Rat brain contains high levels of mannose-6-phosphorylated glycoproteins including lysosomal enzymes and palmitoyl-protein thioesterase, an enzyme implicated in infantile neuronal lipofuscinosis.

Mannose 6-phosphate (Man-6-P) is a posttranslational carbohydrate modification typical of newly synthesized acid hydrolases that signals targeting from the Golgi apparatus to the lysosome via Man-6-P receptors (MPRs). Using iodinated cation independent MPR as a probe in a Western blot assay, we surveyed levels of Man-6-P glycoproteins in a number of different rat tissues. Considerable variation was observed with respect to total amounts and types of Man-6-P glycoproteins in the different tissues. Brain contained 2-8-fold more Man-6-P glycoproteins than other tissues, with relative abundance being brain ≫ testis ≈ heart > lung ≈ kidney ≈ ovary ≈ spleen > skeletal muscle ≈ liver ≈ serum. Analysis of 16 different lysosomal enzyme activities revealed that brain contains lower activities than other tissues which suggested that decreased removal of Man-6-P results in increased levels of Man-6-P glycoproteins. This was directly demonstrated by comparing activities of phosphorylated lysosomal enzymes, purified by immobilized MPR affinity chromatography, with total activities. The phosphorylated forms accounted for a considerable proportion of the MPR-targeted activities measured in brain (on average, 36.2%) but very little in lung, kidney, and liver (on average, 5.5, 2.3, and 0.7%, respectively). Man-6-P glycoproteins were also isolated from rat brain by MPR affinity chromatography on a preparative scale. Of the 18 bands resolvable by SDS-polyacrylamide gel electrophoresis, seven bands were NH2-terminally sequenced and identified as the known lysosomal enzymes cathepsin L, cathepsin A, cathepsin D, α-galactosidase A, arylsulfatase A, and α-iduronidase. One of the major Man-6-P glycoproteins was identified as palmitoyl protein thioesterase, which was not previously thought to be lysosomal. This finding raises important questions about the cellular location and function of palmitoyl protein thioesterase, mutations in which result in the neurodegenerative disorder, infantile neuronal ceroid lipofuscinosis.

Lysosomes contain more than 50 hydrolytic enzymes which act in concert to degrade a wide range of macromolecules. The N-linked oligosaccharides of many newly synthesized lysosomal enzymes are modified to contain mannose 6-phosphate (Man-6-P). 1 This moiety is specifically recognized by Man-6-P receptors (MPRs) which direct the vesicular transport of newly synthesized lysosomal enzymes from the Golgi apparatus to the lysosome (reviewed in Refs. [1][2][3]. In most cultured cells, Man-6-P is a transient targeting signal, being rapidly removed when newly synthesized lysosomal enzymes reach the lysosome (4). This is not, however, always the case. In CI-MPR-deficient mouse cell lines, acid hydrolases that reach the lysosome retain the Man-6-P marker (3,5). Furthermore, dephosphorylation of Man-6-P glycoproteins in a mouse L cell line which contains the CI-MPR varies under different in vitro culture conditions (4,6). These cells dephosphorylate normally when grown at low density, but accumulate Man-6-P glycoproteins when grown at high density or in the absence of serum. In addition, Man-6-P glycoproteins endocytosed by these cell lines in the absence of serum were restricted to a subset of lysosomes. It has been suggested (3) that there are actually two distinct lysosomal populations which are differentiated by the presence or absence of Man-6-phosphatase activity and that these may have specialized functions.
It is possible that such a Man-6-phosphatase-deficient compartment could actually represent a functionally specialized lysosome with a characteristic complement of lysosomal enzymes and a restricted cellular distribution. In this report, we have investigated this possibility by surveying the steady state levels of Man-6-P glycoproteins in different rat tissues and found increased levels in rat brain. This accumulation reflects a higher proportion of lysosomal enzymes containing Man-6-P, which is compatible with targeting of phosphorylated lysosomal enzymes to a Man-6-phosphatase-deficient compartment in brain. We have also identified the major Man-6-P glycoproteins in brain, finding a number of known lysosomal enzymes and one protein not previously known to be Man-6-phosphorylated.

