Cell-Mediated Immunity to NAGLU Transgene Following Intracerebral Gene Therapy in Children With Mucopolysaccharidosis Type IIIB Syndrome.

Mucopolysaccharidosis type IIIB syndrome (Sanfilippo disease) is a rare autosomic recessif disorder caused by mutations in the α-N-acetylglucosaminidase (NAGLU) gene coding for a lysosomal enzyme, leading to neurodegeneration and progressive deterioration of cognitive abilities in affected children. To supply the missing enzyme, several recent human gene therapy trials relied on the deposit of adeno-associated virus (AAV) vectors directly into the brain. We reported safety and efficacy of an intracerebral therapy in a phase 1/2 clinical trial (https://clinicaltrials.gov/ct2/show/NCT03300453), with a recombinant AAV serotype 2/5 (rAAV2/5) coding human NAGLU in four children with MPS IIIB syndrome receiving immunosuppression. It was reported that AAV-mediated gene therapies might elicit a strong host immune response resulting in decreased transgene expression. To address this issue, we performed a comprehensive analysis of cellular immunity and cytokine patterns generated against the therapeutic enzyme in the four treated children over 5.5 years of follow-up. We report the emergence of memory and polyfunctional CD4+ and CD8+ T lymphocytes sensitized to the transgene soon after the start of therapy, and appearing in peripheral blood in waves throughout the follow-up. However, this response had no apparent impact on CNS transgene expression, which remained stable 66 months after surgery, possibly a consequence of the long-term immunosuppressive treatment. We also report that gene therapy did not trigger neuroinflammation, evaluated through the expression of cytokines and chemokines in patients' CSF. Milder disease progression in the youngest patient was found associated with low level and less differentiated circulating NAGLU-specific T cells, together with the lack of proinflammatory cytokines in the CSF. Findings in this study support a systematic and comprehensive immunomonitoring approach for understanding the impact immune reactions might have on treatment safety and efficacy of gene therapies.

Differential Contribution of Acute and Chronic Inflammation to the Development of Murine Mammary 4T1 Tumors.

Based on the notion that inflammation favors tumorigenesis, our experiments comparatively assessed the influence of acute and chronic inflammation on the development of a murine mammary tumor (4T1). In addition, we characterized angiogenic and inflammatory markers in the tumor tissue and systemically. Subcutaneous implantation of polyether-polyurethane sponge discs in Balb/c mice was used to host 4T1 tumor cells (1x10(6)), which were inoculated intraimplant 24 h or 10 days post implantation. Flow cytometric analysis of enzyme-digested implants revealed that, after 24 hours, the population of leukocytes was primarily characterized by neutrophils (42.53% +/- 8.45) and monocytes (37.53% +/- 7.48), with some lymphocytes (16.27% +/- 4.0) and a few dendritic cells (1.82% +/- 0.36). At 10 days, macrophages were predominant (37.10% +/- 4.54), followed by lymphocytes (28.1% +/- 4.77), and monocytes (22.33% +/- 3.05), with some dendritic cells (13.60% +/- 0.55) and neutrophils (11.07% +/- 2.27). A mammary tumor grown in a chronic inflammatory environment was 2-fold when compared with one grown in acute inflammation and 5-fold when compared with tumor alone. The levels of pro-angiogenic cytokine (VEGF-Vascular Endothelial Growth Factor) were higher in implant-bearing tumor when 4T1 cells were grown in 10-day old implants as compared to the VEGF levels of the two other groups. Overall, the levels of the inflammatory markers evaluated (NAG -N-acetylglucosaminidase, TNF-α-Tumor Necrosis Factor-α) were higher in both groups of implant-bearing tumors and in serum from those animals when compared with the tumor alone levels. This inflammation-related difference in tumor growth may provide new insights into the contribution of different inflammatory cell populations to cancer progression.

Feasibility and safety of systemic rAAV9-hNAGLU delivery for treating mucopolysaccharidosis IIIB: toxicology, biodistribution, and immunological assessments in primates.

