Cytosol-to-lysosome Transport of Free Polymannose-type Oligosaccharides

In hepatocellular carcinoma HepG2 cells, free polymannose-type oligosaccharides appearing in the cytosol during the biosynthesis and quality control of glycoproteins are rapidly translocated into lysosomes by an as yet poorly defined process (Saint-Pol, A., Bauvy, C., Codogno, P., and Moore, S. E. H. (1997)J. Cell Biol. 136, 45–59). Here, we demonstrate an ATP-dependent association of [2-3H]mannose-labeled Man5GlcNAc with isolated rat liver lysosomes. This association was only observed in the presence of swainsonine, a mannosidase inhibitor, which was required for the protection of sedimentable, but not nonsedimentable, Man5GlcNAc from degradation, indicating that oligosaccharides were transported into lysosomes. Saturable high affinity transport (K uptake, 22.3 μm, V max, 7.1 fmol/min/unit of β-hexosaminidase) was dependent upon the hydrolysis of ATP but independent of vacuolar H+/ATPase activity. Transport was inhibited strongly by NEM and weakly by vanadate but not by sodium azide, and, in addition, the sugar transport inhibitors phloretin, phloridzin, and cytochalasin B were without effect on transport. Oligosaccharide import did not show absolute specificity but was selective toward partially demannosylated and dephosphorylated oligosaccharides, and, furthermore, inhibition studies revealed that the free reducing GlcNAc residue of the oligosaccharide was of critical importance for its interaction with the transporter. These results demonstrate the presence of a novel lysosomal free oligosaccharide transporter that must work in concert with cytosolic hydrolases in order to clear the cytosol of endoplasmic reticulum-generated free oligosaccharides.

and the glycosylation of glycoproteins (3,4). Whereas phosphorylated free OS (Man 5 GlcNAc 2 -P) are released from dolichol oligosaccharides on the cytoplasmic face of the ER (5), neutral free OS (Glc 3-1 Man 9 -8 GlcNAc 2 ) are liberated from oligosaccharide-lipid in the lumen of the ER (1,6). After deglucosylation, these neutral free OS do not follow the secretory pathway but are routed to the cytosol (7) by an ATP-and calcium-requiring translocation process that can be inhibited by mannose but not by N-acetylglucosamine (8). Glycopepetides (9,10) and misfolded glycoproteins (11,12) are also translocated out of the ER into the cytosol, where they can be deglycosylated by an Nglycanase (13,14), giving rise to further cytosolic free OS.
The rapid appearance of substantial quantities of free OS in the cytosol of cells during the biosynthesis and quality control of glycoproteins suggested that the cytosol must possess elements capable of the disposal of these osmotically active components. In fact, neutral free OS are subject to sequential processing (15)(16)(17) by a previously characterized cytosolic endo-H-like activity (18) or chitobiase (15) and cytosolic mannosidase (19 -21) to yield a free OS containing five residues of mannose and a single reducing N-acetylglucosamine residue. It has now been demonstrated that this free OS is transferred to lysosomes in order to be degraded (17). The fate of cytosolic phosphorylated free OS is less clear, but since this oligosaccharide possesses an identical configuration of mannose residues to that of the free OS that undergoes cytosol-to-lysosome translocation, it may be transferred directly into the lysosome. The physiological importance of the capture of cytosolic free OS by the lysosome is indicated by the fact that up to 40% of the free OS recovered from the urine of people with the lysosomal storage disease, ␣-mannosidosis, have oligosaccharide structures compatible with their having been generated in the cytosol (22,23). However, despite the fact that the cytosol-tolysosome translocation of free OS must represent a vital housekeeping activity in cells engaged in the biosynthesis of glycoproteins, this process remains to be characterized.
Although most molecules destined for degradation in the lysosome arrive by way of a series of vesicular fusion steps after their initial capture, either from the extracellular environment into endocytic vesicles (24) or from within the cell into autophagic vacuoles (25), it has become apparent that the lysosomal membrane itself may be involved in the capture of cytosolic molecules. For instance, cytosolic proteins containing the KFERQ sequence can be translocated into lysosomes by a lysosomally situated translocation machinery comprising both HSC73 and LGP96 molecules (26). In addition, lysosomes may contain specific carriers responsible for the import of cysteine (27) and methotrexate glutamates (28) from the cytosol. In vivo, the cytosol-to-lysosome transfer of free OS was inhibited by agents that reduce cellular ATP levels and by high swainsonine concentrations, which also inhibited the cytosolic demannosy-* This work was supported by institutional funding from INSERM and by a grant from the Association Vaincre les Maladies Lysosomales. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Marie Curie Fellow of the European Community. To whom correspondence should be addressed: INSERM U410, 16 rue Henri Huchard 75018 Paris, France. Tel.: 33 1 4485 6134; Fax: 33 1 4228 8765; E-mail: moore@bichat.inserm.fr. 1 The abbreviations used are: free OS, free oligosaccharide(s); lation of free OS (17). However, apart from being able to rule out macroautophagic sequestration as a possible mechanism to account for this phenomenon, little could be inferred about this process.
