Identification of Residues Essential for Carbohydrate Recognition by the Insulin-like Growth Factor II/Mannose 6-Phosphate Receptor*

Two distinct mannose 6-phosphate (Man-6-P) receptors (MPRs), the cation-dependent MPR (CD-MPR) and the insulin-like growth factor II/MPR (IGF-II/MPR), recognize a diverse population of Man-6-P-containing ligands. The IGF-II/MPR is a type I transmembrane glycoprotein with a large extracytoplasmic region composed of 15 repeating domains that display sequence identity to each other and to the single extracytoplasmic domain of the CD-MPR. A structure-based sequence alignment of the two distinct Man-6-P-binding sites of the IGF-II/MPR with the CD-MPR implicates several residues of IGF-II/MPR domains 3 and 9 as essential for Man-6-P binding. To test this hypothesis single amino acid substitutions were made in constructs encoding either the N- or the C-terminal Man-6-P-binding sites of the bovine IGF-II/MPR. The mutant IGF-II/MPRs secreted from COS-1 cells were analyzed by pentamannosyl phosphate-agarose affinity chromatography, identifying four residues (Gln-392, Ser-431, Glu-460, and Tyr-465) in domain 3 and four residues (Gln-1292, His-1329, Glu-1354, and Tyr-1360) in domain 9 as essential for Man-6-P recognition. Binding affinity studies using the lysosomal enzyme, β-glucuronidase, confirmed these results. Together these analyses provide strong evidence that the two Man-6-P-binding sites of the IGF-II/MPR are structurally similar to each other and to the CD-MPR and utilize a similar carbohydrate recognition mechanism.

Lysosomes, membranous organelles containing numerous acid hydrolases, play an essential role in the degradative metabolism of mammalian tissues and cells. Targeting of newly synthesized acid hydrolases to lysosomes is dependent upon specific recognition in the trans-Golgi network of mannose 6-phosphate (Man-6-P) 1 residues found on the N-linked oligosaccharides of lysosomal enzymes by the two members of the P-type lectin family, the cation-dependent Man-6-P receptor (CD-MPR) and the insulin-like growth factor II/Man-6-P recep-tor (IGF-II/MPR). Enzyme-bound receptors are subsequently transported to acidified, prelysosomal compartments where the low pH (Ͻ6) environment induces release of the enzymes from the receptors. Following enzyme release, the MPRs recycle back to the Golgi to repeat this process or move to the cell surface where the IGF-II/MPR, but not the CD-MPR, binds and internalizes extracellular ligands (1)(2)(3).
In recent years the repertoire of identified extracellular ligands of the IGF-II/MPR has expanded to include a diverse spectrum of Man-6-P-containing target proteins. IGF-II/MPR binding to the mannose 6-phosphorylated latent form of transforming growth factor-␤ at the cell surface results in the proteolytic activation of this critical growth factor that regulates the cellular differentiation and proliferation of many cell types (4 -7). Cell surface recognition of the Man-6-P-modified placental hormone, proliferin, by the IGF-II/MPR is required for proliferin-induced angiogenesis (8 -10). The precursor form of the aspartic proteinase renin (prorenin), a key enzyme of the cardiac renin-angiotensin system, contains Man-6-P residues that enable IGF-II/MPR binding and internalization of prorenin, resulting in its subsequent proteolytic activation in endosomal compartments (11)(12)(13)(14). Circulating levels of the potent cytokine leukemia inhibitory factor are modulated via IGF-II/ MPR-mediated endocytosis and targeting of this Man-6-P-containing protein for degradation in the lysosomes (15,16). Entry into cells and transmission between cells of herpes simplex virus can be facilitated by IGF-II/MPR binding of Man-6-Pmodified herpes simplex viral glycoprotein D (17)(18)(19)(20). The IGF-II/MPR also functions as a death receptor by mediating uptake of mannose 6-phosphorylated granzyme B, a serine proteinase that is essential for the rapid induction of target cell apoptosis by cytotoxic T cells (21,22). Furthermore, internalization of the Man-6-P-containing T cell activation antigen, CD26, by the IGF-II/MPR plays an important role in CD26-mediated T cell costimulation (23). The ability of the IGF-II/MPR to recognize with high affinity the Man-6-P signal found on many functionally distinct ligands underscores the importance of the IGF-II/ MPR and its involvement in a myriad of essential physiological pathways.
The IGF-II/MPR is an ϳ300-kDa type I transmembrane glycoprotein that consists of an N-terminal signal sequence, a large extracytoplasmic region composed of 15 homologous repeating domains, a single transmembrane region, and a Cterminal cytoplasmic domain ( Fig. 1) (24). The ϳ46-kDa CD-MPR is a much smaller type I transmembrane glycoprotein that in some species requires divalent cations for optimal ligand binding (25)(26)(27). Significantly, each of the 15 IGF-II/MPR extracytoplasmic domains displays amino acid sequence identity (14 -38%), similar size (ϳ147 residues), and cysteine distribution to each other and to the single, extracytoplasmic domain of the CD-MPR, giving rise to the prediction that they exhibit similar disulfide bonding and tertiary structures (24). The CD-MPR binds 1 mol of Man-6-P per polypeptide (27), whereas the IGF-II/MPR binds 2 mol of Man-6-P per mol of receptor via two distinct carbohydrate-binding sites localized to extracytoplasmic domains 1-3 and 7-9 (28 -30). The IGF-II/ MPR, unlike the CD-MPR, also binds with high affinity to the nonglycosylated peptide hormone, IGF-II, at a binding site localized to extracytoplasmic domain 11 ( Fig. 1) (31,32).
