The 46-kDa mannose 6-phosphate receptor does not depend on endosomal acidification for delivery of hydrolases to lysosomes

Lysosomal delivery of soluble acid hydrolases depends mostly on the presence of mannose 6-phosphate (M6P) in their carbohydrate moieties. These residues mediate binding to specific M6P receptors (von Figura and Hasilik, 1986; Kornfeld and Mellman, 1989). Two distinct M6P-binding proteins occur in mammalian cells (Kornfeld, 1992), the 300-kDa mannose 6-phosphate/insulin-like growth factor II receptor (M6P/IGF2R) and the 46-kDa mannose 6-phosphate receptor (MPR46). Binding of lysosomal enzymes to M6P receptors is thought to occur in the Golgi complex, followed by delivery of the complexes to the lysosomal pathway. Upon reaching endosomes, the ligands dissociate from their receptors which then return to the Golgi stacks. It has been proposed that this segregation process is triggered by the low luminal pH of endosomes because binding of lysosomal enzymes to M6P receptors in vitro is inefficient under acidic conditions (Distler et al., 1991). Treatment of cells with agents that raise the pH in endosomes and lysosomes has been found to reduce lysosomal enzyme sorting (Braulke et al., 1987). Indeed, this has led to the notion that pH-insensitive targeting of lysosomal enzymes is due to an M6P-independent transport pathway (Capony et al., 1994). Evidence for the existence of alternate sorting receptors has been reported (McIntyre and Erickson, 1993). However, isolation and identification of such a protein has remained elusive.

Mice deficient in MPR46 and/or M6P/IGF2R have been generated (Köster et al., 1993; Ludwig et al., 1993; Ludwig et al., 1994). Studies on MPR46- and/or M6P/IGF2R-negative fibroblasts have indicated that both receptors are necessary for efficient lysosomal targeting. It has been suggested that the two receptors complement each other by binding to distinct subpopulations of lysosomal enzymes (Pohlmann et al., 1995). However, sorting by M6P/IGF2R is generally far more efficient than by MPR46, demonstrating that the former is the main lysosomal targeting receptor in mammalian cells. This can be attributed, at least in part, to the lower affinity of MPR46 for multivalent M6P-containing ligands, compared with M6P/IGF2R (Tong and Kornfeld, 1989; Tong et al., 1989).

We have previously reported that in M6P/IGF2R-deficient murine SCC-VII squamous carcinoma cells, lysosomal enzyme trafficking is largely insensitive to the pH-perturbing agents NH₄Cl and chloroquine (Lorenzo et al., 2000). We now provide evidence that in these and other M6P/IGF2R-deficient murine cells, transport of acid hydrolases to lysosomes is dependent on MPR46, and show that this pathway is functional in the absence of endosomal acidification.

To investigate the role of N-linked oligosaccharides in the biosynthesis of cathepsin B in M6P/IGF2R-deficient cells, SCC-VII cells were metabolically labeled with [³⁵S]methionine in the absence or presence of tunicamycin, an inhibitor of protein N-glycosylation. In the absence of the drug, immunoprecipitation with antibodies against cathepsin B detected a 42-kDa form of the protein in the cell lysate at the start of the chase. This polypeptide corresponds to glycosylated procathepsin B, the latent precursor of the proteinase [Fig. 1A, proCB(3)]. After 4 hours of chase, the fraction retained inside the cells was completely processed to lower molecular mass bands (Fig. 1A, scCB), which represent different glycoforms of the mature proteinase (Lorenzo et al., 2000). However, 60% of the newly synthesized proenzyme was secreted into the medium as a 44-kDa polypeptide.

Upon tunicamycin treatment, procathepsin B was present as a 36-kDa polypeptide in cells and medium before the chase [Fig. 1A, proCB(0)]. This apparent molecular mass is identical to that of enzymatically deglycosylated procathepsin B (Lorenzo et al., 2000) and agrees well with the theoretical molecular mass of the non-glycosylated proenzyme (Chan et al., 1986). After a 4-hour chase, the tunicamycin-treated cells had secreted virtually all (>97%) labeled procathepsin B, with no mature enzyme detectable within the cells even after long exposure times (Fig. 1A). Since procathepsin B delivered to the lysosomal pathway is rapidly converted into the mature forms of the proteinase (Schmid et al., 1999), these results indicate that in SCC-VII cells, biosynthetic transport of cathepsin B to lysosomes strictly relies on the presence of a receptor binding to the carbohydrate moiety of the enzyme.

