Rab5 Association with the Angiotensin II Type 1A Receptor Promotes Rab5 GTP Binding and Vesicular Fusion*

Previous studies have demonstrated that the internalization of the angiotensin II type 1A receptor (AT 1A R) may be mediated by both (cid:1) -arrestin-sensitive and -in-sensitive mechanisms. Therefore, we have used the AT 1A R carboxyl-terminal tail to screen a rat brain yeast two-hybrid expression library for novel AT 1A R-interact-ing proteins that might contribute to the regulation of AT 1A R internalization. We have identified Rab5a as an AT 1A R-binding protein that selectively associates with the AT 1A R and not with the (cid:1) 2 -adrenergic receptor. A Rab5a-S34N mutant defective in GTP binding does not prevent the internalization of the AT 1A R but does pre- vent the trafficking of the AT 1A R into larger hollow cored vesicular structures. Agonist activation of the AT 1A R promotes both the formation of Rab5a (cid:1) AT 1A R pro- tein complexes and Rab5a GTP binding. Rab5a interactions with the AT 1A R are mediated in part by the last 10 amino acid residues of the AT 1A R carboxyl-terminal tail, and although a mutant receptor lacking these residues internalizes normally, it does not redistribute into larger hollow vesicles. Our

The angiotensin II type 1A receptor (AT 1A R) 1 is a member of the large superfamily of G protein-coupled receptors (GPCRs). AT 1A Rs are coupled via G q to the stimulation of phospholipase C␤ leading to increases in intracellular inositol 1,4,5-triphosphate formation, the release of calcium from intracellular stores, and the activation of protein kinase C (1). Agonist activation also initiates the feedback phosphorylation and desensitization of the AT 1A R in response to phosphorylation by both second messenger-dependent protein kinases and G proteincoupled receptor kinases (2,3). G protein-coupled receptor kinase phosphorylation promotes the membrane translocation and binding of ␤-arrestins to the AT 1A R, which serves both to uncouple the receptor from heterotrimeric G proteins and to target the receptor for endocytosis (3)(4)(5)(6)(7). ␤-Arrestins act as intermediary GPCR endocytic adaptor proteins through their association with clathrin and the ␤2-adaptin subunit of the heterotetrameric AP2 adaptor complex (8,9). However, ␤-arrestin-directed internalization of GPCRs in clathrin-coated vesicles may represent only one of many mechanisms contributing to GPCR endocytosis (4,10,11). For example, AT 1A R internalization may also involve a pathway that is relatively insensitive to dominant-negative ␤-arrestin mutants (4).
The AT 1A R is a member of a class of GPCRs that remain associated with ␤-arrestins during clathrin-mediated endocytosis. Moreover, the AT 1A R is targeted to enlarged hollow core vesicular structures (3,12) but is neither dephosphorylated nor recycled efficiently back to the cell surface (3). In contrast, endocytosis is required for the dephosphorylation and resensitization of GPCRs like the ␤ 2 -adrenergic receptor (␤ 2 AR) but does not appear to contribute to the resensitization of AT 1A R responses (3,(13)(14)(15). Although ␤ 2 AR internalization is exquisitely ␤-arrestin-dependent, ␤-arrestin is excluded from ␤ 2 ARbearing endocytic vesicles allowing dephosphorylation in endosomes (3,5,16). The apparent difference in the ability of the ␤ 2 AR and AT 1A R to internalize in a complex with ␤-arrestin appears to be regulated by their carboxyl-terminal tails (3). These observations have led to the suggestion that ␤-arrestins may also regulate the reestablishment of GPCR responsiveness (16). However, although the determinants regulating the stable high affinity association of ␤-arrestin with GPCRs like the AT 1A R have been characterized (12), little is known about the mechanism(s) regulating the intracellular trafficking and retention of these receptors in endosomes. To identify additional molecular components contributing to the regulation of both AT 1A R internalization and intracellular trafficking, we used the AT 1A R carboxyl-terminal tail as bait in the yeast twohybrid system to screen for novel AT 1A R-interacting proteins.
