Primary Structure of a2-Macroglobulin Receptor-associated Protein HUMAN HOMOLOGUE OF A HEYMANN NEPHRITIS ANTIGEN*

The aa-macroglobulin (a2M) receptor complex as purified by affinity chromatography contains three poly- peptides: a 515-kDa heavy chain, an 85-kDa light chain, and a 39-kDa associated protein. Previous studies have established that the 515/85-kDa components are derived from a 600-kDa precursor whose complete sequence has been determined by cDNA cloning (Herz, J., Hamann, U., Rogne, S., Myklebost, O., Gassepohl, H., and Stanley, K. (1988) EMBO J. 7,4119-4127). We have now determined the primary structure of the human 39-kDa polypeptide, termed aZM receptor-as-sociated protein, by cDNA cloning. The deduced amino acid sequence contains a putative signal sequence that precedes the 323-residue mature protein. Comparative sequence analysis revealed that aZM receptor-associ-ated protein has 73% identity with a rat protein re- ported to be a pathogenic domain of Heymann nephritis antigen gp 330 and 77% identity to a mouse heparin- binding protein termed HBP-44. The high overall identity suggests that these molecules are interspecies homologues and indicates that the pathogenic domain, pre- viously thought to be a portion of gp 330, is in fact a distinct

The aa-macroglobulin ( a 2 M ) receptor complex as purified by affinity chromatography contains three polypeptides: a 515-kDa heavy chain, an 85-kDa light chain, and a 39-kDa associated protein. Previous studies have established that the 515/85-kDa components are derived from a 600-kDa precursor whose complete sequence has been determined by cDNA cloning (Herz, J., Hamann, U., Rogne, S., Myklebost, O., Gassepohl, H., and Stanley, K. (1988) EMBO J. 7,4119-4127). We have now determined the primary structure of the human 39-kDa polypeptide, termed a Z M receptor-associated protein, by cDNA cloning. The deduced amino acid sequence contains a putative signal sequence that precedes the 323-residue mature protein. Comparative sequence analysis revealed that aZM receptor-associated protein has 73% identity with a rat protein reported to be a pathogenic domain of Heymann nephritis antigen gp 330 and 77% identity to a mouse heparinbinding protein termed HBP-44. The high overall identity suggests that these molecules are interspecies homologues and indicates that the pathogenic domain, previously thought to be a portion of g p 330, is in fact a distinct protein. Further, the 120-residue carboxylterminal region of aaM receptor-associated protein has 26% identity with a region of apolipoprotein E containing the low density lipoprotein receptor binding domain. Pulse-chase experiments revealed that the newly formed azM receptor-associated protein remains cell-associated, while surface labeling experiments followed by immunoprecipitation suggest that this protein is present on the cell surface forming a complex with the azM receptor heavy and light chains.
a2-Macroglobulin (a2M)' is a plasma glycoprotein capable of inhibiting numerous proteinases. The inhibition is thought * This work was supported by Grants HL30200 and GM42581 and by Research Career Development Award HL-02113 (to D. K. S.). 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 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in thispaper has been submitted to the GenBankl"IEh4BL Data Bank with accession number(s) M639%59. § To whom correspondence and reprint requests should be ad- to occur via a trap mechanism (l), in which limited proteolysis at a specific region of the a2M subunit (2) is followed by conformational changes in the inhibitor (3-5). These conformational alterations result in reduced activity of the trapped proteinase toward large molecular weight substrates and also expose regions within the carboxyl-terminal portion of the inhibitor (6) that are recognized by azM receptors.