We have previously found that in SCC-VII cells lysosomal trafficking of cathepsin B and other acid hydrolases is largely insensitive to the pH-perturbing agents NH₄Cl and chloroquine (Lorenzo et al., 2000). The carboxylic ionophore monensin is also known to dissipate the transmembrane pH gradients in the Golgi and lysosomal compartments (Braulke et al., 1987). Secretion of newly synthesized procathepsin B by SCC-VII cells upon monensin treatment (40%) was only slightly higher than by control cultures (31% in this experiment). Thus, monensin cannot further impede the residual intracellular retention of lysosomal enzymes in these M6P/IGF2R-deficient cells (Fig. 1B). These results mimic the previously reported effects of NH₄Cl and chloroquine on biosynthetic cathepsin B transport in SCC-VII cells (Lorenzo et al., 2000) (Table 1). However, treatment with NH₄Cl or monensin abolished the staining of SCC-VII lysosomes with Lysosensor Yellow/Blue, a dye which specifically accumulates in acidic organelles. The same results were obtained for M6P/IGF2R-positive L-M(TK^-) fibroblasts. This indicates that NH₄Cl and monensin are effective in elevating the lysosomal pH in SCC-VII cells to a similar extent as in M6P/IGF2R-positive cells (Fig. 2).

The above results prompted us to study the effect of NH₄Cl on the intracellular transport of newly synthesized lysosomal enzymes in other murine cell lines either lacking M6P/IGF2R or expressing merely very low levels of the receptor. When the biosynthesis of cathepsin B in Igf2r^-/- murine embryonic fibroblasts (MEFs) was investigated, substantial amounts (55%) of newly synthesized procathepsin B were detected in the culture medium (Fig. 3A). Secretion of the cathepsin B precursor was only modestly stimulated by NH₄Cl (70%). Similar results were obtained for M6P/IGF2R-deficient RAW 264.7 macrophages. By contrast, the low level of procathepsin B secretion by M6P/IGF2R-positive cells was distinctly elevated in the presence of NH₄Cl (Table 1).

We also investigated the effect of NH₄Cl on the secretion and processing of other lysosomal proteinases in M6P/IGF2R-positive NIH/3T3 fibroblasts and Igf2r^-/- MEFs. NIH/3T3 cells secreted only limited amounts of procathepsin D (<1%) and procathepsin L (5%), which were distinctively enhanced by the addition of NH₄Cl (17% and 75%, respectively; not shown). By contrast, Igf2r^-/- MEFs deliver large fractions of their newly synthesized procathepsin D (50%) and L (61%) into the culture medium. Secretion of procathepsin D and L by these cells was only slightly stimulated by the presence of NH₄Cl (57% and 67%, respectively). However, proteolytic maturation of intracellular procathepsin L was dramatically impaired by the addition of the base (Fig. 3B), indicating that NH₄Cl is effective in elevating the lysosomal pH in Igf2r^-/- MEFs.

In line with the results obtained for cathepsin B, D and L, treatment of various M6P/IGF2R-deficient cell lines with NH₄Cl led at best to a slightly stimulated release of the classical lysosomal marker enzyme, β-N-acetylhexosaminidase. In fact, even the NH₄Cl-treated cultures secreted a substantially smaller fraction of their β-N-acetylhexosaminidase activity than MEFs lacking both M6P receptors (Table 2). These data support the hypothesis that pH-perturbing agents such as lysosomotropic amines and monensin generally interfere only weakly with the residual M6P-mediated lysosomal targeting of acid hydrolases in M6P/IGF2R-deficient cell lines.

The biosynthesis of cathepsin B, D and L was also studied in MEFs lacking both known M6P receptors, M6P/IGF2R and MPR46. In contrast to M6P/IGF2R-deficient MEFs (which contain normal amounts of MPR46), M6P receptor (MPR)-negative MEFs secrete more than 90% of their newly synthesized procathepsin B. The same observations were made for procathepsin D and L (Fig. 4A; Table 1). MPR-negative MEFs also secrete more than 90% of their β-N-acetylhexosaminidase activity (Table 2). For cathepsin D and β-N-acetylhexosaminidase, similar results have been reported previously (Pohlmann et al., 1995; Dittmer et al., 1999). These findings are in good agreement with the drastically reduced specific activity of β-N-acetylhexosaminidase (10%) in MPR-negative MEFs compared with normal murine fibroblasts (data not shown) (Pohlmann et al., 1995; Dittmer et al., 1999). Similar results were obtained for cathepsin B, D and L by immunoblot analysis, which also revealed that proteolytic maturation of cathepsin L and to a lesser extent cathepsin D is partially impaired in MPR-negative fibroblasts (Fig. 4B).

