Molecular Basis for the Interaction of Low Density Lipoprotein Receptor-related Protein 1 (LRP1) with Integrin αMβ2

The LDL receptor-related protein 1 (LRP1) is a large endocytic receptor that controls macrophage migration in part by interacting with β2 integrin receptors. However, the molecular mechanism underlying LRP1 integrin recognition is poorly understood. Here, we report that LRP1 specifically recognizes αMβ2 but not its homologous receptor αLβ2. The interaction between these two cellular receptors in macrophages is significantly enhanced upon αMβ2 activation by LPS and is mediated by multiple regions in both LRP1 and αMβ2. Specifically, we find that both the heavy and light chains of LRP1 are involved in αMβ2 binding. Within the heavy chain, the binding is mediated primarily via the second and fourth ligand binding repeats. For αMβ2, we find that the αM-I domain represents a major LRP1 recognition site. Indeed, substitution of the I domain of the αLβ2 receptor with that of αM confers the αLβ2 receptor with the ability to interact with LRP1. Furthermore, we show that residues 160EQLKKSKTL170 within the αM-I domain represent a major LRP1 recognition site. Given that perturbation of this specific sequence leads to altered adhesive activity of αMβ2, our finding suggests that binding of LRP1 to αMβ2 could alter integrin function. Indeed, we further demonstrate that the soluble form of LRP1 (sLRP1) inhibits αMβ2-mediated adhesion of cells to fibrinogen. These studies suggest that sLRP1 may attenuate inflammation by modulating integrin function.

The LDL receptor-related protein 1 (LRP1) 3 is a large endocytic receptor that recognizes numerous ligands (1,2). LRP1 is synthesized as a single chain 600-kDa precursor that is processed by furin to generate a 515-kDa heavy chain and an 85-kDa light chain (3). The heavy chain contains four clusters of LDL receptor type A or "ligand binding" repeats, which are responsible for recognizing most of the ligands that bind to LRP1. The LRP1 light chain includes the transmembrane and cytoplasmic domain, which contains two NPxY motifs, and two dileucine repeats. The second NPxY motif overlaps with a YxxL motif, and along with the two dileucine repeats, contributes to LRP1 endocytosis (4).
In addition to binding soluble ligands, such as ␣ 2 macroglobulin-protease complexes (5) or various serpin-enzyme complexes (6), LRP1 also associates with other transmembrane proteins. These include the amyloid precursor protein (7,8), the platelet-derived growth factor receptor-␤ (9 -12), and ␤ 2 integrins (13,14). The association of these cell surface transmembrane proteins with LRP1 alters their trafficking and functional properties. In the case of ␤ 2 integrins, Spijkers et al. (13) found that in U937 cells, ␣ M ␤ 2 co-precipitates with LRP1, which regulates its cell adhesion properties. Cao et al. (14) found that LRP1 co-localizes with ␣ M ␤ 2 at the trailing edge of migrating macrophages and further demonstrated that macrophage migration depends upon a coordinated effort between LRP1 and ␣ M ␤ 2 along with tissue plasminogen activator and its inhibitor, plasminogen activator inhibitor-1. Together, these studies reveal that LRP1 modulates the function of ␣ M ␤ 2 .
To gain insight into the role of LRP1 in modulating ␣ M ␤ 2 function, it is necessary to define the molecular basis for the interaction between these two molecules. The objective of the current investigation was to identify regions on LRP1 and on ␣ M ␤ 2 that are important for mediating their interaction. To accomplish this, we employed LRP1 mini-receptors, and based on our observation that LRP1 interacts preferentially with ␣ M ␤ 2 but not ␣ L ␤ 2 , we also employed homolog-scanning mutagenesis. In this process, regions from the ␣ L ␤ 2 -I domain were swapped into the ␣ M ␤ 2 -I domain within the heterodimeric receptor. Together, the results identify multiple determinants on LRP1 responsible for binding to ␣ M ␤ 2 and identify a region within the ␣ M ␤ 2 -I domain that contributes to the interaction.