No treatment is currently available for mucopolysaccharidosis (MPS) IIIB, a neuropathic lysosomal storage disease caused by autosomal recessive defect in α-N-acetylglucosaminidase (NAGLU). In anticipation of a clinical gene therapy treatment for MPS IIIB in humans, we tested the rAAV9-CMV-hNAGLU vector administration to cynomolgus monkeys (n=8) at 1E13 vg/kg or 2E13 vg/kg via intravenous injection. No adverse events or detectable toxicity occurred over a 6-month period. Gene delivery resulted in persistent global central nervous system and broad somatic transduction, with NAGLU activity detected at 2.9-12-fold above endogenous levels in somatic tissues and 1.3-3-fold above endogenous levels in the brain. Secreted rNAGLU was detected in serum. Low levels of preexisting anti-AAV9 antibodies (Abs) did not diminish vector transduction. Importantly, high-level preexisting anti-AAV9 Abs lead to reduced transduction in liver and other somatic tissues, but had no detectable impact on transgene expression in the brain. Enzyme-linked immunoabsorbent assay showed Ab responses to both AAV9 and rNAGLU in treated animals. Serum anti-hNAGLU Abs, but not anti-AAV9 Abs, correlated with the loss of circulating rNAGLU enzyme. However, serum Abs did not affect tissue rNAGLU activity levels. Weekly or monthly peripheral blood interferon-γ enzyme-linked immunospot assays detected a CD4(+) T-cell (Th-1) response to rNAGLU only at 4 weeks postinjection in one treated subject, without observable correlation to tissue transduction levels. The treatment did not result in detectable CTL responses to either AAV9 or rNAGLU. Our data demonstrate an effective and safe profile for systemic rAAV9-hNAGLU vector delivery in nonhuman primates, supporting its clinical potential in humans.

Generation of an anti-NAGase single chain antibody and its application in a biosensor-based assay for the detection of NAGase in milk.

Bovine mastitis, an inflammation of the mammary gland in cows, is a major challenge for the dairy industry worldwide as it lowers milk yield, reduces milk quality and increases overall production costs. Early diagnosis is of the utmost importance. N-acetyl-ß-D-glucosaminidase (NAGase) is an enzyme released into milk during inflammation and acts as an early indicator of mastitis. This paper describes the selection of anti-NAGase single chain fragment variable antibodies (scFv) from naïve human antibody libraries and their incorporation into an automated optical biosensor-based immunoassay to detect NAGase in milk. The scFv with the highest affinity for NAGase was first characterized by inhibition ELISA, followed by further evaluation using a surface plasmon resonance platform. Purified NAGase was immobilized on the surface of a CM5 chip and spiked NAGase milk samples were analyzed. The limit of detection for the assay for the assay was determined as 1µg/ml.

Toxic effects of hyperglycemia are mediated by the hexosamine signaling pathway and o-linked glycosylation in early mouse embryos.

