Low Density Lipoprotein Receptor-related Protein-1 Promotes (cid:1) 1 Integrin Maturation and Transport to the Cell Surface*

Low density lipoprotein receptor-related protein-1 (LRP-1) mediates the endocytosis of multiple plasma membrane proteins and thereby models the composition of the cell surface. LRP-1 also functions as a catabolic receptor for fibronectin, limiting fibronectin accumulation in association with cells. The goal of the present study was to determine whether LRP-1 regulates cell surface levels of the (cid:1) 1 integrin subunit. We hypothe - sized that LRP-1 may down-regulate cell surface (cid:1) 1 by promoting its internalization; however, unexpectedly, LRP-1 expression was associated with a substantial increase in cell surface (cid:1) 1 integrin in two separate cell lines, murine embryonic fibroblasts (MEFs) and CHO cells. The total amount of (cid:1) 1 integrin was unchanged because LRP-1-deficient cells retained increased amounts of (cid:1) 1 in the endoplasmic reticulum (ER). Ex- pression of human LRP-1 in LRP-1-deficient MEFs re-versed the shift in subcellular (cid:1) 1 integrin distribution. Metabolic labeling experiments demonstrated that the precursor form of newly synthesized (cid:1) 1 integrin (p105) is converted into mature (cid:1) 1 (p125) more slowly in LRP-1-deficient cells. Although low levels of cell surface (cid:1) 1 integrin, in LRP-1-deficient

The ␤ 1 integrin subunit associates with multiple ␣-subunits to form transmembrane adhesion receptors for extracellular matrix proteins, including collagen, fibronectin, vitronectin, and laminin (1). Once present at the cell surface, mature integrins anchor the plasma membrane to the actin cytoskeleton and promote cell signaling (2,3). Integrin and growth factor-initiated cell signaling responses are integrated by the cell to regulate gene expression, cell migration, cell growth, apoptosis, and development (2). Altered cell surface integrin expression may be particularly important in cancer, impacting on various aspects of cancer metastasis (4,5). Understanding mechanisms that regulate the concentration of integrins in the plasma membrane is an important problem.
The ␤ 1 integrin subunit is synthesized as an 87-kDa polypeptide that undergoes glycosylation in the endoplasmic reticulum (ER) 1 and in the Golgi apparatus (6,7). In the ER, the most prevalent, incompletely glycosylated form of ␤ 1 has a mass of 105 kDa and is thus referred to as p105. Mature ␤ 1 has a mass of ϳ125 kDa (p125). In some cells, p105 is the primary form of ␤ 1 integrin identified; however, p105 is not found at the cell surface and cannot function in cell adhesion or cell signaling (6,8).
Many factors control maturation of ␤ 1 and its transfer to the cell surface, including the availability of ␣-subunits, growth factors such as transforming growth factor-␤ (TGF-␤), and the state of activation of Ras (9 -11). The membrane-bound protein chaperone, calnexin, associates with ␤ 1 in the ER, promoting integrin assembly but inhibiting integrin transfer to the cell surface (12,13). Other protein chaperones, including calreticulin and receptor-associated protein (RAP), may associate with ␤ 1 -containing integrins or integrin-based adhesion complexes; however, a role for these proteins in integrin maturation has not been defined (14,15). Talin, which is best known for its role in focal adhesion assembly, and HEMCAM/gicerin, an immunoglobulin superfamily protein, also may regulate ␤ 1 maturation (16,17).
LRP-1 is a receptor for over 40 soluble ligands, which undergoes rapid and constitutive endocytosis in clathrin-coated pits, delivering most bound ligands to lysosomes for degradation (18). In some cells, LRP-1 may also localize in caveolae/lipid rafts, where it functions in cell signaling (19,20). In addition to soluble ligands, LRP-1 mediates the endocytosis of other plasma membrane proteins, including the urokinase receptor (uPAR), tissue factor, and amyloid precursor protein (APP), and thereby down-regulates the plasma membrane levels of these proteins (21)(22)(23)(24)(25). Regulation of the concentration and activity of membrane proteins represents an indirect mechanism by which LRP-1 may control cell signaling. For example, by down-regulating cell surface uPAR in murine embryonic fibroblasts (MEFs), LRP-1 suppresses activation of the small GTPase, Rac1, and inhibits cell migration (22,26).
