Analysis of ligand recognition by the purified alpha 2-macroglobulin receptor (low density lipoprotein receptor-related protein). Evidence that high affinity of alpha 2-macroglobulin-proteinase complex is achieved by binding to adjacent receptors.

The molecular basis for binding of alpha-macroglobulin-proteinase complexes to the human two-chain 500/85-kDa (alpha/beta) alpha 2-macroglobulin (alpha 2M) receptor (alpha 2MR)/low density lipoprotein receptor-related protein was analyzed. Ligand blotting experiments showed that a 40-kDa protein, present in the affinity-purified alpha 2MR preparation, is bound to the alpha 2MR alpha-chain and released by heparin. Removal of the 40-kDa protein resulted in a 3-5-fold increase in binding of alpha 2M-trypsin. Nitrocellulose-immobilized pure two-chain alpha 2MR was incubated with human alpha 2M-trypsin, containing four identical subunits, and two monovalent ligands: rat alpha 1-inhibitor-3-chymotrypsin and the 18-kDa receptor binding fragment of the alpha 2M subunit. Binding of alpha 2M-trypsin to the alpha-chain of immobilized alpha 2MR was composed of a high (Kd = 40 pM at 4 degrees C) and a low (Kd = 2 nM) affinity component. alpha 1-Inhibitor-3-chymotrypsin bound to the same sites but with one component (Kd = 0.4 nM). Competition-inhibition experiments and dissociation experiments, using ligands with different valences, as well as experiments with alpha 2MR immobilized at different densities, led to the following model. The low (Kd = 2 nM) affinity of alpha 2M-proteinase is prevalent when only one of the four domains binds to alpha 2MR, i.e. when the receptor density is low or when neighboring receptors are occupied. The high (Kd = 40 pM) affinity is achieved by binding of at least two domains to adjacent receptors.

a,M receptors (a2MR) with apparent dissociation constants (&) varying from 40 to 540 p~ at 4 "C have been described in several cell types such as monocyte-macrophages (12, 13), fibroblasts (14,15), hepatocytes (3,4,16,17), and syncytiotrophoblasts (18). Low affinity binding with an apparent K d = 2-100 nM has also been reported (3,18-21). a,MR was recently purified by ligand affinity chromatography from rat liver (22) and human placenta (23,24) as an approximately 500-kDa protein with associated components at 85 and 40 kDa. The 500-kDa protein contains the ligand binding site (22, 23) and high affinity Ca2+-binding sites important for receptor conformation and ligand recognition (25). Sequence analyses of tryptic and cyanogen bromide fragments (26,27) and of amino termini (27) of the 500-and 85-kDa components have recently revealed that a,MR is identical with the low density lipoprotein receptor-related protein presumed to be a receptor for chylomicron remnants and @-migrating very low density lipoproteins (28)(29)(30)(31). Pulsechase experiments with ["Slcysteine have shown that a2MR/ low density lipoprotein receptor-related protein (here designated a2MR) is synthesized as an approximately 600-kDa precursor protein. This is cleaved in a tram-Golgi compartment to generate the two-chain molecule with the 85-kDa membrane-spanning @-chain anchoring the approximately 500-kDa a-chain noncovalently (32). Amino-terminal sequencing of the 40-kDa protein has shown that it is of distinct genetic origin (27).
The purpose of the present study was to analyze binding of the 40-kDa protein to azMR and to elucidate how binding of a2M-proteinase to the two-chain a2MR could result in both high and low affinity binding. We performed binding experiments using anMR immobilized onto nitrocellulose at different densities and the following ligands: human a2M-trypsin, containing four binding domains; al-inhibitor-3-chymotrypsin (al13-chymotrypsin), a rat homologue of the human a,M monomer containing one binding domain (33-35); and the 18-kDa carboxyl-terminal fragment of the a2M-subunit containing the receptor recognition domain (36, 37).

