Proteasomes Mediate Prolactin-induced Receptor Down-regulation and Fragment Generation in Breast Cancer Cells*

Prolactin regulates a variety of physiological processes, including mammary gland growth and differentiation, and recent findings support an important role in breast cancer development and progression. However, little is known about the trafficking of its receptor, a member of the cytokine receptor superfamily. In the present study, we examined the effect of ligand on the endogenous “long” isoform of the prolactin receptor in breast cancer cells. We found that prolactin caused rapid and prolonged down-regulation of this receptor. The prolactin-induced increase in degradation was blocked by inhibitors of both proteasomes and lysosomes. However, the ubiquitin-conjugating system was not required for internalization. Prolactin also resulted in the concomitant appearance of a cell-associated prolactin receptor fragment containing the extracellular domain. This latter process required proteasomal, but not metalloprotease, activity, distinguishing it from ectodomain “shedding” of other membrane receptors, which are secreted as binding proteins. The prolactin receptor fragment was labeled by surface biotinylation and independent of protein synthesis. Together, these data indicated that prolactin binding initiates limited proteasomal cleavage of its receptor, generating a cell-associated fragment containing the extracellular domain. Our findings described a new potential mediator of prolactin action and a novel mechanism whereby proteasomes modulate cellular processes.

Prolactin regulates a variety of physiological processes, including mammary gland growth and differentiation, and recent findings support an important role in breast cancer development and progression. However, little is known about the trafficking of its receptor, a member of the cytokine receptor superfamily. In the present study, we examined the effect of ligand on the endogenous "long" isoform of the prolactin receptor in breast cancer cells. We found that prolactin caused rapid and prolonged down-regulation of this receptor. The prolactin-induced increase in degradation was blocked by inhibitors of both proteasomes and lysosomes. However, the ubiquitin-conjugating system was not required for internalization. Prolactin also resulted in the concomitant appearance of a cell-associated prolactin receptor fragment containing the extracellular domain. This latter process required proteasomal, but not metalloprotease, activity, distinguishing it from ectodomain "shedding" of other membrane receptors, which are secreted as binding proteins. The prolactin receptor fragment was labeled by surface biotinylation and independent of protein synthesis. Together, these data indicated that prolactin binding initiates limited proteasomal cleavage of its receptor, generating a cell-associated fragment containing the extracellular domain. Our findings described a new potential mediator of prolactin action and a novel mechanism whereby proteasomes modulate cellular processes.
The hormone/cytokine prolactin (PRL) 4 is critical for development and differentiation of the mammary gland. It also participates in numerous other processes, including behavior, metabolism, immune function, and osteogenesis (for reviews, see Refs. 1 and 2). Although the kinase cascades and alterations in gene expression that are initiated by PRL binding to its receptors have received considerable study (for reviews, see Refs. 3 and 4), relatively little is known about the effect of ligand on subsequent receptor trafficking, an important determinant of cellular responsiveness. The actions of PRL are mediated by the PRL receptor (PRLR), a member of the class I cytokine receptor superfamily. Alternative splicing results in isoforms of differing cytoplasmic domains (5). However, the so-called "long" isoform (lPRLR), which transmits the most signals, is the best-studied. Ligand binding to the lPRLR initiates internalization (6 -8), which is dynamin-dependent and is at least partially mediated by clathrin in a COS-7 cell model (7). Several cytoplasmic motifs distinct for the lPRLR (7) and short isoforms (7,9) have been identified that are critical for this process. Although PRL has been observed to either up-or down-regulate its receptor in complex in vivo systems, exogenous PRL reduced receptor levels in more defined experimental systems in vitro, a process blocked by inhibitors of lysosomal function (for review, see Ref. 10). Recently, PRL was reported to increase PRLR interaction with SCF ␤-TrCP E3 ubiquitin ligase, increasing PRLR ubiquitination and thereby destabilizing the receptor (11). However, the role of ubiquitination in ligand-stimulated PRLR endocytosis and proteasomes in receptor down-regulation was not examined.
