The Two Mannose 6-Phosphate Binding Sites of the Insulin-like Growth Factor-II/Mannose 6-Phosphate Receptor Display Different Ligand Binding Properties*

The two mannose 6-phosphate (Man-6-P) binding sites of the insulin-like growth factor-II/mannose 6-phosphate receptor (IGF-II/MPR) have been localized to domains 1–3 and 7–9, and studies have shown that Arg435 in domain 3 and Arg1334 in domain 9 are essential for Man-6-P binding. To determine whether the IGF-II/MPR containing a single Man-6-P binding site is functional, clonal mouse L cell lines stably transfected with either mutant bovine IGF-II/MPR cDNA, containing substitutions at position 435 and/or 1334, or the wild type receptor cDNA were assayed for their ability to sort lysosomal enzymes to the lysosome. Mutant receptors containing a single Man-6-P binding site were ∼50% less efficient than the wild type receptor in the overall targeting of lysosomal enzymes to the lysosome. Mutant receptors containing a substitution at Arg1334 (Dom9Ala), in contrast to those containing a substitution at Arg435(Dom3Ala), were unable to target cathepsin D and β-hexosaminidase to the lysosome. Equilibrium binding assays using125I-labeled β-glucuronidase demonstrated that Dom3Ala and Dom9Ala had aK d of 2.0 and 4.3 nm, respectively. In addition, Dom3Ala, unlike Dom9Ala, was unable to completely dissociate from ligand under acidic pH conditions. These data indicate that the two Man-6-P binding sites of the IGF-II/MPR are not functionally equivalent.

somal enzymes are delivered to the lysosome, while the receptors recycle back to the Golgi or move to the plasma membrane, where they function to internalize exogenous ligands.
Two distinct MPRs have been identified and extensively characterized. The bovine insulin-like growth factor-II/MPR (IGF-II/MPR) is a 270-kDa membrane glycoprotein that can bind two distinct classes of ligands: Man-6-P-containing proteins and IGF-II, a nonphosphorylated polypeptide hormone (5,6). The bovine IGF-II/MPR consists of a 2269-residue extracytoplasmic region, a single transmembrane region, and a 163residue carboxyl-terminal cytoplasmic domain (7). The extracytoplasmic ligand-binding region has a repetitive structure due to the presence of 15 contiguous domains, each ϳ150 amino acids in length. The second receptor, the 46-kDa bovine cationdependent MPR (CD-MPR), exhibits optimal ligand binding in the presence of divalent cations (8,9), shares sequence homologies with each of the 15 repeating domains of the IGF-II/MPR (10), and is expressed along with the IGF-II/MPR in most cell types (11). Characterization of the ligand binding properties of the MPRs has revealed that the IGF-II/MPR binds 2 mol, while the CD-MPR binds 1 mol, of Man-6-P per polypeptide (12,13). Since the CD-MPR exists predominantly as a dimer in membranes (14,15), the functional form of the CD-MPR, like that of the IGF-II/MPR, contains two Man-6-P binding sites. However, it is not known whether both Man-6-P binding sites are necessary for the targeting of lysosomal enzymes to the lysosome.
Previous studies using partial proteolysis and expression of truncated forms of the receptor have localized the two Man-6-P binding sites of the IGF-II/MPR to domains 1-3 and domains 7-9 (16,17). Site-directed mutagenesis studies have demonstrated an important role for the conserved arginine residue in domain 3 (Arg 435 ) and domain 9 (Arg 1334 ). In constructs expressing only a single Man-6-P binding site (domains 1-3 or domains [5][6][7][8][9] or both Man-6-P binding sites (domains 1-9), replacement of the conserved arginine residue(s) with alanine or lysine resulted in the complete loss of ligand binding as assessed by pentamannosyl phosphate-agarose affinity chromatography. However, if only one of the arginine residues was mutated in the construct encoding domains 1-9, thus leaving one functional Man-6-P binding site, the ability to bind to a pentamannosyl phosphate-agarose affinity column remained intact (17).
In the current study, we have generated full-length IGF-II/ MPRs containing one functional Man-6-P binding site by replacing a single arginine residue at position 435 or 1334 and have determined their ability to target lysosomal enzymes to the lysosome in vivo as well as their affinity for ligand in vitro.
Our results indicate that IGF-II/MPRs containing a single functional Man-6-P binding site are ϳ50% less efficient than the wild type (WT) IGF-II/MPR in the overall targeting of lysosomal enzymes. In addition, analyses of the sorting of individual lysosomal enzymes, the binding affinity to ␤-glucuron-idase, and the dissociation of ligand at acidic pH provide evidence that the two Man-6-P binding sites of the IGF-II/MPR are not functionally equivalent.