The a2M receptor (a2MR) has been purified from detergent extracts of placenta (7, 8) and liver (9) and has been shown to be identical with LDL receptor-related protein (10, 11) whose complete sequence has been derived from cDNA sequencing (12). In addition to removal of a2M-proteinase complexes, this receptor system is also thought to bind and internalize large cholesterol-containing remnant lipoproteins, enriched with apolipoprotein E (13, 14). a2MR is synthesized as a 600-kDa precursor that is cleaved to give a 515-kDa heavy chain and an 85-kDa light chain (15). The 85-kDa light chain contains a cytoplasmic domain, a hydrophobic membranespanning region and an extracellular portion that binds noncovalently to the heavy chain (15). Associated with the 515/ 85-kDa receptor, we have found a 39-kDa polypeptide (7) that has been termed a2MR-associated protein (a2MRAP) This molecule binds with high affinity (Kd = 18 nM) to the 5151 85-kDa a2MR (10). The objective of the present study was to learn more about the structure and function of a2MRAP. The results indicate that this molecule is complexed with the a2MR on the cell surface and that it shares considerable sequence identity with a rat protein thought to be a pathogenic domain of Heymann nephritis antigen gp 330 (16) and with a mouse heparinbinding protein, termed HBP-44 (17).
The high degree of identity between these molecules suggests that they are interspecies homologues and indicates that the Heymann nephritis antigen previously thought to be a portion of gp 330 (16), is in fact a distinct protein. These studies raise questions regarding the role of a2MRAP and the a2MR complex in the pathogenesis of membranous glomerulonephritis.

EXPERIMENTAL PROCEDURES
Purification of n2MRAP-n2MRAP was purified from detergent extracts of human placenta (obtained from Montgomery General Hospital, Olney, MD) by affinity chromatography over immobilized methylamine-activated a2M followed by anion exchange chromatography over a Mono Q (Pharmacia LKB Biotechnology Inc.) column as previously described (7). n2MRAP was dialyzed overnight versus 50 mM HEPES, 50 mM NaCl, p H 7.4, and applied to a Mono s (Pharmacia LKB Biotechnology Inc.) column equilibrated with the same buffer. The column was eluted with a linear gradient of 50 mM NaCl in 50 mM HhPES, pH 7.4, to 1 M NaCl in 50 mM HEPES, pH 9.6. The eluted a,MRAP was then applied to a Superdex 200 (Pharmacia LKB Biotechnology Inc.) column equilibrated with 50 mM sodium phosphate, pH 7.4. n,MRAP derived from this column was homogeneous upon SDS-PAGE followed by silver staining. Deter-mination of protein concentrations were based on absorption measurements at 280 nm. The absorption coefficient was calculated from the amino acid composition (based on the deduced sequence) as described (18) using the following equation: E(l mol" cm", 280 nm) = 0.97(5690W + 1280Y + 12O(S-S)), where W, Y, and S-S represent the number of Trp, Tyr, and disulfide bonds, respectively. A molecular mass of 37,714 daltons was used to calculate the A:E"" value of 9.29 for a,MRAP. Antibody Production-Antibodies against azMRAP were prepared by intradermal injection of rabbits with 25 pg of SDS gel-purified protein emulsified in Freund's complete adjuvant. Four weeks after the first injection, the rabbit was boosted with 25 pg of protein in incomplete Freund's adjuvant, and the rabbit was then boosted at 6week intervals. Ten days following each boost, the rabbit was bled. The antiserum was specific for a,MRAP and had a 50% maximal binding at a dilution of 1/10,000 in an enzyme-linked immunosorbent assay (19). This antiserum was not reactive with the 515/85-kDa components of the a,MR in enzyme-linked immunosorbent assay or Western blotting assays. Affinity-purified IgG was prepared by mixing 3 ml of antiserum with 1 ml of azMRAP-Sepharose overnight at 4 "C. Following washing, the IgG was eluted with 0.1 M glycine, pH 1.9. The pH of the eluted fractions was immediately raised, and the IgG was stored at 4 "C. Antibodies against the azMR complex (i.e. 515/85-kDa and 38-kDa polypeptides) were prepared in a similar manner using the eluate from the a,M-methylamine affinity column as a source of antigen.
