Hydrolysis of Tay-Sachs Ganglioside by @-Hexosaminidase A of Human Liver and Urine*

A crude /I-hexosaminidase fraction prepared by (NH,),SO( fractionation of human liver extract or urine was found to convert Tay-Sachs ganglioside, GalNAcPl ceramide (G& into NAN& + 3Galfil -+ 4 Glc + ceramide (GMM3). After separation of hexosaminidase A and B by DEAE-cellulose chromatography, only freshly fi-hexosaminidase A hydrolyzed GM2 although both forms were still active toward P-nitrophenyl-fl-D-N-acetylglucosaminide. A heat-stable, nondialyzable preparation obtained from the /?-hexosaminidase fraction of found the hydrolysis of by but isolated both the

A heat-stable, nondialyzable preparation obtained from the crude /?-hexosaminidase fraction of human liver was found to stimulate the hydrolysis of GM2 by /3-hexosaminidase A but not B isolated from both sources. Upon aging, P-hexosaminidase A would only hydrolyze GRIZ in the presence of the heat-stable preparation.
Extensive purification as well as aging tended to reduce the capacity of /?-hexosaminidase A to hydrolyze GM2 even in the presence of the heat-stable preparation.
Our results explain why P-hexosaminidase A has been previously reported by other investigators to hydrolyze Glvrz only with great difficulty.
Our results also relate the inordinate storage of GRIZ to the absence of P-hexosaminidase A in the classical form of Tay-Sachs disease.
Human tissues contain two P-hexosaminidase isosymes, phexosaminidase A and B (I). Okada and O'Brien (2), Sandhoff (3), and Hultberg (4) demonstrated that @-hexosaminidase A was deficient in the classical form of Tay-Sachs disease. This conclusion was based on the measurement of the enzyme activity with synthetic substrates such as p-nitrophenyl-or 4-methylumbelliferyl-P-D-N-acetylglucosaminide. have not beeu extensively studied. Although it has been reported that Ghl?: could be hydrolyzed by highly purified /?-hexosaminidase isolated from various tissues (5)(6)(7), in all cases, the rate of hydrolysis was extremely slow and required the use of radioactive GMZ if hydrolysis was to be detected at all. Wenger et al. (8) recently esamined the substrate specificity of P-hexosaminidase A and 13 isolated from human liver and concluded that neither enzyme hydrolyzed the terminal N-acetylgalactosaminyl unit from GM2. In order to elucidate the relationship between the inordinate storage of GMM2 and the absence of P-hexosaminidase A in Tay-Sachs disease, we have made an extensive examina.tion of the action of the @-hexosaminidases isolated from human liver and urine upon GM2.
EXPERIMENTAL PROCEDURE p-Nitrophenyl-P-D-N-acetylglucosaminide was purchased from Sigma Chemical Company.
GM2 was isolated from normal human brain (9) and GM3 from human plasma (10). Radioactive GMM2 (2300 cpm per nmole), 3H-labeled in the terminal N-acetylgalactosamine, was a gift of Professor Lars Svennerholm of the University of Giiteborg in Sweden. The asialo derivative of GILIZ was prepared by hydrolyzing GhlQ with 1 M HCOOH for 1 hour at 100" as previously described (11). Globoside was prepared from human red cell stroma (12). Analytical thin layer chromatography of gangliosides and neutral sphingoglycolipids was performed on plates coated with a layer (0.25 mm) of Silica Gel G and developed with chloroform-methanol-water (60 : 32 : 7 or 65 : 25 : 4). Gangliosides and neutral sphingoglycolipids were visualized by spraying the plates with anisaldehyde (13) or resorcinol reagent (14). For convenience, p-nitrophenyl-fi-D-N-acetylglucosaminide was used routinely to follow P-hexosaminidase activity during enzyme isolation. The procedure for using p-nitrophenyl-/3-n-N-acetylglucosaminide as substrate has been described elsewhere (15). One unit of enzyme was defined as the amount of enzyme which hydrolyzes 1 /Imole of p-nitrophenyl-P-D-N-acetylglucosaminide per min at 37". The specific activity of the enzyme was expressed as units per mg of protein.
Protein was determined by the method of Lowry et al. (16) with crystalline bovine serum albumin as the standard.
