The Role of Grp 78 in α2-Macroglobulin-induced Signal Transduction

The low density lipoprotein receptor-related protein (LRP) is a scavenger receptor that binds to many proteins, some of which trigger signal transduction. Receptor-recognized forms of α2-Macroglobulin (α2M*) bind to LRP, but the pattern of signal transduction differs significantly from that observed with other LRP ligands. For example, neither Ni2+ nor the receptor-associated protein, which blocks binding of all known ligands to LRP, block α2M*-induced signal transduction. In the current study, we employed α2-macroglobulin (α2M)-agarose column chromatography to purify cell surface membrane binding proteins from 1-LN human prostate cancer cells and murine macrophages. The predominant binding protein purified from 1-LN prostate cancer cells was Grp 78 with small amounts of LRP, a fact that is consistent with our previous observations that there is little LRP present on the surface of these cells. The ratio of LRP:Grp 78 is much higher in macrophages. Flow cytometry was employed to demonstrate the presence of Grp 78 on the cell surface of 1-LN cells. Purified Grp 78 binds to α2M* with high affinity (K d ∼150 pm). A monoclonal antibody directed against Grp 78 both abolished α2M*-induced signal transduction and co-precipitated LRP. Ligand blotting with α2M* showed binding to both Grp 78 and LRP heavy chains in these preparations. Use of RNA interference to silence LRP expression had no effect on α2M*-mediated signaling. We conclude that Grp 78 is essential for α2M*-induced signal transduction and that a “co-receptor” relationship exists with LRP like that seen with several other ligands and receptors such as the uPA/uPAR (urinary type plasminogen activator or urokinase/uPA receptor) system.

goes a major conformational change that exposes receptor recognition sites on the molecule (1). This activated form of ␣ 2 M is designated ␣ 2 M*. Though difficult to reconcile, accumulating evidence has demonstrated that ␣ 2 M*-induced signaling occurs via a receptor that is functionally unique from the previously characterized ␣ 2 M* receptor, the low density lipoprotein receptor-related protein (LRP). LRP is a scavenger receptor (1) that binds to a variety of proteins, many of which trigger signal transduction (2)(3)(4)(5). This pathway requires the activation of a pertussis toxin-sensitive G protein (2)(3)(4)(5). The receptor-associated protein (RAP) that blocks the binding of all known ligands to LRP is also an antagonist for this signaling pathway (2)(3). Whereas ␣ 2 M* binds to LRP, when cells are exposed to ␣ 2 M* a distinct set of signaling events is observed that differ from those induced by other ligands for this receptor (2)(3)(4)(5)(6). These include activation of a different G protein and the lack of antagonism by RAP or Ni 2ϩ (2)(3)(4)(5)(6)(7). In addition, binding studies with a variety of cells in culture demonstrate two classes of binding sites, one of very high affinity (K d ϳ100 pM and 1600 sites/cell) and one of lower affinity (K d ϳ2-5 nM and ϳ70,000 sites/cell). The high affinity binding site is responsible for ␣ 2 M*-induced signal transduction and has been termed the ␣ 2 M* signaling receptor (␣ 2 MSR), whereas the lower affinity binding site has been identified as LRP (8 -12). Numerous studies have suggested that high affinity binding sites alone are involved in ␣ 2 M*-induced signal transduction (6 -13).
In this report we demonstrate by flow analysis that Grp 78 is on the surface of 1-LN human prostate cancer cells and that cell surface Grp 78 is essential for ␣ 2 M*-induced signal transduction in both 1-LN human prostate cancer cells and murine macrophages. A monoclonal antibody directed against Grp 78 completely abolished ␣ 2 M*-induced signal transduction in both 1-LN cells and murine macrophages as judged by the effects of this antibody on the ability of ␣ 2 M* to induce a rise in [Ca 2ϩ ] i . This response is always seen when signal transduction is activated by ␣ 2 M* in these cells (8 -12). This monoclonal antibody co-precipitates Grp 78 and LRP in both 1-LN cells and macrophages. But in the former cells the ratio of Grp 78:LRP is much greater than in macrophages. We also show by ligand binding assays to membranes that ␣ 2 M* binds to Grp 78 and the heavy, but not light, chain of LRP. The latter observation is consistent with the known receptor-binding properties of LRP (1). Using the technique of RNA interference, we find that LRP is not required for ␣ 2 M*-dependent signal transduction. Finally, we demonstrate that highly purified Grp 78 binds to ␣ 2 M* with very high affinity (K d ϳ150 pM).

