Acid Phosphatase 5 Is Responsible for Removing the Mannose 6-Phosphate Recognition Marker from Lysosomal Proteins

Acid phosphatase 5 is responsible for removing the mannose 6-phosphate recognition marker from lysosomal proteins

Pengling Suna, David E. Sleata, Miche` le Lecocqb, Alison R. Haymanc, Michel Jadotb, and Peter Lobela,1

aCenter for Advanced Biotechnology and Medicine and Department of Pharmacology, University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School, Piscataway, NJ 08854; bLaboratoire de Chimie Physiologique, Unite´de Recherche en Physiologie Mole´culaire, Faculte´s Universitaires Notre-Dame de la Paix, 61 rue de Bruxelles, 5000-Namur, Belgium; and cSchool of Clinical Veterinary Science, University of Bristol, Bristol BS40 5DU, UK

Edited by Stuart A. Kornfeld, Washington University School of Medicine, St. Louis, MO, and approved September 12, 2008 (received for review July 31, 2008) Most newly synthesized proteins destined for the lysosome reach Golgi to repeat the process or to the plasma membrane. At the this location via a specific intracellular pathway. In the Golgi, a latter location, the CI-MPR can function in the endocytosis and phosphotransferase specifically labels lysosomal proteins with lysosomal targeting of extracellular Man-6-P glycoproteins. mannose 6-phosphate (Man-6-P). This modification is recognized Early studies indicated that Man-6 phosphorylation was a by receptors that target the lysosomal proteins to the lysosome transient modification of lysosomal proteins that was removed where, in most cell types, the Man-6-P recognition marker is rapidly with a half-life of 1.4 h in mouse lymphoma cells (3) and also removed. Despite extensive characterization of this pathway, the rapidly removed in CHO cells (4). The exact site of dephos- enzyme responsible for the removal of the targeting modification phorylation is not clear, and there is evidence suggesting that it has remained elusive. In this study, we have identified this activity. may occur in both a prelysosomal compartment (5) and a dense Preliminary investigations using a cell-based bioassay were used to postendosomal compartment that probably represents the lyso- follow a dephosphorylation activity that was associated with the some itself (6). The rate and the extent of removal of the lysosomal fraction. This activity was high in the liver, where Man-6-P recognition marker in cell culture are dependent on endogenous lysosomal proteins are efficiently dephosphorylated, cell-type and culture conditions. For example, the Man-6-P but present at a much lower level in the brain, where the modi- modification was found to persist in cells lacking the CI-MPR fication persists. This observation, combined with an analysis of (4). In addition, mouse L cells remove the Man-6-P modification the expression of lysosomal proteins in different tissues, led us to normally when grown at low density but accumulate Man-6-P identify acid phosphatase 5 (ACP5) as a candidate for the enzyme glycoproteins when grown at high density or in the absence of that removes Man-6-P. Expression of ACP5 in N1E-115 neuroblas- serum (7, 8). Man-6-P glycoproteins endocytosed by the L-cells toma cells, which do not efficiently dephosphorylate lysosomal in the absence of serum were restricted to a subset of lysosomes, proteins, significantly decreased the steady state levels of Man6-P suggesting that there may be lysosomal populations with distinct glycoproteins. Analysis of ACP5-deficient mice revealed that levels functions that differ in their ability to remove the Man-6-P of Man-6-P glycoproteins were highly elevated in tissues that modification (8). normally express ACP5, and this resulted from a failure to dephos- Cell-specific differences in the removal of the Man-6-P rec- phorylate lysosomal proteins. These results indicate a central role ognition marker have also been observed under physiological for ACP5 in removal of the Man-6-P recognition marker and open conditions. Although rat liver has higher levels of most lysosomal up new avenues to investigate the importance of this process in cell enzyme activities than rat brain, the actual proportion of these biology and medicine. activities that are represented by the Man-6-P-containing form of each enzyme is dramatically less in the liver than in the brain dephosphorylation ͉ lysosome ͉ protein targeting (9). In the brain, Man-6-P-containing glycoproteins predomi- nantly accumulate within the lysosomes that are present in neuronal cell bodies (10). he lysosome is an acidic membrane-delimited organelle that The cellular activity that is responsible for the removal of the plays a critical role in the cellular digestion of a diverse range T Man-6-P recognition marker has not been identified. It could of macromolecules including proteins, carbohydrates, nucleic conceivably be a glycosidase or isomerase but is most likely a acids, and lipids (1). These catabolic functions are conducted phosphatases, and several studies have investigated the role of primarily by the concerted actions of over 60 soluble luminal two lysosomal hydrolases, acid phosphatase 2 (ACP2, also called hydrolases and accessory proteins, most of which are targeted to lysosomal acid phosphatase) and acid phosphatase 5 (ACP5, also the lysosome by a common pathway (reviewed in ref. 2). Central called tartrate-resistant acid phosphatase or uteroferrin) in the to this pathway is a posttranslational modification that is largely removal of Man-6-P from lysosomal proteins. ACP2 is trans- specific to lysosomal proteins, mannose 6-phosphate (Man-6-P). ported to the lysosome in a membrane-bound form and then is Like other glycoproteins, soluble lysosomal proteins are synthe- proteolytically cleaved to release a soluble phosphatase (11), sized in the endoplasmic reticulum and are cotranslationally whereas ACP5 is a soluble lysosomal protein that is targeted via glycosylated on select asparagine residues. As these proteins the Man-6-P pathway (12). Both enzymes hydrolyze a variety of move through the secretory pathway, the lysosomal proteins are selectively recognized by a phosphotransferase that initiates a two-step reaction that results in the generation of the Man-6-P Author contributions: P.S., D.E.S., M.L., M.J., and P.L. designed research; P.S., M.L., and modification on specific N-linked oligosaccharides. The modi- A.R.H. performed research; A.R.H. contributed new reagents/analytic tools; P.S., D.E.S., fied proteins are then recognized by two Man-6-P receptors M.L., M.J., and P.L. analyzed data; and P.S., D.E.S., and P.L. wrote the paper. (MPRs), the cation-dependent MPR and the cation-indepen- The authors declare no conflict of interest. dent (CI-) MPR. These receptors bind the phosphorylated This article is a PNAS Direct Submission. lysosomal proteins in the near neutral pH environment of the 1To whom correspondence should be addressed. E-mail: [email protected]. transGolgi network and travel to an acidic prelysosomal com- This article contains supporting information online at www.pnas.org/cgi/content/full/ partment in which the low pH promotes dissociation of the 0807472105/DCSupplemental. receptors and ligands. The receptors then recycle back to the © 2008 by The National Academy of Sciences of the USA