In contrast to the CD-MPR, little is known about the amino acid residues involved in carbohydrate recognition by the IGF-II/MPR. Previous biochemical studies (30,33) have identified only a single arginine residue in each of the two IGF-II/MPR Man-6-P-binding sites, Arg-435 in domain 3 and Arg-1334 in domain 9, to be important for Man-6-P binding by the IGF-II/ MPR. However, the recently determined three-dimensional structure of the extracytoplasmic region of the bovine CD-MPR complexed with Man-6-P (34) or pentamannosyl phosphate (35) not only reveals the nature of carbohydrate recognition by the CD-MPR, but it also provides the framework from which to decipher the molecular basis of Man-6-P recognition by the IGF-II/MPR as described in this report. Consequently, sitedirected mutagenesis was used to generate single amino acid substitutions in soluble, truncated forms of the IGFII/MPR in order to test the hypothesis that residues of IGF-II/MPR extracytoplasmic domains 3 and 9 predicted by a structure-based sequence alignment with the CD-MPR to be in the N-and C-terminal Man-6-P-binding pockets of the IGF-II/MPR, respectively, are essential for high affinity Man-6-P binding. By using pentamannosyl phosphate-agarose affinity chromatography, four amino acid residues (Gln-392, Ser-431, Glu-460, and Tyr-465) in domain 3 and four residues (Gln-1292, His-1329, Glu-1354, and Tyr-1360) in domain 9 were identified as essential for carbohydrate recognition by the IGF-II/MPR in addition to the previously determined Arg-435 (domain 3) and Arg-1334 (domain 9) residues. To assess quantitatively the relative contribution of each of the residues essential to the Man-6-P binding ability of the IGF-II/MPR, binding affinity studies of the mutant constructs for the lysosomal enzyme, ␤-glucuronidase, were performed. Together these results provide strong evidence that the two IGF-II/MPR Man-6-P-binding sites utilize a mechanism similar to that of the CD-MPR for high affinity Man-6-P binding and that the N-and C-terminal carbohydrate recognition domains of the IGF-II/MPR are structurally similar to each other and to the CD-MPR.

EXPERIMENTAL PROCEDURES
Materials-The following reagents were obtained commercially as indicated: restriction endonucleases, T4 polynucleotide kinase, T7 DNA polymerase, uracil glycosylase inhibitor, M13K07 helper phage, CJ236, JM109, and NM522 Escherichia coli strains were from New England Biolabs; T4 DNA ligase, BenchMark prestained protein ladder, and trypsin-EDTA were from Invitrogen; Wizard Plus SV plasmid DNA minipreps were from Promega; GeneMate plasmid DNA minipreps were from ISC Bioexpress; acrylamide was from Schwarz/Mann; SDS was from BDH Biochemical; ammonium persulfate and TEMED were from Bio-Rad; glucose 6-phosphate, Man-6-P, -aminoethyl agarose, Dulbecco's modified Eagle's medium, and lactoperoxidase were from Sigma; fetal bovine serum was from HyClone Laboratories or Invitrogen; EXPRE 35 S 35 S 35 S-protein labeling mix (1200 Ci/mmol) was from PerkinElmer Life Sciences; and Rapid Gel 6% acrylamide, Thermo-Sequenase kit, 14  Generation of Wild-type and Mutant IGF-II/MPR N-and C-terminal Man-6-P-binding Site cDNAs-The wild-type constructs encoding the bovine IGF-II/MPR signal sequence and extracytoplasmic domains 7-11 (Dom7-11) or domains 1-3 (Dom1-3His) and domains 7-9 (Dom7-9His) followed by a C-terminal tag of 6 histidine residues (Fig. 1) were generated as described previously (33). Each of these constructs was cloned into the pBluescript SK vector (Stratagene) to prepare single-stranded DNA templates for site-directed mutagenesis or into the pcDNA3 vector (Invitrogen) for expression in mammalian cells. Single amino acid substitutions of residues predicted to form the N-or C-terminal IGF-II/MPR Man-6-P-binding sites were generated according to the method of Kunkel et al. (36). The mutagenic oligonucleotides (Operon) used are as follows (each substitution is indicated by the A polymerase chain reaction-based strategy was used to generate the Dom7-11His construct by the addition of 6 histidine residues (CAC) and a stop codon (TGA) following the C-terminal residue, Glu-1656, of the wild-type Dom7-11 cDNA construct. The oligonucleotide primer used was the reverse complement of the following: 5Ј-TGGCA-CACACCCCTGGCCTGCGAGCACCACCACCACCACCACTGAATTC-TAGACCGG-3Ј. Each of the predicted sequences was confirmed by DNA sequence analysis using an Amersham Biosciences ALF DNA sequencer.
Pentamannosyl Phosphate-Agarose Affinity Chromatography-COS-1 cells (American Type Culture Collection) were cultured as described previously (37) and transiently transfected using the DEAEdextran (Amersham Biosciences) technique (30) with pcDNA3 vectors containing wild-type or mutant Dom1-3His, Dom7-9His, or Dom7-11 cDNAs. At 48 h post-transfection, the cells were metabolically labeled with [ 35 S]methionine/[ 35 S]cysteine for an additional 24 h. The medium was harvested and dialyzed, and the COS-1 cells were solubilized as detailed previously (37). To assay the constructs for ligand binding ability, dialyzed medium containing the secreted wild-type or mutant receptors was subjected to pentamannosyl phosphate-agarose affinity chromatography as described previously (37). Subsequently, the affinity chromatography samples were immunoprecipitated with rabbit polyclonal antisera specific for the bovine IGF-II/MPR plus protein A-Sepharose (Sigma) at 4°C for 16 -24 h as described (37). To determine the percentage of each construct that was secreted into the medium, equal aliquots of cell lysates and dialyzed media were subjected to immunoprecipitation as described above. The immunoprecipitated samples were subjected to SDS-PAGE, and the radioactivity was quantified using a PhosphorImager (Molecular Dynamics Storm 860) and ImageQuant (version 4.1) software.