Here we describe that the interaction of Rab5a with the carboxyl-terminal tail of the AT 1A R not only promotes the vesicular sorting of the AT 1A R into enlarged vesicular structures but also activates the GDP/GTP cycle of Rab5a.

EXPERIMENTAL PROCEDURES
DNA Construction-The truncated AT 1A R mutants were constructed by polymerase chain reaction. A 5Ј-oligonucleotide primer hybridized upstream from a unique EcoRI site within the coding sequence of the AT 1A R-encoding gene, and 3Ј-oligonucleotides introduced a stop codon after residues 319, 329, 339, and 349 in the AT 1A R carboxyl-terminal tail. GST-Rab5 fusion proteins were constructed by subcloning Rab5a, Rab5a-S34N, and Rab5a-Q79L into the eukaryotic expression vector pEBG4. The V2 vasopressin carboxyl-terminal tail (residues 320 -371) and the AT 1A R carboxyl-terminal tail constructs (297-359, 310 -359, and 297-249) were subcloned by PCR in to the yeast vector pAS2-1. Wild-type canine Rab5a was subcloned into the yeast vector pACT2. Sequence integrity of the DNA constructs was confirmed by DNA sequencing. All other cDNA constructs used have been reported previously (3,5,17).
Cell Culture and Transfection-HEK 293 cells and COS7 cells were grown in Eagle's minimal essential medium with Earle's salt (MEM) and Dulbecco's minimal Eagle's medium (DMEM), respectively (Invitrogen) supplemented with 10% (v/v) heat-inactivated fetal bovine serum (Hyclone) and gentamicin (100 g/ml). The cells were seeded at a density of 2.5 ϫ 10 6 cells/100-mm dish (Falcon) and were transiently transfected by a modified calcium phosphate method (21) with the cDNAs described in the figure legends. COS7 cells were transiently transfected using LipofectAMINE reagent (Invitrogen) with the cDNAs described in the figure legends. Following transfection (18 h), the cells were pooled and reseeded into 35-mm glass-bottomed culture dishes (MatTek) for confocal studies, into six well-dishes (Falcon) for receptor internalization and GTP␥S loading studies, into 12-well dishes (Falcon) for receptor radioligand binding assays and 100-mm dishes for coimmunoprecipitation experiments.
Co-immunoprecipitation Experiments-COS7 cells transiently cotransfected to overexpress FLAG-AT 1A R or FLAG-␤ 2 AR plus GSTtagged Rab5a, GST-Rab5a-S34N, or GST-Rab5a-Q79L were incubated for 20 min at 37°C in the presence and absence of 100 nM Ang II in serum-free DMEM containing 2.5 mM dithiobis(succinimidylpropionate) cross-linker (Pierce) that had been dissolved in Me 2 SO (14). The cells were then washed twice with ice-cold phosphate-buffered saline and were lysed in cold lysis buffer containing 50 mM Tris, pH 8.0, 150 mM NaCl, 0.1% Triton X-100 containing protease inhibitors (20 g/ml leupeptin, 20 g/ml aprotinin, and 1 mM phenylmethylsulfonyl fluo-ride). The mixture was gently agitated for 15 min at 4°C and thereafter centrifuged at 13,000 ϫ g for 20 min. Cleared supernatants (500 -600 g of protein) were incubated with 20 l of FLAG M2-agarose affinity beads (Sigma) for 16 h at 4°C. The beads were then washed twice with lysis buffer and twice with phosphate-buffered saline and were solubilized in 3ϫ SDS sample buffer, resolved by 10% SDS-PAGE, and transferred to nitrocellulose membrane. GST-Rab5a was detected via Western blotting with an anti-GST rabbit polyclonal antibody (1:500) followed by horseradish peroxidase-conjugated anti-rabbit IgG secondary antibody (1:2500) and visualized via chemiluminescence using the ECL kit from Amersham Biosciences, Inc.