EXPERIMENTAL PROCEDURES
Materials and Antibodies-Hybridomas expressing the anti-Myc monoclonal 9E10, and the anti-␣ M -I domain antibodies LM2/1 and 44a were obtained from ATCC. Rabbit polyclonal antibody ARC22 directed against the cytoplasmic domain of * This work was supported, in whole or in part, by National Institutes of Health Grants P01 HL054710 (to D. K. S. and L. Z.) and HL050784 and HL072929 (to D. K. S.). 1  the ␤ 2 subunit has been described (14). Monoclonal antibody 8G1 directed against human LRP1 has been described (5), and anti-LRP1 R2629 has been described (9). The ␣ M -I domain was prepared as a fusion protein with GST as described (15). Prior to use, GST was removed by proteolysis to generate the free ␣ M -I domain. Human kidney 293 cells stably expressing ␣ M ␤ 2 or ␣ L ␤ 2 were prepared as described (16,17), whereas human kidney 293 cells expressing mutant ␣ M ␤ 2 or ␣ L ␤ 2 molecules were prepared as described (17). LRP1 mini-receptors were prepared as described (18). Receptor-associated protein (RAP) was prepared as described (19). Transient Transfection of Human Kidney 293 Cells and Coimmunoprecipitation Experiments-Human kidney 293 cells stably transfected with ␣ M ␤ 2 , ␣ L ␤ 2 , or mutant forms of these integrins were grown to 70% confluency and transiently transfected using FuGENE HD transfection reagent (5 g of plasmid DNA/100-mm plate) with either N-terminal Myc-tagged LRP1 light chain (LC), N-terminal Myc-tagged mini LRP1 receptor containing the second cluster of ligand binding repeats (mLRP-II), N-terminal Myc-tagged mini LRP1 receptor containing the fourth cluster of ligand binding repeats (mLRP-IV), or vector control. 40 h following transfection, cell lysates were prepared by adding lysis buffer (50 mM Tris, 150 mM NaCl, 1% Nonidet P-40) containing protease and phosphatase inhibitors. Lysates were precleared with nonimmune IgG-protein G-Sepharose and immunoprecipitated overnight with anti-Myc monoclonal 9E10 IgG (7.5 g/ml) and protein G-Sepharose. After washing in lysis buffer three times, the immunoprecipitates were separated on 4 -12% SDS-PAGE and transferred onto nitrocellulose membranes. The membranes were probed for ␤ 2 subunit using ARC22 (1 g/ml). To measure the immunoprecipitated LRP1, the blots were probed with 125 I-labeled 9E10 IgG.
Binding of LRP1 to ␣ M -I Domain-The purified ␣ M -I domain was coated on 96-well flat-bottomed microtiter plates in 50 mM Tris and 150 mM NaCl (TBS) containing 2 mM CaCl 2 overnight at 4°C. The wells were then washed with 1% BSA in TBS and 2 mM CaCl 2 for 1 h at room temperature. Increasing concentrations of purified LRP1 were then added in 1% BSA, TBS, 2 mM CaCl 2 , and 0.05% Tween 20 and were allowed to incubated overnight at 4°C with the immobilized ␣ M -I domain. Following washing, bound LRP1 was detected with monoclonal antibody 8G1 (0.5 g/ml) and horseradish peroxidase-conjugated goat anti-mouse IgG antibody. Wells were developed using tetramethylbenzidine peroxidase substrate (KPL, Gaithersburg, MD), and absorbance was measured at 590 nm. The binding data were fit to a binding model with a single class of sites using nonlinear regression analysis using SigmaPlot software as described (20).