Here, we demonstrate an ATP-dependent, N-ethylmaleimide-sensitive, saturable import of free OS into isolated rat liver lysosomes. Results are reported to show that the lysosome possesses a novel oligosaccharide importer capable of capturing an array of small polymannose oligosaccharides that possess a reducing GlcNAc residue.
Preparation and Purification of Free Oligosaccharides from Metabolically Radiolabeled Lymphoma Cells-Metabolically radiolabeled neutral free oligosaccharides were prepared from radiolabeled cells as described previously (16), and the resulting [ 3 H]free OS was resolved on plastic thin layer chromatography plates coated with cellulose (Merck), which were developed in pyridine/ethyl acetate/water/acetic acid (5:3: 3:1) for either 3 days or for a single day if monosaccharides were to be examined. Resolved components were visualized by autoradiography prior to their elution from the chromatography plates with water and passed over coupled columns of AG 50 (H ϩ form) and AG 1 (acetate form). Nonretained neutral components were loaded onto charcoal columns as described previously (16), and after washing, free oligosaccharides were eluted from the columns with 30% ethanol. Free oligosaccharide phosphates (2,30) were isolated from the upper phase of chloroform/methanol/100 mM Tris-HCl, pH 7.4, containing 4 mM MgCl 2 (3:2:1) extracts of Thy Ϫ1 cells. Material was dried down and passed over coupled columns of AG 50 over AG 1, and, whereas neutral free oligosaccharides were recovered from the run-through and water washes, oligosaccharide phosphates were eluted from the AG 1 with 2 M NaCl. After desalting this material by Biogel P2 chromatography, it was subjected to thin layer chromatography on silica-coated plates that were developed overnight in 1-propanol/acetic acid/water (3:3:2). Four components were resolved, and after digestion with alkaline phosphatase it was found that the slowest migrating component comigrated with Man 5 GlcNAc 2 .
Structural Analysis of Free Oligosaccharides-Free oligosaccharides were treated with 1 unit of jack bean ␣-mannosidase (Sigma) overnight at 37°C in 40 mM sodium acetate, pH 4.5, or digested overnight at 37°C with 5 microunits of ␣-1,2-mannosidase (Oxford Glycosystems) in 100 mM sodium acetate buffer, pH 5.0. The digestion products were then desalted as described above, concentrated, and resolved by thin layer chromatography. 5 GlcNAc-Large free oligosaccharides ([ 3 H]Man 9 -8 GlcNAc 2 ) that were generated in permeabilized HepG2 cells and recovered from the membrane-bound compartments (7) were treated with endoglycosidase H to yield [ 3 H]Man 9 -8 GlcNAc. These oligosaccharides were then treated with the ␣-1,2-mannosidase (see above) to yield the branched isomer of [ 3 H]Man 5 GlcNAc.
Preparation of Rat Liver Lysosomes-Male Wistar rats were sacrificed, and livers were cut into small pieces before being homogenized in 5 volumes of homogenization buffer (250 mM sucrose, 1 mM EDTA, 20 mM HEPES/KOH, pH 7.3) using a tight fitting Dounce apparatus. After centrifuging the homogenate at 600 ϫ g for 10 min, the supernatant was removed and recentrifuged at 18,000 ϫ g av . The resulting mitochondrial/lysosomal pellet was resuspended in homogenization buffer, 5 ml of this suspension was mixed with 3 ml of an 80% Percoll solution (32), and the gradient was subsequently formed by centrifugation for 35 min at 92,570 ϫ g av (32). Material recovered from the bottom third to quarter of the gradient was washed twice with HEPES-buffered sucrose (250 mM sucrose, 20 mM HEPES/KOH, pH 7.3). The final lysosomal pellet was resuspended in 1-1.5 ml of HEPES-buffered sucrose.