GTP␥S Loading Assay-COS7 cells transiently co-transfected to overexpress FLAG-AT 1A R, FLAG-␤ 2 AR, or empty plasmid (control) along with GST-Rab5a were labeled with 5 M [ 35 S]GTP␥S (10 Ci/ml) for 30 min at 37°C in digitonin buffer (1% digitonin, 20 mM Hepes, pH 7.5, 100 mM NaCl, 5 mM MgCl 2 , 1 mM dithiothreitol) (0.75 ml/well) and then incubated for 15 min at 37°C with or without 100 nM Ang II or 10 M isoproterenol in serum-free DMEM. The cells were then washed twice with ice-cold phosphate-buffered saline and were lysed in cold lysis buffer containing 50 mM Tris, pH 8.0, 150 mM NaCl, 1 M GT␥S, 0.1% Triton X-100 plus protease inhibitors (20 g/ml leupeptin, 20 g/ml aprotinin and 1 mM phenylmethylsulfonyl fluoride). The mixture was gently agitated for 15 min at 4°C and thereafter centrifuged at Rab5 Association with the Angiotensin II Type 1A Receptor 13,000 ϫ g for 20 min. Cleared supernatants were incubated with 40 l of glutathione 4B-Sepharose beads (Amersham Biosciences, Inc.) for 3 h at 4°C. The samples were counted for 35 S activity (cpm) in a ␤-scintillation counter. Aliquots (20 l) of each cleared supernatant were analyzed for relative expression of GST-Rab5a by immunoblotting using an anti-GST antibody.
Receptor Internalization Assays-Radioligand binding measurements of AT 1A R internalization were assessed as described previously (20). Briefly, HEK 293 cells transiently transfected to overexpress FLAG-AT 1A R in the absence and presence of Rab5a constructs were washed with serum-free MEM with 10 mM Hepes, pH 7.4, and then were incubated at 37°C in the same medium containing 100 nM Ang II for 1 h. The cells were then washed in cold MEM containing 10 mM Hepes and were incubated in acid wash solution (90 mM NaCl, 50 mM sodium citrate, pH 5.0) for 20 min to dissociate bound ligand. The cells were then washed and incubated with 250 l [ 125 I-Sar1-Ile 8 ] Ang II for 3-4 h at 14°C. Nonspecific binding values were obtained in the presence of 10 M Losartan. The cells were washed on ice and solubilized in 0.1 N NaOH, and radioactivity was counted using ␤-scintillation. Receptor expression ranged from 250 to 500 fmol/mg protein.
Confocal Immunofluorescence Microscopy-Confocal microscopy was performed on a Zeiss LSM-510 laser scanning microscope using a Zeiss 63ϫ 1.3 NA oil immersion lens. Live cell imaging of HEK 293 cells expressing FLAG-AT 1A R and Rab5a, Rab5a-S34N, or Rab5a-Q79L and ␤-arrestin2-GFP was performed using cells plated on 35-mm glassbottomed culture dishes. The cells were kept warm at 37°C in serumfree MEM on a heated microscope stage as described previously (3,17). FLAG-AT 1A R staining of HEK 293 cells grown on glass coverslips and fixed with 3% paraformaldehyde in Hanks' balanced salt solution with 0.1% Triton X-100 for 20 min was performed using monoclonal anti-FLAG (M2) antibody (1:250 dilution) in conjunction with a rhodamine red-conjugated goat anti-mouse secondary antibody (1:500 dilution) (Molecular Probes). Co-localization studies of GFP-Rab5a and rhodamine-labeled FLAG-AT 1A R fluorescence were performed using dual excitation (488, 568 nm) and emission (515-540 nm, GFP; 590 -610 nm, rhodamine) filter sets. The specificity of labeling and the absence of signal cross-over were established by examination of single-labeled samples.