These results establish that biosynthetic transport of cathepsin B and L in murine fibroblasts relies entirely on the M6P receptor system, as previously observed for cathepsin D and β-N-acetylhexosaminidase (Pohlmann et al., 1995; Dittmer et al., 1999). This implies that in general the residual intracellular retention of lysosomal enzymes in M6P/IGF2R-deficient MEFs is mediated by the second M6P receptor, MPR46.

The synthesis of the M6P recognition marker is initiated by the cis-Golgi enzyme UDP-N-acetylglucosamine-1-phosphotransferase (von Figura and Hasilik, 1986; Tiede et al., 2005). However, generation of exposed M6P residues requires the subsequent action of the `uncovering enzyme' N-acetylglucosamine-1-phosphodiester α-N-acetylglucosaminidase - a protein that resides in the trans-Golgi network (TGN) (Rohrer and Kornfeld, 2001). We have previously shown that SCC-VII cells are capable of forming functional M6P residues on lysosomal enzymes (Lorenzo et al., 2000). To assess the effects of NH<sub>4</sub>Cl and monensin on the phosphorylation status of lysosomal enzymes, SCC-VII cells were metabolically labeled with [<sup>3</sup>H]mannose in the absence (control) or presence of the drugs. <sup>3</sup>H-labeled glycoproteins secreted into the culture medium were then isolated and treated with endo-β-N-acetylglucosaminidase H (Endo H) to release high-mannose and hybrid-type N-linked oligosaccharides. By means of anion-exchange chromatography, the control sample was found to consist of neutral (35%), sialylated (2%) and phosphorylated (63%) N-glycans (Table 3). Of the carbohydrate-associated phosphate groups synthesized by untreated SCC-VII cells, 92% were found to be phosphomonoesters which serve as high-affinity ligands for M6P receptors (Kornfeld and Mellman, 1989). Treatment with NH<sub>4</sub>Cl did not affect the sialylation and phosphorylation of Endo-H-sensitive N-linked oligosaccharides. However, the `uncovering' efficiency of phosphorylated N-glycans was moderately reduced (71%). This is in good agreement with data reported for human and murine fibroblasts (Isidoro et al., 1990; Dittmer and von Figura, 1999). Monensin displayed a more pronounced effect on the composition of Endo-H-sensitive N-glycans, because sialylated structures were not detected upon treatment with the drug. This is in accordance with previous findings (Dittmer and von Figura, 1999). By contrast, phosphorylation of N-glycans was not significantly affected by monensin treatment. However, monensin substantially reduced the `uncovering' of M6P residues, with only 37% of them present as phosphomonoester groups (Table 3). Similar observations have been made previously with murine fibroblasts (Dittmer and von Figura, 1999). These results demonstrate that NH<sub>4</sub>Cl and monensin do not impede the activity of UDP-N-acetylglucosamine-1-phosphotransferase, but inhibit the action of the `uncovering enzyme', N-acetylglucosamine-1-phosphodiester α-N-acetylglucosaminidase. In fact, the moderate reduction of the `uncovering' efficiency by NH₄Cl could account for the slight stimulatory effect of this compound on the secretion of cathepsin B (Table 1) and β-N-acetylhexosaminidase (Table 2) by SCC-VII cells and other M6P/IGF2R-deficient cell lines.

To assess the subcellular distribution of lysosomal enzymes in SCC-VII cells treated with NH₄Cl and monensin, post-nuclear organelles were separated by Percoll density-gradient centrifugation. In all gradients, the activity of the lysosomal marker β-N-acetylhexosaminidase was evenly distributed between heavy (lysosomal) and light (endosomal) fractions (Fig. 5). By contrast, the bulk of the activities of the Golgi enzyme galactosyltransferase and the ER marker NADH-cytochrome-c-reductase were located in the light gradient fractions (not shown). Immunoblot analysis revealed the presence of cathepsin D in heavy and light gradient fractions. Treatment of SCC-VII cells with monensin induced a subtle reduction of the amount of cathepsin D residing in lysosomes. This effect was not observed upon NH₄Cl treatment (Fig. 5). Similar results were observed for the effects of NH₄Cl and monensin on the subcellular distribution of cathepsin L (not shown).

