Search

The AB-variant form of GM2 gangliosidosis is an inherited lysosomal storage disease. Biochemical data have linked its cause to the lack of a functional GM2 activator protein (activator). In the present study we identify a mutation in the gene encoding the activator protein of an AB-variant patient. These data represent direct evidence that the disease in the patient described here is a result of mutations at the Activator gene locus. A T412----C transition was found in the homozygous form in cDNA and genomic DNA from the patient. This nucleotide change would result in the substitution of Cys138 by an Arg residue in the activator protein. Whereas the patient's fibroblasts produce apparently normal levels of activator mRNA, they lack a functional activator protein. Transfection of either a construct containing the normal activator cDNA, pAct1, or a cDNA construct containing the T----C transition caused COS-1 cells to transcribe high levels of activator mRNA. Lysates from cells transfected with pAct1 produced an elevated level of both pro- and mature forms of the activator protein, with an accompanying 11-fold enhancement in the ability of purified hexosaminidase A to hydrolyze GM2 ganglioside. However, lysates from cells transfected with the mutant cDNA construct contained only low levels of the pro-activator protein, which failed to enhance hexosaminidase A activity significantly above the endogenous level of mock transfected COS cells. We conclude that the T412----C transition in the GM2 Activator gene of the patient is responsible for the disease phenotype.

Two cDNAs encoding GM2 activator, pGM2A (648 bp) and GAP (1093 bp), were isolated from human placenta lambda gt11 libraries. The DNA sequence of pGM2A from 1 to 302 was almost identical with GAP, but diverged from 303-648. PCR was used to demonstrate the presence of both species of GM2 activator in placental RNA. Both cDNAs hybridized to mRNAs of approximately 2.3 kb and to identical single bands on genomic Southern blots.

GM2-gangliosidoses are neurological disorders caused by a genetic deficiency of either the beta-hexosaminidase A or the GM2 activator, a glycolipid binding protein. In a patient with an immunologically proven GM2 activator protein deficiency, A T412----C transition (counted from A of the initiation codon) was found in the coding sequence, which results in the substitution of Arg for the normal Cys107 in the mature GM2 activator protein. The remainder of the coding sequence remained entirely normal.

Full-length cDNAs coding for the human GM2-activator protein has been isolated and characterized, and its genomic structure studied in two overlapping clones in lambda-EMBL-4 isolated from a human brain genomic library. Two different cDNAs were found that were identical to the 5'-terminus to nt 1311 (counted from the A of the initiation codon, ATG) including the entire protein coding sequence. However, they were entirely dissimilar in the 3'-non-coding sequences. The genomic clones covered 94% of the full-length cDNA sequence. Three introns were found. The last exon spans contiguously the carboxyl terminus of the protein and the entire 3'-untranslated region of one of the two cDNAs with different 3'-ends. The origin of the 3'-portion of the other cDNA clone is not clear at this time.

The disease locus for the clinically heterogeneous childhood spinal muscular atrophies (SMA) maps to the chromosome 5 subregion, 5q11.2-13.3. The beta-subunit of beta-D-N-acetylhexosaminidase (hexosaminidase) (EC 3.2.1.52) (Hex B) maps to the same region, and the protein required for substrate recognition by this enzyme, GM2-activator protein, likewise maps to chromosome 5. We have investigated the possibility of allelic variation among some forms of SMA and hexosaminidase deficiency. Recombination between the Hex B and SMA loci eliminates this enzyme as a candidate site for defects causing the illness. Furthermore, we show that, despite previous evidence to the contrary, the GM2-activator locus does not map to chromosome 5, thereby eliminating it as a candidate gene for SMA.

We report the construction of a cDNA clone encoding a functional GM2-activator protein. The sequence of the complete 5' end of the coding region was determined by direct nucleotide sequencing of a fragment generated by multiple RACE PCR procedures from Hela cell cDNA. Specific oligonucleotides were synthesized from these data which allowed us to produce a PCR fragment that contained the complete coding sequence of the protein. This was then cloned into a mammalian expression vector. The ability of purified hexosaminidase A (beta-N-acetylhexosaminidase, EC 3.2.1.52) to hydrolyse labeled GM2 ganglioside was enhanced 10-fold more by the addition in the assay mix of lysate from transfected COS-1 cells than by the addition of identical amounts of lysate from mock transfected cells. Direct sequencing of PCR fragments from two sources also identified three polymorphisms.