We previously demonstrated that LRP-1 binds fibronectin and mediates its endocytosis, limiting fibronectin accumulation in association with cell surfaces (27). LRP-1 also is an endocytic receptor for thrombospondin-1 (28). By modifying the composition of the extracellular matrix, LRP-1 may regulate processes such as cell adhesion and migration. In this study, we sought to determine whether LRP-1 regulates cell surface integrin expression. We hypothesized that LRP-1 may down-regulate basal levels of integrin subunits, at the cell surface, by facilitating integrin endocytosis. In support of this hypothesis, Czekay et al. (29) recently reported that plasminogen activator inhibitor-1 promotes ␣ v integrin endocytosis by forming an integrin-containing multiprotein complex that is recognized and internalized by LRP-1. In this report, we demonstrate the unanticipated finding that LRP-1 expression is associated with a substantial increase in cell surface ␤ 1 integrin. The effects of LRP-1 are observed principally in confluent cell cultures and reflect the ability of LRP-1 to directly or indirectly promote maturation of newly synthesized ␤ 1 integrin in the secretory pathway. Regulation of ␤ 1 integrin maturation represents a novel mechanism by which LRP-1 may influence interactions of the cell with its microenvironment.

EXPERIMENTAL PROCEDURES
Reagents and Proteins-Purified fibronectin, vitronectin and ␣ 5 ␤ 1 integrin were purchased from Chemicon International (Temecula, CA). Type I collagen was obtained from BD Biosciences (Palo Alto, CA). Glutathione S-transferase (GST)-RAP was expressed in bacteria and purified as previously described (27) using a construct obtained from Dr. Joachim Herz (University of Texas Southwestern Medical Center, Dallas, TX). As a control, GST without fused RAP, was also expressed and purified from bacteria transformed with the empty vector, pGEX-2T. Endoglycosidase H (Endo-H) and peptide-N-glycosidase F (PNGase F) were purchased from Roche Applied Science (Mannheim, Germany) and Sigma, respectively. Polyclonal antibody PAB1952, which recognizes the C-terminal cytoplasmic domain of ␤ 1 integrin, was obtained from Chemicon International (Temecula, CA). ␤ 1 integrin-specific polyclonal antibody 363 was kindly provided by Dr. Douglas DeSimone (University of Virginia). Polyclonal anti-epidermal growth factor receptor (EGFr) antibody was purchased from Upstate Biotechnology (Lake Placid, NY) and polyclonal anti-extracellular signal-regulated kinase (ERK/MAPK) antibody was from Zymed Laboratories Inc. (San Francisco, CA). TGF-␤1,2,3-neutralizing antibody 1D11 was from R&D Systems. The activity of this antibody was confirmed in endothelial cell growth assays, as previously described by our laboratory (30). Streptavidin-Sepharose, peroxidase-conjugated donkey anti-rabbit IgG and Protein A-Sepharose were from Amersham Biosciences. Trans 35 S-label, for metabolic labeling, was from ICN Biochemicals (Irvine, CA). The membrane-impermeable biotinylation reagent, sulfo-NHS-LCbiotin, was from Pierce. Cell culture media was from Invitrogen Life Technologies, Inc.
In some experiments, MEFs were cultured in serum-free medium in the presence of GST-RAP (200 nM) or an equivalent concentration of GST (negative control) for 3 days. The medium and GST-RAP were replaced daily. To test the effects of different plating substrata, MEFs were cultured in 6-well plates that were pre-coated with 20% FBS, 5 g/ml purified vitronectin, 10 g/ml fibronectin, 25 g/ml type I collagen or 10 g/ml poly-L-lysine for 2 h at 37°C. After precoating and before adding cells, the wells were blocked with 5 mg/ml bovine serum albumin (BSA) for 1 h. To block the function of ␣ 5 ␤ 1 , MEFs were pre-treated with 10 g/ml monoclonal antibody BMA5 (Chemicon).
Biotinylation and Recovery of Cell Surface Integrin Subunits-Monolayer cultures of MEFs and CHO cells were washed three times with ice-cold 20 mM sodium phosphate, 150 mM NaCl, pH 7.4 (PBS) to remove contaminating FBS and other soluble proteins and then treated with the membrane-impermeable biotinylation reagent, sulfo-NHS-LCbiotin (0.1 mg/ml), for 15 min at 22°C, as described previously (27). Biotinylation reactions were terminated by adding 50 mM Tris-HCl, 150 mM NaCl, 100 mM glycine, pH 7.5 for 15 min at 22°C. After washing with PBS, the cells were counted and lysed in extraction buffer. Biotinylated cell surface proteins were precipitated with Streptavidin-Sepharose (Amersham Biosciences). The affinity precipitates were recovered by centrifugation, washed, boiled in SDS sample buffer, and subjected to SDS-PAGE and immunoblot analysis to detect ␤ 1 integrin.
