Cloning, Expression, Purification, and Characterization of the Human Broad Specificity Lysosomal Acid α-Mannosidase

We have cloned and expressed two cDNAs encoding the human lysosomal α-mannosidase (EC 3.2.1.24) by RT-PCR of human spleen mRNA. This enzyme is required for the degradation of N-linked carbohydrates during glycoprotein catabolism in eucaryotic cells. The shorter of the two cDNAs (3 kilobases (kb)) was found to encode an open reading frame of 2964 base pairs and, when expressed in Pichia pastoris, was found to encode an enzyme that could cleave high mannose oligosaccharides, oligosaccharides isolated from α-mannosidosis fibroblasts, and p-nitrophenyl-α-D-mannopyranoside substrates. In addition, the Pichia-expressed enzyme was inhibited by swainsonine, and had a pH optimum, Km, and Vmax characteristic of the enzyme purified previously from human liver. The second, larger RT-PCR product (3.6 kb) was found to contain an insertion and a deletion relative to the 3-kb spleen amplimer and encoded a truncated coding region, indicating that it resulted from alternate transcript splicing. No α-mannosidase activity could be detected in Pichia transformants containing this coding region, indicating that it did not encode a functional enzyme. Antiserum raised to the recombinant product of the 3-kb α-mannosidase cDNA immunoprecipitated lysosomal α-mannosidase activity from human fibroblast extracts. Northern blots identified a 3-kb RNA transcript in all human tissues tested, including α-mannosidosis fibroblasts, while minor transcripts of 3.6 kb were also present in several adult tissues. Human chromosome mapping of the mannosidase gene confirmed that the functional gene maps to the MANB locus on chromosome 19. Sequence comparisons were made to previously published human cDNA sequences encoding a putative lysosomal α-mannosidase (Nebes, V. L., and Schmidt, M. C. (1994) Biochem. Biophys. Res. Commun. 200, 239-245) and several differences were found relative to the functional lysosomal α-mannosidase encoded by the 3-kb spleen cDNA.

The lysosomal catabolism of N-glycans on mammalian glycoproteins occurs through the sequential exoglycosidase digestion of oligosaccharides from the non-reducing terminus down to the carbohydrate-peptide core region (1). Included among the hydrolytic activities responsible for this oligosaccharide degra-dation is a broad specificity exo-␣-mannosidase activity (EC 3.2.1.24) that catalyzes the hydrolysis of ␣1,2-, ␣1,3-, and ␣1,6mannoside linkages present in complex, hybrid, and high mannose Asn-linked glycans (2). This enzyme is distinguished from other ␣-mannosidase activities by a combination of a low pH optimum (pH 4.5), a broad natural substrate specificity, activity toward the synthetic substrate p-nitrophenyl-␣-mannoside, and sensitivity to inhibition by swainsonine (3,4). An enzyme activity with these characteristics has been identified and purified from several species sources including Dictyostelium discoideum (5), and a variety of mammalian tissues (6 -11). Metabolic radiolabeling studies have indicated that the human enzyme is initially synthesized as a polypeptide of ϳ110 kDa that is subsequently processed into two subunits of 40 -46 and 63-67 kDa (12). Proteolytic processing of the precursor appears to be variable in different mammalian species, with reported subunit sizes for the enzyme ranging from 20 to 70 kDa. Even within a single tissue, isozymes of the acidic ␣-mannosidase, designated A and B, have been isolated by ion exchange chromatography differing only in the degree of proteolytic processing from the common precursor (7). The processing of the lysosomal ␣-mannosidase from Dictyostelium has been characterized by metabolic radiolabeling (13), NH 2 -terminal sequencing of the processed subunits, and comparison to the sequence of the ␣-mannosidase gene (5). These studies concluded that the Dictyostelium enzyme precursor is processed by the removal of an NH 2 -terminal signal sequence and two propeptides, one at the NH 2 terminus and one in the center of the precursor, resulting in the mature 60-and 58-kDa subunits.
A human genetic disease characterized by a deficiency in the lysosomal ␣-mannosidase, ␣-mannosidosis, results in varied neural, skeletal, and immune defects, proliferation of lysosomes in most cell types resulting from accumulation of undegraded oligosaccharides, and an elevation of serum and urinary oligosaccharide levels that are characteristics of lysosomal storage diseases (3,14,15). The enzymatic deficiency in the lysosomal ␣-mannosidase in these patients has been long recognized (16 -20), but the molecular basis for the defect has been hampered by the lack of information on the human MANB gene. The gene encoding lysosomal ␣-mannosidase has been isolated from D. discoideum (5) and sequence comparisons (4) with other mammalian ␣-mannosidases identified regions similar to the previously cloned Golgi processing enzyme, ␣-mannosidase II (21), and the endoplasmic reticulum/cytosolic ␣-mannosidase (22). No sequence similarities were detected with the ␣1,2-mannosidases involved in early steps of glycoprotein biosynthesis (23)(24)(25)(26). Thus, a classification of the ␣-mannosidase sequences was proposed whereby the early processing ␣1,2-mannosidases were categorized as Class I ␣-mannosidases, and the Class II ␣-mannosidases were subdivided into Golgi processing enzymes (i.e. ␣-mannosidase II), lysosomal catabolic enzymes (i.e. lysosomal ␣-mannosidase), and the cytosolic/endoplasmic reticulum ␣-mannosidase (4). Regions of sequence similarity between the different subgroups of the Class II ␣-mannosidases allowed the design of degenerate oligonucleotide primers for the isolation of partial murine (27) and full-length human (28) lysosomal ␣-mannosidase cDNAs by the generation of a cDNA probe by RT-PCR 1 and subsequent isolation of full-length clones by plaque hybridization. The human lysosomal ␣-mannosidase cDNA sequence was derived from overlapping partial clones isolated from human retina and muscle cDNA libraries (28) and predicts an open reading frame of 2883 base pairs that is 38% identical to the Dictyostelium lysosomal ␣-mannosidase sequence. Direct evidence that the human cDNA expresses functional enzyme activity with the characteristics of the purified lysosomal enzyme has not been reported, but recently, partial peptide sequence data from the purified human enzyme was found to match the retina/muscle cDNA sequence (29), confirming the identity of the cDNA clones.