Isolation and Sequencing of azMRAP cDNAs-A human placental cDNA hgtll library (20) was immunologically screened (21) using affinity-purified antibody against the a,MR complex (i.e. 515/85 kDa and 39 kDa). Phage that expressed fusion protein reacting with the antibody were cloned to homogeneity, and their insert cDNAs were prepared by EcoRI restriction endonuclease digestion or by polymerase chain reaction (22) and then subcloned into the phage vector M13mp19. Sequencing was performed by the dideoxy-chain termination method (23) using modified T7 polymerase (United States Biochemicals) and synthetic oligonucleotide primers based on derived sequences. All of the sequences reported are derived from sequencing of both strands of the cDNA inserts. Since immunological screening failed to identify a cDNA encoding the initiator methionine, a primer extension library was prepared using a synthetic oligonucleotide primer (nucleotides 393-409) from a region near the 5' end of the longest cDNA (clone 5). The resulting X g t l O library was screened with a '"P-labeled oligonucleotide based on the 5' sequence of clone 5 (nucleotides 17 to 29). Of the 12 positive clones identified in multiple screenings, only a single clone was found having sequence extending 5' of that of clone 5. This cDNA encoded an in-frame methionine and a short 5'-untranslated segment.
Protein Sequencing-30 pg of azMRAP were purified by SDS-PAGE, transferred to nitrocellulose, and digested in situ with trypsin as described (24). The peptides were isolated using an Applied Biosystems Model 130 microbore HPLC and RP300 L8 cartridge. Major peaks were collected and used for sequencing on an Applied Biosystems Model 477A protein sequenator with an on-line Applied Biosystems Model 130 phenylthiohydantoin analyzer.
Circular Dichroism Measurements-Circular dichroism measurements were performed on a JASCO J-5OOC Spectropolarimeter at room temperature. Prior to use, the instrument was calibrated using ammonium d-10-camphorsulfonate (Katayama Chemical Co.) as described (25). Cell pathlength was 0.2 cm, and a baseline spectrum of buffer was subtracted from the data. A mean residue weight of 116, determined from the deduced sequence of a2MRAP, was used to calculate mean residue ellipticities.
Surface Labeling of Fibroblasts and Immunoprecipitation Studies-Human gingival fibroblasts were surface-iodinated, and labeled cell extracts were immunoprecipitated as previously described (7, 10).
Pulse-chase Immunoprecipitation-Human gingival fibroblasts were grown to near confluence in 35-mm culture dishes. Cells were washed three times in leucine-deficient RPMI-1640 (Select-amine, GIBCO), supplemented with ITS (insulin/transferrin/sodium selenite, Collaborative Research) and 10 mM HEPES, pH 7.0, and then incubated for 15 rnin at 37 "C with the leucine-free ITS medium. The cells were then pulse-labeled for 2 min with leucine-free ITS medium, containing 0.37 mCi/ml ['H]leucine. After the 2-min pulse, the medium was removed, and cell layers were washed two times with RPMI, ITS, containing 1.9 mM leucine, 10 mM HEPES, pH 7.0, and allowed to incubate for various periods of time in the same medium at 37 "C. At the indicated time intervals, medium was removed, and the cell layer was extracted with 1 ml of 50 mM Tris, pH 7.4, containing 1% Triton X-100, 0.05% Tween 20, and the proteinase inhibitors 1 mM phenylmethylsulfonyl fluoride, 0.02 mg/ml leupeptin, 0.0015 mg/ml D-Phe-Pro-Arg-CHZC1 (Calbiochem) using a disposable cell scraper. After centrifugation at 100,000 X g, the supernatants from the cell layer and from the media were subjected to immunoprecipitation, followed by SDS-PAGE on gradient gels. Following SDS-PAGE, gels were treated with Enlightning (Du Pont-New England Nuclear), dried, and exposed to x-ray film for 4 months at -70 "C.
