Collagenase-3 Binds to a Specific Receptor and Requires the Low Density Lipoprotein Receptor-related Protein for Internalization*

We have previously identified a specific receptor for collagenase-3 that mediates the binding, internalization, and degradation of this ligand in UMR 106-01 rat osteoblastic osteosarcoma cells. In the present study, we show that collagenase-3 binding is calcium-dependent and occurs in a variety of cell types, including osteoblastic and fibroblastic cells. We also present evidence supporting a two-step mechanism of collagenase-3 binding and internalization involving both a specific collagenase-3 receptor and the low density lipoprotein receptor-related protein. Ligand blot analysis shows that 125I-collagenase-3 binds specifically to two proteins (∼170 kDa and ∼600 kDa) present in UMR 106-01 cells. Western blotting identified the 600-kDa protein as the low density lipoprotein receptor-related protein. Our data suggest that the 170-kDa protein is a specific collagenase-3 receptor. Low density lipoprotein receptor-related protein-null mouse embryo fibroblasts bind but fail to internalize collagenase-3, whereas UMR 106-01 and wild-type mouse embryo fibroblasts bind and internalize collagenase-3. Internalization, but not binding, is inhibited by the 39-kDa receptor-associated protein. We conclude that the internalization of collagenase-3 requires the participation of the low density lipoprotein receptor-related protein and propose a model in which the cell surface interaction of this ligand requires a sequential contribution from two receptors, with the collagenase-3 receptor acting as a high affinity primary binding site and the low density lipoprotein receptor-related protein mediating internalization.

Collagenase-3 (MMP-13) 1 is a member of the matrix metalloproteinase family of enzymes, which participates in extracellular matrix remodeling (1). Members of this family have a number of structural and functional features in common. In addition to sharing a similar domain structure, all are synthesized in inactive form, function at neutral pH, and require intrinsic zinc and calcium ions for their activity. Collagenase-3 is a highly regulated enzyme that cleaves native fibrillar collagens of types I, II, and III. The 57-kDa proenzyme is converted to its active 52-kDa form by the plasmin activation cascade, as well as by cathepsin B, stromelysin, and plasma kallikrein (2,3). Collagenase-3 activity is inhibited by a family of tissue inhibitors of metalloproteinases (4,5). A range of hormones and agents can also regulate expression of collagenase-3 (6). Parathyroid hormone is one of the hormones participating in this process. Osteoblastic cells respond to parathyroid hormone by increasing collagenase-3 synthesis (6 -8), plasminogen activator activity (9), and tissue inhibitors of metalloproteinases expression (8). In addition, experiments with UMR 106-01 rat osteosarcoma cells showed that over 80% of exogenous rat collagenase-3 was removed from the medium after 8 h of incubation (10). This rapid removal of rat collagenase-3 from the medium suggested the existence of a specific receptor that represents another level of regulation of this enzyme. Previous work (10) established the existence of this receptor and showed that it has a high affinity for rat collagenase-3 (K d ϭ 5 ϫ 10 Ϫ9 M), with approximately 12,000 receptors per UMR 106-01 cell. In this report, we show that the receptor is specific for collagenase-3 among the matrix metalloproteinases, demonstrate that binding activity is present on other cells, and describe the two-step process of binding and internalization that requires both a specific 170-kDa collagenase-3 receptor and LDL-receptor-related protein (LRP).

EXPERIMENTAL PROCEDURES
Materials-Cell culture media, fetal bovine serum (FBS), and other cell culture reagents were purchased from the Washington University Tissue Culture Support Center, St. Louis, MO. The following chemicals were purchased from Sigma: ascorbic acid, bovine serum albumin, chloramine T, proteinase E (Pronase), sodium iodide, sodium metabisulfite, Tween 20 and Tween 80, isopropylthio-␤-D-galactoside, glutathione and glutathione-agarose, thrombin inhibitor, CHAPS, insulin, transferrin. Na 125 I and ECL immunoblotting detection kit were purchased from Amersham Pharmacia Biotech. Bovine serum thrombin was purchased from Roche Molecular Biochemicals. SDS-polyacrylamide gel electrophoresis materials and nonfat dry milk were purchased from Bio-Rad and Amresco.
