Protein Kinase C-dependent Ubiquitination and Clathrin-mediated Endocytosis of the Cationic Amino Acid Transporter CAT-1*

Cationic amino acid transporter 1 (CAT-1) is responsible for the bulk of the uptake of cationic amino acids in most mammalian cells. Activation of protein kinase C (PKC) leads to down-regulation of the cell surface CAT-1. To examine the mechanisms of PKC-induced down-regulation of CAT-1, a functional mutant of CAT-1 (CAT-1-HA-GFP) was generated in which a hemagglutinin antigen (HA) epitope tag was introduced into the second extracellular loop and GFP was attached to the carboxyl terminus. CAT-1-HA-GFP was stably expressed in porcine aorthic endothelial and human epithelial kidney (HEK) 293 cells. Using the HA antibody internalization assay we have demonstrated that PKC-dependent endocytosis was strongly inhibited by siRNA depletion of clathrin heavy chain, indicating that CAT-1-HA-GFP internalization requires clathrin-coated pits. Internalized CAT-1-HA-GFP was accumulated in early, recycling, and late endosomes. PKC activation also resulted in ubiquitination of CAT-1. CAT-1 ubiquitination and endocytosis in phorbol ester-stimulated porcine aorthic endothelial and HEK293 cells were inhibited by siRNA knockdown of NEDD4-2 and NEDD4-1 E3 ubiquitin ligases, respectively. In contrast, ubiquitination and endocytosis of the dopamine transporter was dependent on NEDD4-2 in all cell types tested. Altogether, our data suggest that ubiquitination mediated by NEDD4-2 or NEDD4-1 leading to clathrin-mediated endocytosis is the common mode of regulation of various transporter proteins by PKC.

heteromeric amino acid transporters. There are four proven members in the CAT subfamily: CAT-1, CAT-2A, CAT-2B, and CAT-3 (reviewed in Ref. 2). All CATs have a predicted topology of 14 transmembrane domains with cytoplasmic amino and carboxyl termini. The largest, third extracellular loop is glycosylated in all CATs. CAT-2 proteins are also glycosylated in the second loop. CAT-1 is expressed ubiquitously in all adult tissues except the liver (reviewed in Ref. 3). The transport properties of CAT-1 resemble those of system y ϩ , defined by the selectivity for cationic amino acids, K m of 0.1-0.2 mM, Na ϩ and pH independence, and strong trans-stimulation.
CAT-1 is ubiquitously expressed and the main portal of entry for cationic amino acids into mammalian cells. Homozygous knockout of CAT-1 in mice is postnatally lethal (4). CAT-1 activity has been reported to be regulated through activation of protein kinase C (PKC) (5,6). Even though there are some discrepancies, most studies conclude that activation of PKC leads to inhibition of L-Arg uptake due to the decreased activity of CAT-1. Studies of CAT-1 phosphorylation and mutations of the PKC phosphorylation consensus sites in CAT-1 suggested that CAT-1 inhibition by PKC is not due to CAT-1 phosphorylation (7). Hence, PKC activity may regulate amino acid transport through CAT-1 indirectly. A role of endocytic trafficking in mediating the regulation of CAT-1 activity by PKC in Xenopus laevis and human glioblastoma cells was suggested based on the demonstration of PKC-dependent down-regulation of the cell surface CAT-1 (7).
The activity of several mammalian transporters of the SLC6 family, such as dopamine (DAT), norepinephrine, and serotonin transporters, has also been shown to be inhibited by PKC through down-regulation of the surface pool of these transporters (8 -13). Phosphorylation of the intracellular loop of norepinephrine is implicated in the PKC-dependent endocytosis of this transporter (14,15). In contrast, as for CAT-1, phosphorylation of DAT is not necessary for its accelerated endocytosis (16). However, another post-translational modification, ubiquitination, has been found to be critical for the PKC-dependent endocytosis of DAT (17).