EXPERIMENTAL PROCEDURES
Preparation of Macroglobulins-Human a n M was prepared from pooled citrate plasma using Zn" chelate-affinity chromatography as earlier described in detail (38). Purification of the carboxyl-terminal tr,M fragment containing the receptor recognition domain was carried out by papain digestion of methylamine-treated a2M (37). Rat alIs was prepared from pooled EDTA plasma from 200-g male Wistar rats as described (34,39). The ligands were iodinated with 1 mol of '"1 (Amersham, Ltd., United Kingdom)/mol of 200-kDa alia or 180-kDa tryM subunit, as described (3). alia was reacted with 0.2 M of methylamine for 2 h a t 20 "C, dialyzed, and complexed with chymotrypsin (Worthington) by incubation of ligand and proteinase at the ratio 1:l for 2 min a t 20 "C. The enzyme activity was stopped by adding phenylmethanesulfonyl fluoride to a final concentration of 1 mM. tuyM was complexed with trypsin as described earlier (3). The complexed proteins were purified either by Sephacryl S-300 gel filtration or Superose 12 fast protein liquid chromatography. All preparations of native and proteinase-complexed macroglobulins were controlled by reducing and nonreducing sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). More than 80% of the radiolabeled ligands were able to bind to a2MR as tested in the ligand binding assay (see below).
Preparation of Human Placental a2MR and the 40-kDa Associated Protein-Human a2MR was purified by ligand-affinity chromatography of detergent-solubilized membranes from human placentas (23,25). The two-chain a2MR (200-500 pg/ml) was further purified by shaking incubation for 16 h at 4 "C, pH 8.0, with 25 mg of heparin-Sepharose CL-GB (Pharmacia LKB Biotechnology Inc., Sweden)/ml. This treatment removed more than 95% of the associated 40-kDa protein.
In some experiments, two chain a,MR was further purified by Mono-Q (Pharmacia) anion exchange chromatography, as described (24). The 40-kDa associated protein was purified by Mono-S (Pharmacia) cation exchange chromatography of the a2MR preparation, as described (26). The purified proteins were iodinated according to the chloramine-T method, as previously described (3,22). Triton X-114 phase separation of receptor subunits and the 40-kDa protein was performed essentially by the method of Bordier (40). The protein components were treated with 0.2% SDS for 10 min at 20 "C immediately prior to phase separation in 50 volumes of 1% Triton X-114.
Preparation of Polyclonal and Monoclonal Antibodies-Polyclonal rabbit antibodies against the a?MR and 40-kDa protein were raised by immunizing a rabbit with 0.1 mg of a2MR in 50 pl of Freund's incomplete adjuvant with 4-week intervals (2-week interval for the first 8 weeks). For production of monoclonal antibodies, mouse myeloma cells (NS-1) were fused with spleen cells from mice immunized with affinity-purified human aZMR, as described earlier (41). Screening for positive clones was carried out after 7 days by an indirect enzyme-linked immunosorbent assay. Positive clones were isolated twice by limiting dilution. Class and subclass specificities of the antibodies were determined using an anti-mouse antibody typer kit (Bio-Rad, 172-2055). The novel antibodies A2MR@-1 (against a?MR &chain) and S4-D5 (against the 40-kDa protein) belonged to the IgG, subclass and IgM class, respectively. The IgGl antibody, A2MRa-2, against the a,MR a-chain has previously been described as A2MR-2 (25,41). Polyclonal and monoclonal antibodies were purified by protein G-affinity chromatography according to the instructions by the manufacturer (Pharmacia) of protein G-Sepharose. Neither the monoclonal antibodies directed against the a-chain or the @-chain interfered with binding of a2M-proteinase (Ref. 41, data not shown).
Measurement of a?MR and the 40-kDa Associated Protein-Protein concentrations were determined by measuring absorbance a t 280 nm (25,26) and by the following indirect double-layer ("sandwich") enzyme-linked immunosorbent assay. Microtiter wells from NUNC (Denmark) were coated for 16 h a t 4 "C with 100 pl of polyclonal rabbit anti-a2MR antibodies (40 mg of IgG/ml in 50 mM NaHCOa, pH 9.5) directed against the affinity-purified a2MR preparation. After washing three times and blocking with 0.2% nonfatty milk, a2MR preparation was applied a t varying dilutions. The different constituents in the n2MR preparation were then detected with mouse monoclonal antibodies (A2MRa-2, A2MR@-1, or S4-D5, 2 pg/ml) and peroxidase-conjugated rabbit antimouse antibody (diluted 1/1000) from Dakopatts (Denmark). The incubations with antigens and antibodies were carried out for 2 h a t 20 "C in phosphate-buffered saline (PBS; 10 mM NaH2P04, 150 mM NaC1, and 0.6 mM CaCI?, pH 7.4) containing 0.2% nonfatty milk.