RNA Hybridization-Total RNA from human gingival fibroblasts, placenta, human placental choriocarcinoma cells, and human nasal septum tumor cells was electrophoresed in denaturing 1% agarose gels containing formaldehyde (26) and transferred to nitrocellulose (27). RNA hybridization analysis was performed using a "P-labeled 540-base pair a2MRAP cDNA fragment (nucleotides 399-939) as a probe. The fragment was prepared by polymerase chain reaction using a,MRAP cDNA insert and specific, synthetic oligonucleotides as primers. After hybridization, the filters were washed under high stringency (27) and exposed to x-ray film for 48 h.

RESULTS
Isolation of cDNA Clones for the a~MRAP"Xgtl1 phage plaques from a human placental cDNA library (20) were immunologically screened with a rabbit antiserum against the a2MR complex. Five clones were isolated that expressed a fusion protein which was reactive with the antiserum. The fusion protein from these clones absorbed antibodies that reacted specifically with a2MRAP upon Western blotting. Nucleotide sequencing revealed that the cDNA inserts, which ranged from 1.4 to 1.5 kilobases, encoded identical and overlapping sequences. DNA sequencing showed that the composite cDNA contains a single open reading frame of 1071 nucleotides that is followed by a 409-nucleotide 3'-untranslated region, terminating in a poly(A) tail. A potential polyadenylation signal, AATAGA, was detected 16 nucleotides upstream from the poly(A) tail.
Deduced Amino Acid Sequence of the azMRAP-The cDNA and deduced amino acid sequence of azMRAP is shown in Fig. 1. The deduced amino acid sequence was found to contain the amino-terminal sequence obtained from protein sequence analysis of a2MRAP. In addition, two sequences derived from protein sequencing of tryptic fragments isolated from a2MRAP were also found within the cDNA deduced sequence. These findings confirmed the identity of the cDNA clones as encoding a2MRAP. Preceding the amino-terminal sequence of the mature protein is a sequence with properties of a signal peptide which include a core of hydrophobic residues preceded by charged residues (28) and an initiator methionine residue.
The signal peptide cleavage site, however, does not fit the criteria described by Von Heijne (29) for eukaryotic signal cleavage sites in that position -1 consists of a lysine residue rather than the usual Ala, Ser, Gly, Cys, Thr, or Gln. The mature protein contains 323 residues having a molecular mass of 37,714 daltons, excluding the contribution of any carbohydrate. This value is in excellent agreement with the estimate of 39-40 kDa as measured by SDS-PAGE and that of 42 kDa determined by gel filtration chromatography. The sequence contains one potential N-linked glycosylation site (N-X-S/T). A sequence fitting the consensus pattern of a leucine zipper (L-X6-L-xS-L-&-L) (30) at positions 28-49 is present. A tetrapeptide sequence, HNEL, was found at the carboxyl terminus of azMRAP and is similar to the endoplasmic reticulum retention consensus sequence of KDEL (31). In addition, sequence similarity was noted in the regions surrounding each of the 5 tryptophan residues contained in the mature molecule.
A secondary structure analysis (32) of the deduced sequence suggested a relatively high content of a helix. This was confirmed by circular dichroism measurements of the purified protein (data not shown) from which an a-helical content of TGTCATCTTGGCCAAGTATGGTCTGGACGWW\GAAGGACGCTCGGCAGGTGACCAGCAACTCCCTCAGTGGCACCCAGGAAG~CGGGCTGGATGACCCCAGGCTGG~GCTGTGGCA  ~~~A  U  V  G~D  G  K  K  D  A  R  Q~~S  N  S~S  G~Q~D  G~D  D  P  R~~K  TGTCGAGGCAGGAAGGATTGTTTCTGGTGACTGCAGCCGCTGCCGTCGCGACACAGGGCTTGGTGGTGGTAGCATTTGGGTCTGAGATCGGCCCAGCTCTGACT~GGGGCTTGGCTTC 1320  38% was calculated (25). In addition, two regions (residues 189 to 206 and residues 303 to 320) of a,MRAP were found to have potential for forming amphipathic helices.