Ubiquitination has recently emerged as a major molecular signal mediating endocytosis and post-endocytic sorting of transmembrane proteins, including many transport and channel proteins (18,19). Ubiquitination involves covalent attachment of a 76-amino acid polypeptide called ubiquitin mainly to free amino groups and is catalyzed by the sequential action of three enzymes (E1, E2, and E3) (20). The E3 ligase is the last enzyme responsible for the transfer of ubiquitin to the substrate. E3 typically determines substrate specificity of the ubiquitination reaction. Interestingly, two highly homologous E3 enzymes, NEDD4-2 (neural precursor cell expressed, developmentally down-regulated  and to a lesser extent NEDD4-1, have been implicated in ubiquitination of many mammalian transporters and channels (21)(22)(23)(24)(25)(26). These enzymes contain the catalytic HECT (homologous E6-AP carboxyl-terminal) domain. Likewise, Rsp5, the sole representative of the NEDD4 family in yeast, is responsible for ubiquitination and endocytosis of several transporters including amino acid transporters homologous to the SLC7 transporter family (27)(28)(29). NEDD4 family ligases typically recognize their substrates through the interaction of their WW domains with the PPXY motifs in the substrate (30 -34). However, many NEDD4 family substrates do not have these sequence motifs and bind NEDD4/Rsp5 indirectly (26,35). Recently it was proposed that Rsp5 binds to several transporters through ␤-arrestin-like adaptor proteins also called ␣-arrestins (29). These ␣-arrestins contain one or more "PPXY" sequences at the carboxyl termini, capable of direct binding to the WW domains of Rsp5 (36).
In the present study we developed a new assay to monitor and quantitatively analyze CAT-1 endocytosis and demonstrated that PKC activation promotes endocytosis of CAT-1. CAT-1 was also found to be ubiquitinated in a PKC-dependent manner, and this ubiquitination required NEDD4 family E3 ligases.

EXPERIMENTAL PROCEDURES
Antibodies and Chemicals-Antibodies were purchased from the following sources: monoclonal mouse antibody P4D1 to ubiquitin and polyclonal rabbit to human NEDD4-1 (H-135) from Santa Cruz Biotechnology (Santa Cruz, CA); monoclonal mouse to green fluorescent protein (GFP) from Zymed Laboratories Inc. (San Francisco, CA); mouse monoclonal antibody to hemagglutinin epitope HA11 (16B12) from Covance (Berkley, CA); donkey anti-mouse antibodies conjugated with Cy5 or Cy3 from Jackson ImmunoResearch (West Grove, PA); rabbit polyclonal antibody to NEDD4-2 was a kind gift from Dr. Oliver Staub (University of Lausanne, Lausanne, Switzerland). Monoclonal antibody to EEA.1 was from BD Transduction Laboratories (Los Angeles, CA). Rat anti-Lamp1 was from the Developmental Studies Hybridoma Bank (University of Iowa). Monoclonal antibody to clathrin heavy chain TD.1 was from the American Type Culture Collections, Inc. The antibody to human CAT-1 was described previously (6).
Plasmids-Human CAT-1 tagged with the HA epitope tag and GFP (CAT-1-HA-GFP) was generated from hCAT-1-GFP plasmid (7) by inserting the HA sequence (11 amino acids) into the second extracellular loop of the transporter by site-directed mutagenesis. To this end, the sequence AGACCCATCGGG-GAGTTCTCACGGA was replaced by HA sequence TATC-CCTACGACGTGCCCGATTACG using a QuikChange sitedirected mutagenesis kit according to the manufacturer's protocol (Stratagene Cloning Systems, La Jolla, CA). Cyan fluorescent protein (CFP)-tagged Arrdc3 and Arrdc4 were generated by subcloning into pECFP-C1 from Xpress-Arrdc4 or Arrdc3-V5 plasmids (kind gift of Dr. P. Patwari, Harvard Medical School, Boston, MA). Single and multiple amino acid substitutions were made in CFP-Arrdc3 and -4 as templates using site-directed mutagenesis. All mutations were verified by automatic dideoxynucleotide sequencing.
Cell Culture and Transfections-Human HEK293 cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (HyClone, Logan, UT) and antibiotics. Porcine aortic endothelial (PAE) cells were grown in Ham's F-12 medium containing 10% fetal bovine serum and antibiotics. Single cell clones of HEK293 and PAE cells stably expressing CAT-1-HA-GFP were generated by selection in the presence of G418 (400 g/ml). HEK293 stably expressing CFP-HA-DAT were previously described (37). PAE stably expressing YFP-HA-DAT were previously described (38).

