(Received for publication, February 12, 1997)
From the Departments of Biochemistry and Molecular Pharmacology and Microbiology and Immunology, Kimmel Cancer Institute, Thomas Jefferson University, Philadelphia, Pennsylvania 19107
We have recently demonstrated that the nonvisual
arrestins, -arrestin and arrestin3, interact with high affinity and
stoichiometrically with clathrin, and we postulated that this
interaction mediates internalization of G protein-coupled receptors
(Goodman, O. B., Jr., Krupnick, J. G., Santini, F., Gurevich, V. V.,
Penn, R. B., Gagnon, A. W., Keen, J. H., and Benovic, J. L. (1996)
Nature 383, 447-450). In this study, we localized the
clathrin binding domain of arrestin3 using a variety of approaches.
Truncation mutagenesis demonstrated that the COOH-terminal half of
arrestin3 is required for clathrin interaction. Assessment of the
clathrin binding properties of various glutathione
S-transferase-arrestin3 fusion proteins indicated that the predominant
clathrin binding domain is contained within residues 367-385. Alanine
scanning mutagenesis further localized this domain to residues
371-379, and site-directed mutagenesis demonstrated the critical
importance of both hydrophobic (Leu-373, Ile-374, and Phe-376) and
acidic (Glu-375 and Glu-377) residues in the arrestin3/clathrin
interaction. These results are complementary to the observation that
hydrophobic and basic residues in clathrin are critical for its
interaction with nonvisual arrestins (Goodman, O. B., Jr., Krupnick,
J. G., Gurevich, V. V., Benovic, J. L., and Keen, J. H. (1997)
J. Biol. Chem. 272, 15017-15022). Lastly, an
arrestin3 mutant in which Leu-373, Ile-374, and Phe-376 were mutated to
Ala was significantly defective in its ability to promote
2-adrenergic receptor internalization in COS-1 cells
when compared with wild-type arrestin3. Taken together, these results
implicate a discrete region of arrestin3 in high affinity binding to
clathrin, an interaction critical for agonist-promoted internalization
of the
2-adrenergic receptor.
G protein-mediated signal transduction involves agonist activation
of a seven transmembrane-spanning G protein-coupled receptor (GPR)1 which, in turn, activates a heterotrimeric
guanine nucleotide-binding protein (G protein). It is
now well established that both the and
subunits of G
proteins modulate the activity of many effectors including adenylyl
cyclases, phospholipases, ion channels, and cGMP phosphodiesterase (for
review, see Refs. 1 and 2). Within seconds to minutes following agonist
exposure, activated GPRs lose their ability to respond to agonist with
their original sensitivity, a phenomenon commonly referred to as
desensitization. Receptor desensitization is initiated by
phosphorylation of the agonist-activated GPR by a family of enzymes
known as G protein-coupled receptor kinases (GRKs) leading to high
affinity binding of a second class of proteins known as arrestins (for
review, see Refs. 3 and 4). It is thought that arrestin binding to the
phosphorylated GPR sterically inhibits G protein binding (5). To date,
four mammalian arrestins have been identified. These include two visual arrestins (arrestin and C- or X-arrestin) (6-8), which likely regulate
photoreceptors based on their restricted localization, and two
nonvisual arrestins (
-arrestin and arrestin3) (9-12), which are
ubiquitous and likely regulate a wide variety of GPRs.
Another level of regulation of GPR signaling involves internalization of the activated receptor into a compartment distinct from the plasma membrane, a process known as sequestration (for review, see Ref. 13). Utilizing a variety of techniques, many studies have demonstrated that at least one mechanism by which GPRs internalize is via clathrin-coated pits (14-19). Interestingly, recent studies have demonstrated a role for sequestration in resensitization of the activated GPR, suggesting a model in which the desensitized receptor is dephosphorylated by an intracellular vesicle-derived phosphatase and then recycled back to the plasma membrane (20-23).
Recently, several studies have implicated GRKs and arrestins in the
internalization of agonist-activated GPRs. Overexpression of GRK2
(-adrenergic receptor kinase) enhanced sequestration of coexpressed
m2 muscarinic acetylcholine receptor, whereas a dominant-negative
-adrenergic receptor kinase inhibited receptor sequestration (24).