Identity between a2MRAP, a Rat Protein Identified as Heymann Nephritis Antigen and a Mouse Heparin-binding Pro-
tein-When the sequence of mature a2MRAP was compared (33) with that of other molecules in protein databases, two rodent proteins were found that displayed a high degree of identity. A rat protein, identified as a major pathogenic domain of the Heymann nephritis antigen gp 330 (16), has 73% identity with a2MRAP. A mouse protein, termed heparinbinding protein-44 (HBP-44) (17), has 77% identity with a2MRAP. Fig. 2 shows an alignment of the three proteins. We conclude from the high degree of identity between a2MRAP and these molecules that they are interspecies homologues of one another.
Sequence Similarity with Apolipoprotein E-Database searching also revealed similarity between a2MRAP and sev-eral other proteins. Among these, apolipoprotein E was particularly interesting given that this molecule is thought to mediate lipoprotein binding to LDL receptor-related protein (13, 14), now known to be the a2MR (10). An alignment of the sequence of a2MRAP with apolipoprotein E revealed 26% identity between the carboxyl terminal region of azMRAP (residues 203-321) and a region of apolipoprotein E (residues 47-154) that contains the LDL receptor binding domain (34,35). The similarity between these two molecules seems confined not only to their primary structure. The homology also includes regions with the potential to form amphipathic helices (Fig. 3). A similar amphipathic region in apolipoprotein E (residues 132-149) is known to be directly involved with lipoprotein-receptor interactions (34). Whether or not the corresponding region in a2MRAP mediates its interaction with a2MR remains to be established. In addition to serving a possible role in receptor-ligand interactions, amphipathic helices are thought to serve as lipid-associating domains in apolipoproteins (36).
RNA Hybridization Analysis-RNA hybridization analysis was performed using an a2MRAP-specific cDNA fragment. Two transcripts, of approximately 1.6 and 3.2 kilobases, were detected in total RNA prepared from gingival fibroblasts (Fig.  4). Similar profiles (data not shown) were obtained with RNA isolated from placenta, a human placenta choriocarcinoma cell line (JAR), and a human nasal septum tumor cell line (RPMI 2650). The length of transcripts encoding a2M receptor-associated protein corresponded closely to the 1.8-and 3.5-kilobase transcripts reported for mouse HBP-44 (17). The two transcripts of HBP-44 encode overlapping. RNA sequences differing only in the length of their 3"untranslated regions. In the case of HBP-44, the two messages were thought to arise from the presence of two sets of transcription-termination-polyadenylation signals. We assume that a similar mechanism gives rise to the two transcripts hybridizing with the a,MRAP cDNA probe.
Localization of cu2MRAP on the Cell Surface-The deduced amino acid sequence predicts that a2MRAP could be a secretory protein because it possesses a putative signal peptide and lacks any obvious transmembrane domain. Pulse-chase experiments were performed to examine the time course of biosynthesis and potential secretion of a,MRAP. Immunoprecipitation experiments were conducted with ["Hlleucine, rather than ""S-labeled amino acids because the deduced sequence lacks cysteine and contains only 1 methionine resi-due. The results of pulse-chase immunoprecipitation (Fig. 5) show that a,MRAP remains exclusively associated with the cellular fraction with none secreted into the medium, even after 120 min of chase. The high molecular weight polypeptide seen after 15 min of chase presumably corresponds to the heavy chain of the a2MR suggesting that a2MRAP interacts with the newly formed a2MR at an early stage in its biosynthesis, evidently before the complex reaches the cell surface. Within the first minutes of chase, a 56-kDa immunoreactive polypeptide is also detected. The relationship between this polypeptide and a,MRAP, if any, remains to be established.