Data Analysis-The means Ϯ S.E. are shown for the values obtained for the number of independent experiments indicated in the figure legends. The data were analyzed for statistical significance using GraphPad Prism software. Statistical significance was determined by an unpaired two-tailed t test.

RESULTS AND DISCUSSION
Identification of Rab5 as an AT 1A R-interacting Protein-To identify novel proteins that interact with the AT 1A R, we screened ϳ20 ϫ 10 6 independent clones of a rat brain cDNA library using the AT 1A R carboxyl-terminal tail (Ct), amino acid residues 297-359, as bait in the yeast two-hybrid system. Six of the clones isolated for growth on ϪLeu/ϪTrp/ϪAde plates were positive for both partial and full-length Rab5a sequences. Rab5a is a member of a large family of Rab GTP-binding proteins (ϳ60 members) that regulate the trafficking of vesic- ular cargo between intracellular compartments (21). In particular, Rab5a regulates the formation, trafficking, and fusion of clathrin-coated vesicles with early endosomes (22)(23)(24).
Specificity of Rab5a/AT 1A R Interactions-Growth of a canine Rab5a clone on ϪLeu/ϪTrp/ϪAde plates was dependent upon the AT 1A R Ct GAL4 BD fusion protein and was not observed in the presence of either empty vector or a GAL4 BD fusion protein of the tail of the vasopressin type 2 receptor (V2R) ( Table I). To determine whether Rab5a physically interacts with the AT 1A R, GST-Rab5a and FLAG-AT 1A R were both expressed in COS7 cells and the co-immunoprecipitation of these proteins was assessed. We found that GST-Rab5a and not GST alone was co-immunoprecipitated from COS7 cells transfected with FLAG-AT 1A R (Fig. 1A). Furthermore, co-immunoprecipitation of Rab5a was specific to the AT 1A R, because Rab5a was not co-immunoprecipitated with FLAG-␤ 2 AR (Fig.  1B). Taken together, these observations indicate that Rab5a represents a novel AT 1A R-interacting protein.
Effect of Rab5a Mutants on AT 1A R Internalization and Vesicular Distribution-Rab5a has previously been shown to regulate the endocytosis and intracellular trafficking of the ␤ 2 AR, D2 dopamine receptor, and neurokinin 1 receptor (17,25,26). Therefore, we sought to examine whether Rab5a was involved in the endocytosis and vesicular trafficking of the AT 1A R. We find that even though a Rab5a-S34N mutant defective in GTP binding blocks the internalization of several other GPCRs (17,25,26), co-expression of either wild-type Rab5a, Rab5a-S34N, or constitutively active Rab5a-Q79L has no apparent effect on the agonist-stimulated (100 nM angiotensin II) internalization of FLAG-AT 1A R (Fig. 2). In the absence of agonist stimulation, GFP-Rab5a is neither co-localized with FLAG-AT 1A R at the cell surface nor localized to large hollow vesicular structures (Fig.  3A). Rather, GFP-Rab5a is limited to small endocytic vesicles (Fig. 3A). However, upon agonist stimulation of the AT 1A R we observed enlarged GFP-Rab5a-positive vesicular structures and that the AT 1A R is completely co-localized with GFP-Rab5a within these enlarged hollow core vesicular structures (Fig.  3B). These observations suggest that the proposed association of Rab5a with the AT 1A R may contribute to the regulation of AT 1A R trafficking and/or vesicular fusion rather than AT 1A R endocytosis.