Similarly, overexpression of
-adrenergic receptor kinase rescued
sequestration of a
2-adrenergic receptor mutant (Y326A)
defective in its ability to be sequestered, whereas sequestration of
2-adrenergic receptor mutant Y326A lacking
-adrenergic receptor kinase phosphorylation sites was not rescued
(25). Indeed, coexpression of either GRK3, GRK4, GRK5, or GRK6, but not
the photoreceptor-specific GRK1, also enhanced sequestration of
2-adrenergic mutant Y326A (26). It was further observed that overexpression of
-arrestin or arrestin3 (
-arrestin2) alone rescued the sequestration of
2-adrenergic mutant Y326A,
whereas putative dominant-negative mutants of these arrestins inhibited sequestration of the wild-type
2-adrenergic receptor
(27). Interestingly, overexpression of
-arrestin also enhanced
internalization of the angiotensin type II receptor via clathrin-coated
pits, a compartment that is not normally utilized by this receptor
(28).
We recently proposed that arrestins could both desensitize
agonist-activated GPRs and promote their sequestration by interacting not only with the GPR but also with clathrin, the major protein component of the clathrin-based endocytic machinery (29). We observed
that -arrestin and arrestin3, but not visual arrestin, interact
specifically, stoichiometrically, and with high affinity and rapid
kinetics with clathrin. Moreover, immunofluorescence analyses
demonstrated that the activated
2-adrenergic receptor,
-arrestin, and clathrin all colocalize in intact cells upon agonist addition, suggesting that the arrestin/clathrin interaction observed in vitro also occurs in cells in the presence of an
activated receptor. Thus,
-arrestin and arrestin3 appear to act as a
signal for internalization of agonist-activated GPRs by virtue of their ability to target the desensitized receptor to clathrin-coated pits. In
this study, we localize the clathrin binding domain of arrestin3 to a
small stretch of residues in the far COOH terminus of the molecule.
Interestingly, this region is highly conserved between
-arrestin and
arrestin3 but is largely absent in the visual arrestins.
Characterization of the clathrin binding domain in arrestins should
provide a powerful means of disrupting the arrestin/clathrin
interaction in cells.
[3H]Leucine and
[-32P]ATP were purchased from DuPont NEN. pGEX4T-2
vector was obtained from Pharmacia Biotech. All restriction enzymes
were purchased from Boehringer Mannheim. Sepharose 2B and all other
chemicals were purchased from Sigma. SP6 polymerase and rabbit
reticulocyte lysate were prepared as described previously (30, 31).
11-cis-Retinal was generously supplied by Dr. R. K. Crouch.
Purified rhodopsin kinase was generously supplied by Drs. J. Pitcher
and R. Lefkowitz.
Truncation mutants of arrestin3S (409 amino acids) (12) were generated by linearizing the arrestin3 cDNA in pGEM-2 with the restriction enzyme SfuI or PvuII to obtain mRNAs encoding residues 1-375 and 1-183, respectively. Alanine scanning mutants of arrestin3 were generated by a two-step PCR protocol using the arrestin3 cDNA as template. In the first reaction, an ~800-bp product was generated using a forward primer (42 or 43 bp beginning at base ~1120-1165 of arrestin3 cDNA) containing the specific base mutations and a 21-bp reverse primer beginning at base 400 of pGEM-2. This product was purified and used together with an AccI-HindIII restriction fragment of arrestin3 cDNA as template for a second PCR. Here, an ~1,200-bp product was generated using a forward primer (22 bp beginning at base ~800 of arrestin3 cDNA) and the pGEM-2 reverse primer. This product was purified, digested with XhoI and HindIII, and an ~510-bp fragment purified on a 1.2% agarose gel. The ~510-bp fragment was then subcloned into the XhoI-HindIII-digested and phosphatase-treated arrestin3 cDNA. Site-directed mutants of arrestin3 were generated as described above except forward primers for the first PCR, which contained the specific base mutation(s), started at base ~1129-1152 and were 34-46 bp in length. All sequences were confirmed by DNA sequencing (Nucleic Acid Facility, Kimmel Cancer Center, Thomas Jefferson University).
