Sequence Identity between the cy2-Macroglobulin Receptor and Low Density Lipoprotein Receptor-related Protein Suggests That

Ten peptides, derived from az-macroglobulin @zM) by chemical or proteolytic digestion, were sequenced. Comparative analysis revealed all of the resulting sequences present within that show this molecule, like LRP, to contain two polypeptides of approximately 420 and 85 kDa that are noncovalently associated. An additional component of this receptor system is a 39-kDa polypeptide that co-purifies with the (YAM receptor during affinity chromatography. Solid phase binding studies reveal that the 39-kDa polypeptide binds with high affinity (& = 18 nM) to the 420-kDa component of the azM receptor. The ap- parent identity of LRP and the azM receptor suggests that this molecule is a multifunctional receptor with the capacity to bind diverse biological ligands and highlights a possible relationship between two previ- ously unrelated biological processes, lipid metabolism and proteinase regulation.

Recently, the aeM receptor has been purified by ligand affinity chromatography from detergent extracts of placenta (8,9) and liver (10,ll). This procedure identified the receptor as a large glycoprotein with an approximate molecular mass of 420 kDa. This molecule binds to the activated form of (Y*M with an affinity (8) similar to that observed for the binding of CQM to fibroblasts (7), hepatocytes (12), macrophages (13,14), and monocytes (15). In addition to this large polypeptide, two additional components were observed to co-purify. One of these, with an apparent molecular mass of 85 kDa, was thought to originate from proteolysis of the intact receptor (8). The second molecule has an approximate molecular mass of 39 kDa and does not appear to bind to apM.
The present investigation was undertaken to determine the primary structure of the LY~M receptor. The results establish that the sequence of the apM receptor is probably identical to the deduced sequence of a recently described (16) cell surface protein termed low density lipoprotein receptor-related protein (LRP), so named due to its structural similarity to the LDL receptor.
Proteins-The a,M receptor was purified by affinity chromatography from detergent extracts of placenta as described (8)  Structure of the cr2-Macroglobulin Receptor Immunoprecipitation Experiments-Human gingival fibroblasts were grown in 100 X 20-mm culture dishes until almost confluent and washed with cysteine-free RPMI. The cells were then incubated overnight with 10 ml of 25 &i/ml [""S)cysteine in cysteine-free RPMI, washed with TBS containing 1 mM PMSF and 20 rg/ml leupeptin, and extracted with TBS containing 0.1% Triton X-100, 1 mM PMSF, 20 pg/ml leupeptin. The remainder of the steps were carried out as previously described (17), except that protein G-Sepharose was used in place of protein A-Sepharose.
ELISA-An ELISA (18) was performed as previously described (8), except that the blocking step included 1 mg/ml rabbit IgG. The Kd was estimated from the data as previously described (8).

Monoclonal Antibody
Production-The immunization protocol and fusion were carried out as previously described (19). Mice were immunized with polypeptides eluted from the (YAM affinity column. Solid phase screening assays (20) using purified components were utilized to identify antibodies of interest.

SDS-PAGE
and Western Blotting-A discontinuous pH gel system (Laemmli) was used with a 4% polyacrylamide stacking gel and a 5-15% separating gel. For immunoblotting experiments, the gels were run at 4°C. Prior to transfer, the gel was soaked in TBS, 5 mM CaCh. Transfer to immobilon-P (Millipore) was performed overnight at 4 "C at 32 V in 20 mM Tris, 150 mM glycine, 20% (v/v) methanol. After transfer, the membrane was blocked with TBS, pH 8.0, containing 0.1% Tween 20, 2 mM CaC12, and 10% fetal calf serum (Buffer A). Following blocking, the membranes were incubated with various antibodies diluted into buffer A. Antibody binding was detected with an anti-mouse alkaline phosphatase conjugate.