Proteins, Antibodies, and Plasmids-Purified rat collagenase-3 was isolated from media of cultures of post-partum rat uterine smooth muscle cells as described previously (11). The pGEX-receptor-associated protein (RAP) expression construct was a kind gift from Dr. Joachim Herz (University of Texas Southwestern Medical Center, Dallas, TX). The rabbit polyclonal antibody raised against the LRP receptor was a gift from Dr. Dudley Strickland (American Red Cross, Rockville, MD). Human RAP from the pGEX-RAP expression vector was expressed in bacteria and prepared as described previously (12). Human collagenase-3 was produced in a vaccinia virus-based expression system as described (13). 92 kDa and 72 kDa gelatinases were kind gifts from Dr. Howard Welgus (Washington University, St. Louis, MO). Human stromelysin was a generous gift from Dr. Paul Cannon (Syntex, Palo Alto, CA).
Radioiodination of Proteins-Protein labeling with 125 I was done using the chloramine T method (15). The proteins had specific activities ranging from 9 to 27 Ci/g.
Binding Assays-For all binding experiments, cells were seeded into 2.0 cm 2 wells. After the cells reached approximately 95% confluence, the medium was replaced with fresh medium containing 1 mg/ml bovine serum albumin, and the cells were assayed for binding 4 h later. The cells were first washed with maintenance medium, then incubated in the same medium with 0.01% Tween 80 containing 125 I-labeled rat collagenase-3 or other iodinated ligands at 4°C for 2 h. Nonspecific binding was assessed by adding a 50 -100-fold excess of cold ligand to half the wells, while an equivalent volume of buffer was added to the remaining wells. After incubation, the wells were washed three times with ice-cold MEM (0.5 ml). The cells were then lysed with 500 l of 1 M NaOH, and the lysates were counted on a ␥ counter.
Internalization Experiments-After binding 125 I-labeled proteins as above, the cells were washed three times with cold MEM (0.5 ml) to remove unbound ligand. The cells were then warmed to 37°C by the addition of prewarmed MEM (0.25 ml) and incubated at 37°C for selected intervals. At each time point, the media were collected, and the cells were washed once with ice-cold MEM, then incubated with 0.25% Pronase-E in MEM for 15 min at 4°C to strip cell surface proteins. The cell suspension was then centrifuged, and the radioactivity associated with cell pellets (defining internalized 125 I-labeled proteins) was measured at each time point.
Ligand Blotting-UMR 106-01, MEF-1, and MEF-2 cell membranes were prepared by differential centrifugation of homogenized cells at 1,000 ϫ g for 10 min, 10,000 ϫ g for 10 min, 100,000 ϫ g for 40 min in buffer containing 20 mM Tris-HCl, pH 7.5, 2 mM MgCl 2 , 0.25 M sucrose, 1 mM phenylmethylsulfonyl fluoride. The 100,000 ϫ g membrane pellet was then resuspended in buffer containing 50 mM Tris-HCl, pH 8.0, 2 mM CaCl 2 , 80 mM NaCl (16). The samples of cell membranes were subjected to 4 -15% SDS-polyacrylamide gel electrophoresis under nonreducing conditions at 50 V for 3 h and then electrotransferred to polyvinylidene difluoride filters in transfer buffer containing 10% methanol, 192 mM glycine, 56 mM Tris at 15 V for 16 h at 4°C. The filters were blocked with 5% nonfat dried milk in buffer containing 50 mM Tris-HCl, pH 8.0, 80 mM NaCl, 2 mM CaCl 2 , and 0.1% Triton X-100 (binding buffer) for 1 h at room temperature. The filters were then incubated for 16 h at 4°C in the same buffer supplemented with 1% nonfat dried milk in the presence of 20 pmol of 125 I-labeled rat collagenase-3 or 20 pmol of 125 I-GST-RAP in the presence or absence of the same unlabeled ligands (30 -40-fold excess of rat collagenase-3, 170-fold excess of GST-RAP). The filters were then washed with the same buffer, dried, and subjected to autoradiography.