For several membrane receptors, including the closely related growth hormone receptor (GHR), ligand binding has been shown to initiate proteolytic cleavage(s) of the receptor (for reviews, see Refs. [12][13][14][15]. The resulting extracellular domain (ECD) is released into the extracellular milieu, where it competes for ligand binding to cell-associated receptors. In some cases, receptor fragments have been reported intracellularly, where they can potentially serve as signaling molecules (for reviews, see Refs. [13][14][15]. Analogous events have not been reported for the PRLR, although PRL binding proteins in serum and milk have been observed (16 -18), and a 62-kDa chromatin-associated protein that binds PRL was identified in Nb2 cells (19). The importance of these events in modulating signal transduction, surface receptor levels, and therefore cellular responsiveness, as well as generation of potential new mediators, underscores the need to understand these processes and their regulation and the basis of receptor/ligand/cell specificity.
In addition to its role in regulation of physiologic mammary function, PRL also plays an increasingly recognized role in the pathogenesis and progression of breast cancer (for reviews, see Refs. 4,20,and 21). A majority of human tumors express the PRLR (22)(23)(24), and some studies have found higher PRLR levels in tumors than in adjacent normal tissue (25). Mammary epithelial cells are exposed to both circulating PRL as well as endogenously synthesized PRL, which is particularly prominent in primates (for review, see Ref. 4). This local synthesis complicates the study of ligand-induced responses in experimental in vitro systems and also increases PRLR expression (26), convoluting studies of membrane receptor trafficking subsequent to ligand binding. We have derived cell lines from MCF-7 breast cancer cells that are deficient in endogenous PRL production (27), permitting us to examine the effect of exogenous ligand on endogenously expressed lPRLR in clinically relevant target cells rather than transiently transfected model systems. Here we have demonstrated that proteasomal activity is critical for the PRL-initiated dramatic lPRLR down-regulation and generation of a PRLR fragment containing the ECD, although the ubiquitin-conjugating system is not required for internalization. This fragment was derived from surface lPRLR and was not dependent on translation, consistent with postinternalization processing of the lPRLR. Our findings showed the importance of the proteasomal system in mediating ligand effects on the lPRLR and suggest a novel mechanism for PRL signaling.

EXPERIMENTAL PROCEDURES
Reagents-Recombinant human PRL (lot AFP9042) was obtained through the National Hormone and Pituitary Program, NIDDK, National Institutes of Health, and Dr. Parlow. Bovine PRL (U. S. Department of Agriculture (USDA) bPRL B-1, AFP 5300) was obtained from the Animal Hormone Program of the USDA Reproduction Laboratory (Beltsville, MD), and recombinant bovine placental lactogen was a gift from the Monsanto Co. (St. Louis, MO). Protein G-Plus agarose was purchased from Santa Cruz Biotechnology, Inc., Santa Cruz, CA. Antibodies were purchased as follows: anti-PRLR-extracellular domain (ECD, catalog number 35-9200), Zymed Laboratories Inc., South San Francisco, CA; anti-hemagglutinin epitope (HA.11, catalog number MMS-101R) and anti-ubiquitin, Covance Research Products, Inc., Denver, PA); and anti-V5 (catalog number R960-25), Invitrogen. Alexa Fluor-conjugated human transferrin, cholera toxin subunit B, and goatanti-mouse antibody were purchased from Molecular Probes (Eugene, OR). Anti-GHR cyt-AL47 was a gift from Dr. S. J. Frank (28). The ubiquitinated protein enrichment kit was obtained from Calbiochem, and the kit for biotinylation was from Pierce. All inhibitors and culture media were purchased from Sigma with these exceptions: lactacystin and bisindolylmaleimide II and ␥-secretase inhibitor IX (Calbiochem) and matrix metalloproteinases inhibitor GM6001 (Chemicon International, Inc., Temecula, CA).