To demonstrate that the published cDNA sequence encodes a functional ␣-mannosidase enzyme activity and to initiate studies on the molecular basis of the enzyme defect in the lysosomal storage disease, ␣-mannosidosis, we have isolated cDNA sequences from human spleen and fibroblast cDNA sources by RT-PCR. Heterologous expression of the cDNA in Pichia pastoris and characterization of the expressed enzyme purified from the Pichia culture medium confirmed the authenticity of the cDNA clones as encoding the human lysosomal ␣-mannosidase. In addition, transcripts encoding the enzyme were detected at variable levels in all human tissues tested, including ␣-mannosidosis fibroblasts.
Significant sequence differences were detected when these spleen cDNAs were compared with the published composite clone from the human retina and muscle cDNA libraries (28). These differences include numerous deletions, substitutions, and insertions relative to the published human cDNA sequence. Chromosome mapping confirmed the assignment of the lysosomal ␣-mannosidase gene to the MANB locus human chromosome 19 (28, 30 -33). No pseudogenes were detected, and no amplimer products consistent with the retina/muscle cDNA sequence (28) could be identified. Since the 3-kb spleen cDNA sequence encodes a functional lysosomal ␣-mannosidase activity, it will provide a basis for further studies on the molecular basis of the defect in hereditary human ␣-mannosidosis. cDNA Synthesis-cDNA synthesis was performed using the Superscript preamplification kit (Life Technologies, Inc.) essentially as de-scribed by the manufacturer using 2.5 g of poly(A ϩ ) RNA from the indicated tissue source at 42°C for 1.5 h. After the reaction was completed, the mixture was heated to 90°C for 5 min, followed by the addition of 2 units of RNase H and an additional incubation at 37°C for 20 min. The reaction mixture was extracted with phenol/chloroform/ isoamyl alcohol, 25:24:1 (v/v/v), and desalted over a Sephadex G-50 column (Nick column, Pharmacia Biotech Inc.) equilibrated in 100 mM KCl, 10 mM Tris-HCl (pH 8.0). This desalted preparation was used directly in the PCR amplification.

Materials-Restriction
PCR and Subcloning of the PCR Product-The amplification of the human lysosomal ␣-mannosidase was performed in a 25-l reaction volume containing 3 l of desalted cDNA, 1.0 mM MgCl 2 , 30 mM Tris-HCl (pH 8.5-10), 7.5 mM (NH 4 ) 2 SO 4 , 200 M each dNTP, 0.5 M 5Ј and 3Ј primer, 2.5 units of Taq polymerase, and 1.25 units of Pfu polymerase. The 5Ј and 3Ј primer sequences were, respectively, 5Ј-CCCAAGC-TTGGGGGCTTCGGGGTCTGCGCTCGCGGCTG-3Ј and 5Ј-CCCTCT-AGAGTCTCGGTCTGCCCCCGGAGCAGGAGGCTT-3Ј. The samples were placed in an DNA thermal cycler 480 (Perkin-Elmer) programmed for a temperature-step cycle of 94°C (45 s), 65°C (45 s), and 72°C (4 min) for 5 cycles. An additional 25 cycles were then performed with an annealing temperature of 55°C (45 s) in place of the 65°C step above. After the final cycle, the reaction was maintained at 72°C for 7 min. The final reaction products were resolved on a 1% agarose gel containing ethidium bromide (0.5 g/ml). Amplification products were purified from the gel using the Sephaglas DNA purification kit (Pharmacia) and subcloned into the pCR II vector (Invitrogen). Recombinant plasmids were isolated from liquid bacterial cultures using Qiagen columns (Qiagen Inc., Chatsworth, CA) and subjected to DNA sequencing.
DNA Sequence Analysis-Lysosomal ␣-mannosidase cDNA amplimers subcloned into the vector, pCR II, were sequenced by using Taq polymerase in the dideoxy dye-terminator reaction (34) using synthetic primers and analyzed on an Applied Biosystems 373A DNA sequencer (Molecular Genetics Instrumentation Facility, University of Georgia) following the standard protocol as described by the manufacturer. DNA sequence data were assembled into a contiguous sequence data base by the method of Staden (35). Statistical analysis of sequence similarity between protein or DNA sequences was determined using the Bestfit and Pileup programs of the University of Wisconsin Genetics Computer Group (GCG software, version 8.0).
Expression in P. pastoris-Both the 3.0-and 3.6-kb human spleen cDNA fragments that were isolated by RT-PCR and subcloned into the pCR II vector were excised from pCR II by digestion with EcoRI and ligated into the EcoRI site in the Pichia expression vector, pHIL-D2 (Invitrogen). Both expression constructs were transformed into the his4 Pichia host strain, GS115, using transformation procedures and media compositions described in the Pichia expression manual (Invitrogen) and as described previously (36 -38). Ten micrograms of SalI-linearized recombinant vector was used for each transformation. Negative controls using the pHIL-D2 plasmid without an insert or a transformation with no added DNA were included. Positive transformants had presumably undergone a homologous recombination between the his4 locus of the host cells and the HIS4 gene on the pHIL-D2 vector allowing positive auxotrophic selection in His-deficient medium (36). Ten colonies were picked from each of the transformations for methanol induction (36) of the heterologous protein expression. At the end of methanol induction, the clarified culture medium from each of the transformants was collected by centrifugation at 2800 ϫ g for 10 min and assayed for ␣-mannosidase enzyme activity.
Enzyme and Protein Assays-The human lysosomal ␣-mannosidase was assayed using p-nitrophenyl-␣-D-mannopyranoside (pNP-Man) as substrate as described previously (6). Enzyme assays of crude culture medium or of purified enzyme preparations were performed in a 50-l total reaction volume containing 0.1 M sodium acetate (pH 4.5), 4 mM pNP-Man, and 25 l of enzyme solution at 37°C for 2 h or as specified. The enzyme assays were terminated by the addition of 100 l of a buffer containing 133 mM glycine, 67 mM NaCl, 83 mM sodium carbonate (pH 10.4) and the absorbance was determined at 405 nm on a microtiter plate reader (Bio-Tek, Winooski, VT). For determination of the pH optimum, the sodium acetate buffer was varied from pH 3.5 to 7. One unit of lysosomal ␣-mannosidase activity is defined as the amount of enzyme that releases 1 mol of p-nitrophenol in 1 min at 37°C. When used as an inhibitor of the lysosomal ␣-mannosidase, swainsonine was dissolved in H 2 O to a concentration of 1.44 mM and added into the reaction mixture at the final concentrations indicated in Fig. 5C. Protein concentration was determined using the BCA protein assay reagent (Pierce) as described by the manufacturer using bovine serum albumin as standard.