Measurement of L-[ 3 H]Arginine
Transport-HEK293 cells grown to confluence in 96-well plates (poly-D-lysine coated) were washed twice with Locke's solution (in mM: NaCl 154; KCl 5.6; CaCl 2 2; MgCl 2 1; HEPES 10; NaHCO 3 3.6 ; glucose 5.6) prewarmed to 37°C containing 100 M L-arginine. Cells were then incubated with Locke's solution containing 100 M L-[ 3 H]arginine (10 Ci ml Ϫ1 ) for 15 s at 37°C, subsequently washed three times with ice-cold Locke's solution, and lysed in 0.5 M NaOH (50 l well Ϫ1 , 45 min at RT). After neutralization of the lysates with 50 l of 0.5 M HCl and 50 l of Lock's solution without L-arginine, the protein content of each sample was determined using the BCA Protein Assay Kit (ThermoFisher Scientific, Bonn, Germany). Radioactivity was measured by liquid scintillation counting.
To measure transport parameters in X. laevis oocytes, the complete coding sequence of hCAT-1 and hCAT-1-HA-GFP was cloned into pSGEM. 20 ng of each cRNA (in 40 nl of water) transcribed in vitro from these plasmids or 40 nl of water alone (as control) were injected into Xenopus oocytes. 3 days later, oocytes were preincubated for 60 min in buffer containing the indicated arginine concentrations. Then uptake of [ 3 H]arginine at the same respective concentrations was measured for 15 min. The apparent half-saturating substrate concentrations (K m ) and V max values for each carrier were determined by nonlinear regression fitting the data (after subtraction of the values obtained from water-injected oocytes).
HA Antibody Endocytosis Assay-The endocytosis assay using HA11 antibody was performed similarly as previously described in studies of DAT (38). Briefly, the cells grown on glass coverslips were incubated with 1 g/ml of HA11 in binding medium (DMEM or F-12, 0.1% bovine serum albumin) for 15 min and then with Me 2 SO (vehicle) or PMA (1 M) with or without preincubation with BIM (1 M) at 37°C for the indicated times. The cells were then washed with ice-cold Ca 2ϩ ,Mg 2ϩ -free PBS and fixed with freshly prepared 4% paraformaldehyde for 15 min at room temperature. The cells were stained with secondary anti-mouse antibody conjugated with Cy5 (5 g/ml) in CMF-PBS containing 0.5% bovine serum albumin at room temperature for 1 h to occupy surface HA11. After washing, the cells were permeabilized by a 5-min incubation in CMF-PBS containing 0.1% saponin, 0.5% bovine serum albumin at room temperature, and then incubated with the same secondary antibody conjugated with Cy3 (0.5 g/ml) in CMF-PBS containing 0.1% saponin, 0.5% bovine serum albumin for 45 min to stain internalized HA11. Both, primary and secondary antibody solutions were precleared by centrifugation at 100,000 ϫ g for 20 min. After staining, the coverslips were mounted in Mowiol (Calbiochem).
To obtain high resolution three-dimensional images of the cells, a Z-stack of images was acquired through Cy5, Cy3, and FITC filter channels using a Marianas Imaging workstation and SlideBook 4.2 software (Intelligent Imaging Innovation, Denver, CO) as described previously (39). Typically, 15-20 serial two-dimensional images were recorded at 300-nm intervals. All image acquisition settings were identical in each experiment. The Z-stack of images obtained was deconvoluted using a nearest neighbor algorithm of SlideBook4.2. In several experiments a Z-stack of images was obtained using a spinning disk confocal imaging system (Intelligent Imaging Innovation). Quantification of the relative amount of Cy5 and Cy3 fluorescence in the cell was performed using the statistics module of the SlideBook. The background-subtracted three-dimensional images were segmented using a minimal intensity of GFP as a low threshold. The integrated voxel intensity of Cy5 and Cy3 in each cell in the new image was then quantitated. The ratio of the Cy3 to Cy5 fluorescence signal per each cell was considered as the extent of endocytosis of HA11/CAT-1-HA-GFP complexes.