EXPERIMENTAL PROCEDURES
Materials-The following reagents were obtained commercially as indicated: Trans 35  Preparation of Wild Type and Mutant IGF-II/MPR cDNAs-An arginine to alanine substitution at position 435 and/or 1334 was generated as described previously (17) and cloned into the full-length receptor. The WT and mutant bovine IGF-II/MPR cDNAs were placed into the expression plasmid pSFFV-neo (18), which contains a neomycin resistance gene.
Transfection and Metabolic Labeling of L (Rec Ϫ ) Cells-An IGF-II/ MPR deficient mouse L cell line, designated L (Rec Ϫ ), was described previously (19). Cell monolayers were maintained in DMEM/F-12 containing 10% heat-inactivated FBS at 37°C in 5% humidified CO 2 . The L (Rec Ϫ ) cells were transfected with 20 g of XbaI-linearized plasmid using the calcium phosphate precipitation method (20). Cells were selected for resistance to G418 (0.35 mg/ml active concentration). For each construct, 25-35 G418-resistant colonies were isolated 10 -15 days after transfection and were analyzed for receptor expression and homogeneity by indirect immunofluorescence staining (21). Homogeneous clones were expanded for further study and maintained in DMEM/F-12, 10% heat-inactivated FBS, and 0.35 mg/ml G418 (21). For selected experiments, stably transfected cells at 80% confluency were incubated in methionine-free DMEM/F-12 containing 10% FBS plus 50 Ci/ml Trans 35 S-label for 18 h or incubated in methionine-free DMEM/F-12 containing 10% FBS plus 1 mCi/ml Trans 35 S-label (pulse medium) for 1 h followed by incubation with DMEM/F-12 containing 10% FBS plus 1 mM methionine and 1 mM cysteine (chase medium). Unless otherwise indicated, the cells were solubilized for 1 h on ice in column buffer (50 mM imidazole (pH 6.5), 150 mM NaCl, 5 mM ␤-glycerophosphate, and 0.05% (v/v) Triton X-100) containing 1% (v/v) Triton X-100, 0.1% (w/v) sodium deoxycholate, 1% (v/v) aprotinin, 4 g/ml antipain, 20 g/ml benzamidine, and 2 g/ml each of leupeptin, chymostatin, and pepstatin. The total amount of protein in the supernatant was determined using the Bradford protein assay as recommended by the manufacturer (Bio-Rad). For the determination of the WT and mutant IGF-II/MPR half-life, stably transfected cells were pulse-labeled for 1 h followed by a chase for 1, 2, 8, 24, or 48 h. The cell lysates were immunoprecipitated with an IGF-II/MPR-specific antiserum and protein A-Sepharose. The samples were washed and analyzed by SDS-PAGE, and the radioactivity in the receptor band was quantified using a PhosphorImager (Molecular Dynamics Storm 860) and ImageQuant (version 4.1) software.
Western Blot Analysis-The level of receptor expression for each construct was determined by quantitative Western blot analysis as described previously (22) except that following incubation with bovine IGF-II/MPR-specific antiserum, the membranes were incubated for 1 h with 20 mM Tris (pH 7.6), 137 mM NaCl, 3.8 mM HCl, and Tween 20 (0.2% v/v) containing bovine serum albumin (0.1%, w/v) and protein A/horseradish peroxidase conjugate (1:2000, Amersham Pharmacia Biotech). The receptor was detected by enhanced chemiluminescence and quantified using a densitometer, and the results were normalized to total protein. Purified bovine liver IGF-II/MPR (50 -840 pg) was used as the standard and was present on each gel as described previously (22).
Pentamannosyl Phosphate-Agarose Affinity Chromatography-Stably transfected cells were labeled with Trans 35 S-label overnight, and the cell lysates were subjected to pentamannosyl phosphate-agarose chromatography as described previously (17). The flow-through fractions, glucose 6-phosphate (Glu-6-P) eluate, and Man-6-P eluate were immunoprecipitated using rabbit polyclonal antisera specific for the bovine IGF-II/MPR (17) or precipitated with trichloroacetic acid and analyzed by SDS-PAGE. The radiolabeled bands were quantified with an Ambis optical imaging system. In some experiments, WT and mutant IGF-II/MPRs purified from stably transfected L (Rec Ϫ ) cells were subjected to pentamannosyl phosphate-agarose affinity chromatography as described above with the exception that prior to the Glu-6-P and Man-6-P elutions, the columns were eluted with MES buffer (50 mM MES, 150 mM NaCl, 5 mM ␤-glycerophosphate, and 0.05% Triton X-100, pH 4.6).
Iodination of Ligands-Human recombinant IGF-II was iodinated by the lactoperoxidase method as described by Tait et al. (23) to specific activities of 50 -100 Ci/g. Purified human ␤-glucuronidase (50 g) was iodinated with 1 mCi of Na 125 I using soluble lactoperoxidase as described by Jadot et al. (24) except that following gel filtration on Sephadex G-150, 125 I-labeled ␤-glucuronidase was further purified on an IGF-II/MPR affinity column. The purified ␤-glucuronidase (1-2 Ci/ g) was then dialyzed exhaustively to remove the Man-6-P.