Maternal hyperglycemia is believed to be the metabolic derangement associated with both early pregnancy loss and congenital malformations in a diabetic pregnancy. Using an in vitro model of embryo exposure to hyperglycemia, this study questioned if increased flux through the hexosamine signaling pathway (HSP), which results in increased embryonic O-linked glycosylation (O-GlcNAcylation), underlies the glucotoxic effects of hyperglycemia during early embryogenesis. Mouse zygotes were randomly allocated to culture treatment groups that included no glucose (no flux through HSP), hyperglycemia (27 mM glucose, excess flux), 0.2 mM glucosamine (GlcN) in the absence of glucose (HSP flux alone), and O-GlcNAcylation levels monitored immunohistochemically. The impact of HSP manipulation on the first differentiation in development, blastocyst formation, was assessed, as were apoptosis and cell number in individual embryos. The enzymes regulating O-GlcNAcylation, and therefore hexosamine signaling, are the beta-linked-O-GlcNAc transferase (OGT) and an O-GlcNAc-selective beta-N-acetylglucosaminidase (O-GlcNAcase). Inhibition of these enzymes has a negative impact on blastocyst formation, demonstrating the importance of this signaling system to developmental potential. The ability of the OGT inhibitor benzyl-2-acetamido-2-deoxy-alpha-D-galactopyranoside (BADGP) to reverse the glucotoxic effects of hyperglycemia on these parameters was also sought. Excess HSP flux arising from a hyperglycemic environment or glucosamine supplementation reduced cell proliferation and blastocyst formation, confirming the criticality of this signaling pathway during early embryogenesis. Inhibition of OGT using BADGP blocked the negative impact of hyperglycemia on blastocyst formation, cell number, and apoptosis. Our results suggest that dysregulation of HSP and O-GlcNAcylation is the mechanism by which the embryotoxic effects of hyperglycemia are manifested during preimplantation development.

Synovial fluid exoglycosidases are predictors of rheumatoid arthritis and are effective in cartilage glycosaminoglycan depletion.

OBJECTIVE: To analyze enzymes involved in joint damage by simultaneous investigation of glycosidases and matrix metalloproteinases (MMPs) in patients with various joint diseases. METHODS: Activities of glycosidases (beta-D-glucuronidase, beta-D-N-acetyl-glucosaminidase, beta-D-N-acetyl-galactosaminidase, beta-D-galactosidase, and alpha-D-mannosidase) were tested at an acidic pH as well as at the original pH of the synovial fluid (SF) samples in parallel with activities of MMP-1 and MMP-9. RESULTS: Patients with rheumatoid arthritis (RA) were characterized by significantly elevated activities of beta-D-glucuronidase and beta-D-N-acetyl-glucosaminidase in SF compared with patients with osteoarthritis, seronegative spondylarthritis, or acute sports injury. To select the best predictor for distinguishing among patient groups, a stepwise logistic regression analysis was performed; the strongest association was found to be between RA and beta-D-glucuronidase/beta-D-N-acetyl-glucosaminidase activities (measured at the pH of the SF). Further, a significant correlation was observed between the activity of SF beta-D-N-acetyl-glucosaminidase and the level of rheumatoid factor. In vitro digestion of human hyaline cartilage samples revealed that the dominant glycosidases, alone or in combination with MMPs, proved to be effective in depleting glycosaminoglycans (GAGs) from cartilage. CONCLUSION: These results suggest that exoglycosidases, which are present in the SF of RA patients, may contribute to the depletion of GAGs from cartilage and thereby facilitate the invasion of synovial cells and their attachment to cartilage in RA.

Isoforms of an N-acetyl-beta-D-glucosaminidase from the Antarctic krill, Euphausia superba: purification and antibody production.

Two forms of the chitinolytic enzyme N-acetyl-beta-D-glucosaminidase (NAGase, EC 3.2.1.52) have been isolated from the Antarctic krill, Euphausia superba, in order to study their potential role in temperature adaptation processes. A chromatographic protocol was developed that allowed complete separation of the two enzyme forms, named NAGase B and NAGase C. The latter was purified to homogeneity with 600-fold enrichment and a yield of 17%. The molecular mass was 150 kDa. NAGase B showed characteristics of a glycoprotein due to affinity towards concanavalin A sepharose, while NAGase C did not. Highly specific polyclonal antibodies to NAGase C [anti-(E. superba-NAGase C)-IgG] showed only negligible cross-reactivity with NAGase B isoforms. A comparison with the Northern krill, Meganyctiphanes norvegica, revealed a corresponding chromatographic pattern with two main activity peaks, for differentiation named NAGase II and NAGase III. Application of the antibody on M. norvegica revealed a high specificity toward NAGase III and a low cross-reactivity with NAGase II. First indication is given that the two forms are no isoenzymes in a strict sense but instead may have different functions in the metabolism of krill.

