Purification and properties of the hexosaminidase A-activating protein from human liver

Human liver extracts contain an activating protein which is required for hexosaminidase A-catalysed hydrolysis of the N-acetylgalactosaminyl linkage of G_M2 ganglioside [N-acetylgalactosaminyl-(N-acetylneuraminyl) galactosylglucosylceramide]. A partially purified preparation of human liver hexosaminidase A that is substantially free of G_M2 ganglioside hydrolase activity is used to assay the activating protein. The proceudres of heat and alcohol denaturation, ion-exchange chromatography and gel filtration were used to purify the activating protein over 100-fold from crude human liver extracts. When the purified activating protein is analysed by polyacrylamide-gel disc electrophoresis, two closely migrating protein bands are seen. When purified activating protein is used to reconstitute the G_M2 ganglioside hydrolase activity, the rate of reaction is proportional to the amount of hexosaminidase A used. The activation is specific for G_M2 ganglioside and and hexosaminidase A. The activating protein did not stimulate hydrolysis of asialo-G_M2 ganglioside by either hexosaminidase A or B. Hexosaminidase B did not catalyse hydrolysis of G_M2 ganglioside with or without the activator. Kinetic experiments suggest the presence of an enzyme–activator complex. The dissociation constant of this complex is decreased when higher concentrations of substrate are used, suggesting the formation of a ternary complex between enzyme, activator and substrate. Determination of the molecular weight of the activating protein by gel-filtration and sedimentation-velocity methods gave values of 36000 and 39000 respectively.     

Assignment of the human gene for hexosaminidase B to chromosome 5

Mouse-human somatic cell hybrids between different mouse and human cells were studied for the expression of human hexosaminidases A and B activities. The expression of human hexosaminidase B in the hybrids was found to segregate concordantly with the presence of the human chromosome 5. Mouse-human hybrid clones containing either the human chromosomes 5 and 7 only or the human chromosome 7 only were also included in this study. Expression of human hexosaminidase B activity was detected only in those clones containing human chromosome 5. These results indicate that the gene(s) for human hexosaminidase B is located on chromosome 5. No hexosaminidase A activity was detected in clones which contained either human chromosomes 5 and 7 or chromosome 7.     

Biochemical and immunochemical characterization of hexosaminidase P.

Hexasaminidase P, the main isozyme of hexosaminidase in pregnancy serum, was isolated and purified 600--700-fold by a two-step purification procedure--affinity chromatography on Sepharose-bound epsilon-aminocaproyl-N-acetylglucosylamine, followed by ion-exchange chromatography on DEAE-cellulose. The purified enzyme was subjected to biochemical and immunochemical analysis. Its catalytic property, namely, kinetic behavior, is similar to that of the major isozymes of hexosaminidase, A and B. However, it differs from these isozymes in its electrophoretic mobility and in its apparent molecular weight which is around 150 000 compared with 100 000 of the A and B isozymes. Immunochemical analysis indicates that the P isozymes is antigenically cross-reactive with both A and B isozymes, but it does not contain the A-specific antigenic determinants, and exhibits identical antigenic specificity to hexasaminidase B. Two possible structures are suggested that are compatible with the experimental data: (a) a hexosaminidase B like structure with higher extent of glycosylation; (b) a hexameter of beta chain, possibly arranged as three beta2 subunits.     

Purification and characterization of beta-N-acetylhexosaminidase I from human placenta

beta-N-Acetylhexosaminidase (hexosaminidase) I, which has an intermediate charge character between those of hexosaminidases A(alpha beta 2) and B[beta beta)2), was purified 1,500-fold from human placenta by procedures including chromatographies on concanavalin A (Con A)-Sepharose and an immunoadsorbent column. The isolated hexosaminidase I was heat-stable, and antigenically cross-reactive to anti-beta chain-IgG but not to anti-alpha chain-IgG. The results of substrate specificity experiments using 3H-labeled natural substrates indicated that the hexosaminidase I hydrolyzed Gb4Cer to Gb3Cer but not GM2 to GM3. The tryptic peptide map of the hexosaminidase I was similar to that of hexosaminidase B, though some differences were observed. The hexosaminidase I after treatment with neuraminidase or endo-beta-N-acetylglucosaminidase H was partly converted to less acidic forms. Treatment of the hexosaminidase I with acid phosphatase did not change the charge character. Therefore hexosaminidase I is an acidic variant form of hexosaminidase B, possibly resulting from sialylation and the presence of phosphodiester bonds at the carbohydrate moiety.     

