Dual Functions for Cytosolic α-Mannosidase (Man2C1)

Background: Man2C1 regulates apoptosis via an unknown mechanism. Results: Suppression of Man2C1 induces mitochondria-dependent apoptosis independently of its enzyme activity. Conclusion: Man2C1 has dual functions in glycan catabolism and apoptotic signaling. Significance: This study demonstrates that the role of Man2C1 in apoptosis is independent of its α-mannosidase activity. Cytosolic α-mannosidase (Man2C1) trims free oligosaccharides in mammalian cells, and its down-regulation reportedly delays cancer growth by inducing mitotic arrest or apoptosis. However, the mechanism by which Man2C1 down-regulation induces apoptosis is unknown. Here, we demonstrated that silencing of Man2C1 via small hairpin RNAs induced mitochondria-dependent apoptosis in HeLa cells. Expression of CHOP (C/EBP homologous protein), a transcription factor critical to endoplasmic reticulum stress-induced apoptosis, was significantly up-regulated in Man2C1 knockdown cells. However, this enhanced CHOP expression was not caused by endoplasmic reticulum stress. Interestingly, Man2C1 catalytic activity was not required for this regulation of apoptosis; introduction of mutant, enzymatically inactive Man2C1 rescued apoptotic phenotypes of Man2C1 knockdown cells. These results show that Man2C1 has dual functions: one in glycan catabolism and another in apoptotic signaling.

Free oligosaccharides derived from N-glycans on proteins or from their lipid-linked precursors accumulate in the cytosol (5,(7)(8)(9)(10). Currently, two glycosidases (endo-␤-N-acetylglucosaminidase (11) and Man2C1 (12)) are known to be involved in the catabolism of free, high mannose-type oligosaccharides in the cytosol of mammalian cells. The processing of free oligosaccharides by these enzymes is postulated to be important for the subsequent transport of the oligosaccharides into lysosomes by a putative oligosaccharide transporter on the lysosomal membrane (13); this hypothesis is consistent with the observation that down-regulation or inhibition of Man2C1 leads to oligosaccharide accumulation in the cytosol (12, 14 -17).
Reportedly, Man2C1 overexpression induces ER 2 stress (17) and promotes tumor growth and metastasis (18 -20). These findings led us to speculate that this protein is a potential target for anti-cancer therapy. Suppression of the Man2C1 gene was shown to induce growth delay or arrest in nasopharyngeal carcinoma cells (19) and in Hek293 cells (12). Down-regulation of Man2C1 reportedly also results in mitotic arrest and apoptosis in esophageal carcinoma cells (21). These results indicate that Man2C1 has a role in regulating apoptotic signaling for certain cells.
Here, we investigated the mechanism(s) by which Man2C1 suppression causes cell growth arrest and/or apoptosis. We found that Man2C1 knockdown cells underwent apoptosis via a mitochondria-dependent pathway that involved CHOP (C/EBP homologous protein) up-regulation. Our results also show that Man2C1 had a role in apoptosis that was independent of its enzymatic activity; this finding provides convincing evidence that Man2C1 has dual functions: one as a catabolic enzyme for the breakdown of free, cytosolic oligosaccharides and another as a modulator of apoptotic signaling.
Lentiviral Vector-based Small Hairpin RNA-Sequences encoding two distinct shRNAs were each cloned into lentivirus vectors; each construct was introduced into cells either by viral infection or plasmid transfection. The two shRNAs were designed to target and knock down Man2C1 expression. Lentiviral particles (2-15 l, roughly 1 particle/cell, Mission lentiviral transduction particles; Sigma) were added to 5 ϫ 10 4 cells. To screen for stable transfectants, culture medium was replaced with fresh DMEM containing 0.5 g/ml puromycin 48 h after infection. Medium was replaced every 3 or 4 days until resistant colonies were identified. The following two sequences were used for these experiments: For plasmid transfection, 2 g of a shRNA vector or a control vector (SHC001, described above, or MISSION pLKO.1-puro nonmammalian shRNA control plasmid, SHC002; Sigma) was transfected into 1 ϫ 10 6 cells using the Nucleofactor transfected kit (Amaxa) according to the manufacturer's protocols. The cells were harvested 48 h after transfection for further examination. The sequences of Man2C1 shRNAs were the same as those above.