New sandwich ELISA for human urinary N-acetyl-beta-D-glucosaminidase isoenzyme B as a useful clinical test.

We have developed a new ELISA for quantifying N-acetyl-beta-D-glucosaminidase (NAG) isoenzyme B in human urine after raising monoclonal antibodies against the isoenzyme from human placenta. Though the obtained antibodies reacted not only to isoenzyme B but also to A, we could detect isoenzyme B selectively by a two-step sandwich ELISA with a pair of selected antibodies at low pH in the first reaction. The detected limit was 0.5 microgram/L for a sample volume of 25 microL. Within-run CVs ranged from 2.5% to 5.4% and between-run CVs ranged from 6.2% to 9.1%. Recoveries of NAG isoenzyme B added to each of three urine samples ranged from 91% to 114%. The dilution curves of urine samples showed good linearity. The cross-reactivity of NAG isoenzyme A was practically negligible (2-3%). The mean value for NAG isoenzyme B in spot urines from healthy adults was 2.9 micrograms/g creatinine. This ELISA method is rapid and precise enough for routine determination of NAG isoenzyme B in human urine.

An autolysin ring associated with cell separation of Staphylococcus aureus.

atl is a newly discovered autolysin gene in Staphylococcus aureus. The gene product, ATL, is a unique, bifunctional protein that has an amidase domain and a glucosaminidase domain. It undergoes proteolytic processing to generate two extracellular peptidoglycan hydrolases, a 59-kDa endo-beta-N-acetylglucosaminidase and a 62-kDa N-acetylmuramyl-L-alanine amidase. It has been suggested that these enzymes are involved in the separation of daughter cells after cell division. We recently demonstrated that atl gene products are cell associated (unpublished data). The cell surface localization of the atl gene products was investigated by immunoelectron microscopy using anti-62-kDa N-acetylmuramyl-L-alanine amidase or anti-51-kDa endo-beta-N-acetylglucosaminidase immunoglobulin G. Protein A-gold particles reacting with the antigen-antibody complex were found to form a ring structure on the cell surface at the septal region for the next cell division site. Electron microscopic examination of an ultrathin section of the preembedded sample revealed preferential distribution of the gold particles at the presumptive sites for cell separation where the new septa had not been completed. The distribution of the gold particles on the surface of protoplast cells and the association of the gold particles with fibrous materials extending from the cells suggested that some atl gene products were associated with a cellular component extending from the cell membrane, such as lipoteichoic acid. The formation of a ring structure of atl gene products may be required for efficient partitioning of daughter cells after cell division.

The use of monoclonal antibodies for studying the biological properties of Staphylococcus aureus endo-beta-N-acetylglucosaminidase.

Staphylococcus aureus endo-beta-N-acetylglucosaminidase (SaG) has been suggested to function as a virulence determinant which interferes with the host cellular immune response. To further characterize the biological properties of SaG, monoclonal antibodies (mAbs) were raised against purified SaG. Four IgG1 subclass mAbs were obtained, none of which reacted with the reduced, sodium dodecyl sulphate pretreated or boiled enzyme. The ability of the mAbs to react with the enzymes present in supernatants obtained from 197 S. aureus strains indicated that they recognized epitopes which are highly conserved; bacteriolytic enzymes produced by staphylococci other than S. aureus did not show any cross-reactivity. After pretreatment of SaG with mAbs (mAb-SaG molar ratios varying from 1 to 20), it was shown that all selected mAbs caused, at a mAb:SaG molar ratio of 10, a 90% inhibition of SaG bacteriolytic activity and a statistically significant reduction of its ability to interfere with phagocytosis by human polymorphonuclear leukocytes. All selected mAbs reacted with several commercially available exo-beta-N-acetylglucosaminidases; mAb C1/10-11 also reacted with chicken and turkey egg muramidases and, at a mAb:SaG molar ratio of 10, inhibited their bacteriolytic activity by 97%. This suggests that one or more epitopes present in the above exo-glucosaminidases and muramidases share some degree of homology with others present in SaG.