Purification of Soluble LRP1 (sLRP1)-sLRP1 was purified from human plasma by affinity chromatography over monoclonal antibody 8G1-Sepharose followed by ion-exchange chromatography over a Mono-Q anion exchange column. Briefly, 1 liter of fresh frozen plasma was thawed, and the following were added: 1 mM CaCl 2 , 1 M Phe[D]-Pro-Arg-chloromethylketone, 20 g/ml PMSF, 40 g/ml pepstatin, 1.25 g/ml leupeptin, and 10 g/ml benzamidine. The plasma was applied to a 5-ml Sepharose Cl 6B precolumn and then applied to a 10-ml 8G1-Sepharose column. The column was washed with TBS, pH 7.5, 20 mM EDTA and then washed with TBS, pH 7.5, containing 1 mM CaCl 2 . sLRP1 was eluted with 0.1 M sodium acetate, 0.5 M NaCl, pH 4.5. The sLRP1 was then applied to a mono-Q anion exchange column and eluted as described previously for full-length LRP1 (21).
Adhesion Assays-24-Well non-tissue culture polystyrene plates were coated with fibrinogen (100 l, 50 g/ml). After 90 min at 22°C, 400 l of blocking buffer (1% BSA containing 0.05% polyvinyl-pyrrolidone in PBS) was added. After washing twice with PBS, 2 ϫ 10 6 293 cells stably transfected with ␣ M ␤ 2 in 400 l of Hank's buffered salt solution containing 5 mM HEPES, 1 mM CaCl 2 , and 1 mM MgCl 2 were incubated with or without sLRP1 (20 nM) for 10 min at room temperature. The cells were then added to fibrinogen-coated wells. After incubating at 37°C for 30 min, non-adherent cells were removed by washing. The number of adherent cells were quantified by crystal violet staining and measuring the absorbance at 570 nm.
Effect of Integrin Activation of Association with LRP1-Thioglycolate-ellicited macrophages were placed into four different tubes (2.5 ϫ 10 6 cells/tube). Two tubes were placed on ice and served as control cells. LPS (500 ng/ml) was added to the remaining two tubes, which were then incubated at 37°C for 30 min. Cell extracts from all four tubes were used for immunoprecipitation with 10 g of mouse IgG as a control or 10 g of anti-LRP1 monoclonal antibody 5A6. Following immunoprecipitation, proteins were separated on a 4 -12% Tris glycine gel and analyzed for the presence of ␤ 2 integrin and LRP1 by immunoblot analysis.
Measurement of ␣ M ␤ 2 Internalization in LRP1-expressing and LRP1-deficient Macrophages-Mice with LRP1 deleted in macrophages, macLRP1 Ϫ/Ϫ , were generated by crossing LysMCre mice (22) (kindly provided by I. Förster) with LRP1 flox/flox mice (23) (kindly provided by J. Herz) as described (24). To prepare primary macrophages, bone marrow was flushed from the femur and tibia with DMEM medium and dispersed into single-cell suspension. After lysis of red blood cells with ammonium chloride (8.3 g/liter in 10 mM Tris-HCl, pH 7.4), bone marrow cells were plated in 10-cm tissue culture Petri dishes in DMEM/10% fetal bovine serum and incubated in a humidified incubator with 5% CO 2 at 37°C for 2 to 4 h. Suspension cells were collected and cultured in 10-cm Petri dishes at a density of 2 ϫ 10 5 cells/ml in DMEM with 10% FBS and 10% L929 cell-conditioned medium at 37°C and 5% CO 2 for 7 days. The maturity and purity of the differentiated macrophages were verified by flow cytometry based on their positive staining for F4/80 and M1/70, and negative staining for CD11c and by morphological examination of Hema3 (Fisher Scientific)stained Cytospin (Shandon) smears. Internalization of ␣ M ␤ 2 was performed using FACS analysis as described (14). Briefly, 10 6 macrophages in Hank's balanced salt solution plus 1 mM Mg 2ϩ were incubated with 10 ng/ml LPS and then stained with mAb M1/70 for 60 min at 4°C. The temperature was then raised to 37°C for indicated time periods to allow integrin internalization. At the indicated times, the cells were chilled to 4°C. ␣ M ␤ 2 remaining on the cell surface was measured by using an Alexa 488 conjugate of anti-rat IgG at 4°C. Single color FACS analysis was performed for WT and deficient cells using