In Vitro Transcription and Translation of Arrestin3Linearization for generation of arrestin3 truncation mutants was as described above. Wild-type, alanine scanning, and site-directed mutant cDNAs of arrestin3 were linearized with HindIII and purified on a 1% agarose gel. In vitro transcription and translation of arrestin3 proteins were performed exactly as described previously (32). The specific activities and concentrations of in vitro synthesized proteins were calculated to be ~500-1,200 dpm/fmol and ~5-20 nM, respectively.
Generation of Glutathione S-Transferase (GST)-Arrestin3 Fusion ProteinsSpecific sequences within the COOH-terminal half of bovine arrestin3 were amplified by PCR using a forward primer (~30 bp) containing an engineered SmaI restriction site and a reverse primer (~30 bp) containing both an engineered SmaI restriction site and a stop codon. PCRs were purified, digested with SmaI, repurified, and then subcloned into SmaI-cut and phosphatase-treated pGEX4T-2. All sequences were confirmed by DNA sequencing.
Production of GST-arrestin3 fusion proteins was carried out according
to the manufacturer's instructions (Pharmacia). Briefly, Escherichia coli BL21 cells, transformed with specific
GST-arrestin3 fusion protein constructs, were grown at 37 °C to an
OD600 ~1 and then induced with 100 µM
isopropyl-1-thio--D-galactopyranoside for 2 h at
room temperature. Cells (1 liter) were centrifuged for 30 min at 5,000 rpm, 4 °C and resuspended in 15 ml of ice-cold phosphate-buffered
saline (PBS), 1 mM phenylmethylsulfonyl fluoride, 20 µg/ml leupeptin, 200 µg/ml benzamidine (buffer A). The cells were
incubated 10 min on ice following the addition of lysozyme (1 mg/ml),
and then Triton X-100 (1%) and dithiothreitol (5 mM) were
added, and the cells were lysed by two freeze/thaw cycles. DNase was
added to ~50 µg/ml, and the cells were incubated on ice for 30 min
and then centrifuged for 30 min at 50,000 rpm at 4 °C to obtain a
cleared lysate. The lysate was incubated with glutathione-Sepharose 4B
(0.8 ml) for 60 min at 4 °C on a rotator. The sample was centrifuged
for 1 min at 1,000 rpm, 4 °C, washed three times with buffer A
containing 0.1% Triton X-100, two times with buffer A, and then either
resuspended in 0.8 ml of buffer A (for experiments involving
GST-arrestin3 beads) or eluted with 10 mM reduced
glutathione, 50 mM Tris-HCl, pH 8.0, 1 mM
phenylmethylsulfonyl fluoride. Protein concentrations were determined
by dye binding (Bio-Rad) using bovine serum albumin as standard.
The preparation of clathrin and clathrin cages (33) and the clathrin cage binding assay (34) have been described previously. Briefly, in vitro synthesized radiolabeled arrestins (~0.5 nM) or GST-arrestin3 fusion proteins (15 nM) were incubated with or without clathrin cages for 30 min at 22 °C in a total volume of 50 µl in 100 mM Na-MES, pH 6.8; 1 mM dithiothreitol; 1 µg/ml each of leupeptin, pepstatin, and antipain; and 0.1 mg/ml bovine serum albumin (in the case of GST-arrestin3 fusion proteins). Samples were then cooled on ice, loaded onto a 75-µl 0.2 M sucrose cushion in 100 mM Na-MES, pH 6.8, and centrifuged for 5 min at 100,000 rpm, 4 °C in a TLA100.1 rotor. Pellets were solubilized in 20 µl of SDS sample buffer, heated at 100 °C for 5 min, and run on 10% SDS-polyacrylamide gels. For in vitro synthesized radiolabeled proteins, gels were stained with Coomassie Blue, destained, enhanced with 20% 2,5-diphenyloxazole in glacial acetic acid, dried, and autoradiographed. Quantitation was carried out by PhosphorImager analysis (Molecular Dynamics). After subtracting out nonspecific sedimentation in the absence of clathrin cages (~9% of input), specific binding was normalized to the binding for wild-type arrestin. For GST-arrestin3 fusion proteins, proteins were transferred to a nitrocellulose membrane at 100 V for 1 h. Blots were blocked for 30 min with 5% nonfat dry milk in 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.05% Tween 20 (buffer B) and then incubated for 1 h with a 1:250 dilution of PF2 antibody (recognizes GST) in buffer B. Blots were washed three times for 5 min each with buffer B, incubated for 1 h with a 1:2,000 dilution of goat anti-rabbit secondary antibody in buffer B, washed four times for 5 min each with buffer B, incubated with ECL reagent (Amersham) for 1 min, and then exposed to film for 5-10 s. Films were scanned by densitometry (Molecular Dynamics) to quantitate binding. Nonspecific sedimentation was assessed in the absence of clathrin cages (~8% of input) and subtracted from all experimental values.