We have previously demonstrated that ␤-arrestin2-GFP is internalized in a physical complex with the AT 1A R making this a useful tool for following the temporal dynamics of AT 1A R internalization and trafficking in live cells (3,12). In the absence of agonist stimulation, ␤-arrestin2-GFP is diffusely distributed throughout the cytosol of FLAG-AT 1A R-expressing HEK 293 cells (Fig. 4A). In response to agonist stimulation, we observe that AT 1A R⅐␤-arrestin2-GFP complexes first redistribute to clathrin-coated pits (Fig. 4A, 5 min) and are then localized in small endocytic vesicles (Fig. 4A, 10 min), which subsequently fuse into the enlarged hollow core vesicular structures (Fig. 4A, 15 min) in which the receptor was previously demonstrated to co-localize with GFP-Rab5a (Fig. 3B). To determine whether Rab5a may regulate the intracellular trafficking and fusion of AT 1A R-containing endocytic vesicles, we utilized ␤-arrestin2-GFP to follow the subcellular localization of the AT 1A R in the presence of wild-type Rab5a, dominantnegative Rab5a-S34N, or constitutively active Rab5a-Q79L. In either the absence (Fig. 4A, 15 min) or presence (Fig. 4B) of wild-type Rab5a, ␤-arrestin2-GFP-labeled FLAG-AT 1A Rs are localized to large hollow core endocytic vesicular structures. In contrast, the expression of Rab5a-S34N prevents the agoniststimulated redistribution of ␤-arrestin2-GFP labeled FLAG-AT 1A Rs to enlarged hollow core vesicular structures (Fig. 4C). Instead, ␤-arrestin2-GFP fluorescence remains limited to smaller endocytic vesicles (Fig. 4C). The expression of Rab5a-Q79L did not alter the formation of enlarged vesicles contain-

Rab5 Association with the Angiotensin II Type 1A Receptor
ing ␤-arrestin2-GFP-labeled FLAG-AT 1A Rs (Fig. 4D). Taken together, these data suggest that the GTP-bound form of Rab5a is required for fusion of AT 1A R containing endocytic vesicles into enlarged hollow cored endocytic structures.
Functional Consequence of the Agonist-stimulated Formation of AT 1A R⅐Rab5 Complexes-We find that Rab5a forms a physical complex with the AT 1A R in the absence of agonist and also regulates the agonist-promoted intracellular trafficking and fusion of AT 1A R-bearing endocytic vesicles. Therefore, we examined whether AT 1A R activation stimulates both GST-Rab5a association with the receptor and GTP␥S loading of GST-Rab5a in COS7 cells. We observe that agonist stimulation increases the relative amount of GST-Rab5a, GST-Rab5a-S34N, and GST-Rab5a-Q79L co-immunoprecipitated with the FLAG-AT 1A R (Fig. 5A). Consistently more Rab5a-S34N than Rab5a and Rab5a-Q79L is co-immunoprecipitated with the FLAG-AT 1A R under basal conditions (Fig. 5, A and B), which may reflect preferential interaction of the receptor with the GDPbound form of Rab5a. Angiotensin II stimulation of the AT 1A R increases GTP␥S binding to GST-Rab5a by 6.6 Ϯ 1.7-fold above agonist-stimulated GST-Rab5a expressing control cells lacking AT 1A R (Fig. 4C). Basal GTP␥S binding to GST-Rab5a in AT 1A R-expressing cells is also increased 3.2 Ϯ 1.6-fold above GST-Rab5a-expressing control cells lacking AT 1A R (Fig. 4C). The increased basal GST-Rab5a GTP␥S binding likely reflects the observed agonist-independent association of the GTPase with the AT 1A R. Nonetheless, agonist activation of the AT 1A R increases GTP␥S binding to GST-Rab5a 2.4 Ϯ 0.9-fold above basal. In contrast, ␤ 2 AR expression and activation has no effect on GTP␥S binding to GST-Rab5a (Fig. 4C). These data not only demonstrate that the association of Rab5a with the AT 1A R is regulated by agonist activation of the receptor but also suggest that the AT 1A R may function as a guanine nucleotide exchange factor for Rab5a, thereby serving a function that is analogous to its role as a heterotrimeric G protein-coupled receptor.