Preparation of IGF-II/MPR Man-6-P-binding Site Constructs for Binding Affinity Analyses-To prepare Dom1-3His and Dom7-9His receptors for binding affinity studies, COS-1 cells were transiently transfected using the DEAE-dextran technique (30) with pcDNA3 vectors containing the wild-type or mutant Dom1-3His or Dom7-9His cDNAs. Forty eight hours after transfection the cells were rinsed briefly with phosphate-buffered saline (pH 7.3) and incubated for 24 h in Waymouth's MB 752/1 serum-free medium (Invitrogen). The medium was harvested and dialyzed against buffer containing 20 mM Tris (pH 8.0), 0.5 M NaCl, and 8 mM imidazole (binding buffer). Batch binding of the His 6 -tagged receptors was performed by adding glycerol (10% final concentration), Nonidet P-40 (0.1% final concentration), and nickelnitrilotriacetic acid (Ni-NTA, Qiagen) agarose (ϳ200 l resin/50 ml of medium) to the dialyzed medium and incubating at 4°C for 16 -24 h. The Ni-NTA-agarose was transferred to a column and washed extensively with binding buffer and then eluted with binding buffer containing 250 mM imidazole (pH 8.0). Aliquots of each Ni-NTA eluate were subjected to the Bio-Rad (Bradford) protein assay using bovine serum albumin standards to determine the approximate amount of each Dom1-3His or Dom7-9His protein obtained. An average yield of 11 g of protein was purified from the medium of each transfected 150-cm 2 cell culture flask. The relative amounts of Dom1-3His or Dom7-9His proteins obtained was confirmed by performing Western blot analysis as detailed previously (33) using bovine IGF-II/MPR-specific antiserum and protein A/horseradish peroxidase conjugate (Amersham Biosciences) or ImmunoPure goat anti-rabbit IgG/horseradish peroxidase conjugate (Pierce) followed by detection using SuperSignal West Pico chemiluminescent substrate (Pierce).
Binding Affinity Analysis-125 I-Labeled (1-2 Ci/g) human ␤-glucuronidase was prepared and used to perform binding affinity analyses as described previously (38). Briefly, increasing concentrations of iodinated ␤-glucuronidase were incubated with purified wild-type or mutant Dom1-3His or Dom7-9His, and the receptors and bound ligand were immunoprecipitated with bovine IGF-II/MPR-specific antiserum prebound to protein A-Sepharose beads. The beads were washed, and bound ␤-glucuronidase was specifically eluted from the receptors by incubation with 10 mM Man-6-P. The results were analyzed by nonlinear regression (SigmaPlot version 5.05, SPSS Science).

Site-directed Mutagenesis of the N-and C-terminal Man-6-P-binding
Sites of the IGF-II/MPR-A structure-based sequence alignment of the extracytoplasmic region of the CD-MPR with domains 3 and 9 of the IGF-II/MPR and molecular modeling identified several residues in domain 3 (Tyr-368, Gln-392, Ser-430, Ser-431, Arg-435, Thr-458, Glu-460, and Tyr-465) and domain 9 (Tyr-1264, Gln-1292, His-1329, Arg-1334, Leu-1352, Glu-1354, and Tyr-1360) that are located at positions equivalent to CD-MPR residues that serve as potential hydrogen bond donors/acceptors with Man-6-P ( Fig. 2A) (34). We predicted that these residues are important for Man-6-P recognition by the IGF-II/MPR. To test this hypothesis, sitedirected mutagenesis was used to generate single amino acid substitutions in cDNA constructs encoding either the N-(Dom1-3His) or the C-terminal (Dom7-9His or Dom7-11) Man-6-P-binding site of the IGF-II/MPR (Figs. 1 and 2B). The wild-type Dom1-3His, Dom7-9His, and Dom7-11 constructs were demonstrated previously (33) to bind the lysosomal enzyme, ␤-glucuronidase, with affinities similar to that of the full-length wild-type receptor. In addition, the wild-type Dom7-11 construct was shown to contain a functional IGF-IIbinding site (domain 11, Fig. 1), providing additional evidence that this recombinant protein is folded into its native state (33). Two criteria were used to select the single amino acid substitutions that were generated. First, substitutions were selected that would eliminate or significantly perturb hydrogen bonding potential. Second, only amino acid substitutions considered "safe" according to the exchange matrices of Bordo and Argos (39) were selected to reduce the possibility of disrupting the structural integrity of the IGF-II/MPR constructs. In summary, 20 different point mutations were generated to identify and characterize the role of residues predicted to be important for carbohydrate recognition by the two distinct Man-6-P-binding sites of the IGF-II/MPR (Fig. 2B).
Expression of the N-and C-terminal Man-6-P-binding Sites of the IGF-II/MPR-The use of truncated IGF-II/MPR constructs allows us to study separately the two IGF-II/MPR Man-6-P-binding sites. In addition, since Dom1-3His, Dom7-9His, and Dom7-11 receptors lack the transmembrane region of the IGF-II/MPR, these soluble recombinant proteins are secreted into the medium, facilitating their efficient separation from endogenous MPRs when expressed in mammalian cells. Furthermore, previous studies in our laboratory (30) indicate that secreted MPRs are folded into a conformation conducive to ligand binding, whereas misfolded MPRs are retained in the cells. Thus, the efficiency of secretion of each IGF-II/MPR construct provides an indirect assessment of correct folding. Consequently, the wild-type and mutant Dom1-3His, Dom7-9His, and Dom7-11 constructs were expressed in transiently transfected COS-1 cells, and the efficiency of secretion of each construct was assayed (data not shown). With the exception of the Q1292N Dom7-9His mutant (no detectable secretion) and the Q1292N (secretion efficiency ϳ29%) and L1352A (secretion efficiency ϳ14%) Dom7-11 mutants, all of the IGF-II/MPR mutants were efficiently secreted (i.e. Ͼ50% of the total receptor secreted at steady state).