Binding of 125 I-IGF-II to WT and Mutant IGF-II/MPRs-Bovine WT and mutant IGF-II/MPRs purified from stably transfected L (Rec Ϫ ) cells by D. discoideum lysosomal enzyme affinity chromatography (8) were incubated with increasing concentrations of iodinated human recombinant IGF-II for 2 h at 25°C as described by Dahms et al. (22). The receptor-ligand complex was precipitated by the addition of three volumes of ice-cold 16.6% polyethylene glycol 8000, 0.75 mg/ml ␥-globulin, and 0.1 M Tris, pH 7.4 (5). Reaction mixtures without IGF-II/MPR were used as a measure of nonspecific binding. In all cases, nonspecific binding was Յ11% of the added counts. The results were analyzed by nonlinear regression (Sigma Plot 4.0) and Scatchard analysis (25).
Internalization of 125 I-IGF-II by Transfected Cells-Transfected cell lines seeded in 12-well plates were incubated at 37°C in phosphatebuffered saline, 1% bovine serum albumin plus 10 mM Man-6-P for 10 min to dissociate any bound ligands from the IGF-II/MPR. The cells were incubated with 1-2 nM 125 I-IGF-II in DMEM/F-12, 2% bovine serum albumin for 1 h at 37°C. The medium was then harvested, and the radioactivity was precipitated with 10% trichloroacetic acid. The plates were placed on ice, the surface ligands were removed by two 10-min incubations with 1 ml of 0.2 M acetic acid (pH 3.5)/0.5 M NaCl, and the cells were solubilized in 0.1 N NaOH. The cells transfected with vector alone (Neo) were used as a control for nonreceptor binding. The internalization index was calculated by dividing the sum of the intracellular and degraded radioactivity by the surface-bound radioactivity (26).
Analysis of Lysosomal Enzyme Sorting-The following experiments were performed in the presence of Man-6-P (10 mM) to prevent cell surface IGF-II/MPRs from internalizing secreted lysosomal enzymes as described previously (27).
Secretion of Total Lysosomal Enzymes-Stably transfected cells were pulse-labeled for 1 h followed by a chase for 4 h. The cells were solubilized in 0.1 M NaOH, and the total amount of protein was determined using the Bio-Rad DC protein assay. A fraction of the dialyzed medium sample was passed over bovine liver IGF-II/MPR affinity columns (24). After washing, the columns were eluted sequentially with 5 mM Glu-6-P in column buffer followed by 5 mM Man-6-P in column buffer. The eluates were precipitated with trichloroacetic acid and subjected to SDS-PAGE, and the radioactivity was quantified using an Ambis radioanalytical imaging system.
Sorting of Cathepsin D-Stably transfected cells were pulse-labeled for 1 h followed by a chase for 4 h. Aliquots of cells and medium were immunoprecipitated with cathepsin D-specific antiserum plus protein A-Sepharose overnight at 4°C (28). Man-6-P was added to each sample to a final concentration of 20 mM in order to displace any cathepsin D bound to soluble IGF-II/MPRs present in the FBS (21). The samples were washed and analyzed by SDS-PAGE, and the amount of cathepsin D present in the cells and medium was quantified using a densitometer.
Binding Affinity-Since L (Rec Ϫ ) cells contain endogenous CD-MPR (19), it was necessary to separate the IGF-II/MPR from the endogenous CD-MPR prior to performing the binding studies. The IGF-II/MPR, unlike the CD-MPR (13), can bind phosphodiesters (31,32). WT and mutant IGF-II/MPRs were purified by passing cell lysates over an affinity column substituted with D. discoideum lysosomal enzymes, which contain Man-6-P residues diesterified to methyl groups (33). The columns were washed and eluted with 10 mM Man-6-P, and the Man-6-P eluates were dialyzed extensively to remove Man-6-P. The amount of IGF-II/MPR present was determined by quantitative Western blot analysis described above. To confirm that there was no contaminating CD-MPR, Man-6-P eluates (aliquot size was 10 -50 times more than that used for the binding studies) were subjected to Western blot analysis using CD-MPR-specific antisera. In all cases, the amount of CD-MPR in the Man-6-P eluate was below the level of detection (Ͻ0.1 ng).