Ganglioside GD1a-GalNAc was isolated from Tay-Sachs brain, tritium-labeled in its sphingosine moiety, and its enzymic degradation studied in vitro and in cultured fibroblasts. When offered as micelles, GD1a-GalNAc was almost not hydrolyzed by Hex A or Hex B, while after incorporation of the ganglioside into the outer leaflet of liposomes, the terminal GalNAc residue was rapidly split off by Hex a. In striking contrast to ganglioside GM2, the major glycolipid substrate of Hex A, the enzymic hydrolysis of GD1a-GalNAc was not promoted by the GM2 activator protein, although the activator protein did bind GD1a-GalNAc to form a water-soluble complex. Pathobiochemical studies corroborate these results. After incorporation of [3H]GD1a-GalNAc into cultured skin fibroblasts from healthy subjects and from patients with different variants of GM2 gangliosidosis, its degradation was found to be strongly attenuated in mutant cells with Hex A deficiencies such as variant B (Tay-Sachs disease), variant B1 and variant 0 (Sandhoff disease), while in cells with variant AB (GM2 activator deficiency), its catabolism was blocked only at the level of GM2. In line with these metabolic studies, a normal content of GD1a-GalNAc was found in brains of patients who had succumbed to variant AB of GM2 gangliosidosis whereas in brains from variants B, B1, and 0, its concentration was considerably elevated (up to 19-fold). Together with studies on the enzymic degradation of GM2 derivatives with modifications in the ceramide portion, these results indicate that mainly steric hindrance by adjacent lipid molecules impedes the access of Hex A to membrane-bound GM2 (whose degradation therefore depends on solubilization by the GM2 activator) and in addition that the interaction between the GM2. GM2 activator complex and the enzyme must be highly specific.

The complete amino-acid sequences of human ganglioside GM2 activator protein and cerebroside sulfate activator protein have been established by Edman degradation. The GM2 activator is composed of 162 amino acids, the first two serine residues being present in only 20% of the material. A single carbohydrate chain is N-glycosidically linked to Asn32. Three hydrophobic alpha-helices may contribute to its lipid-binding site. Three amino acids differ from those found by cDNA sequencing which may be due to a polymorphism. The cerebroside sulfate activator consists of 80 amino acids and carries one N-linked carbohydrate chain at Asn21. The C-terminal valine residue is lacking in about 80% of the material. In spite their similar functions, both activator proteins show no sequence or structural similarities.

Hexosaminidase A (HA) was isolated from human kidney and purified to an electrophoretically homogeneous state. The purification procedure included ion-exchange chromatography on DEAE-cellulose, gel filtration on Toyopearl HW-55 and chromatofocusing on PBE 94 (enzyme yield 26.6%, 1133.6-fold purification). The physico-chemical and kinetic properties of HA are as follows: Mr of the purified enzyme is approximately 100,000; Km for 4-methylumbelliferyl-2-acetamido-2-deoxy-beta-D-glucopyranoside is 0.6 mM; pH optimum is at pH 4.4-4.6; pI is 5.0. The amino acid composition of the purified enzyme was determined. A specific anti-HA antiserum was raised, which did not immunoprecipitate with fibroblast extracts characterized by a mutational blockade of HA synthesis. GM2 was isolated and purified from murine liver as well as from the brain of a female patient who died of Tay-Sachs disease. The label was introduced by way of treatment of GM2 with tritiated acetic anhydride. The specific radioactivity of [3H]GM2 was 521 and 2065 Ci/M, respectively. The label was introduced into the N-acetylneuraminic acid and GalNAc residues of these GM2 preparations. An activator protein capable of solubilizing the natural substrate of HA was isolated from human kidney and partially purified (with a 19.9% yield and 480-fold purification). The Mr of the purified activator protein was approximately 21,000. Purified HA hydrolyzed [3H]GM2 only in the presence of the activator protein. An addition of the activator to the incubation medium containing normal fibroblast culture extracts and [3H]GM2 caused an increase in the rate of substrate hydrolysis, tenfold, on the average.

A sensitive assay was developed to assess the ability of extracts from cultured fibroblasts to catabolize ganglioside GM2, in the presence of the natural activator protein but without detergents. This method, which permitted the reliable determination of residual activities as low as 0.1% of normal controls, was then used to measure ganglioside GM2 hydrolase activities in fibroblasts from several hexosaminidase variants. The residual activities thus determined correlated well with the clinical status of the respective proband: infantile Tay-Sachs (0.1% of normal controls), late-infantile (0.5%), and adult GM2 gangliosidoses (2%-4%) and healthy probands with "low hexosaminidase" (11% and 20%). In contrast, beta-hexosaminidase A levels as measured with the synthetic substrate 4-MU-GlcNAc could not be relied on for diagnostic purposes (the late-infantile patient studied retained 80% of the activity of controls).