Immunoprecipitation-10" cpm of 12sII-labeled azMR (15 ng) was immunoprecipitated with 5 pg of A2MRa-2 linked to Sepharose 4B (Pharmacia). After incubation for 1 h a t 4 "C, the antibody-Sepharose beads (2 mg of protein/ml) were sedimented by a brief spin (15 s) in a microcentrifuge, washed 5 times in PBS, and boiled for 2 min in the sample buffer for SDS-PAGE.
Blotting Procedures-SDS-PAGE was performed in 100 x 100-mm and 1-mm-thick slab gels with a 4-16% polyacrylamide gradient and 2% SDS in the sample buffer. The gels were stained with Coomassie Brilliant Blue. Proteins were electroblotted onto nitrocellulose (Sartorius, Germany) using a semidry electroblotting system (JCA aps, Denmark). Nitrocellulose strips were blocked for 2 h a t 20 "C with 2% nonfatty milk in PBS. For immunoblotting, the strips were incubated for 16 h a t 4 "C with monoclonal antibody (2 pg/ml) in PBS with 2% nonfatty milk (buffer A), washed 3 X 10 min in PBS containing 0.05% Tween 20 (buffer B), incubated for 2 h a t 20 "C with 2 ~1 0 % cpm of "'I-labeled rabbit anti-mouse antibody (4000 cpm/ng) in buffer A, and finally washed 3 X 10 min in buffer B. For ligand blotting with the l"'II-labeled 40-kDa protein, the strips were incubated for 16 h at 4 "C in buffer A with 2 X 10" cpm of "'I-protein (5000 cpm/ng) and washed 3 X 10 min in buffer B. Autoradiography of immuno-and ligand blots was performed on Amersham Hyperfilm for 24-48 h a t -70 "C using an intensifying screen.
Binding of a-Macroglobulins-Binding of radiolabeled ligands to placental membranes (approximately 300 pg/ml) was performed in 200 pl of incubation buffer containing 1% bovine serum albumin as previously described (23). Binding was measured after 18 h a t 4 "C by pelleting the membranes in microcentrifuge tubes. Ligand binding to the n,MR immobilized to nitrocellulose (Sartorius) discs was measured essentially as described (25). 25 fmol (if not stated otherwise) of aZMR in 2 X 2 pl was applied to the discs, followed by drying for 30 min a t 4 "C and soaking for another 30 min in the incubation buffer containing 1% Tween 20. The filters were washed in buffer and incubated with ligand in 200 pl for 16 h at 4 "C with constant shaking. Thus, the a,MR concentration was 125 PM, if not otherwise indicated. Control experiments with radiolabeled a2MR showed more than 99% association to the nitrocellulose and no loss during the incubation procedure. The incubations were stopped by rapid wash in 3 X 5 ml of ice-cold buffer, and the discs were assayed for radioactivity. 1-276 of the radioactive ligand was associated with the filters dotted with eluate containing no receptor. The same value was obtained when filters were incubated in the presence of a saturating concentration of ligand. This blank value was measured in all experiments and subtracted from all values.

RESULTS
Properties of the Receptor Preparation-Monoclonal antibodies were used for identification and enzyme-linked immunosorbent assay measurements of the a2MR a-and pchains and the 40-kDa protein present in the azMR preparation. Fig. 1 shows immunoblots using A2MRa-2 directed against the a-chain (25,41), and two new antibodies, A2MRp-1 directed against the p-chain and S4-D5 directed against the 40-kDa associated protein. No single chain azMR (approximately 600 kDa) was detected in the preparation by A2MRa-2 or A2MRP-1.
The following experiments were performed to further characterize the 40-kDa protein and its binding to a2MR.   preparation using A2MRa-2. The coprecipitation of the constituents confirms the tight binding between two-chain apMR and the 40-kDa associated protein. Lanes 2 and 3 show that the membrane-spanning P-chain, in contrast to the a-chain, was extracted by Triton X-114, whereas the 40-kDa protein was not, indicating a hydrophilic nature for this molecule. The 40-kDa protein was removed from the a2MR preparation by incubation with heparin-Sepharose (lane 4 ) . Control experiments with incubation of Sepharose alone showed no removal (data not shown). Lune 5 shows purification of the 40-kDa protein using retention on a Mono-S cation exchange column, followed by elution a t high salt. Two-chain atMR was retained on an Mono-Q anion exchange column (data not shown). Fig. 3, lane 1, shows by ligand blotting technique that the a-chain bound '*'I-labeled 40-kDa protein. Addition of heparin completely abolished this binding (lane 2 ) , whereas EDTA had no effect (lane 3 ) . This is in contrast to the complete inhibition of the Cat+-dependent apM-trypsin binding by EDTA (25). Reduced a2MR did not bind the 40-kDa protein (lane 4 ) .