Western Blot Analysis-The filters used for ligand blot analysis were wetted with methanol for 2 s, rinsed with H 2 O, and equilibrated with buffer containing 20 mM Tris-HCl, pH 7.6, 137 mM NaCl, 0.1% Tween 20. The filters were incubated 2 h at room temperature in the same buffer containing 5% nonfat dried milk. Subsequently, the filters were incubated with anti-LRP antibodies (1:2,000) in the same buffer containing 1% nonfat dried milk for 16 h at 4°C. A 1:10,000 dilution of horseradish peroxidase-conjugated goat anti-rabbit IgG in the same buffer containing 1% nonfat dried milk was incubated with the filters for 1 h at room temperature to detect the primary antibodies. Detection was performed using an ECL kit.

RNA Isolation and Northern Blot
Analysis-Poly(A ϩ ) RNA was isolated from 2 ϫ 10 8 each of UMR 106-01, MEF-1, and MEF-2 cells using the mRNA purification kit from Invitrogen. Five g of mRNA from each of UMR 106-01, MEF-1, and MEF-2 cells was separated by electrophoresis in 0.5% agarose formaldehyde (2.2 M) gel. The RNA was UVcross-linked to a Zeta-Probe GT membrane (Bio-Rad) after upward capillary transfer. The 5.99-kilobase fragment of LRP in pGEM-4 vector (ATCC 65430) was used as a probe for identification of LRP mRNA. The plasmid with the insert was labeled using the nick translation kit from Promega. ␤-Actin cDNA was labeled by random priming using Promega Prime-a-Gene kit. Prehybridization and hybridization of both LRP and ␤-actin probes was carried out at 42°C in 50% formamide, 5 ϫ SSC (1 ϫ SSC ϭ 0.15 M NaCl and 0.015 M sodium citrate, pH 7.0), 0.2% each of bovine serum albumin, Ficoll and polyvinylpyrrolidene, salmon sperm DNA (250 g/ml), 0.1% SDS and sodium pyrophosphate, pH 6.5 (50 mM), with 10 6 cpm/ml of each probe for 16 h. The filter was washed in 2 ϫ SSC, 0.1% SDS for 4 ϫ 5 min at room temperature, followed by 0.1 ϫ SSC, 0.1% SDS for 15 min at 50°C.

RESULTS
Previous work has shown that the rat collagenase-3 receptor is present on UMR 106-01 rat osteosarcoma cells (10). To characterize this receptor, we first tested a variety of other cells for rat collagenase-3 binding to ensure that this was not just restricted to a transformed rat osteosarcoma line. Specific binding was found in normal rat osteoblasts, rat embryo fibroblasts, and rat osteosarcoma cells, ROS 17/2.8 ( Fig. 1). The binding of rat collagenase-3 to normal rat osteoblasts and normal rat embryo fibroblasts was higher than binding to the UMR 106-01 cells. We observed very low levels of binding in rat epithelial breast carcinoma cells, BC-1, mouse NIH 3T3 fibroblasts, and human osteosarcoma cells, SAOS-2.