Plasmids-The expression plasmid for the human long PRLR isoform was a gift from C. Clevenger (29). MT123 (HA-Ub) (30) was used to assess receptor-associated ubiquitination. The HA epitope, YPYD-VPDYA, was introduced between the signal peptide and the first amino acid of the mature bovine lPRLR as follows. First, 30 cycles of PCR were used to synthesize two fragments overlapping with the HA sequence: Fragment 1, starting from a unique 5Ј EcoRI site through the signal peptide, ending with the HA tag with two oligonucleotides, 5Ј-GTGC-TGGAATTCCGGGCAAATG-3Ј, 5Ј-GGCGTAATCGGGCACGTC-ATAGGGGTATCCATTCAGAAGGCTGACAC-3Ј; Fragment 2, starting with the HA tag followed by the first amino acid of the mature PRLR to the unique BstEII site with two other oligonucleotides, 5Ј-TA-CCCCTATGACGTGCCCGATTACGCCCAGTCACCTCCTGAA-AAACC-3Ј, 5Ј-GGATGCAGGTCACCATGCTA-3Ј. Second, five cycles of PCR supplied with dNTPs were used to anneal and fill the ends of the two fragments. Third, 25 cycles of PCR supplied with two outside oligonucleotides containing unique sites were used to amplify the whole fragment. The PCR product was digested with EcoRI and BstEII and then substituted into the wild-type PRLR with the same restriction enzymes.
MCF-7 and NIH-3T3-F442A Cell Culture-PRL-deficient cells derived from the human mammary adenocarcinoma cell line, MCF-7, were grown in RPMI 1640 containing 10% horse serum and 50 M ganciclovir as reported previously (27). NIH-3T3-F442A cells were a gift from Dr. P. Bertics; these cells were grown in Dulbecco's modified Eagle's medium F12 containing 10% fetal bovine serum. For some experiments, 10 6 cells/60-mm plate were incubated in serum-free medium or serum-free medium supplemented with 0.5% bovine serum albumin (MCF-7 and 3T3-F442A, respectively), for 24 -48 h prior to treatments. Cells were then pretreated with vehicle or inhibitor for 30 min or 1 h followed by the addition of vehicle, 4 nM PRL, or 1 g/ml 4␤-phorbol 12-myristate 13-acetate (PMA) for an additional incubation. In some experiments, media were harvested, floating debris were removed by centrifugation, and protein was precipitated by incubation with 3 volumes of acetone at Ϫ20°C.
Ubiquitination-To examine ubiquitin associated with the PRLR, 3.9 ϫ 10 6 Chinese hamster ovary cells were transfected with the HA-Ub expression vector together with the human lPRLR expression plasmid using SuperFect reagent (Qiagen Inc., Santa Clarita, CA). Cells were grown in serum-free media overnight after the removal of transfection reagent and then treated with or without 4 nM PRL for 10 min and rinsed twice with PBS and scraped into immunoprecipitation assay buffer (10 mM Tris, pH 8, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.5% sodium deoxycholate, 0.05% SDS, containing 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, 1 g/ml aprotinin, and 5 mM N-ethylmaleimide). The cell extract was centrifuged at 4°C, 14,000 rpm for 10 min, and 1 mg of the supernatant was precleared by incubating with 16 l of Protein G-Plus agarose and 0.25 g of normal mouse IgG at 4°C for 1 h. Cell lysate was centrifuged at 2,500 rpm for 5 min at 4°C, and the supernatant was collected and incubated with V5 antibody (1:1000) at 4°C for 1 h, and 20 l of Protein G-Plus-agarose was added and incubated overnight. The pellet was collected by centrifugation at 2,500 rpm for 5 min at 4°C and washed once with ice-cold immunoprecipitation assay buffer and two times with ice-cold PBS and then resuspended in 30 l of 2ϫ sample buffer prior to examination by Western analysis as below. To investigate PRLR ubiquitination using another approach, we employed the GST⅐UIM kit (Calbiochem). 5 ϫ 10 6 PRL-deficient MCF-7 cells were serum-starved for 24 h and then treated with or without 4 nM PRL for 10 min. Cells were washed twice in PBS and harvested in cell lysis buffer (25 mM Tris, pH 8.0, 2 mM EDTA, pH 2.0, 10% glycerol, 1% Triton-X-100, 2 mM Na 3 VO 4 , 20 mM NaF, and 5 mM N-ethylmaleimide). 1 mg of total protein was incubated with either the UBA domain of yeast Rad23 protein conjugated to glutathione (31) or control glutathione alone, immobilized on agarose beads for 3 h at 4°C. Samples were then centrifuged at 4°C for 3 min at 1000 ϫ g. Pellets were washed three times and resuspended in 2ϫ sample buffer, boiled for 5 min, and analyzed by Westerns.