Detection of Man-6-P Glycoproteins in Rat Tissue Extracts-Tissues
were obtained from 6 -8-week-old Sprague-Dawley rats and were frozen on dry ice immediately after dissection and stored at Ϫ70°C prior to use. Frozen tissues were pulverized using a Bessmann homogenizer, and protein extracts were prepared from powdered samples essentially * This study was supported by National Institutes of Health Grant DK45992 and National Science Foundation Grant DCB-9118681. 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.
Measurement of Enzyme Activities-Frozen rat tissues were pulverized and stored as powders at Ϫ70°C before use. Samples were thawed on ice, placed in 50 volumes (w/v) of 0.15 M NaCl, 0.1% Triton X-100 and homogenized with a Brinkmann Polytron homogenizer. A soluble supernatant was prepared by centrifugation at 12,000 ϫ g for 25 min at 4°C. Glycosidase, phosphatase, and sulfatase activities were measured using 4-methylumbelliferyl (4-MU) substrates with modifications of the conditions described by Kolodny and Mumford (13). Protease assays using amino-4-methylcoumarine (AMC) substrates were as described elsewhere (14,15) adjusted to the indicated pH using sodium hydroxide, acetic acid, or HCl, respectively, and contained 0.1% Triton X-100. Substrates were purchased from Sigma and were prepared freshly in reaction buffer (4-MU-phosphate and 4-MU-sulfate) or as stocks in dimethyl sulfoxide (remaining substrates) that were added to reaction buffer immediately prior to assay. Samples added to substrate solutions after addition of termination buffer were used as blanks. Fluorescent reaction products were measured using a CytoFluor II (PerSeptive Biosystems, Framingham, MA) fluorescence multiwell plate reader with exitation at 360 nm and emission 460 nm. Cathepsin D was measured using 1 mg/ml hemoglobin as a substrate in acetate, pH 3.5, for 120 min (16), and trichloroacetic acid-soluble degradation products measured by the method of Lowry et al. (8). Enzyme activities were linear with respect to input sample.
Determination of Enzyme Phosphorylation State-Man-6-P glycoproteins were purified on an analytical scale to determine the relative proportion of lysosomal enzyme activities represented by the phosphorylated forms. 150 mg of frozen tissue were homogenized using a Polytron homogenizer in 1.5 ml of 50 mM phosphate buffer, pH 7.2, containing 0.15 M NaCl and 0.1% Triton X-100 (homogenization buffer, HB). A postnuclear supernatant was prepared by centrifugation at 40,000 ϫ g for 30 min and loaded onto a 1-cm 3 bed volume column of HB-equilibrated coupled sCI-MPR which was then washed with 6 ml of HB, 3 ml of HB containing 10 mM glucose 6-phosphate and specifically eluted with 6 ml of 10 mM Man-6-P in HB. Recovery of Man-6-P glycoproteins was monitored at each stage by SDS-PAGE and affinity blotting (7,12) as described earlier. Enzyme activities were determined on samples of the supernatant, pooled flow-through and washes, glucose 6-phosphate eluate, and Man-6-P eluate. Activities were determined as described above, except samples were diluted in HB and were normalized to contain equivalent glucose 6-phosphate and Man-6-P concentrations.