EXPERIMENTAL PROCEDURES
Materials-The sources of thioglycollate, cell culture material, and endotoxin-free ␣ 2 M* have been described previously (2)(3)(4)6). ␣ 2 M* was prepared by reaction of human ␣ 2 M with methylamine as in previous * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  studies (2)(3)(4)6). [ 3 H]Myoinositol was from ARC, St. Louis, MO. Antibodies against the heavy (515 kDa) and light (85 kDa) chains of LRP were procured from American Diagnostica Inc. (Greenwich, CT). Antibodies against Grp 78 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Bacitracin and ␤-octylglucoside were procured from Sigma. Polyclonal anti-human antibodies against Grp 78 and normal goat IgG employed for flow cytometry were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). ChromPure Goat IgG, Fc Fragment, fluorescein isothiocyanate (FITC)-conjugated Rabbit Anti-goat IgG, F(abЈ) 2 Fragment Specific, and Rabbit Gamma Globulins were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). Purified anti-mouse H-2K b (AF6 -88.5) was purchased from Pharmingen. Affinity chromatography materials and ECF kits were from Amersham Biosciences. All other reagents used were procured locally and were of the highest available grade. Murine macrophages and 1-LN human prostate cancer cells were cultured as previously described (2-4, 6, 13). The 1-LN cell line was a kind gift from Dr. Philip Walther (Duke University Medical Center, Durham, NC).
Receptor Purification-Cell membranes from murine macrophages and 1-LN prostate cancer cells were obtained and solubilized in ␤-octylglucoside and chromatographed on an ␣ 2 M*-agarose affinity column as previously described (14). By this procedure both LRP and Grp 78 were isolated as demonstrated by several techniques (see below). Grp 78 was also purified from the cell membranes of 1-LN prostate cancer cells by suspending the membranes in 20 mM Tris⅐HCl, pH 8.0 containing 1% Triton X-100 and chromatography on anti-Grp 78-agarose. This procedure yielded Grp 78 free of LRP as demonstrated by several techniques (see below). SDS-PAGE was employed to characterize the proteins obtained from the affinity column. Bands were either directly stained with Coomassie Brilliant Blue or subjected to blotting for probing with either antibodies or ␣ 2 M*. Techniques for blotting assays are described in detail elsewhere (15,16). For some studies, immunoprecipitates were obtained from solubilized cell membranes of 1-LN cell and macrophage membranes employing a monoclonal antibody to Grp 78. These precipitates were subjected to SDS-PAGE (10%) and blotted to Hybrid P ® membranes (Amersham Biosciences) prior to exposure to either appropriate antibodies or ␣ 2 M*. When ␣ 2 M* was employed as a ligand, after appropriate washing the bound ␣ 2 M* was identified by an anti-␣ 2 M* antibody and use of ECF (Amersham Biosciences). A Storm ® 860 Phosphoimager (Molecular Dynamics, Sunnyvale, CA) was employed to detect the bound ␣ 2 M*. The use of these techniques is described in detail elsewhere (9).
