Identification of an Insulin-responsive, Slow Endocytic Recycling Mechanism in Chinese Hamster Ovary Cells*

In adipocytes, the insulin-regulated aminopeptidase (IRAP) is trafficked through the same insulin-regulated recycling pathway as the GLUT4 glucose transporter. We find that a chimera, containing the cytoplasmic domain of IRAP fused to transmembrane and extracellular domains of the transferrin receptor, is slowly recycled and rapidly internalized in Chinese hamster ovary cells. Morphological studies indicate that the chimera is slowly trafficked through the general endosomal recycling compartment rather than being sorted to a specialized recycling pathway. A chimera in which a di-leucine sequence within the cytoplasmic domain of IRAP has been mutated to alanines is rapidly internalized and rapidly recycled, indicating that this di-leucine is required for the slow recycling but not for the rapid internalization. Insulin stimulates a 2–3-fold increase in the recycling of the chimera and only a 1.2-fold increase in the recycling of the transferrin receptor. The effect of insulin on the recycling of the chimera is blocked by wortmannin, a phosphatidylinositol 3′-kinase inhibitor. GTPγS (guanosine 5′-3-O-(thio)triphosphate) increases the recycling of the chimera by 50% but has no effect on the recycling of the transferrin receptor. In these studies, we have identified in Chinese hamster ovary cells a novel, slow endocytic recycling mechanism that is regulated by insulin.

GLUT4 is a glucose transporter isoform predominantly expressed in fat and muscle cells (reviewed in Ref. 1). In the absence of insulin, ϳ90% of GLUT4 is localized to intracellular compartments. Exposure of fat and muscle cells to insulin stimulates a translocation of GLUT4-containing vesicles to the plasma membrane, shifting the distribution to a new steady state in which 50% of GLUT4 is on the cell surface. In both the basal and insulin-stimulated states GLUT4 is continually internalized and recycled back to the cell surface (2,3). Insulin induces the redistribution of GLUT4 to the cell surface predominantly by increasing the recycling rate of GLUT4 by 3-9-fold (2)(3)(4)(5). Insulin also affects general membrane trafficking; however, these effects are small compared with the effect on GLUT4 (6). Both the carboxyl-and amino-terminal cytoplasmic domains of GLUT4 contain information that targets GLUT4 to the insulin-responsive pathway (7)(8)(9). When GLUT4 is expressed in cell types other than fat or muscle it is concentrated intracellularly in the absence of insulin, yet insulin does not cause a large translocation to the cell surface (see, e.g., Refs. 10 -14).
A type II membrane protein with aminopeptidase activity has been isolated from GLUT4-containing vesicles (15,16). This protein, originally named vp165 or gp160, is now referred to as IRAP, for insulin-regulated aminopeptidase (17). In fat and muscle cells, IRAP has the same trafficking characteristics as GLUT4 (15,16,18,19). The information required for targeting IRAP to the insulin-regulated membrane trafficking pathway is likely to be contained within its single 109-amino acid cytoplasmic domain because, as noted above, the information that regulates GLUT4 trafficking is found within its cytoplasmic domains.
The internalization and recycling of the transferrin receptor (TR) 1 has been extensively characterized. Briefly, TR binds diferric transferrin (Tf) and is rapidly internalized through clathrin-coated pits into acidic endosomes. At the acidic intralumenal pH of endosomes, Fe is released from Tf and the iron-free Tf remains bound to its receptor (20). The Tf⅐TR complex is returned to the plasma membrane, and Tf is released from the TR at the neutral extracellular pH. We and others have used TR chimeras as reporter molecules to study endocytic trafficking (see, e.g., Refs. 21 and 22). Advantages of using the TR for these studies are that substitution of cytoplasmic domains for the cytoplasmic domain of the TR does not affect the binding of Tf to the TR, and therefore Tf can be used in biochemical and microscopy studies to characterize the internalization, intracellular trafficking, and recycling back to the cell surface of the chimeras. The only known endocytic trafficking motif of the TR is a cytoplasmic YTRF internalization motif (23). Unlike internalization, efficient recycling back to the plasma membrane does not require a specific recycling motif (24 -26). Consequently, the transmembrane and extracellular domains of the TR do not contribute to the trafficking characteristics of chimeras containing these domains.
In this report, we characterize the behavior of chimeras between IRAP and the human TR. In one chimera, vp⅐TR, the cytoplasmic domain of IRAP was substituted for the single amino-terminal cytoplasmic domain of the TR, and in the other, vpTM⅐TR, the cytoplasmic and transmembrane domains of IRAP were fused to the extracellular domain of the TR. Both chimeras are recycled at about one-third the rate of the wild type TR and internalized at the same rate as the TR. These data demonstrate that the cytoplasmic domain of IRAP contains information that promotes rapid internalization and slow recycling. Morphological studies indicate that the vp⅐TR is slowly trafficked through the general endosomal recycling compartments rather than being sorted to a specialized recycling pathway. The slow recycling of vp⅐TR is dependent on a dileucine sequence located at position 76 in the cytoplasmic domain of IRAP because a mutant in which alanines have been substituted for these leucines is recycled at the same rate as wild type TR. Insulin stimulates the recycling of vp⅐TR by 2-3-fold and causes a much smaller increase, 1.2-fold, in the recycling rate of the wild type TR. The effect of insulin on the recycling of the chimera is blocked by wortmannin, a phosphatidylinositol 3Ј-kinase inhibitor. GTP␥S increases the recycling of the chimera but has no effect on the recycling of the TR. We conclude that CHO cells possess a specialized, slow recycling mechanism that is regulated by insulin.