Preparation of Rod Outer Segment (ROS) MembranesBovine ROS
membranes containing >95% rhodopsin were prepared as described
previously (32). Phosphorylated rhodopsin was generated by incubating
ROS membranes (~150 µg of rhodopsin) with purified recombinant
rhodopsin kinase in a total volume of 500 µl in 20 mM
Tris-HCl, pH 7.5, 2 mM EGTA, 7.5 mM
MgCl2, 100 µM ATP for 1 h at 30 °C
under bright light illumination. Phosphorylated ROS membranes were
centrifuged for 45 min at 50,000 rpm, 4 °C, washed with 1 ml of 20 mM Tris-HCl, pH 7.5, 2 mM EGTA, resuspended in
200 µl of 20 mM Tris-HCl, pH 7.5, 2 mM EGTA,
2 mM dithiothreitol, sonicated twice for 10 s on ice,
and then regenerated with a 3-fold molar excess of
11-cis-retinal for 40 min at 37 °C in the dark followed
by another 3-fold molar excess of 11-cis-retinal for 2 h at 37 °C in the dark. The stoichiometry of phosphorylation was
assessed by incubating 5 µl of the initial phosphorylation reaction
with 5-10 µCi of [-32P]ATP under the exact
conditions as described above. The incubation was stopped with SDS
sample buffer, and the sample was electrophoresed on a 10%
SDS-polyacrylamide gel. The gel was stained with Coomassie Blue,
destained, dried, and autoradiographed, and pmol of phosphate incorporated was determined by excising and counting the rhodopsin band. The stoichiometry of phosphorylation was calculated to be ~2.4
mol of phosphate/mol of rhodopsin.
The rhodopsin binding assay was carried out essentially as described (35). Briefly, in vitro synthesized radiolabeled arrestins (~0.5 nM) were incubated with ROS membranes (200 nM phosphorylated rhodopsin) for 5 min at 37 °C in total volume of 50 µl in 50 mM Tris-HCl, pH 7.5, 0.5 mM MgCl2, 100 mM potassium acetate, 1.5 mM dithiothreitol. Samples were then cooled on ice and receptor-bound arrestins separated from unbound arrestins by Sepharose 2B column chromatography. Specific binding was calculated after subtracting out nonspecific counts in the absence of ROS membranes (~1.8 fmol).
Internalization Experiments in COS-1 CellsThe bovine
arrestin3 expression construct in pBC12BI was described previously
(29); the arrestin3-LIFA mutant (Leu-373, Ile-374, and Phe-376 mutated
to Ala) was made by replacing an ~600-bp
EcoRV-XhoI restriction fragment of pBC-arrestin3
with the corresponding fragment of arrestin3-LIFA. The human
2-adrenergic receptor expression construct in pBC12BI
was generously provided by Dr. B. Kobilka. COS-1 cells were grown in
T75 flasks at 37 °C in a humidified atmosphere containing 5%
CO2 in Dulbecco's modified Eagle's medium supplemented
with 10% fetal calf serum, 100 units/ml penicillin, and 100 µg/ml
streptomycin. The cells were grown to ~90% confluence and then
transfected with the constructs using LipofectAMINE according to the
manufacturer's instructions (Life Technologies, Inc.). Briefly,
12-13.5 µg of pBC-
2-adrenergic receptor and 3 µg of
pBC-arrestin3 or pBC-arrestin3-LIFA was incubated with 65 µl of
LipofectAMINE in 5 ml of Dulbecco's modified Eagle's medium for 30 min at 22 °C. Five ml of Dulbecco's modified Eagle's medium was
then combined with the DNA/LipofectAMINE and added to a T75 flask of
Dulbecco's modified Eagle's medium-rinsed COS-1 cells. After 5 h
at 37 °C the DNA-containing medium was replaced with complete
medium.