Identification of IGF-II/MPR Residues Essential for Man-6-P Recognition by Pentamannosyl Phosphate-Agarose Affinity
Chromatography-To identify residues important for carbohydrate recognition by the IGF-II/MPR COS-1 cells were transfected with wild-type or mutant Dom1-3His or Dom7-11 constructs. Subsequently, the cells were metabolically labeled with [ 35 S]methionine/[ 35 S]cysteine, and the medium containing the secreted receptors was harvested and subjected to pentamannosyl phosphate-agarose affinity chromatography. Wild-type Dom1-3His (ϳ70% Man-6-P binding) and wild-type Dom7-11 (ϳ66% Man-6-P binding) exhibited similar specific binding to pentamannosyl phosphate-agarose affinity columns (Fig. 3). In contrast, specific binding to the affinity columns was eliminated when single amino acid substitutions were made of Gln-392, Ser-431, Glu-460, or Tyr-465 in domain 3 or of Gln-1292, His-1329, Glu-1354, or Tyr-1360 in domain 9. Although substitution of Leu-1352 with alanine also eliminated the Man-6-P binding ability of Doms7-11, this mutant construct was poorly secreted (secretion efficiency ϳ14%) from COS-1 cells, suggesting that this mutation disrupts folding instead of having a specific effect on binding. This prediction is supported by the results obtained with the L1352V mutant, which is efficiently secreted (secretion efficiency ϳ57%) and retains Man-6-P binding ability. Binding to pentamannosyl phosphate affinity columns was also retained when Thr-458 in domain 3 was replaced, whereas substitutions of Tyr-368 or Ser-430 in domain 3 or Tyr-1264 in domain 9 resulted in only a partial inhibitory effect on Man-6-P binding. In summary, these studies identified four residues (Gln-392, Ser-431, Glu-460, and Tyr-465) in domain 3 and four residues (Gln-1292, His-1329, Glu-1354, and Tyr-1360) in domain 9 as essential for Man-6-P recognition by the IGF-II/MPR (Fig. 3) in addition to the previously identified Arg-435 (domain 3) and Arg-1334 (domain 9) residues (30,33).
Preparation of Purified IGF-II/MPR Man-6-P-binding Site Constructs for Binding Affinity Analyses-In preparation for binding affinity studies, nickel-nitrilotriacetic acid (Ni-NTA)agarose affinity chromatography was used to purify wild-type and selected Dom1-3His mutants that exhibited a partial inhibitory effect (Y368F) or a complete elimination (Q392N, S431A, E460D, and Y465F) of Man-6-P binding ability. Unlike the Dom1-3His construct, the Dom7-11 construct used to identify essential Man-6-P-binding site residues in domain 9 of the IGF-II/MPR does not contain a His 6 tag. Consequently, a Dom7-11His construct was generated by adding a tag of six histidine residues to the C terminus of the wild-type Dom7-11 construct. Surprisingly, the Dom7-11His construct was retained within transiently transfected COS-1 cells 2 in contrast to the wild-type Dom7-11 construct, which was efficiently secreted (ϳ89%) into the medium (data not shown). In addition, Dom7-11His from solubilized COS-1 cells did not bind to Ni-NTA-agarose affinity columns although denatured Dom7-11His could be detected by Western blotting using an anti-His antibody (Tetra-His antibody, Qiagen). 2 Dom7-11His from solubilized COS-1 cells also did not bind to pentamannosyl phosphate-agarose affinity columns. 2 Together these results suggest that the presence of a His 6 tag at the C terminus of the Doms7-11 construct disrupts its structural integrity. Therefore, selected single amino acid substitutions that partially inhibited (Y1264F) or eliminated (Q1292N, H1329N, E1354D, and Y1360F) the Man-6-P binding ability of the Dom7-11 construct were generated in the previously characterized Dom7-9His construct (33). To confirm the effect of the selected amino acid substitutions on the Man-6-P binding ability of the Dom7-9His construct, transiently transfected COS-1 cells were metabolically labeled with [ 35 S]methionine/[ 35 S]cysteine, and the medium containing the secreted receptors was harvested and subjected to pentamannosyl phosphate-agarose affinity chromatography. Consistent with the results obtained with the Dom7-11 construct, the Y1264F mutation had a partial inhib-2 B. Janowiak, M. Hancock, and N. Dahms, unpublished data.

FIG. 2. Residues of the IGF-II/MPR targeted for mutagenesis.
A, structurebased amino acid sequence alignment of the bovine CD-MPR with domain 3 (Dom3) and domain 9 (Dom9) of the bovine IGF-II/MPR (34). The secondary structure of the CD-MPR is shown above the sequence (wavy line represents the single ␣-helix, and arrows represent ␤-strands). The 9 residues that compose the carbohydrate-binding pocket of the CD-MPR are indicated (q). The residues of the IGF-II/MPR that were subjected to site-directed mutagenesis are shaded. The arginine residues, Arg-345 in domain 3 and Arg-1334 in domain 9, that were previously mutated (30) 1.0 cm). The columns were washed extensively and then eluted first with 10 mM glucose 6-phosphate (nonspecific sugar) and then with 10 mM Man-6-P. The run through (RT), wash (W), glucose 6-phosphate (G) eluates, and Man-6-P (M) eluates were immunoprecipitated with IGF-II/MPR-specific antisera and analyzed by SDS-PAGE. The radioactive bands were visualized using a Phosphor-Imager, and a representative gel of each construct is shown. The percent, representing the mean Ϯ S.E. for a minimum of three independent determinations, of each construct found in the Man-6-P eluate is also indicated. itory effect, whereas the H1329N, E1354D, and Y1360F single amino acid substitutions abolished the Man-6-P binding ability of the Dom7-9His construct (Fig. 3). Interestingly, wild-type Dom7-9His (ϳ40% Man-6-P binding) and the Y1264F Dom7-9His mutant (ϳ25% Man-6-P binding) each exhibited significantly lower specific binding to pentamannosyl phosphate-agarose affinity columns as compared with wild-type Dom7-11 (ϳ66% Man-6-P binding) and the Y1264F Dom7-11 mutant (ϳ44% Man-6-P binding), respectively. These results suggest that the presence of a His 6 tag at the C terminus of the Dom7-9His construct and/or the absence of domains 10 and 11 negatively influence the Dom7-9His construct such that less than half of the total Dom7-9His receptor population is capable of specific binding to pentamannosyl phosphate-agarose affinity columns under the conditions employed in these experiments. Nevertheless, previous studies (33) that analyzed the inhibition of ␤-glucuronidase binding to wild-type Dom7-9His by Man-6-P as well as the binding affinity analyses presented below demonstrate that the wild-type Dom7-9His construct maintains very similar affinities for Man-6-P and for ␤-glucuronidase as reported for Dom7-11 and the full-length IGF-II/ MPR (33,40). Therefore, in preparation for binding affinity studies, purified wild-type and selected mutant Dom7-9His receptors were obtained by Ni-NTA-agarose affinity chromatography, with the exception of the Q1292N Dom7-9His mutant which exhibited no detectable secretion from transiently transfected COS-1 cells (data not shown).