The procedures for hydro-after exhaustive dialysis against distilled water and lyophilizalyzing sialic acid from GM3 with cholera neuraminidase and tion, was designated as the heat-stable preparation. This galactose from lactosylceramide with jack bean fi-galactosidase preparation alone does not have any activity toward GILIZ or have been previously described (10, 11).
p-nitrophenyl-,&N-acetylglucosaminide. Isolation of @Hexosaminidases from Human Liver-Enzyme isolation was carried out at O-5". Human liver (400 g) obtained post mortem was diced, homogenized with 2400 ml of 0.1% (v/v) Triton X-100, in a Waring blendor at 12,500 rpm for 2 min and centrifuged at 8,000 X g to obtain a clear brown crude extract.
Solid citric acid was added to adjust the pH of the extract to 4.3 while stirring.
After standing overnight, the precipitate was removed by centrifugation and discarded. The pH of the supernatant was adjusted with solid tris(hydroxymethyl)aminomethane to pH 7.0 and was then brought to 30% saturation with solid (NH.&S04.
After settling overnight, the sample was centrifuged to remove the precipitate, and the supernatant was brought to 507, saturation with solid (NH&SO+ The precipitated protein was harvested the next day by centrifugation and resuspended in 35 ml of 0.01 M sodium phosphate buffer, pH 6.0. This preparation was designated as the crude fi-hexosaminidase.
No neuraminidase activity capable of releasing sialic acid from GRI1, GM2, or GM8 was detectable in this preparation.
Partially purified @-hexosaminidase A and B were prepared from the crude fi-hexosaminidase by DEAE-cellulose chromatography according to the procedure described by Robinson and Stirling (1). Partially purified P-hexosaminidase A was further purified by CM-cellulose chromatography.
After dialysis against 0.05 M sodium citrate buffer, pH 4.4, the enzyme was applied to a CM-cellulose column equilibrated with the same buffer. After washing the column with the same buffer, the enzyme was eluted with a linear salt gradient from 0 to 0.4 M NaCl in the same buffer. The fractions containing fl-hexosaminidase activity were pooled, precipitated by reverse dialysis against (NH&SO+ and designated as highly purified liver flhexosaminidase A. This procedure is summarized in Table I.

Isolation of &Hexosaminidases from Human
Urine-Fresh urine from healthy male subjects was pooled and made 60% saturated with (NH&S04. After standing in the cold (4") overnight, the clear supernatant was removed by decantation and the thick slurry at the bottom of the container was centrifuged to obtain a packed dark brown precipitate.
The yield of this precipitate was about 10 to 15 g/10 liters of urine.
The precipitate obtained from 100 liters of urine was pooled and extracted twice with 500 ml of 0.05 M sodium phosphate buffer, pH 7.0. The P-hexosaminidase in the extracts was then precipitated between 35 to 65% saturation with (NH&SO+ The precipitate was dissolved in 100 ml of 0.1 M sodium phosphate buffer, pH 7.0, and was designated as crude urine extract. The crude urine extract (20 ml) was applied to a Sephadex G-ZOO column (5 X 90 cm) previously equilibrated with 0.1 M sodium phosphate buffer, pH 7.0. The column was eluted with the same buffer. Those fractions containing fl-hexosaminidase activity were pooled and designated as the crude urine ,&hexosaminidase.
Partially purified urine P-hexosaminidase A and B were also prepared from the crude fl-hexosaminidase fraction by DEAE-cellulose chromatography according to Robinson and Stirling (1) fi-Hexosaminidase A obtained at DEAE-chromatography was pooled and dialyzed against 0.05 M sodium acetate buffer, pH 4.4, and applied to a CM-cellulose column (2 x 30 cm) equilibrated previously with the same buffer. After washing the column with the starting buffer and then with 0.05 M sodium acetate, pH 4.9, the P-hexosaminidase was eluted with 0.01 M citrate buffer, pH 6.2. This preparation was designated as highly purified urine fi-hexosaminidase A. This procedure is summarized in Table I. aminidase-The crude P-hexosaminidase isolated from human liver was heated at 100" for 5 min and centrifuged at 17,000 X g Details of the purification steps are described in the text.