Binding of ␣ 2 M* to Immobilized Grp 78-To study the specific binding of ␣ 2 M* to immobilized Grp 78, 96-well culture plates were coated with Grp 78 (1 g/ml in 0.1 M sodium carbonate, pH 9.6, 200 l/well at 37°C for 2 h). After coating, plates were rinsed with 10 mM sodium phosphate/100 mM NaCl, pH 7.4, containing 0.05% Tween 80 (PBS-Tween) to remove unbound protein. Nonspecific sites were blocked by incubation with PBS-Tween containing 2% (w/v) bovine serum albumin at room temperature for 1 h. Plates were rinsed twice with PBS-Tween, air-dried, and stored at 4°C. For assays, increasing concentrations of ␣ 2 M* were added to triplicate wells and incubated at 37°C for 1 h. An ELISA procedure was used to calculate the amount of bound ligand to Grp 78. This procedure involves the construction of calibration curves in 96-well culture plates coated with ␣ 2 M* as described above and then titration with a specific rabbit anti-␣ 2 M*-IgG followed by detection with a secondary anti-rabbit IgG conjugated to alkaline phosphatase. Values from this curve were used to determine the amount of ␣ 2 M* bound to immobilized Grp 78.
Binding of LRP and Grp 78 to Immobilized ␣ 2 M*-Binding of LRP and Grp 78 from cell lysates to immobilized ␣ 2 M* was carried out by an ELISA on 96-well culture plates. Plates were first coated with ␣ 2 M* (5 g/ml) in 0.1 M sodium carbonate, pH 9.3, 0.01% sodium azide. The plates were rinsed with 10 mM sodium phosphate, pH 7.3, containing 0.15 M sodium chloride and 0.05% (v/v) Tween 80 (PBS-Tween) and incubated for 2 h at room temperature with 3% bovine serum albumin in PBS-Tween to block nonspecific sites, followed by three more rinses in PBS-Tween to remove excess reagents. The plates were then dried at room temperature and stored at 4°C until further use. Binding of LRP from cell lysates was performed by incubating the plates with increasing concentrations of the lysate at 37°C for 90 min. The plates were then rinsed with PBS-Tween and incubated for 60 min at 37°C with 200 l of a solution containing an anti-LRP mouse IgG (1 g/ml) in PBS-Tween. The plates are then rinsed with PBS-Tween and incubated with 200 l of a solution containing the alkaline phosphatase substrate p-nitrophenyl phosphate (1 m/ml) in 20 mM Tris-HCl, pH 8.5, 1 mM zinc chloride. The absorbance was monitored at a wavelength of 405 nm with an Anthos Labtec plate reader, and the LRP bound was expressed as⌬A 405 nm . Binding of Grp 78 was also assessed on the same lysates by incubating the ␣ 2 M* plates with a rabbit anti-Grp 78 IgG after reaction with the lysate and by using an anti-rabbit alkaline phosphataseconjugated secondary antibody. The amount of Grp 78 bound was monitored as described above and also expressed as ⌬A 405 nm .
Flow Analysis of 1-LN Human Prostate Cancer Cells for Cell Surfaceassociated Grp 78 -1-LN cells were grown at 37°C and 5% CO 2 in RPMI 1640 supplemented with 10% fetal bovine serum (FBS), 1% penicillin/streptomycin, and 35 nM human insulin. Cells were detached by incubation for 5 min at 37°C with Ca 2ϩ -and Mg 2ϩ -free PBS containing 4 mM EDTA and then pelleted. Cells were washed once with PBS before being resuspended in ice-cold staining buffer (Phenol Redfree Hank's balanced salt solution (HBSS), 1% bovine serum albumin, 0.3 mg/ml rabbit gamma globulins, 0.4 mg/ml goat IgG Fc fragments, 0.01% NaN 3 ) at a concentration of 1 ϫ 10 7 cells/ml. Aliquots (100 l) of these cell suspensions were incubated at 4°C for 30 min with an appropriate dilution of goat polyclonal anti-Grp 78 or normal goat IgG (control). Cells were washed three times with ice-cold staining buffer, pelleted, and resuspended in 100 l of ice-cold staining buffer. These cell suspensions were then incubated in the dark with a FITC-conjugated rabbit anti-goat IgG, F(abЈ) 2 for 30 min. Cells were then washed three times with ice-cold staining buffer, resuspended in ice-cold 1% paraformaldehyde, and stored in the dark at 4°C until analysis by flow cytometry. The mean relative fluorescence after excitation at 495 nm was determined for each sample on a FACS Vantage S.E. flow cytometer, and data was analyzed using CellQuest software (BD Pharmingen).