Binding Affinities of the Wild-type and Mutant Dom1-3His and Dom7-9His Constructs-To assess quantitatively the relative importance and contribution of each of the essential residues of domain 3 and domain 9 to the Man-6-P binding ability of the IGF-II/MPR, binding affinity studies were performed using the well characterized lysosomal enzyme, ␤-glucuronidase (41). In addition, the binding affinities of two mutants, Y368F (domain 3) and Y1264F (domain 9), which exhibited a partial inhibitory effect on Man-6-P binding, were analyzed. Increasing concentrations of iodinated ␤-glucuronidase were incubated with wild-type or Y368F Dom1-3His or with wildtype or Y1264F Dom7-9His receptors. Identical K d values were obtained for wild-type Dom1-3His (0.5 Ϯ 0.1 nM, Fig. 4A) and wild-type Dom7-9His (0.5 Ϯ 0.1 nM, Fig. 4C), which compare well with the previously determined K d values for wild-type Dom1-3 (1.0 Ϯ 0.3 nM) and wild-type Dom7-9 (2.5 Ϯ 0.9 nM) (33) as well as to the affinity of the full-length wild-type IGF-II/MPR (K d ϭ 2 nM) (40). Consistent with a partial inhibitory effect on Man-6-P binding ability, the Y368F Dom1-3His (K d ϭ 31 Ϯ 10 nM) and Y1264F Dom7-9His (K d ϭ 87 Ϯ 26 nM) mutants exhibited 62-and 174-fold, respectively, lower binding affinities for ␤-glucuronidase than the wild-type receptors (Fig.  4, B and D). In contrast, K d values could not be obtained for each of the selected Dom1-3His (Q392N, S431A, E460D, and Y465F) and Dom7-9His (H1329N, E1354D, and Y1360F) mutants that displayed no significant Man-6-P binding ability by pentamannosyl phosphate-agarose chromatography (Fig. 3). However, binding affinity analyses at two increasing concentrations of iodinated ␤-glucuronidase indicated that each of these point mutations results in a Ͼ1200-fold inhibitory effect on carbohydrate recognition by Dom1-3His or Dom7-9His (Fig. 4, A and C). DISCUSSION The two members of the P-type lectin family, the CD-MPR and the IGF-II/MPR, recognize with high affinity and specificity an expanding list of Man-6-P-containing ligands involved in a number of physiologically essential pathways. The binding specificity of the MPRs for Man-6-P is evidenced by the ϳ10,000-fold lower affinity each has for mannose and for the closely related structural analog, glucose 6-phosphate, which contains an equatorial 2-hydroxyl instead of the axial 2-hydroxyl found in Man-6-P (27,28). These results indicate that the presence of the axial 2-hydroxyl and the phosphate moiety of Man-6-P are essential for Man-6-P recognition by the MPRs. Furthermore, based on these data it was calculated that each of these two Man-6-P substituents contributes ϳ4 -5 kcal/mol of Gibbs free energy, an amount consistent with a single strong hydrogen bond or a few weak hydrogen bonds (27). However, until recently, little was known about the residues involved in Man-6-P binding by either the CD-MPR or the IGF-II/MPR. The recently obtained crystal structures of the bovine CD-MPR complexed with Man-6-P (34) or pentamannosyl phosphate (35) identified the residues that form its Man-6-P-binding pocket. All nine amino acids (Tyr-45, Gln-66, Asp-103, Asn-104, His-105, Arg-111, Glu-133, Arg-135, and Tyr-143) that form the carbohydrate-binding pocket of the bovine CD-MPR (34,35) are also in the human, mouse, and rat CD-MPR sequences (Fig. 5). Previous biochemical studies identified only a single arginine residue in each of the two IGF-II/MPR Man-6-P-binding sites, Arg-435 in domain 3 and Arg-1334 in domain 9, to be important for Man-6-P binding (30). In this report, site-directed mutagenesis was used to test the hypothesis that residues of domain 3 (Tyr-368, Gln-392, Ser-430, Ser-431, Arg-435, Thr-458, Glu-460, and Tyr-465) and domain 9 (Tyr-1264, Gln-1292, His-1329, Arg-1334, Leu-1352, Glu-1354, and Tyr-1360) predicted by a structure-sequence alignment with the CD-MPR form the N-and C-terminal Man-6-P-binding pockets of the IGF-II/ MPR, respectively.
In the CD-MPR, Tyr-45 is in proximity to interact with the 1-hydroxyl of the terminal Man-6-P, which is involved in an O-glycosidic linkage, and the 4-hydroxyl of the penultimate mannose residue (35). Previous site-directed mutagenesis studies of the CD-MPR (37) demonstrated that replacement of Tyr-45 with phenylalanine decreased binding to pentamannosyl phosphate columns by ϳ33%, indicating that this residue is not essential for carbohydrate recognition. In this report, replacement of Tyr-368 in domain 3 or Tyr-1264 in domain 9 with phenylalanine resulted in a similar partial inhibition of binding (ϳ33-38% decrease) to pentamannosyl phosphate-agarose affinity columns as compared with the wild-type IGF-II/MPR constructs (Fig. 3). These results suggest that like Tyr-45 in the CD-MPR, Tyr-368 and Tyr-1264 may play a similar nonessential role in ligand recognition by the N-and C-terminal Man-6-P-binding sites of the IGF-II/MPR (Fig. 6).
