Properties of a Particle-bound Enzyme from Rat Intestine that Cleaves Sialic Acid from Tay-Sachs Ganglioside

Adeno-associated virus-mediated expression of β-hexosaminidase prevents neuronal loss in the Sandhoff mouse brain

Sandhoff disease, a GM2 gangliosidosis caused by a deficiency in β-hexosaminidase, is characterized by progressive neurodegeneration. Although loss of neurons in association with lysosomal storage of glycosphingolipids occurs in patients with this disease, the molecular pathways that lead to the accompanying neurological defects are unclear. Using an authentic murine model of GM2 gangliosidosis, we examined the pattern of neuronal loss in the central nervous system and investigated the effects of gene transfer using recombinant adeno-associated viral vectors expressing β-hexosaminidase subunits (rAAV2/1-Hex). In 4-month-old Sandhoff mice with neurological deficits, cells staining positively for the apoptotic signature in the TUNEL reaction were found in the ventroposterior medial and ventroposterior lateral (VPM/VPL) nuclei of the thalamus. There was progressive loss of neuronal density in this region with age. Comparable loss of neuronal density was identified in the lateral vestibular nucleus of the brainstem and a small but statistically significant loss was present in the ventral spinal cord. Loss of neurons was not detected in other regions that were analysed. Administration of rAAV2/1-Hex into the brain of Sandhoff mice prevented the decline in neuronal density in the VPM/VPL. Preservation of neurons in the VPM/VPL was variable at the humane endpoint in treated animals, but correlated directly with increased lifespan. Loss of neurons was localized to only a few regions in the Sandhoff brain and was prevented by rAAV-mediated transfer of β-hexosaminidase gene function at considerable distances from the site of vector administration.

Tay-Sachs disease in Jacob sheep

Autopsy studies of four Jacob sheep dying within their first 6-8 months of a progressive neurodegenerative disorder suggested the presence of a neuronal storage disease. Lysosomal enzyme studies of brain and liver from an affected animal revealed diminished activity of hexosaminidase A (Hex A) measured with an artificial substrate specific for this component of β-hexosaminidase. Absence of Hex A activity was confirmed by cellulose acetate electrophoresis. Brain lipid analyses demonstrated the presence of increased concentrations of G(M2)-ganglioside and asialo-G(M2)-ganglioside. The hexa cDNA of Jacob sheep was cloned and sequenced revealing an identical number of nucleotides and exons as in human HexA and 86% homology in nucleotide sequence. A missense mutation was found in the hexa cDNA of the affected sheep caused by a single nucleotide change at the end of exon 11 resulting in skipping of exon 11. Transfection of normal sheep hexa cDNA into COS1 cells and human Hex A-deficient cells led to expression of Hex S but no increase in Hex A indicating absence of cross-species dimerization of sheep Hex α-subunit with human Hex β-subunits. Using restriction site analysis, the heterozygote frequency of this mutation in Jacob sheep was determined in three geographically separate flocks to average 14%. This large naturally occurring animal model of Tay-Sachs disease is the first to offer promise as a means for trials of gene therapy applicable to human infants.

Spontaneous appearance of Tay-Sachs disease in an animal model

Tay-Sachs disease (TSD) is a progressive neurodegenerative disorder due to an autosomal recessively inherited deficiency of beta-hexosaminidase A (Hex A). Deficiency of Hex A in TSD is caused by a defect of the alpha-subunit resulting from mutations of the HEXA gene. To date, there is no effective treatment for TSD. Animal models of genetic diseases, similar to those known to exist in humans, are valuable and essential research tools for the study of potentially effective therapies. However, there is no ideal animal model of TSD available for use in therapeutic trials. In the present study, we report an animal model (American flamingo; Phoenicopterus ruber) of TSD with Hex A deficiency occurring spontaneously in nature, with accumulation of G(M2)-ganglioside, deficiency of Hex A enzymatic activity, and a homozygous P469L mutation in exon 12 of the hexa gene. In addition, we have isolated the full-length cDNA sequence of the flamingo, which consists of 1581 nucleotides encoding a protein of 527 amino acids. Its coding sequence indicates approximately 71% identity at the nucleotide level and about 72.5% identity at the amino acid level with the encoding region of the human HEXA gene. This animal model, with many of the same features as TSD in humans, could represent a valuable resource for investigating therapy of TSD.