Quantitative Gene Expression Analyses via PCR-For real time PCR analysis, total RNA was isolated from cells with the Qiagen RNeasy kit; 4 g of each RNA sample was reverse-transcribed with random hexamers using the SuperScript III RT kit (Invitrogen) according to the manufacturer's protocols. Each real time PCR was performed using a TaqMan Universal PCR Master Mix system (Roche Applied Science). Each final 20-l real time PCR contained 10 l of TaqMan Mix buffer, 1 l of the probes provided by Roche Applied Science, and 9 l of cDNA (0.2 g). The cDNA was amplified by an initial step at 95°C for 10 min, followed by 40 cycles at 95°C for 5 s and at 55°C for 20 s; an ABI PRISM 7900HT sequence detection system (Applied Biosystems) was used. The samples were analyzed in duplicate; the mean number of cycles to the threshold of fluorescence detection was calculated for each sample. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression in each sample was measured to normalize the amount of cDNA in each sample. The Man2C1 and GAPDH probes used for real time PCR were purchased from Roche Applied Science.
To examine XBP1 transcripts via RT-PCR, total RNA was isolated and reverse-transcribed as described above. The following two primers were used to assess XBP1 expression: sense primer, 5Ј-CCTTGTAGTTGAGAACCAGG-3Ј, and antisense primer, 5Ј-GGGGCTTGGTATATATGTGG-3Ј. To distinguish spliced XBP1 from unspliced XBP1, PCR products were incubated with 1 l of PstI (15 units/l) at 37°C for 30 min before being resolved on a 2% agarose gel (22).
FITC-Annexin V/Propidium Iodide (PI) Double Staining Analysis-Cultured cells were harvested, washed twice with ice-cold PBS, and suspended in 1ϫ binding buffer (10 mM HEPES buffer, 0.14 M NaCl, 2.5 mM CaCl 2 , pH 7.4) at a concentration of 1 ϫ 10 6 cells/ml. Samples of 100 l (1 ϫ 10 5 cells) were each transferred into separate 5-ml FALCON tubes; 5 l of FITC-annexin V (BD Biosciences) and 10 l of PI (50 g/ml) were added to each tube. Each cell suspension was mixed gently via vortex and then incubated for 15 min at room temperature (25°C) in the dark; 400 l of 1ϫ binding buffer was then added to each suspension. The cells were subjected to flow cytometry analysis within 1 h.
PI Cell Cycle Analysis-Cultured cells (1 ϫ 10 6 ) were harvested, then fixed, and permeabilized via treatment with 70% (w/v) ice-cold ethanol overnight at 4°C; the cells were then washed with PBS and resuspended in 1 ml of 0.1% (v/v) Triton X-100/PBS containing 20 g/ml PI and 500 units/ml boiled RNase A. After incubation for 15 min in the dark at room temperature, the cells were subjected to flow cytometry analysis.
Fractionation of Cytosol and Mitochondria-HeLa cells (1 ϫ 10 6 ) and Man2C1 knockdown HeLa cells were washed twice with PBS, then harvested, and finally separated into cytosolic fractions and mitochondrial fractions using a Qproteome mitochondria isolation kit (Qiagen) according to the manufacturer's protocol.
Cytosolic ␣-Mannosidase Assay-The isolation of a cytosolic fraction from cell lysates was described above. For each 40-l ␣-mannosidase assay reaction, 1 mM CoCl 2 , 20 l of 200 mM Mes/NaOH buffer, pH 6.7, 10 l of substrate (p-nitrophenyl-␣-D-mannoside final concentration, 1 mM; Sigma), and 10 l of cytosolic fractions of the indicated cell type were mixed gently and incubated at 37°C for 1 h. Each reaction was stopped by adding 900 l of 0.2 M Na 2 CO 3 (pH 11.4). Liberated p-nitrophenol was measured with a spectrophotometer set at 410 nm; one unit of enzyme was defined as the amount required to release 1 mol of p-nitrophenol from p-nitrophenyl-␣-D-mannoside in 1 h at 37°C under the experimental condition. Total protein concentration in each crude enzyme preparation was determined via the bicinchoninic acid method (Pierce) according to the manufacturer's instructions; bovine serum albumin was used as the standard.