Chemical Synthesis of dsRNA Homologous in Sequence to the Target LRP-1 Gene Sequence-The chemical synthesis of dsRNA homologous to the target LRP gene amino acid sequence 146 -152 nucleotide base sequence 5ЈAAAACATGTAAGGACTTTGAT3Ј (Swiss-Prot, LRP primary accession number Q0754) was performed by Ambion sequence ID 285, Austin, TX. For making dsRNA, the sense (5ЈAACAUGUAAG-GACUUUGAUtt3Ј) and antisense (5ЈAUCAAAGUCCUUACAU-GUUtt3Ј) oligonucleotides were annealed according to the manufacturer's instructions. Throughout the entire period of experimentation handling of reagents was performed in an RNase-free environment. Briefly, equal amounts of sense and antisense oligonucleotides were mixed and heated at 90°C for 1 min and then for 1 h at 37°C in an incubator. The dsRNA preparation was stored at Ϫ20°C before use.
Transfection of Murine Peritoneal Macrophages with dsRNA Homologous in Sequence to the Target LRP-1 Gene Sequence-This procedure has been described in detail elsewhere (16). In brief, thioglycollateelicited murine peritoneal (1 ϫ 10 6 cells/well in a 6-well plate) were lavaged as above and allowed to adhere for 2 h in RPMI 1640 medium containing 10% FBS, penicillin (12.5 units/ml), streptomycin (6.5 g/ ml), and 2 mM glutamine at 37°C in a CO 2 (5%) humidified incubator at 37°C. The non-adherent cells were aspirated, monolayers were washed twice with Hank's balanced salt solution containing HEPES (HHBSS), 2 ml of DMEM medium containing 10% FBS and above antibiotics added, and cells incubated as above for 16 h. For each transfection, 2 g of dsRNA was diluted into 100 l of serum-free DMEM in a tube. In another tube, 10 l of LipofectAMINE was diluted into 100 l of serumfree medium. The two solutions were combined, mixed gently, and incubated for 45 min at room temperature followed by the addition of 800 l of serum-free and antibiotic-free medium to each tube. The monolayers were washed twice with serum-free DMEM medium, layered in each well with 1 ml of LipofectAMINE-DMEM (10 l/ml) or lipid-dsRNA mixtures, dsRNA (26 g), gently mixed, and incubated for 5 h at 37°C in a humidified CO 2 incubator. At the end of the incubation, 1 ml of antibiotic-free DMEM containing 10% FBS was added to each well and cells were incubated for 16 h as above. Microscopic observation of the monolayers did not show obvious evidence of toxicity. The cells did not take up Trypan Blue, and their morphology was not altered by the procedure (data not shown). The medium was replaced with DMEM medium containing antibiotics and 10% FBS 24 h after the start of transfection. The monolayers were washed with the above DMEM medium once, and Western blotting of LRP and Grp 78 in cell lysates was performed. Equal amounts of proteins from both groups of lysates were electrophoresed and transferred to membranes. LRP and Grp 78 were detected by ELISA as described above.