Characterization of Rab5a Interactions with the AT 1A R Tail and AT 1A R Internalization-To identify the regions of the AT 1A R carboxyl-terminal tail contributing to Rab5a binding, we used GAL4 BD fusion proteins of different regions of the AT 1A R Ct in the yeast two-hybrid system with Rab5a fused to the GAL4 AD. We find that the removal of either residues 297-310 or residues 349 -359 prevents the transactivation of yeast reporter genes in a ␤-galactosidase activity assay (Fig.  6A). This observation suggests that the association of Rab5a with the AT 1A R requires Rab5a binding to multiple regions of the AT 1A R Ct. Furthermore, amino acid residues 349 -359 do not contribute to AT 1A R endocytosis, because the truncation of the AT 1A R by 10 and 20 amino acid residues had no effect on the agonist-stimulated internalization of the AT 1A R (Fig. 6B). The internalization of the AT 1A R was not altered until 30 amino acids were deleted from the carboxyl-terminal tail, AT 1A R-(1-329) (Fig. 6B). Therefore, the regions critical for Rab5a binding to the AT 1A R Ct are distinct from those residues (residues 327-332) required for stable ␤-arrestin binding and AT 1A R endocytosis (16).
Effect of AT 1A R Carboxyl-terminal Truncations on AT 1A R Trafficking-We observed that the removal of either residues 297-310 or residues 349 -359 from the AT 1A R Ct prevented the interaction of Rab5a with the AT 1A R Ct. However, because the removal of residues 297-310 would result in a nonfunctional receptor, we tested whether the deletion of the last 10 amino acid residues from the AT 1A R carboxyl-terminal tail might alter the agonist-stimulated redistribution of the mutant AT 1A R into enlarged hollow core vesicular structures. Although the truncation of the last 10 amino acid residues from the AT 1A R carboxyl-terminal tail has no effect on the internaliza-tion of AT 1A R⅐␤-arrestin2-GFP complexes (Fig. 7A), ␤-arres-tin2-GFP labeled AT 1A R-(1-349) is no longer observed to redistribute into large hollow core vesicles (Fig. 7A). Rather, the ␤-arrestin bound AT 1A R remains localized in small endocytic vesicular structures (Fig. 7A). Moreover, similar to what was previously observed for the ␤ 2 AR (17), the AT 1A R (1-349) mutant remains localized to small endocytic vesicles and no longer completely co-localizes with GFP-Rab5a (Fig. 7B). Consequently, the redistribution of the AT 1A R into enlarged endocytic structures appears to be dependent both of Rab5a activity and the last 10 amino acids of the AT 1A R carboxyl-terminal tail that contributes to AT 1A R/Rab5a interactions.
Taken together, our data indicate that Rab5a is an AT 1A R binding protein and that the association of Rab5a with the AT 1A R appears to regulate the intracellular trafficking and fusion of AT 1A R-bearing endocytic vesicles into enlarged hollow core vesicular structures. The association between the AT 1A R and Rab5a appears to be interactive, in that activation of the AT 1A R increases Rab5a GTP binding and activated Rab5a is required for homotypic fusion of AT 1A R-bearing endocytic vesicles into enlarged hollow core vesicular structures. Rab5a binding to the AT 1A R requires both the distal and proximal ends of the AT 1A R Ct. Because the proximal end of the AT 1A R Ct is implicated in heterotrimeric G protein coupling (27), it stands to reason that its carboxyl-terminal tail may allow the AT 1A R to act as a guanine nucleotide exchange factor for Rab5a. As a consequence, acti- vated AT 1A R exhibits the capacity to promote the Rab5a-dependent homotypic fusion of endocytic vesicles.