Purified WT or mutant IGF-II/MPRs were incubated with increasing concentrations of 125 I-labeled ␤-glucuronidase. The MPR and bound ligand were immunoprecipitated with an anti-IGF-II/MPR polyclonal antibody that was prebound to protein A-Sepharose. After a total incubation of Ն4 h, the beads were pelleted and washed. The supernatant and washes were collected, and the radioactivity was counted as a direct measure of unbound ligand in each reaction. Bound ␤-glucuronidase was specifically eluted from the antibody-bound MPRs by incubation at 4°C for 1 h with 10 mM Man-6-P. The protein A-Sepharose beads were counted as nonspecific binding and represented Յ1% of the total radioactivity added. To validate this binding assay, experiments were performed that demonstrated that ␤-glucuronidase did not interfere with the immunoprecipitation of the IGF-II/MPR and that sufficient antibody and protein A-Sepharose were present to immunoprecipitate all of the IGF-II/MPR. Control experiments were also performed to show that equilibrium binding was reached within 2 h, that ␤-glucuronidase binding was dependent on the concentration of receptor present, and that 10 mM Man-6-P was sufficient to elute Ն95% of the ␤-glucuronidase bound to the IGF-II/MPR.

Expression of WT and Mutant IGF-II/MPRs in Stably Transfected Cell
Lines-To determine whether mutant IGF-II/MPRs containing a single Man-6-P binding site, rather than two are capable of targeting lysosomal enzymes to the lysosome, we took advantage of our previous studies, which identified Arg 435 in domain 3 and Arg 1334 in domain 9 as essential residues for high affinity Man-6-P binding (17). Site-directed mutagenesis was used to generate an arginine to alanine substitution at position 435 (Dom3 Ala ) or 1334 (Dom9 Ala ) to construct IGF-II/ MPRs with one functional Man-6-P binding site. Also, a double mutant construct was generated that contained alanine substitutions at both essential arginine residues (Dom39 Ala ) to create a receptor that lacks the ability to bind Man-6-P. Table  I summarizes the different constructs of the bovine IGF-II/ MPR that were expressed in mouse L (Rec Ϫ ) cells, a cell line that lacks endogenous IGF-II/MPRs (19), resulting in the secretion of the majority of its lysosomal enzymes into the medium.
To verify that the mutant receptors folded properly, their ability to bind both Man-6-P-containing proteins and IGF-II, to function in endocytosis, and to remain stable in the cells was assayed. Previous studies using a construct encoding domains 1-9 of the IGF-II/MPR showed that constructs with a single point mutation either at Arg 435 or Arg 1334 retained their ability to bind to a pentamannosyl phosphate affinity column, while the construct encoding point mutations at both Arg 435 and Arg 1334 could not (17). To determine if the full-length receptor, exhibited similar properties, lysates from stably transfected cells were subjected to pentamannosyl phosphate-agarose affinity chromatography (Fig. 1). The WT and mutant receptors with one functional Man-6-P binding site (Dom3 Ala or Dom9 Ala ) were retained on the affinity column and eluted specifically with Man-6-P, whereas Ͻ5% of the construct containing the double mutant (Dom39 Ala ) bound to the pentamannosyl phosphate column. These results confirm that the full-length IGF-II/MPR is similar to the truncated receptors in that replacement of Arg 435 and Arg 1334 dramatically diminishes the ability to bind Man-6-P, thus confirming that these two arginine residues are essential for high affinity Man-6-P binding. To evaluate the integrity of the IGF-II binding site, equilibrium binding assays using 125 I-IGF-II were performed on the purified receptors. Analysis of the binding curves (data not shown) by nonlinear regression demonstrated a K d of 0.3, 0.2, and 0.1 nM for the WT, Dom3 Ala , and Dom9 Ala constructs, respectively, which are all very similar to the K d of 0.2 nM reported for the bovine IGF-II/MPR (5). These results demonstrate that the IGF-II binding site, which has been localized to domain 11 (22,34), has not been altered by the single amino acid substitution in either domain 3 or domain 9. To confirm that the mutant receptors were expressed on the cell surface and participated in endocytosis of ligands, stably transfected cells were incubated with 125 I-IGF-II, and the cell surface and internalized radioactivity was determined. The internalization indexes were determined to be 21, 13, 16, and 18 for the WT, Dom3 Ala , Dom9 Ala , and Dom39 Ala constructs, respectively, demonstrating that all four constructs are expressed on the cell surface and internalize IGF-II at a similar rate. Finally, the half-life of the WT, Dom3 Ala , Dom9 Ala , and Dom39 Ala constructs was determined to be 25, 37, 22, and 32 h, respectively. Taken together, these results demonstrate that the single amino acid substitution in domain 3 or domain 9 does not dramatically affect the stability or structure of the IGF-II/MPR.
Secretion of Man-6-P-containing Polypeptides by Transfected  3 and/or domain 9 were substituted with alanine, which has been shown previously to inhibit Man-6-P binding (17). Stably transfected cell lines were generated for each construct, and the expression ranged between 40 and 600 pg of receptor/g of total protein.  35 S-labeled L (Rec Ϫ ) cells transfected with vector alone (Neo) or IGF-II/MPR constructs were passed over pentamannosyl phosphate-agarose columns. The columns were washed in column buffer, and a total of five 3-ml fractions were collected. The columns were eluted with 5 mM Glu-6-P (G) and then with 5 mM Man-6-P (M). The wash fractions and eluates were immunoprecipitated with IGF-II/ MPR-specific antisera, analyzed on 7.5% SDS-polyacrylamide gels under reducing conditions, and quantified using an Ambis radioanalytical imaging system. The numbers to the right are the mean of two independent experiments and represent the amount of radioactive IGF-II/ MPR specifically eluted with Man-6-P as a percentage of the total receptors in that sample.