Endosomal Marker Co-localization Experiments-To determine the extent of co-localization of CAT-1-HA-GFP with EEA.1 and LAMP1, the cells were preincubated with leupeptin (21 M) for 1 h at 37°C to block lysosomal degradation. After treatment the cells were incubated with PMA for 30 min, fixed, and stained with the corresponding primary (EEA.1 or LAMP1) antibody followed by secondary antibody conjugated with Cy3, all incubations with antibodies in the presence of 0.1% saponin. Live-cell imaging was used to determine co-localization of CAT-1-HA-GFP with Texas Red-conjugated transferrin. The cells were incubated with 5 g/ml of Texas Red-conjugated transferrin during treatment with PMA, and imaged through Cy3 and FITC filter channels. Image acquisition was performed as described above for the HA internalization assay.
Immunoprecipitation-The cells grown in 35-mm dishes were placed on ice and washed with CMF-PBS, and the proteins were solubilized in Triton X-100/glycerol/HEPES lysis buffer supplemented with 100 mM NaCl, 10 mM N-ethylmaleimide and protease inhibitors for 20 min at 4°C (37). The lysate was then centrifuged at 100,000 ϫ g for 20 min to remove the insoluble material. Lysates were incubated with appropriate antibodies overnight and antibodies were precipitated with protein A-or protein G-Sepharose. Immunoprecipitates and aliquots of cell lysates were denatured in the sample buffer at 95°C, resolved by electrophoresis, and probed by Western blotting with various antibodies followed by the chemiluminescence detection. Several x-ray films exposed for different times were analyzed to determine the linear range of the chemiluminescence signals, and the quantifications were performed using densitometry and ImageJ software analysis.
Statistical Analysis-The statistical significance of the data were analyzed by unpaired or paired t tests. Significant differences were defined as those with p Ͻ 0.05.

PKC Activation Promotes Endocytosis of CAT-1-Down-reg-
ulation of the CAT-1 activity upon PKC activation has been demonstrated in Xenopus oocytes and a human glioblastoma cell line (7). To elucidate the mechanisms of this PKC-dependent down-regulation of CAT-1 and examine CAT-1 distribution in the cell, an HA epitope was introduced into the second extracellular loop of the CAT-1-GFP construct by replacing 11 amino acid residues without changing the size of the loop (Fig.  1A). Attachment of GFP at the carboxyl terminus of CAT-1 does not affect the transporter function (7,40). To test the functionality of the HA-tagged CAT-1-GFP (CAT-1-HA-GFP), HEK293 cells were transfected with this construct. HEK293 cells have endogenous cationic amino acid transporters but the uptake of L-[ 3 H]Arg was significantly increased in cells stably expressing CAT-1-HA-GFP (Fig. 1B), indicating that CAT-1-HA-GFP mediates substrate transport. Because cationic amino acid transporters are expressed in all mammalian cells, transport parameters of HA-tagged and wild-type CAT-1 were compared in Xenopus oocytes. The apparent K m values of each carrier were 0.37 mM for CAT-1 and 0.32 mM for CAT-1-HA-EGFP (supplemental Fig. S1). These experiments confirmed normal functionality of the CAT-1-HA-GFP mutant. Therefore, single cell clones of HEK293 and PAE cells stably expressing CAT-1-HA-GFP were used in subsequent experiments. PAE cells are well suited for microscopic analysis due to their flattened shape and low background autofluorescence. On the other hand, HEK293 cells, unlike PAE cells, allow high efficiency transient expression of proteins.
To examine the effect of PKC activation on CAT-1-HA-GFP localization, the HA antibody endocytosis assay was performed in PAE/CAT-HA-GFP cells untreated or treated with PMA. In this assay the cells were allowed to bind and internalize the HA11, and then localization of HA11/CAT-1-HA-GFP complexes was detected before and after cell permeabilization with secondary antibodies conjugated to Cy5 and Cy3, respectively. In the resulting images, GFP fluorescence corresponds to the total cellular CAT-1-HA-GFP, whereas Cy5 and Cy3 signals correspond to plasma membrane and internalized CAT-1-HA-GFP, respectively. In vehicle-treated cells, GFP fluorescence was diffusely distributed throughout the plasma membrane with some accumulation of GFP in the perinuclear area (newly synthesized/Golgi-localized transporter) ( Fig. 1C and supplemental Fig. S2A). Very few intracellular structures containing Cy3 were observed in this clone of PAE/CAT-1-HA-GFP cells under steady-state growth conditions ( Fig. 1C and supplemental Fig. S2A). Predominant localization of CAT-1-HA-GFP at the plasma membrane indicates that the anterograde transport of CAT-1-HA-GFP through the endoplasmic reticulum and Golgi is efficient.