16590–16595 ͉ PNAS ͉ October 28, 2008 ͉ vol. 105 ͉ no. 43 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0807472105 Downloaded by guest on September 25, 2021 phosphomonoesters at acid pH in vitro, but their physiological brain. The liver extract allows the N1E-115 cells to dephosphor- substrates are unknown. ACP2 was overexpressed by a factor of ylate Ϸ80% of the Man-6-P containing forms of lysosomal 100-fold in mouse L cells, but this failed to increase the slow rate proteins, whereas dephosphorylation in cells cultured with the of removal of Man-6-P and purified ACP2 did not dephosphor- brain extract was considerably less (Ϸ20%) (data not shown). ylate arylsulfatase A in vitro (13). These data suggest that ACP2 This finding is congruent with our measurements of steady-state is not involved in the removal of Man-6-P under physiological phosphorylation in vivo, indicating physiological relevance. conditions. In contrast, ACP5 can efficiently dephosphorylate Second, the dephosphorylation activity is associated with arylsulfatase A in vitro and was originally considered a likely lysosomes. We conducted classical analytical subcellular frac- candidate for the Man-6-phosphatase (6). However, subsequent tionation to determine what membrane-delimited cellular com- studies on fibroblasts from mice that were deficient in both partment contains the dephosphorylation activity. From analysis ACP5 and ACP2 concluded that the absence of both of these of membrane fractions from postnuclear supernatants, the ac- enzymes had no effect on the dephosphorylation of endocytosed tivity was enriched most highly in the differential centrifugation arylsulfatase A (14). Thus, the current view is that ACP2 and fraction L, similar to that of lysosomal markers (Fig. S4). ACP5 do not play a significant role in the dephosphorylation of Although these results are consistent with lysosomal localization, lysosomal proteins under physiological conditions. peroxisomes are also enriched in liver L fractions. We therefore In a recent proteomic study of lysosomal proteins in rat tissues, conducted isopycnic centrifugation on liver organelles prepared we made the observation that ACP5 protein was present at low from control mice and mice injected with Triton WR-1339 4 days levels in the brain but relatively high levels in the liver (15). This before subcellular fractionation. Triton WR-1339 induces a finding is consistent with the measurement of a relative abun- lysosomal lipidosis and results in a selective shift in the buoyant dance of ACP5 transcripts in rodents (16). Given that the levels density of lysosomes but not of other organelles. As demon- of ACP5 appeared inversely correlated with levels of glycopro- strated in Fig. S5, the dephosphorylation activity undergoes a teins containing Man-6-P in rat liver and brain, these observa- shift similar to that of lysosomal markers. These findings are tions suggest that this enzyme could play a role in the removal consistent with the conclusion of others (6) that the activity that of this carbohydrate modification. In this study, we have exam- removes the Man-6-P modification resides in a late endocytic/ ined the potential role of ACP5 in removal of the Man-6-P lysosomal compartment. recognition marker and demonstrate that this enzyme is responsible for this cellular function. Tissue Expression of Lysosomal Proteins. The observation that the activity that removes Man-6-P resides within the lysosome led us Results to consider known and potential lysosomal proteins as candi- dates for this enzyme. Given that the Man-6-P marker is largely Bioassay for the Activity that Removes Man-6-P. We initially devel- retained in the brain and removed in the liver, the protein oped a bioassay for the activity responsible for the removal of the responsible for this activity would be expected to be expressed Man-6-P-targeting modification from lysosomal proteins [sup- at relatively low levels in the former and high levels in the latter. porting information (SI) Fig. S1]. This assay entailed monitoring In a recent proteomics study of lysosomal proteins in rat tissues, the phosphorylation state of endogenous lysosomal enzymes in we made the observation that the distribution of ACP5 fits this N1E-115 cells, a neuroblastoma cell line that does not efficiently profile (15). This finding is also consistent with the measurement dephosphorylate lysosomal proteins. Briefly, N1E-115 cells were of a relative abundance of ACP5 transcripts in mice by Northern cultured in the presence of an analyte [e.g., tissue extract or blotting (16) and in mice, rats, and humans found by using the fractions from a purification procedure for the dephosphoryla- Unigene EST expression profiling tool (www.ncbi.nlm.nih.gov/ tion activity (Fig. S2)]. The phosphorylation state of lysosomal sites/entrez?dbϭunigene). Given that ACP5 has also been enzymes was determined by comparing the activities of lysoso- demonstrated to dephosphorylate Man-6-P glycoproteins in mal enzymes that were retained on immobilized MPR affinity vitro (6), we therefore decided to investigate this enzyme columns with the total activities, as previously described (9). further as a candidate for the enzyme that dephosphorylates Approximately 70% of each lysosomal hydrolase retains Man- lysosomal proteins. 6-P in N1E-115 cells (data not shown). Culturing the cells in the presence of a mouse liver differential centrifugation fraction Dephosphorylation of Man-6-P Glycoproteins in N1E-115 Cells by enriched in lysosomes resulted in a significant increase in ACP5. If ACP5 were responsible for the removal of the Man-6-P dephosphorylation of cellular lysosomal activities, and this was recognition marker, increased levels of this enzyme should result essentially complete after7h(Fig. S3A). There was a lag period in increased dephosphorylation of lysosomal proteins. We tested of 1 h, which was presumably the time period required for the this prediction by overexpressing ACP5 in N1E-115 neuroblas- endocytosis and delivery of the enzyme that removes Man-6-P to toma cells. As a control, we also transfected these cells with the compartment in which dephosphorylation takes place. De- another lysosomal phosphatase, ACP2, that is not thought to be phosphorylation of endogenous lysosomal enzymes depended involved in the removal of Man-6-P (13). We identified trans- on the amount of liver extract added (Fig. S3B). fected cells by visualization of the cotransfection marker GFP Two lines of investigation indicate that the activity present in and visualized Man-6-P glycoproteins with biotinylated sCI- the liver extracts is physiologically relevant to lysosome function MPR (Fig. 1). In cells transfected with GFP alone (data not and does not reflect some other stress condition or signal shown) or cotransfected with GFP and ACP2, Man-6-P glyco- transduction pathway that indirectly transforms the N1E-115 proteins were detected at similar levels as observed in untrans-