MATERIALS AND METHODS
Cell Culture-TRVb, a CHO cell line that does not express functional endogenous TR, was used for transfection studies (27). The cells were transfected with the vp⅐TR chimera using Lipofectin (Life Technologies, Inc.) according to the manufacture's recommendations. The G418 resistance plasmid, pSV3Neo, was co-transfected as a selectable marker. Cells were selected in 1 mg/ml G418, and individual colonies were isolated with cloning cylinders. The isolation and characterization of TRVb cells expressing the wild type human TR and the ⌬3-59 TR deletion construct have been described (27,28). Cells were grown in either Ham's F-12 medium or McCoy's 5A medium containing 5% fetal bovine serum, penicillin-streptomycin (Life Technologies, Inc.), and 220 mM sodium bicarbonate (Sigma).
Ligands-Human Tf was obtained from Sigma and further purified by Sephacryl S-300 gel filtration. Diferric Tf, 125 I-Tf, and 55 FeCl 3 were prepared as described previously (21). 125 I and 55 Fe 2 were purchased from NEN Life Science Products. Tf was labeled with the fluorescent dye Cy3 (Biological Detection Systems, Pittsburgh, PA) according to the manufacturer's instructions.
Chimera Construction-Oligonucleotides were purchased from Operon Technologies (Alameda, CA) or Life Technologies, Inc. cDNAs of IRAP (16) and the human TR cloned into pUC8 vector (27) were used as templates. The overlapping oligonucleotide PCR scheme previously used to create chimeras in which heterologous cytoplasmic domains were substituted for the TR cytoplasmic domains was used to create the vp⅐TR chimera (21). The oligonucleotides bridging the cytoplasmic domain sequences of IRAP and the transmembrane sequences of the TR were: 5Ј-CCAGATGAGAATGACAAAAGGTGTAGTGGAAG and its complement. The underlined nucleotides are the membrane proximal cytoplasmic sequences of the IRAP, and the remainder are from the transmembrane domain of the human TR. The oligonucleotides used to fuse the cytoplasmic and transmembrane domains of IRAP to the extracellular domain of the TR were: 5Ј-TATCTACTGAAAGGGGTAC and 5Ј-CTACCCCTTTCAGTAGATAAATC. The underlined sequences are from IRAP. The oligonucleotides used to mutate the double leucines at positions 53 and 54 were: 5Ј-TCGAGGGCAGCAGTGAGAGGTGTTG-GTG and its complement. The underlined nucleotides are the mutated codons. The oligonucleotides used to mutate the double leucines at positions 76 and 77 were: 5Ј-CTGCCAAGGCAGCAGGCATGTCCTTC and its complement. The underlined nucleotides are the mutated codons. The cytoplasmic and transmembrane domains of the PCRgenerated products were sequenced by the Rockefeller University sequencing facility.
TR Trafficking Assays-For each assay, cells were grown in either six-well or 24-well clusters. Nonspecific binding of Tf was determined in duplicate for each data point by including a 100-fold excess of unlabeled Tf in the incubation medium. All data have been corrected for nonspecific binding, which was typically less than 10% of the total. The buffer used for experimental incubations was McCoy's 5A medium, 220 mM sodium bicarbonate, 20 mM HEPES, pH 7.4 (medium 1). The neutral wash buffer was 150 mM NaCl, 20 mM HEPES, 1 mM CaCl 2 , 5 mM KCl, 1 mM MgCl 2 , pH 7.4 (medium 2). All incubations at 37°C were performed in a humidified atmosphere with 5% CO 2 . The kinetic assays used to characterize the rates of internalization, recycling, steady-state surface-to-internal distribution, and iron accumulation have been de-scribed in detail elsewhere (29,30).
Steady-state Distribution-The amount of chimera on the surface was determined using a previously described assay (30).
Internalization Assay-Cells were incubated for 1 h in medium 1 at 37°C, followed by an incubation in medium 1 with 3 g/ml iodinated Tf at 37°C for varying times (2-8 min). The cells were placed on ice, washed, and incubated for 5 min at 4°C in 0.5 M NaCl, 0.5 M glacial acetic acid, pH 3.0. The cells were then washed with medium 2 (4°C), solubilized and the radioactivity measured. This value represents the Tf internalized during the incubation at 37°C. The steady-state amount of Tf bound to the cell surface is determined in parallel dishes by incubating cells with iodinated Tf for 2 h at 4°C. The slope of a plot of the ratio of internal Tf to surface Tf versus time is the internalization rate constant. 55 Fe Accumulation-Cells were incubated in medium 1 for 1 h at 37°C, and then incubated in 3 g/ml 55 Fe 2 Tf at 37°C. At the times indicated cells were washed three times with medium 2, solubilized, and cell-associated 55 Fe determined. To compare Fe accumulation data among cells expressing varying amounts of chimera or TR, the Fe accumulation data were normalized to surface Tf binding of 30 ng/well. Actual surface Tf binding was determined in parallel dishes as described above.
Recycling Assay-Cells were preincubated for 1 h in medium 1 at 37°C, followed by an incubation in medium 1 with 3 g/ml iodinated Tf for Ͼ2 h at 37°C. The cells were washed with medium 2, incubated for 2 min in 0.2 M NaCl, 50 mM MES, pH 5.0, followed by three washes with medium 2 supplemented with 3 g/ml unlabeled Tf. Incubation in the acidic medium releases Tf bound to TR on the cell surface. The cells were incubated at 37°C in medium 1 with 3 g/ml unlabeled Tf and 100 M of the iron chelator desferroxamine (efflux medium). At the times indicated, the medium was collected and the cells solubilized. The radioactivity in the medium is the Tf released from the cells during the incubation, and the cell-associated radioactivity is the Tf remaining inside cells. The recycling rate constant is the slope of the natural logarithm of the percent Tf remaining cell-associated versus time. To examine the effect of insulin on recycling, cells were treated as above except that the efflux medium was made 100 nM insulin. For experiments measuring the effect of wortmannin on recycling, 100 nM wortmannin was added to cells for the last 30 min of incubation with the iodinated Tf, and 100 nM wortmannin was added to all subsequent buffers used in the assay. For experiments measuring the effect of nocodazole on the recycling of the vp⅐TR chimera, 10 g/ml nocodazole was added for the last 30 min of incubation with iodinated Tf and 10 g/ml nocodazole was included in all subsequent incubations.