PMA treatment for 30 min resulted in CAT-1-HA-GFP accumulation in vesicular, endosome-like structures containing both GFP and Cy3 but not containing Cy5 (Fig. 1C). The extent of endocytosis measured by the Cy3/Cy5 ratio was significantly increased by PMA (Fig. 1D). Pre-treatment of cells with an inhibitor of conventional PKCs, BIM, prevented PMAinduced endocytosis of CAT-1-HA-GFP, indicating that the effects of PMA are indeed PKC-dependent (supplemental Fig.  S2, C and D). These data suggested that activation of PKC leads to accumulation of the transporter in endosomes due to either accelerating CAT-1 internalization or inhibiting the recycling of constitutively internalized transporter, or affecting both processes. However, PMA significantly increased endosomal accumulation of CAT-1-HA-GFP in cells treated with an inhibitor of recycling, monensin (41), suggesting that activation of PKC does accelerate the internalization step of endocytosis in the absence of recycling. Increased accumulation of HA-CAT-1-GFP in PMA-treated cells was also observed in HEK293 cells, although substantially more of constitutive internalization of the transporter was observed in these cells as compared with PAE cells (data not shown). The specificity of CAT-1-HA-GFP detection by HA11 antibody is demonstrated in supplemental Fig. S2B.

CAT-1 Is Internalized by Clathrin-mediated Endocytosis into
Early, Recycling, and Late Endosomes-To characterize the mechanisms of PKC-dependent endocytosis of CAT-1-HA-GFP, we tested whether this endocytosis is mediated by clathrin-coated pits and vesicles, using siRNA to target the clathrin heavy chain (CHC) (41). Knockdown of CHC in PAE cells significantly reduced PKC-dependent accumulation of CAT-1-HA-GFP in endosomes (Fig. 2, A-C). A residual endocytosis of CAT-1-HA-GFP in CHC-depleted PMA-treated cells was likely due to incomplete CHC knockdown in a population of cells. These data suggest that PKC-dependent endocytosis of CAT-1 is mostly clathrin-dependent.
To further define the endocytic pathway followed by CAT-1 after PKC activation, localization of CAT-1-HA-GFP in PMAtreated PAE cells was compared with that of markers of early, recycling, and late endosomes. Immunofluorescence microscopy showed partial but substantial co-localization of CAT-1-HA-GFP with Texas Red-conjugated transferrin, a specific marker of early and recycling endosomes (Fig. 2D). CAT-1-HA-GFP was also partially co-localized with the early endosomal antigen 1 (EEA1) and the lysosomal-associated membrane protein (LAMP)-1, a resident protein of late endosomes and lysosomes (Fig. 2B). Altogether, these data demonstrate that PKC activation leads to internalization of CAT-1-HA-GFP into early endosomes and to subsequent sorting of a pool of the nonrecycled transporter to late endosomes and lysosomes.
PKC-dependent Ubiquitination of CAT-1-The clathrin-dependent endocytosis and pattern of localization of internalized CAT-1-HA-GFP in PMA-treated cells were reminiscent of the PKC-dependent endocytosis of DAT, a member of another transporter family, SLC6 (41). Because PKC-dependent endocytosis of DAT is mediated by its ubiquitination and because the endocytosis of yeast homologues of the SLC7 family transporters is also mediated by their ubiquitination, we examined whether CAT-1 is ubiquitinated in a PKC-dependent manner. CAT-1-HA-GFP ubiquitination was examined by immunoprecipitation of the transporter from untreated and PMA-treated PAE cells followed by probing the immunoprecipitates by Western blotting with the ubiquitin antibody. In vehicletreated cells, the very weak ubiquitin signal was detected in CAT-1-HA-GFP immunoprecipitates (Fig. 3A). When the cells were treated with PMA, the amount of ubiquitin in CAT-1-HA-GFP immunoprecipitates was dramatically increased (Fig.  3, A and B). The single ubiquitinated form of CAT-1-HA-GFP (Ub-CAT-1) was migrating on SDS-PAGE as a ϳ115-120-kDa band, whereas non-ubiquitinated CAT-1-HA-GFP ran as a ϳ90-kDa band. CAT-1-HA-GFP appears as a smeared band, which is characteristic of highly hydrophobic and glycosylated proteins. PMA-induced ubiquitination of CAT-1-HA-GFP was abolished by BIM, confirming that ubiquitination as well as endocytosis of this transporter are PKC-dependent. Similar results were obtained in HEK/CAT-1-HA-GFP cells (Fig. 3C).