Indirect Immunofluorescence-PRL-deficient MCF-7 cells were grown as subconfluent monolayer cultures on glass coverslips overnight. To examine internalization of the lPRLR, cells were transfected with the HA-lPRLR expression plasmid. After the removal of the transfection reagent, cells were grown in serum-free RPMI 1640 overnight. Cells were treated with inhibitors for 20 min and then cooled on ice for 15 min, prior to incubation with antibody HA.11 (1:500) together with the same inhibitor in the blocking media (RPMI 1640 with 1% bovine serum albumin) on ice for 2 h. Cells were then washed twice with cold PBS and incubated with prewarmed RPMI 1640 with or without 8 nM PRL for 0, 5, or 15 min. Cells were washed with PBS, fixed with 4% paraformaldehyde/PBS, and permeabilized with 1% Triton/PBS. Cells were incubated at 4°C with Alexa Fluor 568-conjugated goat-antimouse antibody (1:400) overnight and then washed five times with PBS, fixed with 4% paraformaldehyde/PBS at room temperature for 15 min, washed two times with PBS, and mounted using VECTASHIELD mounting media (Vector Laboratories, Burlingame, CA). A Zeiss Axiovert 200 motorized fluorescence microscope with deconvolution (Axiovision 3.1) and cooled CCD camera (Zeiss Axiocam HRm) was used to capture the images.
Role of the Ubiquitin-conjugating System in Internalization-The Chinese hamster lung cell line (ts20) and ts20-GHR cells, which stably express the rabbit GHR, were gifts from Ger J. Strous (Utrecht, the Netherlands). lPRLR was transfected into ts20 cells to generate the stable cell line, ts20-lPRLR. Ts20 cells were grown in minimal essential medium ␣ containing 10% heat-inactivated fetal bovine serum, penicillin, streptomycin, and 450 g/ml G-418 at 30°C in 5% CO 2 . For internalization studies, 3 ϫ 10 5 cells were plated in 24-well plates and grown overnight at 30°C. Cells were washed once with 1 ml of binding media (Dulbecco's modified Eagle's medium/F12 supplemented with 10 mM MgCl 2 , 1 mM CaCl 2 , 1% bovine serum albumin, 1% penicillin-streptomycin, pH 7.6) followed by incubation in prewarmed binding media containing I 125 bovine placental lactogen (150,000 cpm, about 0.34 nM of 30 Ci/g of labeled protein, near the K d of these receptors), with or without a 300-fold excess (180 nM) of unlabeled PRL or bovine placental lactogen in 0.3 ml of binding media for 0, 30, 60, or 90 min at permissive (30°C) or non-permissive (42°C) temperature. Cells were then washed three times with 1 ml of ice-cold binding medium. Cell surface bound and the internalized ligand were harvested as described previously (7). Specific binding in each fraction was determined as the difference between 125 I ligand detected in the presence and in the absence of an excess of unlabeled hormone. The internalization ratio was expressed as a percentage of the specific internalized fraction with respect to total specific binding at 90 min as described (33).
Biotinylation-Cell surface proteins were biotinylated using a cell surface protein labeling kit (Pierce) with some adaptations. PRL-deficient MCF-7 cells were plated at a density of 5 ϫ 10 6 in 100-mm Petri dishes and serum-starved overnight. Cells were washed twice in icecold PBS and incubated at 4°C in 200 mg/ml sulfo-NHS-SS biotin for 30 min. Following removal of unbound biotin with quenching solution and two washes with ice-cold PBS, cells were stimulated with 4 nM hPRL at 37°C for 2 h. Lysates were harvested with the buffer provided in the kit supplemented with 1 g/ml leupeptin, 1 g/ml aprotinin, and 1 mM Na 3 VO 4 . Lysates were cleared by centrifugation, and 1 mg of total protein was incubated with 100 l of streptavidin-Sepharose (50%, v/v, in lysis buffer) for 1 h at room temperature. Pellets were washed three times, resuspended in 2ϫ sample buffer, boiled for 5 min, and analyzed by Western blot.

