Nature of the Accumulated Glycoconjugates
The deficient activity of α-N-acetylgalactosaminidase (EC 220.127.116.11) in all disease subtypes leads to an accumulation of glycoconjugates with terminal α-N-acetylgalactosaminyl moieties in various tissues and fluids. Table 139-1 lists some of the major α-N-acetylgalactosaminyl-containing glycoconjugates detected in normal human tissues. α-N-Acetylgalactosaminyl residues occur in O-linked and N-linked glycopeptides and glycoproteins (e.g., mucins, blood group A substances), in glycosphingolipids (e.g., Forssman glycolipid), and in proteoglycans (e.g., cartilage keratan sulfate II).39–41 Potentially, each of these compounds can be the source of substrates that accumulate in this disease. However, the nature and concentrations of the accumulating glycoconjugate substrates in tissues of patients with α-N-acetylgalactosaminidase deficiency remain the subject of investigation, because all of the identified patients with this disease are living (except for a type I patient diagnosed retrospectively and not autopsied15), and the only sources available for analysis have been urine, blood, tissue biopsies, and cultured cells.
Table 139-1: Various Glycoconjugates Containing α-N-acetylgalactosaminyl Residues
Glycopeptides and Glycoproteins.
Glycopeptides and glycoproteins with O-linked oligosaccharide moieties accumulate in patients with α-N-acetylgalactosaminidase deficiency.22,24,25,42 Various structures containing O-linked α-N-acetylgalactosaminyl moieties are shown in Table 139-1. The mucins most commonly contain a Galβ(1→3)GalNAcα(1→O)Ser/Thr core structure and additional N-acetylneuraminyl, fucosyl, N-acetylglucosaminyl and/or galactosyl residues.43–46 In addition, a variety of N-linked glycopeptides and glycoproteins contain α-N-acetylgalactosaminyl moieties, including the predominant N-linked glycopeptide structure in human erythrocytes, which has a complex-type triantennary oligosaccharide composed of 20 to 30 residues with a terminal A blood group trisaccharide.39
High-resolution thin-layer chromatography of the excreted urinary glycopeptides accumulated in patients with all three subtypes revealed four major compounds that were identified by comigration with known standards1,3,10,13,14,22 (Fig. 139-11); these glycopeptides were further characterized by structural analyses.24,25,42,47,48 For example, the structure of the tetrasaccharide monopeptide below was determined at high resolution without prior derivatization from urine of the two type I-disease siblings by electrospray mass spectrometry.42 The structures of the accumulated glycopeptides are shown in Table 139-2.
Thin-layer chromatographic profiles of urinary oligosaccharides (A) and glycopeptides (B) from a normal individual (lane 1), the affected brothers with type I disease (lanes 2 and 3), and the Japanese female with type II disease (lane 4). The four major abnormal glycopeptide bands in the patients' urine comigrated with the glycopeptide standards, NeuNAcα(2 → 3)Galβ(1 → 3) [NeuNAcα(2 → 3)Galβ(1 → 4)GlcNAcα(2 → 6)]GalNAcα(1 → O)-Ser/Thr (lane 5), NeuNAcα(2 → 3)Galβ(1 → 3)[Neu-NAcα(2 → 6)]GalNAcα (1→O)-Ser/Thr (lane 6), NeuNAcα(2→3)Galβ(1→3)GalNAcα(1→O)-Ser/Thr (lane 7), GalNAc (lane 8A), and GalNAcα(1 → O)-Ser/Thr (lane 8B). (From Schindler et al.22 Used with permission.)