To further evaluate whether or not alMRAP is secreted, conditioned medium was concentrated 5-fold and subjected to Western blot analysis using affinity-purified antibody to a,MRAP. These experiments failed to detect any mediaderived immunoreactive material, but a2MRAP was readily detected in detergent extracts of the cells. In earlier studies, we found that a,MRAP, like HBP-44 (17), binds to heparin-Sepharose. We utilized this affinity matrix to concentrate any a2MRAP in serum, plasma, and conditioned media. Immunoblot analysis of initial and eluted fractions failed to detect any a2MRAP in these fractions. Thus, a,MRAP does not appear to accumulate in cell culture media, nor does it appear to be a component of normal human plasma or serum.
Immunoprecipitation of cell extracts prepared from surfacelabeled fibroblasts was performed to establish whether or not n,MRAP is present on the cell surface. The results showed that indeed a radiolabeled band with mobility corresponding to a,MRAP was precipitated with affinity-purified a2MRAP antibodies (Fig. 6). Co-precipitating with the a,MRAP were the 515/85-kDa heavy and light chains of the a2MR suggesting that a2MRAP and the a2MR form a complex, thus supporting previous data documenting their interaction (7, 10). Overall, the results from these experiments suggest that a2MRAP is located on the cell surface and associates with the (u~MR.

DISCUSSION
This paper reports the complete primary structure of a2MRAP, a molecule that co-purifies with the a2MR during affinity chromatography (7). The prominent structural features of a,MRAP include a putative signal sequence, a leucine were subject,ed to immunoprecipitation using 10 pg of affinity-purified anti-trZMRAI' IgG. Immunoprecipitat.es were analyzed by SDS-PAGE on gradient gels using a discontinuous pH gel system (Laemmli) with a 4% polyacrylamide stacking gel and a 4-12.67; separating gel. Following electrophoresis, the gels were dried and exposed to x-ray film.  6. Immunoprecipitation analysis of cell extracts from '2RI-surface-labeled human gingival fibroblasts. Affinity-purifled anti-n,MRAP IgG (5 pg) was used in conjunction with protein G-Sepharose to immunoprecipitate polypeptides from a 0.1% Triton extract of labeled fibroblasts. Immunoprecipitates were analyzed by SDS-PAGE on gradient gels using a discontinuous p H gel system (Laemmli) with a 4% polyacrylamide stacking gel and a 4-12.5% separating gel. Following electrophoresis, the gel was dried and exposed to x-ray film for 48 h. Arrows designate mobility of the 515-kDa, 85-kDa, and 39-kDa components of the n2MR complex.
zipper motif, regions with potential for forming amphipathic helices, a putative N-linked glycosylation site, and a carboxylterminal sequence similar to the endoplasmic retention consensus sequence. Furthermore, analysis of the deduced sequence suggested a relatively high content of a-helix, which was confirmed by circular dichroism measurements.
An unexpected finding in the present study is that arMRAP appears to be the human homologue of a rat polypeptide identified as a Heymann nephritis antigen (16). Heymann nephritis is a rat model of human membranous glomerulonephritis that is characterized by the accumulation of immune deposits within the glomerular basement membrane. One target of autoimmunity in this disease is gp 330 (37), a large glycoprotein that is isolated from renal proximal tubule brush borders. Using antibodies eluted from glomeruli of rats with An examination of the available amino acid sequence derived from sequencing partial cDNAs encoding gp 330 (38) reveals that gp 330 has considerable homology with both LDL receptor and a2MR, clearly indicating that these molecules are members of a receptor family. The homology between these proteins includes the overall organization of structural and functional motifs, such as repeated cysteine-rich regions, a transmembrane domain, and one or more consensus sequences (NPXY) within the cytoplasmic domain known to be involved in internalization of the LDL receptor (39). The interaction of a,MRAP with the a2MR is now well documented (7, lo), and it is possible that a2MRAP interacts with additional members of the LDL receptor family given the similarities between these molecules. The fact that a2MRAP appears to be a homologue of a rat protein shown to be an autoantigen in Heymann nephritis, suggests that this molecule, perhaps in complex with the a2MR, may have some role in the pathogenesis of membranous glomerular nephritis. Given the role of the a2MR in the uptake of a2M-proteinase complexes, and probably certain lipoproteins, it is interesting to note that elevated levels of a2M (40,41) and hyperlipoproteinemia (42)(43)(44) are associated with the nephrotic syndrome.