Extraction and Labeling of Free Oligosaccharides with 2-Aminopyridine (PA) Recovered from Cytosol-Individual aliquots of 1 ϫ 10 8 cells were each washed twice with cold PBS; the cells were then harvested and lysed in 800 l of lysis buffer (10 mM HEPES/NaOH buffer, pH 7.4, 5 mM DTT, 250 mM mannitol, 1 mM EDTA, pH 8.0, 1ϫ complete protease inhibitor mixture, and 1 mM Pefabloc). Each cell lysate was homogenized and centrifuged at 1,000 ϫ g for 10 min at 4°C to remove nuclei and unbroken cells. Each supernatant was then centrifuged at 100,000 ϫ g for 1 h at 4°C; the resulting supernatants contained the cytosolic fractions. To precipitate proteins, ethanol (150% v/v) was added to the cytosolic fraction; each solution was mixed well and centrifuged at 17,000 ϫ g for 20 min at 4°C. Each supernatant, which contained the free oligosaccharides, was desalted using a PD-10 column (GE Healthcare) according to the manufacturer's instructions. Detailed methods for PA labeling and for removal of excess reagent were described before (23); each sample thus prepared was used for subsequent size fractionation HPLC analysis.

Down-regulation of Man2C1 by shRNA Induces HeLa Cell
Apoptosis-To investigate the function of Man2C1 in growth arrest and/or cellular apoptosis, we used shRNA to down-regulate Man2C1 expression in HeLa cells. We achieved Man2C1 knockdown (49 or 51%) using two distinct shRNA vectors specific for Man2C1 (Fig. 1A); Man2C1 knockdown caused suppression of ␣-mannosidase activity (ϳ70 or 65%, respectively) in the cytosol (Fig. 1B). These Man2C1 knockdown cells exhibited increased accumulation of free oligosaccharides in the cytosol ( Table 1), indicating that, as with previous observations (12, 14 -17), processing of cytosolic free oligosaccharides was impaired by Man2C1 down-regulation. In these Man2C1 knockdown cells, apoptosis was evident, as shown by FITCannexin V/PI double staining. Higher numbers of annexin V ϩ /PI Ϫ cells were observed in both groups of Man2C1 knockdown cells, indicating up-regulation of apoptotic signal, when compared with control cells (Fig. 1C). Moreover, PI cell cycle analysis revealed accumulation of cells in sub-G 1 phase (Fig.  1D), also indicating that the number of apoptotic cells was higher among the Man2C1 knockdown cells. Apoptotic bodies were frequently apparent in the nuclei of Man2C1 knockdown cells (supplemental Fig. S1).
Although apoptotic cells were common in stable transfectants expressing either Man2C1 shRNA, we were concerned that we had selected clones with special characteristics. This consideration was of particular concern if Man2C1 down-regulation was critical for normal cell growth and survival. We therefore tested the effect of transient transfection of Man2C1 shRNA into HeLa cells. Transient expression of either Man2C1 shRNA suppressed the Man2C1 mRNA level (56 or 58% of control) in HeLa cells ( Fig. 2A). We also observed higher numbers of annexin V ϩ /PI Ϫ cells in transiently transfected Man2C1 knockdown cells (Fig. 2B). These results confirm that more HeLa cells underwent apoptosis when Man2C1 was suppressed than when it was not.