The question arose whether the 40-kDa protein might interfere with a2M-proteinase binding to the a2MR n-chain. Fig. 4 shows that its removal from the apMR preparation with heparin-Sepharose resulted in a high increase in binding of 5 pM '""I-apM-trypsin. In four experiments, using different npMR preparations, binding of 1PRI-a2M-trypsin was increased 3-5-fold. Incubation of the pure two-chain a2MR with 40-kDa protein reduced '2"I-a2M-trypsin binding by up to 90% with half-maximal inhibition at approximately 200 PM, and heparin abolished this inhibition. Conversely, a2M-trypsin (200 nM) reduced the binding of "'I-labeled 40-kDa protein, although by only about 50% (data not shown). Thus, the 40-kDa protein is a ligand of the a-chain that clearly interferes with binding of apM-proteinase. The affinity-purified a2MR preparation is, to a large extent, occupied with the 40-kDa protein, and the experiments described below were therefore performed using apMR preparations further purified by incubation with heparin-Sepharose and essentially devoid of the 40-kDa associated protein (cf. Fig. 2, lane 4 ) .
Properties of the Ligands -Fig. 5, lane 1, shows that apMtrypsin migrates both as a dimeric (approximately 360 kDa) and a tetrameric species when subjected to nonreducing SDS-PAGE. The reason is that the proteinase molecules establish covalent cross-links between lysine residues of the proteinase and glutamyl residues of the cleaved apM thiol esters, either within or between a2M half-molecules (42). apM-trypsin migrates as a single band in nondenaturating gels (43). Moreover, electron microscopic examination of the present preparation demonstrated a homogeneous population of the typical H-shaped complexes (data not shown). Thus, the apM-trypsin complexes present in the incubation buffers all had four receptor-binding domains. However, in view of the shape of the complex and the location of the receptor-binding domains at the tip of each "leg" (44), it seems likely that a2M-trypsin is functionally divalent when interacting with receptors on a the Purified a2-Macroglobulin Receptor cell surface or immobilized onto nitrocellulose. Lune 2 shows SDS-PAGE of the 200-kDa al13-chymotrypsin complex. This is homologous to the azM subunit and has one receptorbinding domain. Lane 3 shows the 18-kDa carboxyl-terminal fragment of azM previously shown to possess receptor binding activity (36, 37), i.e. a monovalent analogue. The slightly faster migrating band is due to carbohydrate heterogeneity (42).
Receptor Binding Affinities of al13-Chymotrypsin and aZM-Trypsin- Fig. 6 shows the Scatchard transformations of binding of al13-chymotrypsin and a2M-trypsin, respectively, to nitrocellulose-immobilized two-chain a2MR. The curve representing a113-chymotrypsin binding is linear, in agreement with binding to a single class of receptors, whereas the curve representing azM-trypsin binding is concave upward, giving the illusion of at least two classes of receptors or negative cooperativity. The intercepts with the abscissa show binding of about 0.6-0.7 mol of ligand molecule/mol of receptor. The dissociation constant ( K d ) for binding of a113-chymotrypsin was determined as 440 k 65 pM in five experiments. By fitting a two-receptor model to the data for azM-trypsin, the apparent Kd values were calculated to 37 f 9 pM and 2.2 k 0.4 nM ( n = 5). Indistinguishable results were obtained when azMchymotrypsin or aZM-methylamine were used as ligands (data not shown). Results closely similar to those shown in Fig. 6 were obtained when measurements were performed with a2MR bound to immobilized monoclonal antibody A2MRa-2 and when experiments were performed using placental membrane fractions (data not shown). Fig. 7A shows that azM-trypsin at high concentrations inhibited binding of 10 PM 'Z51-al13-~hymotrypsin with halfmaximal inhibition at 2 nM, i.e. equivalent to K d for the apparent low affinity azM-trypsin binding. The 18-kDa fragment inhibited binding of '2sI-al13-chymotrypsin according to a one-receptor model, and about 100 nM was required for halfmaximal inhibition. Fig. 7B shows that 0.4 nM a1I3-chymotrypsin (or 100 nM 18-kDa fragment) was required to inhibit binding of 5 PM 1251-azM-trypsin half-maximally.