cells into a dephosphorylation-competent state. fected N1E-115 cells (Fig. 1 Upper). However, cells transfected CELL BIOLOGY First, the liver contains high levels of the dephosphorylation with ACP5 showed significantly reduced levels of Man-6-P activity compared with the brain. Previous studies from our glycoproteins compared with untransfected cells (Fig. 1 Lower). laboratory (9) indicate that the liver contains higher levels of These results indicate increased removal of the Man-6-P recog- lysosomal enzyme activities than the brain, but the proportion of nition marker when ACP5 is overexpressed in a cell-based these enzymes that contain Man-6-P is much lower. From this model. observation, it is reasonable to predict that the activity that removes Man-6-P activity is high in the liver and low in the brain. Glycoproteins Containing Man-6-P in ACP5 Deficient Mice. To inves- We tested this prediction by culturing N1E-115 cells with tigate the potential role of ACP5 in the dephosphorylation of equivalent amounts of tissue extracts from the liver and the lysosomal proteins under physiological conditions, we examined

Sun et al. PNAS ͉ October 28, 2008 ͉ vol. 105 ͉ no. 43 ͉ 16591 Downloaded by guest on September 25, 2021 GFP Man6-P proteins 1.5 Brain

1.0 ACP2

0.5

ACP5 0.0 1.5 Liver

1.0 Fig. 1. Expression of recombinant ACP5 in mouse neuroblastoma cells. N1E-115 cells were cotransfected with plasmids expressing ACP2 or ACP5 and GFP as a marker for transfected cells (also see arrows, Right). Images were 0.5 obtained at 40ϫ magniﬁcation. Enzyme Activity, fraction of wildtype control wildtype of fraction Activity, Enzyme 0.0 how the loss of this protein affected levels of Man-6-P glyco- TPPI DPPI proteins in an ACP5-knockout mouse model. Should ACP5 play DPPII a physiological role in the removal of this modification, levels of legumain -fucosidase cathepsin B α -glucosidase β -mannosidase -mannosidase