The Transport Assay-Radioactive substrates were dried down into Eppendorf tubes, and the following additions were made, on ice, in the indicated order: 5 l of water (or, where indicated, test substances); 5 l of ATP and an ATP-regenerating system (50 mM ATP, 0.5 mg/ml creatine kinase, 70 mM creatine phospate); 20 l of incubation buffer (350 mM sucrose, 125 mM KCl, 25 mM MgCl 2 , 20 mM HEPES/KOH, pH 7.3). After mixing, 20 l of lysosomes were added to make a final volume of 50 l. Subsequent to further mixing, the tubes were incubated on ice for 10 min prior to being transferred to a water bath at 25°C. Incubations were terminated by the addition of 1 ml of ice-cold SH, and lysosomes were recovered by centrifugation at 11,600 ϫ g av for 15 min at 4°C. The resulting supernatant was removed, made 1% with respect to Triton X-100, and kept for ␤-hexosaminidase determinations, and the lysosomal pellet was washed with a further 1 ml of SH prior to being solubilized in 1 ml of SH containing 1% Triton X-100. Radioactivity associated with the lysosomal lysate was determined by scintillation counting. In experiments where [ 3 H]free OS transported into lysosomes was to be analyzed, the transport assays were terminated by the addition of 1 ml of ice-cold MH (250 mM mannitol, 20 mM HEPES/KOH, pH 7.3). Lysosomes were also washed, as described above, with MH.
Measurement of Lysosomal pH-Lysosomal pH was measured as described previously (33). Briefly, rat liver lysosomes were loaded with FITC-dextran in vivo prior to purification as described above. Fluorescence measurements were made by incubating approximately 1 mg of lysosomes under conditions identical to those used for the transport assay (see above).
Protein, Sugar, and Enzymatic Assays-Lysosomal ␤-D-hexosaminidase activity was measured using p-nitrophenyl N-acetylglucosamine as described (34). Protein was measured by the bicinchoninic acid method using a kit purchased from Sigma.

Preparation of Radiolabeled Substrates Suitable for Examining the Cytosol-to-lysosome Transport of Free
Oligosaccharides-In order to develop an in vitro assay for the lysosomal transport of free oligosaccharides, we first sought an abundant source of radiolabeled linear Man 5 GlcNAc, the component that has been demonstrated to be the major free OS that accumulates in the lysosomes of HepG2 cells treated with the vacuolar ATPase inhibitor CCM A (17). Since the isomeric configuration of the mannose residues in this free OS (17) is known to be the same as that occurring in the Man 5 GlcNAc 2 intermediate that appears during the biosynthesis of dolichol oligosaccharides, we chose to harvest free OS from the mutant mouse lymphoma cell line Thy Ϫ1 (35), which, lacking functional dolichol-P-Man synthase, is unable to further mannosylate this oligosaccharide-lipid intermediate (36,37). Thin layer chromatographic analysis of free OS isolated from Thy Ϫ1 cells after metabolic radiolabeling with [2-3 H]mannose for 4 h, shown in Fig. 1A, reveals four oligosaccharides (free OSa-d), the largest of which comigrates with authentic Man 5 GlcNAc. Since it is known that free OS isolated from mammalian cells possess either a di-Nacetylchitobiose moiety or a single residue of N-acetylglucosamine at their reducing termini (1, 4), we wanted to verify that the free OS comigrating with Man 5 GlcNAc was not Man 4 GlcNAc 2 or a mixture of these two components. This issue was addressed by digesting free OSa with jack bean ␣-mannosidase, and Fig. 1A (lane 6) shows that this treatment yielded mannose and the disaccharide Man␤1-4GlcNAc, confirming the parent structure's identity as Man 5 GlcNAc and not Man 4 GlcNAc 2 . Furthermore Fig. 1A (lane 7) reveals that free OSa was sensitive to an ␣-1,2-mannosidase, yielding mannose and a structure that comigrated Man 3 GlcNAc, confirming that free OSa contains two ␣-1,2-linked mannose residues. Taking into account these results and what is known of the dolichol oligosaccharide biosynthetic pathway in Thy Ϫ1 cells (36,37), free OSa possesses the structure shown in Fig. 1B. Similar structural analyses (results not shown) of free OSb-d indicated that these three oligosaccharides are likely to have the structures indicated in Fig. 1B  containing the lysosomal marker enzyme ␤-hexosaminidase. This association was abolished when incubations were quenched with 200 M unlabeled Man 5 GlcNAc. These results, although suggesting a specific association of [ 3 H]free OS with lysosomal membranes, indicated that because of the high nonspecific association of radioactive components with permeabilized cells another assay system would be required.