RESULTS
Multiple Determinants within LRP1 Are Involved in Interaction with ␣ M ␤ 2 -To identify regions within LRP1 that are responsible for interacting with ␣ M ␤ 2 , we used LRP1 mini-receptors ( Fig. 1A), which have been successfully employed to map out ligand binding sites for this receptor (18,25). Human kidney 293 cells stably expressing ␣ M ␤ 2 (17) were transfected with empty vector, or plasmids expressing the LRP1 light chain (LC), LRP1 mini-receptor II (mLRP-II) or LRP1 mini-receptor IV (mLRP-IV). Following transfection, cell lysates were subjected to immunoprecipitation using anti-Myc IgG. The immunoprecipitates were then analyzed by immunoblotting with ARC22, an antibody specific for the ␤ 2 subunit (14). The results (Fig. 1B) reveal that ␣ M ␤ 2 co-immunoprecipitates with the LRP1 LC as well as mLRP-II and mLRP-IV. To confirm that the LRP1 constructs were expressed, the immunoblots were also probed with 125 I-labeled 9E10 (Fig. 1B, middle panel). These results reveal a somewhat lower expression of mLRP-II and mLRP-IV when compared with the LC, and from this, we concluded that the LRP1 LC binds more weakly to ␣ M ␤ 2 than do the LRP1 constructs that express the clusters of ligand binding repeats. These data reveal that determinants within the LRP1 ectodomain play a major role in interacting with ␣ M ␤ 2 , whereas determinants within the LRP1 LC contribute less to the binding interaction.
Because the LRP1 LC contains two NPxY motifs that are recognized by a number of adaptor proteins (26), which in turn, may also interact with the integrin cytoplasmic domain, we performed experiments using various mutant LC molecules in which either the first NPxY motif (NPxY-1) or second NPxY (NPxY-2) motifs were mutated. We also used a mutant LC, termed 16T3S, in which the serine and threonine phosphorylation sites on the LC were mutated (27). The results from this experiment (Fig. 1C) reveal that ␣ M ␤ 2 co-immunoprecipitated with all of the mutant LC molecules as effectively as WT LC suggesting that the NPxY motifs do not appear to contribute to integrin binding.
LRP1 Appears Specific for ␣ M ␤ 2 and Does Not Interact with ␣ L ␤ 2 -To identify regions on ␣ M ␤ 2 that interact with LRP1, we first examined whether LRP1 is also capable of interacting with ␣ L ␤ 2 . Similar to the ␣ M subunit, the ␣ L subunit also contains an I domain, which has been implicated in the binding of ligands to these integrins (28,29). Human kidney 293 cells stably expressing either ␣ M ␤ 2 or ␣ L ␤ 2 were transfected with constructs expressing either the LC, mini-LRP1-II, or mini-LRP1-IV, and co-immunoprecipitation experiments were performed. The results confirm the experiments in Fig. 1, revealing that ␣ M ␤ 2 co-immunoprecipitates with the LRP1 LC as well as mLRP1-II and mLRP1-IV ( Fig. 2A). In contrast, we note very little coimmunoprecipitation of ␣ L ␤ 2 with the LRP1 LC or with mLRP1-II and mLRP1-IV (Fig. 2B). This is not due to low expression of ␣ L ␤ 2 in the transfected cells, as immunoblot analysis using a ␤ 2 -specific antibody confirmed expression of ␣ L ␤ 2 in the transfected cells (Fig. 2B, lower panel). These results are in contrast to those reported earlier (13), where LRP1 was reported to also associate with ␣ L ␤ 2 .