Down-regulation of Man2C1 Induces Mitochondria-dependent Apoptosis-To Down-regulation of Man2C1 induces mitochondria-dependent apoptosis. To identify which apoptotic signaling pathway(s) were regulated by Man2C1, caspase 3 cleavage was examined because it is central to all caspase-dependent apoptotic signal (see Fig. 9 for a schematic view of the intrinsic and extrinsic apoptotic signaling pathways). Levels of pro-caspase 3, the inactive form of caspase 3, evidently decreased in Man2C1 knockdown cells, indicating enhanced activation of this protein (Fig. 3A, compare lanes 3 and 4 with  lanes 1 and 2). We could not detect the active (cleaved) form of caspase 3 even in the presence of MG-132, a positive inducer of cell apoptosis (24 -26); therefore, we examined the processing of poly(ADP-ribose) polymerase (PARP), a major substrate of activated caspase 3 (27). A drastic increase of processed PARP (ϳ89 kDa) was evident in cells stably expressing either Man2C1 shRNA (Fig. 3B, lanes 3 and 4) and in cells transiently transfected with either Man2C1 shRNA (Fig. 3C, compare lanes 2  and 3 with lane 1). Taken together, these results clearly indicated that Man2C1 suppression resulted in caspase 3 activation and in turn increased the occurrence of the cleaved form of PARP.
Having confirmed caspase 3 activation, we next studied several upstream apoptotic signals. First, processing of two initiator caspases (caspases 8 and 9) was examined. Several major signaling pathways induce apoptosis in mammalian cells (28 -31). In the extrinsic pathway, a signaling pathway is initiated by binding between the death receptor and the apoptotic ligands; pro-caspase 8 is then recruited and activated to further activate caspase 3 (32,33). In the intrinsic pathway, death signals stimulate translocation, oligomerization, and insertion of Bax and/or Bak into the outer mitochondrial membrane; this insertion causes the release of several proteins, including cytochrome c, from the mitochondria and leads to pro-caspase 9 activation and subsequent caspase 3 activation (34). Caspase-3 is ultimately activated in both pathways but via distinct initiator  caspases: caspase 8 for the extrinsic pathway and caspase 9 for the intrinsic pathway. Pro-caspase 8 levels did not differ between control and Man2C1 knockdown cells, nor did levels of cleaved (active) caspase 8 (Fig. 3D, top panel, lanes 3 and 4); these findings clearly indicated that the Man2C1 knockdown-induced apoptotic signal did not occur via the death receptor-mediated extrinsic pathway. In sharp contrast, levels of the active form of caspase 9 (37 kDa/35 kDa) were higher in Man2C1 knockdown cells than in control cells (Fig. 3D, middle panel, lanes 3 and 4), suggesting that Man2C1 knockdown-induced apoptosis operated via the mitochondria-dependent intrinsic pathway.

TABLE 1 Size fractionation HPLC measurements of free oligosaccharides in the cytosol of HeLa cells with modulated expression of wild-type or mutant Man2C1
To further confirm that mitochondria-mediated cell death occurred in Man2C1 knockdown cells, mitochondrial cytochrome c release to cytosol was assessed in these cells because this phenomenon precedes activation of caspase 9-dependent apoptosis (35). The levels of cytosolic cytochrome c were higher in Man2C1 knockdown cells than in control cells, and conversely, levels of mitochondrial cytochrome c were lower in knockdown cells than in control cells (Fig. 3E, lanes 3 and 4). Taken together, these results indicated that Man2C1 downregulation stimulated release of mitochondrial cytochrome c to the cytosol and thereby induced the caspase 9-dependent apoptotic signaling pathway. (36 -38). CHOP, often used as a marker for activation of ER stress, is believed to be the key molecule mediating mitochondria-dependent apoptosis triggered by ER stress (36 -39). Having shown that Man2C1 downregulation caused mitochondria-dependent apoptosis, we next investigated whether CHOP was involved in this phenomenon. Surprisingly, shRNA-mediated Man2C1 down-regulation led to a drastic increase in CHOP expression in HeLa cells (Fig. 4A,   3 and 4 with lanes 1 and 2), indicating that CHOP was involved in the apoptosis induced by Man2C1 down-regulation.

Man2C1 Down-regulation Causes Enhanced CHOP Expression without Triggering ER Stress-The intrinsic apoptotic pathway can be induced by accumulation of unfolded or misfolded proteins in the ER
Although CHOP increased in Man2C1 knockdown cells, ER stress signaling was not evident because sXBP1 (the spliced form of the XBP1 gene product and a marker for ER stress (37)) was not evident in these cells (Fig. 4B, lanes 3 and 4). Similarly, BiP, a major ER chaperone that is up-regulated upon ER stress, was not up-regulated in Man2C1 knockdown cells (Fig. 4C,  lanes 3 and 4). Taken together, these results indicated that ER stress may not be involved in the Man2C1 knockdown-induced apoptosis and that up-regulation of CHOP could occur in an ER stress-independent fashion.