Recycling in Permeabilized Cells-The assay to measure recycling in streptolysin O-permeabilized cells has been described in detail (31). Streptolysin O was purchased from Murex, Inc. (Dartford, United Kingdom). Donor cytosol was prepared from mouse liver and used at a protein concentration of 9 mg/ml (31). GTP␥S was used at a concentration of 50 M.
Microscopy-Cells, grown on coverslip bottom dishes, were incubated with 3 g/ml Cy3-Tf in medium 1 at 37°C for Ͼ2 h. The cells were labeled with bodipy-sphingomyelin (SM) as described previously (26). Briefly, the cells were pre-incubated with fluorescent Tf (Cy3-Tf) for 3 h at 37°C, chilled to 4°C, washed once with medium 2, and the plasma membrane labeled with fluorescent SM by incubating with bodipy-SM liposomes for 30 min at 4°C (26). The cells were then washed three times with medium 2 and incubated at 37°C for 20 min in medium 2 containing 3 g/ml Cy3-Tf. Following this incubation, the cells were placed on ice and washed three times, 5 min each wash, with 5 mg/ml fatty acid-free bovine serum albumin. To examine the steady-state distributions of Tf and SM, the cells were fixed for 20 min in 3.7% formaldehyde in medium 2. To examine the recycling of Tf and SM, the cells were incubated at 37°C for 30 min in medium 2 with 5 mg/ml fatty acid-free bovine serum albumin and 3 g/ml Tf. At the end of the incubation, the cells were washed once with medium 2 and fixed in 3.7% formaldehyde in medium 2. For the efflux experiments, cells were singly labeled with either Cy3-Tf or bodipy-SM and processed in parallel. For experiments using nocodazole, 10 g/ml nocodazole was added to the cells during the first incubation with Cy3-Tf. Nocodazole was also added to the buffers used for incubations at 37°C. The cells were examined on a Leitz DM-IRB fluorescence microscope with a cooled CCD camera (Princeton Instruments Inc., West Chester, PA). Images were collected using Metamorph Image processing software (Universal Imaging Corporation, West Chester, Pa). Immunofluorescence with a monoclonal antibody specific for the extracellular domain of the human TR (B3/25; Boehringer Mannheim) was performed as described previously (21).

RESULTS
Expression of the vp⅐TR Chimera in CHO Cells-To determine the role that the cytoplasmic and transmembrane domains play in determining the endocytic trafficking of IRAP, two chimeras were constructed. In one, vp⅐TR, the 109-amino acid cytoplasmic domain of IRAP was fused to the transmembrane and extracellular domains of the human TR, and in the other, vpTM⅐TR, the cytoplasmic and transmembrane domains of IRAP were fused to the extracellular domain of the human TR ( Fig. 1). TRVb CHO cells were transfected with these constructs and single colonies were isolated. TRVb CHO cells do not express functional endogenous TR and therefore provide a TR-free background for studies of the trafficking of transfected TR constructs (27). Individual clonal lines expressing the chimeras were identified by immunofluorescence using a monoclonal antibody against the extracellular domain of the human TR, B3/25 (27). The wild type human TR and a cytoplasmic deletion mutant of the human TR (⌬3-59 TR) have previously been transfected into TRVb CHO cells (25,27). Uptake of iodinated Tf was examined for each clone and clonal lines that express less of the chimeras than control cells express of the wild type TR were chosen for study. Several independently isolated clonal lines expressing the chimeras were studied, and in no case were there significant differences among clones expressing the same construct.
vp⅐TR and vpTM⅐TR Are More Concentrated Intracellularly than the Wild Type TR-Steady-state conditions were established as follows. Cells were incubated with iodinated Tf at 37°C, and the amount of cell-associated radioactivity was determined as a function of time. Cell-associated radioactivity asymptotically approaches a plateau level in cells expressing the vp⅐TR chimera ( Fig. 2A). Binding of Tf to TR on the surface saturates within 5 min at 37°C (20), and the increase in cellassociated radioactive Tf reflects the delivery of internal chimeras to the surface which at the start of the experiment are unoccupied with iodinated Tf. Thus, the time required to reach the plateau level is the time required for the internal population of the chimera to cycle through the plasma membrane. In cells expressing the vp⅐TR chimera or the vpTM⅐TR chimera, the plateau level is reached within 100 min. In subsequent experiments, incubations with Tf of at least 2 h were used to measure the steady-state trafficking kinetics of the chimeras.
The steady-state distributions of the vp⅐TR and vpTM⅐TR chimeras between the surface and the cell interior were determined. At steady-state, ϳ10% of the vp⅐TR and vpTM⅐TR chimeras are on the surface, whereas ϳ30% of the wild type TR and ϳ70% of the ⌬3-59 TR are on the surface (Fig. 2B). The distribution of the ⌬3-59 TR is shifted to the cell surface, relative to the TR, because deletion of the cytoplasmic domain of the TR removes the internalization motif and thereby slows internalization. The observation that the distribution of vp⅐TR is different from the distribution of the ⌬3-59 TR demonstrates that the cytoplasmic domain of IRAP contains information that regulates endocytic trafficking. The finding that the two chimeras have the same distribution indicates that the transmembrane domain of IRAP does not contain information that affects endocytic trafficking in CHO cells. This conclusion was con-firmed upon further characterization of the trafficking parameters (see below). The observation that the distribution of vp⅐TR is different from that of the TR indicates differences in the trafficking of the vp⅐TR relative to the TR. The greater intracellular concentration of vp⅐TR, relative to the TR, could be due to an increased rate of internalization, a decreased rate of recycling, or changes in both parameters.