LRP1 Binds to Purified ␣ M -I Domain-The ␣ M -I domain is responsible for interacting with numerous ligands, including ICAM-1, fibrinogen, C3bi, and neutrophil inhibitory factor, and thus could mediate the interaction of ␣ M ␤ 2 with LRP1. To determine whether LRP1 recognizes this portion of the integrin, solid phase assays were performed in which the purified ␣ M -I domain was first coated to microtiter wells, and then increasing concentrations of purified LRP1 were added to the wells. The results (Fig. 3A) reveal a tight association of purified LRP1 with microtiter wells coated with the ␣ M -I domain. Nonlinear regression analysis was employed to estimate an apparent . Following transfection, cell lysates were immunoprecipitated with anti-Myc IgG, and the immunoprecipitated proteins were analyzed by immunoblotting with ARC22, an antibody specific for the ␤ 2 subunit (upper panels). Total LC and mLRP immunoprecipitated was detected by incubating the blot with 125 I-labeled 9E10 (middle panels). Cell lysates were also immunoblotted for ␤ 2 levels (bottom panels). Data shown are representative of three independent experiments. psec, empty plasmid; WB, Western blot. FIGURE 2. The interaction of LRP1 mini receptors with ␤ 2 integrins is specific for ␣ M ␤ 2 . Human kidney 293 cells stably expressing either ␣ M ␤ 2 (A) or ␣ L ␤ 2 (B) were transfected with constructs expressing the LC, mLRP-II, or mLRP-IV. Following transfection, LC, mLRP-II, and mLRP-IV were precipitated with anti-Myc IgG, and immunoprecipitated proteins were analyzed for ␤ 2 subunit by immunoblot analysis (upper panels). Total LC and mLRP immunoprecipitated was detected by incubating the blot with fluorescently labeled 9E10 (FL-9E10; middle panels). Cell lysates were also immunoblotted for ␤ 2 levels (bottom panels). Data shown are representative of three independent experiments. psec, empty plasmid. K D of 46 nM. As expected, no binding of LRP1 to microtiter wells coated with BSA was detected. As an additional control for these experiments, we also measured the binding of two well characterized conformation-dependent antibodies to the ␣ M -I domain: antibody LM2/1 (30) and 44a (31). The results (Fig. 3B) reveal that both of these antibodies bind avidly to the immobilized ␣ M -I domain, confirming the integrity of the recombinant ␣ M -I domain. In summary, these studies confirm that the ␣ M -I domain is recognized by LRP1.
Reconstruction of LRP1 Binding Site in ␣ L ␤ 2 -The demonstration that ␣ L ␤ 2 does not co-precipitate with mini-receptors of LRP1 afforded the opportunity to test the hypothesis that replacing the ␣ L -I domain with that from ␣ M might restore, at least partially, the LRP1 binding site in the integrin. To test this hypothesis, we used human kidney 293 cells stably transfected with a mutant ␣ L ␤ 2 molecule in which the ␣ L -I domain was replaced with that from ␣ M . The results, shown in Fig. 4, reveal that the hybrid integrin ␣ L (I/␣ M )␤ 2 co-immunoprecipitated with mLRP-IV, although not to the same extent seen with ␣ M ␤ 2 . These results confirm that the ␣ M -I domain contributes to LRP1 binding. The lack of complete restoration of binding by I domain replacement suggests that other regions within the ␣ M ␤ 2 molecule also are involved in binding to LRP1.
Homolog-scanning Mutagenesis Identifies Region on I Domain That Interacts with LRP1-Our prior work investigating the binding of the RAP D3 domain to LRP1 demonstrated an essential role of lysines 256 and 270 in this interaction (32). These two lysine residues are located 21 Å apart on helix ␣8 of RAP (33). A crystal structure of the RAP D3 domain with two LDLa repeats from the LDL receptor revealed that lysines 270 and 256 are each docked within acidic pockets located within the LDLa repeats (34) and provide a model for ligand binding to LRP1 in which lysine residues play a critical role. Thus, to locate binding sites within the ␣ M -I domain, we focused on regions with surface exposed lysine residues. For these experiments, a homolog-scanning mutagenesis strategy was employed. This approach, which has been successfully utilized to identify the neutrophil inhibitory factor-binding site in the ␣ M -I domain (17), involves switching segments within the ␣ M -I domain with those from the ␣ L -I domain. In the current investigation, we focused on regions that are rich in lysine. A loss of binding to mLRP1-IV was noted for a mutant, H1B, as assessed by coimmunoprecipitation analysis (Fig. 5A). In this mutant, residues 162-170 (EQLKKSKTL) were replaced with those (KKL-SNTSYQ) from the ␣ L -I domain. The structure (35) of the ␣ M -I domain is shown in Fig. 5B, and residues 162-170 are located in a loop that connects the ␣1 helix with the B-␤ strand of the I domain (Fig. 5B, red).