Table 139-2: Accumulated Urinary Compounds in Types I and II Schindler Disease |Favorite Table|Download (.pdf) Table 139-2: Accumulated Urinary Compounds in Types I and II Schindler Disease
| Oligosaccharide |
|Blood group A trisaccharide* ||Ga1NAα(1→3) Gal (2?1) αFuc |
| Glycopeptides |
|1 GalNAcα (1→O)-Ser/Thr |
|2 NeuNAcα (2→3) Galβ (1→3) GalNAcα (1→O)-Ser/Thr |
|2a Galβ(1→3) [NeuNAcα (2→6] GalNAcα(1→O)-Ser/Thr |
|3 NeuNAcα(2→3) Galβ (1→3) [NeuNAcα (2→6)] GalNAcα (1→O)-Ser/Thr |
| ||Pro |
|3a NeuNAcα (2→3) Galβ (1→3) [NeuNAcα (2→6)] GalNAcα (1→O)-Thr/Thr |
| ||Pro |
|4 NeuNAcα (2→3) Galβ (1→3) [NeuNAcα (2→3)] GalNAcβ (1→4) GlcNAcβ (1→6) GlcNAcα (1→O)-Ser/Thr |
|4a Galβ (1→3) [NeuNAcα (2→3] Galβ (1→3) [Galβ (1→4) GlcNAcβ (1→6)] GalNAcα (1→O)-Ser/Thr |
|4b NauNAcα(2→3) Galβ (1→3) [Galβ (1→4) GlcNAcβ (1→6)] GalNAcα (1→O)-Ser/Thr |
|5 (NeuNAc)2 (Galβ (1→4) GlcNAc)2 Galβ (1→3) GlcNAcα(1→O)-Ser/Thr |
The type I proband who was blood group A-positive also excreted the blood group A trisaccharide at a concentration about fivefold greater than that in age-matched normal urines1 (Table 139-2). The urine from the type I and II patients also contained the dipeptide tetrasaccharide, NeuNAcα(2→3)Galβ(1→3) [NeuNAcα(2→6)]GalNAcα(1→O)-Thr-Pro.24,25,47 In addition, amino acid analysis of the peptide moiety of this urinary sialoglycopeptide from the type I patients revealed threonine, serine, proline, glycine, and glutamic acid in a molar ratio of 6.3:2.7:2.6:1.7:1.0.25 It is assumed that the structures of the excreted glycopeptides and oligosaccharides in the type III patients are identical to those in the type I and II patients. However, the amounts of the accumulating glycopeptides and oligosaccharides in the urine of patients with each subtype (except type I) have not been reported, nor have the nature and concentrations of the glycopeptides and/or oligosaccharides accumulating in tissues been reported.
The structures of the accumulated sialoglycopeptides in the urines of the type I and II patients were remarkably similar to those excreted by patients with sialidosis (isolated α-neuraminidase deficiency)49 and galactosialidosis (combined α-neuraminidase and β-galactosidase deficiencies) due to mutations in the gene encoding the protective protein.50 The fact that most of the same sialoglycopeptides accumulate in these disorders suggests that O-linked sialoglycopeptides of the mucin type are hydrolyzed by the α-neuraminidase/β-galactosidase/protective protein complex, which also appears to contain α-N-acetylgalactosaminidase.51,52 It is intriguing to speculate that mutations in any one of the enzymes can alter the ability of this complex to degrade substrates that would normally be sequentially hydrolyzed by each of these exoglycosidases (see below).
Glycosphingolipids with α-N-acetylgalactosaminyl residues include those with blood group A activity and the Forssman glycolipid. The blood group A glycosphingolipids are characterized by the terminal blood group A-specifying trisaccharide, GalNAcα(1→3)Gal(2←1)αFuc.53 They contain varying numbers of internal Galβ(1→3)GlcNAcβ(1→3) (type I) and/or Galβ(1→4)GlcNAcβ(1→3) (type 2) units and the core structure, Galβ(1→4)Glcβ(1→1′)Cer.39,54,55 In humans, Forssman antigen has been detected in low amounts in normal lung and was increased fivefold in lung carcinomas, regardless of blood type.56,57 Forssman antigen also has been found in intestinal mucosal carcinoma, but not in normal mucosal cells.58 The complete chemical synthesis of this neutral glycosphingolipid has been accomplished.59
The individual neutral glycosphingolipids in plasma, erythrocytes, urinary sediment, and a cortical biopsy from the type I German sibs were quantitatively analyzed and were found to be within their respective normal ranges.3 The metabolism of the endogenous neutral glycosphingolipids in cultured fibroblasts from the type I patients has been investigated in greater detail using [14C]galactose to label these compounds metabolically.60 There were no significant differences in glycosphingolipid biosynthesis or degradation in the cultured cells from the type I patients and normal individuals. High-resolution thin-layer chromatographic analysis of the individual glycolipids revealed virtually identical steady state compositions. Forssman glycosphingolipid was not detected in normal or type I disease fibroblasts using a very sensitive immuno-overlay technique with monoclonal anti-Forssman antibody.60 The addition of α-N-acetylgalactosaminyl-containing glycolipids including [ceramide-3H]Forssman glycosphingolipid, the more water-soluble fluorescent C6-NBD-lyso-Forssman glycolipid, and the blood group A derivative neoglycolipid, GalNAc(Fuc)GalPEN [14C]Ac, to cell extracts or cultured fibroblasts from the type I patients resulted in the accumulation of these substrates, consistent with the enzymatic defect.