Preparative Isolation of Man-6-P Glycoproteins-Man-6-P glycoproteins were purified from rat brain on a preparative scale for identification by N-terminal sequence analysis. 28 brains (48 g total from Sprague-Dawley rats, 6 -8 weeks old) were homogenized using a Waring Blender in 300 ml of homogenization buffer 2 (HB2: 1 ϫ phosphatebuffered saline, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, 1.4 g/ml pepstatin, 1 mM sodium orthovanadate, and 5 mM ␤-glycerophosphate) for 30 s (low setting) and 30 s (high setting), then sonicated for three times for 10 s at 100 Watts. The homogenate was centrifuged at 1300 ϫ g for 5 min at 4°C, and the post nuclear supernatant was transferred to fresh tubes. The extract was then centrifuged at 100,000 ϫ g for 60 min at 4°C, and the supernatant was decanted. The pellet was reextracted with 100 ml of HB2 containing 0.1% Triton X-100 and centrifuged at 100,000 ϫ g for 60 min at 4°C. The high speed supernatants were pooled and applied to a HB2-equilibrated affinity column consisting of 30 ml of Affi-Gel-10 (Bio-Rad) coupled with bovine sCI-MPR (2 mg/cm 3 ). The column was washed with 3 column volumes of HB2, then mock eluted with 2 volumes of HB2 containing 5 mM glucose 6-phosphate. Man-6-P glycoproteins bound to the immobilized sCI-MPR were specifically eluted with 2 volumes of HB2 containing 5 mM Man-6-P. Fractions containing the highest protein content were pooled and concentrated using Centriprep-10 and Centricon-10 (Amicon) spin concentrators. The total yield of Man-6-P glycoproteins was 244 g.
NH 2 -terminal Sequencing-Affinity-purified proteins (20 g) were fractionated by 12.5% acrylamide SDS-PAGE and transferred to Immobilon P sq polyvinylidene difluoride membrane (Millipore) using a semidry transfer apparatus (Bio-Rad) as described previously (13,14). The blot was washed for 10 times for 30 s with HPLC grade water (Baker) to remove excess glycine, stained for 5 min with 0.01% Coomassie Blue in 40% methanol (HPLC grade, EM), 10% glacial acetic acid (Baker) and destained with several changes of HPLC grade water. Individual bands were excised and stored at Ϫ70°C prior to N-terminal sequence analysis. Samples were sequenced on an Applied Biosystems model 477A protein/peptide sequencer with an on-line Applied Biosystems model 120A analyzer. The apparent molecular weights of purified proteins are averages of two independent determinations from Coomassie Bluestained gels using calibrated prestained molecular weight markers (Life Technologies, Inc., high range).
Sequence Analysis-Sequence data were analyzed primarily using the TFasta and Find programs in the Genetics Computer Group Sequence Analysis Software Package, Version 7.2 (17).

Survey of Mannose 6-Phosphorylated Glycoproteins in Rat
Tissues-Man-6-P containing glycoproteins were detected in a number of rat tissues in a quantitative biochemical assay as described (7,12). Briefly, soluble extracts of rat tissues were normalized for protein content, fractionated by SDS-PAGE and transferred to nitrocellulose. Iodinated sCI-MPR was used as an affinity probe for Man-6-P glycoproteins and was detected using a PhosphorImager. Representative blots are shown in Fig. 1 and quantitative data are presented in Fig. 2. Man-6-P (10 mM) abrogated binding of the sCI-MPR to most of the immobilized proteins indicating that the probe was specifically recognizing Man-6-P glycoproteins. In contrast, interaction of the sCI-MPR with a doublet of proteins of ϳ28 and 30 kDa found in lung, spleen, and ovary extracts, or with ϳ35and ϳ45-kDa bands in skeletal muscle and heart, respectively, was not inhibited by Man-6-P (Fig. 1, panel A). However, addition of 350 nM unlabeled sCI-MPR to the probe mix did not decrease these signals, indicating that binding to these proteins is nonspecific (data not shown).
While all solid tissues examined contained Man-6-P glycoproteins, the content of the different rat tissues differs widely in both quantitative and qualitative respects. In terms of total levels (i.e. the sum of all Man-6-P glycoproteins in each extract), brain contained the highest amounts (7-8-fold higher than liver) of Man-6-P glycoproteins and small intestine, skeletal muscle, and liver the lowest. Bands of approximately 30, 36, and 45-50 kDa are particularly prominent in brain extracts. Both the pattern and levels of Man-6-P glycoproteins in male and female rats were similar ( Fig. 2; data not shown).
Levels of Man-6-P glycoproteins in rat serum were very low with no obvious protein species detected.