The trafficking of proteins between intracellular organelles is a highly regulated process involving several membrane budding and fusion events between donor and acceptor membranes (22)(23)(24)28). Rab GTPases are important regulators of intracellular vesicular transport and endosomal fusion (22)(23)(24)28). Recent studies have indicated that Rab5 contributes to the formation, trafficking, and fusion of clathrin-coated vesicles with early endosomes (29). The expression of the dominantnegative Rab5a-S34N mutant blocks the internalization and trafficking of ␤ 2 AR-bearing endocytic vesicles (17), but Rab5a does not associate with the ␤ 2 AR. In contrast, although Rab5 binds the AT 1A R, similar to what is observed for ␤-arrestin dominant-negative, the internalization of the AT 1A R appears to also be insensitive to inhibition by a dominant-negative Rab5a-S34N mutant. The reason why the internalization of the AT 1A R is insensitive to Rab5a-S34N remains unknown. It is possible that the ␤ 2 AR and AT 1A R are internalized by distinct endocytic mechanisms and/or vesicular populations or that the AT 1A R recruits additional proteins involved in clathrin-mediated endocytosis.
Multiple effector proteins regulate the distribution of Rab5 between membrane compartments and the cytosol. For example, the Rab GDP dissociation inhibitor regulates Rab5 membrane localization where the exchange of GDP for GTP is regulated by specific guanine nucleotide exchange factors (30,31). We propose that the AT 1A R also functions as a novel Rab5a guanine nucleotide exchange factor linking AT 1A R signaling and intracellular trafficking. In particular, AT 1A R activation promotes Rab5a GTP binding, which is correlated with enhanced vesicular fusion. Agonist activation of the epidermal growth factor receptor (EGFR) also results in the recruitment of intracellular proteins involved in endocytic trafficking. In particular, EGFR stimulation leads to the activation and plasma membrane translocation of Rab5a (32). However, unlike what we observe for the AT 1A R, EGFR internalization is blocked by Rab5a-S34N. Furthermore, EGFR-dependent Rab5a activation requires the stimulation of a tyrosine kinase signal transduction cascade that leads to the activation of uncharacterized intermediary Rab5a regulatory proteins (32). In contrast, the activation of Rab5a by the AT 1A R appears to involve the direct association of the receptor with the GTPase.
There is emerging evidence to suggest that Rab GTPases may be required for selecting cargo proteins for sequestration into clathrin-coated vesicles (33)(34)(35). Recently, Carroll et al. (36) demonstrated that Rab9 sequesters cargo proteins into vesicles shuttling from endosomes back to the trans-Golgi network. Specifically, this involves an intermediary Rab9 effector protein, TIP47, which acts as a molecular link between the mannose-6-phosphate receptor cargo protein and Rab9 (36,37).
These studies indicate that Rab GTPases not only function to recruit vesicular effector proteins but may also regulate their activities. In contrast, the observation that Rab5/AT 1A R interactions may stimulate Rab5-dependent homotypic fusion of early endosomes suggests that vesicular cargo proteins, such as the AT 1A R, may control their own targeting between intracellular compartments by controlling the activity of specific components of the intracellular trafficking machinery in the absence of an intermediary effector protein.
It is now appreciated that the termination of GPCR signaling via heterotrimeric G proteins does not abrogate the potential for "desensitized" GPCRs to signal via other cascades. Through their association with regulatory proteins such as ␤-arrestins, GPCRs serve as scaffolds for the assembly and compartmentalization of multi-molecular signaling protein complexes (38 -41). Understanding why and how GPCRs regulate the redistribution of these complexes into different intracellular compartments will improve our comprehension of the molecular and physiological consequences of GPCR signaling. The identification of an interdependent relationship between AT 1A R activation and Rab5a-mediated intracellular vesicular trafficking and fusion provides the first indication that GPCR subtypes may specify the compartment to which these signaling complexes are mobilized. Because there is evidence that multiple small GTP-binding proteins interact with GPCRs (42), future studies should reveal whether GPCRs both interact with and regulate additional components of the intracellular trafficking machinery.