist binding. This finding may indicate a requirement for the agonist-bound conformation of the receptor or for agonist-induced receptor-G protein coupling. Subsequent receptor sequestration into intracellular vesicles was
independent of agonist binding (30). Once agonist-bound receptors
translocated to clathrin-coated pits, dissociation of the
receptor-ligand complexes with an excess of antagonist did not inhibit
2AR sequestration into vesicles. Further studies are necessary to
determine whether the receptor-G protein coupling-independent TRHR
internalization step is analogous to this agonist-independent step for
2AR sequestration. In addition, studying agonist-induced TRHR
internalization in cells that do not express Gq
is the
most direct way to compare Gq
coupling-dependent and -independent steps in TRHR
endocytosis. Such Gq
knockout cell strains are not yet available.
*
This investigation was supported in part by Research Grant
DK 11011 from the National Institute of Diabetes, Digestive and Kidney
Diseases. The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement" in
accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
¶
To whom correspondence should be addressed: Department of
Molecular and Cellular Toxicology, Harvard School of Public Health, 665 Huntington Ave., Boston, MA 02115. Tel: 617-432-1177; Fax: 617-432-1780; E-mail: tashjian{at}hsph.harvard.edu.
1
The abbreviations used are: 2AR,
2-adrenergic receptor; HA, hemagglutinin; HBSS/Ca2+,
Hepes-buffered salt solution plus 0.5 mM Ca2+;
PBS, phosphate-buffered saline; TRH, thyrotropin-releasing hormone; MeTRH, 3-methyl His2-TRH; TRHR, TRH receptor; HA-TRHR,
HA-epitope-tagged TRHR; FITC, fluorescein isothiocyanate.
2
C. Petrou and A. H. Tashjian, Jr., submitted for
publication.
3
L. Chen and A. H. Tashjian, Jr., manuscript in
preparation.
©1997 by The American Society for Biochemistry and Molecular Biology, Inc.