In order to understand the etiology of Type AB GM2 gangliosidosis, we have purified and characterized the activator protein (GM2 activator) specific for the enzymic hydrolysis of GM ganglioside from normal human brain. The purified activator from human brain moved as one major protein band in various electrophoretic systems. We have also prepared the antiserum against this activator. The levels and the nature of GM2 activator and beta-hexosaminidase A were examined in the brains of five cases of GM2 gangliosidosis-one Type B, two O, and two Type AB. We found that the levels of GM2 activator in both Type B and Type O cases were markedly elevated, and that the two Type AB cases were the results of different causes. One case had a defective beta-hexosaminidase A and an elevated level of GM2 activator. Although this defective beta-hexosaminidase A could hydrolyze synthetic substrates, it was inactive in the cleavage of natural glycosphingolipids in the presence of the GM2 activator. It could, however, hydrolyze asialo-GM2 and GbOse4Cer in the presence of sodium taurodeoxycholate. The other case had normal beta-hexosaminidase A, but had a very low level of GM2 activator when analyzed by in vitro assay, suggesting the deficiency of this activator. By immunoelectrophoresis, this case was found to be completely devoid of the protein that cross-reacts with the antiserum against the GM2 activator.

The GM2 activator protein is a glycolipid-binding protein required for the lysosomal degradation of ganglioside GM2. A human fibroblast cDNA library was screened with mixtures of oligonucleotide probes corresponding to four different areas of the amino acid sequence. A putative clone (821 bp) which gave positive signals to all four probe mixtures was purified and sequenced. The sequence was colinear with the sequence of 160 amino acids of the mature GM2 activator protein. Availability of the cDNA clone should facilitate investigation into function of the GM2 activator protein and also into genetic abnormalities underlying GM2 gangliosidosis AB variant.

The physiological activator protein for the degradation of ganglioside GM2 by hexosaminidase A has been employed to assess the capability of cultured human fibroblast extracts to catalyze this ganglioside. This method permits a more reliable diagnosis of the different variants of GM2 gangliosidoses than the methods hitherto used. These either rely on artificial substrates or, when natural substrates are used, on detergents. Our method avoids a number of possible sources of error introduced by the unphysiological detergents, such as alteration of the isoenzymes' substrate specificity or inactivation of the enzymes. The range of application of the new method is discussed.

Gangliosidoses are very rare neurological diseases based on specific enzyme defects. They constitute models for the disruption of specific metabolic pathways and cellular functions with the ultimate consequence of manifest clinical symptoms. The investigation of the various steps involved in the generation of a given syndrome can therefore lead to a more profound understanding of the cell biology of the nervous system. In the present synopsis we try to briefly summarize some aspects of the present knowledge of pathophysiological mechanisms in GM1- and GM2-gangliosidoses.

A case of infantile GM2 gangliosidosis with high residual beta-hexosaminidase A activity toward the synthetic substrate 4-methylumbelliferyl-2-acetamido-2-deoxy-beta-D-glucopyranoside was diagnosed prenatally. Extracts from cultured amniotic fluid cells of the fetus had a hexosaminidase A activity of 27% of total hexosaminidase but were almost completely unable to degrade [3H]ganglioside GM2 (less than 0.5% of control values) when assayed in the presence of the natural activator protein. These results were confirmed by analyses of fetal muscle fibroblasts, liver, and brain. All tissues examined showed a profound deficiency of ganglioside GM2 galactosaminidase despite hexosaminidase A levels in the heterozygote range. In brain tissue, ganglioside GM2 content was elevated more than 4-fold. Hydrolysis of p-nitrophenyl glucosaminide-6-sulfate, a substrate specific for hexosaminidases A and S, by tissue extracts was also markedly reduced but the residual activities found (5% in liver, 12% in fibroblasts, and 16% in brain) were much higher than those with the physiological lipid substrate, ganglioside GM2.

An enzyme-linked immunosorbent assay was developed for the quantitation of the human activator protein for ganglioside GM2 degradation by beta-N-acetyl-D-hexosaminidase A. With this assay, various tissues and body fluids of normal subjects and of patients with variant forms of GM2 gangliosidosis were screened for their GM2-activator content. The highest content of this protein was found in kidney. Tissues from patients with variant AB of infantile GM2 gangliosidosis contained only low levels (up to 5% of that of normal controls) of cross-reacting material. On subcellular fractionation of normal human skin fibroblasts, the activator was found to co-distribute with the lysosomal marker enzyme beta-hexosaminidase.