The subunits of human hexosaminidase A.

Previous studies of the subunit structure of hexosaminidase gave ambiguous results, but suggested that the enzyme was composed of six equally sized subunits. Dissociation of hexosaminidase A with p-chloromercuribenzoate produces an alkylated fragment with mol.wt. approx. 50000, which is converted into hexosaminidase S by treatment with dithiothreitol. Treatment of native hexosaminidase A with sodium dodecylsulphate results in the formation of a large and a small fragment. However, although the native enzyme has a sedimentation coefficient of 5.8S, dissociation by S-carboxymethylation and maleic anhydride treatment results in subunits exhibiting a single schlieren boundary on analytical ultracentrifugation with a sedimentation coefficient of 2.18S. These results indicate that the enzyme is composed of four subunits, each with molwt. approx. 25000-27000. The mol.wt. of the native enzymes is calculated to be approx. 110000. Our data are consistent with the subunit structures of hexosaminidases A, B and S as being alpha2beta2, beta4 and alpha4 respectively.     

The biochemical genetics of the hexosaminidase system in man.

Tay-Sachs disease and related GM2 ganglioside storage disorders result from the absence of one form of hexosaminidase, HEX A. The persistence of a second major hexosaminidase isozyme, HEX B, does not protect against the lethal accumulation of GM2 ganglioside in the central nervous system. Using immunologic and biochemical techniques, it has been demonstrated that the two major isozymes of hexosaminidase, HEX A and HEX B, share a common subunit, the structure of HEX A being designated (alpha beta)n and the structure of HEX B being designated as (beta2)n. The minor isozyme, HEX S, is an alpha chain homopolymer designated (alpha2)n, and HEX C seems unrelated to the HEX A, B, S system. The structures of other minor isozymes have not been totally resolved, but HEX I1, I2, and P (which may be identical to I2) appear to represent forms of HEX B.     

Purification and properties of human kidney-cortex hexosaminidases A and B.

Hexosaminidases (EC 3.2.1.30) A and B from human kidney cortex were purified to homogeneity by using concanavalin A affinity chromatography, ion-exchange chromatography and gel filtration. The yield of homogeneous isoenzymes improved approx. 20-fold, giving preparations of hexosaminidases A and B with specific activities of about 200 and 325 units/mg of protein respectively. The kinetic and structural properties of kidney hexosaminidase isoenzymes were studied and compared with the hexosaminidase isoenzymes from human placenta. The amino acid composition of hexosaminidase A was significantly different from that of hexosaminidase B. In the event of success in developing enzyme-replacement therapy for Tay-Sachs and Sandhoff's diseases, this modified procedure can furnish larger amounts of homogeneous isoenzymes.     

Irreversible inhibition of hexosaminidase C by medium-chain monocarboxylic acids and Triton X-100

The neutral beta-N-acetylhexosaminidase (hexosaminidase C) from human brain was partially purified (separated from lysosomal beta-N-acetylhexosaminidases by chromatography on a Con A-Sepharose column). Hexosaminidase C was inhibited by medium-chain fatty acids (monocarboxylic acids with chain-length between C6 and C9), whereas shorter-chain monocarboxylic acids showed no inhibitory effect. Studies on the inhibition mechanism showed an irreversible and pH-dependent inhibition which progresses with time and which is not reversed by the removal of fatty acids (by Bio-Beads SM-2). Similar inhibitory effects were also obtained using Triton X-100 (but not with homologous alkylamines). These results suggest that the hexosaminidase C inactivation is related to the hydrophobic properties of the inhibitor which acts as a denaturing agent mainly at acidic pH. The possibility has been discussed that this inactivation effect of monocarboxylic acid on hexosaminidase C could constitute a molecular model of the toxicity of medium-chain-length fatty acids.     

Studies on the substrate specificity of hexosaminidase A and B from liver

β-N-Acetylhexosaminidase A and B were partially purified from normal human liver using DEAE-cellulose column chromatography. Hexosaminidase B was also purified from the livers of patients who had died of Tay-Sachs disease. The hexosaminidase fractions were tested for their ability to hydrolyze the amino sugar moiety of synthetic substrates and of three amino sugar-containing glycolipids, GA₂, globoside, and GM₂.     