The vp⅐TR Chimera Is Rapidly Internalized from the Plasma Membrane-The internalization rate constants of the chimeras were measured (Fig. 3A). The vp⅐TR and vpTM⅐TR chimeras are internalized at essentially the same rate as the TR, which is ϳ7-fold faster than the rate of ⌬3-59 TR internalization. These data show that the cytoplasmic domain of IRAP contains a motif that promotes rapid internalization of the vp⅐TR and vpTM⅐TR.
Delivery of Fe to cells is the biological function of Tf internalization. Efficient Fe accumulation by cells is dependent on internalization of Tf into acidic endosomal compartments, where Fe is released from Tf. To determine whether Tf internalized by the vp⅐TR chimera is capable of delivering Fe to FIG. 2. A, data from a representative experiment in which the total cell-associated iodinated Tf in cells expressing the vp⅐TR chimera is plotted as a function of time. Cells were incubated with 3 g/ml iodinated Tf at 37°C for the times indicated. The cells were then washed and the cell-associated radioactivity determined. Each data point is the average of triplicate measurements Ϯ S.D. Similar results were obtained using other clonal lines expressing the vp⅐TR chimera. B, the percent of wild type TR, vp⅐TR, vpTM⅐TR, and ⌬3-59 TR on the cell surface at steady-state is shown. The data are the mean values Ϯ S.E. of at least five independent measurements. The percent cell surface was measured as described under "Materials and Methods." The data shown are from clonal lines expressing the different constructs. Similar results were obtained using other clonal lines expressing the vp⅐TR or vpTM⅐TR chimeras. cells, the ability of cells expressing vp⅐TR to accumulate Fe from diferric Tf was measured (Fig. 3B). Cells expressing the vp⅐TR chimera accumulate Fe from diferric Tf as efficiently as do cells expressing the TR, indicating that vp⅐TR delivers Tf to acidic endosomal compartments.
The vp⅐TR and vpTM⅐TR Chimeras Are More Slowly Recycled than the TR-We next determined the rate at which the chimeras are recycled. The results from a representative experiment are shown in Fig. 4 (A and B). In cells expressing the wild type TR, greater than 80% of the Tf is released from cells within 60 min. In cells expressing the vp⅐TR chimera, the decrease in cell-associated Tf is slow relative to that observed in cells expressing the TR. Although the release of Tf is slower in cells expressing the vp⅐TR chimera, Tf is recycled from these cells to the same degree as it is in cells expressing the TR, with greater than 80% of the Tf released by 120 min. The slow recycling of Tf in cells expressing the vp⅐TR chimera is not due to inefficient Fe release from Tf, inasmuch as Fe is efficiently released from Tf internalized into cells by vp⅐TR, as shown in Fig. 3B. Greater than 90% of the Tf released from cells expressing either vp⅐TR or wild type TR was precipitated with 15% trichloroacetic acid at 4°C. These data indicate that Tf is recycled intact and it is not being significantly targeted to degradative endosomal compartments.
To quantify the differences in recycling rates between vp⅐TR and the TR, the initial recycling rate constants were calculated by plotting the natural log of the percent cell-associated Tf versus time (Fig. 4B). The mean recycling rate constants for the TR, the ⌬3-59 TR, and the vp⅐TR and vpTM⅐TR chimeras are presented in Fig. 4C. The recycling rate of the vp⅐TR and vpTM⅐TR chimeras is ϳ3-fold slower than the recycling rate of the TR. The ⌬3-59 TR is recycled at the same rate as the TR because the recycling of membrane proteins back to the cells surface is by a bulk flow process (26,32). Thus, finding that vp⅐TR and vpTM⅐TR are more slowly recycled than either the ⌬3-59 TR or the TR demonstrates that the cytoplasmic domain of IRAP contains information that slows recycling.
At steady state, the distribution of a membrane protein between the surface and interior of cells is determined by the ratio of the recycling to internalization rate constants (25). The measured rates of internalization and recycling for the vp⅐TR and vpTM⅐TR chimeras are consistent with 8% of the chimeras being on the cell surface at steady state, which is comparable to the measured values of ϳ10% (see Fig. 2B). Thus, the slower recycling of the vp⅐TR and vpTM⅐TR chimeras, relative to the TR, accounts for the more pronounced intracellular localization of the chimera.
Because we find no differences in the kinetics of trafficking of vp⅐TR and vpTM⅐TR, unless noted otherwise, the following microscopy experiments were performed on clonal lines expressing the vp⅐TR chimera.
The vp⅐TR Chimera Traffics through the General Endosomal System-Two general mechanisms could account for the slow recycling of the vp⅐TR relative to the wild type TR. The vp⅐TR chimera could traffic through different intracellular endosomal compartments than the TR, in which case the cytoplasmic domain of vp⅐TR would contain information that targets the chimera to a slowly recycling pathway. Alternatively, vp⅐TR could traffic more slowly through the same endosomal compartments as the TR. In this case, the slow recycling of vp⅐TR would reflect regulated trafficking of vp⅐TR through the general endosomal recycling pathway, and the cytoplasmic domain of IRAP would contain information that slows traffic of vp⅐TR through the endosomal recycling system. In the former model, the vp⅐TR chimera would be found in different intracellular compartments than the TR, whereas in the latter model, vp⅐TR would be in the same intracellular compartments as the TR.