Cells were harvested by trypsinization 48-72 h after transfection,
washed twice with ice-cold PBS and resuspended in 1.2 ml of ice-cold
PBS containing 0.1 mM ascorbic acid. The cells (0.5 ml)
were incubated with or without 1 µM ()isoproterenol at
37 °C for 15 min, washed twice with 50 ml of ice-cold PBS, and
resuspended in 0.4-0.5 ml of cold PBS. Cell surface receptors were
measured directly by incubation with 6-9 nM
[3H]CGP-12177 (Amersham) for 3 h at 14 °C with
nonspecific binding assessed in the presence of 10 µM
(
)alprenolol. Bound ligands were separated on glass fiber filters
(Whatman, GF/C) by vacuum filtration on a Brandel cell harvester. In
these studies,
2-adrenergic receptor expression levels
(mean ± S.E. in pmol/mg of protein) were 4.7 ± 0.7 for
control, 4.9 ± 0.4 for arrestin3-, and 4.8 ± 0.6 for
arrestin3-LIFA-coexpressing cells.
To assess arrestin3 expression, 0.1-0.2 ml of the trypsinized washed cells was pelleted by centrifugation, resuspended in 0.2 ml of cold lysis buffer (20 mM Hepes, pH 7.2, 10 mM EDTA, 0.2 M NaCl, 1% Triton X-100, 0.1 mg/ml benzamidine, 1 mM phenylmethylsulfonyl fluoride), and then lysed by freeze/thaw and vortexing. The sample was centrifuged to remove insoluble material, and 20 µg of the supernatant protein was electrophoresed on a 10% SDS-polyacrylamide gel and then transferred to a nitrocellulose membrane. Immunoblotting was carried out exactly as described above for GST-arrestin3 fusion proteins except the primary monoclonal antibody was F4C1 (36) at a 1:2,000 dilution, and the secondary antibody was goat anti-mouse antibody. Protein concentrations were determined by Bio-Rad assay using bovine serum albumin as a standard.
We demonstrated previously that the nonvisual arrestins
-arrestin and arrestin3, but not visual arrestin, interact
specifically with clathrin cages (29), an assembled form of clathrin
resembling that found in coated pits. Thus, our initial approach to
localize the clathrin binding domain in nonvisual arrestins entailed
assessing the ability of
-arrestin/arrestin chimeras to interact
with clathrin cages. A chimera consisting of the
NH2-terminal half of arrestin and the COOH-terminal half of
-arrestin bound to clathrin cages comparably to wild-type
-arrestin, whereas a chimera consisting of the
NH2-terminal half of
-arrestin and the COOH-terminal
half of arrestin did not bind (data not shown). Moreover, a chimera consisting of the first 345 residues of arrestin with the last 78 amino
acids of
-arrestin retained significant clathrin binding, whereas a
chimera consisting of predominantly
-arrestin with the
NH2-terminal 47 and COOH-terminal 59 amino acids of
arrestin bound weakly to clathrin cages (29). These results implicate the COOH-terminal half of
-arrestin, and more specifically the far
COOH terminus, as being critical for its high affinity interaction with
clathrin.
We chose to do a more detailed characterization of the clathrin binding
domain with arrestin3 since it has an ~6-fold greater affinity for
clathrin than -arrestin (29). We initially sought to determine
whether the clathrin binding domain in arrestin3 is also localized to
the COOH-terminal domain by generating two truncation mutants, one
lacking the last 34 amino acids and the other lacking the entire
COOH-terminal half (Fig. 1A). Wild-type arrestin3 and the two truncation mutants were in vitro
translated with rabbit reticulocyte lysate in the presence of
[3H]leucine and then assessed for interaction with
clathrin cages. While truncation of the last 34 amino acids partially
reduced clathrin binding (~40%), truncation of the entire
COOH-terminal half of arrestin3 essentially eliminated clathrin binding
(~90% reduced) (Fig. 1A). These results thus demonstrate
that the clathrin binding domain in arrestin3, as described for
-arrestin, is contained within the COOH-terminal half of the protein
and that at least a portion of this domain is localized within the last
34 amino acids of arrestin3.