Man-6-P glycoproteins should be increased in its absence. Sol- -galactosidase β -glucuronidase α β β -hexosaminidase

uble extracts were prepared from tissues, and Man-6-P glycop- β roteins were detected with a radiolabeled Man-6-P receptor Fig. 3. Lysosomal enzyme activities in ACP5-deﬁcient mice. Activities of 12 affinity probe after fractionation of proteins by SDS/PAGE and lysosomal enzymes were measured in brain and liver from Acp5ϩ/ϩ and transfer to nitrocellulose (17). Acp5Ϫ/Ϫ mice. Results were obtained from 2 animals of each genotype and Loss of ACP5 had little or no significant effect on the levels expressed as activity in Acp5 mutant tissue normalized to wild-type control or pattern of Man-6-P glycoproteins in brain and heart (Fig. 2). tissue with the line at y ϭ 1 representing control activity. Mean activities are This finding is consistent with the fact that these tissues normally shown with error bars representing the range. express low levels of ACP5. However, loss of ACP5 had a dramatic effect on levels of Man-6-P glycoproteins in tissues that normally express higher levels of this enzyme. In the knockout between these possibilities, we measured the activity of 12 mice compared with controls, the patterns of Man-6-P glyco- representative lysosomal enzymes in tissues (brain, liver, spleen, proteins were similar in such tissues but levels were clearly kidney, heart, and serum) from ACP5-deficient and wild-type increased in liver, kidney, spleen, testis, lung, and small intestine control mice (Fig. 3 and Fig. S6). For most of the enzymes (Fig. 2). Total levels of Man-6-P glycoproteins were strikingly measured, there was little or no alteration in activity in the absence of ACP5. In cases where there was an increase in high in the spleen and testis in mutant mice. activity, elevations were modest (less than Ϸ1.5-fold) and did not appear sufficient to account for the increased levels of Man-6-P Lysosomal Enzyme Activities in ACP5-Deficient Mice. The finding that glycoproteins. These results suggest that the increase in levels of levels of Man-6-P glycoproteins were elevated in the absence of Man-6-P glycoproteins in the absence of ACP5 cannot be ACP5 is consistent with a role for this enzyme in removal of the attributed to a generalized increase in lysosomal enzyme activ- Man-6-P recognition marker. In this case, the proportion of a ities. Interestingly, ␣-L-fucosidase was significantly decreased in given lysosomal protein that contains Man-6-P would be in- all of the ACP5-deficient tissues tested. creased. However, an alternative possibility is that the increase in Man-6-P glycoproteins in the absence of ACP5 may simply Dephosphorylation of Lysosomal Proteins in ACP5-Deficient Mice. reflect an elevation in lysosomal protein expression with the Our results suggest that removal of the Man-6-P recognition proportion of proteins that contain the Man-6-P modification marker is impaired in the ACP5-deficient mice and in this case, being the same in the mutant and control mice. To distinguish the relative proportion of a given lysosomal protein that retains the targeting modification will be increased. To address this

intestine possibility directly, we conducted analytical affinity purification small spleen kidney brain heart testis lung liver of Man-6-P glycoproteins on immobilized sCI-MPR and mea- Tissue: sured the proportion of lysosomal activities that contain Man- ACP5: + - + - + - + - + - + - + - + - 6-P as the fraction that was specifically eluted from the column 112 compared with the total activity. We measured 4 representative lysosomal enzyme activities, 3 of which contain Man-6-P (␤- MW (kD) ␤ ␤ 59 galactosidase, -hexosaminidase, and -glucuronidase) and a control activity (␤-glucosidase, also called glucocerebrosidase) that does not contain Man-6-P and is targeted to the lysosome

30 by an independent pathway (18). In the liver, the Man-6-P modification is normally efficiently removed and levels of the 25 Man-6-P containing lysosomal proteins containing this modification were very low in the control mice (Fig. 4). However, in the absence of ACP5, the proportion of these enzymes that con- Fig. 2. Levels of Man-6-P glycoproteins in tissues from ACP5-deﬁcient mice. ␤ Twenty microgram protein equivalents of each tissue extract were fraction- tained Man-6-P was elevated greatly. Levels of -glucosidase ated by SDS/PAGE, transferred to nitrocellulose, and probed with 2 nM that were specifically purified on the immobilized sCI-MPR were 125I-CI-MPR to speciﬁcally detect glycoproteins containing Man-6-P. Tissue unchanged. In the brain, where the Man-6-P modification is extracts were prepared from Acp5Ϫ/Ϫ and Acp5ϩ/ϩ 129SvEv mice. retained, the proportion of each activity represented by the

16592 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0807472105 Sun et al. Downloaded by guest on September 25, 2021 TPPI Man6-P proteins Merge 0.6 Brain A