The degradation of blood group A glycosphingolipid, A-6-2, GalNAcα(1→3)(Fucα(1→2)(Galβ(1→4)GlcNAcβ(1→3) Galβ (1→4)Glcβ(1→1′)Cer, [3H]-labeled in its ceramide moiety, also has been studied in cultured fibroblasts from patients with α-N-acetylgalactosaminidase deficiency.61 In normal cells, the expected metabolites included various less-polar products and some lipids resulting from reuse of the the liberated sphingosine, mainly sphingomyelin and phosphatidylcholine. In α-N-acetylgalactosaminidase-deficient cells, blood group glycolipid A-6-2 was not degraded. It has been estimated that approximately 0.5 mg of the blood group glycolipids are degraded per day.61 Thus, a significant amount of blood group A-active glycolipids would accumulate in blood type A or AB patients. Although overt neutral glycosphingolipid accumulation was not detected in cultured fibroblasts, plasma, erythrocytes, urinary sediment, or cerebral cortex from the blood type A type I patient, it is possible that individual α-N-acetylgalactosaminyl-containing glycosphingolipids, including the blood group A glycolipids, may accumulate in various tissues and cause disease pathology. However, disease manifestations due to such glycosphingolipid accumulation may be insignificant, but the analysis of the tissue glycolipids in type I patients is required to provide this information.
A keratan sulfate type II mucopolysaccharide containing an O-linked α-N-acetylgalactosaminyl residue has been identified in normal human skeletal cartilage.41 To date, the accumulation of this mucopolysaccharide in skeletal cartilage from α-N-acetylgalactosaminidase-deficient patients has not been investigated. In addition, the urinary glycosaminoglycans in patients with all three disease subtypes were not increased, and the glycosaminoglycan metabolism in cultured fibroblasts from type I patients was normal when assayed by [35S]sulfate pulse-chase studies.3
α-N-Acetylgalactosaminidase: The Enzymatic Defect
The deficient activity of lysosomal α-N-acetylgalactosaminidase is the specific enzymatic defect in types I, II, and III Schindler disease.1–5,7,11,12 Initially, the enzymatic defect was identified in the type I patients by determining the pNP-αGalNAc activity in plasma, leukocytes, and cultured fibroblasts from the affected sibs and other family members.1,2 Using this assay, the affected type I sibs had apparent residual activities of less than 2 percent of the respective normal mean values, and their heterozygous parents had reduced activities at the lower end of the normal range. Subsequent assays performed with the more sensitive fluorogenic substrate, 4-methylumbelliferyl-α-D-N-acetylgalactosaminide (4MU-αGalNAc),3,62 detected 0.5 to 2 percent residual activity in plasma, cultured fibroblasts, or cultured lymphoblasts from the affected type I, II, and III patients and provided significantly improved heterozygote discrimination3,11–14,63 (Table 139-3).