Cells-To examine the ability of mutant IGF-II/MPRs with one Man-6-P binding site to target lysosomal enzymes to the lysosome, medium from clonal cell lines metabolically labeled with Trans 35 S-label for 1 h followed by a 4-h chase was subjected to affinity chromatography on columns substituted with IGF-II/ MPR (Fig. 2). A similar pattern of 10 distinct Man-6-P-containing proteins was observed to be secreted by each of the clonal cell lines. Mouse L (Rec Ϫ ) cells transfected with the double mutant, Dom39 Ala , which does not contain a functional Man-6-P binding site, secreted a similar level of total phosphorylated lysosomal enzymes when compared with cells transfected with vector alone (Neo). In contrast, clonal cell lines expressing the WT, Dom3 Ala , or Dom9 Ala IGF-II/MPRs secreted only 54, 77, or 83%, respectively, of the enzymes secreted by cells transfected with vector alone. These data indicate that IGF-II/MPRs with one Man-6-P binding site are less efficient than the WT IGF-II/MPR in the overall targeting of lysosomal enzymes to the lysosome. Analysis of the individual bands demonstrated that clonal cell lines expressing the WT, Dom3 Ala , or Dom9 Ala receptor secreted varying amounts of the different lysosomal enzymes. These data also indicate that, although the Dom3 Ala and Dom9 Ala constructs are similar in their ability to target lysosomal enzymes overall, the two receptors differ in their ability to target certain lysosomal enzymes to the lysosome (Fig. 2, arrows). To test this hypothesis, cells transfected with WT and mutant IGF-II/MPRs were characterized for their ability to target the individual lysosomal enzymes cathepsin D, ␤-hexosaminidase, and ␤-glucuronidase.
Sorting of Individual Lysosomal Enzymes by Transfected Cells-Delivery of cathepsin D to the lysosome can be monitored by the appearance of a lower molecular weight species that results from the proteolytic conversion of the precursor to the mature form (35). The results of a typical experiment in which cell lysates and media were analyzed from cells incubated with Trans 35 S-label for 1 h and chased for 4 h to allow the newly synthesized cathepsin D to be phosphorylated and To assess if mutant IGF-II/MPRs with one Man-6-P binding site could effectively sort ␤-hexosaminidase, intracellular and extracellular activities were measured after a 48-h culture period (Fig. 4). Mouse L (Rec Ϫ ) cells transfected with vector In contrast, clonal cell lines that express the Dom9 Ala construct were similar to Neo cells and secreted ϳ65% of their ␤-hexosaminidase. These results suggest that receptors with a mutation in domain 9, in contrast to receptors with a mutation in domain 3, cannot efficiently target this lysosomal enzyme to the lysosome.
The intracellular and extracellular activities of ␤-glucuronidase were also determined for the stably transfected mouse L (Rec Ϫ ) cells (Fig. 5). Cells transfected with vector alone or with the double mutant (Dom39 Ala ), secreted ϳ45-50% of their lysosomal enzyme, ␤-glucuronidase. Clonal cell lines expressing increasing levels of the WT IGF-II/MPR secreted decreasing amounts of ␤-glucuronidase in the media, with 12-20% secreted in clonal cell lines expressing over 340 pg of receptor/g of total protein. Analysis of clonal cell lines that express the Dom3 Ala construct showed a decrease in the secretion of ␤-glucuronidase to 38 or 22%, depending on the level of receptor expression. A clonal cell line expressing 320 pg of Dom9 Ala receptor/g of total protein decreased the amount of ␤-glucuronidase secreted into the medium to ϳ33%. These data suggest that mutant IGF-II/MPRs with one functional Man-6-P binding site can target ␤-glucuronidase to the lysosome, but when expressed at levels less than 400 pg receptor/g total protein, they are not as efficient as the WT IGF-II/MPR.
Binding Affinity of WT and Mutant IGF-II/MPRs from Transfected Cells-The analysis of total and individual lysosomal enzyme secretion suggests that IGF-II/MPRs with one Man-6-P binding site are not as efficient in sorting lysosomal enzymes as the WT receptor. The above data also suggest that the extent of sorting depends on the enzyme analyzed and the location of the Man-6-P binding site. In order to test if the decrease in sorting efficiency may be due to a decrease in the affinity for ligand, WT or mutant IGF-II/MPRs were passed over pentamannosyl phosphate-agarose columns and eluted stepwise with increasing concentrations of Man-6-P (Fig. 6). WT IGF-II/MPR remained associated with a pentamannosyl phosphate-agarose column with little or no receptor being M Man-6-P, and the majority eluted with less than 800 M Man-6-P. These data suggest that receptors with one functional Man-6-P binding site have a lower affinity for pentamannosyl phosphate than the WT IGF-II/MPR. The elution profiles also show that the majority of the Dom9 Ala receptor elutes at a lower concentration of Man-6-P than the Dom3 Ala receptor, suggesting that the two Man-6-P binding sites have different affinities for ligand.