In addition to endosomes, the light Percoll gradient fractions contain a range of other cellular compartments such as the Golgi complex. Hence, it could be envisioned that treatment with NH₄Cl or monensin causes a partial relocation of endosomal cathepsin D to the TGN. However, this possibility was ruled out by indirect immunofluorescence analysis of SCC-VII cells treated with these compounds. As in control cells, immunocytochemical detection of cathepsin D revealed numerous vesicles distributed throughout the cytoplasm, a staining pattern reminiscent of endosomes and lysosomes. This was verified by the extensive colocalization of cathepsin D and the endosomal/lysosomal marker LAMP-1 (Fig. 6). By contrast, the TGN marker TGN38 was found concentrated in the perinuclear region, as typically observed for Golgi-derived elements. Importantly, significant overlap of the subcellular distributions of cathepsin D and TGN38 was not detected either in controls or in NH₄Cl- or monensin-treated cells (Fig. 7A). These results demonstrate that treatment with either NH₄Cl or monensin does not result in a significant relocation of soluble endosomal and lysosomal hydrolases to the TGN and other regions of the Golgi apparatus.

It has been reported that in acidification-incompetent human cancer cells, M6P/IGF2R is mainly found in endosomes and lysosomes. Similarly, NH₄Cl treatment of acidification-competent human cancer cells led to a redistribution of Golgi-associated M6P/IGF2R to endosomal and lysosomal structures (Kokkonen et al., 2004). When untreated SCC-VII cells were double-stained with antibodies to MPR46 and LAMP-1, no co-localization of the two antigens was observed. Furthermore, neither NH₄Cl nor monensin led to significant colocalization of MPR46 and LAMP-1 in SCC-VII cells (Fig. 7B). As in untreated cells, MPR46 was found concentrated in perinuclear Golgi-like structures and extensively colocalized with the TGN marker TGN38 (data not shown). These findings indicate that endosomal dissociation of MPR46 and its ligands is not prevented by NH₄Cl and monensin.

We demonstrate in this report that the intracellular targeting of cathepsin B and other lysosomal proteinases in M6P/IGF2R-negative murine fibroblasts and SCC-VII cells relies on the presence of MPR46. Similar results have been reported for M6P/IGF2R-deficient rat Morris 7777 hepatoma cells, where treatment with blocking antibodies against MPR46 induced the secretion of the majority of the residual intracellular acid hydrolases (Stein et al., 1987). This supports the concept that at least in the cell types investigated, biosynthetic transport of acid hydrolases to lysosomes strictly requires the presence of the M6P recognition marker and the expression of at least one of the two M6P receptors.

M6P-dependent sorting of lysosomal hydrolases is thought to be initiated in the TGN. In this compartment, M6P-containing ligands first associate with M6P receptors before packaging into clathrin-coated vesicles for transport to late endosomes where the receptor-ligand complexes dissociate (von Figura and Hasilik, 1986; Kornfeld and Mellman, 1989). This process might be triggered by differences in the luminal pH of these compartments. The TGN displays a pH of 6.0-6.5, whereas late endosomes have a pH of 5.0-5.5 (Mellman et al., 1986; van Weert et al., 1995; Miesenböck et al., 1998). Binding of ligands to immobilized M6P receptors is strongly reduced at pH 5.4 compared with the optimal pH range for this interaction (pH 6.2-6.6) (Distler et al., 1991; Pohlmann et al., 1995). Hence, carboxylic ionophores and lysosomotropic amines that increase the luminal pH of lysosomes to pH 6.0-6.3 (Yoshimori et al., 1991; Moriyama et al., 1992) should prevent the displacement of the enzymes from their receptors and thus impede lysosomal hydrolase sorting. We now show that such compounds fail to prevent the MPR46-dependent residual lysosomal targeting of M6P-modified proteinases in M6P/IGF2R-deficient cells. This suggests that the pH-gradient in the endosomal/lysosomal system is not solely responsible for dissociation of MPR46 and its ligands in situ. Other factors, such as the structural properties of the receptor or intrinsic features of the endosomal/lysosomal system, are probably also involved in this process.