Rat liver chitobiase: purification, properties, and role in the lysosomal degradation of Asn-linked glycoproteins.

Chitobiase, the lysosomal glycosidase responsible for splitting the GlcNAc beta-D-(1-4)GlcNAc moiety in Asn-linked glycoproteins, was purified over 600-fold from frozen rat livers utilizing an assay with di-N-acetylchitobiose as the substrate. The final preparation showed a major polypeptide of Mr 43,000 (sodium dodecylsulfate-polyacrylamide gel electrophoresis) that was determined to be the chitobiase by an immunological method. The purified chitobiase also hydrolyzed tri- and tetrasaccharides of chitin, which like di-N-acetylchitobiose were not substrates if first reduced by NaBH4. The initial products formed during hydrolysis of the tetrasaccharide were trisaccharide and GlcNAc. These results imply that chitobiase is a "reducing-end exohexosaminidase" which cleaves single GlcNAc units only from the reducing end of oligosaccharides. Fucose, typically found linked to the reducing-end GlcNAc in complex oligosaccharide chains, was found to block this reaction. Additional substrates that were hydrolyzed included GlcNAc beta-D-(1-4)MurNAc, the repeating structure from bacterial cell wall peptidoglycan, and the Man beta-D-(1-4)GlcNAc reducing-end component of glycoproteins. Km and Vm for hydrolysis of these substrates were of similar magnitude as for di-N-acetylchitobiose (6.3 mM and 15 mumol/min/mg protein, respectively). Liver tissues from nin mammalian species were surveyed for the presence of chitobiase activity. The activity was found in rat, mouse, rabbit, and guinea pig liver (Stirling [(1974) FEBS Lett. 39, 171-175] previously observed the enzyme in human liver), but not in dog, sheep, pig, cat, and cow liver. The presence or absence of chitobiase so far observed was found to exactly correlate with the type of oligosaccharide fragments found to accumulate in animals containing genetic or inhibitor-induced lysosomal storage pathologies. The presence of the chitobiase corresponds to the occurrence of one GlcNAc unit at the reducing end of stored oligosaccharides, while the absence of this glycosidase yields fragments with an intact GlcNAc beta-D-(1-4)GlcNAc moiety. These results verify our previous proposal that lysosomal disassembly of glycoproteins to free amino acids and sugars is an ordered, bidirectional pathway in which chitobiase (when present) catalyzes the last step during digestion of the protein-oligosaccharide linkage region.

A locus regulating N-acetylglucosaminidase synthesis during development in dictyostelium.

The cellular specific activity of N-acetylglucosaminidase increases during development in Dictyostelium discoideum. A monoclonal antibody which specifically recognizes Mr 68,000 and 67,000 forms of N-acetylglucosaminidase was used to show that changes in the relative rate of enzyme synthesis during development parallel the pattern of enzyme accumulation. Developmental and regulatory mutants were isolated to study the relationship between development and enzyme accumulation. No evidence was obtained for any dependence of enzyme accumulation on those genes that are required for aggregation. However, a separate regulatory locus was identified which is involved in enzyme accumulation. Mutations in this gene, nagC, prevent enzyme accumulation during development by preventing an increase in the relative synthetic rate of N-acetylglucosaminidase. The accumulation of other enzymes is unaffected and the mutation causes no developmental defects other than those caused by the loss of N-acetylglucosaminidase activity. The nagC mutation, which is recessive, maps to linkage group VI and is therefore unlinked to the structural gene for N-acetylglucosaminidase.

The use of tolerization in the production of monoclonal antibodies against minor antigenic determinants.