Activities of Lysosomal Enzymes in Rat Tissues-There are two possible explanations for the elevated levels of Man-6-P glycoproteins in brain. The proportion of some or all lysosomal enzymes containing Man-6-P could be increased in brain. Alternatively, the proportion of each lysosomal enzyme containing Man-6-P could be the same in different tissues but brain could contain higher overall levels of lysosomal enzymes. As a first step toward differentiating between these two models, we have measured the activities of a number of lysosomal and nonlysosomal enzymes in rat tissues. Specific activities of 16 lysosomal enzymes were measured, including acid phosphatase and ␤-glucosidase, both of which are synthesized as membranebound precursors and whose targeting is MPR-independent. Activities of the nonlysosomal proteases, thimet and alanine aminopeptidase, were also determined. Results are presented in Fig. 3.
In Fig. 4, the average relative activities of enzymes whose targeting is MPR-dependent and MPR-independent are shown. Overall, kidney, spleen, and ovary had the highest average relative enzyme activities, while brain and heart had the lowest. Interestingly, lysosomal enzyme activities in the different tissues tended toward inverse correlation with the levels of Man-6-P glycoproteins (Fig. 5). In most cases, this tendency was weak, if existent, but for some, e.g. ␣-mannosidase, cathepsin B, acid phosphatase, and ␤-glucuronidase, there is potential inverse correlation (p Ͻ 0.1). With the exception of acid phosphatase, MPR-independent or non-lysosomal enzymes showed no significant correlation with levels of Man-6-P glycoproteins in the different tissues.
It is clear from these results that brain does not contain higher lysosomal enzyme activities (thus presumably mass) than other tissues. Elevated levels of Man-6-P glycoproteins is therefore likely to reflect an increase in the proportion of each lysosomal enzyme containing Man-6-P. We tested this possibility directly by purifying Man-6-P containing proteins from different tissues by MPR-affinity chromatography and determining the proportion of total activity in each tissue represented by the phosphorylated forms. The percentage of each enzyme containing Man-6-P was determined in brain, liver, lung, and kidney and is shown in Fig. 6. As expected, little (Ͻ2.3% of total) binding to the column was observed for the negative controls (MPR independently targeted lysosomal enzymes and nonlysosomal enzymes). Importantly, and as predicted from Fig. 5, the percentage of each MPR-targeted enzyme activity binding to the column was considerably higher in brain (on average, 36.2%) compared to lung, kidney, and liver (on average, 5.5, 2.3, and 0.7%, respectively).
Purification of Man-6-P Glycoproteins from Rat Brain- Fig.  1 clearly demonstrates that total levels of Man-6-P glycoproteins are elevated in brain compared to other tissues, and the profile of brain Man-6-P glycoproteins is distinct. Similar results were also obtained using murine, bovine, and human brain samples (data not shown). As a first step toward determining the biological significance of these observations, we undertook a large scale purification of Man-6-P glycoproteins from rat brain with the aim of individually identifying each protein by microsequencing. Rat brain proteins were purified by affinity chromatography on immobilized sCI-MPR and specifically eluted with Man-6-P. Parallel samples were resolved by SDS-PAGE, and proteins were detected either by Coomassie Blue (CB) staining or affinity blotting (AB) using the iodinated sCI-MPR probe. An illustrative gel is shown in Fig. 7. The two profiles are similar, indicating that most of the purified proteins contain Man-6-P.  1. A survey of Man-6-P glycoproteins in rat tissues. 80-g protein equivalents of each extract were fractionated by SDS-PAGE, transferred to nitrocellulose, and probed with 2 nM iodinated sCI-MPR. Representative blots probed in the presence (left) and absence (right) of 10 mM Man-6-P are presented. Standards were 5-, 15-, and 50-l secretions from CI-MPR-deficient D9 cells.

FIG. 2. Quantitation of levels of Man-6-P glycoproteins in rat tissues.
Error bars represent standard errors of the mean from duplicate blots from tissue extracts from one female and one male rat. In the case of the gonads, two rats of the appropriate gender were used.