Purification of Pichia-expressed Human Lysosomal ␣-Mannosidase- The Pichia transformant, LM9, expressing the highest levels of ␣-mannosidase activity in the small scale cultures was prepared for induction in a 1-liter shake-flask culture. The culture was initiated by inoculating LM9 into 10 ml of BMGY medium (36) and incubating at 30°C for 2 days with vigorous agitation. This culture was used to inoculate a 1-liter shake flask culture in BMGY medium for 2 days at 30°C. Cells were harvested by centrifugation at 2800 ϫ g for 30 min and resuspended in 200 ml of BMMY induction medium, followed by incubation at 30°C for 5 days with vigorous shaking (36). After 2 days of culture in the BMMY medium, 50% methanol (v/v in H 2 O) was added to the culture medium to a final concentration of 0.5%, and the culture was supplemented daily with methanol to maintain a concentration of 0.5%. After 5 days of induction in BMMY medium, the medium was harvested by centrifugation at 2800 ϫ g for 30 min, and the clarified supernatant was concentrated to 30 ml by ultrafiltration through a YM-100 membrane (Amicon). The concentrated medium was dialyzed against 50 mM potassium phosphate (pH 6.0) (phosphate buffer), clarified by centrifugation at 16000 ϫ g for 30 min, and applied at a flow rate of 1 ml/min to a Q-Sepharose column (26 mm ϫ 130 mm, Pharmacia) pre-equilibrated with the same buffer. The column was washed with 50 ml of phosphate buffer, and ␣-mannosidase activity was eluted with a 170-ml step gradient of 0 -1 M NaCl in phosphate buffer Endoglycosidase Deglycosylation-Purified human lysosomal ␣-mannosidase was denatured in a solution of 0.5% SDS and 1% ␤-mercaptoethanol at 100°C for 10 min and subsequently incubated overnight in 50 mM sodium citrate (pH 5.5) with 3000 units of endo H at 37°C. Endoglycosidase digestion of undenatured lysosomal ␣-mannosidase was performed in 50 mM sodium citrate (pH 5.5) with 3000 units of endo H at 20°C overnight without heat denaturation in SDS/␤-mercaptoethanol. Aliquots of the deglycosylated material were resolved on SDSpolyacrylamide gels or assayed for enzyme activity.
Lectin Detection of Glycoprotein-Proteins resolved on SDS-PAGE gels were electroblotted onto polyvinylidene difluoride membranes (Immobilon-P, Millipore) as described previously (40). Lectin detection of glycosylated protein bands on the blot was accomplished using the DIG glycan differentiation kit (Boehringer Mannheim) using a method adapted from the manufacturer's instruction. Following the electrophoretic transfer, the blot was incubated for at least 30 min in blocking solution (0.5% casein in 0.05 M Tris-HCl, 0.15 M NaCl (pH 7.5) (TBS)) and washed twice for 10 min each in TBS and then once with Buffer 1 (1 mM MgCl 2 , 1 mM MnCl 2 , 1 mM CaCl 2 in TBS) at room temperature. After blocking and washing, the blot was incubated with a digoxigeninlabeled Galanthus nivalis agglutinin (GNA) solution (1 g/ml in Buffer 1) for 1 h at room temperature. Following three washes for 10 min each in TBS, the blot was incubated with a polyclonal sheep anti-digoxigenin Fab fragment conjugated with alkaline phosphatase. After a 1 h incubation at room temperature, the blot was washed three times in TBS. The alkaline phosphatase visualization solution was composed of 7.7 mg of nitro blue tetrazolium, 3.75 mg of X-phosphate in 20 ml of 0.1 M Tris-HCl (pH 9.5), 50 mM MgCl 2 , 0.1 M NaCl. Following the staining reaction, the blot was rinsed in H 2 O and photographed.
Preparation of Antibodies to the Human Lysosomal ␣-Mannosidase, and Immunoprecipitation of Enzyme Activity-The purified human lysosomal (300 g) was emulsified in an equal volume of Freund's complete adjuvant was used to immunize a male New Zealand White rabbit. At 3-week intervals additional booster immunizations were given with 300 g of enzyme and Freund's incomplete adjuvant. Immunoprecipitation studies were performed using either the purified enzyme expressed in Pichia or enzyme activity from crude cell extracts from human fibroblast cultures. Fibroblast extracts were generated from a washed cell pellet (1.2 ϫ 10 7 cells) by homogenization in a buffer containing 2% Triton X-100, 0.5 M NaCl, 20 mM potassium phosphate (pH 7.5), followed by centrifugation at 145,000 ϫ g for 30 min. Immunoprecipitation was performed by preincubation of the indicated amount of antibody with 100 l of Protein A-Sepharose for 2 h at 4°C on a rotating mixer. Unbound antibody was removed by washing the beads twice with 1 ml of phosphate-buffered saline. Cell extracts or pure protein preparations were then added to the washed beads and allowed to incubate at 4°C for 2 h. The beads were sedimented by centrifugation at 16,000 ϫ g for 1 min, and residual activity in the supernatant was determined using the pNP-Man substrate as described above. Enzyme activity associated with the beads was determined following two washes with 1 ml of phosphate-buffered saline by direct addition of the substrate solution to the beads. Following a 6-h incubation at 37°C, the stop solution was added, the beads were sedimented by centrifugation at 16,000 ϫ g for 1 min, and the absorbance (at 400 nm) of the supernatant solution was determined.
Isolation of Oligosaccharides from ␣-Mannosidosis Fibroblasts-Human ␣-mannosidosis fibroblasts (19 confluent 100-mm culture dishes of the D.B. cell line) were removed from the culture dish by trypsinization, washed once with Hank's balanced salt solution and resuspended in 2 ml of sonication buffer (0.5 M NaCl, 10 mM Tris-Cl (pH 7.0)), sonicated for four 30-s bursts at a setting of 30 watts with a Sonifier Cell Disruptor model 185 (Branson Ultrasonics Corp., Danbury, CT), and clarified by centrifugation at 17,000 ϫ g for 30 min. The pellet was resuspended in 2 ml of sonication buffer and was subjected to sonication and centrifugation as above. The two supernatants were pooled, and two volumes of ethanol were added. The sample was incubated overnight at Ϫ20°C and then centrifuged at 17,000 ϫ g for 30 min to remove the precipitated proteins. The supernatant solution was dried, resuspended in water, and sequentially passed through a C-18 SepPak cartridge (Waters Corp., Milford, PA) and a column (0.6 cm ϫ 2 cm) of mixed bed ion exchange resin (equal volume of AG1-X8 and AG50W-X8, Bio-Rad) to remove any hydrophobic and charged components, respectively. The remaining material in the run-through solution was dried and fluorescently tagged with pyridylamine (PA) using the procedure of Hase et al. (41,42). The tagged oligosaccharides were desalted on a column of Toyopearl HW40 F (1 ϫ 66 cm, TosoHaas, Montgomeryville, PA) equilibrated in 10 mM ammonium acetate (pH 6.0), at a flow rate of 0.5 ml/min. Relevant fractions were pooled, dried, and used as substrate for digestion with the lysosomal ␣-mannosidase. Man 9 -GlcNAc was isolated from soybean agglutinin as described previously (43,44). Man 5 -GlcNAc was generously provided by Dr. Annette Herscovics (McGill Cancer Center, Montreal).