[  30 and 25°C, and as can be seen, incubations performed at the lower temperatures greatly increased the stability of lysosomes over the longer incubation periods. Although these reduced incubation temperatures lead to slower initial rates of association of [ 3 H]free OS with membranes, this association is more sustained and reaches higher maximal levels at the two reduced temperatures. These results demonstrate a time-, temperature-, and ATP-dependent association of [ 3 H]free OS, but not [ 3 H]inulin, with the membranes of a lysosome-enriched particulate fraction. Because this association was ATP-and temperature-dependent and appeared to correlate with the intactness of lysosomes in the lysosomal preparation, it seemed possible that [ 3 H]free OS were being transported into lysosomes.
[ 3 H]Free OS Transported into Lysosomes-If free OS are being transported into functional lysosomes, they should be subject to complete catabolism by lysosomal ␣and ␤-mannosidases, yielding mannose and N-acetylglucosamine. Since both of these monosaccharides can be transported out of lysosomes (38,39), the observed transport of free OS into lysosomes should be greatly increased in the presence of SW, a potent inhibitor of lysosomal mannosidases. If, on the other hand, free OS are simply adsorbed onto the outer surface of membranebound organelles, then this process would not be expected to be affected by either the absence or presence of SW. In order to distinguish between these two possibilities, we examined the importance of SW in the assay mixtures, and as shown in Fig.  4A (40), this organelle would not be expected to be permeable to larger free OS. Accordingly, the lysosomal membrane should possess free OS transport machinery showing both specificity and saturability. The saturability of the transport process was examined by incubating lysosomes with increasing concentrations of [ 3 H]Man 5 GlcNAc (1 ϫ 10 5 cpm/nmol). These incubations were performed at 30°C for 10 min, conditions where transport was observed to be linear (see Fig. 3). Results demonstrate (Fig. 5) that the ATP-dependent transport of [ 3 H]Man 5 GlcNAc into lysosomes shows saturability, and typical Michaelis-Menten saturation kinetics (K m , 22.3 M; V max , 7.1 fmol/min/unit of ␤-hexosaminidase) were observed.
Lysosomal Free OS Transport Shows Selectivity toward Trimmed, Nonphosphorylated Free OS-In order to examine the selectivity of free OS transport into isolated rat liver lysosomes, we conducted transport assays using the free OS that are generated physiologically by the ER during glycoprotein biosynthesis. First, phosphorylated Man 5 GlcNAc 2 (Man 5 GlcNAc 2 -P; see Fig. 6A) released from dolichol-PP-Man 5 GlcNAc 2 on the cytosolic face of the ER (5), was tested, and as can be seen in Fig. 6A we were unable to detect lysosomal sequestration of this compound unless it had been previously treated with alkaline phosphatase, indicating that the specific activity of our preparation of phosphorylated [ 3 H]Man 5 GlcNAc 2 was sufficiently high to observe transport. In addition, these results show that free OS terminating with a reducing di-Nacetylchitobiose moiety can also be transported into rat liver lysosomes. Second, we were not able to detect lysosomal import of the large free polymannose oligosaccharides ([ 3 H]Man 9-8 GlcNAc 2 ; see Fig. 6A) that are transported out of the ER during glycoprotein biosynthesis (7,8). Furthermore, treatment of these components with endoglycosidase H, generating [ 3 H]Man 9-8 GlcNAc, did not affect their ability to be sequestered by lysosomes. However, treatment of this free OS preparation with endoglycosidase H and an ␣-1,2-mannosidase generated a branched isomer of [ 3 H]Man 5 GlcNAc, which was found to be transport-competent, indicating that the specific activity of the parent [ 3 H]Man 9-8 GlcNAc 2 was sufficient in order to observe lysosomal import. The above result indicated that free OS uptake is not dependent upon the ␣-1,2linked mannoses of the free OS, suggesting that the Man 4-3 GlcNAc structures shown in Fig. 1 may also be transport substrates. This is confirmed in Fig. 6B, where it is shown that these oligosaccharides are good transport substrates. In summary, these results demon-strate that the free OS and phosphorylated free OS that are generated by the ER during glycoprotein biosynthesis are not efficiently imported into lysosomes unless they are processed by cytosolic glycosidases and/or phosphatases.