Recent site-directed mutagenesis studies have confirmed that four of the CD-MPR-binding site residues (Gln-66, Arg-111, Glu-133, and Tyr-143), which together form all but one of the contacts between the CD-MPR and the 2-, 3-, and 4-hydroxyls of the mannose ring of Man-6-P, are essential for Man-6-P binding (Fig. 6) (37). Comparison of the amino acid sequence of the CD-MPR to domain 3 and domain 9 of the IGF-II/MPR indicated that three (Gln-66, Arg-111, and Tyr-143) of the four residues are at identical positions in the two Man-6-P-binding sites of the IGF-II/MPR (Fig. 2). Arg-435 in domain 3 and Arg-1334 in domain 9, which correspond to Arg-111 in the CD-MPR, were confirmed previously to be critical for Man-6-P binding in the IGF-II/MPR (30,33). In this study, single amino acid substitutions of the IGF-II/MPR domain 3 (Gln-392 and Tyr-465) or domain 9 (Gln-1292 and Tyr-1360) residues that align with Gln-66 and Tyr-143 of the CD-MPR, also eliminated Man-6-P binding ability (Fig. 3), confirming the prediction that these residues are essential to carbohydrate recognition by the IGF-II/MPR.
The CD-MPR structure-based sequence alignment (Fig. 2) also predicted the presence of a threonine (Thr-458) in domain 3 and a leucine (Leu-1352) in domain 9 of the IGF-II/MPR at the sequence position corresponding to the essential CD-MPRbinding pocket residue, Glu-133. However, with the exception of the L1352A mutation that appeared to disrupt folding rather than have a specific effect on binding, replacement of Thr-458 or Leu-1352 had little or no significant effect on Man-6-P binding by the IGF-II/MPR (Fig. 3). On the other hand, substitutions made of the glutamate residues, Glu-460 in domain 3 or Glu-1354 in domain 9, predicted to align with Arg-135 of the CD-MPR abolished binding to pentamannosyl phosphate columns (Fig. 3). Previous studies (37) demonstrated that Man-6-P binding was retained when amino acid substitutions were made of Arg-135, which interacts with the 4-hydroxyl of Man-6-P in the CD-MPR-binding pocket (Fig. 6). Together these disparate results have led us to propose a refined CD-MPR structure-based amino acid sequence alignment in which Glu-460 (domain 3) and Glu-1354 (domain 9) align with Glu-133 of the CD-MPR, whereas Arg-135 does not align with any domain 3 or domain 9 residues (Fig. 5). Most importantly, the refined alignment predicts that the two Man-6-P-binding sites of the IGF-II/MPR contain identical, essential residues (Gln-392, Arg-435, Glu-460, and Tyr-465 in domain 3; Gln-1292, Arg-1334, Glu-1354, and Tyr-1360 in domain 9) that in the threedimensional structure of the CD-MPR (Gln-66, Arg-111, Glu-133, and Tyr-143) form the base of the Man-6-P-binding pocket and provide critical contacts to the 2-, 3-, and 4-hydroxyl groups of mannose (Figs. 5 and 6) (34,35).
In the three-dimensional structure of the CD-MPR, the phosphate moiety of Man-6-P is almost completely buried within the binding pocket by a loop containing three residues (Asp-103, Asn-104, and His-105), which are missing in IGF-II/MPR domains 3 and 9, that provide main chain and side chain interactions to the phosphate (Figs. 2 and 6) (34,35). Previous mutagenesis studies demonstrated that substitution of His-105, which provides a side chain contact with the phosphate moiety of Man-6-P, had only a partial inhibitory effect on the Man-6-P binding ability of the CD-MPR (37). Consequently, the majority of the binding energy associated with the phosphate moiety of Man-6-P likely results from a combination of side chain and main chain contacts made by the CD-MPR with the phosphate group (34,35). To identify potential residues responsible for binding the phosphate moiety of Man-6-P in the N-and C-terminal IGF-II/MPR-binding pockets, we generated single amino acid substitutions of Ser-430 and Ser-431 in domain 3 and His-1329 in domain 9 (Fig. 2). Whereas substitution of Ser-430 in domain 3 resulted in only a partial inhibitory effect on ligand binding, replacement of Ser-431 in domain 3 or His-1329 in domain 9 eliminated Man-6-P recognition (Fig. 3), indicating that Ser-431 and His-1329 play an important role in carbohydrate recognition by the IGF-II/MPR. It is likely that adjacent residues in domain 3 and domain 9 also provide important main chain and/or side chain contacts to the phosphate moiety of Man-6-P as observed in the CD-MPR (Fig. 6) (34, 35). In addition, residue differences, such as the presence of small residues (Ser-430, Ser-431, and Gly-432) in domain 3 as compared with bulky residues (His-1329, Lys-1330, and Val(Ile)-1331) in domain 9 (Figs. 2 and 5), may account for the ability of the IGF-II/MPR N-terminal Man-6-P-binding site, but not the C-terminal site, to recognize the Man-6-P methyl ester and mannose 6-sulfate modifications found on the lysosomal enzymes from Dictyostelium discoideum in addition to the phosphomonoester, Man-6-P (33).