Natural Substrate Digestions and Chromatography-Enzyme reactions with oligosaccharide substrates were carried out in 100 mM sodium acetate (pH 4.5) in a volume of 220 l. At individual time points, aliquots were removed and adjusted to 70% acetonitrile and 50-l aliquots were run on a Hypersil APS-2 NH 2 high performance liquid chromatography column (250 ϫ 4.6 mm, Keystone Scientific, Bellefonte, PA) with an initial buffer of 70% acetonitrile, 30% 100 mM sodium phosphate (pH 4.0). The column was developed with a linear gradient over 60 min from the initial buffer to a buffer containing 49% acetonitrile, 51% 78 mM sodium phosphate (pH 4.0), at a flow rate of 1 ml/min. 2 Peak area was used to quantitate the percent of each intermediate at each time point. Oligosaccharide size standards for the smaller oligosaccharides were obtained from G. Alvarez-Manilla (University of Georgia) and were composed of the disaccharide Gal-GlcNAc, the tetrasaccharide Gal-GlcNAc-Gal-Glc, the pentasaccharide (Fuc)-Gal-GlcNAc-Gal-Glc, and the hexasaccharide (Fuc) 2 -Gal-GlcNAc-Gal-Glc. The positions of the fucosyl residues in the above oligosaccharides have not been determined. The high mannose oligosaccharide size standards, Man 9 -5 GlcNAc, were obtained from Annette Herscovics ((McGill Cancer Center, Montreal).
Northern Blot Analysis-Poly(A ϩ ) RNA was isolated from human fibroblast cell lines by extraction of the cells in an SDS/protease mixture and purification on oligo(dT) as described by the manufacturer (Fast-Track, Invitrogen). One microgram of RNA was resolved on a 1% formaldehyde-agarose gel and transferred by capillary blotting to a Zetaprobe nylon membrane (Bio-Rad) as described previously (24). Northern blots containing poly(A ϩ ) RNA from various human tissues were purchased from Clontech. Both the human fibroblast Northern blots and the human tissue Northern blots were prehybridized, hybridized, and washed as described (21) using the 3.0-kb human spleen cDNA amplimer as the radiolabeled probe. The blots were subsequently hybridized with a 32 P-labeled human ␤-actin control probe (Clontech) to act as an RNA load control for the blots. 32 P-Labeled DNA probes were generated using [ 32 P]dCTP and the Megaprime labeling system (Amersham). Blots were either exposed to a x-ray film for 4 days at room temperature (for the human multiple Northern blot, Fig. 7) or visualized with a PhosphorImager (Molecular Dynamics) after a 1-day exposure (for the human fibroblast Northern blot, Fig. 8).
Chromosome Mapping of Human Lysosomal ␣-Mannosidase Gene-Somatic cell hybrid DNAs (BIOSMAP, BIOS Laboratories, Inc.) were screened for the lysosomal ␣-mannosidase gene by PCR. Two human/ rodent somatic cell hybrid panels were tested. The first hybrid panel contains 20 human/rodent hybrid cell lines, each carrying one or more specific human chromosomes. The second hybrid panel includes 24 human/rodent hybrid cell lines, each containing a single human chromosome. Rodent and human genomic DNAs were also tested as negative and positive controls, respectively. The PCR reactions for chromosome mapping were performed in a 25-l volume containing 60 mM Tris-HCl (pH 9.0), 15 mM (NH 4 )SO 4 , 2 mM MgCl 2 , 200 M dNTP, 0.5 M 3Ј and 5Ј primers, 50 ng of chromosomal DNA from the hybrid panels, and 2.5 units of Taq polymerase. The 5Ј and 3Ј primer sequences were: 5Ј-GGCAGATCCCACCCGTCGCTTC-3Ј and 5Ј-GGGGTCGCCCATC-ATTGCCAAA-3Ј, respectively, and mapped to positions 258 bp and 499 bp relative to the start of the ␣-mannosidase cDNA open reading frame. A thermal step cycle of 94°C (1 min), 65°C (1 min), and 72°C (2 min) was repeated for five cycles. An additional 30 cycles were then performed with an annealing temperature of 60°C (1 min) in place of the 65°C step above. The final reaction products were resolved on a 2% agarose gel. An amplified fragment was obtained from human genomic DNA controls or human/rodent hybrids containing human chromosome 19, but not from hamster or mouse chromosome DNA. The amplified fragment from the human genomic DNA source was 600 bp in length, containing 242 bp of exon sequence and a 358-bp intron (data not shown).

Isolation of cDNA Clones Encoding the Human Lysosomal
␣-Mannosidase-We have previously isolated partial cDNA clones encoding the murine lysosomal ␣-mannosidase (27) by a PCR-based approach using conserved sequences between the murine Golgi ␣-mannosidase II (21) and the D. discoideum lysosomal ␣-mannosidase (5). A similar approach was taken in the isolation of partial clones encoding an ␣-mannosidase-related sequence from human retina and muscle cDNA libraries (28). Using the published composite sequence of the human retina and muscle ␣-mannosidase cDNA clones (28), we designed exact-match primers flanking the ␣-mannosidase coding region to isolate the full-length open reading frame by RT-PCR. cDNA sources for the amplification reactions included human placenta, liver, spleen, and cultured fibroblasts. A single PCR product of 3 kb was obtained from placenta, liver, and fibroblasts, while an additional amplimer of 3.6 kb was detected from the spleen cDNA template. The 3.0-and 3.6-kb spleen cDNA amplimers (Fig. 1) and the 3.0-kb fibroblast amplimer (data not shown) were subcloned and fully sequenced.