To determine whether the vp⅐TR chimera traffics through the general endosomal recycling compartments, the intracellular compartments containing Tf internalized by the vp⅐TR chimera were compared with those containing fluorescent sphingomyelin internalized from the plasma membrane. Previous studies have shown that fluorescent sphingomyelin inserted into the plasma membrane is trafficked through the same endosomal compartments as the TR (26,33). In cells expressing vp⅐TR, the intracellular compartments labeled with the fluorescent Tf internalized from the medium also contain bodipy-SM internalized from the plasma membrane (Fig. 5, A  and B). The majority of Tf and bodipy-SM are concentrated as a single fluorescent patch located in the peri-centriolar region of the cell (arrows). In CHO cells, this juxta-nuclear compartment has been previously characterized as a compartment along the general endosomal recycling pathway (see, e.g., Refs. 26,34,and 35). Tf and other recycling molecules accumulate in this compartment because the rate-limiting step in return to the surface is transport from this compartment back to the cell surface (26). Tf and bodipy-SM are also observed in peripheral, small punctate structures which are early sorting endosomes (36). Finding co-localization of Tf and bodipy-SM indicates that the vp⅐TR chimera is trafficked through the general endosomal recycling system. A similar co-localization of the vpTM⅐TR chimera and bodipy-SM was observed (not shown). The distributions of the Tf and bodipy-SM in cells expressing the wild type TR are similar to the distributions observed in cells expressing the chimera (Fig. 5, C and D), providing further support for the conclusion that the vp⅐TR chimera is slowly recycled through the general recycling system rather than being targeted to a distinct slowly recycling pathway.
A number of membrane compartments are localized to the juxta-nuclear region of cells, including the Golgi, TGN, and lysosomes (37). To more rigorously demonstrate a co-localization between bodipy-SM and the chimera, we examined the co-distribution of vp⅐TR and bodipy-SM in cells in which the microtubule cytoskeleton had been depolymerized with nocodazole. The peri-centriolar, endocytic recycling compartment is vesiculated and dispersed when cells are treated with nocodazole (34). The peri-centriolar localization of vp⅐TR is fragmented and dispersed by treatment of cells with nocodazole, and most significantly, the fluorescent Tf remains co-distributed with the bodipy-SM (Fig. 5, E and F). The wild type TR and the bodipy-SM also co-distribute in nocodazole-treated cells (Fig. 5, G and H). These results indicate that the vp⅐TR chimera and the bodipy-SM are in the same compartment because if they were in distinct, nocodazole-sensitive peri-centriolar compartments, they would not necessarily remain colocalized following nocodazole treatment. Treatment of cells with nocodazole does not promote aberrant mixing of distinct compartments located in the peri-centriolar region, as there is no mixing of the peri-centriolar, endocytic recycling compartment and the TGN in cells treated with nocodazole (not shown). Although nocodazole treatment disrupts the morphology of the recycling compartment, it has no effect on the recycling rate of the wild type TR (34). The recycling rate constant for vp⅐TR is unchanged by nocodazole, demonstrating that an intact microtubule cytoskeleton is not required for slow recycling of the chimera (data not shown).
Endogenous Hamster IRAP Co-localizes with TR in CHO Cells-To examine the distribution of endogenous IRAP, CHO cells expressing the wild type human TR were incubated with Cy3-Tf to label endosomal compartments, and the endogenous IRAP was localized using polyclonal antibodies against the cytoplasmic domain of IRAP (16). The endogenous IRAP colocalizes with the TR (Fig. 6). This finding indicates that the endogenous IRAP, like the vp⅐TR chimera, is in the general endosomal system. It is not possible to examine the co-distribution of the vp⅐TR chimera and endogenous IRAP, inasmuch as the only available antibody to IRAP recognizes the cytoplasmic domain and it therefore reacts with the vp⅐TR chimera.
Expression of vp⅐TR Does Not Slow General Membrane Recycling-We next sought to determine whether the slow recycling of vp⅐TR is characteristic of the chimera or if expression of the chimera slows the recycling of other molecules. To address this question, we examined the recycling of bodipy-SM in cells expressing the vp⅐TR chimera. Endosomal compartments were labeled with bodipy-SM as described for Fig. 5, and the transport of bodipy-SM from the peri-centriolar, endocytic recycling compartment to the cell surface was examined (26). In parallel dishes, the recycling of Cy3-Tf was examined. Following 30 min of efflux, a significant amount of Cy3-Tf remains cell-associated in cells expressing the chimera, whereas almost no Cy3-Tf is detectable in cells expressing the TR (Fig. 7, A-D). In contrast, the majority of bodipy-SM is released from cells expressing the chimera as well as from those expressing the TR during 30 min of efflux (Fig. 7, E-H). This qualitative analysis, demonstrating that fluorescent Tf is more slowly recycled from cells expressing the vp⅐TR chimera than from cells expressing the TR, agrees with the biochemical analysis of recycling of the vp⅐TR chimera. These data also demonstrate that vp⅐TR is trafficked more slowly through the recycling compartments than is bodipy-SM, and consequently that the vp⅐TR chimera is specifically retained within the general endosomal recycling compartment.
The Di-leucine Sequence at Positions 76 -77 of the Cytoplasmic Domain of IRAP Plays a Role in the Slow Recycling of the vp⅐TR Chimera-The above results demonstrate that the cytoplasmic domain of IRAP contains information that promotes rapid internalization and slows recycling. One class of endocytic trafficking motifs is based on di-leucine sequences (38). These motifs can function as both internalization and intracellular trafficking motifs. The cytoplasmic domain of IRAP contains two di-leucine motifs, one at positions 53-54 and the other at positions 76 -77 (Fig. 1). To investigate the role of these sequences in the trafficking of vp⅐TR three constructs were created: LL53AA, in which alanines were substituted for the leucines at positions 53-54; LL76AA, in which alanines were substituted for the leucines at positions 76 -77; and LL53,76AA, in which alanines were substituted for the leucines at positions 53-54 and 76 -77. The mutant cDNAs were transfected into TRVb cells, clonal lines isolated and the endocytic behaviors of the constructs characterized. The mean internalization rate constants derived from a number of independent experiments are shown in Fig. 8A. Neither of the di-leucine sequences are necessary for internalization as all three mutant constructs are internalized at the same rate as vp⅐TR. These A, B, E, and F) or human TR (C, D, G, and H). findings demonstrate that sequences other than the di-leucine sequences are responsible for the rapid internalization of the vp⅐TR chimera.