The pool of ubiquitinated CAT-1-HA-GFP was rather small because the ubiquitinated transporter species was not detected by blotting with GFP antibody. However, when CAT-1-HA-GFP accumulated at the cell surface due to the blockade of endocytosis in cells depleted of clathrin, PKC-dependent CAT-1 ubiquitination was dramatically elevated (Fig. 3D). This allowed detection of a smeared GFP immunoreactivity corresponding to ubiquitinated CAT-1-HA-GFP (Fig. 3D). Altogether, the data in Fig. 3 demonstrate that activation of PKC results in ubiquitination of CAT-1.

PKC-dependent CAT-1 Ubiquitination Is Mediated by NEDD4 Family E3
Ligases-Ubiquitination of many transporters is catalyzed by NEDD4-2 and more rarely by NEDD4-1, members of the NEDD4 family of HECT-containing E3 ligases (26). We have previously reported that Nedd4-2 is essential for DAT ubiquitination (37,38). Therefore, we tested whether NEDD4-2 is also the E3 ubiquitin ligase involved in the PKCdependent ubiquitination of CAT-1. Fig. 4, A and B, shows that siRNA knockdown of NEDD4-2 in PAE cells significantly reduced the amount of ubiquitinated CAT-1-HA-GFP in PMA-treated cells, much like what was observed for YFP-HA-DAT (Fig. 4E). The antibody uptake assay and quantitative microscopy analysis revealed that depletion of NEDD4-2 by siRNA abolished PKC-dependent endocytosis of CAT-1-HA-GFP (Fig. 4, C and D). These data suggested that NEDD4-2 controls both endocytosis and ubiquitination of CAT-1 in PAE cells. Surprisingly, knockdown of Nedd4-2 did not significantly affect CAT-1-HA-GFP ubiquitination in HEK293 cells (Fig. 5,  A and B). Instead, PKC-induced ubiquitination of CAT-1-HA-GFP in HEK293 cells was substantially reduced by siRNA depletion of NEDD4-1 (Fig. 5, A and B). In the same experiments, depletion of NEDD4-1 had no effect on PKC-dependent ubiquitination of DAT in HEK293, whereas the presence of NEDD4-2 was critical to DAT ubiquitination (Fig. 5, C and D). Therefore, these data demonstrate that E3 ligases of the NEDD4 family are responsible for ubiquitination of transporters like CAT-1 and DAT. These experiments also demonstrate unexpected cell specificity of the mechanisms of CAT-1 ubiquitination with the predominant role of either NEDD4-2 or NEDD4-1 in different cell types.
Arrestin-like Arrdc Proteins Are Not Involved in PKC-dependent CAT-1 and DAT Endocytosis-CAT-1 does not have WW binding motifs within its cytoplasmic sequences and the mechanism of its interaction with NEDD4 ligases is unknown. Recently, it was reported that arrestin-like proteins serve as adaptors to link Rsp5 to yeast transporters lacking PPXY motifs  and promote ubiquitination and endocytosis of these transporters (29,42). Therefore, we examined if human arrestin-like proteins of the Arrdc family have the same function in CAT-1 regulation.