The 3-kb PCR amplimers from human spleen and fibroblast cDNA sources were essentially identical in sequence, the differences being nine single-base substitutions between the two sequences, five of which result in amino acid substitutions. Sequence differences at this frequency could readily result from either allelic differences or PCR artifacts. The clones encoded an open reading frame of 2964 bp with 55 base pairs of 5Ј-untranslated region and 60 base pairs of 3Ј-untranslated region, respectively. Translation of the spleen cDNA coding region predicted a polypeptide of 988 amino acids (Fig. 2) with 10 putative N-glycosylation sites and an NH 2 -terminal sequence containing characteristics of a cleavable signal se-quence (45). The 3.6-kb human spleen cDNA sequence is essentially identical to the 3-kb spleen sequence, except that it contains a 83-bp deletion and a 691-bp insertion within the coding region (Fig. 1). A frameshift introduced by the deletion in the 3.6-kb cDNA causes a premature translational termination to generate a polypeptide of 323 amino acids (Fig. 2). The sequences bordering the insertion conform to consensus transcript splicing donor and acceptor sites (46) and the 691-bp insertion sequence contains extended regions of poly(dT) sequence, indicating that it is unlikely to encode a polypeptide. In addition, the insertion matches the sequence obtained from human lysosomal ␣-mannosidase genomic clones, 3 indicating that it represents an unspliced intron sequence. The 3.6-kb amplimer does not represent an unspliced transcript, however, since it contains an 83-bp deletion within the coding region and is missing several other introns identified in genomic clones. 3 These sequence characteristics of the 3.6-kb spleen amplimer indicate that it may represent an alternative splicing product of human lysosomal ␣-mannosidase transcript in spleen tissues.
Comparison between Human Spleen and Retina/Muscle cDNA Sequences-Comparison of the deduced polypeptide translations of our spleen and fibroblast cDNA sequences with the published human retina and muscle cDNA clones (28) revealed that, while the sequences were highly similar, they were not identical (90.5% identity). The retina/muscle composite cDNA sequence contains 16 base deletions, 11 base substitutions, 4 base insertions, and a 69-base gap relative to the consensus of the human spleen cDNA sequences (Figs. 1 and 2). The 69-bp deletion in the retina/muscle cDNA is at the same position as the 691-bp insertion in the 3.6-kb spleen transcript. The sequences bordering the deletion in the retina/muscle sequence conform to consensus transcript splicing donor and acceptor sites (46), indicating that it also presumably arises from alternative splicing of the RNA transcript at this sequence position to result in the excision of a portion of a coding exon. 3  The additional deletions and insertions found in retina and muscle cDNA sequences (Figs. 1 and 2) cause several frameshifts, resulting in significant differences in the predicted protein sequence between our spleen cDNAs and the published retina/muscle cDNAs. Sequence data from independent spleen RT-PCR products as well as equivalent regions of the human 3.6-kb spleen cDNA, human genomic clones spanning this sequence region, 3 and the cDNA sequence of the murine lysosomal ␣-mannosidase 4 were all consistent with the reading frame and sequence of the 3-kb human spleen cDNA and in disagreement with the previously published human retina/muscle composite cDNA sequence (28).
Protein Expression in P. pastoris-Sequence differences between the translation products of our spleen cDNAs and the previously published retina and muscle cDNA sequences (28) indicated that it is likely that only one of the sequences will result in the expression of functional enzymatic activity. In order to demonstrate the expression of functional lysosomal ␣-mannosidase activity from our spleen cDNA sequences, we prepared constructs for expression in the methylotrophic yeast, P. pastoris. Both the 3-kb and the 3.6-kb spleen cDNA fragments were individually subcloned into the expression plasmid, pHIL-D2, a vector that employs the methanol-inducible promoter of the Pichia alcohol oxidase gene (AOX1) (37). The constructs were prepared with the assumption that the NH 2terminal signal sequence on the human lysosomal ␣-mannosi-dase coding region would direct the co-translational translocation of the polypeptide into the yeast secretory pathway. Overexpression of the polypeptide via the AOX promoter would be expected to result in secretion of the enzyme into the culture medium where it could be detected using the synthetic substrate, p-nitrophenyl-␣-D-mannopyranoside (pNP-Man). Plasmid constructs containing the human spleen ␣-mannosidase cDNAs were linearized by digestion with SalI and used to transform spheroplasts of the his4 Pichia host strain, GS115. Homologous recombination of the expression vector-encoded HIS4 gene at the host his4 locus resulted in a reversion of the auxotrophic phenotype and growth on minimal medium. Of the several hundred Pichia transformants, 10 colonies of each of the two spleen cDNA constructs and vector controls without inserts were isolated and assessed for expression of ␣-mannosidase activity. Transformant cultures were induced for 2 days in medium containing methanol as the sole carbon source and enzyme activity in the medium was determined. Only the transformants containing the 3-kb spleen cDNA expressed detectable ␣-mannosidase activity in the medium. The medium of 5 of the 10 transformants with the 3-kb spleen construct contained ␣-mannosidase activity that was inducible with an approximately linear time course over 6 days of culture (data not shown). No activity could be detected in the medium of the 3.6-kb spleen cDNA constructs or the vector control transformants during the 6 days of culture. The transformant that expressed the highest level of inducible ␣-mannosidase activity, LM9, was subjected to large scale culture for enzyme purification and characterization. Purification of Human Lysosomal ␣-Mannosidase from the Pichia Culture Medium-The human lysosomal ␣-mannosidase transformant, LM9, was induced in a 1-liter shake flask culture for 5 days as described under "Experimental Procedures," and the ␣-mannosidase was purified to homogeneity by a combination of ultrafiltration, ion exchange chromatography, hydrophobic chromatography, and gel permeation chromatography ( Fig. 3 and Table I) as described under "Experimental Procedures." The purified enzyme preparation was resolved by SDS-PAGE and migrated as a diffuse band with an apparent molecular mass of ϳ210 kDa (Fig. 4, panel A, lane 1), almost twice the expected size of the translated spleen cDNA sequence (predicted polypeptide molecular weight ϭ 111,484). The protein band stained poorly with Coomassie R-250 (Fig. 4, panel A, lane 1) and silver staining, but could be readily detected on gel blots using a digoxigenin-labeled GNA lectin (Fig. 4, panel B,  lane 1), a mannose-specific lectin (47), indicating that the protein was highly glycosylated.