Endoglycosidase Digestion-For PNGase F treatment, MEF extracts were diluted to a concentration of 1 mg/ml in 100 mM Tris-Cl, pH 7.4 with protease inhibitors, 50 mM ␤-mercaptoethanol and 0.1% SDS. The samples were then boiled for 2 min. After neutralization with Triton X-100 (0.5%), the samples were incubated with PNGase F (0.025-2.5 units/ml) at 37°C for 12 h. A second dose of enzyme was added for an additional 12 h. For Endo-H treatment, MEF extracts were diluted to a concentration of 1 mg/ml in 100 mM sodium acetate, pH 5.5, containing 0.02% SDS and protease inhibitors. Samples were treated with Endo-H (30 milliunits/ml) at 37°C for 12 h and then with a second dose of Endo-H for an additional 12 h. Enzymatic reactions were terminated by protein precipitation with ice-cold acetone. Pellets were air-dried. The samples were then neutralized and resuspended in SDS sample buffer for SDS-PAGE and immunoblot analysis.
Metabolic-labeling of ␤ 1 Integrin-MEFs were seeded in 60-mm dishes in serum-containing medium, and allowed to grow until 100% confluent. Depletion of intracellular methionine was achieved by culturing in L-methionine/L-cysteine-free Dulbecco's modified Eagle's medium containing 5% FBS for 1 h. The FBS was predialyzed to remove associated L-cysteine and L-methionine. The cells were then labeled with 0.15 mCi/ml of TRAN 35 S-Label (which includes radioactive Lmethionine and L-cysteine) for 1 h in the same medium. After labeling, the cells were harvested immediately (time 0) or chased for the indicated times in DMEM supplemented with 10% FBS and excess of non-radioactive L-methionine/L-cysteine (2 mM). Cell extracts were prepared by boiling for 10 min in 1% SDS, 20 mM Tris-HCl, 150 mM NaCl (TBS), 5 mM EDTA. The extracts were diluted 1:10 in TBS with 1% (v/v) Triton X-100 and protease inhibitor mixture (Roche Applied Science), subjected to centrifugation, and precleared protein-A Sepharose (Amersham Biosciences). ␤ 1 integrin subunit was recovered by immunoprecipitation with polyclonal antibody 363. The precipitates were then subjected to SDS-PAGE on 6% gels. The gels were fixed, dried, and exposed to intensifying screens. Radiolabeled ␤ 1 integrin subunit was detected using a Storm 860 PhosphorImager and analyzed using Im-ageQuant software (Molecular Dynamics).
Cell Adhesion-Costar 96-well plates were coated with various concentration of fibronectin in PBS overnight at 4°C, rinsed, and then blocked with 2% (w/v) BSA (Sigma) in PBS for 2 h at room temperature. CHO K1, CHO 14-2-1, PEA10 and MEF-2 cells were harvested in calcium/magnesium-free PBS containing 0.5 mM EDTA. The cells were pelleted at 1000 ϫ g for 5 min and re-suspended at a density of 10 6 cells/ml in Puck's saline medium A, supplemented with 10 mM HEPES, pH 7.4, 0.5 mM CaCl 2 , 0.5 mM MgCl 2 , and, when indicated, 1 mM MnCl 2 . Cells were allowed to adhere for different periods of time at 37°C in a humidified atmosphere. Non-adherent cells were removed by washing with PBS. Adherent cells were fixed for 20 min with 4% formaldehyde in PBS, rinsed with PBS, and stained with 0.2% crystal violet in 2% ethanol for 30 min. Excess stain was washed away. Cell associated stain was then released in 1% SDS (50 l per well). The absorbance at 595 nm was measured. Each value represents the mean of 18 separate replicates, divided among three different experiments. To evaluate cell spreading, cells were allowed to adhere for 30 min and then photographed using an Axiovert microscope with a Contax camera.