Dissociation of al13-Chymotrypsin and azM-Trypsin-Dissociation of prebound '251-al13-~hymotrypsin followed a single exponential curve (0-7 h) in the absence or presence of unlabeled ligands at large concentrations (Fig. 8A). On the other hand, most of the bound 'z51-azM-trypsin dissociated very slowly in the absence of unlabeled ligand, and the process was markedly accelerated by excess unlabeled ligand (Fig.  8B). Experiments with placental membranes gave similar results. Dissociation of azM-trypsin was markedly accelerated, whereas dissociation of alI3-chymotrypsin was not (data not shown).
Relationship between Receptor Density and Binding of Labeled Ligand-The following experiments were performed because the different behavior of al13-chymotrypsin and azMtrypsin might be explained by their different valences. We therefore wanted to determine whether azM-trypsin might bind to two sites at the same receptor or to adjacent receptors. If the latter was the case, binding of azM-trypsin, but not al13-chymotrypsin, should depend on the density of nitrocellulose-immobilized a2MR. Fig. 9 shows that binding of Iz5I-al13-chymotrypsin increased proportionally with the receptor concentration (density), whereas binding of 12sI-azM-trypsin was comparatively lower at low receptor densities and higher at high receptor densities.

DISCUSSION
The present data show that the 40-kDa molecule is a heparin-binding protein that binds to the azMR a-chain and inhibits binding of 5 PM '2sI-a2M-trypsin by up to 90%. Ligand (azM-methylamine) affinity-purified azMR preparations contained enough associated 40-kDa protein to reduce azMtrypsin binding by up to 80%. The nature of the mutual interaction between binding of the two ligands remains to be elucidated; they may interact with overlapping sites or binding of the 40-kDa protein may change the overall affinity of ~z M R to azM-proteinase by an allosteric mechanism. The function of the 40-kDa protein is presently not known. Two-dimensional gel electrophoresis and immunoblotting of cell lysates have revealed* that the monoclonal antibody, S4-D5, cross-reacts with another heparin-binding protein, a 70-kDa nuclear protein that recognizes the 3' splice site of premessenger RNA (47). Electron microscopic immunocytochemistry has shown antigen recognition in the nucleus and cytoplasm, but not on the cell surface. tetrameric a-macroglobulin-proteinase to a2MR. The monomeric culI,t-chymotrypsin and tetrameric npM-trypsin molecules are drawn schematically according to their electron microscopic ultrastructure (45, 46). n,MR 0-and @-chains are drawn in "cigar" forms according to their molecular volumes since no ultrastructural information is yet available. A, monovalent binding of nlI:l-chymotrypsin. B, monovalent binding ("one leg") binding of nrM-trypsin thought to occur predominantly when receptors are scattered or neighboring receptors are occupied. C, dimeric binding of nnM-trypsin thought to occur predominantly to receptors a t high densities and a t low occupancies.
protein may have functions normally unrelated to the asMR and may be associated with the a-chain in the process of a2MR purification. The heparin-releasable association of the 40-kDa protein is interesting in view of the reported binding of apolipoprotein E-rich lipoproteins and liposomes to a2MR (under the name low density lipoprotein receptor-related protein (29)(30)(31)). Clusters of basic amino acids in apolipoprotein E (residues 142-158) are essential for heparin-releasable binding to the cysteine-rich repeats of the low density lipoprotein receptor (48), which is highly homologous to the cysteine-rich repeats of a2MR (28). Thus, the 40-kDa protein and apolipoprotein E-rich lipoproteins might share binding domains in the arMR.