Table 139-3: -N-Acetylgalactosaminidase Activity in Schindler Disease Subtypes |Favorite Table|Download (.pdf) Table 139-3: -N-Acetylgalactosaminidase Activity in Schindler Disease Subtypes
|Source || Plasma || Lymphoblasts || Fibroblasts |
| ||(Percent of Normal Enzyme Activity) |
| Type I disease || || || |
| German family 3,5 || || || |
|Affected homozygote, sib ||0.5 ||0.5 ||0.7 |
|Affected homozygote, sib ||1.1 ||0.7 ||1.2 |
|Normal, brother ||120.0 ||77.8 ||69.3 |
|Heterozygote, mother ||53.0 ||31.5 ||24.2 |
|Heterozygote, father ||53.0 ||29.4 ||41.2 |
|Heterozygote, maternal uncle ||54.9 ||39.8 ||— |
|Affectd homozygote, cousin ||1.1 ||— ||1.0 |
| Type II disease || || || |
| Japanese family 11,12 || || || |
|Affected homozygote ||1.1 ||0.4 ||0.7 |
|Normal husband ||93.4 ||117.0 ||— |
|Heterozygote, son ||65.2 ||39.1 ||23.2 |
|Heterozygote, daughter ||64.1 ||22.0 ||— |
|Heterozygote, sib ||26.0 ||36.4 ||45.2 |
|Heterozygote, sib ||55.4 ||19.6 ||51.7 |
|Heterozygote, sib ||49.0 ||32.6 ||67.5 |
| Spanish family 13 || || || |
|Affected homozygote, sib ||0.5 ||— ||0.7 |
|Affected homozygote, sib ||— ||— ||1.6 |
| Type III disease || || || |
| Dutch family 14 || || || |
|Affected homozygote, sib ||4.0 ||— ||5.0 |
|Affected homozygote, sib ||3.0 ||— ||— |
|Normal sister ||104.0 ||— ||— |
|Heterozygote, mother ||32.0 ||— ||— |
|Heterozygote, father ||26.0 ||— ||— |
| French family 16 || || || |
|Affected homozygote ||— ||— ||2.0 |
| Normal Individuals* || || || |
| ||(nmol/h/ml) ||nmol/h/mg protein) |
| || || |
|Mean ± 1 SD ||18.4 ± 5.1 ||44.5 ± 11.2 ||213.5 ± 56.1 |
|Range ||10.9 – 30.0 ||21.7 – 63.4 ||141.3 – 350.7 |
|n ||104 ||50 ||34 |
Immunologic studies have been conducted to further characterize the nature of the enzymatic defect in the patients with type I, II, and III probands using monospecific rabbit antihuman α-N-acetylgalactosaminidase antibodies raised against the homogeneous lung enzyme (see “Purification” under “Human α-N-Acetylgalactosaminidase” below).3,11 Immunoblotting studies revealed two bands of cross-reactive immunologic material (CRIM), estimated to be 48 and 117 kDa, respectively, in the purified enzyme preparation. The same two bands were present in fibroblast extracts from obligate heterozygotes and normal individuals (Fig. 139-12), consistent with the occurrence of monomeric and homodimeric forms of the normal enzyme (see “Physical Properties” under “Human α-N-Acetylgalactosaminidase” below). Immunoblotting revealed the absence of detectable enzyme protein in fibroblast extracts from patients with all three subytpes;3,11,15 that neither immunoreactive band was detectable suggested that the two enzyme forms were encoded by the same gene (see Fig. 139-12). These findings also indicated that the enzymatic deficiency in all three subtypes resulted from different mutations that impaired enzyme synthesis or stability, consistent with the fact that only a small amount (<2 percent) of α-N-acetylgalactosaminidase activity was detected in cells from the affected patients.
Immunoblot of human α-N-acetylgalactosaminidase. A, Fibroblast extracts from family members with type I disease: father (lane 1), mother (lane 2), unaffected brother (lane 3), and affected brothers (lanes 4 and 5), a normal individual (lane 6), and purified lung enzyme (lane 7). B, Fibroblast extracts from the Japanese family members with type II disease: proband (lane 1), her son (lane 2), and a normal individual (lane 3). Note that both the 48- and 117-kDa enzyme bands are absent in the type I and II probands. (A from Schindler et al.3 and B from Kanzaki et al.11 Used with permission.)