Role of beta Arg211 in the active site of human beta-hexosaminidase B

Tay-Sachs or Sandhoff disease results from a deficiency of either the alpha- or the beta-subunits of beta-hexosaminidase A, respectively. These evolutionarily related subunits have been grouped with the "Family 20" glycosidases. Molecular modeling of human hexosaminidase has been carried out on the basis of the three-dimensional structure of a bacterial member of Family 20, Serratia marcescens chitobiase. The primary sequence identity between the two enzymes is only 26% and restricted to their active site regions; therefore, the validity of this model must be determined experimentally. Because human hexosaminidase cannot be functionally expressed in bacteria, characterization of mutagenized hexosaminidase must be carried out using eukaryotic cell expression systems that all produce endogenous hexosaminidase activity. Even small amounts of endogenous enzyme can interfere with accurate K(m) or V(max) determinations. We report the expression, purification, and characterization of a C-terminal His(6)-tag precursor form of hexosaminidase B that is 99.99% free of endogenous enzyme from the host cells. Control experiments are reported confirming that the kinetic parameters of the His(6)-tag precursor are the same as the untagged precursor, which in turn are identical to the mature isoenzyme. Using highly purified wild-type and Arg(211)Lys-substituted hexosaminidase B, we reexamine the role of Arg(211) in the active site. As we previously reported, this very conservative substitution nevertheless reduces k(cat) by 500-fold. However, the removal of all endogenous activity has now allowed us to detect a 10-fold increase in K(m) that was not apparent in our previous study. That this increase in K(m) reflects a decrease in the strength of substrate binding was confirmed by the inability of the mutant isozyme to efficiently bind an immobilized substrate analogue, i.e., a hexosaminidase affinity column. Thus, Arg(211) is involved in substrate binding, as predicted by the chitobiase model, as well as catalysis.     

The expression of hex A and hex B isozymes of hexosaminidase in parental and experimental human fibroblast cells and their components

The expression of the two major isozyme forms of hexosaminidase (EC 3.2.1.30), hesoxaminidase A and hexosaminidase B, has been examined. The parental cells and/or cellular components of parental cells are individually fused using inactivated Sendai virus with the aid of a micromanipulator. The progeny cells produced from such hybrids are subjected to a microenzymatic assay which allows measurements at the single cell level. The lysosomal-deficient cells used in this study are Tay-Sachs and Sandhoff fibroblasts, and the normal cells used are WI-38 (fetal lung fibroblasts), amniotic fluid cells (GM 473), and JASD₃ (normal human foreskin). The results show that the ratio of cell components which are fused to form the experimental cell affects the percentage of hexosaminidase A expressed in the progeny cells. Furthermore, our results imply the presence of a “factor” in the Sandhoff cell's cytoplasm which, together with the Tay-Sachs nucleus, is necessary for hexosaminidase A expression in the experimental cell's progeny.     

Separation and properties of human brain hexosaminidase C

Hexosaminidase C was separated from human brain supernatant by immunoadsorption of the A and B forms on to a column of immobilized antibody followed by preparative starch-block electrophoresis. There were some differences in the properties of hexosaminidase C preparations after each of these stages, shown by comparison of their heat-inactivation characteristics and filtration through Bio-Gel P-200. The C form prepared by both separation steps had properties which differed markedly from those of the A and B isoenzymes; its molecular weight was much larger, greater than 200000, it had optimum activity between pH6 and 7 and could not be successfully eluted from DEAE-cellulose, even with high salt concentrations, or from Sephadex G-200. These results seem to support the proposal that the C form is under a separate genetic control from the others.     

Purification and characterization of an activator protein for the degradation of glycolipids GM2 and GA2 by hexosaminidase A

The activator protein for the degradation of glycolipids GM2 and GA2 by hexosaminidase A was purified some 2 500-fold from normal human kidney. It has a molecular weight of approximately 25 000 is heat-stable up to 60 degrees C, possesses an isoelectric point of pH 4.8 and is digestible by proteases. Enzymic degradation of the lipid substrates in the presence of this activator proceeds optimally at pH 4.2. The mode of action of the activator was also studied: the protein most probably complexes lipid molecules and presents them to the enzyme which otherwise cannot attack the aggregates formed by the lipids in aqueous solution. The hydrolysis of water-soluble synthetic substrates is not affected by the activator protein. The activator is highly specific for hexosaminidase A: hydrolysis of glycolipids GA2 and GM2 by the hexosaminidase B isoenzyme is almost not enhanced by this protein. The isoenzymes' lipid substrate specificity measured in the presence of the activator is entirely different from that obtained with detergents and can satisfactorily account for the lipid storage pattern observed in patients with variant forms of infantile GM2- gangliosidosis.     