Two exceptions discussed below are bands 1 and 2, which are detectable by Coomassie staining but not by the sCI-MPR probe. (Band 1 is a low abundance species not readily apparent in Fig. 7 but clearly visible by Coomassie staining when larger quantities of protein were resolved for preparative purposes.) In addition, while the affinity blotting profile of the purified preparation (Fig. 7) is considerably sharper than that of the crude brain homogenate (Fig. 1), it shows a similar pattern, indicating that it is a good representation of the starting material. Sequence Analysis of Purified Man-6-P Glycoproteins-The purified Man-6-P glycoproteins from rat brain were fractionated by SDS-PAGE and transferred to Immobilon P sq polyvinylidene fluoride membrane (Millipore), and select individual bands were NH 2 -terminally sequenced. While some bands yielded a single sequence, others contained a mixture of multiple sequences, indicating incomplete resolution by SDS-PAGE. Nonetheless, with all sequence mixtures except one, it was possible to unambiguously identify a number of different proteins (Table I).
Band 1 was unambiguously identified as the cathepsin A (lysosomal protective protein) light chain. The protein sequence was identical to the predicted mouse sequence (precathepsin A residues 321-331) at 9 out of 10 positions (18). The amino terminus of band 1 corresponds to the known start of the human cathepsin A light chain (19). Finally, the apparent size of band 1 (ϳ19 kDa) is in excellent agreement with the reported size of the rat cathepsin A light chain (20 kDa) (18,20). It is noteworthy that band 1 was not detectable by activity blotting, consistent with the observation of others that only the heavy chain of cathepsin A contains Man-6-P (21).
Band 2 contained a single protein sequence which was iden-tical to rat cathepsin L (22,23). The sequence obtained here starts at residue 114 of prepro cathepsin L, which is the propeptide cleavage site (22). Note that size of band 2 (ϳ25 kDa) is consistent with the reported (24) size of single chain cathepsin L (23 to 24 kDa). One interesting finding is that although band 2 was isolated by affinity chromatography on immobilized sCI-MPR, it is not recognized by the iodinated sCI-MPR probe in the Western blot assay. There are two potential explanations for this. First, like band 1, band 2 may be associated with another Man-6-phosphorylated glycoprotein rather than contain Man-6-P itself. Second, band 2 may have relatively low affinity for the sCI-MPR such that it would be retained on the high density sCI-MPR column ([sCI-MPR] ϳ 5 M), but would not be readily detectable in the blotting assay ([sCI-MPR] ϳ 2 nM). We favor the latter possibility given that cathepsin L is known to have low affinity for the CI-MPR compared to other phosphorylated lysosomal enzymes (25,26). Band 3 (ϳ30 kDa) yielded a single amino-terminal sequence. No close matches were obtained by data base searching, so this at present remains unidentified.
Band 4 was unambiguously identified as the cathepsin A heavy chain. The protein sequence was identical to the predicted mouse sequence at 9 out of 9 positions and started at the signal sequence cleavage site (18). Furthermore, the apparent size of band 4 (ϳ31 kDa) is in good agreement with the reported size of 32 kDa for the mouse and human cathepsin A heavy chain (18). FIG. 5. Correlation between each enzyme activity and the levels of Man-6-P glycoproteins. Unshaded bars represent nonlysosomal enzymes or lysosomal enzymes whose targeting is MPR-independent. Average Pearson correlation coefficients were obtained by plotting enzyme activity values for each animal (Fig. 3) versus Man-6-P glycoproteins content (Fig. 2).
FIG. 6. Percentage of lysosomal enzyme activities represented by Man-6-phosphorylated forms in male rat brain, kidney, liver, and lung. The percentage of each activity represented by the phosphorylated form was calculated as the activity in the Man-6-P eluate/total recovered activity (flow through/wash ϩ glucose 6-phosphate eluate ϩ Man-6-P eluate). Average recovery of input activity was 83%.