The present results provide the first data for binding of aM-proteinase complexes to two-chain a2MR without interference of other proteins. The distinctive high affinity binding of tetrameric apM-proteinase is explained according to the model shown in Fig. 10 A-C. Due to the H-shaped structure of the complex, dimeric binding of a2M-proteinase seems most likely, but binding of some complexes to more than two receptors cannot be excluded. It is important to note that the monomeric alIn-chymotrypsin (Fig. 1OA) binds according to a "simple" model, i.e. a linear Scatchard plot with a stoichiometry indicating that most of the receptors are in a conformation capable of binding the ligand (see Fig. 6 and its legend), dissociation following an exponential course with or without excess unlabeled ligand in the medium (Fig. 8A), and binding proportional with the receptor density (Fig. 9). These are necessary premises for drawing conclusions from the more complicated binding of azM-trypsin. The upward concave Scatchard plot (Fig. 6) is explained as the result of primarily dimeric and, thus, high affinity binding a t low occupancies and primarily monomeric and low affinity binding at high occupancies. The acceleration of '"I-labeled a2M-trypsin dissociation in the presence of excess monomeric or dimeric ligand (Fig. 8B) is explained as the result of frequent dissociation of either one of the two legs, followed by binding of unlabeled ligand to the exposed receptor; the labeled molecule is now bound in the low affinity form and will dissociate rapidly. The low binding of a2M-proteinase at low receptor density (Fig. 9) is explained as the result of infrequent occurrence (Fig. 10B) of pairs of receptors capable of binding two legs (Fig. 1OC). Finally, the model is also supported by the finding that a,M-trypsin is a poor competitor of al13-chymotrypsin binding (Fig. 7 A ) , i.e. corresponding to its low affinity constant.
The findings that the 40-kDa protein interferes with azMtrypsin binding and that the overall affinity depends on the density of immobilized azMR could explain previously reported apparent affinities for a,M-proteinase in the nanomolar range (24) and around 500 PM (22,23). The dimeric binding model also helps to explain previous observations with the receptor binding domain (18-kDa fragments) of amacroglobulins. Enghild et al. (49) reported a Kd for the alIsderived fragment of 10 nM, i.e. an affinity of 2-3% of that of the native molecule (Fig. 6). The Kd for the a2M-derived fragment is about 100 nM (Fig. 7) (37, 49), i.e. about 2% of the affinity of a2M-proteinase when binding in the monomeric mode. Thus, the higher intrinsic affinity of a,IS-chymotrypsin appears to be reflected in its receptor-binding domain.
When a2MR is in a solubilized state, a2M-methylamine should theoretically not be able to gain the affinity advantage of divalent binding characteristic of a2MR immobilized to a solid support. al13-chymotrypsin should therefore be more suitable for affinity chromatography due to its higher intrinsic affinity. Accordingly, we have found that the receptor output from a al13-chymotrypsin-Sepharose column was more than twice as high as that from a parallel column with the same amount of immobilized azM-methylamine. 3 The dimeric high affinity mode of a,M-trypsin binding is important also in cell membranes, as judged from the nearly identical Kd values obtained when using human placental membranes. Moreover, high and low affinity binding of azMtrypsin (3) and a relatively poor competition of a2M-trypsin with '2sI-al13-chymotrypsin (34) have been reported in rat hepatocytes. The dimeric binding mode requires a high density of receptors, and electron microscopic gold immunocytochemistry has revealed high concentrations of a2MR in coated pits of human fibroblasts (41). Thus, the high affinity may be a mechanism for modulating clearance of a,M-proteinase complexes that may bind preferentially to receptors located in coated pits.
The dimeric mode of binding resembles that of IgG molecules obtaining a "bonus" of high affinity through binding of the two Fab fragments to two pairs of surface antigens (50). It is also known that interleukin-2 obtains a high affinity through binding to different neighboring receptors, Tac and non-Tac proteins (51). The a,M-proteinase complex is an unusual ligand, with four exposed receptor binding domains, at least two of which are capable of binding to receptors S. K. Moestrup and L. Sottrup-Jensen, unpublished results. suitably distributed at the cell surface. Its dimeric binding is the basis for nonlinear Scatchard plots and for ligand-induced accelerated dissociation. This phenomenon is frequently observed in receptor biology and explained as negative cooperativity among receptors. It seems possible that cobinding of two (or more) sites a t one ligand to distinct sites a t cell surface molecules may explain accelerated dissociation in other receptor-binding systems.
In conclusion, we have provided evidence that azM-proteinase can bind both with high and low affinity to a2MR; high affinity when binding is dimeric to adjacent receptors, and low affinity when binding is monomeric due to scattered receptors or high receptor occupancy.