PRL Down-regulates the lPRLR in the PRL-deficient MCF-7 Cells-
Receptor down-regulation permits peptide ligands to regulate availability of their receptors on the cell surface. However, some target cells, including breast cancer cells, express PRL endogenously, making it difficult to observe the effect of circulating ligand on the PRLR. Using a genetic selection method, we derived cell lines from MCF-7 cells that are deficient in endogenous PRL production (27), which allowed us to examine the effect of exogenous PRL on the PRLR. lPRLR is the predominant isoform expressed in this cell line; other isoforms are not detected using an antibody to the shared  ECD (data not shown), consistent with previous reports of the parent cell line (34). As shown in Fig. 1, A and B, exogenous PRL rapidly down-regulated the lPRLR, which remained low through at least 24 h. This downregulation was dependent on ligand concentration, as shown in Fig. 1C. However, lower concentrations of PRL were able to elicit substantial lPRLR down-regulation after longer incubations (Fig. 1, compare D with A).
The lPRLR Undergoes Rapid Turnover-Levels of receptor expression are determined by the balance between synthesis and degradation.
To determine the effect of ligand on turnover of the lPRLR, we used cycloheximide to inhibit protein synthesis. In the absence of PRL, cycloheximide rapidly decreased cellular lPRLR levels with a half-life of about 100 min, indicating high turnover of the lPRLR even in the absence of ligand (Fig. 2). In the presence of PRL, the apparent half-life of the lPRLR fell to ϳ25 min. This rate of degradation of endogenous lPRLR is comparable with surface receptors in rabbit mammary explants (35) and in Chinese hamster ovary cells stably expressing the rabbit PRLR (6), suggesting that rapid turnover of the lPRLR is common across cell type and experimental system.
Down-regulation of the lPRLR by PRL Is Mediated by Proteasomes and Lysosomes-Both proteasomes and lysosomes are important mediators of degradation of cellular proteins. To determine their respective roles in PRL-induced down-regulation of the lPRLR, we used pharmacological inhibitors. Cells were pretreated with proteasomal inhibitors with distinct mechanisms of action (MG132 or lactacystin) or the lysosomal inhibitor, chloroquine, for 1 h followed by the addition of PRL or vehicle for an additional 2 h. As shown in Fig. 3, A and B, inhibition of both processes reduced PRL-induced down-regulation of the lPRLR.
Over this time frame, these inhibitors had little effect on lPRLR levels in the absence of ligand.
The lPRLR and/or Associated Proteins Are Ubiquitinated-The involvement of proteasomes in ligand-induced degradation of the lPRLR suggested ubiquitination of the receptor or associated proteins. Ubiquitination of the PRLR was initially demonstrated in a baculovirusinsect system (36), and more recently, in transiently transfected mammalian cells (11). To examine the endogenously expressed lPRLR in MCF-7 cells for this modification, we employed a GST⅐UIM fusion immobilized on agarose beads. As shown in Fig. 3C, no ubiquitination of the endogenously expressed lPRLR was evident with this approach; longer exposures did not reveal additional signals. However, this method favored capture of polyubiquitinated proteins, so monoubiquitinated receptor may not be detected. To examine the lPRLR for this modification in a more sensitive system, we transfected HA-tagged ubiquitin (HA-Ub) and V5-tagged lPRLR constructs into Chinese hamster ovary cells. As shown in Fig. 3D, ubiquitin associated with the lPRLR both in the presence and in the absence of ligand. These data did not discriminate between ubiquitination of the lPRLR itself and that of associated proteins.