Ganglioside GM2 is substrate for a sialidase in MDCK cells

GM1 ganglioside carrying a fluorescent fatty acid in substitution of the natural one, has been administered to cultured Madin-Darby canine kidney (MDCK) cells for different pulse times (0.5-24 h), and its metabolic fate was followed. The fluorescent GM2, asialo-GM2, asialo-GM1 and ceramide were the only detectable metabolites. The complete absence of fluorescent GM3 is consistent with the presence in these cells of a sialidase working on GM1 and GM2 gangliosides. After treatment of the cells with chloroquine the fluorescent GM1 remained essentially undegraded, indicating a catabolic processing at lysosomal level.

Monomer-micelle transition of the ganglioside GM1 and the hydrolysis by Clostridium perfringens neuraminidase

The action of Clostridium perfringens neuraminidase on the ganglioside Gm1 tritiated in the ceramide moiety was studied. The rates of hydrolysis of the Gm1 ganglioside were determined from radioactivity in the neutral glycolipid product, which was separated from the substrate on DEAE-Sephadex columns. In order to study the physical state of the substrate in the conditions used in the neuraminidase treatment, the critical micelle concentrations of the Gm1 ganglioside were determined using formation of the triiodide anion in aqueous iodine solution as an indicator. The critical micelle concentrations were also obtained by determining the non-sedimenting radioactivity at different concentrations of the labeled ganglioside per total volume used in ultracentrifugation experiments. In addition, the concentrations of the monomeric ganglioside were concluded from the results of the ultra-centrifugation studies. The increase in the reaction rate of the Gm1 hydrolysis as the function of the substrate concentration was leveled off at 25-28 microM ganglioside. The abrupt change at this concentration is interpreted as reflecting the monomer-micelle transition of the ganglioside in the conditions used (50mM sodium acetate buffer, pH 4.6). The critical micelle concentration was 29 microM on the basis of the triiodide test, and ultracentrifugation revealed the critical micelle concentration 28 microM. The reaction velocity of the hydrolysis was decreased immediately above the critical micelle concentration, and became constant at higher concentrations of the ganglioside. A close correlation to these changes in the reaction rate is suggested to exist in the concentrations of the monomeric Gm1 ganglioside. Saturation of the buffer used in the neuraminidase assays with butanol effected a striking change in the plot of reaction rate versus ganglioside concentration. The reaction rate increased up to 100-110 microM Gm1 ganglioside. The shift of the inflexion point in the rate plot from 25-28 microM to 100-110 microM ganglioside concentration is suggested to be due to a respective change in the critical micelle concentration effected by butanol. N-Acetylneuraminyllactosyl ceramide, lactosyl ceramide and asialo-Gm1 ganglioside had an inhibitory effect on the reaction. In contrast, N-acetylneuraminyllactose, lactose and some other free saccharides were not inhibitory. The results demonstrate that factors other than the saccharide structure must be taken into account when substrate specificity of a glycosidase is studied using competition experiments. It is suggested that the inhibition effected by the glycolipids is due to an increase in the micellar state of the Gm1 ganglioside.

Studies on brain cytosol neuraminadase. I. Isolation and partial characterization of two forms of the enzyme from pig brain

1. Two forms of cytosol neuraminidase (EC 3.2.1.18) (neuraminidase A and neuraminidase B) were isolated and purified from pig brain homogenate, by proceeding through the following steps: centrifugation of brain homogenate at 105 000 X g (1h); ammonium sulphate fractionation (35-55% saturated fraction); column chromatography on Biogel A 5 m; column chromatography on hydroxy apatite/cellulose gel; affinity chromatography on Affinose-tyrosyl-p-nitrophenyloxamic acid. The separation of the two forms of neuraminidase was provided by chromatography on hydroxylapatite/cellulose gel. Neuraminidase A was purified about 500-fold; neuraminidase B about 400-fold. 2. The pH optima and the maximum activities in various buffers were different for neuraminidase A and B (for instance the pH optimum was in sodium acetate/acetic acid buffer, 4.7 for neuraminidase A and 4.9 for neuraminidase B). Ions affected in a different way the two enzymes: K+ activated neuraminidase A but not neuraminidase B; Na+ and Li+ inhibited neuraminidase A at a higher degree than neuraminidase B. Neuraminidase B seemed to be moderately activated by some bivalent cations (Ca2+; Mg2+; Zn2+); neuraminidase A did not. The Km values for sialyllactose were different: 2.2-10(-3) M for neuramindase A; 0.46-10(-3) M for neuraminidase B.