RESULTS
Sequencing Analysis-To determine the primary structure of the crzM receptor, protein sequencing was performed. The receptor was purified by affinity chromatography, followed by anion exchange chromatography (8). By utilizing these procedures it is possible to obtain highly purified preparations that migrate as a single 420 kDa band upon SDS-PAGE (8). The amino terminus of the receptor was blocked, and, therefore, peptides were generated from this molecule by either chemical or proteolytic digestion. This resulted in the sequence of 10 peptides (Fig. 1). The identity of a total of 146 from the purified 420-kDa crzM receptor with corresponding sequences in LRP (16).
Lys C peptides were generated by Lys C digestion of the reduced and alkylated receptor. Three of the peptides (CNBr 17/Lys C 29, CNBr 19/Glu C 2, and CNBr 19/Glu C 10) were first digested with CNBr, purified by HPLC, and further digested with Glu C or Lys C. Amino-terminal residues enclosed inparentheses represent expected residues based on the method used to generate the peptide. The boxed residues represent differences between the chemical sequence and that deduced from the cDNA for LRP (16). An X indicates a cycle in which no amino acid could be identified. In the case of Lys C 39 (peptides 1 and 4), two sequences with similar yields were obtained. The yield from the first cycle for peptides l-10 and the repetitive yield for each sequencing run was: 1) 36 pmol, 95%; 2) 14 pmol, 94%; 3) 16 pmol, 90%; 4) 35 pmol, 93%; 5) 40 pmol, 86%; 6) 40 pmol, 88%; 7) 15 pmol, 88%; 8) 47 pmol, 89%; 9) 15 pmol, 91%; 10) 8 pmol, 94%. residues was determined. All resulting sequences were found at various locations throughout the cDNA-deduced sequence of LRP (16). Only 5 out of the 146 residues differed between the sequence deduced from cloning and that determined from chemical sequencing; the differences could represent either sequencing errors or polymorphisms. Overall, these findings provide convincing evidence that LRP and the CQM receptor are the same molecule.
Immunoprecipitation of the a2M Receptor-Herz et al. (21) examined the biosynthesis of LRP and found this molecule to be synthesized as a 600-kDa precursor that is cleaved in the Golgi to form a heavy and light chain, with approximate molecular masses of 500 and 85 kDa, respectively. These two chains are proposed to associate noncovalently to form the functional receptor (21). Immunoprecipitation experiments, using a monoclonal antibody, were employed to determine if both the 420-and 85-kDa components of the anM receptor co-precipitated.
The specificity of the monoclonal antibody was evaluated by immunoblotting. The results shown in Fig. 2 (lane 2) demonstrate that antibody 8Gl binds specifically to the 420-kDa polypeptide but not to the 85-or 39-kDa polypeptides, while two other monoclonal antibodies, 5A6 and 8B8, were selective for the 85-kDa polypeptide (Fig. 2, lanes 4 and 6). The immunoblotting results are also in agreement with solid phase binding experiments demonstrating that 8Gl binds selectively to microtiter wells coated with the 420-kDa polypeptide. In immunoprecipitation experiments, 8Gl precipitated not only the expected 420-kDa polypeptide but a second polypeptide with an apparent molecular mass of 85 kDa (Fig.  3). These data suggest that the 420-and 85-kDa components are associated by relatively strong non-covalent interactions, as has been suggested for LRP, and provide additional evidence that these two molecules are similar. Parallel immunoprecipitation of media fractions indicates that neither the 420-nor 85-kDa polypeptides are secreted.  Protein G-Sepharose was utilized to precipitate the antibody-antigen complex. Following incubation with Protein G-Sepharose, the resin was extensively washed with TBS containing 0.1% Triton X-100, 1 mM PMSF, and 0.5 M NaCl prior to SDS-PAGE under reducing conditions. Following electrophoresis, the gel was exnosed to Kodak XAR film for 4 davs at -70 "C. Arrows denote the migration position of the 420-and 85-kDa polypeptides. Microtiter wells were coated with the 39-kDa polypeptide (0) or with bovine serum albumin (m), and increasing concentrations of (Y?M receptor were added. Following incubation and washing, the amount of bound receptor was determined by using monoclonal antibody 8Gl at a 1:lOOO dilution. Each point represents the average of duplicate determinations. The curve represents the best fit to a single class of binding sites, with & = 18 nM, determined by nonlinear regression.
kDa, was also detected in samples of ligand affinity-purified CQM receptor. Since this polypeptide did not appear to bind to any form of cr2M (8), its interaction with the 420-kDa polypeptide was examined. The results of Fig. 4 demonstrate a specific and high affinity interaction (apparent Kd = 18 nM) between the 39-kDa polypeptide and the purified 420-kDa component of the (Y*M receptor. In a control experiment, the 420-kDa receptor did not bind to microtiter wells coated with bovine serum albumin.