An initial attempt to prepare monoclonal antibodies specific for the Dictyostelium discoideum lysosomal enzyme beta-glucosidase was unsuccessful. All of the antibodies resulting from this fusion recognized an extremely immunogenic epitope that is present on all of the lysosomal enzymes of Dictyostelium. In two succeeding fusions, changes in the immunization schedule intended to increase the immune response to enzyme-specific epitopes were not entirely successful. Although nine hybridomas producing antibodies specific for beta-glucosidase resulted from these two fusions, most (70%) of the cell lines isolated secrete antibodies that recognize the shared, immunodominant epitope. Moreover, the nine beta-glucosidase-specific antibodies proved to be of limited utility since none recognize the native enzyme. Therefore, we attempted to tolerize a BALB/c mouse to the common epitope by injecting the lysosomal enzyme, N-acetylglucosaminidase, within 40 h after birth. As an adult, this animal was immunized with beta-glucosidase. Fusion of the spleen cells from this mouse with myeloma cells resulted in the isolation of nine hybridoma lines that produce antibodies specific for beta-glucosidase. No antibodies reactive with the common epitope were detected. These results suggest that tolerization may provide a means whereby an undesired class of antibody-producing cell lines can be selectively eliminated from the products of a fusion.

Biosynthesis of alpha-N-acetylglucosaminidase in cultured human kidney carcinoma cells.

The biosynthesis of alpha-N-acetylglucosaminidase was studied in cultured human kidney carcinoma cells by labeling cells with 35S-methionine, isolation of the enzyme by immunoprecipitation and analysis on gel electrophoresis of the denatured polypeptide(s) and analysis of the native enzyme on linear sucrose gradient centrifugation. The enzyme is synthesized as precursor forms of apparent molecular weight 82,000-86,000. Processing of these precursors yields a polypeptide of apparent molecular weight of 80,000. The precursor-product relationship was indicated by pulse-chase as well as endocytosis experiments. Sucrose gradient centrifugation of the native enzyme shows that, extracellularly, the molecule is present with a molecular weight of 80,000; intracellularly, 80-90% of the enzyme is present with an apparent molecular weight of 240,000. We suggest that this is a polymeric form and that polymerization of alpha-N-acetylglucosaminidase is a late event of the maturation process.

An alternative hypothesis of cellular transport of lysosomal enzymes in fibroblasts. Effect of inhibitors of lysosomal enzyme endocytosis on intra- and extra-cellular lysosomal enzyme activities.

Recapture of lysosomal enzymes secreted by fibroblasts was inhibited by growing the cells in the presence of either free or immobilized antibodies against lysosomal enzymes or in the presence of phosphorylated carbohydrates known to interact with the cell-surface receptors for lysosomal enzymes. The following results were obtained. 1. Conditions that prevent recapture of released lysosomal enzymes increase the rate of extracellular accumulation of these enzymes up to twice that of controls. 2. Growing cells for 12 days in the presence of 0.5mm-mannose 6-phosphate, which decreases beta-N-acetylglucosaminidase endocytosis to less than 10% of that of controls, has no effect on the intracellular activity of this and four other lysosomal enzymes. 3. Growing cells for 4 days in the presence of 50mm-mannose 6-phosphate, which is a 1000-fold higher concentration than that required for 50% inhibition of lysosomal enzyme endocytosis, leads to a 4-fold increase in extracellular beta-N-acetylglucosaminidase accumulation and a decrease in intracellular enzyme. These results give evidence that, in fibroblasts, transfer of lysosomal enzymes into lysosomes does not require secretion before a receptor-mediated recapture [Hickman & Neufeld (1972) Biochem. Biophys. Res. Commun.49, 992-999]. We propose that (a) lysosomal enzymes are present in a receptor-bound form in those vesicles that fuse with the cell membrane, (b) the major part of the lysosomal enzyme cycles via the cell surface in a receptor-bound form and (c) only a minor part of the lysosomal enzyme is released into the extracellular space during its life cycle.

Characterization of proteins structurally related to human N-acetyl-beta-D-glucosaminidase.