0.4 Acp5+/+

0.2

0.0 Acp5-/- 0.6 Liver

0.4 B TPPI Man6-P proteins Merge

0.2

Fraction Mannose 6-PhosphorylatedFraction Acp5+/+

0.0

Acp5-/- -glucosidase β -galactosidase -glucuronidase β β -hexosaminidase β

Fig. 4. Phosphorylation state of lysosomal proteins in control and ACP5- Fig. 5. Localization of Man-6-P glycoproteins in ACP5-deficient mice. TPPI ϩ/ϩ deficient mouse brain and liver. Tissue homogenates from Acp5 (open and Man-6-P glycoproteins were detected in the brains (A) and livers (B)of Ϫ/Ϫ symbols) and Acp5 (filled symbols) mice were fractionated by affinity wild-type and mutant mice. In the merged images, the signals from GFP and chromatography on immobilized sCI-MPR (9). Enzyme activities attributable Man-6-P proteins are in green and red, respectively, with yellow indicating to the Man-6-P glycoforms were calculated as the activity in the Man-6-P colocalization. Images were obtained at 63ϫ magnification. eluate/total recovered activity (flow through/washϩglucose-6 phosphate eluateϩMan-6-P eluate). Bars represent mean of measurements from 2 animals of each genotype and error bars represent the range. compelling hypothesis is that ACP5 itself is directly responsible for the removal of Man-6-P targeting modification in many cell Man-6-P-containing form was unchanged in the absence of types. ACP5. These results indicate that the increased levels of Man- An earlier study (14) concluded that neither ACP5 nor ACP2 6-P glycoproteins measured in the ACP5-deficient mice do not was essential for dephosphorylating lysosomal proteins. This reflect an increase in lysosomal protein synthesis but result from conclusion was based on the observation that the rate of decreased removal of the targeting modification. dephosphorylation of exogenously added Man-6-P-containing arylsulfatase A appeared similar after endocytosis by wild-type Localization of Man-6-P Glycoproteins in ACP5-Deficient Mice. We or Acp2Ϫ/Ϫ/Acp5Ϫ/Ϫ embryonic murine fibroblasts. It is possible investigated the location of Man-6-P glycoproteins in the liver that there is a compensatory dephosphorylation activity in and brain from ACP5-deficient mice by using biotinylated cultured fibroblasts that is not present in the tissues analyzed sCI-MPR as a specific affinity probe. As a lysosomal marker, we here. Alternatively, endocytosed arylsulfatase A may be deliv- also performed immunodetection for a known lysosomal pro- ered to a dephosphorylation-incompetent or -deficient lysoso- tein, TPPI. There was no significant difference in the levels of mal/endosomal compartment (8). It will therefore be of interest Man-6-P glycoproteins between the ACP5-deficient and wild- to determine whether the dephosphorylation of lysosomal pro- type brain sections (Fig. 5), which is consistent with our previous teins depends on their route of delivery to the lysosome (e.g., findings (Fig. 2). Man-6-P glycoproteins were not detectable in endocytosis or intracellular targeting). the livers of wild-type mice. However, in the absence of ACP5, A variety of biological functions has been assigned to ACP5 Man-6-P glycoprotein levels were clearly elevated. In both the (reviewed in ref. 19), including roles in iron metabolism and brain and liver, TPPI staining was unaffected by the loss of immune function, but it is best known for its participation in ACP5. Superimposed images showed that Man-6-P glycopro- skeletal development and collagen metabolism and its impor- teins colocalized with TPPI in both the brains and livers of tance in these processes are clear from studies of the ACP5- ACP5-deficient mice. These results indicate that Man-6-P gly- deficient mutant mouse (20, 21). However, the observation that coproteins accumulate within the lysosomal compartment in the ACP5 is expressed at different levels in mammalian tissues and absence of ACP5. that its expression appears to be related to the Man-6- Discussion phosphorylation status of lysosomal proteins suggests that ACP5 may have more widespread physiological function. It is also In this study, we have used 2 parallel approaches to demonstrate CELL BIOLOGY that ACP5 is essential for removal of the Man-6-P recognition worth noting that there is evidence for multiple populations of marker from lysosomal proteins. First, in mice that are deficient lysosomes within a cell that differ in their respective abilities to in ACP5, we show that levels of Man-6-P glycoproteins are remove the Man-6-P modification (8). Distribution of hydrolases dramatically elevated in tissues that normally contain low levels between these compartments appears to be regulated by serum of these proteins. Second, in a murine neuroblastoma cell line factors (7). The precise significance of the presence of both that retains the Man-6-P modification on lysosomal proteins, phosphorylation-competent and -incompetent lysosomes within levels of Man-6-P glycoproteins are clearly reduced when ACP5 the cell is not clear but this intracellular compartmentalization is overexpressed. Given that ACP5 has been reported to de- may be intricately linked to ACP5 expression, processing, or phosphorylate Man-6-P glycoproteins in vitro (6), the most location.