B41 cells are MEF-2 cells that were transfected for stable expression of full-length human LRP-1. Expression of human LRP-1 in MEF-2 cells increased the fraction of p125 ␤ 1 integrin (Fig. 1B). In four separate experiments, p125/p125ϩp105 was increased by 2.5 Ϯ 0.3-fold in the B41 cells compared with MEF-2 cells (p Ͻ 0.05). To confirm the identity of p105 as an intracellular ␤1 precursor in the ER, we treated MEF extracts with Endo-H. This enzyme dissociates N-linked mannose-rich glycans, which become Endo-H-resistant after modification by glycosyltransferases in the Golgi apparatus. As anticipated, Endo-H totally eliminated p105, replacing this band with a new band that migrated near unglycosylated ␤ 1 core protein. p125 was not modified by Endo-H, confirming that this species is the mature form of the integrin subunit.
Because LRP-1-deficient MEF-2 cells have decreased amounts of p125, we hypothesized that these cells have decreased amounts of cell surface ␤ 1 integrin. To test this hypothesis, we biotinylated cell surface proteins in MEFs, using the membrane-impermeable biotinylation reagent, sulfo-NHS-LCbiotin. Biotinylated proteins were affinity-precipitated with streptavidin-Sepharose and probed for ␤ 1 integrin. As shown in Fig. 1C (representative of four separate experiments), the LRP-1-positive cells (MEF-1 and B41) had increased amounts of biotinylated ␤1 compared with MEF-2 cells. As anticipated, biotinylated p105 was never recovered in the affinity precipitates, confirming that p105 is not found on cell surfaces in MEFs.
We considered a number of previously described mechanisms whereby LRP-1 may indirectly control maturation of p105. TGF-␤ establishes an autocrine pathway in some cell cultures (33) and has been shown to promote ␤1-chain maturation (10,11). Furthermore, LRP-1 may function as a TGF-␤ receptor (34). To test whether alterations in endogenously produced TGF-␤ activity are responsible for differences in ␤ 1 integrin maturation in LRP-1-positive and -negative MEFs, we treated PEA10 and MEF-2 cells with TGF-␤-neutralizing antibody (20 g/ml) or with an equivalent concentration of preimmune mouse IgG for 12 h in serum-free medium. Fig. 1D shows that TGF-␤-neutralizing antibody had no effect on ␤ 1 distribution between p125 and p105 forms. In separate experiments, plating LRP-1-positive and -negative cells on different substrata, including vitronectin, fibronectin, type I collagen, and poly-Llysine (negative control) had no effect on the distribution of ␤ 1 integrin into the p125 and p105 bands (results not shown). Furthermore, ␣ 5 ␤ 1 function-blocking antibody (10 g/ml for 48 h) failed to alter p105 and p125 in LRP-1-positive and -negative MEFs. However, culture confluency did affect the distribution of ␤ 1 integrin into p125 and p105, and this result was only observed in LRP-1-positive cells.
We applied two strategies to test the effects of culture confluency on ␤ 1 integrin maturation. In one set of experiments, cells were plated at different densities and cultured for 24 h. By visual inspection, cells plated at 10 5 /well remained sub-confluent. Cells plated at 10 6 /well or 2 ϫ 10 6 /well were confluent; however, as anticipated, the cells plated at 2 ϫ 10 6 /well became confluent sooner. In a second set of experiments, cells were plated at the equivalent density (2 ϫ 10 6 /well) and analyzed after culturing for increasing periods of time. In both experiments, the fraction of ␤ 1 integrin migrating as mature p125 increased under conditions that favored development of cell FIG. 1. LRP-1 expression regulates ␤ 1 integrin in MEFs. A, equal amounts of cellular protein were recovered from detergent-soluble cell extracts of MEF-1, PEA10, and MEF-2 cells. The cell cultures were confluent at the time of extraction. Samples were subjected to immunoblot analysis for ␤ 1 integrin. p125 is labeled ␤1 mature. p105 is labeled ␤1 precursor. B, equal amounts of cellular protein from whole cell extracts of MEF-1 cells, MEF-2 cells, and B41 cells were treated with Endo-H for 24 h at 37°C (ϩ) or with vehicle (Ϫ). The samples were subjected to immunoblot analysis for ␤ 1 integrin. C, streptavidin affinity precipitates were subjected to immunoblot analysis with a specific ␤ 1 integrin antibody. D, PEA10 and MEF-2 cells were treated with TGF-␤-neutralizing antibody (20 g/ml) or with an equivalent concentration of preimmune mouse IgG for 12 h in serum-free medium. The samples were subjected to immunoblot analysis for ␤ 1 integrin. culture confluency only in LRP-1-expressing cells (Fig. 2).