A low-molecular-weight protein cross-reacting with human liver N-acetyl-β-d-glucosaminidase

Antisera were raised to preparations of hexosaminidase isoenzymes A and B purified from human liver. Protein that cross-reacted with the liver hexosaminidase was detected by an antibody-consumption method. A cross-reacting protein with a low molecular weight (20000) was partially characterized and purified from control human liver. This protein is also present in the liver of patients with Tay-Sachs disease or with Sandhoff's disease. Hexosaminidases A and B gave an immunological reaction of partial identity with the low-molecular-weight protein. The possible identity of the low-molecular-weight cross-reacting protein as a subunit of hexosaminidase is discussed.     

Purification and characterization of beta-N-acetylhexosaminidase I2 from human liver.

beta-N-Acetylhexosaminidase I2 was purified from human liver by a combination of concanavalin A chromatography, DEAE-cellulose chromatography, gel filtration and affinity chromatography on 2-acetamido-N-(6-aminohexanoyl)-2-deoxy-beta-D-glucopyranosylamine coupled to CNBr-activated Sepharose 4B. Its specific activity was 130 mumol/min per mg of protein compared with values of 150 and 320 mumol/min/mg of protein for beta-N-acetylhexosaminidases A and B purified from the same tissue. Km values for I2, A and B were 1.0 mM, 0.8 mM and 0.74 mM respectively. On gradient gel electrophoresis under non-denaturing conditions, hexosaminidase I2 behaved similarly to A and appeared to have an Mr between 100 000 and 110 000. beta-N-Acetylhexosaminidase I2 was resolved into two major polypeptides, of Mr 56 000 and 29 000, on SDS/polyacrylamide-gel electrophoresis under denaturing conditions. Immunoblotting with anti-(hexosaminidase alpha-subunit) serum confirmed that the 56 000-Mr component was the alpha-subunit and anti-(hexosaminidase B) serum reacted with the 29 000 Mr component. beta-N-Acetylhexosaminidase I2 more closely resembles form A than B, but the features of its structure that allow it to be separated from A on the basis of net charge have not yet been found.     

Human placental N-acetyl-beta-D-hexosaminidase isozymes. Activity toward native hyaluronic acid.

The activity of purified human hexosaminidases A and B toward hyaluronic acid (HA) isolated from cultured human skin fibroblasts was investigated. The cleavage of N-acetylglucosaminyl residues to monosaccharide N-acetylglucosamines by hexosaminidase isozymes was determined in the presence and absence of purified human β-glucuronidase. The pH optima of this reaction, with and without β-glucuronidase, were 4.5 for hexosaminidase A and 4.0 for hexosaminidase B. The hydrolysis of HA by both hexosaminidase isozymes proceeds linearily for at least 18 h in the presence of β-glucuronidase. Concentrations of 0.5–5 units of either isozyme showed a linear relationship with rate of hydrolysis. Without β-glucuronidase, hexosaminidase only cleaved the terminal N-acetylglucosamine residue. However, under optimal conditions, with β-glucuronidase, the hydrolytic activity of hexosaminidase B was about 30% as efficient as that of hexosaminidase A. Approximately 70% of the HA could be degraded by 5 units of hexosaminidase A in the presence of 0.5 unit of β-glucuronidase, as opposed to 25% degraded by hexosaminidase B. These results probably reflect intrinsic differences in the activities of the two isozymes. Since the substrate (HA) did not inhibit the hydrolysis of a synthetic substrate (4-methylumbelliferyl-β-glucosaminide) by hexosaminidase B, the linear kinetics of HA hydrolysis implies no product inhibition. These data indicate that native HA can be hydrolyzed by the combined activities of β-glucuronidase with hexosaminidase A or hexoaminidase B.     

Purification and chemical characterization of human hexosaminidases A and B.