Sun et al. PNAS ͉ October 28, 2008 ͉ vol. 105 ͉ no. 43 ͉ 16593 Downloaded by guest on September 25, 2021 Why the Man-6-P modification is removed in some cell types 0.1% Triton X-100, 5 mM ␤-glycerophosphate, and 1 mM EDTA] were added, (e.g., hepatocytes) but retained in others (e.g., neurons) and samples were homogenized on ice by using a Polytron homogenizer remains an intriguing question. Studies using cultured fibro- (Brinkmann Instruments). The resulting homogenate was cleared by centrif- blasts suggest that retention of the Man-6-P modification may ugation at 14,000 rpm at 4 °C for one hour in a BHG Hermle Z229 microcen- trifuge and supernatants transferred to fresh tubes. Protein concentration allow lysosomal enzymes to be anchored at the plasma mem- was measured by using Bio-Rad protein assay (Bio-Rad LaboratoriesInvitro- brane by interaction with the CI-MPR, allowing for their gen), transferred to nitrocellulose membrane, and probed with 125I labeled extracellular function in a locally concentrated and spatially sCI-MPR receptor as described previously (17). Radioactive signal was de- controlled manner (22, 23). In the brain, it is conceivable that tected by using a PhosphorImager (GE Healthcare) and quantified by using lysosomal enzymes tethered at the cell surface may play an ImageQuant 5.2 software (Molecular Dynamics). important role in axon growth and neuronal migration. Con- versely, retention of the Man-6-P modification may have a Enzyme Activity Measurements. Fluorescence was measured with a CytoFluor protective role in some cell types. In most cell types, lysosomal 4000 multiwell plate reader (PerSeptive Biosystems) with excitation at 360 nm proteins are largely retained within this compartment after and emission at 460 nm. Measurements from different dilutions were aver- intracellular targeting, which limits potential deleterious ef- aged after correction for dilution factor. Enzyme activity measurements were fects associated with inappropriate release of hydrolases. conducted by using fluorescent substrates either as described (9) or with the following protocols. Individual reactions conditions for each enzyme were as However, in some cell types that are engaged in extensive follows (enzyme, substrate/buffer/incubation time): ␤-glucosidase,5mM vesicular trafficking (e.g., neurons), one unintended conse- 4-MU-␤-glucoside/acetate, pH 5.0, 0.25% sodium taurocholate/60min; ␣-man- quence may be inappropriate release of lysosomal hydrolases nosidase, 1 mM 4-MU-␣-mannoside/acetate, pH 4.0/60 min; ␤-glucuronidase, into the extracellular milieu. In such cases, retention of the 1 mM 4-MU-␤-glucuronide/acetate, pH 4.0/30 min; ␣-fucosidase, 0.2 mM Man-6-P modification may facilitate the rapid reuptake of 4-MU-␣-fucoside/citrate, pH 5.0/120 min; cathepsin B, 0.2 mM Z-Arg-Arg-AMC/ secreted lysosomal proteins and their delivery to the lysosome acetate with 5 mM DTT and 2.5 mM EDTA, pH 6.0/30 min; dipeptidylpeptidase via CI-MPR-mediated endocytosis. I, 0.5 mM Gly-Arg-AMC/acetate with 5 mM DTT and 2.5 mM EDTA, pH 5.5/60 Alterations in the lysosomal system accompany and may play min; dipeptidylpeptidase II, 0.5 mM Lys-Ala-AMC/acetate, pH 5.5/120 min; a mechanistic role in widespread disorders, including cancer. tripeptidylpeptidase I, 0.25 mM Ala-Ala-Phe-AMC/acetate pH 4.5/60 min; and legumain, 0.2 mM Z-Ala-Ala-Asn-AMC/citrate-phosphate with 1 mM DTT, pH Studies on human breast cancer specimens revealed a marked 5.8/120 min. Buffers consisted of 0.1 M citric acid or 0.1 M sodium acetate, increase in the levels of Man-6-P containing glycoproteins in adjusted to the indicated pH by using sodium hydroxide or acetic acid, approximately 1/3 of tumors, and this was associated with the respectively, and contained 0.1% Triton X-100. For all reactions, substrates more aggressive cases (24). This finding does not appear to were prepared as stocks in dimethyl sulfoxide that were stored at Ϫ20 °C and simply reflect increased lysosomal enzyme synthesis but proba- diluted shortly before use. bly reflects decreased dephosphorylation. Retention of the Man-6-P modification in malignant cells could allow presenta- Immunohistochemistry. Tissues from 8-week-old mice were collected and fixed tion of these hydrolases on the cell surface via interaction with overnight in Bouin’s reagent (Electron Microscopy Sciences), then treated with the CI-MPR, and this could play a role in pathogenic processes 15%, followed by 30% sucrose. Tissues were then embedded in Tissue-Tek ␮ such as metastasis (22, 23, 25). It may be relevant in this respect OCT medium (Sakura) and frozen in liquid nitrogen. Serial sections (10 m) that ACP5 has been reported to be down-regulated in some were cut and mounted on glass slides. Sections were incubated with the rabbit polyclonal antisera Rb72–5 (to detect the lysosomal protein TPPI) and then forms of hepatocarcinoma (26). probed with Alexa Fluor 488 goat anti IgG at a dilution of 1:400. Man-6-P In summary, we have identified a key enzyme responsible for proteins were probed with biotinylated sCI-MPR as described (10) by using removal of the Man-6-P targeting modification from lysosomal Alexa Fluor 555 strepavidin at a dilution of 1:1,000 as a secondary detection proteins. This identification should allow in-depth investigation reagent. Slides were mounted by using Aqua Poly/Mount (Polysciences). Flu- regarding the biomedical importance for removal and retention orescence was detected by using a Zeiss LSM 410 confocal microscope. of this posttranslational modification. Transfection of N1E-115 Cells. N1E-115 cells were grown to 50% confluence in Materials and Methods 4-well chamber slides and were cotransfected with 0.05 ␮g of pCS2-GFP and 0.5 ␮g of either pCMV-SPORT6-ACP5 or pCMV-SPORT6-ACP2 using lipo- Materials. N1E-115 mouse neuroblastoma cells (27) and eukaryotic pCMV- fectamine (Invitrogen). After 24 h, cells were fixed with Bouin’s reagent for 15 SPORT6 expression plasmids encoding murine ACP5 (MGC:7956) and ACP2 min and then permeabilized with PBS containing 0.5% Triton X-100. Probing (MGC:31030) were obtained from the American Type Culture Collection. The was as described above except that anti-GFP antibody (1:1,000 dilution) was eukaryotic expression plasmid pCS2-GFP was constructed and kindly provided used instead of anti-TPPI antibody. Images were captured by using a Nikon by P. Matteson (University of Medicine and Dentistry of New Jersey, Piscat- Eclipse E600 fluorescence microscope (Nikon Instruments) at magnification away, NJ). Rabbit polyclonal antisera Rb72–5 raised against tripeptidylpepti- 40ϫ. Images were taken by using a SPOT software version 4.5.9.9 (Diagnostic dase 1 (TPPI) and biotinylated and 125I-labeled derivatives of the soluble Instruments). cation-independent mannose 6-phosphate receptor (sCI-MPR) were described previously (17, 24, 28). Rabbit anti-GFP was from Molecular Probes. Goat anti-rabbit IgG conjugated with Alexa Fluor 488 and streptavidin conju- Other. Determination of the phosphorylated proportion of lysosomal proteins gated with Alexa Fluor 555 were purchased from Invitrogen. ACP5- was conducted by analytical purification of Man-6-P glycoproteins as de- deficient-knockout mice in a 129SvEv genetic background were described scribed previously (9). The cell-based dephosphorylation bioassay is described previously (20). in Fig. S1. Subcellular fractionation is described in Fig. S2.