It has been demonstrated that osteoblastic cells in vitro can secrete a number of matrix metalloproteinases including collagenase-3 (8,(17)(18)(19), 72-kDa and 92-kDa gelatinase (19 -22), and stromelysin-1 (19).These proteinases are thought to play an active role in extracellular matrix remodeling in bone tissue. As we have shown in our previous competition experiments (10), none of the various proteins we used was able to compete with rat collagenase. However, of the matrix metalloproteinases (MMPs), only human collagenase-1 (MMP-1) had been FIG. 1. 125 I-Labeled rat collagenase-3 binding to different cell types. Seven cell types were plated in 24-well plates. All of them except normal rat osteoblasts were allowed to reach confluence and then incubated at 4°C for 2 h in binding medium with 3 nM 125 I-labeled rat collagenase-3. Normal rat osteoblasts were grown to confluence, and the medium was subsequently replaced to allow differentiation and mineralization. After 21 days of culture, a binding assay was conducted. The binding of 125 I-labeled rat collagenase-3 to different cell types is shown as means Ϯ S.E. for triplicate wells and expressed as a percentage of UMR 106-01 binding. tested. Since the publication of our initial observations on the rat collagenase receptor (10), human collagenase-3 (MMP-13) had been cloned and shown to be homologous to the rat and mouse collagenases (13). In addition, supplies of the other MMPs had become available. To show the specificity of the rat collagenase receptor in UMR 106-01 rat osteosarcoma cells, we investigated the ability of these cells to bind other MMPs. Ligand binding assays were performed using rat collagenase-3 (rat MMP-13), human fibroblast collagenase-1 (MMP-1), human stromelysin-1 (MMP-3), human collagenase-3 (human MMP-13), human 92-kDa gelatinase (MMP-9), and human 72-kDa gelatinase (MMP-2). As shown in Table I, only human collagenase-3 was comparable to rat collagenase-3 in binding to UMR cells. This was expected since human collagenase-3 has 86% homology to rat collagenase-3 (13). Human collagenase-3 also competes effectively with 125 I-labeled rat collagenase-3 for binding to the collagenase receptor (Fig. 2). This result argues for the existence of a specific receptor for collagenase-3 on osteoblastic cells, in contrast to collagenase-1, which has never been observed to be produced by these cells nor to bind or compete for binding. We next conducted a binding assay on UMR 106-01 cells using 125 I-labeled rat collagenase in the presence and absence of Ca 2ϩ to investigate the requirements of ligand-receptor interaction for this ion (Table II). The results showed that Ca 2ϩ is necessary for rat collagenase-3 binding to its receptor.
To determine the molecular weight of the rat collagenase-3 receptor, we next performed a ligand blot assay using partially purified UMR 106-01 cell membranes. It was found that 125 Ilabeled rat collagenase-3 bound to two proteins with molecular masses of about 600 kDa (q) and 170 kDa (Fig. 3, panel 1, *). 125 I-Collagenase binding was highly specific, since a 40-fold excess of unlabeled rat collagenase abolished binding to both proteins (Fig. 3, panel 2).
As described previously (23), rat collagenase-3 undergoes a process of binding, internalization, and degradation following secretion from UMR 106-01 cells. We hypothesized that the mechanism might be similar to the internalization of the members of the low density lipoprotein (LDL) receptor superfamily (24,25). Therefore, we proposed that one of the proteins that showed collagenase-3 binding on ligand blot analysis might be a member of the LDL receptor superfamily. Among members of this superfamily, only two have molecular masses around 600 kDa: LRP and gp330/megalin. None of the LDL superfamily receptors has a molecular mass of about 170 kDa (26 -32). We identified the 600-kDa protein as the large subunit of the LRP by Western blotting using anti-LRP antibodies (Fig. 3, panel 3,  q). Anti-LRP antibodies also detected the small subunit of the LRP (Fig. 3, panel 3, E).
To exclude the possibility that the collagenase-3 receptor is the LRP, we used two cell lines of mouse embryo fibroblasts: wild-type (MEF-1) and LRP-null (MEF-2). Northern blot analysis showed that both UMR 106-01 and MEF-1 cells express LRP, whereas MEF-2 cells do not (Fig. 4). Ligand blot and Western blot analyses showed that 125 I-labeled rat collagenase-3 specifically binds to the large subunit of the LRP in UMR 106-01 and MEF-1 but not MEF-2 cell membranes (Fig. 5,  panels 1, 2, and 5, q). Also, 125 I-RAP binds to only the large subunit of the LRP in UMR 106-01 and MEF-1 cell membranes (Fig. 5, panels 3-5, q). Furthermore, all three of these cell lines show binding of 125 I-collagenase-3 to the 170-kDa protein (Fig.  5, panel 1, *).We have also noticed that both MEF-1 and MEF-2 cells have an additional protein with molecular mass of approximately 200 kDa, which specifically binds 125 I-labeled rat collagenase-3 (Fig. 5, panel 1, OE).