N-Acetyl-beta-hexosaminidases A and B were purified to homogeneity from human placenta. In the initial step of purification, the enzymes were adsorbed on concanavalin A-Sepharose 4B and eluted from the column with alpha-methyl D-mannosides. Subsequent purification steps included DEAE-cellulose column chromatography, QAE-Sephadex [diethyl-(2-hydroxypropyl)aminoethyl-Sephadex] column chromatography, Sephadex G-200 gel filtration and preparative disc polyacrylamide-gel electrophoresis, followed by another QAE-Sephadex chromatography for the hexosaminidase A preparation, and DEAE-cellulose column chromatography, calcium phosphate gel chromatography, Sephadex G-200 gel filtration, QAE-Sephadex chromatography and CM-cellulose chromatography for the hexosaminidase B preparation. The purified preparations, particularly hexosaminidase A, had significantly higher specific enzyme activities than previously reported. The preparations moved on polyacrylamide-gel electrophoresis as single protein bands, which also stained for enzyme activity. Sedimentation-equilibrium centrifugation indicated homogenous dispersion of the enzymes, and the molecular weight was estimated as about 110000 for both enzymes. Complete amino acid and carbohydrate compositions of the two isoenzymes were determined, and, in contrast with previous suggestions, no sialic acid was found in the enzymes.     

Cleavage of the (1 goes to 3)-2-acetamido-2-deoxy-beta-D-glucopyranosyl linkage present in keratan sulfate. The A and B isoenzymes of human liver hexosaminidase (EC 3.2.1.30)

The disaccharide 2-acetamido-2-deoxy-beta-D-glucopyranosyl-(1 goes to 3)-D-[1-3H]-galactitol, prepared from keratan sulfate, was rapidly hydrolyzed by the A and B isoenzymes of normal human liver hexosaminidase (EC 3.2.1.30), and by the B isoenzyme prepared from the liver of a patient who had died of Tay-Sachs disease. The disaccharide substrate was also hydrolyzed by extracts of normal, cultured-skin fibroblasts, and fibroblasts of patients with Tay-Sachs disease, whereas it was not hydrolyzed by fibroblast extracts of patients with Sandhoff disease. Thus, effective degradation of keratan sulfate, secondary to a defect of the beta subunits present in the A and B isoenzymes of hexosaminidase, may contribute to the appearance of skeletal lesions in patients affected by Sandhoff disease.     

Studies on complementation of beta hexosaminidase deficiency in human GM2 gangliosidosis.

Complementation of beta hexosaminidase A (hex A) deficiency was obtained by Sendai virus-mediated somatic cell hybridization of cultured skin fibroblasts from two unrelated patients with Tay-Sachs disease (TSD) and one patient with Sandhoff-Jatzkewitz disease (SJD). The newly formed hex A was identified by its electrophoretic mobility in three different systems, heat lability, and reactivity with an antiserum against the unique antigenic determinant, alpha of hex A. The percentage of heterokaryons obtained by virus treatment of TSD and SJD fibroblast mixtures showed good correlation with the observed percentage of hex A activity. It is concluded that, in these two forms of GM2 gangliosidosis, beta hexosaminidase deficiency results from two different mutations. All of the current models of beta hexosaminidase structure are compatible with the observed complementation. No complementation was detected in 13 Sendai virus-induced fusions of cultured skin fibroblasts from seven unrelated patients with SJD. The enzyme deficiency in these patients may be due to very similar allelic mutations, not capable of undergoing complementation; or to different structural mutations, all coding for unstable beta hexosaminidase molecules.     

Immunological properties of N-acetyl-β-d-glucosaminidase of normal human liver and of GM2-gangliosidosis liver

Antisera were raised to a partially purified preparation of human liver hexosaminidase and to highly purified preparations of hexosaminidase isoenzymes A and B. All the antisera precipitated the enzyme in an enzymically active form, which could be located on immunodiffusion and immunoelectrophoretic gels by using a histochemical substrate. The antisera to the purified isoenzymes were shown to react with hexosaminidase from human liver, kidney, brain and spleen, but did not cross-react with human liver beta-glucosidase, beta-galactosidase, alpha-mannosidase, beta-xylosidase, arylsulphatase or acid phosphatase. Hexosaminidases A and B were immunologically identical. The immunological properties of the hexosaminidases from livers of patients with three types of GM(2)-gangliosidoses were closely similar. No evidence could be found for cross-reacting material in enzyme-deficient states.