Previously, human α-N-acetylgalactosaminidase was known as lysosomal α-galactosidase B.64–70 The confusion concerning the substrate specificity of this enzyme originally resulted from the use of artificial substrates in the identification and determination of the enzymatic defect in Fabry disease.71–75 In 1970, Kint71 was the first to use p-nitrophenyl-α-D-galactopyranoside (pNP-αGal) and 4-methylumbelliferyl-α-D-galactopyranoside (4MU-αGal) to demonstrate the marked, but not total, deficiency of α-galactosidase activity in hemizygotes with Fabry disease (see Chap. 150). Kint71,76,77 and others72–75,78–80 subsequently demonstrated that two enzymes cleaved the synthetic, water-soluble pNP-αGal and 4MU-αGal substrates. One enzyme, designated α-galactosidase A, was thermolabile and inhibited by myoinositol, while the other, designated α-galactosidase B, was thermostable, had a higher Km toward artificial galactosides, and was not inhibited by myoinositol. α-Galactosidase A was deficient in Fabry disease, whereas α-galactosidase B was responsible for the residual α-galactosidase activity in Fabry hemizygotes when water-soluble substrates were used.71–75,81,82
Electrophoretic and isoelectric focusing studies revealed multiple forms of α-galactosidase A in various sources from normal individuals, the major components having pI values of 4.3 to 5.1.79,83–85 In contrast, only a single form of α-galactosidase B was observed with a pI value of about 4.4.79,83–85 Neuraminidase treatment of normal α-galactosidase A or B in crude tissue extracts and in purified preparations converted the multiple α-galactosidase A forms to a single activity band at pI 5,83,86 whereas the migration of α-galactosidase B was not altered by neuraminidase treatment.74,83,85–89 These and other findings initially suggested that α-galactosidases A and B were enzyme forms with similar substrate specificities that differed in their glycosylation.73,74,76,77,89 In addition, partially purified human α-galactosidase B was shown to hydrolyze, albeit very inefficiently, the natural glycosphingolipid substrate, globotriaosylceramide, when incubated in the presence of a heat-stable glycoprotein activator.87,90 However, subsequent studies of these enzymes64,79,85,91 demonstrated kinetic, structural, and immunologic differences, which indicated that they were not glycoforms, but were distinct proteins.
In 1977, Dean et al.65,66 and Schram et al.64 independently showed that α-galactosidase B actually was an α-N-acetylgalactosaminidase that was competitively inhibited by α-N-acetylgalactosamine. Purified α-galactosidase B was shown to have a higher affinity for α-N-acetylgalactosaminides (Km ≈ 1 mM) than for α-galactosides (Km ≈ 20 mM), as well as the same physical properties (e.g., thermostability and pH optimum) as human α-N-acetylgalactosaminidase purified from liver.92 Thus, it was concluded that α-galactosidase B was actually an α-N-acetylgalactosaminidase. This enzyme was not a focus of interest or investigation until it was found deficient in type I Schindler disease.1–5
Table 139-4 summarizes the purification and physical properties of human α-N-acetylgalactosaminidase isolated from various sources. The enzyme has been purified to homogeneity from placenta,79,85 liver,90,92 spleen,93 and lung.94 The purification strategy was similar to that for α-galactosidase A (see Chap. 150). Concanavalin A Sepharose chromatography was typically employed as a first step to enrich the glycoprotein fraction from crude tissue extracts. DEAE-cellulose chromatography afforded additional purification and nearly complete separation of