FIG. 7. Affinity blotting and Coomassie staining of purified rat brain Man-6-P glycoproteins. The numbering used for the purified bands is indicated. Proteins were detected by Coomassie Blue (CB) or by affinity blotting with iodinated sCI-MPR (AB).
Band 5 was identified as arylsulfatase A by comparison with the human and mouse sequences (27,28). The rat protein sequence was identical to the invariant mouse and human sequences at 6 out of 6 positions, starting at the predicted signal sequence cleavage site (27). By SDS-PAGE, band 5 appears somewhat smaller (ϳ34 kDa) than the single chain (58 kDa) or two-chain (50 and 7 kDa) forms of arylsulfatase A (29). This suggests that the rat brain enzyme may contain additional COOH-terminal cleavage sites.
Band 6 was unambiguously identified as palmitoyl protein thioesterase (PPT). The protein sequence was identical to the published rat sequence at 14 out of 15 positions and started at the signal sequence cleavage site (30). Furthermore, the apparent size of band 6 (ϳ36 kDa) is in excellent agreement with the reported size of 37 kDa (31). Identification of PPT as a Man-6-P glycoprotein was surprising as this protein was not previously thought to have a lysosomal function or localization (see "Discussion").
Band 7 contained single-chain cathepsin D. Amino-terminal sequencing yielded a mixture which was identified as three forms of cathepsin D by comparison with the rat sequence (32). These sequences varied at their NH 2 termini around the propeptide cleavage site predicted from the known porcine and human sites (33,34). This suggests that either cleavage of the propeptide can occur at one of three adjacent amino acids, or cleavage occurs at a single site and is followed by exopeptidase digestion to remove one or two amino acids, generating three NH 2 termini. Interestingly, while the rat sequence is quite different from the human (35) in this region, this result parallels previous studies showing three NH 2 termini at corresponding positions of human cathepsin D (34). Cathepsin D exists as either a single-chain, 48-kDa protein that can be further processed to yield an amino-terminal 14-kDa light chain and a carboxyl-terminal 34-kDa heavy chain (33,36). The size of band 7 (ϳ45 kDa) indicated that it represents the single-chain form. Brain extracts and purified proteins were also immunoblotted for cathepsin D, confirming our sequence assignment for band 7 (data not shown).
Band 8 yielded a mixed sequence derived from three species. The first of these (band 8a, Table I) was identified as ␣-galactosidase A. The protein sequence was identical to the published mouse sequence at 9 out of 9 residues (37) and starts at the known propeptide cleavage site (38). In addition, the size of band 8a (ϳ47 kDa) agrees well with published studies reporting that ␣-galactosidase migrates as a diffuse band of ϳ47 kDa (39).
In addition, band 8 may also contain ␣-iduronidase. When the ␣-galactosidase residues are subtracted from the mixed sequence, the remaining sequence is identical to 6 out of 6 invariant residues in the published dog (40), mouse (41), and human (42) sequences. Interestingly, like cathepsin D, the ␣-iduronidase sequence has a heterogeneous NH 2 terminus with one form which starts at the known amino terminus of mature canine and human ␣-iduronidase (40,42) and another which starts at one amino acid before it. It is possible that the reported NH 2 terminus of ␣-iduronidase (Gln 26 ) actually represents an NH 2 -terminally clipped form, especially given that Leu 24 -Ala 25 is the most favorable predicted signal cleavage site for mouse and dog ␣-iduronidase (SigCleave) (17). While the apparent size of band 8 (ϳ47 kDa) is smaller than the reported intact dog and human ␣-iduronidase (68 and 74 kDa, respectively (40,42), smaller (52 kDa) degradation products of ␣-iduronidase have been reported (43). Band 8 may contain this, or a similar, protein. DISCUSSION We have demonstrated that rat brain contains considerably higher levels of Man-6-P glycoproteins than other tissues and organs. One possible explanation for this observation is that synthesis of lysosomal enzymes is higher in brain than in other tissues, resulting in detection of more Man-6-P glycoproteins in transit to the lysosome. In this report, however, we demonstrate that, while brain contains lower levels of lysosomal enzymes than most other tissues, it has a much higher proportion   containing Man-6-P.