Binding affinity analyses using the lysosomal enzyme, ␤-glucuronidase, were performed to characterize further the contribution to binding affinity made by each of the residues identified by pentamannosyl phosphate-agarose affinity chromatography to be essential for carbohydrate recognition by the IGF-II/MPR. In addition, the binding affinities of two additional mutants, Y368F (domain 3) and Y1264F (domain 9), which exhibited a partial FIG. 4. Binding of ␤-glucuronidase to wild-type and mutant Dom1-3His and Dom7-9His constructs. Increasing concentrations of iodinated ␤-glucuronidase were incubated with wild-type or mutant Dom1-3His or Dom7-9His, and the receptors and bound ligand were immunoprecipitated with bovine IGF-II/MPR-specific antiserum prebound to protein A-Sepharose beads. The beads were washed, and bound ␤-glucuronidase was specifically eluted from the receptors by incubation with 10 mM Man-6-P. The results were analyzed by nonlinear regression (SigmaPlot version 5.05, SPSS Science). A, wild-type Dom1-3His (q) and Q392N, S431A, E460D, and Y465F Dom1-3His mutants (E); B, Y368F Dom1-3His mutant (q); C, wild-type Dom7-9His (q) and H1329N, E1354D, and Y1360F Dom7-9His mutants (E); and D, Y1264F Dom7-9His mutant (q). Approximately equivalent amounts of wild-type and mutant receptors were used.  (54) using Lasergene (version 5, DNASTAR) sequence analysis software. The output was modified such that gaps and insertions were placed primarily in variable loop regions. Further refinements to the alignment were generated in accordance with results from previous site-directed mutagenesis studies of the CD-MPR (37) as well as from results of the current studies of the IGF-II/MPR. The secondary structure of the bovine CD-MPR (34) is shown above the sequence (wavy line represents the single ␣-helix, and arrows represent ␤-strands). The amino acid sequence identity between the carbohydrate recognition domains of the IGF-II/MPR and the CD-MPR from various species is indicated (gray shading). In addition, the two cysteine residues present in domains 3 and 9 of the IGF-II/MPR, but not in the CD-MPR, are indicated (gray shading). The 9 residues that compose the Man-6-P-binding pocket of the CD-MPR complexed with Man-6-P (34) are shown (q). Additional residues found in the binding pocket of the CD-MPR complexed with pentamannosyl phosphate (35) are also shown (E). The sequence positions of the human cancer-associated missense mutations, C1262S (47,50) and G1296R (52), are also indicated (*). The CD-MPR sequences, with their corresponding NCBI protein data base accession numbers, shown are as follows: (bov) bovine (A27068) (25); (hum) human (A32700) (55); (mo) mouse (A40399) (56); and (rat) rat (57). The IGF-II/MPR sequences, with their corresponding NCBI protein data base accession numbers, shown are as follows: inhibitory effect on Man-6-P binding, were assayed. Consistent with a partial inhibitory effect on Man-6-P recognition, the Y368F Dom1-3His (K d ϭ 31 Ϯ 10 nM) and Y1264F Dom7-9His (K d ϭ 87 Ϯ 26 nM) mutants displayed 62-and 174-fold, respectively, lower binding affinities for ␤-glucuronidase than the wildtype receptors (K d ϭ 0.5 Ϯ 0.1 nM) (Fig. 4). In contrast, with the exception of the Q1292N Dom7-9His mutant that was not secreted efficiently from COS-1 cells and therefore was not assayed, a Ͼ1200-fold lower affinity for ␤-glucuronidase was observed for each of the selected Dom1-3His (Q392N, S431A, E460D, and Y465F) and Dom7-9His (H1329N, E1354D, and Y1360F) mutants that displayed no binding to pentamannosyl phosphate affinity columns (Figs. 3 and 4). Previous quantitative studies had demonstrated a similar Ͼ1000-fold decrease in binding affinity for ␤-glucuronidase when Arg-435 in domain 3 or Arg-1334 in domain 9 was replaced with an alanine residue (33). The binding affinity data indicate that each of the residues identified to be essential for carbohydrate recognition contributes Ͼ3 kcal/ mol of energy to ligand binding, consistent with the prediction that each of these residues forms critical hydrogen bond and/or ionic interactions with Man-6-P.
The three-dimensional structure of the bovine CD-MPR bound to pentamannosyl phosphate shows that the binding site of the receptor encompasses the phosphate group and the terminal three mannose rings (35). Significantly, the vast majority (15 of the 21 potential hydrogen bond interactions) of the binding site contacts are made to the terminal Man-6-P moiety (35), consistent with inhibition studies that demonstrated that the presence of Man-6-P at a terminal position represents the major determinant of receptor binding (42,43). Three binding site amino acids (Asp-43, Tyr-45, and Gln-68) are responsible for providing the six remaining contacts, four of which are made to the penultimate mannose ring and two of which are made to the prepenultimate mannose ring of pentamannosyl phosphate (35).
In the CD-MPR, Asp-43 interacts with the 3-and 4-hydroxyls of the penultimate mannose ring (35). Interestingly, in our refined sequence alignment (Fig. 5) both of the IGF-II/MPR-binding sites have a glutamate residue at this position that might perform a similar carbohydrate recognition function. As discussed previously, both of the IGF-II/MPR-binding sites contain a tyrosine residue at the sequence position corresponding to Tyr-45 (Fig. 5), which interacts with the 1-hydroxyl of the terminal Man-6-P, which is involved in an O-glycosidic linkage, and the 4-hydroxyl of the penultimate mannose ring in the CD-MPR. However, none of these tyrosine residues is essential for carbohydrate recognition by their respective binding sites since replacement of each tyrosine with phenylalanine results in only a partial inhibitory effect on Man-6-P binding (Figs. 3 and 4) (37). Gln-68 was shown to provide contacts to the penultimate and prepenultimate mannose rings of pentamannosyl phosphate (35). However, it is unclear whether these interactions are maintained in the wild-type CD-MPR since the structure of the CD-MPR bound to pentamannosyl phosphate was obtained using a glycosylation-deficient form of the receptor in which Gln-68 replaced the asparagine residue found in the wild-type receptor (35). Furthermore, a dramatic difference is observed at that position in the two Man-6-P-binding sites of all of the IGF-II/MPRs sequenced to date; with the exception of the N-terminal binding site of the colugo, all have a lysine residue at the sequence position corresponding to Asn-68 in the wild-type CD-MPR (Fig. 5). Consequently, further studies will be needed to identify which, if any, IGF-II/MPR residues are involved in binding to sugars of the bound oligosaccharide other than the terminal Man-6-P moiety.