To specifically address possible differences in the binding affinity of the individual Man-6-P binding sites, the affinity of the WT, Dom3 Ala , and Dom9 Ala IGF-II/MPRs for an endogenous ligand was directly determined. The ligand chosen was the lysosomal enzyme, ␤-glucuronidase, which has been well characterized and has a reported K d of 2 nM (36) for WT human IGF-II/MPR. Purified WT and mutant IGF-II/MPRs were incubated with increasing concentrations of iodinated human recombinant ␤-glucuronidase. Analysis of the binding curves by nonlinear regression demonstrated a K d of 2.9 Ϯ 0.3, 2.0 Ϯ 0.6, and 4.3 Ϯ 1.5 nM for the WT, Dom3 Ala , and Dom9 Ala constructs, respectively (Fig. 7). Analysis of these data by the method of Scatchard (25), resulted in a linear plot (Fig. 7, inset) that extrapolated to 0.8, 1.0, and 0.6 mol of ␤-glucuronidase bound per mol of monomeric receptor for the WT, Dom3 Ala , and Dom9 Ala constructs, respectively. These data are consistent with the stoichiometry of 1 mol bound per mol of IGF-II/MPR reported for the lysosomal enzyme, ␤-galactosidase (12). These results indicate that the mutant receptors are similar to the WT receptor in their affinity for ␤-glucuronidase. However, the mutant receptors differ slightly from each other in that Dom9 Ala has ϳ2-fold lower affinity for ␤-glucuronidase than Dom3 Ala .
Dissociation of WT and Mutant IGF-II/MPRs from Pentamannosyl Phosphate-Agarose Affinity Columns-The targeting of lysosomal enzymes by MPRs is a multistep process that involves both the binding of ligand in the Golgi and the release of the lysosomal enzyme in the acidic prelysosomal compartment (1,37). To test the ability of the WT, Dom3 Ala, and Dom9 Ala constructs to dissociate ligand at acidic pH, purified IGF-II/MPRs were passed over pentamannosyl phosphate-agarose affinity columns (Fig. 8). Although the majority of WT and Dom9 Ala receptors dissociated from the column at acidic pH (pH 4.6), only 45% of the Dom3 Ala receptor was present in the acidic pH eluate. Similar results were observed when the purified IGF-II/MPRs were passed over pentamannosyl phosphate-agarose affinity columns that were eluted with buffer at pH 5.5 (data not shown). These results demonstrate that the Dom3 Ala construct is impaired in its ability to dissociate ligand at acidic pH and that the Dom3 Ala and Dom9 Ala constructs display differences in their ligand dissociation properties.

DISCUSSION
MPRs mediate the lysosomal targeting of soluble acid hydrolases, a heterogeneous population of Ͼ40 enzymes that differ in size, oligomeric state, number of N-linked oligosaccharides, extent of phosphorylation of their oligosaccharide chains, and/or the position of the Man-6-P moiety and its linkage to the penultimate mannose residue in the oligosaccharide chain. The CD-MPR binds 1 mol of Man-6-P/polypeptide. However, since the receptor exists as a dimer, the functional form of the CD-MPR, like that of the IGF-II/MPR, contains two Man-6-P binding sites. The conservation of two Man-6-P binding sites in both MPRs suggests that the receptors have evolved to contain two carbohydrate binding sites to efficiently target the diverse array of acid hydrolases to the lysosome. To test this hypothesis, the ability of IGF-II/MPRs containing a single functional Man-6-P binding site to target lysosomal enzymes to the lysosome was evaluated. To carry out these studies, we took advantage of our previous work in which a single amino acid substitution in each binding site was shown to inhibit Man-6-P binding as assessed by pentamannosyl-phosphate agarose affinity chromatography (17). cating that the presence of a guanidinium group at these positions is essential for ligand binding. The three-dimensional structure of the bovine CD-MPR reveals that Arg 111 , which aligns with Arg 435 and Arg 1334 of the IGF-II/MPR, directly coordinates Man-6-P (39). Thus, it is very likely that the analogous arginine residue in domain 3 and domain 9 also coordinate Man-6-P, as supported by our modeling studies on the IGF-II/MPR. Therefore, it is not surprising that the arginine substitutions specifically affect the binding of Man-6-P-containing ligands and not the stability, trafficking, or IGF-II binding of the receptor as demonstrated by our control experiments.