Assuming that lysosomal enzymes are initially phosphorylated to similar extents in different tissues, changes in the proportion of lysosomal enzymes that contain Man-6-P are most likely to result from decreased dephosphorylation, which could reflect either decreased Man-6-phosphatase activity or targeting to a specialized phosphatase-deficient compartment. Support for the latter possibility comes from histochemical studies. Man-6-P glycoproteins are prominent in the cell bodies of multiple neurons throughout the nervous system (44), but do not colocalize with lysosome-associated membrane protein 1, a marker of "classical" lysosomes and late endosomes. The nature of this compartment and the function for the retention of the Man-6-P marker remain to be determined. However, it may be significant that brain expresses less CI-MPR than other tissues (45,46).
One important result of this study is the finding that not only is PPT Man-6-phosphorylated in brain but that it is one of the most abundant Man-6-P glycoproteins. PPT is a 37-kDa glycoprotein which cleaves palmitoyl-CoA and removes covalently attached palmitate (and other fatty acid) moieties from proteins (30,31). It is clear from Fig. 1 that brain contains the highest levels of Man-6-phosphorylated PPT. While bands approximating in size to the Man-6-phosphorylated PPT in brain are also seen in kidney, lung, and heart ( Fig. 1), their identity remains equivocal, and they are comparatively minor. Although brain contains the highest levels of Man-6-phosphorylated PPT, it does not contain the highest PPT activity. Using 3 H-labeled, palmitoylated H-Ras as a substrate, PPT activity has been demonstrated to be higher in testis, seminal vesicle, and spleen than brain (30).
The presence of Man-6-P residues on PPT strongly suggests that this protein is lysosomal. However, although the intracellular distribution of PPT has not yet been rigorously defined, it has been tentatively assigned a nonlysosomal, possibly extracellular, localization. This conclusion was based on: 1) overexpression studies in COS cells where most of the PPT was secreted into the media and 2) the observation that PPT has a neutral pH optimum. These do not necessarily preclude a lysosomal localization. First, expression studies in vitro may not necessarily reflect distribution in vivo. For example, known lysosomal enzymes, such as cathepsin D, are secreted when overexpressed in cell lines (47). Moreover, the perinuclear and punctate cytoplasmic staining of intracellular PPT in transfected COS cells (48) is not incompatible with a lysosomal localization. Second, some established lysosomal enzymes such as sialic acid specific 9-O-acetylesterase and glycosyl-N-asparaginase have neutral pH optima (49).
There are some nonlysosomal enzymes which are reported to contain significant amounts of Man-6-P, e.g. renin (50), transforming growth factor ␤1 precursor (51), epidermal growth factor receptor (52), proliferin (53), thyroglobulin (54), and herpes simplex glycoprotein D (55). The physiological relevance of the Man-6-phosphorylated forms of these proteins remains questionable as they have only been detected when overexpressed in cell culture. Should PPT eventually prove to be a nonlysosomal Man-6-P glycoprotein, then it will be the first to be directly observed in vivo. However, it is clear that PPT contains the Man-6-P lysosomal targeting signal and thus most likely represents an authentic lysosomal enzyme.
In a recent study, mutations inactivating PPT have been shown to result in infantile neuronal ceroid lipofuscinosis (INCL) (48), which is a progressive and fatal degenerative encephalopathy. The later stages of INCL are characterized by the appearance of storage bodies containing lipofuscin in the brain and other organs. Lipofuscin is an autofluorescent cellu-lar pigment that represents an accumulation of insoluble polymeric matter within the lysosomes. This material is complex and contains protein and oxidized lipid components. A lysosomal role for PPT actually provides a molecular explanation for INCL. Defects in PPT could lead to accumulation of its lipoprotein substrates within the lysosome, impairing cellular function and resulting in neurodegeneration. INCL may actually therefore represent a classical lysosomal storage disorder.