The results presented thus far suggest a model in which LRP-1 regulates the relative abundance of ␤ 1 integrin glycoforms in MEFs. To confirm that the total amount of ␤ 1 integrin is not altered, we treated extracts of MEF-1, MEF-2, and B41 cells with PNGase F, which hydrolyzes aspartyl-glycosamine bonds, releasing nearly all N-linked glycans. As shown in Fig.  3A, PNGase F modified both p125 and p105, converting both species into a common product that migrated near the mobility of core protein. The amount of product observed with each of the three cell lines was essentially identical, supporting the conclusion that LRP-1 does not regulate the total amount of ␤ 1 integrin in the cell. In control experiments, we also determined that the total amount of EGFr (Fig. 3B) and ERK/MAPK (Fig.  3C) were equivalent in the MEF protein extracts. Thus, the LRP-1-dependent changes in p125 and p105 ␤ 1 integrin are specific.
LRP-1 Expression Regulates p125/Cell Surface ␤ 1 Integrin in CHO Cells-As a second model system to assess the effects of LRP-1 on the subcellular distribution of ␤ 1 integrin, we compared LRP-1-positive CHO K1 cells and LRP-1-deficient CHO 13-5-1 cells. We also studied CHO 14-2-1 cells, which have a mutation in either LRP-1 or an unidentified gene product that is necessary for LRP-1 transport in the secretory pathway so that, in this cell line, LRP-1 does not transfer to the cell surface. The CHO 13-5-1 and 14-2-1 cells were cloned from CHO K1 cells treated with the mutagen, ethyl methane sulfate, and selected based on resistance to Pseudomonas exotoxin A (32). As shown in Fig. 4A (left panel), the LRP-1-deficient CHO 13-5-1 cells demonstrated decreased p125 and increased p105, compared with CHO K1 cells. Thus, the effects of LRP-1 expression on ␤ 1 distribution into p105 and p125 were similar in CHO cells and MEFs. In the CHO 14-2-1 cells, p125 was almost entirely absent.
To confirm these results, cell surface ␤ 1 integrin was compared in the three CHO cell lines using our surface protein biotinylation method (Fig. 4A, right panel). Once again, only p125 was detected in streptavidin-affinity precipitates, as anticipated. The amount of biotinylated ␤ 1 integrin was reduced by 60% in CHO 13-5-1 cells, compared with CHO K1 cells, and to trace levels in CHO 14-2-1 cells (n ϭ 5, Fig. 4B). The results of our experiments with CHO 13-5-1 cells confirm, in a second model system, that cell surface ␤ 1 integrin is decreased when LRP-1 is not expressed. The profound results obtained with CHO 14-2-1 cells may suggest that either LRP-1 accumulation in the ER dramatically retards maturation of ␤1 integrin or that these cells are missing a gene product, which is necessary for transfer of both LRP-1 and ␤ 1 integrin to the cell surface. In separate experiments, we demonstrated that culture confluency promotes accumulation of p125 in CHO K1 cells but not in CHO 13-5-1, confirming the results obtained with MEFs (results not shown).
LRP-1 Expression Promotes ␤ 1 Integrin Maturation-To explain the results presented thus far, we hypothesized that LRP-1, either directly or indirectly, promotes maturation and transport of ␤ 1 integrin in the secretory pathway. An alternative hypothesis, which was considered less likely, was that LRP-1 stabilizes cell surface ␤ 1 integrin against catabolism and that in LRP-1-deficient cells, there is compensatory up-regulation of ␤ 1 integrin expression.
To test whether LRP-1 expression alters ␤ 1 integrin maturation in the secretory pathway, metabolic labeling experiments were performed in LRP-1-positive PEA10 cells and in LRP-1-negative MEF-2 cells. After pulse exposure to radioactive L-methionine/L-cysteine, radiolabeled p105 ␤ 1 integrin was detected in both cell types, as anticipated (Fig. 5). p105 ␤ 1 integrin disappeared more slowly in the LRP-1-negative MEF-2 cells, compared with the PEA10 cells; more than 70% of the p105 persisted in MEF-2 cells after 5 h. In the PEA10 cells, the mature form of ␤ 1 integrin (p125) was detected at an earlier time (2 h) and persisted over the 8 h time course as a substantially higher fraction of the total amount of ␤ 1 integrin (p125ϩp105). These results were interpreted as supporting the hypothesis that LRP-1 promotes maturation of ␤ 1 integrin in the secretory pathway.