We further localized the clathrin binding domain of arrestin3 by
assessing the ability of GST-fusion proteins containing various portions of the arrestin3 COOH terminus (schematically illustrated in
Fig. 1B) to interact with clathrin cages (Fig.
1B). A GST-arrestin3 fusion protein containing the entire
COOH-terminal half of arrestin3 (residues 182-409) bound to clathrin
cages with an affinity comparable to that for recombinant arrestin3
(29), further confirming that the clathrin binding domain of arrestin3
is localized to its COOH terminus. In contrast, GST alone or a
GST-arrestin3 fusion protein containing the NH2-terminal
half of arrestin3 (residues 1-210) did not bind clathrin (data not
shown). Although truncation of the 182-409 construct to residue 385 did not affect clathrin binding, truncation to residues 332 or 366 essentially eliminated clathrin binding. These results suggest that the
predominant clathrin binding domain in arrestin3 is localized between
residues 367 and 385. Indeed, GST-fusion proteins containing the far
COOH terminus of arrestin3 (residues 318-409 or 340-409) also bound
well to clathrin cages (Fig. 1B and data not shown). A
similar pattern of clathrin binding was observed when GST-arrestin3
fusion proteins were analyzed for binding to free clathrin trimers, as
the 182-409 and 318-409 constructs bound clathrin trimers, whereas
the 182-332 and 182-366 constructs did not (data not shown).
Moreover, the GST-arrestin3 fusion proteins that did bind clathrin
cages demonstrated rapid kinetics (t1/2 1 min) (data not shown), similar to the kinetics of clathrin binding
observed for purified recombinant arrestin3 (29).
A comparison of the far COOH-terminal domains of the four mammalian
arrestins is depicted in Fig. 2A. The
shaded portion represents the putative clathrin binding
domain between residues 368 and 385 of arrestin3. Interestingly,
-arrestin and arrestin3 have greater than 80% homology in this
region. In contrast, visual arrestin lacks a significant portion of
this region and has less than 30% homology in the remaining residues,
whereas cone arrestin lacks greater than 70% of this region. This
likely explains the poor ability of visual arrestin to interact with
clathrin (29). Furthermore, it is tempting to speculate that cone
arrestin, similar to arrestin, will also not interact well with
clathrin. The observation that
-arrestin and arrestin3 contain a
putative clathrin binding domain but the visual arrestins do not
suggests that during evolution either the nonvisual arrestins gained
the ability to interact with clathrin, a function critical for their
ability to promote internalization of GPRs (27-29), or that the visual
arrestins lost this ability. The poor ability of arrestin to interact
with clathrin is interesting in light of the apparent lack of
internalization of rhodopsin as a mechanism of its regulation, a
mechanism that is likely critical for the regulation of signaling of
many other GPRs (17, 37-41).
Our initial approach toward identifying specific amino acid residues within region 368-385 which are critical for clathrin interaction involved alanine scanning mutagenesis in which three consecutive residues were mutated to alanine. The six alanine scanning mutants of arrestin3 generated are depicted in Fig. 2A (shaded boxes) and are named according to the three residues that were mutated. The mutants were in vitro translated with rabbit reticulocyte lysate in the presence of [3H]leucine and then assessed for binding to clathrin cages (Fig. 2B). Although the alanine scanning mutants PVD, YAT, and DDD were modestly reduced in clathrin binding (15-40%), the mutants TNL, IEF, and ETN were ~70-85% reduced in clathrin binding. These results thus more precisely localize the clathrin binding site in arrestin3 to residues 371-379. To confirm a native tertiary structure of these alanine scanning mutants, their ability to bind to phosphorylated agonist-activated receptor was also analyzed. Since arrestin3 recognizes the phosphorylated and activated state of the photoreceptor rhodopsin with high affinity (42), we used phosphorylated light-activated rhodopsin to assess receptor binding. Previous studies have demonstrated that the primary sites in arrestins for interaction with the phosphorylated agonist-activated receptor are localized to the NH2-terminal half of the molecule (42-44). Thus, we did not expect any of the alanine scanning mutants to be defective in receptor binding in the absence of alterations in overall protein conformation. Indeed, each of these mutants of arrestin3 bound to phosphorylated light-activated rhodopsin comparably to wild-type arrestin3, although the binding of IEF and DDD was reduced modestly (Fig. 2C).