To test the ability of mutant IGF-II/MPRs to target lysosomal enzymes to the lysosome, mouse L (Rec Ϫ ) cells, which do not contain endogenous IGF-II/MPR, were used to generate stably transfected cell lines expressing WT and mutant IGF-II/MPRs. Our results demonstrated that mutant IGF-II/MPRs containing a single Man-6-P binding site (Dom3 Ala and Dom9 Ala ) were only ϳ50% as effective as the WT receptor in targeting the total phosphorylated population of lysosomal enzymes to the lysosome (Fig. 2). These data support the hypothesis that the IGF-II/MPR evolved two Man-6-P binding sites in order to efficiently target the heterogeneous population of lysosomal enzymes to the lysosome. The decrease in the overall targeting efficiency of IGF-II/MPRs with one Man-6-P binding site may be attributed to either a decrease in the affinity for all lysosomal enzymes or the inability to bind certain enzymes. Our results suggest that IGF-II/MPRs with one Man-6-P binding site display a general decrease in the affinity for Man-6-Pcontaining ligands. This decrease in affinity may be due to the lack of bivalent binding in the mutant receptors that contain a single Man-6-P binding site. It is well documented that the presence of multiple carbohydrate binding sites in a lectin can contribute to significantly higher binding affinities, as has been shown for the rat asialoglycoprotein receptor, the chicken hepatic lectin, and the macrophage mannose receptor (40,41). The analysis of the individual bands secreted by stably transfected cells also suggests that mutant receptors with a functional Man-6-P binding site in domains 1-3 differ from receptors with a single Man-6-P binding site in domains 7-9 in their ability to target certain lysosomal enzymes.
Examination of the individual enzymes, ␤-hexosaminidase and ␤-glucuronidase, both of which are heavily glycosylated oligomers, and the monomeric cathepsin D has shown that the efficiency of sorting by IGF-II/MPRs with one Man-6-P binding site varied from enzyme to enzyme and depended on the loca-tion of the single functional Man-6-P binding site. IGF-II/MPRs with a functional Man-6-P binding site in domains 7-9 (Dom3 Ala ) were only 35% as efficient as the WT receptor in the targeting of cathepsin D, a monomeric aspartyl protease that contains only two potential N-linked oligosaccharides (42). In contrast, the sorting of cathepsin D in cell lines expressing IGF-II/MPRs containing a functional Man-6-P binding site in domains 1-3 (Dom9 Ala ) was not significantly different from cell lines transfected with vector alone (Fig. 3). Analysis of the sorting of ␤-glucuronidase, a more heavily glycosylated lysosomal enzyme, revealed that both the Dom3 Ala and Dom9 Ala constructs could target this enzyme to the lysosome ϳ30% as efficiently as the WT receptor. However, at higher receptor expression levels, Dom3 Ala was able to target ␤-glucuronidase to a similar extent as the WT receptor (Fig. 5). Murine ␤-glucuronidase is a homotetramer, and each monomer contains four potential N-linked oligosaccharides of which three sites are utilized and partially phosphorylated (43). Thus, ␤-glucuronidase contains 12 N-linked oligosaccharides that are phosphorylated, whereas cathepsin D contains only two potential sites. Our data indicate that the sorting ability of the IGF-II/ MPRs with one Man-6-P binding site may correlate with the extent of glycosylation and phosphorylation on the ligands: the greater the number of N-linked oligosaccharides, the higher the probability that a phosphorylated mannose residue will be in a proper location and conformation for a mutant IGF-II/MPR to recognize, which may explain why IGF-II/MPRs with one Man-6-P binding site can target ␤-glucuronidase better than cathepsin D. These data also suggest that IGF-II/MPRs may have evolved two Man-6-P binding sites to increase the probability of recognizing a heterogeneous family of ligands.
Surprisingly, our studies demonstrated differences between the Dom3 Ala and the Dom9 Ala constructs in their ability to target certain lysosomal enzymes. The most dramatic example of differences between the two Man-6-P binding sites was seen in the analysis of ␤-hexosaminidase targeting. ␤-hexosaminidase is a dimer composed of a combination of ␣ and ␤ chains, which contain four and five potential sites for N-linked oligosaccharides, respectively (44,45). This heavily glycosylated enzyme is targeted by the Dom3 Ala construct as efficiently as the WT receptor, whereas the sorting of ␤-hexosaminidase in cell lines expressing the Dom9 Ala construct was no different from cell lines transfected with vector alone (Fig. 4). These results demonstrate that the two mutant IGF-II/MPRs containing a single Man-6-P binding site display differences in their ability to target different lysosomal enzymes, suggesting that the ligand binding properties of the two carbohydrate recognition domains differ.