We assumed that p105 disappeared, during the course of the metabolic labeling, due to conversion into p125; however, in all of our experiments, the amount of p125 generated was less

FIG. 2. Cell culture confluency promotes accumulation of p125 only in LRP-1-expressing cells.
A, PEA10 or MEF-2 cells were plated in 35-mm wells at different densities, as indicated below the lanes, and cultured under standard conditions for 24 h. The samples were then subjected to immunoblot analysis for ␤ 1 integrin. Densitometry was performed to quantitate p125 as a fraction of total ␤ 1 integrin (p105ϩp125) in this representative study (bar graph to the right). B, 2 ϫ 10 6 cells were plated in each well and cultured for the indicated times. The samples were then subjected to immunoblot analysis for ␤ 1 integrin. Densitometry was performed to quantitate p125 as a fraction of total ␤ 1 integrin (p105ϩp125) in this representative study (bar graph to the right).
than the initial amount of p105 detected. This result may suggest that p125 is degraded during the time course of our experiment, that some p105 is targeted for degradation instead of conversion into p125, or that our antibody is less effective for recovery of p125 by immunoprecipitation. In control experiments, total ␤ 1 integrin was assessed in MEF-1 cells by immunoblot analysis or isolated, from the same cells, by immunprecipitation and then subjected to immunoblot analysis. The fraction of p125 was significantly reduced when the immunoprecipitation step was included (results not shown). Thus, the limited amount of p125 detected in our metabolic labeling experiments is at least partially explained by the function of our antibody in immunoprecipitation.
To test whether LRP-1 alters the stability of cell surface ␤ 1 integrin, we cultured LRP-1-expressing MEF-1 cells in the presence of GST-RAP (200 nM) or an equivalent concentration of GST (negative control) for 3 days. RAP binds to the ligandbinding sites in LRP-1, blocking its interaction with all extracellular ligands and reversing its effects on the cell surface levels of uPAR and APP (24,35). In three separate experiments, RAP did not cause any detectable change in the level of cell surface ␤ 1 integrin or in the fraction of mature p125 (p125/ p125ϩp105) (results not shown). These results suggest that the function of LRP-1 as an endocytic receptor is not coupled to the mechanism by which LRP-1 expression alters p105 and p125. Similar experiments were carried out using LRP-1-expressing HT-1080 human fibrosarcoma cells and again, GST-RAP had no effect on the amount of mature cell surface ␤ 1 integrin.
Adhesion to Fibronectin Is Decreased in LRP-1-deficient MEFs and in CHO 14-2-1 Cells-Because cell surface ␤ 1 integrin was decreased in LRP-1-deficient cells, we compared attachment of PEA10 and MEF-2 cells to immobilized fibronectin. Decreased attachment of MEF-2 cells was observed at various fibronectin-coating densities (Fig. 6A). In wells that were precoated with fibronectin at 1.0 g/ml, the difference in attachment was significant at the p Ͻ 0.05 level.
Adhesion assays were also performed with CHO K1 and CHO 14-2-1 cells since the CHO 14-2-1 cells demonstrated the most profound decrease in cell surface ␤ 1 integrin. As shown in Fig. 6B, CHO 14-2-1 cells demonstrated a significant decrease in attachment to 1.0 g/ml fibronectin (p Ͻ 0.05). When the cells were allowed to attach in the presence of 0.5 mM MnCl 2 , to promote increased integrin activation, the difference related to LRP-1 expression was more profound (p Ͻ 0.01); under these conditions, significant differences in cell attachment were observed using multiple fibronectin coating concentrations.
Finally, when CHO K1 and CHO 14-2-1 cells were examined 30 min after addition to fibronectin-coated wells (1.0 g/ml), the CHO 14-2-1 cells demonstrated substantially decreased spreading (Fig. 6C). This effect was no longer observed when the fibronectin coating density was increased or the time of plating was increased beyond 1 h. DISCUSSION LRP-1 gene disruption in the mouse is embryonic lethal (36); however, in cell culture model systems, LRP-1 regulates cell signaling and cell physiology by diverse mechanisms. A unique property of LRP-1 is its ability to facilitate endocytosis of other plasma membrane proteins, including uPAR, APP, and tissue factor, and thereby down-regulate cell surface levels of these proteins (21)(22)(23)(24)(25). This process probably requires bifunctional ligands or adaptor proteins, such as Fe65, that bridge LRP-1 to other plasma membrane proteins so that the complexes undergo endocytosis as units (18,21,37). By decreasing the level of uPAR, LRP-1 indirectly controls cell signaling pathways leading to the activation of Ras, ERK/MAPK, and Rac1 (23,26). In mice, inactivation of the LRP-1 gene in vascular smooth muscle cells increases cellular levels of PDGF ␤ receptor and its degree of tyrosine phosphorylation (38). Thus, there is evidence that LRP-1 controls the activity of cell signaling receptors in vivo. LRP-1 may also directly regulate cell signaling through interactions involving its intracytoplasmic tail and proteins such as Shc, Disabled-1, and JNK-interacting protein (37,39,40).