Quantification of [ 3 H]IP 3 Formation in Macrophages after LRP Gene
Silencing-The quantification of [ 3 H]inositol (1, 4, 5)trisphosphate [ 3 H]IP 3 formed from [ 3 H]myoinositol-labeled, dsRNA-transfected macrophages upon stimulation with ␣ 2 M* (100 pM) was performed as described earlier (2,6). Briefly, 2-h adhered macrophages in 6-well plates (1 ϫ 10 6 cells/well) were transfected with dsRNA homologous in sequence to target the LRP gene as above. The transfected cells in DMEM medium containing 10% FBS and antibiotics were incubated with [ 3 H]myoinositol (10 Ci/ml) overnight. The cells were washed with HHBSS containing 10 mM LiCl, 1 mM CaCl 2 , and 1 mM MgCl 2 thrice. A volume of the wash buffer was added and cells were incubated at 37°C for 5 min for temperature equilibration. The cells were stimulated with ␣ 2 M* (100 pM) for varying periods of time. The reaction was stopped by aspirating the medium and adding a volume of 6.25% HClO 4 . The cells were scraped into respective glass tubes, 1 ml of freon:octylamine (1:1, v/v) was added, and the mixture was vortexed for 10 s and centrifuged at 1500 rpm for 10 min. The aqueous phase was applied onto a 1 ml Dowex resin (AG 1-X8-formate), and an inositol phosphates batch was eluted as described previously (6). The untransfected cells were processed similarly and were the controls. In experiments where the effect of pertussis toxin was studied on the formation of [ 3 H]IP 3 upon stimulation with ␣ 2 M, the cells were incubated with pertussis-toxin (1 g/ml/ 12h) 24 h after initial transfection. Other details of quantification were the same as above.

Purification of ␣ 2 M* Binding Proteins from 1-LN Human
Prostate Cancer-LRP has been purified from a variety of cells; however, isolation from 1-LN human prostate cancer cells has not been reported. These cells were chosen because insulintreated 1-LN cells signal briskly when exposed to ␣ 2 M* while expressing very little LRP (13). In these cells the predominant ␣ 2 M* binding site demonstrates a K d of ϳ100 pM (13). Following affinity chromatography of solubilized cell membrane preparations, LRP was found in low yield, but the predominant protein isolated demonstrated a molecular weight of ϳ80 kDa (Fig. 1A). NH 2 -terminal sequence analysis yielded NH 2 -EEED-KKEDVGT-, identical to that of Grp 78 as identified in the Swiss-Prot data base. When macrophage membranes were employed for purification, as expected (19), the majority of the protein obtained was LRP, and only a trace of Grp 78 could be identified by immunoblotting with anti-Grp 78 (data not shown). The association of Grp 78 with LRP was demonstrated by experiments in which antibody to Grp 78 was used to immunoprecipitate the protein from solubilized membranes of The blot demonstrates that LRP co-precipitated with Grp 78. B, the binding of ␣ 2 M* to the proteins in the membrane preparation from 1-LN prostate cancer cells and macrophages. Immunoprecipitation was performed with anti-Grp 78 as in panel A, lanes 5 and 6. After membrane transfer, ligand binding of ␣ 2 M* to the proteins on the membrane was performed as previously described (15). Lane 1 is from 1-LN prostate cancer cell membranes, whereas lane 2 is from macrophages. Lane 3 is purified Grp 78 obtained as described under Experimental Procedures for comparison. C, in a separate experiment performed as in A and B, the immunoprecipitated Grp 78 was probed for MHC class I 45-kDa protein as previously described (17). both 1-LN cells and macrophages (Fig. 1A). In these studies, LRP heavy and light chains were co-precipitated with Grp 78. Ligand blot analysis of solubilized membranes from either 1-LN cells or macrophages immunoprecipitated with anti Grp 78 showed ␣ 2 M* binding to Grp 78 and the heavy, but not light, chain of LRP (Fig. 1B). The latter result is expected because ␣ 2 M* binds only to the heavy chain of LRP (1). Previous studies have demonstrated that Grp 78 is a co-receptor with MHC class I histocompatibility antigens (17). We were able to demonstrate that the Grp 78 immunoprecipitate from murine membranes also contained MHC class I ϳ45-kDa protein (Fig. 1C) consistent with this report (17). Based on these observations, we next determined whether Grp 78 is expressed on the surface of 1-LN cells. After staining the cells with an antibody against Grp 78, cell-surface expression was confirmed by flow cytometry (Fig.  2A). In light of these observations, we then purified Grp 78 free of LRP from 1-LN cell membranes as described under "Experimental Procedures." ␣ 2 M* binding to the purified preparation was detected by ELISA. The data fit a simple hyperbolic curve, indicating a single class of binding sites, K d ϳ150 pM (Fig. 2B). These results are consistent with previous observations of an ␣ 2 M* high affinity binding site on these cells, K d ϳ100 pM (13).