The Cytosol-to-lysosome Free OS Transporter Recognizes a Reducing N-Acetylglucosamine-containing Sugar Motif-In order to compare lysosomal free OS transport with other sugar transport processes, we wanted to characterize the sugar motif that is recognized by the lysosomal transporter. Knowing that the ER-to-cytosol transport of free OS is inhibited by mannose and various mannosides (8), we conducted lysosomal transport assays in the presence of several commonly occurring monosaccharides. Fig. 7A shows that when lysosomes are incubated with [ 3 H]Man 5 GlcNAc in the presence 50 mM glucose, galactose, mannose, fucose, or N-acetylglucosamine (GlcNAc), each provokes a 15% loss of ␤-hexosaminidase from lysosomes; in contrast, whereas glucose, galactose, mannose, and fucose cause about 20% transport inhibition, GlcNAc completely abolishes [ 3 H]Man 5 GlcNAc transport. Furthermore, dose-response curves to GlcNAc, N-acetylgalactosamine, and N-acetylmannosamine (see Table I Fig. 3 (SEDIMENTABLE RADIOACTIVITY). The amount of sedimentable ␤-hexosaminidase has been expressed as a percentage of that occurring in the absence of swainsonine. B, after incubation of lysosomes with [ 3 H]Man 5 GlcNAc, radioactive components were recovered from the supernatant (top panel), and particulate (bottom panel) fractions derived from incubations conducted in either the absence or presence of 100 M swainsonine and analyzed by thin layer chromatography.

FIG. 5. Kinetics of ATP-dependent [ 3 H]free OS transport into lysosomes.
Upper panel, lysosomes were incubated either in the absence or presence of ATP at 30°C for 10 min with increasing concentrations of [ 3 H]Man 5 GlcNAc (specific activity 1 ϫ 10 5 cpm/nmol). After terminating the incubations and washing the particulate fraction twice, aliquots of the lysed lysosomal pellet were taken for scintillation counting and ␤-hexosaminidase assay as described under "Experimental Procedures." The saturable transport component (curve without experimental points) was computed by subtracting the ATP-independent sedimentable radioactivity from the ATP-dependent sedimentable radioactivity. Lower panel, double reciprocal analysis of the saturable, ATP-dependent, transport component. compared with that of GlcNAc, it is clear that the N-acetyl group of the GlcNAc residue plays an important role in oligosaccharide recognition by the transport machinery (Table I). Fig. 7B and Table I demonstrate two important structural features required for increased inhibitory potency of sugars. First, the compound must possess an intact or free reducing GlcNAc residue; accordingly, N-acetylglucosaminitol and ␣-benzyl GlcNAc have reduced inhibitory capacities when compared with that of GlcNAc, and di-N-acetylchitobiitol and ␤-benzyl di-N-acetylchitobioside are much less potent than di-N-acetylchitobiose at inhibiting [ 3 H]Man 5 GlcNAc transport into lysosomes. Second, substituting the reducing GlcNAc residue with Gal, Man, or GlcNAc at position 4 results in 7-, 19-, and 240-fold increases, respectively, in the inhibitory potency of the GlcNAc, indicating that although a reducing GlcNAc moitey is the most important determinant of the structural motif recognized by the transport machinery, a second sugar linked ␤1-4 to the GlcNAc greatly enhances the interaction of the substrate with the transporter. In addition, Table I shows that the linear isomer of Man 5 GlcNAc is about 2-fold more efficient at inhibiting [ 3 H]Man 5 GlcNAc transport than is Man␤1-4GlcNAc, indicating that the transporter recognizes not only this disaccharide unit but also other more distal mannose residues. Finally, Table I shows that whereas the branched isomer of Man 5 GlcNAc (see Fig. 6) interacts with the transporter less efficiently than its linear counterpart, the same oligosaccharide devoid of its reducing GlcNAc residue had no effect on [ 3 H]Man 5 GlcNAc transport over the concentration range tested. In another type of experiment we treated [ 3 H]Man 5 GlcNAc with sodium borohydride and found that the reduced oligosaccharide was no longer a substrate for lysosomal import (results not shown). Taken together with the transport specificity studies shown in Fig. 6 and the transport inhibition experiments summarized in Table I, these results demonstrate that the lysosomal transporter recognizes predominantly the reducing Man␤1-4GlcNAc disaccharide moiety of free OS.
Lysosomal Free Oligosaccharide Transport Is Not Blocked by Various Commonly Used Sugar Transport Inhibitors-The selectivity of the lysosomal free OS transport process suggested that it is carried out by a novel transporter distinct from the hexose and GlcNAc/GalNAc transporters already described in mammalian cells. In order to gain more information on the characteristics of this transporter, we conducted transport assays in the presence of various classical sugar transport inhibitors. Cytochalasin B, unlike its close relative cytochalasin D, inhibits the plasma membrane glucose carrier (41), the lysosomal hexose carrier (42), and also the lysosomal N-acetylhexosamine transporter (38). Phloretin and phloridzin inhibit the plasma membrane glucose carrier and the Na ϩ -dependent active hexose transporter, respectively (41, 43), whereas verapamil is an excellent substrate for the multidrug resistance protein (44). However, as can be seen from the upper portion of Table II, none of the above compounds possess significant inhibitory action toward [ 3 H]free OS transport into lysosomes.