FIG. 7. Efflux of Cy3-Tf (A-D) and bodipy-SM (E-H) from cells expressing vp⅐TR (
The mean recycling rate constants derived from a number of independent experiments are presented in Fig. 8B. The LL53AA mutant is recycled at the same rate as vp⅐TR, indicating that the di-leucine sequence at positions 53-54 is not required for the slow recycling of vp⅐TR. Mutation of the dileucine sequence at positions 76 -77, however, increases the recycling rate to approximately the recycling rate of the wild type TR. This finding demonstrates that the di-leucine at positions 76 -77 of the cytoplasmic domain of IRAP is involved in regulating the recycling of the vp⅐TR chimera. This conclusion is further supported by the finding that the construct containing mutations of both di-leucine sequences, LL53,76AA, is also rapidly recycled. In fluorescence microscopy analysis, both the rapidly recycled LL76AA and the slowly recycled LL53AA, like the vp⅐TR chimera, are localized to the peri-centriolar, endocytic recycling compartment (not shown). Therefore, mutation of the dileucine at positions 76 -77 affects the kinetics of trafficking without detectably affecting the intracellular itinerary of the chimera. These data are consistent with the proposal that the slow recycling of vp⅐TR is achieved by regulating its trafficking through the general endosomal recycling pathway.
Recycling of the vp⅐TR Chimera Is Regulated by Insulin-In fat and muscle cells, IRAP trafficking is regulated by insulin (15,16,18,19). To examine the effect of insulin on the trafficking of vp⅐TR in CHO cells, the recycling rate constant of the vp⅐TR chimera and the TR were measured in the presence of 100 nM insulin. Data from representative experiments on the effect of insulin on the recycling of the wild type TR and vp⅐TR are shown in Fig. 9 (A and B), and a summary of a number of experiments on the effect of insulin on the recycling of the TR, vp⅐TR, vpTM⅐TR, LL53AA, and LL76AA are presented in Fig.  9C. Insulin had a large effect on the recycling rate constant of the vp⅐TR and vpTM⅐TR chimeras, increasing the rate constant by 2-3-fold, whereas insulin had a small effect on the recycling of the TR, increasing the recycling rate constant by 1.2-fold. These data demonstrate that the recycling of the vp⅐TR chimera is regulated by insulin. Analysis of the effect of insulin on the recycling of the LL53AA and LL76AA chimeras demonstrates that slow recycling is linked to the insulin response, as insulin has a large effect on the recycling of the slowly recycled LL53AA mutant and a small effect on the rapidly recycled LL76AA mutant (Fig. 9C).
The effect of insulin on expression of the vp⅐TR chimera and the TR on the cell surface was also examined. Cells were incubated for 10 min in the presence or absence of 100 nM insulin, chilled to 4°C, and surface iodinated Tf binding measured at 4°C. Insulin caused a 2.3 Ϯ 0.3-fold (Ϯ S.D., n ϭ 4) increase in surface expression of the vp⅐TR chimera, and a 1.3 Ϯ 0.4-fold (Ϯ S.D., n ϭ 4) increase in surface expression of the TR. The increased surface expression of vp⅐TR can be accounted for by the insulin-induced increase in the recycling rate of the vp⅐TR chimera, and this indicates that insulin does not have a large effect on the internalization rate of the chimera.
The Insulin-stimulated Recycling of the vp⅐TR Chimera Is Inhibited by Wortmannin-The insulin-induced translocation of GLUT4 to the cell surface is inhibited by wortmannin, a phosphatidylinositol 3Ј-kinase inhibitor (see, e.g., Ref. 39). We therefore examined whether the insulin-regulated recycling of the vp⅐TR chimera in CHO cells is also inhibited by wortmannin. Treatment of cells with 100 nM wortmannin, a concentration that inhibits the phosphatidylinositol 3Ј-kinase in CHO cells (40), blocks the insulin-induced increase in recycling of the vp⅐TR chimera (Fig. 10A).
The Recycling of the vp⅐TR Chimera Is Stimulated by GTP␥S-In permeabilized cells, GTP␥S induces the translocation of GLUT4 to the surface (41,42). We examined the effects of insulin and GTP␥S on the recycling of the vp⅐TR chimera reconstituted in permeabilized CHO cells. We have previously shown that GTP␥S does not affect TR recycling reconstituted in semi-permeabilized CHO cells (31). Both GTP␥S and insulin increased the recycling rate of the vp⅐TR chimera in permeabilized cells by about 50%, and their effects on recycling were not additive (Fig. 10B). Neither insulin, GTP␥S, nor insulin and GTP␥S together had a large effect on the recycling of the TR in semi-permeabilized CHO cells, providing evidence that the effect of insulin on vp⅐TR trafficking is not a consequence of alterations in general endocytic recycling. DISCUSSION In this report, we characterized the endocytic trafficking of chimeras that contain either the cytoplasmic domain of IRAP and the transmembrane and extracellular domains of the human TR (vp⅐TR), or the cytoplasmic and transmembrane domains of IRAP and the extracellular domain of the TR (vpTM⅐TR). When expressed in CHO cells, vp⅐TR and vpTM⅐TR, like the TR, are rapidly internalized, trafficked to the peri- centriolar, endocytic recycling compartment, and recycled back to the cell surface. The trafficking of the chimeras differ from that of the TR in that the chimeras are recycled 3-fold more slowly. We do not observe any differences in the trafficking of vp⅐TR and vpTM⅐TR, indicating that the transmembrane domain of IRAP does not contain information which regulates endocytic trafficking. Thus, the 109-amino acid cytoplasmic domain of IRAP contains information that promotes rapid internalization from the plasma membrane as well as information that slows transport from the endocytic recycling compartment to the plasma membrane. A chimera in which the dileucine at positions 76 -77 is mutated to di-alanine is rapidly recycled and rapidly internalized, demonstrating that this dileucine is required for slow recycling but not for rapid internalization. Insulin increases the recycling of vp⅐TR and vpTM⅐TR to near the rate observed for the wild type TR. These studies demonstrate that CHO cells possess the machinery that allows for the insulin-regulated dynamic retention of membrane proteins within the endocytic system.