The identification of amino acids essential for Man-6-P recognition in domains 3 and 9 of the IGF-II/MPR is consistent with the observation that the best sequence alignment occurs between the CD-MPR and domains 3, 5, and 9 of the IGF-II/MPR (24). However, the N-and C-terminal Man-6-Pbinding sites of the IGF-II/MPR have each been localized to a minimum of three domains, 1-3 and 7-9, respectively ( Fig. 1) (29,30). Furthermore, various attempts to express a functional domain 3 or domain 9 alone have been unsuccessful, 3 leading to the prediction that residues in adjacent domains are important either directly for carbohydrate recognition or, more likely, indirectly for proper folding of the receptor to enable Man-6-P binding to occur (30). Importantly, each of the four residues identified as essential for carbohydrate recognition by the bovine CD-MPR (Gln-66, Arg-111, Glu-133, and Tyr-143) is found in domains 3 and 9 of all of the IGF-II/MPRs sequenced to date (Fig. 5). It is interesting to note that these four essential Man-6-P-binding site residues are also found in the sequence of domain 5 (24), although it is clear that domain 5 does not harbor a high affinity Man-6-Pbinding site in contrast to domains 3 and 9 (29,30,38). One likely explanation for the inability of domain 5 to bind Man-6-P with high affinity is the absence of two cysteine residues, corresponding to Cys-106 and Cys-141 in the CD-MPR sequence, necessary to form a critical disulfide bond that in the CD-MPR brings two loops (loop between ␤-strands 6 and 7 and loop between ␤-strands 8 and 9) together to form the Man-6-P-binding pocket (Fig. 5) (34, 35). Nevertheless, the possibility that domain 5 acts as a low affinity carbohydratebinding site has not been definitively ruled out.
Several lines of evidence support the hypothesis that the IGF-II/MPR acts as a growth inhibitor and that loss of IGF-II/ MPR function is associated with progression of tumorigenesis (44,45). In particular, the ability of the IGF-II/MPR to regulate targeting of lysosomal enzymes to the lysosome (1-3), facilitate 3 N. Dahms, unpublished data.
FIG. 6. Schematic representation of the potential interactions between Man-6-P and the binding pocket residues of the CD-MPR in comparison with predicted corresponding IGF-II/MPR domain 3 and domain 9 residues. The potential hydrogen bond interactions between Man-6-P (yellow) and the binding pocket residues of the CD-MPR (34) are shown. Directly beneath the CD-MPR residues (underlined) are listed the corresponding Man-6-P-binding site residues of domain 3 and domain 9 of the IGF-II/MPR, as predicted by the refined CD-MPR structure-based sequence alignment (Fig. 5). Blue, CD-MPR residues that have not been subjected to site-directed mutagenesis (Asn-104) or when mutated retained wild-type Man-6-P binding ability (Asp-103). Gray, single amino acid substitutions of CD-MPR residues or corresponding domain 3 or domain 9 residues that resulted in a partial inhibitory effect on Man-6-P recognition. Purple, single amino acid substitutions of CD-MPR residues or corresponding domain 3 or domain 9 residues that abolished Man-6-P binding. A water molecule (red sphere, H 2 O) that coordinates the manganese (Mn) divalent cation (purple sphere) is also shown. activation of the growth inhibitor transforming growth factor-␤ (4 -7), and modulate circulating levels of the potent cytokine leukemia inhibitory factor (15,16) due to its Man-6-P recognition function in addition to binding and targeting IGF-II for degradation (31,32,46) indicates that the IGF-II/MPR plays an important role in tumor suppression. To date, nine human cancer-associated missense mutations that occur in the extracytoplasmic region of the IGF-II/MPR have been identified as follows: two are located in Man-6-P recognition domain 9 (C1262S and G1296R); three are in domain 10 (Q1445H, G1449V, and G1464E); and four are in IGF-II binding domain 11 (G1564R, I1572T, A1618T, and G1619R) (47)(48)(49)(50)(51)(52). Significantly, with the exception of G1464E, all of the missense mutations (C1262S, Q1445H, G1449V, G1464E, and I1572T) that have been characterized to date result in receptors with altered Man-6-P and/or IGF-II binding properties (50,51). One of the domain 9 mutations, C1262S, results in altered Man-6-P and IGF-II binding functions (50). This effect is likely due to a conformational disruption of the IGF-II/MPR since, based on the CD-MPR, Cys-1262 (Cys-1271 in domain 9 of the bovine IGF-II/MPR, Fig. 5) is predicted to form a critical disulfide bond (24,34,53). The other C-terminal Man-6-P-binding site mutation, G1296R (Gly-1305 in domain 9 of the bovine IGF-II/MPR, Fig. 5), has not yet been characterized. However, based on the CD-MPR structure, G1296R is predicted to generate significant conformational changes by altering the interaction between two orthogonal ␤-sheets and consequently disrupt the normal ligand binding function of the IGF-II/MPR (34,52). Furthermore, by analogy to the loss of normal IGF-II/MPR ligand binding function observed with the cancer-associated missense mutations that have been characterized to date, our identification of several residues that are essential for carbohydrate recognition by the IGF-II/MPR implicates these residues as potential targets of carcinogenesis.
In summary, we report the identification of four amino acid residues (Gln-392, Ser-431, Glu-460, and Tyr-465) in domain 3 and four residues (Gln-1292, His-1329, Glu-1354, and Tyr-1360) in domain 9 that are essential for carbohydrate recognition by the IGF-II/MPR in addition to the previously determined Arg-435 (domain 3) and Arg-1334 (domain 9) residues (30). The results from these studies provide strong evidence that the two IGF-II/MPR Man-6-P-binding sites utilize a mechanism similar to that of the CD-MPR for high affinity Man-6-P binding and that the N-and C-terminal carbohydrate recognition domains of the IGF-II/MPR are structurally similar to each other and to the CD-MPR (Fig. 6). Future structural studies will be needed to verify the predictions made from these studies as well as to identify other residues and structural features necessary for high affinity carbohydrate recognition by the IGF-II/MPR.