Those proteins of human liver that cross-reacted with antibodies raised to apparently homogenous hexosamindases A and B were detected by immunodiffusion. Cross-reacting proteins with high molecular weights (greater than 2000000) and intermediate molecular weights (70000--200000) were present both in the unadsorbed fraction and in the 0.05--0.2M-NaCl eluate obtained by DEAE-cellulose chromatography at pH7.0. The unadsorbed fraction also contained a cross-reacting protein of low molecular weight (10000--70000). The possible structural and functional relationships between hexosaminidase and the cross-reacting proteins are discussed. An apparently cross-reacting protein present in the 0.05M-NaCl eluate from the DEAE-cellulose column was serologically unrelated to hexosaminidase, but it gave a reaction of immunological identify with one of the apparently cross-reacting proteins having the charge and size characteristics of hexosaminidase A. It is suggested that immunochemical methods may provide criteria for the homogeneity of enzyme preparations superior to those of conventional methods.

Characterization of beta-D-N-acetylhexosaminidase isoenzymes in man-Chinese hamster somatic cell hybrids.

A series of man-Chinese hamster hybrids were investigated with the use of an anti-Chinese hamster hexosaminidase serum, a specific anti-human hex A serum and an anti-human hex B serum. The expression of human hex A was found to be dependent on the presence of hex B. A heteropolymeric molecule is formed independently of hex B, which consists of Chinese hamster and specific hex A moieties. It has an electrophoretic mobility nearly identical to hex A. A relationship between the absence and presence of the heteropolymeric molecule, mannosephosphate isomerase (MPI), and pyruvate kinase (PK-3), assigned to chromosome 15, was established. With respect to the two locus subunit model, the gene coding for the alpha subunit, specific for hex A, has been localized on chromosome 15.

Interrelationship of hexosaminidases A and B: conformation of the common and the unique subunit theory.

Human kidney hexosaminidase A (beta-N-acetylglucosaminidase; 2-acetamido-2-deoxy-beta-D-glucoside acetamidodeoxyglucohydrolase; EC 3.2.1.30) is a heteropolymer of two immunologically distinct subunits designated as alpha and beta. Hexosaminidase B, however, is a homopolymer comprised entirely of beta subunits. When human kidney hexosaminidase A was dissociated into its subunits by p-hydroxymercuribenzoate, three distinct proteins having isoelectric points of pH 7.2.5.4, and 4.3 were isolated. The fraction having an isoelectric point of pH 7.2, designated as beta fraction, was electrophoretically and immunologically identical to hexosaminidase B and was enzymatically active. The proteins having isoelectric points of pH 5.4 and 4.3, designated as hexosaminidase Ai and alpha fractions, respectively, were enzymatically inactive and crossreacted with antiserum against hexosaminidase A and not with antiserum against hexosaminidase B. Upon incubation of p-hydroxymercuribenzoate-treated hexosaminidase A with dithiothreitol,, hexosaminidase A activity, as well as antigenicity, was regenerated, indicating that alpha and beta subunits hybridize to form hexosaminidase A. Antibodies raised in rabbits against beta fractions reacted with both hexosaminidase A and B, whereas the antibodies against alpha and hexosaminidase Ai fractions reacted only against hexosaminidase A. This would indicate that both fractions are composed only of subunits unique to hexosaminidase A. The molecular weights of alpha,beta, and hexosaminidase Ai fractions were estimated to be 47,000, 120,000, and 180,000 respectively, by Sephadex gel filtration. Upon urea-sodium dodecyl sulfate polyacrylamide electrophoresis, each of the three fractions dissociated into a single polypeptide having a molecular weight of approximately 18,000. It is concluded that p-hydroxymercuribenzoate dissociates hexosaminidase A, (alphabeta)3, into its subunits, and the beta subunits can reassociate to form relatively stable hexosaminidase B, (betabeta)3, while the alpha subunits reassociate in both the dimeric state, alpha2, and a polymeric state, alpha8.