DISCUSSION
Protein sequencing studies of the (YAM receptor have provided strong evidence that this molecule is probably identical to LRP. The sequences of a total of 10 peptides were obtained. These peptides were prepared from different preparations of the (Y*M receptor by either chemical or proteolytic cleavage, or both. The fact that all of the peptides that were sequenced are located within the deduced sequences of LRP strongly argues against the possibility that LRP represents a minor contaminant of the apM receptor preparation. In addition to this structural evidence, biochemical analysis reveals additional similarities between LRP and the apM receptor. This includes identification of an 85-kDa polypeptide that associates with the 420-kDa component of the LY*M receptor during affinity purification and immunoprecipitation experiments. These results are consistent with the proposal that the CY~M receptor, like LRP, consists of at least two polypeptides that are non-covalently associated to form a functional receptor. The relationship between these two polypeptides was established by Herz et al. (1990) who demonstrated that the 85 kDa polypeptide originates from the carboxyl-terminal region of a single chain precursor LRP molecule following cleavage in the Golgi during receptor processing. Interestingly, blotting experiments in the present investigation, using monoclonal antibodies specific for the 85-kDa component, suggest that the active anM receptor (i.e. receptor eluted from the ligand affinity column) is present exclusively in the two-chain form. Whether or not the single chain form of the receptor retains the ability to bind to ligand remains to be determined. Discrepancies between our estimate of 420 kDa for the molecular mass of the heavy chain of the (Y*M receptor and the estimate of 500 kDa for the heavy chain of LRP are likely due to uncertainties in determining the mass of large proteins from their mobility in SDS-PAGE. Thus all of the biochemical evidence currently available indicates that the OVUM receptor and LRP are the same molecule.
The present investigation has identified a third molecule that associates with this receptor. This polypeptide, which has an apparent molecular mass of 39 kDa, co-purifies with the LY*M receptor during affinity chromatography and binds to the 420-kDa polypeptide of the LY*M receptor with high affinity. This polypeptide would not be detected by immunoprecipitation of [35S]cysteine-labeled cell extracts due to the absence of cysteine residues within the protein.' Whether this molecule represents a receptor subunit or an additional ligand for this receptor remains to be determined, and studies are currently under way to delineate the function of this unique molecule.
The apparent identity between LRP and the crzM receptor raises important questions about the ligand for this receptor. Ligand binding studies with the purified azM receptor clearly establish that the activated form of aaM is a ligand. The interaction is of high affinity (Kd = 0.6-12 nM) (8-lo), is specific for the activated form of a2M (8), depends upon calcium (8-lo), and is inhibited when the pH is reduced to 5.0 (8-10) as expected for this interaction (22). Furthermore, cross-linking studies (9,10) and purification of the heavy chain (8) of the cuzM receptor provide additional evidence that the interaction between activated (YAM and the receptor involves the larger 420-kDa polypeptide chain. It has also been suggested that LRP might function as a receptor for lipoproteins that contain apoE (23)(24)(25)(26)  Structure of the cu2-Macroglobulin Receptor ration and that the stimulation could be blocked by an antibody against LRP. However, it was not possible to measure directly the binding, uptake, and degradation of apoE-enriched lz51-labeled /WLDL by these cells. More direct evidence for an interaction between apoE-and apoB-containing liposomes and LRP was derived from cross-linking studies (24) and ligand blotting studies (25, 26) which demonstrated a Ca*+-dependent binding of various lipoproteins to LRP immobilized on nitrocellulose. Thus, while it is not possible to estimate the affinity of these interactions from studies of this kind, there is considerable evidence for the interaction of certain lipoproteins with LRP.
It is likely that the ability of this receptor to bind to apparently structurally unrelated ligands is a unique feature of the "modular" structure of this molecule. LRP was originally identified (16) by screening a murine lymphocyte cDNA library with an oligonucleotide derived from the "class A cysteine-rich" motif of low density lipoprotein receptor. The complete amino acid sequence of LRP (16) reveals that the extracellular domain contains 31 copies of the "cysteine-rich" motif that is also found in the LDL receptor and terminal complement components, as well as 22 copies of repeats with homology to the epidermal growth factor precursor. The involvement of the various "modules" in binding to different ligands has been confirmed by an investigation of the regions on the LDL receptor responsible for binding to apolipoprotein E and apolipoprotein B (27). This analysis substantiated that these two ligands bind to different regions of LDL receptor. Interestingly, another receptor, the insulin-like growth factor II receptor, also binds to diverse biological ligands. The receptor, which contains 15 repeat sequences and a domain homologous to the type II repeat found in fibronectin, is identical to the cation-independent mannose 6-phosphate receptor (2% 30). Bindings studies indicate that the two distinct ligands for this receptor (i.e. insulin-like growth factor II and lysoso-ma1 enzymes containing phosphomannosyl residues) bind to different regions of the molecule (30). Thus, it appears that certain receptors are able to bind structurally diverse ligands through distinct binding sites on the molecule.
The apparent ability of the tiZM receptor to bind distinct ligands also raises questions regarding the overall function of this receptor. Studies of the binding of cr*M-proteinase complexes or methylamine-activated o(~M to macrophages suggest that this binding regulates many cell functions such as proteinase secretion, antigen processing, and activation of the respiratory burst by phorbol esters (31-33). These observations suggest an expanded role for this receptor beyond its catabolic function to modulate levels of proteinase activity. It is conceivable that various ligands (e.g., cu*M-proteinase complexes, lipoproteins) may have a concerted role in the process of modulating receptor function. For example, it is certainly possible that (YAM may in some manner facilitate the binding of various lipoproteins to this receptor. In this regard, it is interesting to note that several vitamin K-dependent zymogens and proteinases, such as thrombin, have been reported to bind to VLDL (34,35). The potential cooperative role that ligands may have in the process of modulating the levels of proteinases and lipoproteins and how the different functions of the receptor are integrated remain to be investigated.