The predicted translation of the human spleen lysosomal ␣-mannosidase cDNA sequence indicated 10 potential N-glycosylation sites. The contribution of Asn-linked oligosaccharides to the apparent molecular mass of the enzyme was determined by treatment with endoglycosidase H. Treatment of the purified enzyme under denaturing conditions with endo H as de-scribed under "Experimental Procedures" resulted in a decrease in apparent molecular mass to ϳ150 kDa, a sharpening of the band on SDS gels, and an increased sensitivity to staining by Coomassie R-250 (Fig. 4, panel A, lane 2). A similar mobility shift was obtained if the purified protein was treated with endoglycosidase H without denaturation at 20°C for 16 h (data not shown). ␣-Mannosidase enzyme activity levels, K m , and V max were not affected by the endoglycosidase treatment under non-denaturing conditions (data not shown). The 150-kDa product of the endoglycosidase digestion was still larger than that predicted by the translation of the cDNA, and it also stained heavily with the GNA lectin (Fig. 4, panel B, lane 2), indicating residual oligosaccharides were still attached to the polypeptide. Further digestion with endoglycosidase H or Nglycanase were without effect (data not shown). We hypothesize that the additional oligosaccharides represent extensive addition of O-linked mannose structures that are not susceptible to endo H digestion, similar to those commonly observed on glycoproteins from Saccharomyces cerevisiae (48). Further characterization of the residual oligosaccharide structures is presently under way.
The specific activity of the purified enzyme (Table I) is considered to be an estimate based on the difficulty in generating consistent data with different protein determination techniques. The heavy glycosylation of the secreted recombinant enzyme would be a likely cause for the inconsistency. The final enzyme preparation was enriched 719-fold from the crude medium to a specific activity of 0.287 units/mg. This specific activity is approximately 1/10 of the published specific activity for the purified human liver lysosomal ␣-mannosidase (10), but the fact that the secreted form is a heavily glycosylated precursor form of the enzyme could significantly influence both the enzyme activity and the protein determination measurements.
Enzymatic Characterization of the Expressed Protein-The purified human lysosomal ␣-mannosidase expression product was further characterized using pNP-Man as the substrate. The expressed enzyme shows catalytic activity toward the synthetic substrate between pH 4.25-6.0, with the highest activity at pH 4.5 (Fig. 5A). The enzyme activity follows normal Michaelis-Menten kinetics with a K m of 2.4 mM for pNP-Man (Fig. 5B). Swainsonine, a potent inhibitor of Class II ␣-mannosidases (4), also significantly inhibited the expressed enzyme activity with a IC 50 ϭ 0.11 M (Fig. 5C).
The purified enzyme was also tested for substrate specificity in the cleavage of high mannose oligosaccharides, including oligosaccharides accumulated in ␣-mannosidosis fibroblasts. The genetic deficiency in the lysosomal ␣-mannosidase causes the accumulation of predominantly linear oligosaccharides extended on the ␣1,3Man branch of the tri-mannosyl oligosaccharide core of N-linked glycans (49). Neutral free oligosaccharides were isolated from ␣-mannosidosis fibroblasts, labeled with pyridylamine, and subjected to enzymatic cleavage with the purified lysosomal ␣-mannosidase (Fig. 6, left panels). The ␣-mannosidosis oligosaccharide sample contained a mixture of structures eluting at positions corresponding to Man 7-2 -GlcNAc-PA standards, as would be predicted based on the structures of oligosaccharides in the urine of ␣-mannosidosis patients (49). The enzymatic digestion resulted in a progressive cleavage down to a structure with the elution position of Man-GlcNAc-PA, the limit digestion product of the lysosomal ␣-mannosidase from human fibroblasts (2). The purified recombinant lysosomal ␣-mannosidase was also tested with pyridylaminederivatized Man 9 GlcNAc and Man 5 GlcNAc substrates. The Man 9 GlcNAc-PA structure could be cleaved down to structures smaller than Man 5 GlcNAc-PA and the Man 5 GlcNAc-PA substrate could be readily cleaved down to a Man 2 GlcNAc-PA  (Fig. 6, center and right panels, respectively). These data indicate that the recombinant enzyme is capable of cleaving ␣1,2Man, ␣1,3Man, and ␣1,6Man linkages on Man 5 GlcNAc and Man 9 GlcNAc substrates, while not recognizing one of the ␣-linked mannose residues remaining on the ␤-linked core mannose. Previous studies have shown that the broad specificity lysosomal ␣-mannosidase is inefficient in the cleavage of the single ␣-linked mannose residue on the trisaccharide Man␣1,6Man␤1,4GlcNAc. This residue is instead recognized by the ␣1,6 core-specific lysosomal ␣-mannosidase (50,51). This cleavage pattern is in contrast to the product of the lysosomal ␣-mannosidase digestion of the ␣-mannosidosis oligosaccharides where a disaccharide, presumably Man␤1,4GlcNAc, is produced. Oligosaccharides accumulating in ␣-mannosidosis fibroblasts have a very low content of the Man␣1,6Man linkage because of the action of the ␣1,6 core-specific mannosidase in these cells (2,52). All of the other ␣-linked mannose residues in the accumulated oligosaccharides are susceptible to cleavage by the broad specificity lysosomal ␣-mannosidase. These data indicate that the spleen cDNA encodes a functional enzyme with a natural substrate specificity similar to the mammalian broad specificity lysosomal ␣-mannosidase.
Immunoprecipitation of Fibroblast ␣-Mannosidase Enzyme Activity from Cell Extracts-Antiserum was raised to the recombinant enzyme derived from the human spleen 3.0-kb cDNA and was used to immunoprecipitate the recombinant lysosomal ␣-mannosidase enzyme activity as well as the activity from cell extracts of human fibroblasts (Fig. 5D). The antibody, bound to Protein A-Sepharose, was shown to deplete ␣-mannosidase activity from fibroblast cell extracts and the enzyme activity was shown to be quantitatively associated with the pelleted Protein A-Sepharose beads.
Tissue Distribution of Lysosomal ␣-Mannosidase RNA-Transcript levels of the human lysosomal ␣-mannosidase were determined by Northern blot analysis using the full-length 3-kb spleen amplimer as a radiolabeled probe. A major lysosomal ␣-mannosidase transcript of 3 kb was found in all tissues, although the abundance varied among tissues (Fig. 7). The transcript was most abundant in peripheral blood leukocytes and barely detectable in brain. In addition to the major 3-kb transcript, a second minor transcript of 3.6 kb could also be detected in several tissues, but most prominently in spleen, thymus, and leukocytes, consistent with our isolation of a 3.6-kb alternate transcript from spleen cDNA. In testis, an additional 2-kb transcript was detected using the spleen 3-kb amplimer as a probe. The nature and origin of this transcript is presently under investigation, since the size of the transcript is smaller than the size of the coding region of any known lysosomal ␣-mannosidase.