␣ M ␤ 2 Integrin Activation Enhances Its Interaction with LRP1-Treatment of macrophages with LPS increases the amount of ␣ M ␤ 2 integrin on the cell surface (36) and also leads to increased activation of the integrin (37). We conducted experiments to determine whether LPS treatment of macrophages would enhance the interaction of ␣ M ␤ 2 with LRP1. The results of this experiment reveal that a substantial increase in ␣ M ␤ 2 integrin co-immunoprecipitated with LRP1 when the cells were treated with LPS (Fig. 6A), revealing that integrin activation enhances the association of ␣ M ␤ 2 with LRP1.
LRP1 Modulates ␣ M ␤ 2 Integrin Function-To test the effect of LRP1 on the functional properties of ␣ M ␤ 2 , we used soluble forms of LRP1 purified from plasma to measure the impact on adhesive properties of cells. In this experiment, 293 cells stably  transfected with ␣ M ␤ 2 were preincubated with 20 nM sLRP1 prior to measuring their ability to adhere to fibrinogen-coated wells. The results of this experiment reveal that sLRP1 markedly reduces the adherence of these cells to fibrinogen (Fig. 7A).
Our prior work demonstrated that LRP1 and ␣ M ␤ 2 co-localize on the cell surface of primary macrophages and that RAP blocks the LPS-stimulated internalization of this integrin (14). As RAP is known to interact with virtually all LDL receptor family members and is thus not specific for LRP1, we used macrophages in which the Lrp1 gene was genetically deleted (24). In the current experiments, macrophages from WT or macLRP1 Ϫ/Ϫ mice were employed. Immunoblot analysis confirmed effective deletion of LRP1 from macrophages (Fig. 7B). Upon LPS stimulation, ϳ20% of ␣ M ␤ 2 was internalized in macrophages expressing LRP1. In contrast, no ␣ M ␤ 2 was internalized in macLRP1 Ϫ/Ϫ macrophages, which lack LRP1 (Fig.  7C). These results establish that in macrophages, LRP1 is important for mediating the internalization of this integrin.

DISCUSSION
During an acute inflammatory response, macrophages accumulate at the injury site where they participate in wound repair processes. To complete this process, inflammation has to be resolved, which requires migration of macrophages from the injury site into the lymphatics in a process dependent upon ␣ M ␤ 2 (38). This process also appears to be regulated by LRP1, which associates with ␣ M ␤ 2 and alters the adhesive properties of this integrin (13) and modulates its ability to mediate cell migration (14). To gain insight into the interaction of LRP1 and ␣ M ␤ 2 at the molecular level, we initiated studies with LRP1 mini-receptors and with mutants of ␣ M ␤ 2 . Our results identify several important insights regarding the interaction of LRP1 with ␤ 2 integrins and the functional consequences of this interaction.