α-N-acetylgalactosaminidase from α-galactosidase A, the latter being eluted at a lower salt concentration.70,79,83,94,95 Hydroxylapatite chromatography completely separated the two enzymes and provided a significant purification step.83,90,94 Gel filtration85,90 or chromatography using ampholyte displacement,90 SP-Sephadex,85 or butyl agarose85 has been used as a final purification step. However, affinity chromatography on α-galactosylamine coupled to Sepharose provided the most efficient purification step for both α-N-acetylgalactosaminidase and α-galactosidase A. The former can be specifically eluted with N-acetylgalactosaminide,93,94 whereas both can be selectively eluted with D-galactose.83
Table 139-4: Purification and Properties of Human -N-Acetylgalactosaminidase |Favorite Table|Download (.pdf) Table 139-4: Purification and Properties of Human -N-Acetylgalactosaminidase
| || || || || || ||Molecular Weight || |
| || || || || || || || |
|Source ||Specific Activity (nmol/h/mg) ||Substrate ||pH Optimum ||pl || Km (mM) ||Native (kDa) ||Subunit (kDa) ||Reference |
|Placenta ||67,000 ||4MU-αGal ||4.5 ||4.4 || ||150 || ||79 |
|Placenta ||271,000 ||4MU-αGal ||4.4 ||4.4 ||13.0 ||117 ||47.7 ||85 |
|Liver ||4,580 ||pNP-αGalNAc ||4.3 || || || || ||92 |
|Liver ||4,740 ||pNP-αGalNAc || || ||1.0 ||110 || ||64,69 |
|Liver ||14,800 ||4MU-αGal || ||4.9 ||11.7 || ||86 |
|Liver ||394,000 ||4MU-αGal ||4.8 ||4.5 ||6.8 ||90 || ||70,90 |
|Liver ||2,900,000 ||oNP-αGalNAc ||4.3 || ||1.3 ||90 || ||70,90 |
|Lung ||370,000 ||4MU-αGal ||4.6 ||4.6 || ||117 ||45 ||94 |
|Liver ||7 ||4MU-αGal ||4.6 ||4.3 ||4.0 ||90 || ||74 |
|Liver ||251,000 ||4MU-αGal || || || || ||46.8 ||93 |
Human α-N-acetylgalactosaminidase is a relatively thermostable enzyme.64,67,79,83,92 At 50°C, the half-life for the enzyme at pH 4.8 to 7.0 is greater than 2 h.64,67,70,75,82,92,96 The enzyme has been estimated to have a native molecular weight of 90 to 117 kDa by gel filtration64,69,70,74,79,93 (see Table 139-4). The subunit molecular weights of the enzyme from human placenta and liver have been reported to be 47.7 kDa and 46.8 kDa, respectively.85,93 SDS polyacrylamide gel electrophoresis of homogeneous α-N-acetylgalactosaminidase from human lung revealed species of ≈48 and ≈117 kDa.17 When the purified lung enzyme was subjected to native polyacrylamide gel electrophoresis, only a single species of ≈117 kDa was detected. Immunoblotting of the purified enzyme or crude human fibroblast extracts (after SDS gel electrophoresis) with rabbit antihuman lung α-N-acetylgalactosaminidase antibodies demonstrated two immunoreactive species of 48 and 117 kDa, respectively. Notably, both species were present in immunoblots of normal human fibroblasts, but neither was detected in fibroblasts from the patients with type I, II, or III disease,3,11,15,97 suggesting that both species were encoded by a single gene (see Fig. 139-12). Microsequencing of both the 48- and 117-kDa species revealed identical N-terminal amino acid sequences. Subsequent tryptic mapping of the 48- and 117-kDa species determined that the maps were identical and that the 117-kDa species was the homodimeric form of the 48-kDa species.17 These findings indicated that the enzyme is a homodimer which is highly resistant to denaturation. Evidence for the homodimeric structure also is based on biosynthetic studies of α-N-acetylgalactosaminidase in human fibroblasts.98,99 Further support for the homodimeric structure is inferred from studies of the porcine enzyme, which was a dimer with native and subunit molecular weights of 102 and 52 kDa, respectively.100
N-Linked Oligosaccharide Structures.