One such factor might be the oligomerization status of MPR46 as oligomeric forms of the receptor seem to display a higher affinity to M6P-containing ligands than dimeric MPR46. Although the receptor exists mostly as a homodimer, substantial amounts of tetramers can be detected in purified MPR46 preparations (Waheed and von Figura, 1990). In vitro, tetramer formation is favored, amongst other factors, by a high receptor concentration and the presence of Mn²⁺ ions (Waheed et al., 1990; Sun et al., 2005). Because MPR46 is mainly localized in the TGN (Klumperman et al., 1993) and thus occurs there at a higher local concentration than in endosomal compartments, it could be that TGN-derived MPR46 tetramers spontaneously dissociate upon arrival in endosomes and thus release bound ligands. A similar scenario could be envisaged for Mn²⁺-driven differences between the quaternary structures of MPR46 in the TGN and endosomes. However, no major changes in the tetramer/dimer ratio of MPR46 could be detected during intracellular recycling and ligand binding (Punnonen et al., 1996).

Another potential means to trigger pH-insensitive receptor-ligand dissociation could be based on the intrinsic ligand-binding properties of MPR46. MPR46 binds multivalent M6P-containing ligands less avidly than M6P/IGF2R (Tong and Kornfeld, 1989; Tong et al., 1989; Distler et al., 1991). Thus, the secretion of substantial amounts of lysosomal hydrolases by M6P/IGF2R-deficient cells could be due to low-affinity interactions of the enzymes with MPR46. Equilibrium dialysis experiments have revealed that MPR46 binds to bovine testis β-galactosidase and a high-mannose oligosaccharide carrying two phosphomonoesters only with a K_d of 2×10^-7 M (Tong and Kornfeld, 1989; Distler et al., 1991). Lysosomal hydrolases carrying one or two phosphorylated oligomannosidic N-glycans such as procathepsin B (Lorenzo et al., 2000) should display a similar affinity to MPR46. The concentration of procathepsin B in the TGN can be estimated to reach 1×10^-6 M, assuming a cellular synthesis rate of 1×10^-19 mol procathepsin B per hour (Coulibaly et al., 1999), a cellular volume of 1×10^-12 liter of which 10% is accounted for by the compartments involved in secretion (Griffiths et al., 1989), and a transition time of 1 hour in the secretory pathway (Hanewinkel et al., 1987). This concentration would be sufficient to permit binding to MPR46 and thus enforce segregation to the lysosomal pathway. Once delivered to endosomes, ligand displacement could be initiated by dilution of the complexes, thus enforcing enhanced ligand dissociation. Subsequent dephosphorylation because of the action of endosomal phosphatases would prevent reassociation of ligands and MPR46. It has been shown that M6P-modified acid hydrolases can be quickly dephosphorylated upon delivery to late endosomes (Einstein and Gabel, 1991). Thus, it is possible that at least three factors contribute to MPR46-ligand dissociation in the absence of a pH-gradient: the oligomerization status of the receptor, differences in the concentration of M6P-modified lysosomal hydrolases in situ, and the presence of phosphatases in the recipient compartments.

Tran[³⁵S]label metabolic labeling reagent (>1000 Ci/mmol) was obtained from ICN Pharmaceuticals (Costa Mesa, CA). [³H]mannose (10-20 Ci/mmol), UDP-[³H]galactose (5-20 Ci/mmol), ¹⁴C-methylated molecular mass standards, Amplify™ fluorographic reagent, Percoll and protein-A-Sepharose 4B were provided by Amersham Biosciences (Rainham, UK). Hygromycin B, NADH (disodium salt) and tunicamycin were from Roche Diagnostics (Mannheim, Germany). Prestained molecular mass standards and enhanced chemiluminescence western blotting reagents were obtained from Bio-Rad (Richmond, CA). E-64, leupeptin, 4-methylumbelliferyl-β-N-acetylglucosaminide, monensin, phenylmethylsulfonyl fluoride (PMSF) and sodium β-glycerophosphate were purchased from Sigma (St Louis, MO). All other chemicals were of reagent grade.

A rabbit antiserum against human cathepsin B (Mach et al., 1992) has been shown to crossreact with all forms of the murine enzyme (Coulibaly et al., 1999). The other rabbit antisera used in this study were: anti-mouse cathepsin D [(Pohlmann et al., 1995) or John S. Mort, Shriners Hospital for Children, Montreal, Canada]; anti-mouse procathepsin L (Ann H. Erickson, University of North Carolina, Durham, NC); anti-bovine M6P/IGF2R (Bernard Hoflack, Technische Universität Dresden, Germany); anti-mouse MPR46 (Annette Hille-Rehfeld, Georg-August Universität Göttingen, Germany). The rat monoclonal antibody to mouse LAMP-1 (clone 1D4B) developed by J. Thomas August (Johns Hopkins University, Baltimore, MD) was obtained as culture supernatant from the Developmental Studies Hybridoma Bank maintained by the Department of Biological Sciences (University of Iowa, Iowa City, IA). Affinity-purified sheep antibodies to rat TGN38 were from Serotec (Oxford, UK).