The gene coding for the GM2 activator protein has been mapped to human chromosome 5, using an enzyme-linked immunoadsorbent assay (ELISA) to identify the human protein in human-mouse somatic cell hybrids.

The biosynthesis and secretion of lysosomal GM2-activator was studied in fibroblasts from controls and patients of GM2 gangliosidosis metabolically labelled with [3H]-leucine. Immunoprecipitation was performed with affinity-purified antibodies to human kidney GM2-activator protein. Normal fibroblasts and fibroblasts of variant B and O of GM2 gangliosidosis secrete GM2-activator protein as a 24-kDa polypeptide, which is able to stimulate degradation of ganglioside GM2 by beta-hexosaminidase A in the in vitro assay. In the presence of 10mM NH4Cl the rate of secretion is twice as high as in normal fibroblasts. Intracellularly, GM2-activator protein is represented in these cell lines by polypeptides with apparent molecular masses ranging from 21 kDa-22.5 kDa. Under the same labelling conditions, in two cell lines of patients with variant AB of infantile GM2 gangliosidosis intracellularly only traces of GM2-activator were detectable, whereas significant amounts of polypeptides with molecular masses between 25 and 26.5 kDa could be precipitated from the media of these fibroblasts.

After the revelation of the presence of ganglioside GM2 as the major ganglioside in the roe of striped mullet, Mugil cephalus [Li, Hirabayashi, DeGasperi, Yu, Ariga, Koerner & Li (1984) J. Biol. Chem. 259, 8980-8985], we have continued to investigate the catabolism of GM2 in this tissue. We have found that mullet roe contains a specific activator protein which stimulates the hydrolysis of GM2 carried out by the beta-hexosaminidase isolated from the same tissue. This activator has been purified by using conventional procedures including ammonium sulphate fractionation and chromatography on Sepharose 6B, DEAE-Sephadex A-50, octyl-Sepharose and Matrex Gel Blue A columns. This activator protein is also able to stimulate the hydrolysis of GM2 carried out by human beta-hexosaminidase A. Unlike human GM2-activator, the roe activator protein does not stimulate the hydrolysis of GgOse3Cer or GbOse4Cer. The molecular mass (18 kDa) of the roe activator protein was found to be similar to that of human GM2-activator; however, the pI (pH 4.1) was found to be lower than that of human GM2-activator. This is the first report on the presence of a GM2-activator protein in a source other than mammalian tissues.

Competition experiments were carried out on the hydrolysis of different substrates by beta-hexosaminidase A isolated from human liver. The results show that ganglioside GM2 in the presence of the GM2 activator protein and a new synthetic substrate, 4-methylumbelliferyl-beta-N-acetylglucosaminide 6-sulfate, are hydrolyzed at the same active site on the alpha subunit of beta-hexosaminidase A, whereas 4-methylumbelliferyl-beta-N-acetylglucosaminide is degraded predominantly by a different active site on the beta-subunit. This finding provides for the first time a possible molecular basis for the observation that, in variant B1 of the GM2 gangliosidoses, beta-hexosaminidase A has lost its activity toward GM2 ganglioside and the sulfated artificial substrate while being still able to hydrolyze the unsulfated artificial substrate at a normal rate. Furthermore, the finding that the GM2 activator protein inhibits the degradation of the sulfated substrate by beta-hexosaminidases A and S indicates that the alpha subunit common to both isoenzymes might provide a binding site for the activator protein.

Proteins which bind glycolipids with high specificity are tentatively divided into two groups. One group consists of activator proteins involved in the catabolism of glycolipids by acid lysosomal hydrolases. Two activator proteins, GM2-activator and sphingolipid activator protein-1, are critically appraised on their glycolipid-binding properties and on their activity to facilitate the transfer of glycolipids. These proteins are glycoproteins localized in the lysosomes. Their molecular weights are in a range of 21 000-27 000, and isoelectric points are 4-5. Glycolipid transfer protein (GLTP) is included in the other group. GLTP purified from pig brain has a molecular weight of about 20 000 and an isoelectric point of 8.3. GLTP facilitates the transfer of various glycosphingolipids and glyceroglycolipids between membranes. The protein does not facilitate the transfer of phospholipids or cholesterol. GLTP binds galactosylceramide. The galactosylceramide-GLTP complex participates in the transfer reaction as the intermediate. Each protein in both groups binds glycolipids with a characteristic specificity to the sugar moiety. A stoichiometry of 1 mol of lipid per mol of protein has been found in all three proteins. Proteins in both groups seem to have a hydrophobic region on their surface, since all three proteins have been efficiently purified by hydrophobic chromatography.