The carbohydrate content and oligosaccharide moieties present in the human enzyme have been reported.85,98,101 The human placental enzyme contained 5 percent neutral sugars and 0.3 percent sialic acid.85 Analysis of the radiolabeled oligosaccharides from immunoprecipitated human fibroblast α-N-acetylgalactosaminidase revealed high-mannose-type oligosaccharide structures (Man8–9GlcNAc) on both precursor and processed enzyme forms; only the precursor had phosphorylated mannose residues.98 The virtual absence of complex-type oligosaccharide moieties containing terminal sialic acid residues was supported by the previously mentioned neuraminidase studies, which did not alter the pI of the enzyme.83
More recently, the structures of the N-linked oligosaccharides on the purified secreted forms of α-N-acetylgalactosaminidase overexpressed in Chinese hamster ovary cells were determined.101 The structures were liberated by hydrazinolysis and were analyzed by HPLC resolution and digestion by a series of exoglycosidases. Hydrazinolysis liberated neutral (46.1 percent), monosialyl (21.3 percent), disialyl (6.2 percent), trisialyl (1.6 percent), monophosphoryl (17.1 percent), diphosphoryl (5.5 percent), and monosialyl monophosphoryl oligosaccharides (2.2 percent). The neutral fraction contained six types—high mannose, hybrid, and mono-, bi-, tri- and tetra-antennary complex-type oligosaccharides—in which Man5–6GlcNac2, GalGlcNAcMan6GlcNAcFucGlcNAc, 2,6-branched Gal3GlcNAc3Man3GlcNAcFucGl-cNAc, and Gal2GlcNAc2Man3GlcNAc(Fuc)0–1GlcNAc were predominant. The sialic acid residues were exclusively α2,3-linked, and the desialylated oligosaccharides predominantly had structures of fucosylated and fully galactosylated di-, 2,4 and 2,6-branched tri-, and tetraantennary structures. Phosphate groups were exclusively in the monoester form. The major dephosphorylated oligosaccharides were Man6–7GlcNAc2 and GalGlcNAcMan6GlcNAcGlcNAc. The monosialyl monophosphoryl oligosaccharides had core structures of GalGlcNAcMan6GlcNAc(Fuc)0–1GlcNAc. When compared to the oligosaccharide structures on the secreted recombinant human α-galactosidase A, which is evolutionarily related and highly homologous, both enzymes' oligosaccharides had similar phosphorylation patterns. However, α-N-acetylgalactosaminidase contained more sialylated structures (8.5 vs. 28.9 percent of total oligosaccharides), indicating that the recombinant enzyme had more complete complex-type structures.
Human lysosomal α-N-acetylgalactosaminidase had an acidic pH optimum of about 4.6 for various synthetic chromogenic or fluorogenic substrates (Table 139-5). The apparent Km values for the enzyme toward synthetic α-galactosides were high, approximately 12 to 20 mM for 4MU-αGal and 20 to 30 mM for pNP-αGal. In contrast, the enzyme’s affinity for the α-N-acetylgalactosamine glycon was about tenfold lower, with apparent Km values for oNP-α-N-acetylgalactosaminide (oNP-αGalNAc) ranging from 1 to 2 mM. The apparent Km toward the fluorogenic substrate, 4MU-αGalNAc, for the purified human lung enzyme was 1.6 mM.63 Neither pNP-αGalNAc nor 4MU-αGalNAc was hydrolyzed by purified human α-galactosidase A.63 For purified human liver α-N-acetylgalactosaminidase, the turnover number using oNP-αGalNAc as substrate was calculated to be 6600 min−1 using a molecular weight for the native enzyme of 110 kDa.84 The apparent Km of human liver α-N-acetylgalactosaminidase toward the Forssman glycosphingolipid (GbOseCer, globopentaosylceramide) was 0.59 mM with a Vmax of 1.6 × 105 nmol/h/mg at the pH optimum, 3.9 (see Table 139-5). Other possible natural substrates, including O-linked glycopeptides with terminal α-N-acetylgalactosaminyl residues, have not been used to characterize the kinetic properties of this enzyme. Moreover, kinetic studies of the human enzyme have been limited to characterization of its exoglycosidase activity. Use of substrates with internal α-N-acetylgalactosaminyl moieties could definitively determine if the enzyme can also function as an endoglycosidase.