Because of high sequence homology and possible functional redundancy between Arrdc proteins (43), the approach of Arrdc mutant overexpression was used to test for the role of these proteins in CAT-1 endocytosis. We focused on Arrdc3 and Arrdc4 as only these two proteins among Arrdc proteins available to us were found to be located in cytoplasmic vesicular structures and plasma membrane (data not shown). CFPtagged forms of mArrdc3 and mArrdc4 (CFP-mArrdc3-wt and CFP-mArrdc4-wt) were localized mainly in EEA1 containing endosomes and in the plasma membrane when transiently expressed in PAE cells (Fig. 6 and data not shown). Both CFP-mArrdc4-wt and CFP-mArrdc3-wt also co-localized with CAT-1 and NEDD4-2 in endosomal structures ( Fig. 6 and data not shown). A truncated form of Arrdc4 that lacks the last 68 amino acids of the carboxyl terminus containing the two PPXY sequences motifs (CFP-mArrdc4-⌬Ct), and the CFP-mArrdc4-G109A mutant, in which the glycine residue important for interaction with the cargo was mutated, were generated (Fig.  6A). Fig. 6B shows that CFP-mArrdc4-⌬Ct was not located in endosomes and was mainly cytosolic, whereas localization of CFP-mArrdc4-G109A was not different as compared with that of CFP-mArrdc4-wt. These data suggest that PPXY interactions are important for the endosomal localization of ␣-arrestins.
As shown in Fig. 6, B and C, overexpression of these Arrrdc4 mutants had no effect on CAT-1 endocytosis. In similar experiments performed in PAE cells stably expressing YFP-HA-DAT, overexpression of mArrdc4 mutants did not have an inhibitory effect on DAT endocytosis (Fig. 6C). Furthermore, we observed no effects on DAT ubiquitination by CFP-mArrdc4-wt or CFP-mArrdc4-⌬Ct (data not shown). Experiments with similar Arrdc3 mutants confirmed no role of Arrdc proteins in PKC-dependent endocytosis and ubiquitination of CAT-1 and DAT.

DISCUSSION
In this study we examined the molecular mechanisms of CAT-1 endocytosis. In particular, we focused on the PKC-dependent down-regulation of CAT-1. To quantitatively analyze endocytosis of the transporter, an HA epitope tag was inserted in the second extracellular loop of CAT-1. Based on the substrate uptake assay (Fig. 1B and supplemental Fig. S1) and the direct comparison of the subcellular localization of CAT-1-GFP with and without HA, the CAT-1-HA-GFP chimeric protein is functional and can be used to study its regulation by PKC.
Several studies demonstrated down-regulation of CAT-1 activity by PKC (5,6,44,45). These studies suggested that CAT-1 phosphorylation affects its functional properties. Earlier work demonstrated an increased cationic amino acid transport activity in macrophages upon long-term PMA treatment, although such up-regulation could involve transporters other than CAT-1 (46). The work by Rotmann and co-workers (7,47) demonstrated that the PKC-induced decrease in CAT-1 activity is due to down-regulation of the cell surface CAT-1. Our studies using an HA11 antibody endocytosis assay are consistent with the notion that PKC accelerates endocytosis of CAT-1 and its redistribution from plasma membrane to endosomes.
The development of the quantitative assay to examine CAT-1 endocytosis allowed us to begin the analysis of the mechanism of this process. Knockdown of clathrin had a strong effect on CAT-1 internalization suggesting that CAT-1 is internalized mainly via clathrin-coated pits in PKC-activated cells. CAT-1 has been reported to associate with caveolin and detergent-insoluble membrane microdomains in several types of cells, but such association was not observed in other cells (48 -50). Moloney murine leukemia virus that utilizes mouse CAT-1 as a cellular receptor is internalized independently of dynamin II (51). Thus, the mechanism and the pathway of virus-induced endocytosis of CAT-1 could be different from that of the PKCdependent internalization of CAT-1. The presence of CAT-1 in the same endocytic compartments with the classical marker of clathrin-mediated endocytosis, the transferrin receptor, together with the strong effect of clathrin siRNA argue that CAT-1 is internalized via a clathrin-mediated pathway at least in PAE and HEK293 cells with active PKC.
Acceleration of CAT-1 endocytosis by PKC correlated with the dramatically increased ubiquitination of CAT-1, suggesting that ubiquitination may serve as the internalization signal in this system. Increased PKC-dependent ubiquitination of CAT-1 in the absence of endocytosis (Fig. 3D) suggested that transporter ubiquitination does occur at the plasma membrane. The observation of CAT-1 ubiquitination provides further support to the hypothesis that ubiquitination is the common effect of PKC activation on different families of transporters. For instance, mammalian transporters of the SLC6 family, such as DAT and glycine transporters, were also shown to be ubiquitinated upon PKC activation (17,52).