H]Free OS Transport into Isolated Rat Lysosomes Requires the Hydrolysis of ATP and Is NEM-sensitive but Independent of the Ability of the Vacuolar ATPase to Acidify Lysosomes-In
order to further characterize the import of free OS into isolated rat liver lysosomes, we have investigated the role of ATP in this process using the in vitro transport assay. First, we tested whether free OS transport into isolated rat liver lysosomes required the hydrolysis of ATP, and as shown in Fig. 8A the nonhydrolyzable ATP analogs ATP␥S, AMP-PNP, and AMP-PCP were not efficient promoters of free OS transport. We went on to verify that, under our transport assay conditions, ATP hydrolysis leads to acidification of lysosomes, and, as shown in Fig. 8B, the addition of ATP to FITC-dextran-loaded lysosomes led to a carbonyl cyanide m-chlorophenylhydrazone-reversible quenching of fluorescence, indicating lysosomal acidification (33); furthermore, this acidification was inhibited by CCM A and NEM, two inhibitors of the vacuolar ATPase (45,46). Next, we examined the nature of the ATPase activity responsible for the coupling of ATP hydrolysis to lysosomal free OS import by testing a panel of well known ATPase inhibitors in the transport assay. Table II shows that whereas [ 3 H]free OS transport is sensitive NEM, a nonspecific thiol-reactive ATPase inhibitor, the specific vacuolar ATPase inhibitor, CCM A, is without effect. Furthermore, Table II demonstrates that whereas vanadate, a known inhibitor of P-type ATPases (47) and the multidrug-resistant glycoprotein (44), weakly inhibits free OS transport into lysosomes, sodium azide, a potent inhibitor of the mitochondrial F 0 -F 1 -type ATPase (45), was able to significantly potentiate [ 3 H]free OS transport into lysosomes. This effect may be due to the latter compound's ability to inhibit phosphatases in the incubation mixtures, thereby protecting the ATP and ADP during assays. DISCUSSION Initial transport experiments, carried out using permeabilized HepG2 cells (Fig. 2), suggested that lysosomal mem-branes have the ability to interact with [ 3 H]free OS and led us to test the ability of partially purified rat liver lysosomes to sequester these components. Our lysosome purification procedure led to a 30 -40-fold enrichment of ␤-hexosaminidase activity over that observed in the initial homogenate (results not shown), and in general, greater than 90% of the ␤-hexosaminidase was found to be sedimentable (see percentage of sedimentable ␤-hexosaminidase at t ϭ 0 in Fig. 3). In addition, we were able to show, as have others (33), that lysosomes remain acidic during their purification. Thus, the addition of either the proton ionophore, carbonyl cyanide m-chlorophenylhydrazone, or valinomycin, a potassium ionophore, individually to FITC-dextran-loaded lysosomes resulted in modest alkalinizations, whereas the addition of both reagents together induced a rapid and substantial alkalinization of lysosomes (results not shown), indicating that the lysosomal membranes in our lysosome preparation had low permeabilities to both hydrogen and potassium ions and were more acid than the surrounding incubation medium (33).
Much evidence indicates that not only do free OS associate with the lysosomes in the lysosome-enriched preparation, but they are also transported into functional lysosomes. First, there appeared to be a correlation between association of free OS with the membranes of the lysosomal preparation and the stability of the lysosomes in this preparation (see Fig. 3). Second, we found that association of free OS with the particulate fraction of the lysosomal preparation was greatly enhanced by the addition of low concentrations of SW to the incubation mixtures. As SW had no effect on the stability of free OS not associated with membranes, we were able to conclude that SW is protecting sedimentable free OS from latent mannosidase activities. Since the lysosome is the only known organelle possessing enzymes capable of completely hydrolyzing free OS into their constituent monosaccharides, we conclude that free OS are being transported into the lysosomes of our lysosome-enriched fraction and not into other contaminating organelles.