The possibility that LRP-1 functions to regulate protein trafficking in the secretory pathway was first raised by Pietrzik et al. (41). These investigators demonstrated that LRP-1 stabilizes a 10-kDa proteolytic fragment of APP, which includes its transmembrane domain. The APP fragment was stabilized to a greater extent in CHO 14-2-1 cells, suggesting that LRP-1 in the ER plays a critical role. The results reported here further define the function of LRP-1 as a regulator of plasma membrane protein trafficking in the secretory pathway. We show, for the first time, that LRP-1 expression is associated with a major shift in the degree of maturation and subcellular localization of ␤ 1 integrin.
To determine the effects of LRP-1 on ␤ 1 distribution, we initially compared ␤ 1 integrin expression in LRP-1(ϩ/ϩ) MEF-1 cells, LRP-1(ϩ/Ϫ) PEA10 cells, and LRP-1-deficient MEF-2 cells. In each case, LRP-1 expression was associated with an increase in p125, a decrease in p105, and an increase in cell surface ␤ 1 integrin, as determined by affinity precipitation of biotinylated proteins. In addition, we were able to prove that a decrease in mature ␤ 1 integrin is associated with the absence of LRP-1, by using a gain-of-function model, B41 cells, in which MEF-2 cells are transfected for stable expression of full-length human LRP-1. We confirmed that p105 is localized in the ER, by its susceptibility to Endo-H. Furthermore, we demonstrated that the relationship between LRP-1 expression and ␤ 1 integrin maturation is not cell type-specific by analyzing a second model system, the CHO cell. Once again in this cell type, LRP-1 expression was associated with increased p125, decreased p105, and increased cell surface ␤ 1 integrin.
The ability of LRP-1 to promote accumulation of ␤ 1 integrin at the cell surface represents a novel mechanism by which LRP-1 may regulate cell adhesion, migration, and cell signaling. Our metabolic labeling experiments supported a model, in which LRP-1 promotes trafficking of ␤ 1 integrin from the ER to the cell surface; however, the results of the metabolic labeling experiments do not preclude other mechanisms by which LRP-1 may affect the level of cell surface ␤ 1 . To study whether LRP-1 facilitates ␤ 1 integrin endocytosis, we cultured cells for 3 days in the presence of GST-RAP. In previous studies, GST-RAP has been shown to increase cell surface levels of uPAR and APP (23,35); however, GST-RAP failed to affect levels of p125 or p105. Because the ␤ 1 integrin subunit is a component of multiple integrins, our RAP experiments may not be sufficiently sensitive to detect an effect of LRP-1 on one specific integrin. Furthermore, LRP-1-facilitated integrin endocytosis may require specific conditions, such as activation of uPAR in the presence of plasminogen activator inhibitor-1, as previously described (29).
The ability of LRP-1 to promote ␤ 1 integrin p125 formation was most profound in cultures that became confluent, suggesting that formation of cell-cell contacts facilitates transfer of ␤ 1 integrin to the cell surface in a LRP-1-dependent manner. An exception was the CHO 14-2-1 cell line, which demonstrated a profound defect in ␤ 1 integrin transport, coupled with changes Equal amounts of cellular protein from detergent-soluble cell extracts of CHO K1, CHO 13-5-1, and CHO 14-2-1 cells were subjected to immunoblot analysis for ␤ 1 integrin in the three lanes on the left. In the three lanes to the right, cell surface proteins were biotinylated and streptavidin affinity-precipitated prior to immunoblot analysis for ␤ 1 integrin. B, densitometry was performed to determine the relative intensities of p125 and p105. The fraction, p125/p125ϩp105, obtained from five separate replicates, is shown (mean), standardized against that observed in wild type CHO K1 cells.