First, our data suggest that LRP1 appears specific for ␣ M ␤ 2 and does not appear to associate with ␣ L ␤ 2 . This is in contrast to the study by Spijkers et al. (13) who found that 500 nM recombinant ␣ L -I domain bound to LRP1 immobilized onto a CM5sensor chip. However, the potential interaction of full-length heterodimeric ␣ L ␤ 2 receptor expressed in mammalian cells with LRP1 was not investigated in the Spijkers et al. (13) study. It is thus possible that the interaction between LRP1 and the ␣ L -I domain is of relatively low affinity, and thus, the weak interaction between ␣ L ␤ 2 and LRP1 is not detected readily by co-immunoprecipitation analysis.  (35)) showing the location of Glu-160 -Leu-170, which is colored red in the structure. This region contains three lysine residues (Lys-165, Lys-166, and Lys-168) that may be involved in the interaction of ␣ M -I domain with LRP1. The figure was generated with PyMOL software. FIGURE 6. LPS stimulation of macrophages increases association of ␣ M ␤ 2 with LRP1. A, peritoneal macrophages were incubated with or without LPS for 30 min at 37°C prior to preparing cell extracts. Cell extracts were immunoprecipitated with control IgG or anti-LRP1 monoclonal 5A6, and the immunoprecipitated proteins were analyzed for the presence of the ␤ 2 subunit and for LRP1 by immunoblot analysis. B, cell extracts were analyzed for total ␤ 2 subunit and for LRP1 by immunoblot analysis. WB, Western blot.  (20 nM). Following incubation, the ability of the cells to adhere to fibrinogen-coated wells was measured. *, p ϭ 0.003. B, primary macrophages isolated from LRP1 ϩ/ϩ or macrophage-specific LRP1-deficient mice (macLRP1 Ϫ/Ϫ ) mice were analyzed for LRP1 expression by immunoblot analysis using antibody R2629. C, ␣ M ␤ 2 internalization upon incubation of macrophages with LPS (10 ng/ml) at 37°C was measured by FACS analysis using an FITC conjugate of anti-rat IgG. The mean fluorescence intensity at time 0 was assigned 100%. The data shown represent the means Ϯ S.D. of a duplicate experiment. *, p Ͻ 0.05. Second, our data reveal that both mini-LRP-II and mini-LRP-IV, which contain the second and fourth ligand binding clusters of LRP1, respectively, bind avidly to ␣ M ␤ 2 as assessed by co-immunoprecipitation analysis. In addition, additional interactions are also evident between the LRP1 light chain and ␣ M ␤ 2 . Together, these data suggest that the major interaction sites on LRP1 involve the ligand binding regions of this receptor but that additional interactions with determinants in the LRP1 light chain contribute to the interaction.
Third, our data reveal that LRP1 interacts with the ␣ M -I domain and that this domain represents a major recognition interface. This was confirmed by binding studies using purified components, which revealed a high affinity interaction between LRP1 and the ␣ M -I domain coated on microtiter wells. Furthermore, we were able to demonstrate binding in a gain-in-function experiment by generating a mutant ␣ L ␤ 2 in which the ␣ M -I domain replaced the ␣ L -I domain. By employing homologscanning mutagenesis (17) in which sequences within the ␣ M -I domain were switched with those from ␣ L -I domain, we identified a region corresponding to residues Glu-162 to Leu-170 that contribute to the LRP1 interaction. This region of ␣ M -I domain is of interest, as Zhang and Plow (39) demonstrated that this loop region may represent a regulatory region that controls integrin activation and function. Indeed, these investigators demonstrated previously that perturbation of this specific sequence results in an enhanced adhesive activity of ␣ M ␤ 2 (39). Moreover, they showed that two function-blocking mAbs 44a and 2LPM19c inhibit the adhesive function of ␣ M ␤ 2 allosterically by binding to this same sequence (40). Interestingly, our studies confirm that sLRP1 inhibited ␣ M ␤ 2 -mediated adhesion to fibrinogen. This effect occurred at relatively low concentrations of sLRP1 (20 nM). sLRP1 can be detected in the circulation at levels around 10 nM (41), and it has been documented that the levels of sLRP1 rise dramatically during inflammation (42). Thus, sLRP1 may represent a physiological mechanism to modulate macrophage function.
In summary, using molecular biology approaches, we have localized the functional regions within LRP1 and ␣ M ␤ 2 for their reciprocal recognition. We found that LRP1 binding to ␣ M ␤ 2 is mediated via multiple domains within both receptors. Most importantly, our results suggest the possibility that LRP1 regulates macrophage inflammatory activity by modulating the functions of ␣ M ␤ 2 . Given the importance of LRP1-␣ M ␤ 2 recognition in macrophage efflux and thus the resolution of acute inflammation, the information provided from this work may help us to better understand the role of LRP1 in inflammation. In this regard, conditional macrophage LRP1 knock-out mice exhibit proinflammatory phenotypes in animal models of atherosclerosis (43,44). Thus, this work may help us understand the role of LRP1 and ␣ M ␤ 2 in macrophage-mediated inflammatory response, its proper resolution, and the initiation of wound healing.