Down-regulation of Man2C1-induced Apoptosis Is Independent of the Man2C1 ␣-Mannosidase Activity-Man2C1 is the ␣-mannosidase involved in the catabolic pathway that processes free oligosaccharides in the cytosol (5,7,8,15,17,40). We and others previously identified in vivo inhibitors of this enzyme (14,16). Interestingly, although we saw increased free oligosaccharides in the cytosol, we did not observe any cell growth defect in the presence of this inhibitor (16). This result is in sharp contrast to the gene suppression-induced growth arrest/apoptosis observed in previous studies (12,21), as well as in this study. Therefore, we hypothesized that the apoptosis induced by Man2C1 knockdown was not related to the enzymatic function of Man2C1.
To test this hypothesis, we introduced an enzymatically inactive form of Man2C1, in which Asp-372 was converted to Ala, into Man2C1 knockdown cells to see whether this mutant rescued the apoptotic phenotype caused by the shRNA. The mutated aspartic acid residue is reportedly a catalytic residue that is conserved among family 38 ␣-mannosidases (41). After introduction of a wild-type or mutant Man2C1 gene, mRNA levels of wild-type and mutant Man2C1 increased significantly in the respective cells (Fig. 5A). This observation was consistent with the observations that cytosolic ␣-mannosidase activity increased in transfected cells expressing wild-type Man2C1, but not in cells expressing the catalytically inactive mutant Man2C1 (D372A) (Fig. 5B). Consistent with the cytosolic ␣-mannosidase activity, high mannose-type free oligosaccharides accumulated in Man2C1 knockdown cells were processed by introduction of wild-type Man2C1, but not by introduction  Table 1). Interestingly, expression of either wild-type or mutant Man2C1 (D372A) clearly rescued the apoptotic phenotype in Man2C1 knockdown cells, as shown by annexin V/PI staining (Fig. 5D). Moreover, immunoblot analysis indicated that PARP cleavage decreased in transfectants carrying either Man2C1 transgene, irrespective of the presence or absence of the mannosidase-inactivating mutation (Fig. 5E, lanes 3 and 4). Taken together, these results clearly demonstrated that the regulatory function of Man2C1 in apoptosis was independent of the mannosidase activity.
Recently, He et al. (20) reported that full-length Man2C1 or N-or C-terminal truncations of Man2C1 attenuate phosphatase and tensin homolog (PTEN) functions by binding to PTEN. To determine whether truncated fragments of Man2C1 could rescue the apoptosis induced by Man2C1 shRNA, we constructed V5-tagged forms of full-length Man2C1 and of N-and C-terminally truncated Man2C1 (Fig. 6A). Each of the V5-tagged Man2CI fusion proteins were expressed in Man2C1 knockdown cells (Fig. 6B); neither N-Man2C1 nor C-Man2C1 rescued apoptosis, as indicated by PARP cleavage (Fig. 6C, lanes  3 and 4); in sharp contrast, full-length Man2C1 suppressed the cleavage of PARP (Fig. 6C, lane 2).

Suppression of Man2C1 Also Results in Apoptosis in Other
Cell Types-To further confirm that down-regulation of Man2C1 caused apoptosis in mammalian cells, we tested the effect of shRNA-mediated Man2C1 knockdown in two other types of human-derived cultured cells, MKN45 cells (gastric adenocarcinoma) and HepG2 cells (hepatocellular carcinoma). Each type of Man2C1 shRNA was independently transiently transfected into each cell type. Transient transfection of either Man2C1 shRNA caused decreased Man2C1 expression in both cell types (Fig. 7, A and C), and in both cell types, increased PARP cleavage was observed (Fig. 7, B and D, lanes 2 and 3), clearly indicating that suppression of Man2C1-induced apoptosis was not specific to HeLa cells.