125 I-Collagenase-3 binding assays were performed with MEF-1, MEF-2, and UMR 106-01 cells. The results showed no significant difference in binding between wild-type and LRPdeficient cells, suggesting that the LRP is not required for collagenase-3 binding to these cells (Fig. 6). We have also

FIG. 2. Inhibition of rat collagenase-3 binding to UMR cells by human collagenase-3.
Confluent UMR 106-01 cells were incubated with 3 nM 125 I-labeled rat collagenase-3 at 4°C for 2 h. Increasing concentrations of rat collagenase-3 (q) or human collagenase-3 (E) were added concurrently. Ligand alone is shown as 100%, and the samples with unlabeled rat and human collagenases are shown as a percentage of this binding.

Binding of collagenase-3 to its receptor requires Ca 2ϩ
Binding to UMR cells was conducted as described under "Experimental Procedures." Half of the wells were incubated with 3 nM 125 I-rat collagenase-3 in phosphate-buffered saline containing 0.9 mM CaCl 2 and 0.49 mM MgCl 2 , pH 7.4. The other half of the wells were incubated with 3 nM 125 I-rat collagenase in Ca 2ϩ -and Mg 2ϩ -free PBS, pH 7.4. Nonspecific binding was determined by the addition of a 100-fold excess of nonlabeled rat collagenase to other wells. shown that RAP does not inhibit 125 I-labeled rat collagenase binding to the UMR cells, although it is known to inhibit binding of most ligands for the LRP (Fig. 7). These data suggest that the 170-kDa protein is a specific receptor for collagenase-3 in UMR 106-01 cells. Although the LRP is not required for rat collagenase-3 binding to the cell, it might be required for ligand internalization. Therefore, we performed internalization assays with 125 I-labeled rat collagenase-3 using MEF-1 and MEF-2 cells. The results showed that despite equal binding, MEF-2 cells cannot internalize rat collagenase-3. This suggests that the LRP is required for collagenase-3 internalization (Fig. 8). It is known that RAP inhibits internalization of ligands by the LRP (12,(33)(34)(35). Therefore, we next performed internalization assays using 125 I-labeled rat collagenase-3 as a ligand and RAP as a competitor. Our results showed that internalization of 125 Ilabeled rat collagenase-3 was inhibited by RAP by approximately 70% in UMR 106-01 cells (Fig. 9). Kinetic studies also suggested that RAP acts as a competitive inhibitor for rat collagenase-3 binding to LRP for internalization (Fig. 9, inset).
We next compared the ability of RAP to inhibit internaliza-tion of 125 I-labeled rat collagenase-3 in UMR 106-01 osteoblastic cells and normal rat osteoblasts. The presence of 100 mM RAP in binding medium reduced the intracellular accumulation of 125 I-collagenase by 79% in UMR 106-01 cells and by 43% in normal mineralizing rat osteoblasts (Table III). The difference in inhibition might be explained by the presence of mineralized extracellular matrix in normal rat osteoblast cultures as well as the possibility that an additional mechanism may also operate for internalization of rat collagenase-3 in these cells. However, inhibition of collagenase-3 internalization by  3 and 4). After washing the membranes, bound radioactivity was visualized by autoradiography. Panel 5 shows Western blot analysis using 1:2000 dilution of anti-LRP antibodies. Horseradish peroxidase-conjugated goat antirabbit IgG (1:10,000 dilution) was used as a secondary antibody. Final detection was performed using ECL kit (see "Experimental Procedures").
FIG. 6. Binding of collagenase-3 to mouse embryo fibroblasts with and without LRP. Binding was conducted as described under "Experimental Procedures." MEF-1 cells are embryonic fibroblasts from wild-type mice, whereas MEF-2 are embryonic fibroblasts from LRPnull animals. All three cell lines were incubated with increasing concentrations of 125 I-labeled rat collagenase-3 (or 3 nM 125 I-labeled rat collagenase-3, inset) in binding media for 2 h at 4°C. Nonspecific binding was determined by the addition of a 100-fold excess of unlabeled rat collagenase-3 to one-third of the wells. The values displayed represent specific binding ϮS.E. for triplicate wells. RAP in both transformed osteoblastic cells and normal osteoblasts suggests that the same type of receptor operates in both cell types.