Ubiquitin-conjugating System Is Not Required for Internalization of the lPRLR-The ubiquitin-conjugating system has been shown to be required for internalization of the closely related GHR (37), as well as other membrane receptors (38). The Chinese hamster lung cell line ts20 (39), which contains a thermolabile ubiquitin-activating enzyme E1, has proven useful for these studies (37, 40, 41). We generated a stable line expressing the lPRLR (ts20-lPRLR) and examined the role of the ubiq- Asterisks denote significant differences between vehicle and PRL-treated samples using the Student's t test, p Ͻ 0.05. C, serum-starved MCF-7 cells were treated Ϯ4 nM PRL for 10 min. Cell lysates were collected and incubated with either UIM⅐GST or GST alone conjugated to agarose beads, and complexes were examined by Western analysis with for the lPRLR (left). The blot was reprobed with anti-ubiquitin as a control (right). IB, immunoblot. D, ubiquitination of lPRLR and/or associated proteins. Chinese hamster ovary cells were transfected with the V5-lPRLR and HA-Ub, serum-starved for 24 h, and then treated with or without 4 nM PRL for 10 min, prior to immunoprecipitation (IP) of lysates with anti-V5 followed by Western analysis with HA.11. the Western blot was reprobed with anti-V5 to identify the location of the lPRLR. uitin-conjugating system in internalization of the lPRLR by modulating the thermolabile E1 by culture at either the permissive temperature (30°C) or the non-permissive temperature (42°C). As shown in Fig. 4A  (left panel), incubating ts20-lPRLR cells at 42°C, inactivating ubiquitinactivating enzyme E1, did not inhibit lPRLR internalization. In contrast, these conditions inhibited GHR endocytosis in parallel experiments (Fig. 4A, right). Proteasomal function is also required for endocytosis of the GHR (33). To evaluate this requirement in ligand-induced internalization of the lPRLR, we transiently transfected HA-lPRLR into MCF-7 cells. Surface receptor was labeled by anti-HA on ice before permitting ligand-induced internalization at 37°C (Fig. 4B, left panel). PRL-induced endocytosis of the lPRLR was not blocked by MG132 treatment (Fig. 4B). Together, our findings indicated that ubiquitin-conjugating and proteasomal activities are not required for internalization of the lPRLR, in contrast to the GHR.
lPRLR Fragment Is Generated by Proteasomes in Response to PRL Treatment-When cells were treated with PRL, a fragment of about 50 kDa was detected with the antibody to the PRLR ECD, temporally correlated with down-regulation of the lPRLR. This fragment was evident as early as 30 min after PRL treatment (not shown), peaked at 1-2 h, and was virtually undetectable after 24 h (Fig. 5A). To understand how this fragment was generated, cells were treated with proteasomal or lysosomal inhibitors. As shown in Fig. 5B, this fragment did not appear in cells treated with MG132 or lactacystin but was unaltered by chloroquine, suggesting that it is generated from the lPRLR by proteasomes in response to PRL. However, treatment with chloroquine resulted in a modest accumulation of this fragment (3.95 Ϯ 1.05-fold above vehicletreated PRL-induced controls, mean Ϯ S.D., n ϭ 3 independent experiments), consistent with eventual degradation in lysosomes. We used several methods to examine the population of PRLR that undergoes this limited proteolysis. To confirm that it did not occur cotranslationally on polysomes, as reported for Nf-B (for review, see Ref. 42), cycloheximide was used to inhibit translation. As shown in Fig. 5C, although this compound reduced levels of full-length receptor, as expected from the short half-life shown in Fig. 2, it did not alter the ability of ligand to induce the fragment. To determine whether the fragment was generated from receptor that had been on the plasma membrane, surface proteins were labeled with biotin, purified, and then examined by Western analysis. As shown in Fig. 5D, the PRLR fragment was precipitated by streptavidin, consistent with this origin.