Analysis of the elution profiles of the mutant and WT receptor from a pentamannosyl phosphate-agarose affinity column indicated that IGF-II/MPRs with one functional Man-6-P binding site have a lower affinity than the WT receptor for ligand and that the Dom9 Ala construct has a lower affinity than the Dom3 Ala construct (Fig. 6). To quantitatively determine the differences in binding affinity, equilibrium binding experiments using ␤-glucuronidase demonstrated that the Dom3 Ala construct has a K d of 2.0 nM, which is very similar to that of the WT IGF-II/MPR. The Dom9 Ala construct has an ϳ2-fold decrease in affinity in comparison with the Dom3 Ala construct. Although the affinities of the Dom3 Ala and Dom9 Ala constructs for ␤-glucuronidase are not dramatically different from each other or that of the WT receptor in vitro, their ability to target this enzyme to the lysosome in vivo is significantly different from that of the WT receptor (Fig. 5). One explanation for the minimal differences in binding affinity may be that the IGF-II/MPR is oligomeric. The oligomeric state of this receptor is controversial, since there are conflicting reports, one that suggests it is a monomer (46) and another that suggests it is an oligomer (47). If the IGF-II/MPR is oligomeric, than the mutant receptors may also participate in multivalent binding of ligands and explain the high affinity binding for ␤-glucuronidase. It is also possible that the differences observed in binding affinity were minimized, since the human ␤-glucuronidase used for the in vitro K d determinations may be a better ligand than the endogenous mouse enzyme assayed in sorting. Human ␤-glucuronidase contains four N-linked oligosaccharides on each monomer, and it is known that two sites are preferentially phosphorylated (48). On the other hand, murine ␤-glucuronidase contains three N-linked oligosaccharides, and each is partially phosphorylated, with the majority existing as phosphodiesters (43). Tong et al. (12) has shown that the affinity of the IGF-II/MPR for phosphodiesters is 50-fold lower than the affinity for phosphomonoesters. The in vitro binding studies are consistent with the targeting data in that the Dom9 Ala construct has the lowest affinity for ␤-glucuronidase and is the least efficient in targeting this enzyme (Fig. 5). Although the decrease in affinity is minor (ϳ2-fold), this may be significant in the context of intracellular events. It has been shown that a small difference in affinity determined in vitro can have a significant impact on intracellular targeting. Sleat and Lobel (49) reported a K d determined in vitro for cathepsin D of 2.3 and 7 nM for the IGF-II/MPR and the CD-MPR, respectively. Although the difference between affinities is only 3-fold, the targeting efficiency of each receptor is significantly different. Mouse embryonic fibroblasts expressing only the CD-MPR target 10% of their cathepsin D to the lysosome, in contrast to the 30% targeted by cells expressing only the IGF-II/MPR (50).
Since the Dom3 Ala construct's affinity for ␤-glucuronidase was not significantly different from that of the WT receptor but its ability to target this enzyme in vivo was impaired at lower levels of receptor expression, it is possible that another step in the lysosomal enzyme targeting pathway may be affected. The targeting of newly synthesized acid hydrolases involves both the binding and release of ligand. The receptor-enzyme complex forms in the Golgi and is transported to a prelysosomal compartment. The low pH of this endosomal compartment induces dissociation of the complex by causing a change in the conformation of the receptor (51). A change in the ability to dissociate ligand can have an overall impact in the targeting pathway. This is demonstrated in the treatment of cells with the lysosomotropic amines, chloroquine or NH 4 Cl, which raise the intraorganelle pH and impair the receptor-enzyme dissociation. The inability to dissociate ligand results in constantly occupied receptors and thus stimulates the secretion of newly synthesized acid hydrolases (52,53). Our results indicate that the response to low pH is impaired in the Dom3 Ala construct but not in the Dom9 Ala construct (Fig. 8). In comparison with the WT receptor, impairment of ligand dissociation at acidic pH by the Dom3 Ala construct provides a possible explanation for the observed lower efficiency in the overall targeting of lysosomal enzymes. Furthermore, the different extent to which the Dom3 Ala and Dom9 Ala constructs were eluted from the pentamannosyl phosphate-agarose affinity columns at pH 4.6 and 5.5 indicate that the two Man-6-P binding sites differ in their ligand dissociation properties and provide a possible mechanism for the differences observed in their ability to target lysosomal enzymes to the lysosome in vivo.
In summary, IGF-II/MPRs containing one, rather than two, functional Man-6-P binding site are less efficient than the WT receptor in targeting lysosomal enzymes, suggesting that the IGF-II/MPR has evolved two Man-6-P binding sites to target its heterogeneous population of ligands and/or to allow bivalent binding and thus more efficient targeting of ligands. The differences observed in the ability to target select lysosomal enzymes to the lysosome, the affinity of binding ␤-glucuronidase, and the dissociation of ligand suggest that the two individual Man-6-P binding sites are not functionally equivalent. Future studies are being directed toward a detailed characterization of the binding and dissociation properties of the two Man-6-P binding sites of the IGF-II/MPR.