Enzymatic Hydrolysis of GMa, Asialo G,z, and Globoside-Enzyme activity was assayed in the presence or absence of 2 mg (dried weight) of the heat-stable preparation.
When GM2 was used as substrate, 20 nmoles in 100 ~1 of chloroform-methanol (1: 1) were evaporated to dryness under a stream of nitrogen. Then, 0.7 ml of 0.05 M sodium citrate buffer, pH 4.0, was added and the mixture was sonicated in a Cole-Parmer model 8845-3 ultrasonic bath for 3 min. The reaction was started by adding 0.1 ml of enzyme solution containing 1 to 10 units of enzyme and incubated at 37" for 2 to 16 hours. The reaction was stopped by the addition of 4 volumes of chloroform-methanol (1 :I). After thorough mixing, the chloroform phase and the aqueous phase were separated by centrifugation.
Under these conditions, more than 9Og, of the GM2 and Gh13 was recovered in the lower phase. The lower phase was carefully removed, evaporated to dryness under a stream of nitrogen and analyzed by thin layer chromatography.
When asialo GMM2 or globoside was used as substrate, the procedure was essentially the same as above except that the buffer contained 100 pg of sodium taurocholate per 100 ~1 and the reaction was terminated by the addition of 4 volumes of chloroform-methanol (2:l) as previously described (11)

AND DISCUSSION
Electrophoretic examination (18) revealed that /3-hexosaminidase A and B of liver and urine were completely separated at the DEAE-chromatography step. With p-nitrophenyl-P-n-N-acetylglucosaminide as substrate, the pH optimum was between 4.0 and 4.4 for both ,&glucosaminidase A and B of human liver and urine.
For hydrolysis of GMz by the crude fl-hexosaminidase of human liver and urine, the pH optimum was between 3.8 and 4.0.
All @-hexosaminidase preparations from liver and urine obtained at different purification steps hydrolyzed p-nitrophenyl-P-n-N-acetylglucosaminide, globoside, and the asialo derivative of GhlZ. However, only the crude @-hexosaminidase obtained at the (NH&S04 fractionation step efficiently converted GM2 into G M3. This conversion of GM2 to GM3 by the crude P-hexosaminidase of human liver is shown in Fig. 1. Hydrolysis as a function of GMz concentration was also investigated. Visual examination of the plate clearly indicated that the amount of product was a function of substrate concentration as in a typical enzyme reaction.
Hydrolysis was roughly linear with time at a given GMz concentration.
At the substrate concentration of 11 MM, about 30% of the GMz was converted to GM3 by 5 units of crude liver /3-hexosaminidase in 16 hours under the standard assay conditions which may not be the optimal condition for hydrolyzing GM2 by liver and urine hexbsaminidase.
In Fig. 2A, the slow hydrolysis of GM2 to GM3 by partially purified liver fl-hexosaminidase A is illustrated. The increased hydrolysis of GhlZ by @-hexosaminidase A in the presence of the heat-stable preparation is also shown. Similar results were also FIG. 1. Conversion of GM% to GM~ by the crude p-hexosaminidase of human liver.
The enzyme (5 units) was incubated with 20 nmoles of GM~ in 0.8 ml of 0.05 M sodium citrate buffer, pH 4.0, at 37" for 15 hours. Detailed incubation conditions are described -. _. . . obtained with the corresponding urine fl-hexosaminidase A preparation (Fig. 2B). While freshly prepared @-hexosaminidase A of urine or liver obtained by DEAE-cellulose chromatography still retained some ability to cleave GM2, the corresponding @hexosaminidase B was not able to hydrolyze G,% even in the presence of the heat-stable preparation.
It is of interest to note that highly active crystalline jack bean P-hexosaminidase (15) could not hydrolyze GM2 under any circumstances. Sodium taurocholate could not replace the heat-stable preparation in stimulating the hydrolysis of GNIZ by P-hexosaminidase. Liver P-hexosaminidase A, after further purification by CMcellulose chromatography, was completely inactive toward GM2. The ability of this preparation to hydrolyze GM2, however, could be restored by adding the heat-stable preparation when the highly purified enzyme was fresh. After storage at 4" for 5 days, this preparation could no longer be stimulated by the heatstable factor to hydrolyze GMz, although it still retained the activity toward p-nitrophenyl-P-D-N-acetylglucosaminide.