Metalloproteases Do Not Play a Role in PRLR Fragment Generation, nor Is the PRLR Fragment Secreted into the Media-Because of the key role of metalloproteases in "shedding" of the ECD of multiple transmembrane proteins, including the GHR, interleukin-6 receptor, L-selectin, as well as others (for reviews, see Refs. 12 and 43-45), we examined components of this pathway to ascertain any role in generation of this PRLR fragment using several approaches. GM6001, a broad spectrum hydroxamic acid inhibitor of metalloproteinases, including tumor necrosis factor-␣-converting enzyme (TACE, also known as "a disintegrin and metalloproteinase," ADAM-17), which catalyzes the cleavage of these other membrane receptors, had no effect (Fig. 6A). Inhibition of the ␥-secretase complex, which releases soluble intracellular domains of some of these same proteins, also failed to perturb this event (data not FIGURE 4. Ubiquitin-conjugating system and proteasomes are not involved in internalization of the lPRLR. A, ts-20 cells stably expressing the lPRLR (ts20-lPRLR, left panel) or the rabbit GHR (ts20-rabGHR, right panel) were incubated at either the permissive temperature (30°C) or the non-permissive temperature (42°C) for 1 h. 125 I bovine placental lactogen (150,000 cpm) with or without excess non-labeled ligand was added to the media, and the incubation was continued for the indicated time at the same temperature. Cells were harvested, and surface-bound and intracellular ligand was determined as described under "Experimental Procedures." Internalization ratios were calculated as the internalized fraction divided by total cell-associated binding at 90 min. Each point represents the mean Ϯ S.E. of triplicate measurements from at least 3 independent experiments. When not shown, S.E. bars are smaller than the symbol. B, PRL-deficient MCF-7 cells transiently transfected with HA-lPRLR were treated with vehicle or 20 M MG132, and internalization of the HA-lPRLR after 15 min incubation with PRL was observed by indirect fluorescence microscopy as described under "Experimental Procedures." Representative micrographs from 3 separate experiments are shown. shown). Activation of protein kinase C family members with PMA, which stimulates generation of a secreted ECD fragment from the GHR (for review, see Ref. 12), modestly reduced lPRLR levels in the absence of ligand. However, it did not induce this limited proteolysis in the absence of PRL, suggesting that protein kinase C modulated other aspects of receptor processing. Consistently, inhibition with bisindolylmaleimide II (GF109203X), which inhibits multiple protein kinase C isoforms (46), failed to block the appearance of this fragment. Epidermal growth factor, which initiates signals activating several of these cleavage enzymes, failed to induce detectable cleavage (data not shown). Furthermore, examination of the media for polypeptides containing the extracellular domain of the lPRLR revealed the accumulation of immunoreactive protein of the appropriate size but no regulation by PRL (Fig. 6B) or PMA (data not shown). Comparison of the levels of this fragment secreted in 24 h by 10 6 cells to that associated with 10 5 cells after 2 h of exposure to ligand (lanes 1 and 2 versus lanes 3 and  4) demonstrates that the relative amount of lPRLR "shed" by these cells into the media is very low. These findings demonstrated that the lPRLR cleavage event described here is quite different from metalloprotease-mediated release of the ectodomains of other receptors, such as the GHR. PMA-stimulated shedding of the extracellular domain of the GHR by ADAM-17-mediated proteolysis, which is inhibited by ligand (12), is reduced by GM6001, and the remaining cell-associated GHR "remnant" is stabilized by proteasomal inhibitors, such as MG132 (Fig. 6C) (28).

DISCUSSION
Binding of agonist to surface receptors initiates kinase cascades, modulating signaling pathways and gene expression within the tar-get cell. However, ligand can also alter trafficking of the receptor, in ways specific to the ligand, receptor isoform, and cell type. For many membrane receptors, ligand binding is followed by increased rates of internalization and processing. The pathways and subsequent events have immediate consequences for signal transduction (for reviews, see Refs. [47][48][49][50][51]. Our system, permitting study of endogenously expressed receptors rather than transiently transfected receptors, avoids complications that may be secondary to overexpression, such as a high proportion of improperly folded or mistrafficked receptors (for review, see Ref. 52). Here we demonstrated that PRL dramatically down-regulates the endogenous lPRLR in these breast cancer cells. This fall in lPRLR levels was temporally linked to the appearance of a cell-associated ECD-containing receptor fragment, which is dependent on proteasomal, but not lysosomal or metalloprotease, activities. These findings demonstrated complex effects of ligand on lPRLR trafficking and pointed to a novel role for proteasomes in limited proteolysis resulting in generation of a stable lPRLR cleavage product.