FIG. 5. ␤ 1 integrin maturation as determined by metabolic labeling. A, PEA10 and MEF-2 cells were pulse-exposed to TRAN 35 S-Label for 1 h. Cell extracts were immediately prepared (0 h) or the cultures were maintained for 2, 5, or 8 h in Dulbecco's modified Eagle's medium supplemented with 10% FBS and excess non-radioactive Lmethionine/L-cysteine. ␤ 1 integrin was recovered by immunoprecipitation with polyclonal antibody 363 and analyzed by SDS-PAGE. 35 Slabeled ␤ 1 the integrin subunit was detected using a Storm 860 PhosphorImager. B, the amount of ␤ 1 integrin precursor p105 at various times is shown as a fraction of that present at time 0. The results of two equivalent experiments were averaged to generate the graph. Similar results were obtained in separate experiments using PAB1952 to immunoprecipitate ␤ 1 integrin.
in cell adhesion and spreading, under all cell culture conditions, despite LRP-1 expression. Because LRP-1 is expressed but retained in the secretory pathway, in CHO 14-2-1 cells, the profound defect in ␤ 1 integrin transport may be explained if LRP-1 associates with ␤ 1 integrin in the ER, either directly or indirectly, through a bifunctional adaptor protein/chaperone. We were not able to support the hypothesis that ␤ 1 integrin associates with LRP-1 by co-immunoprecipitation analysis. Furthermore, when we overexpressed a candidate LRP-1 chaperone, RAP, in CHO K1 cells, CHO 14-2-1 cells, MEF-1 cells, and MEF-2 cells, high levels of RAP expression were demonstrated by immunoblotting; however, p125 and p105 were unchanged (results not shown). Other proteins that could bridge LRP-1 to ␤ 1 integrin include calreticulin, Fe65, Hsp90, and ICAP-1 (37,(42)(43)(44)).
An alternative model to explain the defect in ␤ 1 integrin maturation, in CHO 14-2-1 cells, is that both LRP-1 and ␤ 1 integrin require the same gene product for proper post-translational trafficking, which is absent in this mutagen-treated cell line. Boca/MESD is a recently described ER chaperone that plays a critical role in the transport of multiple LDL receptor family members through the secretory pathway (45,46). At this time, it is not known whether Boca/MESD func-tions in the trafficking of proteins outside the LDL receptor family. The significant decrease in cell surface ␤ 1 integrin, in CHO 14-2-1 cells, makes this cell line an attractive model for further study.
␤ 1 integrin associates with multiple ␣-subunits to form integrin heterodimers that have partially redundant activities in cell adhesion (1). The effects of LRP-1 on individual ␤ 1 -subunitcontaining integrins remains to be determined. Such data may clarify certain results presented here. For example, although the major fibronectin-binding integrins, ␣ 5 ␤ 1 and ␣ 4 ␤ 1 , both contain the ␤ 1 -subunit, we found that adhesion and spreading of CHO 14-2-1 cells was delayed but not blocked. Redundancy in the activity of integrins may be operational here, because ␣ v ␤ 3 also serves as a fibronectin receptor (1). Furthermore, although absolute cell surface levels of integrins may affect adhesion and spreading, the state of integrin activation and clustering are also very important (2).
From these studies, a model emerges in which LRP-1 serves to landscape the plasma membrane. It is known that LRP-1 may decrease cell surface levels of membrane proteins by facilitating their endocytosis (22)(23)(24)(25). Our results now suggest that LRP-1 also may increase levels of cell surface proteins based on its ability to facilitate membrane protein transport in the secretory pathway. Together, these novel LRP-1 activities may substantially regulate how the cell responds to extracellular stimuli.
FIG. 6. Adhesion of LRP-1-positive and -negative cells to fibronectin. A, cell culture wells were precoated with the indicated concentrations of fibronectin (FN) or blocked with BSA. PEA10 or MEF-2 cells were allowed to adhere to the wells for 30 min. Adhesion was determined by crystal violet staining. B, CHO K1 and CHO 14-2-1 cells were suspended in medium that included 1 mM MnCl 2 (ϩMn 2ϩ ) or was MnCl 2 -free (ϪMn 2ϩ ). The cells were then plated into wells that were pre-coated with different concentrations of fibronectin (FN) or blocked with BSA, and allowed to adhere for 30 min. Adhesion was determined by crystal violet staining (*, p Ͻ 0.05; **, p Ͻ 0.01). C, CHO cells were allowed to adhere to 1 g/ml fibronectin for 30 min and then imaged by phase-contrast microscopy and photographed at equivalent magnification.