A characteristic feature of a2MRAP deserves mention in view of the potential involvement of this molecule in membranous glomerular nephritis. Notably, the ability of this molecule to bind heparin is of interest since heparin sulfate proteoglycans are components of basement membranes (45) and elevated expression of proteoglycans, along with transforming growth factor-@ and fibronectin, have been reported in an experimental model of glomerulonephritis (46). Furukawa et al. (17) found that message levels of the murine homologue, HBP-44, were considerably higher in kidneys than in other organs examined. It is apparent that an understanding of the distribution of this molecule, and the a2MR, in normal and pathological kidneys is important for understanding the role of a,MRAP in this disease process. The availability of specific monoclonal antibodies and cDNA probes should greatly facilitate these studies.
Comparative sequence analysis also revealed similarity between a2MRAP and apolipoprotein E. These include the relatively high content of a-helix, the potential of several regions to form amphipathic helixes, and sequence homology with a region of apolipoprotein E that is involved in receptor recognition. The similarity with apolipoprotein E is of interest since exogenously added apolipoprotein E appears to be required for the binding of large cholesterol-containing remnant lipoproteins to LDL receptor-related protein (13, 14), now known to be identical with the a2MR. The structural similarity suggests that a2MRAP may function in a similar manner and mediate the binding of large, cholesterol-containing remnant lipoproteins to a2MR. We have been unable to detect a2MRAP in conditioned media of metabolically labeled cells or in plasma. Further, pulse-chase experiments revealed that the newly formed a2MRAP remains cell-associated, and surface labeling experiments document that this molecule is labeled, suggesting that a2MRAP is located on the cell surface. The detection of a2MRAP on the cell surface is in agreement with the observations of Furukawa et al. (17) who noted a cell-surface staining pattern in indirect immunofluorescence studies when PYS-2 parietal endoderm cells were stained with antibodies prepared against the a2MRAP murine homologue, HBP-44. All of the data obtained thus far support the proposal that a2MRAP is located on the cell surface as a component of the aZMR complex, perhaps representing a new ligand.
If not a new ligand, a2MRAP may in some manner influence ligand binding by the 515/85-kDa polypeptides of the a2MR. Another role of a2MRAP may be in regulating the internalization of receptor-ligand complexes during the process of endocytosis.
The carboxyl terminal sequence of a2MRAP, HNEL, was noted to be similar to the carboxyl-terminal sequence of KDEL that marks proteins for retention in the endoplasmic reticulum (31). Since several variations of the KDEL sequence direct intracellular retention of proteins (47), the HNEL sequence of a,MRAP might serve a similar function. The fact that a2MRAP can be detected on the cell surface is not entirely consistent with the observation that other proteins bearing an endoplasmic retention sequence are retained intracellularly. Given that a,MRAP interacts with a receptor involved in endocytosis, it is possible that interactions involving this carboxyl terminal sequence on a2MRAP play an important role in a2MR trafficking.
In summary, the present investigation has determined the primary structure of a component of the a2MR complex. The results indicate that this molecule can be detected on the cell surface and is a human homologue of a rat protein implicated as an autoimmune antigen in Heymann nephritis. The function of a2MRAP in the a2MR complex, its potential interaction with gp 330 and possibly other members of the LDL receptor family, and its putative role in membranous glomerulonephritis remain to be established.