Our data are consistent with vp⅐TR trafficking through the same endosomal recycling compartment as other more rapidly recycled molecules. This conclusion is based on the observation that vp⅐TR co-localizes in the peri-centriolar region of the cells with fluorescent-SM internalized from the plasma membrane. Previous work has shown that fluorescent-SM is recycled through the same peri-centriolar, endocytic recycling compart-ment as is the TR (26). Importantly, the vp⅐TR chimera remains co-localized with the bodipy-SM following nocodazoleinduced vesiculation of the peri-centriolar recycling compartment. This result argues against the chimera residing in a compartment adjacent to the recycling compartment that is not resolved as a distinct compartment in light microscopy, because if that were the case, vp⅐TR and bodipy-SM would not necessarily co-distribute when the compartments were vesiculated. The endogenous hamster IRAP is also in the peri-centriolar recycling compartment, providing evidence that the vp⅐TR chimera behaves like endogenous IRAP.
The steady-state localization of vp⅐TR in the peri-centriolar recycling compartment demonstrates that the rate-limiting step in the recycling of the vp⅐TR chimera is transport from this compartment, as has been previously shown for other recycled molecules (26,32). Our results show that different membrane proteins can be transported from the recycling compartment at different rates. A number of previous observations demonstrate that the kinetics of trafficking from the peri-centriolar, endocytic recycling compartment can be slowed. Transport of the TR from the recycling compartment is slowed when endosomes are alkalinized by pharmacological treatments or in acidificationdefective mutant cells (25,29,32). Inhibition of endosome acidification has a greater effect on the recycling of the TR than it does on the recycling of lipid, indicating that transport of the TR from the peri-centriolar, endocytic recycling compartment is slowed relative to lipid (32). Another example of regulated transport from the peri-centriolar, endocytic recycling compartment is the inhibition of TR recycling by binding to multimeric Tf complexes (43). In the case of the vp⅐TR chimera, a dileucine-based motif in the cytoplasmic domain of IRAP (a protein normally expressed in CHO cells) slows recycling of the vp⅐TR chimera, indicating that CHO cells dynamically retain certain proteins within the peri-centriolar, endocytic recycling compartment. These results are the first report of the regulated dynamic retention of a transmembrane protein within the endocytic system. Because the physiological role of IRAP is not known, it is not prudent to speculate on why IRAP is dynamically retained within the recycling compartment.
Di-leucine-based motifs have been shown to function as membrane protein trafficking motifs (see, e.g., Ref. 38). The molecular mechanism by which these motifs regulate membrane protein trafficking is not known, although they are believed to regulate targeting via a specific interaction with proteins that regulate formation of transport vesicles (e.g. coat and/or adaptin proteins). The di-leucine sequence required for the dynamic retention of vp⅐TR in CHO cells (positions 76 -77) is in a region of the cytoplasmic domain of IRAP (residues 55-82), which, based on microinjection studies, interacts with the cellular machinery that regulates IRAP and GLUT4 distribution in fat cells (44). Thus, the results are consistent with the di-leucine at 76 -77 being involved in insulin-regulated trafficking in both CHO and fat cells. GLUT4 has a di-leucine motif on its carboxyl cytoplasmic domain, which functions as an internalization motif but does not regulate recycling (7,8,9,30,45,46). As discussed below, it is not known whether IRAP and GLUT4 traffic through the same pathway in CHO cells; therefore, a comparison in CHO cells of the function of targeting motifs of GLUT4 to those of IRAP is not warranted at this time.
Two possible models accounting for the slow recycling of vp⅐TR are depicted in Fig. 11. In one model, insulin-regulated retention, flow of vp⅐TR into the recycling vesicles that transport TR and other rapidly recycled molecules back to the cell surface is impeded, thereby slowing recycling of the chimera. In this case, an interaction of the cytoplasmic sequences of IRAP (based in part on the di-leucine sequence at positions 76 -77) with proteins in the cytosol would prohibit the vp⅐TR chimera from efficiently entering recycling vesicles formed from the endocytic recycling compartment (Fig. 11A). This model is conceptually similar to the mechanism proposed for the dynamic retention of proteins in the TGN (see, e.g., Ref. 47). Insulin would disrupt the interaction between the cytoplasmic domain of IRAP and the cytosolic proteins, freeing the vp⅐TR chimera to enter the recycling vesicles and to be rapidly recycled to the cell surface (Fig. 11B). An alternative model, insulin-regulated budding, is that vp⅐TR is transported to the surface in vesicles that are slowly formed from the endocytic recycling compartment (Fig. 11C). These vesicles are different from those that transport TR and other rapidly recycled molecules back to the cell surface. In this model, the cytoplasmic sequences of IRAP would interact with proteins that concentrate the vp⅐TR chimera in the more slowly forming vesicles. This model requires the formation of two different types of vesicles from the pericentriolar, endocytic recycling compartment, and is conceptually similar to the formation of distinct vesicles from the TGN that are destined for the apical or basolateral membrane (49). In this model, insulin would increase recycling of the vp⅐TR chimera by increasing the rate of formation of the vesicles enriched in vp⅐TR (Fig. 11D).