Although a role for lysosomes in degradation of the PRLR has been established (for review, see Ref. 10), the current studies indicate a role for proteasomes in the post-ligand binding fate of the lPRLR as well. Our data were compatible with the emerging story of many receptor tyrosine kinases and G-protein coupled receptors, which employ ubiquitination to direct trafficking in response to ligand, terminating in degradation in lysosomes (for reviews, see Refs. 38 and 53-57). Many details remain to be worked out, including the identity of the ubiquitinated moiety(s), the nature and location of the ubiquitin chains, the PRL-initiated signal(s) leading to this modification, and the relationship between this event(s) and lysosomal degradation. SCF ␤-TrCP has been implicated in ubiquitination of the lPRLR (11). Whether this is the only E3 ubiquitin ligase involved in ubiquitin-mediated trafficking of the lPRLR remains to be determined. Other E3 ubiquitin ligases have been implicated in ubiquitination of cytokine receptors, including p33 RUL for the erythropoietin receptor (58), and several, including c-Cbl and Nedd4 family members, mediate trafficking of the epidermal growth factor receptor by direct ubiquitination of the receptor or associated proteins (for reviews, see Refs. 57 and 59). Unlike the closely related GHR and another cytokine receptor, the interferon-␣ receptor, the lPRLR does not require the ubiquitin-conjugating system for internalization (37,60). Our current studies are consistent with our previous report of the requirement for multiple motifs within the cytoplasmic domain of the lPRLR for optimal internalization, as opposed to the single the Ub-dependent endocytosis motif of the GHR (7). Thus the PRLR appears to more closely resemble the erythropoietin receptor (61,62) and interleukin-2 receptor ␤ (41, 63), other members of the cytokine receptor family, in this regard.
Our studies also revealed that PRL-initiated proteasomal activity is critical for generation of a cell-associated 50-kDa PRLR fragment containing the ECD. The size of this fragment suggested that it results from a cleavage event(s) in the vicinity of the transmembrane domain. Although this PRLR-ECD fragment suggested a "sheddase," such as that reported for multiple other transmembrane receptors, including the cytokine receptors, GHR (12), the interleukin-6 receptor (64), ErbB4 (13), and Notch (65), the apparent regulation of this lPRLR proteolytic event is quite different. Cleavage of the lPRLR was insensitive to protein kinase C agonists and inhibitors, as well as inhibition of the metalloproteases implicated in secretion of these other receptor fragments. Interestingly, the etiology of this cell-associated PRLR-ECD fragment is strikingly similar to that reported for some membrane bound transcription factors (for reviews, see Refs. 42 and 66). Proteasomal cleavage of these transmembrane proteins occurs in polysomes or the endoplasmic reticulum. However, the origination of this lPRLR fragment from surface receptor, and independence from concurrent translation, were not consistent with these locations. Our data suggested that the lPRLR is cleaved after ligand-stimulated internalization, which would be a novel site for this limited proteasomal action. A proteasome-dependent fragment from the common ␤ (␤c) subunit of hematopoietic cytokine receptors displays intriguing similarities, although few details are known, suggesting that the PRLR may not be alone in this respect (67). Clearly, the identity and function of this PRLR fragment require further study. Whether it is merely an intermediate in degradation of the PRLR or transmits additional signals, such as that proposed for other mitogenic receptor fragments (for reviews, see Refs. [13][14][15], is under investigation. Although early reports offered tantalizing suggestions of PRLR translocation into the nucleus (68), this has remained controversial (69). Our data suggested that this possibility should be reevaluated.
The mechanism whereby PRL initiates PRLR cleavage is currently under study. Inhibition of multiple kinase cascades did not block formation of the PRLR-ECD fragment. 5 Whether PRL induces a conformational change in the lPRLR, permitting cleavage, or initiates this process by signaling pathways not examined remains to be determined. Little is known about PRL-initiated signals that may result in serine/threonine phosphorylation of the lPRLR and/or associated proteins. These modifications mediate many interactions important for post-ligand trafficking and facilitate ubiquitination of many membrane receptors (56,57). Indeed, phosphorylation of Ser-349 was required for PRLR recognition by SCF ␤-TrCP in vitro (11).
In conclusion, we have provided evidence that proteasomes mediate ligand-stimulated generation of an ECD-containing fragment and lPRLR degradation in breast cancer cells. Differences in PRLactivated signaling pathways (3,4) and endocytic motifs among the alternatively spliced PRLR isoforms (7,9) suggest that post-ligand trafficking of the short isoforms will be distinct from the lPRLR. Understanding these processes, their regulation, and their interrelationships with signaling cascades in different target cells will increase our knowledge of PRL actions in physiologic activities and provide the basis to modulate its actions in pathological conditions, including breast cancer.