DISCUSSION
Man2C1 is a cytosolic ␣-mannosidase commonly found in vertebrates and is known to catabolize free oligosaccharides that have accumulated in the cytosol (12, 14 -17). Overexpression and suppression of the Man2C1 gene each reportedly have multiple phenotypic consequences. In one investigation, transgene expression of Man2C1 promoted tumor progress in mice (18); later it was found that Man2C1 overexpression could attenuate PTEN function in prostate cancer cells (20). Con-   relative to control subjects (43). None of these reports, however, successfully tied these phenotypic consequences with the Man2C1 enzymatic activity.
Bernon et al. (17) recently reported that Man2C1 overexpression could lead to increased intracellular mannose concentration and a consequent reduction in protein glycosylation, which in turn causes ER stress; these findings indicate that Man2C1 enzymatic activity does affect ER stress phenotypes. In contrast, we herein provide convincing evidence that Man2C1 also has an additional, enzyme-independent role as a key molecule regulating the intrinsic apoptosis signaling pathway.
Here, we demonstrated that suppression of Man2C1 caused HeLa cells to undergo mitochondria-dependent apoptosis through the up-regulation of CHOP. Interestingly, CHOP upregulation did not depend on ER stress, because Man2C1 knockdown-in sharp contrast to Man2C1 overexpression (17)-did not elicit an unfolded protein response. CHOP is often used as a marker for ER stress activation, and it is a major player in ER stress-triggered apoptosis (36 -38); nevertheless, the ER stress response can be induced independently of CHOP (44). Conversely, CHOP expression can be up-regulated independently of ER stress (45). Therefore, it appears possible that, following suppression of Man2C1 expression, CHOP is somehow up-regulated independently of ER stress.
Most strikingly, the effect of Man2C1 on apoptosis was found to be independent of Man2C1 enzymatic activity, because a catalytically inactive form of Man2C1, like the wild-type protein, rescued the apoptotic phenotype in Man2C1 knockdown cells. Our results indicated that Man2C1 had a nonenzymatic function in addition to its enzymatic function. Notably, orthologs of a cytoplasmic peptide:N-glycanase, which itself is involved in the catabolic pathway for free oligosaccharides, are reported to be catalytically inactive, although mutant forms of these orthologs exhibited severe phenotypes (46,47). These observations led us to speculate that such enzyme-independent functions are not rare but are possibly an unrecognized feature of many cytosolic enzymes.
Man2C1 was recently identified as a PTEN-binding protein (20). He et al. (20) suggest that Man2C1, by binding to PTEN, attenuates PTEN functions; thus, Man2C1 up-regulation could promote tumorigenesis in prostate cancer cells. We did not observe any impaired PTEN signal, as judged by the level of phosphorylated Akt in two distinct lines of Man2C1 knockdown cells (Fig. 8). Consistent with this result, we found that an N-terminal fragment of Man2C1 did not have anti-apoptotic activity; notably, this fragment was previously shown to be enough to inhibit the PIP3 phosphatase activity of PTEN (20). Currently, the role of PTEN, if any, in inducing apoptosis upon Man2C1 knockdown remains unknown.
In summary, we have clearly shown that Man2C1 down-regulation induced apoptosis in HeLa cells via a mitochondria-dependent pathway. Notably, ER stress was not involved in CHOP up-regulation in the Man2C1 knockdown cells (Fig. 9). This anti-apoptotic function of Man2C1 was shown to be independent of ␣-mannosidase activity; together, these findings provide convincing evidence that Man2C1 has a role in addition to that as a catabolic enzyme. Man2C1 up-regulation is reportedly involved in tumorigenesis of various cancer types; therefore, this protein has potential as an attractive target for anti-cancer therapy (16,19). Further clarification of the mechanism by which Man2C1 down-regulation induces apoptosis will provide clues for designing Man2C1-based anti-cancer reagents in the future.  . Schematic representation of the extrinsic and intrinsic apoptosis signal pathways. Man2C1 suppression seemed to activate the intrinsic pathway and act upstream of CHOP up-regulation, cytochrome c release from the mitochondria, caspase-9 activation, and caspase-3 activation.