Comparison between the Insulin-regulated Trafficking in CHO and Fat Cells-Our data demonstrate that CHO cells, like fat and muscle cells, have a slow endocytic recycling mechanism that is regulated by insulin. In CHO cells, insulin increases the recycling of the TR by 1.2-fold and the recycling of vp⅐TR by 2-3-fold. It is this large difference which shows that the chimera is trafficked by a specialized insulin-regulated mechanism, because if the effects of insulin on the trafficking of the vp⅐TR chimera were due to effects on general membrane flow, then insulin would not have a greater effect on the recycling of vp⅐TR than on the recycling of the TR. In 3T3-L1 adipocytes, insulin has a small effect on the trafficking of the TR (ϳ2-fold increase in recycling) compared with its large effect on GLUT4 and IRAP (ϳ9-fold increase) (see, e.g., Refs. 3, 6, and 18). As is the case in CHO cells, this large difference in response to insulin is one feature which establishes that GLUT4 and IRAP are trafficked by a specialized mechanism. In addition to these kinetic similarities between the pathways in fat and CHO cells, two observations suggest that the molecular mechanisms of these insulin-regulated pathways are similar. 1) Wortmannin inhibits the insulin-stimulated recycling of the vp⅐TR chimera in CHO cells as well as the insulin-stimulated translocation of GLUT4 to the cell surface in fat cells (see, e.g., Ref. 39). 2) GTP␥S stimulates recycling of the vp⅐TR chimera in permeabilized CHO cells as well as the translocation of GLUT4 in permeabilized fat cells (41,42). Although these are consistent with mechanistic similarities in the insulin-regulated pathways in fat and CHO cells, inhibition of phosphatidylinositol 3Ј-kinase affects a number of trafficking pathways, including recycling of the TR (40,49,50), and the molecular mechanism underlying the effect of GTP␥S on translocation of GLUT4 is not known.
Although the parallels between insulin-regulated trafficking in CHO and fat cells are provocative, there are a number of important distinctions. 1) As noted above, the magnitude of the effect of insulin on the recycling of the chimera in CHO cells is small compared with the effect of insulin on the recycling of IRAP in fat cells. 2) In fat cells, IRAP appears to be segregated away from the general recycling pathway to an insulin-regulated pathway (53)(54)(55), indicating that the IRAP dynamic retention mechanism in fat cells is different from that in CHO cells.
Are GLUT4 and IRAP Trafficked by the Same Pathway in CHO Cells?-In fat and muscle cells, IRAP and GLUT4 are trafficked by the same insulin-responsive mechanism (15,16,18,19,44). Unlike IRAP, GLUT4 is not endogenously expressed by fibroblasts. When expressed in CHO cells, GLUT4, like vp⅐TR, is dynamically retained within cells in the absence of insulin, and slow recycling is the determining step in achieving this dynamic intracellular concentration (see, e.g., Ref. 51). However, a number of observations indicate that GLUT4 and the vp⅐TR may not be trafficked by the same pathway in CHO cells. Epitope-tagged GLUT4 expressed in CHO cells was found enriched in a population of small vesicles that are not enriched in TR (56). This is not in agreement with our findings that both vp⅐TR and hamster IRAP are in the general (TR-containing) recycling compartment. In addition, a number of studies have shown that GLUT4 expressed in CHO cells is not translocated to the surface by insulin (12,14,56,57), in contrast to our findings that insulin increases the recycling of vp⅐TR in CHO cells. Taken together, these findings indicate that GLUT4 and vp⅐TR are not trafficked by the same mechanism in CHO cells. In this regard, it has recently been shown in atrial cardiomyocytes that IRAP and GLUT4 do not completely co-distribute, indicating that in cell types other than CHO, IRAP and GLUT4 FIG. 11. Models for the dynamic retention of the vp⅐TR chimera in the peri-centriolar, endocytic recycling compartment of CHO cells. In the top panels, the insulin-regulated retention model is shown. A, in the basal state, traffic of vp⅐TR into recycling vesicles is inhibited by an interaction between its cytoplasmic domain and a cytoplasmic protein. The behavior of the LL76AA mutant vp⅐TR indicates that the di-leucine at positions 76 -77 is required for this interaction. B, insulin disrupts this interaction, allowing the vp⅐TR chimera to be recycled back to the plasma membrane in the same vesicles carrying the TR and other recycled molecules. In the bottom panels, the insulinregulated budding model is shown. C, in the basal state vp⅐TR is recycled to the plasma membrane by different vesicles than those that carry the TR and other rapidly recycled molecules. The di-leucine at positions 76 -77 would be involved in concentrating vp⅐TR into these vesicles. The vesicles that carry vp⅐TR form more slowly, in the basal state, than those that carry the TR. D, insulin increases the formation of the vp⅐TR-containing vesicles, thereby increasing recycling. The structure labeled ERC is the peri-centriolar endocytic recycling compartment; PM denotes the plasma membrane. may be differentially trafficked (54). Insulin regulation of GLUT4 surface expression in CHO cells has been observed when CHO cells when the insulin receptor is co-transfected with the epitope-tagged GLUT4 (57,58). It is not known why the insulin receptor must be overexpressed to detect an effect of insulin on GLUT4 distribution in CHO cells. The CHO cells we use have not been transfected with the insulin receptor; thus, the effect we see on the trafficking of vp⅐TR is based on the endogenous level of insulin receptor expression.
CHO Cells Possess Specialized Membrane Trafficking Pathways-A number of recent reports demonstrate specialized membrane trafficking pathways in CHO cells. In one report, a Ca ϩ -regulated secretion pathway was identified (59), and in another report, apical-and basolateral-like transport vesicles formed from the TGN of CHO cells were identified (60). Like the insulin-regulated trafficking we have observed, Ca ϩ -regulated exocytosis and polarized sorting from the TGN were thought to be characteristic of specialized cell types. Our findings, together with these other reports, indicate that CHO cells have specialized membrane trafficking mechanisms.