Murine SCC-VII squamous carcinoma cells (Coulibaly et al., 1999) were propagated in Minimal Essential Medium supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin. The following cell lines were maintained in Dulbecco's Modified Eagle's medium with the same supplements: murine RAW 264.7 macrophages and L-M(TK^-) fibroblasts (American Type Culture Collection, Manassas, VA); M6P/IGF2R-deficient murine L cells (clone D9; Jack Rohrer, Universität Zürich, Switzerland); immortalized Igf2r^-/- mouse embryonic fibroblasts (Erwin Wagner, Research Institute of Molecular Pathology, Vienna, Austria); Simian virus 40 (SV40)-transformed mouse embryonic fibroblasts lacking M6P/IGF2R and MPR46 (Kasper et al., 1996); SV40-transformed mouse embryonic fibroblasts lacking cathepsin B, D or L (Christoph Peters, Albert-Ludwigs-Universität, Freiburg, Germany). Murine NIH/3T3 (American Type Culture Collection) and 3T3-L1 fibroblasts (David James, University of Queensland, Brisbane, Australia) were cultured in Dulbecco's Modified Eagle's medium supplemented with 10% bovine calf serum, 2 mM glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin. Murine P388D₁ macrophages (American Type Culture Collection) were grown in RPMI 1640 medium containing 10% fetal bovine serum, 2 mM glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin. All tissue culture reagents were purchased from Invitrogen (Carlsbad, CA).

All molecular biology techniques were performed following standard procedures. A cDNA encoding the complete human M6P/IGF2R open reading frame (American Type Culture Collection) was subcloned into the plasmid pCDM8 (Invitrogen). This construct (obtained from Vladimir Leksa and Hannes Stockinger, Institute of Immunology, Medical University of Vienna, Austria) was digested with the restriction enzyme XbaI (Fermentas, St Leon-Rot, Germany) to excise its 8-kb insert. T4 DNA ligase (Fermentas) was used to ligate this fragment into the XbaI site within the multiple cloning region of pAHygCMV2 (Coulibaly et al., 1999). SCC-VII cells were transfected with the resulting vector using Lipofectin (Invitrogen) according to the manufacturer's instructions. Selection of stably transfected cells was achieved by virtue of their ability to grow in the presence of 350 μg/ml hygromycin B. Drug-resistant clones were isolated after 14 days and tested for M6P/IGF2R production by immunofluorescence analysis. One of the positive clones (clone 15) was selected for further experiments.

Confluent cell monolayers were metabolically labeled for 1 hour with [³⁵S]methionine (100 μCi/ml), then chased for 4-5 hours as described (Hanewinkel et al., 1987; Lorenzo et al., 2000). When added, the following compounds were present throughout the entire experimental procedure: NH₄Cl (10 mM), monensin (1 μM) and tunicamycin (10 μg/ml). In the case of monensin and tunicamycin, the control cultures were treated with the solvent (dimethyl sulfoxide) at a final concentration of 1% (v/v).

Immunoprecipitation of antigens from equivalent amounts of labeled cell and medium extracts followed by SDS-PAGE and fluorography was performed as reported (Hanewinkel et al., 1987; Mach et al., 1992). Bands nonspecifically precipitated by the antisera used were identified in control experiments using antigen-deficient mouse embryonic fibroblasts. The quantity of individual labeled polypeptides was estimated by densitometric analysis of exposed films using ImageQuaNT v4.2 software (Molecular Dynamics, Sunnyvale, CA). The raw data were corrected for the number of methionine residues present in each polypeptide (see http://merops.sanger.ac.uk/ for the sequences of mouse procathepsin B, D and L and their processing products). The fraction of secreted proenzyme is expressed as percentage of the sum of intracellular and secreted antigen present.