In addition to the Northern blot of RNA from human tissues,   circles and squares). The bottom panel shows the enzyme activity associated with the Protein A-Sepharose beads following sedimentation and washing twice with phosphatebuffered saline (filled boxes and squares). Conditions for the immunoprecipitation and enzyme assay are as described under "Experimental Procedures." a blot was prepared with RNA isolated from normal and ␣-mannosidosis fibroblasts (Fig. 8). A single transcript of 3.0 kb was detected in all of the fibroblast lines. Despite the variable RNA content in the lanes, when the bands hybridizing with the ␣-mannosidase probe were normalized to the actin control (Fig.  8, panel C), the abundance of the ␣-mannosidase transcripts was found to be similar or slightly reduced in the ␣-mannosidosis fibroblasts. These data, in combination with immunocytochemistry data detecting the lysosomal ␣-mannosidase polypeptide in ␣-mannosidosis fibroblasts using the antibody to the recombinant enzyme (data not shown), indicate that the deficiency in ␣-mannosidase activity in these cells does not result from defective transcription or translation, but instead results from an inactive polypeptide.
Chromosome Mapping of the Lysosomal ␣-Mannosidase Gene-Previous studies have indicated that the human lysosomal ␣-mannosidase gene, MANB, is present on chromosome 19 (28, 30 -33). The differences in sequence between the previously published retina/muscle cDNA and the spleen and fibroblast cDNAs presented here indicate that one of the sequences may represent an expressed pseudogene. To identify the presence of alternate genes on chromosomes other than human chromosome 19, we employed a PCR screening approach using primer pairs that should anneal to both our cDNA as well as the retina/muscle genomic sequences. These primer pairs were used to screen DNA samples from two human/rodent somatic cell hybrid panels. Data from the somatic cell hybrid panel screening indicated that amplification products could only be obtained from cell lines containing human chromosome 19 (concordance 98%, Table II). Sequence data from the amplimers obtained from the chromosome 19 somatic cell hybrids as well as data obtained from total human genomic DNA were identical to the spleen cDNA sequence (data not shown). These results indicate that there were no detectable sequences within either human genomic DNA or within the cDNAs tested that were consistent with the previously published retina/muscle cDNA sequences. Arrows indicate the significant signals corresponding to the human lysosomal ␣-mannosidase transcripts, with the major transcript being ϳ3 kb in most tissue. There is an additional minor band of ϳ2 kb in human testis. An additional minor band of 3.6 kb in several tissues is coincident with the alternatively spliced spleen transcript isolated by RT-PCR (3.6-kb transcript in Figs. 1 and 2).

DISCUSSION
The catabolism of mammalian glycoproteins requires the action of a broad specificity lysosomal ␣-mannosidase to fully degrade high mannose, hybrid, and complex oligosaccharides. In some mammalian species, an alternate ␣-mannosidase activity that is specific for ␣1,6-mannose linkages also appears to play an accessory role in oligosaccharide catabolism (50). Failure to hydrolyze ␣-mannose linkages through a genetic defect in the expression of the broad specificity lysosomal ␣-mannosidase activity can result in the accumulation of undegraded oligosaccharides in tissues, serum, and urine and cause the appearance of the clinical symptoms of the lysosomal storage disease, ␣-mannosidosis (16 -20).
Lysosomal ␣-mannosidase has been purified from a number of mammalian sources (6 -11) and is generally isolated as a hetero-oligomer of multiple subunits. Biosynthetic radiolabeling studies in human fibroblasts demonstrated that the enzyme is synthesized as a 110-kDa precursor and is proteolytically clipped into fragments of 40 -46 kDa and 63-67 kDa upon transport to lysosomes (12). Cultured fibroblasts also secrete low amounts of the uncleaved precursor into the medium in a manner similar to the secretion of precursors of other lysosomal enzymes (15). Similar processing events have been identified in the maturation of the lysosomal ␣-mannosidase in D. discoideum (13) and protein sequence data from the purified subunits of this enzyme eventually led to the cloning of the D. discoideum lysosomal ␣-mannosidase gene (5).
The unexpected sequence similarities between the D. discoideum lysosomal ␣-mannosidase and other mammalian ␣-mannosidases involved in glycoprotein processing has led to a proposed classification system for the ␣-mannosidase genes into two broad categories based on sequence similarities, size of the encoded polypeptides, substrate specificities, and inhibition by alkaloids (4). The lysosomal ␣-mannosidase has been categorized as a Class II ␣-mannosidase, with sequence similarity to Golgi ␣-mannosidase II and the cytosolic/endoplasmic reticulum ␣-mannosidase. In addition, the sequence similarities have allowed the design of degenerate oligonucleotide primers for the isolation of ␣-mannosidase cDNA probes from a variety of species sources (27). This degenerate primer RT-PCR approach (53) has allowed us to isolate cDNA clones encoding the murine lysosomal ␣-mannosidase 4 and others to isolate clones encoding a human homolog (28). The sequence of the putative human lysosomal ␣-mannosidase cDNA was assembled as a composite of two clones isolated from retina and muscle cDNA libraries, respectively. Although the sequence similarity to the D. discoideum lysosomal ␣-mannosidase was noted, there was no direct demonstration that the composite of these clones encoded a functional ␣-mannosidase activity.
The original goal of our studies was to demonstrate that the previously published human retina/muscle cDNA sequence encoded a functional enzyme activity. In the initiation of these studies we isolated cDNAs by RT-PCR from human spleen, liver, placenta, and fibroblast cDNA sources. Sequence analysis of the subcloned products identified several sequence differences between our cDNAs and the published retina/muscle cDNA composite sequence. Several frameshifts, amino acid substitutions, and a 23-amino acid deletion resulted in a pep- FIG. 8. RNA blot of the human lysosomal ␣-mannosidase in normal and ␣-mannosidosis fibroblasts. One g of poly(A ϩ ) RNA was isolated from normal (WT) and ␣-mannosidosis fibroblasts (M.S. and D.B.), resolved on a 1% agarose/formaldehyde gel and blotted onto a nylon membrane and probed with the radiolabeled 3-kb spleen cDNA (panel A) or the ␤-actin probe as a load control (panel B). The blot was subjected to PhosphorImager analysis after exposure in a cassette overnight. Since the actual amount of RNA loaded in each lane varied, the relative band intensity for the lysosomal ␣-mannosidase was calculated relative to the ␤-actin load control (panel C). The relative band intensity in each lane was compared and expressed as a ratio of the intensity of the lysosomal ␣-mannosidase transcript to the intensity of the actin transcript. The similarity of the ratios of the bars in the three lanes indicates that the relative transcript levels for the lysosomal ␣-mannosidase in the three cell lines are approximately equivalent. tide translation with a 90.5% identity to the retina/muscle composite sequence. Sequence differences of this magnitude would not be expected to be accounted for by PCR error or allelic variation. We confirmed that the sequences isolated from spleen cDNA represented an authentic lysosomal ␣-mannosidase coding region by additional sequencing of independent cDNA and genomic amplimers, as well as comparison to the full-length murine lysosomal ␣-mannosidase. 4 Additional confirmation that the 3-kb spleen cDNA encoded the functional lysosomal ␣-mannosidase activity was accomplished by heterologous expression of the cDNA in P. pastoris. The expressed enzyme was purified from the culture media and was found to contain all of the catalytic characteristics of the enzyme purified from human tissues.