Ubiquitination typically targets the internalized cargo to the lysosomal degradation pathway. Accordingly, activation of PKC resulted in accumulation of CAT-1 in late endosomes and lysosomes, suggesting that a pool of CAT-1 undergoes lysosomal degradation (Fig. 3). This does not contradict to the earlier demonstration of the recycling of the internalized CAT-1 upon removal of the PKC-activating stimulus (7). After PKC inactivation, the transporter could be deubiquitinated and subsequently recycled. Significant co-localization of CAT-1 and transferrin (Fig. 3) suggests that a pool of CAT-1 is recycled back to the plasma membrane. It can be proposed that the PKC activity increases the relative amount of ubiquitinated CAT-1 in endosomes, which in turns increases the probability of internalized CAT-1 to escape recycling (default pathway) and be sorted to late endosomes and lysosomes. Thus, we suggest that PKC activation may have two effects on CAT-1 trafficking: (i) acceleration of internalization (Fig. 1), and (ii) reduction in the extent of recycling of the internalized transporter to the plasma membrane.
Our siRNA experiments demonstrated the role of NEDD4-2 and NEDD4-1 E3 ubiquitin ligases in CAT-1 ubiquitination (Fig. 5). The strong effect of NEDD4-2 depletion in PAE cells on both ubiquitination and endocytosis provides other evidence that ubiquitination is necessary for endocytosis. These experiments also indicate that NEDD4 family ligases are the common component of the PKC-dependent ubiquitination of transporters such as CAT-1 and DAT (Fig. 4). In contrast to DAT that is ubiquitinated by NEDD4-2 in all cell lines tested, CAT-1 ubiquitination was mediated by NEDD4-2 in PAE cells but by NEDD4-1 in HEK293 cells (Fig. 5). The nature of such cell specificity is unknown. The level of expression of NEDD4-1 is higher in HEK293 than in PAE cells. NEDD4-2 is present in the same amounts in these two cell types. Another difference that we observed is the significant constitutive endocytosis of CAT-1 in HEK293 cells as compared with PAE cells. Finally, it should be noted that endogenous CAT-1 is expressed ubiquitously including PAE and HEK293 (3,53), whereas DAT is expressed only in dopaminergic neurons and not normally expressed in the cell lines used. Thus, it is possible that CAT-1 evolutionally developed the ability to be regulated by the abundant E3 ligases in a cell-dependent manner. Interestingly, injection of NEDD4-2 mRNA did not affect arginine transport in Xenopus oocytes (54), which is in agreement with the observation that NEDD4 ligases mediate only PKC-dependent ubiquitination of CAT-1.
Because CAT-1 does not have PPXY motifs, we hypothesized that the NEDD4-1/2 interaction with this transporter can be mediated by ␣-arrestins. However, using overexpression of the Arrdc mutants approach, we were unable to demonstrate the role of these adaptor proteins in CAT-1 ubiquitination and endocytosis. Likewise, no interaction with DAT, another transporter that does not have a PPXY motif, or a role in DAT ubiquitination was observed. Thus, even though a role for arrestinrelated trafficking adaptors in regulation of SLC7 transporters in yeast was demonstrated (29), it is possible that mammalian homologues of arrestin-related trafficking adaptor proteins have different functions. In fact, a bulk of Arrdc3 and Arrdc4 was located in endosomes rather than at the plasma membrane (Fig. 6) where the process of transporter ubiquitination begins. During the preparation of our manuscript, two studies reported the involvement of Arrdc3 in the ubiquitination and degradation (but not in internalization) of ␤2-adrenergic receptor and ␤4-integrin (55,56).
In summary, we propose that ubiquitination of various transport proteins and possibly other transmembrane proteins that lead to their endocytosis and lysosomal targeting, is the general outcome of PKC activation. The principal components of this signaling pathway are NEDD4-2 and, in some cases, NEDD4-1 E3 ubiquitin ligases and the clathrin internalization machinery.