Confluent cell monolayers (∼10⁷ cells) were incubated for 24 hours at 37°C in 10 ml of the respective culture medium containing 10% fetal bovine serum heat-inactivated for 30 minutes at 70°C (to inactivate any β-N-acetylhexosaminidase activity) and 10 mM NH₄Cl. Control cultures were incubated without addition of NH₄Cl. The supernatants were then aspirated and cleared by centrifugation (5 minutes, 320 g) and subsequent passage through 0.22-μm filters. The cell monolayers were scraped into 500 μl of 20 mM sodium acetate buffer (pH 5.5) containing 150 mM NaCl and ultrasonicated, followed by addition of 0.1% Triton X-100 and incubation for 30 minutes at 0°C. After centrifugation (10 minutes, 15,000 g), the β-N-acetylhexosaminidase activity of cell extracts and conditioned media was determined spectrofluorimetrically with 4-methylumbelliferyl-β-N-acetylglucosaminide as substrate (Storrie and Madden, 1990). The fraction of secreted β-N-acetylhexosaminidase is expressed as percentage of the sum of intracellular and secreted enzyme present.

For microsomal extracts, confluent cell monolayers (∼10⁷ cells) were scraped into 1 ml of 3 mM imidazole buffer (pH 7.4) containing 0.25 M sucrose, gently lysed with a Dounce homogenizer and then centrifuged (5 minutes at 400 g). Microsomes were then pelleted (60 minutes at 105,000 g) and extracted for 30 minutes at 0°C with 500 μl of 0.1 M sodium phosphate buffer (pH 6.0) containing 2 mM EDTA and 0.1% Triton X-100. After centrifugation (10 minutes at 15,000 g), the supernatants were subjected to immunoblot analysis.

SCC-VII cells were incubated in complete culture medium in the absence or presence of either 10 mM NH₄Cl or 1 μM monensin for 10 hours at 37°C. Post-nuclear supernatants were obtained and fractionated by density-gradient centrifugation in 18% (v/v) Percoll gradients (initial density: 1.055 g/ml) as described (Schmid et al., 1999). The gradients were divided into ten fractions that were then analyzed for marker enzyme activities. The activity of the lysosomal marker β-N-acetylhexosaminidase was determined spectrofluorimetrically with 4-methyl-umbelliferyl-β-N-acetylglucosaminide (Storrie and Madden, 1990). The activity of the Golgi enzyme galactosyltransferase was measured with UDP-[³H]galactose and chicken ovalbumin (Sigma) as described by Rome et al. (Rome et al., 1979) with minor modifications. The distribution of the endoplasmic reticulum (ER) marker NADH-cytochrome-c-reductase was determined in 75 mM potassium phosphate buffer (pH 7.5) containing 1 mM potassium cyanide, using 0.1 mg/ml NADH (disodium salt) and 20 μM cytochrome c (Sigma) as substrates. The increase of the optical density at 550 nm was then continuously monitored at room temperature. The buoyant density of each gradient fraction was determined either gravimetrically (Schmid et al., 1999) or with a refractometer. For detection of cathepsins D and L, fractions 1-3 (heavy fraction), 4-7 (intermediate fraction) and 8-10 (light fraction) were pooled and treated with 0.1% Triton X-100 for 30 minutes at 0°C. Residual Percoll was then removed by ultracentrifugation for 2 hours at 105,000 g. Proteins were precipitated with 10% (w/v) trichloroacetic acid and 0.04% (w/v) sodium deoxycholate (30 minutes, 0°C) and recovered by centrifugation (10 minutes at 15,000 g). The pellets were washed with ethanol and acetone, air-dried, dissolved in SDS-PAGE sample buffer and then subjected to SDS-PAGE and immunoblot analysis.

SDS-PAGE was performed in 12.5% gels run under reducing conditions. Separated polypeptides were electrophoretically transferred onto a nitrocellulose membrane (Hybond-C, Amersham Biosciences) as reported (Mach et al., 1992). After probing the membrane with the respective primary rabbit antibodies, bound immunoglobulins were visualized with goat anti-rabbit IgG immunoglobulins conjugated to horseradish peroxidase (Jackson ImmunoResearch, West Grove, PA) and chemiluminescence-based detection reagents (Coulibaly et al., 1999). Densitometric analysis of immunoblots was done using ImageQuaNT v4.2 software.