Detection of Man-6-P Glycoproteins in Mouse Tissue Extracts. Twelve-week-old ACKNOWLEDGMENTS. We thank Mukarram El-Banna and Istvan Sohar for their help and advice with this study and Drs. Paul Matteson, James Millonig, male wild-type and ACP5-knockout mice were killed, and tissues were rapidly and Loren Runnels (University of Medicine and Dentistry of New Jersey– Ϫ frozen in liquid nitrogen and stored at 80°C before use. Frozen tissue Robert Wood Johnson Medical School) for their valuable advice with micro- powders were prepared by using a Bessmann tissue pulverizer (Thermo Fisher scopic imaging. This work is supported in part by National Institutes of Health Scientiﬁc), and 100–150 mg of each were transferred to a 1.5-ml microcentri- Grant DK054317 (to P.L.) and Fonds de la Recherche Fondamentale Collective fuge tube. Ten volumes (wt/vol) homogenization buffer [PBS (PBS) containing Grant 2.4543.08 (to M.J.).

1. Holtzman E (1989) Lysosomes (Plenum Press, New York). 4. Gabel CA, Goldberg DE, Kornfeld S (1983) Identification and characterization of cells 2. Kornfeld S, Sly WS (2001) in The Metabolic and Molecular Bases of Inherited Disease, deficient in the mannose 6-phosphate receptor: Evidence for an alternate pathway for eds Scriver CR, et al. (McGraw–Hill, New York), pp 3469–3482. lysosomal enzyme targeting. Proc Natl Acad Sci USA 80:775–779. 3. Gabel CA, Goldberg DE, Kornfeld S (1982) Lysosomal enzyme oligosaccharide phos- 5. Gabel CA, Foster SA (1986) Mannose 6-phosphate receptor-mediated endocytosis of phorylation in mouse lymphoma cells: Specificity and kinetics of binding to the acid hydrolases: Internalization of ␤-glucuronidase is accompanied by a limited de- mannose 6-phosphate receptor in vivo. J Cell Biol 95:536–542. phosphorylation. J Cell Biol 103:1817–1827.