To investigate the mechanism by which RAP regulates internalization of collagenase-3, we performed an experiment where excess unlabeled RAP or rat collagenase-3 was prebound to UMR 106-01 cells. Binding and internalization of 125 I-labeled rat collagenase-3 and RAP were then allowed to proceed. The data showed that although prebound RAP inhibited rat collagenase-3 internalization, prebound rat collagenase-3 had almost no effect on RAP internalization (Fig. 10, A and B). DISCUSSION In this paper we describe collagenase-3 interaction with the cell and show that it involves two receptors; the specific collagenase-3 receptor acts as the primary binding site, whereas the LRP is required for internalization. The LRP belongs to the LDL receptor superfamily (36). This superfamily consists of endocytotic receptors that mostly participate in the recognition and endocytosis of lipoproteins (25). The receptors have high affinity for their ligands and broad specificity. They recognize not only lipoproteins but also a variety of nonlipoprotein ligands, including urokinase and tissue plasminogen activator with their inhibitors and participate in different physiological processes (26,(37)(38)(39)(40)(41)(42)(43)(44)(45). Ten members of this family are known to date: the LDL receptor itself, ␣ 2 -macroglobulin receptor/low density lipoprotein receptor-related protein (␣ 2 MR/LRP), very low density lipoprotein receptor, Heymann nephritis antigen/ megalin/gp330, chicken vitellogenin receptor, Drosophila Yolkless, chicken LR8B, placental calcium sensor protein (29), the newly discovered apolipoprotein E receptor 2 (apoER2) (30), and LR11 (31). These receptors share a similar structure, with a single transmembrane domain and numerous ligand binding domains organized as cysteine-rich repeats arranged in clusters, followed by two epidermal growth factor-like repeats separated from a third one by a spacer region containing a YWTD consensus sequence and an NPXY internalization signal in the cytoplasmic domain.
Binding assays showed that the collagenase-3 receptor is present mostly in osteoblasts and fibroblasts. Interestingly, cell surface binding of collagenase-3 does not necessarily correlate with expression of collagenase-3 by these cells. For example, ROS 17/2.8 cells do not express collagenase-3, but the binding of the enzyme to ROS 17/2.8 cells was comparable to that of Pronase-E to remove cell surface-bound 125 I-labeled rat collagenase-3. Cells were collected, and radioactivity was measured.

TABLE III
Inhibition of 125 I-rat collagenase-3 internalization by RAP in osteosarcoma cells and normal osteoblasts UMR 106 -01 cells and osteoblasts isolated from rat calvariae were plated in 24-well plates. After rat osteoblasts reached confluence at 7 days, the media were replaced with differentiation medium and cells incubated for 14 days at 37°C to allow differentiation. For the experiment, cells were incubated for 2 h at 4°C in the presence of 3 nM 125 I-rat collagenase-3 with RAP (100 nM) or without RAP. After washing, the cells were warmed to 37°C. After 15 min, the overlying media were removed, and the cells were treated with Pronase-E to separate cell surface 125 I-rat collagenase-3 from internalized ligand. The cells were collected, and radioactivity was counted to determine the amount of internalized 125 I-rat collagenase-3. The values displayed represent means Ϯ S.E. for triplicate wells. Normal rat osteoblasts UMR 106-01 cells. At the same time, the binding to BC-1 cells, which secrete collagenase-3 at a high constitutive level, was very low. 2 Based on these data, we conclude that this receptor might bind enzyme secreted by neighboring cells or play other roles in addition to regulation of the extracellular abundance of collagenase-3.
We have assayed UMR 106-01 cells for their ability to bind different metalloproteinases. Although the members of the metalloproteinase family share a number of general functional and structural features, our data demonstrate high specificity of the collagenase receptor for rat collagenase-3 and human collagenase-3 but almost no binding of other MMPs. Similarly, we have shown that mouse collagenase-3 binds equally well as the rat enzyme (data not shown). Nevertheless, we cannot rule out the possibility that the receptor may have ligands other than collagenase-3.