Confluent cell monolayers (∼10⁷ cells) were metabolically labeled with [³H]mannose (1 mCi/ml) in 5 ml glucose-poor culture medium (1 mM glucose) for 8 hours at 37°C in the absence (control) or presence of 10 mM NH₄Cl or 1 μM monensin. Medium proteins were precipitated with 0.5 g/ml ammonium sulfate, redissolved in 2 ml 20 mM sodium acetate buffer (pH 5.5) containing 5 mM sodium β-glycerophosphate, and dialyzed against the same buffer. The retentate was concentrated by ultrafiltration (10 kDa cut-off) and incubated with 5 mU endo-β-N-acetylglucosaminidase H (Roche) as described (Dong and Sahagian, 1990). ³H-labeled oligosaccharides thus released were then isolated by ultrafiltration as above. The ultrafiltrate was sequentially passed through a minicolumn filled with Dowex 50×8 cation-exchange resin (Sigma) and a Sep-Pak C₁₈ reverse-phase cartridge (Waters, Milford, MA), both eluted with water. The unbound fraction was lyophilized and then desalted on a 1.5×7 cm column of Sephadex G-10 (Amersham Biosciences) eluted with water, and finally fractionated on diethyl-(2-hydroxypropyl)-aminoethyl-Sephadex (Sigma) according to Dong and Sahagian (Dong and Sahagian, 1990). To remove phosphomonoester groups, ³H-labeled oligosaccharides were treated with 10 U alkaline phosphatase from calf intestine (Sigma) in 0.1 M Tris-HCl buffer, pH 8.0, for 1 hour at 37°C. Phosphodiester linkages were cleaved by incubation in 2 M acetic acid for 2 hours at 80°C. Sialic acid residues were removed by incubation with 5 mU Arthrobacter ureafaciens sialidase (Roche) in 0.1 M sodium acetate buffer (pH 5.5) for 16 hours at 37°C.

SCC-VII cells grown on glass coverslips were treated in the absence or presence of either 10 mM NH₄Cl or 1 μM monensin in complete culture medium for 10 hours at 37°C. The cells were then fixed for 10 minutes in 4% paraformaldehyde dissolved in phosphate-buffered saline (PBS). After blocking with PBS containing 2 mg/ml bovine serum albumin for 1 hour, the cells were incubated for 1 hour with the following antibodies diluted with PBS containing 0.1% saponin: rabbit anti-mouse cathepsin D (1:100), rabbit anti-bovine M6P/IGF2R (1:100), rabbit anti-mouse MPR46 (1:100), rat anti-mouse LAMP-1 (1:100), sheep anti-rat TGN38 (1 μg/ml). After a second blocking step in PBS containing 0.1% saponin and 5% fetal bovine serum (1 hour), bound primary antibodies were detected by incubation for 1 hour with the respective FITC- or Cy3-conjugated affinity-purified secondary antibodies (Jackson ImmunoResearch) at 5 μg/ml in PBS containing 0.1% saponin. All steps were performed at room temperature. The immunostained cells were examined using a Leica TCS SP2 confocal laser-scanning microscope equipped with diode, Ar and He/Ne lasers. Images from the confocal system were imported into Adobe Photoshop 5.5 for coloration.

SCC-VII cells and L-M(TK^-) fibroblasts grown on glass coverslips were treated in the absence or presence of either 10 mM NH₄Cl or 1 μM monensin in complete culture medium for 2 hours at 37°C. The cells were quickly rinsed and then incubated in PBS (supplemented with NH₄Cl and monensin where appropriate) containing 0.5 μg/ml 4′,6-diamidino-2-phenylindole (DAPI; Sigma) for 10 minutes at 37°C before addition of 5 μM LysoSensor Yellow/Blue DND-160 (Molecular Probes, Eugene, OR). Cells were incubated for an additional 5 minutes at 37°C, then briefly rinsed with PBS and immediately observed by confocal laser-scanning microscopy (see above). Upon excitation at 405 nm, fluorescence emitted in the range of 420-500 nm (DAPI) and 500-580 nm (LysoSensor) was recorded simultaneously. The acquired images were then imported into Adobe Photoshop 5.5 for overlay and pseudocoloration.

Total protein was determined by the Bradford method using the Bio-Rad Protein Assay Kit (Bio-Rad, Richmond, CA), using bovine serum albumin (Sigma) as a standard.

The expert technical support of Martina Schwarz and Jason L. Clark is gratefully acknowledged. We express our gratitude to Ann Erickson, Bernard Hoflack, David James, Vladimir Leksa, John Mort, Christoph Peters, Jack Rohrer, Hannes Stockinger and Erwin Wagner for reagents and cell lines. We also thank Melinda Abas for critical reading of the manuscript and Markus Auly for his assistance in some experiments. This study was supported by the Australian Research Council (grants 04/15/412/170, 04/15/412/284 and 04/15/412/343), the Jubiläumsfonds der Stadt Wien für die Österreichische Akademie der Wissenschaften (project J-7/2005) and the Austrian Science Fund (project P16925), all awarded to L.M.