16594 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0807472105 Sun et al. Downloaded by guest on September 25, 2021 6. Bresciani R, Von Figura K (1996) Dephosphorylation of the mannose-6-phosphate 17. Valenzano KJ, Kallay LM, Lobel P (1993) An assay to detect glycoproteins that contain recognition marker is localized in later compartments of the endocytic route. Identi- mannose 6-phosphate. Anal Biochem 209:156–162. fication of purple acid phosphatase (uteroferrin) as the candidate phosphatase. Eur 18. Reczek D, et al. (2007) LIMP-2 is a receptor for lysosomal mannose-6-phosphate- J Biochem 238:669–674. independent targeting of ␤-glucocerebrosidase. Cell 131:770–783. 7. Einstein R, Gabel CA (1989) Serum factors alter the extent of dephosphorylation of 19. Hayman AR (2008) Tartrate-resistant acid phosphatase (TRAP) and the osteoclast/ ligands endocytosed via the mannose 6-phosphate/insulin-like growth factor II recep- immune cell dichotomy. Autoimmunity 41:218–223. tor. J Cell Biol 109:1037–1046. 20. Hayman AR, et al. (1996) Mice lacking tartrate-resistant acid phosphatase (Acp 5) have 8. Einstein R, Gabel CA (1991) Cell- and ligand-specific dephosphorylation of acid hydro- disrupted endochondral ossification and mild osteopetrosis. Development 122:3151– lases: Evidence that the mannose 6-phosphatase is controlled by compartmentaliza- 3162. tion. J Cell Biol 112:81–94. 21. Roberts HC, et al. (2007) Altered collagen in tartrate-resistant acid phosphatase 9. Sleat DE, Sohar I, Lackland H, Majercak J, Lobel P (1996) Rat brain contains high levels (TRAP)-deficient mice: A role for TRAP in bone collagen metabolism. Calcif Tissue Int of mannose-6-phosphorylated glycoproteins including lysosomal enzymes and palmi- 80:400–410. toyl-protein thioesterase, an enzyme implicated in infantile neuronal lipofuscinosis. 22. Brauker JH, Roff CF, Wang JL (1986) The effect of mannose 6-phosphate on the J Biol Chem 271:19191–19198. turnover of the proteoglycans in the extracellular matrix of human fibroblasts. Exp Cell 10. Jadot M, et al. (1999) Subcellular localization of mannose 6-phosphate glycoproteins Res 164:115–126. in rat brain. J Biol Chem 274:21104–21113. 23. Roff CF, Wozniak RW, Blenis J, Wang JL (1983) The effect of mannose-6-phosphate on 11. Waheed A, et al. (1988) Human lysosomal acid phosphatase is transported as a the turnover of cell surface glycosaminoglycans. Exp Cell Res 144:333–344. transmembrane protein to lysosomes in transfected baby hamster kidney cells. EMBO 24. Sleat DE, Chen TL, Raska K, Jr, Lobel P (1995) Increased levels of glycoproteins con- J 7:2351–2358. taining mannose 6-phosphate in human breast carcinomas. Cancer Res 55:3424–3430. 12. Baumbach GA, Saunders PT, Bazer FW, Roberts RM (1984) Uteroferrin has N- 25. Wood RJ, Hulett MD (2008) Cell surface-expressed cation-independent mannose asparagine-linked high-mannose-type oligosaccharides that contain mannose 6-phos- 6-phosphate receptor (CD222) binds enzymatically active heparanase independently phate. Proc Natl Acad Sci USA 81:2985–2989. of mannose 6-phosphate to promote extracellular matrix degradation. J Biol Chem 13. Bresciani R, Peters C, von Figura K (1992) Lysosomal acid phosphatase is not involved 283:4165–4176. in the dephosphorylation of mannose 6-phosphate containing lysosomal proteins. Eur 26. Chan KY, et al. (2005) Transcriptional profiling on chromosome 19p indicated frequent J Cell Biol 58:57–61. down-regulation of ACP5 expression in hepatocellular carcinoma. Int J Cancer 14. Suter A, et al. (2001) Overlapping functions of lysosomal acid phosphatase (LAP) and 114:902–908. tartrate-resistant acid phosphatase (Acp5) revealed by doubly deficient mice. Devel- 27. Amano T, Richelson E, Nirenberg M (1972) Neurotransmitter synthesis by neuroblas- opment 128:4899–4910. toma clones (neuroblast differentiation-cell culture-choline acetyltransferase- 15. Sleat DE, Della Valle MC, Zheng H, Moore DF, Lobel P (2008) The mannose 6-phosphate acetylcholinesterase-tyrosine hydroxylase-axons-dendrites). Proc Natl Acad Sci USA glycoprotein proteome. J Proteome Res 7:3010–3021. 69:258–263. 16. Hayman AR, Bune AJ, Bradley JR, Rashbass J, Cox TM (2000) Osteoclastic tartrate- 28. Lin L, Lobel P (2001) Expression and analysis of CLN2 variants in CHO cells: Q100R resistant acid phosphatase (Acp 5): Its localization to dendritic cells and diverse murine represents a polymorphism, and G389E and R447H represent loss-of-function muta- tissues. J Histochem Cytochem 48:219–228. tions. Hum Mutat 18:165. CELL BIOLOGY

Sun et al. PNAS ͉ October 28, 2008 ͉ vol. 105 ͉ no. 43 ͉ 16595 Downloaded by guest on September 25, 2021