To identify specific residues critical for arrestin3 interaction with
clathrin, we performed site-directed mutagenesis within the three
consecutive alanine scanning triplets TNL, IEF, and ETN. To assess the
contribution of charged residues to the arrestin3/clathrin interaction,
we mutated Glu-375 and Glu-377 to Lys, individually as well as in
combination. To assess the contribution of hydrophobic amino acids to
this interaction, we mutated Leu-373, Ile-374, and Phe-376 to Ala.
Finally, to assess the potential contribution of hydrogen bonding, we
generated mutants in which Thr-371, Thr-378, Asn-372, and Asn-379
individually were changed to alanine. These site-directed arrestin3
mutants were in vitro translated in the presence of
[3H]leucine and then assessed for clathrin cage binding
(Fig. 3A). Whereas mutation of the individual
Glu residues to Lys only modestly reduced clathrin binding
(~20-30%), mutation of both Glu residues to Lys significantly
reduced binding (~60% reduction). Similarly, mutation of individual
Thr or Asn residues only modestly reduced clathrin binding
(~10-30%). In contrast, an arrestin3 mutant containing Leu, Ile,
and Phe mutated to Ala decreased clathrin binding by ~90%. As
expected, none of the site-directed arrestin3 mutants was significantly
affected in its ability to interact with phosphorylated light-activated
rhodopsin, although the N379A mutant was reduced ~40% (Fig.
3B). These results thus implicate the hydrophobic residues Leu-373, Ile-374, and Phe-376 and the charged residues Glu-375 and
Glu-377 as playing a primary role in high affinity interaction of
arrestin3 with clathrin.
The results obtained in this study correlate well with results involving characterization of the arrestin binding domain in clathrin (46). In the latter study, a small region containing highly conserved residues in the clathrin heavy chain NH2-terminal domain was found to be critical for clathrin interaction with nonvisual arrestins. Clathrin residues most important for arrestin binding included two highly conserved lysines and an invariant glutamine. Thus, it is likely that ionic and hydrophobic and/or hydrogen bonding interactions contribute to high affinity binding between nonvisual arrestins and clathrin.
Since previous studies have implicated arrestin/clathrin interaction in
receptor internalization (29), we next sought to determine whether an
arrestin3 mutant defective in clathrin binding would also be defective
in its ability to promote internalization of the
2-adrenergic receptor. The arrestin3 mutant containing Leu-373, Ile-374, and Phe-376 mutated to Ala (arrestin3-LIFA) was used
for these studies since it was ~90% defective in clathrin binding
(Fig. 3A) but unaffected in its ability to interact with receptor (Fig. 3B). COS-1 cells transfected with
2-adrenergic receptor alone or together with wild-type
arrestin3 or arrestin3-LIFA were treated with agonist for 15 min and
then analyzed for cell surface receptors. When both arrestins were
expressed at a comparable level (Fig. 4A),
wild-type arrestin3 enhanced agonist-specific
2-adrenergic receptor sequestration ~4-fold, whereas
arrestin3-LIFA-promoted
2-adrenergic receptor
sequestration was ~65% reduced compared with the wild-type protein
(Fig. 4B). These results thus confirm the critical
requirement of clathrin binding for the ability of arrestin3 to promote
internalization of agonistactivated GPRs.
In this study, we have localized the predominant clathrin binding domain of arrestin3 to a short stretch of residues in the far COOH terminus of the molecule (residues 373-377). These findings should prove useful for designing dominant-negative arrestins to block specifically the arrestin/clathrin interaction in an effort to modulate the regulation of GPR signaling in cells. Since internalization of GPRs has been implicated in desensitization (37, 38, 45), and more recently, resensitization (20-23) of GPRs, the opportunity exists to prolong or attenuate receptor responsiveness to agonist using such arrestin mutants.
We thank Dr. V. Gurevich for providing the
cDNAs for the -arrestin/arrestin chimeras, Dr. R. Sterne-Marr
for providing the GST-arrestin3 (318-409) construct, Drs. J. Pitcher
and R. Lefkowitz for supplying purified recombinant rhodopsin kinase,
L. Tran for purification of clathrin, and C. Carman and Drs. L. Kallal
and A. Gagnon for helpful discussions.