␣ 2 M*-induced Increases in [Ca 2ϩ ] i -We have previously demonstrated that ␣ 2 M* binding to macrophages or 1-LN cells triggers IP 3 synthesis followed by an increase in [Ca 2ϩ ] i (13). Fig. 4 demonstrates the expected increase in [Ca 2ϩ ] i after macrophages or 1-LN cells were exposed to ␣ 2 M* (100 pM) (Fig.  4). When the cells were pretreated 20 min prior to ␣ 2 M* addition with either anti-Grp 78 or with bacitracin, the increase in [Ca 2ϩ ] i was completely blocked (Fig. 3). Bacitracin was employed in these studies because it was previously shown to block binding of ␣ 2 M* to an uncharacterized high affinity (K d ϳ200 pM) class of receptors on the cell surface (20).
The Role of LRP in ␣ 2 M*-mediated Signal Transduction-The studies presented above demonstrate that Grp 78 is essential for ␣ 2 M*-induced signal transduction and that LRP coprecipitates with Grp 78. They do not, however, establish a role for LRP in ␣ 2 M*-mediated signal transduction. To address this issue, dsRNA homologous to the target LRP gene was employed to silence LRP expression. An ELISA was then utilized to demonstrate that LRP expression was completely silenced in these cells, whereas Grp 78 expression was unchanged (Fig. 4). Having established that LRP expression was ablated, we next studied signal transduction in these macrophages in comparison to the untreated cells. For these studies, IP 3 synthesis was directly measured (Fig. 5A). ␣ 2 M* treatment of macrophages causes a significant and transient rise in IP 3 , which then causes Ca 2ϩ mobilization from the endoplasmic reticulum (3, 4, 6 -12). We chose to directly measure IP 3 rather than [Ca 2ϩ ] i for two reasons. First, the rise in [Ca 2ϩ ] i is directly attributable to IP 3 synthesis (3-6), and second, it would be very difficult to perform the technique of RNA interference with macrophages plated on coverslips as is required for digital imaging microscopy. As can be seen in Fig. 5A, exposure of either macrophages treated with dsRNA homologous to LRP or the control cells resulted in a comparable increase in IP 3 , although the duration of the increase appears prolonged in macrophages where LRP is not produced. This increase was pertussis toxin-insensitive (Fig. 5B), as is a characteristic of ␣ 2 M*-induced synthesis of IP 3 (2-4, 6 -12).

DISCUSSION
Studies by Pastan and co-workers (20 -22) first reported isolation of an ␣ 2 M* receptor by classical purification methods in the 1980's. They demonstrated that a protein of 85,000 molecular weight bound to 125 I-␣ 2 M* (20 -22). Moreover, they observed two classes of binding sites on cells: one of high affinity (K d ϳ200 pM and 10,000 sites/cells) and one of lower affinity (K d ϳ100 nM and 600,000 sites/cell) (20 -22). They also reported that bacitracin blocked binding to the high, but not low, affinity sites. The studies were performed at 4°C, thus ruling out the possibility that these results represented the well known effect of bacitracin on endocytosis at 37°C (20 -22). These observations were not pursued, and this protein was never identified. LRP was then identified by Strickland et al. (18) in 1990 as the ␣ 2 M* receptor. Subsequent studies indicated that this receptor is a multidomain, multiligand receptor that is capable of interacting with a large variety of proteins (1,18,19). The receptor consists of a 515-kDa heavy chain and an 85-kDa light chain. Many subsequent reports have supported the identification of LRP as the ␣ 2 M* receptor (for example, Refs. 1 and 18).