Ligand and Western blot analyses showed that rat collagenase-3 can specifically bind to the large subunit of the LRP receptor and a protein with a molecular mass of approximately 170 kDa, which is present in membranes of UMR 106-01, MEF-1, and MEF-2 cells. Equal levels of rat collagenase-3 binding to UMR 106-01, wild-type (MEF-1), and LRP-null (MEF-2) cells suggested that the collagenase-3 receptor is present in all of these cell lines and that the LRP receptor does not participate in primary binding of collagenase-3 to the cell surface. Although MEF-1 and MEF-2 cells bind rat collagenase-3 equivalently, our experiments showed that MEF-2 cells cannot internalize the bound ligand. We have also observed that rat collagenase-3 internalization by UMR 106-01 cells was abolished in the presence of RAP. Therefore, we conclude that collagenase-3 interaction with the cell is a two-step process. First, a specific collagenase receptor of 170 kDa acts as a primary binding site for collagenase-3 on the cell surface. Interaction between the LRP and the enzyme-receptor complex then occurs, resulting in internalization of collagenase-3. A similar process has been reported for urokinase-type plasminogen activator-plasminogen activator inhibitor type 1, tissue plasminogen activator-plasminogen activator inhibitor type 1, and urokinase-type plasminogen activator-recombinant protease nexin-1 complexes (46 -48). In each case, the serine protease binds to a specific receptor as a primary event. The inhibitor then binds to the receptor-ligand complex, which leads to its rapid internalization and degradation by the LRP. In our studies, this latter process is similarly inhibited by RAP, which implicates the LRP. The exact mode by which transfer and internalization of the collagenase-3 ligand-receptor complex takes place is not completely understood. It is possible that the ligand dissociates from the specific receptor to bind to the LRP. Alternatively, the LRP might bind and internalize the entire receptor-ligand complex. Similarly, it is not known if collagenase-3 is internalized alone or in complex with its inhibitor and/or specific receptor. Recent data suggest that LRP-mediated internalization of the plasminogen activators involves the endocytosis of the primary binding receptor, which may then be recycled to the cell surface (49,50).
Ligand blot studies showed that mouse embryo fibroblasts have an additional protein with a molecular mass of approximately 200 kDa, which also specifically bound 125 I-labeled rat collagenase-3. We thus concluded that in these cells, three membrane proteins might be involved in collagenase-3 clearance, indicating that our proposed mechanism might vary somewhat in different cell types.
The results of inhibition studies showed that RAP abolished rat collagenase-3 internalization in UMR 106-01 cells, whereas collagenase-3 does not change the level of RAP internalization. Thus, collagenase-3 does not compete for binding to RAP sites on the LRP. In addition RAP may be a physiological modulator of collagenase-3 internalization by the LRP. It has been shown that RAP is coexpressed with either LRP or gp330 (51). However, it is still unknown whether RAP is expressed in osteoblastic cells. Further experiments may show the presence of RAP in bone tissue.
In conclusion, the two-step mechanism of collagenase-3 interaction with the cell may serve as one more link in the chain of fine regulation of collagenase-3 activity contributing to homeostasis of the extracellular matrix.

FIG. 10. Comparison of collagenase-3 and RAP internalization.
A, analysis of collagenase-3 internalization. UMR 106-01 cells were plated in 24-well plates. Cells were incubated in the presence (E) or absence (q) of 100 nM RAP in binding medium for 2 h at 4°C. Cells were then washed three times with binding medium, 2 nM 125 I-labeled rat collagenase-3 was added to all wells, and incubation was continued an additional 2 h. Internalization was then conducted as described under "Experimental Procedues." B, analysis of RAP internalization. The experiment was conducted as above, except the cells were incubated in the presence (E) and absence (q) of 100 nM collagenase-3 for 2 h at 4°C. After the cells were washed, 2 nM 125 I-RAP was added to all wells.