In an attempt to identify the source of the retina/muscle cDNA sequence and to determine if it represented an expressed pseudogene, we used a PCR-based chromosome mapping approach. Using primer pairs in regions of DNA sequence identity between the retina/muscle cDNA sequence and our spleen cDNAs, we bracketed regions of sequence difference between the two cDNA sources. The chromosome mapping identified a single genetic locus on chromosome 19 consistent with the previous assignment of the MANB gene to human chromosome 19 (30 -33). Sequence analysis of the amplimer product was identical to the spleen cDNA sequence. Although we have been unsuccessful in our attempts to identify the source of the retina/muscle cDNA sequence in the human genome, we cannot rule out the possibility that the sequence represents an expressed gene or pseudogene that is unique to retina and muscle and is resistant to amplification from human genomic DNA. A definitive test of the authenticity of the retina/muscle cDNA clones would require a comparative expression of this composite cDNA in a host such as Pichia that we have shown can efficiently express functional lysosomal ␣-mannosidase activity.
Partial amino acid sequence data have been recently presented from the purified human lysosomal ␣-mannosidase as a confirmation of the authenticity of the retina/muscle cDNA clones (29). These sequence data are identical with residues 551-569 of the retina/muscle cDNA (28) as well as residues 579 -597 of our spleen cDNA and therefore could not be used to distinguish between the two sequences.
The most significant of the characteristics of the recombinant enzyme was the specificity for natural high mannose oligosaccharides and oligosaccharides from ␣-mannosidosis fibroblasts. Cleavage of high mannose oligosaccharides by the mammalian broad specificity lysosomal ␣-mannosidase down to Man 2 GlcNAc is consistent with the inefficient cleavage by this enzyme of the Man␣1,6Man linkage on the ␤-linked core mannose. This inefficient cleavage was the basis for the discovery of a novel ␣1,6 core-specific ␣-mannosidase in lysosomes that recognizes this substrate for cleavage (50). In contrast, ␣-mannosidosis fibroblasts contain the ␣1,6 core-specific lysosomal ␣-mannosidase and accumulate oligosaccharides susceptible to cleavage by the broad specificity lysosomal ␣-mannosidase down to a Man␤1,4GlcNAc disaccharide. These characteristics of substrate specificity were also found for the recombinant expression product of the spleen cDNA. These data indicate that the spleen 3.0-kb cDNA encodes the broad specificity ␣-mannosidase that is targeted to mammalian lysosomes.
Analysis of the lysosomal ␣-mannosidase transcript expression by Northern blotting identified a major band of 3 kb that was highly expressed in most tissues except brain. In addition, a second minor band of 3.6 kb could be identified in several tissues consistent with an RT-PCR amplimer sequence isolated from the spleen cDNA source. The cDNA sequence corresponding to this transcript contained a deletion and an insertion within the coding region indicating that it could represent an alternatively spliced transcript. Since the deletion causes a frameshift and a premature termination of the polypeptide, attempts were made to express the 3.6-kb cDNA in Pichia. The lack of secreted ␣-mannosidase activity in the Pichia transformants containing the 3.6-kb cDNA, in contrast to the 3-kb cDNA, suggests that the 3.6-kb cDNA does not encode a functional enzyme. However, we cannot rule out the possibility that the truncated enzyme exhibits an alternate catalytic activity that can act on natural substrates but does not recognize the synthetic substrate, pNP-Man.
A third transcript of ϳ2 kb was detected in human testis. A sperm-associated ␣-mannosidase has been previously identified and purified from a rat sperm plasma membrane fraction and shown to play an important role in the maturation of sperm and necessary for sperm-egg interaction (54,55). In addition, an epididymis-specific ␣-mannosidase was purified, and the cDNA sequence encoding this enzyme was cloned from a porcine corpus epididymis cDNA source. Both enzymes have the biochemical characteristics of Class II ␣-mannosidases (4), but both polypeptides are larger than could be encoded by a 2-kb transcript. We are presently attempting to isolate the cDNA clones corresponding to the 2-kb human testis transcripts in order to identify the differences from the major lysosomal ␣-mannosidase transcripts in other tissues.
The detection of transcripts for the lysosomal ␣-mannosidase in ␣-mannosidosis fibroblasts is in agreement with the immunocytochemical detection of cross-reacting material in lysosomes (data not shown) and the previous detection of residual trace ␣-mannosidase activities in cells from ␣-mannosidosis patients. Kinetic analyses of the residual activities in lymphocytes of type I or type II ␣-mannosidosis patients indicated that there was an inverse correlation between the severity of the disease and the affinity of the enzyme for its substrate (14). These results suggested that the heterogeneous presentation of this disorder and the distinctions between the clinical forms of the disease probably reflect multiple allelic missense mutations that result in differences in catalytic characteristics of the enzyme. The detection of normal transcript levels and crossreactive material in lysosomes of ␣-mannosidosis fibroblasts are consistent with these conclusions.
A characterization of the molecular defect in patients with human lysosomal storage disease, ␣-mannosidosis, has been hampered by the lack of information on the human MANB gene. In this study we have identified a cDNA sequence encoding a functional lysosomal ␣-mannosidase activity. The enzyme has been expressed, purified, and characterized in quantities sufficient for potential use in enzyme replacement therapies. In addition, the identification of the cDNA sequence of the functional ␣-mannosidase cDNA will aid in the identification of the molecular basis of the lysosomal storage disease and potentially lead to therapies for affected patients.