The Lysosomal Free OS Transporter Displays a Unique Specificity That Distinguishes It from the ER-to-cytosol Free OS Transporter-Our results show that large polymannose structures (Man 9 -8 GlcNAc 2 ), irrespective of whether they are free or linked to peptide, protein, or lipid, interact poorly with the transporter. On the other hand, less mannosylated free OS  a The percentage of inhibition of free oligosaccharide transport was evaluated as described under "Experimental Procedures." b Sodium azide was found to potentiate lysosomal transport of free oligosaccharides possibly on account of its ability to protect ATP from degradation by phosphatases during transport assays.
(Man 5-3 GlcNAc) were found to be transport substrates, indicating that the peripheral mannose residues of the larger structures are inhibitory to transport. Inhibition studies with monoand disaccharides ( Fig. 7 and Table I) revealed that the transporter binds predominantly to the GlcNAc-containing reducing terminus of the free OS and that a second sugar linked ␤-1,4 to the reducing GlcNAc enhances affinity. Furthermore, since linear Man 5 GlcNAc was better at quenching transport than Man␤1-4GlcNAc and [ 3 H]Man 3 GlcNAc appeared to be a better transport substrate than [ 3 H]Man 5 GlcNAc, either or both of the mannosyl residues linked ␣-1,3 and ␣-1,6 to the Man␤1-4-GlcNAc core must also add to the binding affinity. To summarize, a possible hypothesis for the interaction of free OS with the lysosomal transporter is that small free oligosaccharides are able to enter a pocket in the transport molecule and, once inside, bind tightly to the transporter via their reducing termini. More fully mannosylated or branched free oligosaccharides, although possessing the same high affinity binding motif, are sterically impeded from entering the pocket and are therefore neither good substrates for transport nor good transport inhibitors. Surprisingly, we noted that the most potent inhibitors of [ 3 H]free OS import into lysosomes were small chitooligosaccharides containing 2-4 residues of GlcNAc, and the parallelism of the dose-response curves shown in Fig. 7B suggests that GlcNAc, chitobiose, N-acetyllactosamine, and Man␤1-4-GlcNAc inhibit free OS transport by a common mechanism. These results suggest that the high affinity binding pocket of the lysosomal free OS transporter may be similar to those of many chitin-binding proteins and chitinases that possess a cysteine and glycine-rich structural motif known as the hevein domain (48). Lectins possessing this structural motif such as wheat germ agglutinin (49) and Urtica dioca agglutinin (50) bind GlcNAc and small chitooligosaccharides in the following order: chitotetraose Ն chitotriose Ͼ chitobiose Ͼ Ͼ GlcNAc. These observations led investigators to propose that the sugar binding sites of such lectins comprise three closely situated subsites (49,50). However, our transport inhibition studies (Table I) demonstrate that chitooligosaccharides inhibit [ 3 H]free OS transport with the following order of potency: chitotetraose Ն chitotriose Ն chitobiose Ͼ Ͼ GlcNAc, suggesting that the free OS binding pocket of the lysosomal transporter may possess only two high affinity sugar binding sites. It is also noteworthy that both ␣and ␤-N-acetylglucosaminisides bind more efficiently to wheat germ agglutinin than does GlcNAc, whereas our transport inhibition studies show that blocking of the reducing terminus of GlcNAc and di-N-acetylchitobiose with ␣and ␤-benzyl moieties, respectively, greatly reduced their capacities to quench free OS transport (Table I), indicating further differences between the oligosaccharide binding domain of the lysosomal free OS transporter and that of heveintype lectins. At present, it is not known whether or not the lysosomal free OS transporter can import chitooligosaccharides into lysosomes, but it is noteworthy that similar structures not only exist in invertebrates and fungi but may exist in vertebrates and play a role in their development (51,52).