In 1993 and 1994 we observed that exposure of macrophages to ␣ 2 M* triggered a typical IP 3 -dependent signaling pathway (2,3,6). We then demonstrated that a number of other ligands for LRP also activated signal transduction (2,3). Whereas we initially believed that ␣ 2 M*-induced signal transduction resulted from ligation of LRP, more detailed studies suggested that a unique ␣ 2 M* signaling receptor must exist (2-4, 6 -12). The basis for this hypothesis has been extensively discussed elsewhere (2-4, 6 -12). In brief, ␣ 2 M*-dependent signal transduction requires activation of a different G protein than seen with other ligands for LRP, and signaling is not blocked by either Ni 2ϩ or a very high molar excess of RAP, which is an antagonist for LRP-dependent signal transduction (2,3,7). In contrast to other ligands for LRP, ␣ 2 M* selectively activates phospholipase D, increases cytosolic pH, and activates p21 Ras and PI 3-kinase (23-26).
Bacskai et al. (27) reported that ␣ 2 M* binding to LRP on the surface of neurons mediates signaling by activation of N-methyl-D-aspartate receptors, which presumably represent a "coreceptor." More recently, Qiu et al. (28) have suggested that a negative relationship may exist between ␣ 2 M* binding to neurons and the signal induced by N-methyl-D-aspartate. Other recent studies have shown that the wingless/frizzled (Wnt/Frz) signaling pathway requires that Wnt simultaneously bind to both Frz and LRP-5 to induce signal transduction (29 -32).
A co-receptor relationship also exists between the uPA/uPAR system and LRP (1). This relationship may represent a model for the Grp 78/LRP relationship. In this case uPA signaling depends exclusively on uPAR binding; however, the association of the uPA/uPAR complex, which also probably includes the inhibitor of uPA, binds to LRP and is internalized, thus regulating the signaling of uPAR (1).
We chose to purify ␣ 2 MSR from the membranes of 1-LN human prostate cancer cells, because these cells express very little LRP but significant amounts of ␣ 2 MSR (13). By contrast, in macrophages ␣ 2 MSR represent only 5% or less of the membrane receptors for ␣ 2 M* with the remaining binding sites being LRP (13). By this approach, we purified Grp 78. Subsequent studies demonstrated that purified Grp 78 binds to ␣ 2 M* with very high affinity (K d ϳ150 pM). Moreover, antibodies directed against Grp 78 blocked ␣ 2 M*-induced increases in [Ca 2ϩ ] i . A similar result was obtained with bacitracin, suggesting perhaps that Pastan and co-workers (20 -22) isolated Grp 78 in the 1980s as their 85,000 molecular weight ␣ 2 M* binding protein. Presumably, LRP was lost during the various steps of classical protein purification that they employed to purify the receptor.
Grp 78 is a typical stress protein that is a member of the HSP 70 superfamily of heat shock proteins (33). Grp 78 has been found in the endoplasmic reticulum, and there it can be associated with low density lipoprotein (LDL) receptors as demonstrated with a number of mutated forms of the receptor (34). These mutant forms of LDL receptor are retained in the endoplasmic reticulum after Grp 78 binding, perhaps because of modified conformations (35). Grp 78 has also been located on the cell surface in several recent studies (35,36). That Grp 78 associates with LRP on the cell surface is demonstrated by the co-precipitation of LRP with Grp 78 when either macrophage or 1-LN membrane fractions are treated with antibody to Grp 78. A precedent for the involvement of heat shock proteins in a signaling receptor complex is known. The glucocorticoid receptor is capable of binding its ligand with low affinity, but only when associated with hsp 90 is high affinity binding observed (34,37). The latter complex, when it binds its ligand, triggers signal transduction, whereas the lack of hsp 90 on the cell surface results in a conformation of the glucocorticoid receptor that cannot activate cell signaling (37,38). Our model for the ␣ 2 MSR is shown in Fig. 6. We speculate that during trafficking through the endoplasmic reticulum a small fraction of LRP becomes associated with Grp 78 and that this complex reaches the cell surface. A similar hypothesis has been made to explain the association of Grp 78 and MHC class 1 on the surface of cells (17).
In summary, we demonstrate that Grp 78 on the surface of cells is necessary and sufficient for ␣ 2 M*-mediated signal transduction. Whereas LRP appears to be present in complex with Grp 78, it is not essential for signal transduction.