Table 139-5: Kinetic Properties of Human -N-Acetylgalactosaminidase |Favorite Table|Download (.pdf) Table 139-5: Kinetic Properties of Human -N-Acetylgalactosaminidase
|Source ||Substrate ||pH Optimum || K m (mM) || V max (nmol/h/mg ||Reference |
|Liver ||4MU-αGal ||4.8 ||6.8 ||1.1 × 106 ||65,90 |
|Liver ||oNP-αGalNAc ||4.3 ||1.3 ||3.6 × 106 ||65,90 |
|Liver ||GbOse5 * ||4.4 ||3.7 ||5.5 × 105 ||65 |
|Liver ||GbOse5Cer ||3.9 ||0.59 ||1.6 × 105 ||65,90 |
|Liver ||GbOse3Cer ||4.3 ||0.35 ||1.1 × 104 ||65,90 |
|Placenta ||4MU-αGal ||4.4 ||13.0 || ||83 |
|Lung ||4MU-αGalNAc ||4.5 ||0.37 || ||63 |
|Liver ||pNP-αGalNAc ||4.3 ||3.1 || ||92 |
α-N-Acetylgalactosaminidase hydrolysis of globotriaosylceramide (GbOseCer), the primary glycosphingolipid substrate for human α-galactosidase, occurs at an extremely low rate, if at all, in vitro. α-N-Acetylgalactosaminidase did not cleave GbOse3Cer in cultured fibroblasts or tissues from Fabry hemizygotes,75,91 and minimal, if any, hydrolysis was detected using radiolabeled GbOse3Cer and/or highly purified enzyme65,84 (see Table 139-5). Inhibitors can be used to discriminate α-N-acetylgalactosaminidase and α-galactosidase activities in assays with chromogenic or fluorogenic α-galactosides. α-N-Acetylgalactosaminidase activity is inhibited about 95 percent by 100 mM N-acetylgalactosamine (GalNAc), while it has no effect on α-galactosidase A activity.102 Conversely, 500 mM myoinositol inhibits about 50 percent of α-galactosidase A activity without affecting the activity of α-N-acetylgalactosaminidase.79,96 The Ki of α-N-acetylgalactosaminidase for GalNAc was 3.1 mM with pNP-αGalNAc.92
For comparison, purified α-N-acetylgalactosaminidase from porcine and bovine liver had a pH optimum of 4.3 to 4.7.103,104 The bovine α-N-acetylgalactosaminidase specifically cleaved pNP-αGalNAc with a Km of 6.5 mM and had a Ki of 10 mM with GalNAc.103 Additionally, partially purified α-N-acetylgalactosaminidases from bovine liver and ox spleen cleaved the terminal α-N-acetylgalactosaminyl residues that were O-linked to serine or threonine in desialylated ovine and bovine submaxillary mucins.103,105
Early studies of the biosynthesis and processing of human α-N-acetylgalactosaminidase indicated that the enzyme was synthesized in cultured fibroblasts as a 65-kDa precursor and then was processed to a 48-kDa mature form.98 Subsequent biosynthetic studies in cultured fibroblasts using normal and mutant cell lines indicated that the normal enzyme precursor was 52 kDa and that the 65-kDa polypeptide was not a glycoprotein like the other lysosomal enzymes, and probably was not related to α-N-acetylgalactosaminidase.99 After a 4-h pulse labeling and 17-h chase, the 52-kDa precursor glycopolypeptide was completely replaced by a mature enzyme subunit of ≈49 kDa. Treatment of both the 52-kDa precursor and the 49-kDa mature polypeptides with endoglycosidase F resulted in the formation of a deglycosylated 42-kDa species. These findings indicated that after cleavage of a signal peptide, the enzyme subunit was normally modified by carbohydrate (rather than proteolytic) processing, thereby accounting for the 3-kDa difference between the precursor and the mature enzyme subunits.
Notably, the E325K mutation (see below) in type I patients did not interfere with the synthesis of the 52-kDa precursor. However, the mutant precursor apparently was not phosphorylated and disappeared during the 17-h chase. Because the mutant precursor was not detected in the media, the absent conversion to the immunologically detectable mature enzyme subunit suggested that the mutation in the type I patients caused an unstable precursor polypeptide that was subject to rapid intracellular degradation. These results presumably were limited by the sensitivity of immunodetection, because recent biosynthetic studies of the E325K type I enzyme expressed in COS-1 cells indicated that the mature 48-kDa enzyme subunit was synthesized, but unstable.106 Similar findings were observed for the biosynthesis of the mature mutant R329W enzyme subunit expressed in COS-1 cells, but the Japanese type II mutant polypeptide was slightly more stable than the German type I polypeptide.106 Interestingly, studies of cultured fibroblasts from a Dutch type III patient revealed normal synthesis of the precursor and the production of a small amount of mature enzyme subunit.15