The ER-to-cytosol transport of large free polymannose-type OS can be inhibited by mannose (IC 50 , 4.9 mM) and mannosides (benzyl mannoside, IC 50 , 0.8 mM) but neither by N-acetylglucosamine (50 mM) nor di-N-acetylchitobiose (10 mM) (8), indicating that the ER and lysosomal transporters recognize different structural motifs of free OS. Whereas the interaction of large free polymannose-type oligosaccharides with the ER-tocytosol transport apparatus has similarities to ER-situated N-glycan-lectin/enzyme interactions (calcium dependence/inhibition by mannosides; see Ref. 8), the interaction of free OS with the lysosomal transporter bears similarities to the interactions between free OS and cytosolic glycosidases. Accordingly, the cytosolic chitobiase (15) by definition interacts with the reducing di-N-acetylchitobiose moiety of free OS, but more surprisingly, the cytosolic mannosidase has been found to act only on free OS bearing a single GlcNAc residue at their reducing termini (20). Therefore, these two enzymes and the cytosol-to-lysosome free OS transporter interact with free OS in a manner that involves surveillance of the reducing terminus of the oligosaccharide. Perhaps this system has developed to promote an ordered processing of free OS in the cytosol and to ensure that glycoproteins, glycopeptides, and dolichol oligosaccharides do not interfere with the sequestration of free OS by lysosomes. FIG. 8. ATP hydrolysis leads to both free OS transport and activation of the proton pumping ATPase. A, transport assays were carried out in the presence of 5 mM ATP or the nonhydrolyzable ATP analogues ATP␥S, AMP-PNP, and AMP-PCP (all 5 mM). The transport of [ 3 H]free OS observed in the presence of 5 mM ATP has been set at 100%. Lysosome disruption was monitored as described in the legend to Fig. 6, and the percentage of sedimentable ␤-hexosaminidase observed in the presence of 5 mM ATP has been set at 100%. B, FITCdextran-loaded lysosomes were prepared as described under "Experimental Procedures" and incubated under standard transport assay conditions at room temperature (22°C) in a stirred cuvette. Fluorescence recordings were initiated, and, after stabilization of the base line, water (upper trace), NEM (1 mM final concentration, middle trace), or CCM A (100 nM final concentration, lower trace) was added prior to the addition of ATP (6.25 mM final concentration). After observing the effect of ATP, the proton ionophore carbonyl cyanide m-chlorophenylhydrazone was added (25 M final). A downward deflection of the trace is a measure of the fluorescence quenching caused by lysosomal acidification, whereas an upward deflection represents an alkalinization of the lysosomal interior.
A Novel Transporter Is Responsible for the Import of Free Oligosaccharides into Lysosomes-How do the kinetic characteristics of lysosomal free OS transport compare with those of other lysosomal and plasma membrane sugar transporters? Here we observed that lysosomal free OS sequestration was temperature-dependent and displayed a Q 10 of about 2.5 (calculated from the ATP-dependent transport observed after 5-min incubations at 37, 30, and 25°C; see Fig. 3), which is similar to other lysosomal sugar transport systems (N-acetylhexosamines (38), Q 10 ϭ 2.3; sialic acid (53), Q 10 ϭ 2.4). We were unable to detect saturable free OS transport into lysosomes unless ATP was included in the incubation mixtures; furthermore, we show that free OS transport into lysosomes is critically dependent upon ATP hydrolysis, yet independent of a functional lysosomal proton pump. In contrast, other mammalian lysosomal transporters influenced by ATP such as those described for sialic acid and sulfate show saturability in the absence of ATP and have been shown to be stimulated only about 2-fold by the addition of ATP (39,54). The ATP-dependent transport of free OS into lysosomes showed Michaelis-Menten-type saturation kinetics displaying a K uptake of 22.3 M. This value is significantly lower than the K uptake values obtained for lysosomal hexose (Glc, 75-90 mM; Man, 50 -75 mM (42)), N-acetylhexosamine (GlcNAc, 4.4 mM, GalNAc, 4.4 mM (38)), and sialic acid (0.24 mM (39)) carriers. Our K uptake value must be considered as an approximation, since 10-min incubations at 30°C represent the limit of the linearity of the initial transport rate; however, the time chosen was a compromise, since shorter incubation times would have led to very low levels of radioactivity becoming associated with lysosomes. Nevertheless, another cytosolic activity implicated in the clearance of ER-generated free OS from the cytosol, the neutral cytosolic endoglycosidase H-like activity, was recently found to have a K m value of 25 M for pyridylaminated Man 6 GlcNAc 2 (55). Here, we demonstrate that free OS import into lysosomes was sensitive to NEM, a thiol-reactive drug that inhibits many ATPases (33,56). In contrast, free OS import into lysosomes was only weakly inhibited by vanadate, an inhibitor of P-type ATPases (47) and not at all by the F 0 -F 1 type ATPase inhibitor sodium azide (45). Furthermore, we were unable to demonstrate inhibition of lysosomal free OS import by other well known sugar transporters such as cytochalasin B, phloretin, and phloridzin. Taken together with the unique selectivity of the transport process, these results demonstrate that the sequestration of free OS into lysosomes is accomplished by a novel transporter. The ability of lysosomes to continually clear ER-generated free polymannose-type oligosaccharides from the cytosol indicates that this free OS transporter must be considered an important housekeeping activity of cells engaged in the biosynthesis of glycoproteins.