Similar results were also obtained for the corresponding P-hexosaminidase A preparation isolated from urine. Finally, the recombination of P-hexosaminidase A and B without the heatstable preparation did not restore the capacity to hydrolyze GM2. By using radioactive GMs, the partially purified P-hexosaminidase A isolated from human liver hydrolyzed 0.3 nmoles of GMz per hour per unit of enzyme in the presence of the heat-stable preparation.
In the absence of the heat-stable preparation, it only hydrolyzed 0.06 nmole of G,z per hour per unit of enzyme. The freshly prepared P-hexosaminidase A purified by CM-cellulose chromatography hydrolyzed 0.06 nmole of GMz per hour per unit of enzyme in the presence of the heat-stable preparation. In the absence of the heat-stable preparation, no significant hydrolysis was observed.
Under the same condition, phexosaminidase B did not hydrolyze GM2 in the presence or absence of the heat-stable preparation.
Sandhoff (6) reported that a highly purified fl-hexosaminidase A (enriched 3000-fold) of human liver hydrolyzed G,s at a rate of 0.04 nmole per hour per unit of enzyme in the presence of sodium taurocholate.
The reason for the low rate of hydrolysis could be due to the removal of the heat-stable factor during the course of purification.
Sandhoff also found that GMZ was not hydrolyzed by liver @-hexosaminidase 13. Wenger et al. (8) recently reported that they could not detect the hydrolysis of G M2 by either P-hexosaminidase A or B isolated from human liver.
The reason for their negative result could be due to the fact that their enzymes were isolated from the frozen tissues stored at -20" for 1 to 3 years. We found that freezing and aging of the tissues greatly reduced the ability of fi-hexosaminidase A to hydrolyze GMz. As can be seen from Figs. 1 and 2, the mobility of GM3 produced from brain GM2 was slightly faster than the standard GM3 isolated from human plasma. This is clearly due to the differences in their fatty acid composition.
While stearic acid is the major fatty acid of sphingoglycolipids isolated from brain, nervonic and palmitic acids are the major fatty acids found in the sphingoglycolipids of plasma (10). The identity of the product as Gh13 was confirmed by the following: (a) hydrolysis of the chloroform-methanol extract of the reacted mixture with 1 M HCOOH for 1 hour produced asialo GhlS from unreacted Gh12 and lactosylceramide from GM3; (b) the standard GMB isolated from human plasma (10) and the Gh13 produced from GMM2 were both converted to lactosylceramide by neuraminidase from Vibrio cholera; and (c) the lactosylceramide produced by neuraminidase, in both cases, was further converted to glucosylceramide by jack bean P-galactosidase (11).
The fact that the heat-stable preparation activates fl-hexosaminidase A to hydrolyze GM2 suggests the presence of a "factor" in this preparation This factor is apparently removed during the purification of the enzyme. Work on the isolation and characterization of this factor is in progress and will be published elsewhere.
It is of interest to note that Ho et al. (19) have reported a low level of glucocerebrosidase activity (toward glucocerebroside) in a particulate fraction of human spleen which is greatly stimulated by the addition of a soluble glycoprotein factor.
They also observed a discrepancy between the ability 4-methylumbelliferyl-P-glucoside.
Similarly, it appears that full /3-hexosaminidase activity towards GM2 may require the interaction of two components, the purified P-hexosaminidase A enzyme and a heat-stable factor.
It is evident that a glycosidase whose purification has been followed using synthetic subst,rates may not necessarily act on natural substrates. This fact cannot be overemphasized.
Related to this is the fact that storage of purified P-hexosaminidase A in the cold has an adverse effect on its ability to hydrolyze GM2 but not p-nitrophenyl-B-D-iv-acetylglucosaminide.
Our results explain why previous reports by other investigators indicated hydrolysis of GM2 only with great difficulty when using highly purified P-hexosaminidase A. There are two possible catabolic pathways for GMM%. GalNAcpl + 4GalPl + 4Glc + Cer Our results suggest the existence of pathway I in human liver. The enzyme responsible for this pathway is fl-hexosaminidase A. In addition, our results establish why the absence of fl-hexosaminidase A causes the accumulation of GR?2 in various tissues of the classical Tay-Sachs patient.