gm1 ganglioside @-galactosidase a · gm1 ganglioside @-galactosidase a purification and studies of...

THE JOURNAL OP BIOLOQICAL CHEMISTRY Vol. 249, No. 24, Issue of December 25, pp. 7969-7976, 1974

Printedin U.S.A.

GM1 Ganglioside @-Galactosidase A

PURIFICATION AND STUDIES OF THE ENZYME FROM HUMAN LIVER*

ANTHONY G. W. NORDEN, LINDA L.

From the Department of Neurosciences,

(Received for publication, April 28, 1974)

TENNANT, AND JOHN S. O’BRIEN

University of California at San Diego, La Jolla, California 92037

SUMMARY

GMM1-ganglioside fi-galactosidase A (EC 3.2.1.23) has been purified 17,000-fold from human liver. The enzyme appeared as a single band of protein on polyacrylamide gel electrophoresis. Double immunodiffusion of rabbit anti-p- galactosidase A antiserum against either GM1 /3-galactosidase A or liver supematant gave a single precipitin band. The apparent molecular weight was 65,000 to 75,000 by gel filtration of the native enzyme and 72,000 by sodium dodecyl sulfate gel electrophoresis of the denatured, reduced, and carboxymethylated enzyme. The purified enzyme liberated D-galactose from synthetic chromogenic and fluorogenic P-D-

galactosides, as well as lactose, N-acetyllactosamine, and the glycoprotein asialofetuin. The same catalytic site(s) appeared to be responsible for the hydrolysis of GMI-ganglioside and 4-methylumbehiferyl-&r+galactopyranoside. The enzyme also cleaved B-D-fucoside and a-L-arabinoside link- ages. Lactulose, galactosyl-hydroxylysine, galactosyl-hy- droxylysyl peptides, galactocerebroside, lactosylceramide, and monogalactosyldiglyceride were very poor susbtrates. Anti-/3-galactosidase A antiserum quantitatively precipitated both /3-galactosidase A and /3-galactosidase B. Neutral P-glucosidase (“nonspecific” /3-galactosidase), galactocerebroside /3-galactosidase, and lactosylceramide @-galactosidase were not precipitated by anti-P-galactosidase A antiersum.

A large number of mammalian @-n-galactosidasesl (EC 3.2.1.23) has been detected in various tissues (see Refs. 1 and 2 for review). These have been implicated in the degradation of carbohydrates (3, 4), glycolipids (5, 6), glycoproteins (7), and glycosaminoglycans (8). Studies of the aglycone specificity of these enzymes are incomplete. Such studies are difficult due to the multiplicity of P-galactosidases present, each of which may hydrolyze a number of substrates. We have identified previously two GMl-ganglioside2 P-galactosidases in human liver,

* This project was aided by the National Institutes of Health Grant NS 08682. United States Public Health Service Program Project Grant GM 17702, the National Foundation-March of Dimes Grant CRBS-242, and the A. P. Sloan Foundation.

1 All monosaccharides in this report have the n-glycopyranose configuration unless otherwise stated.

2 The abbreviations used are: Gnl and GXZ, gangliosides Gnr and Gnz according to the nomenclature of Svennerholm (9): Gni,

a major form termed “A” (mol wt 60 to 70 x lOa) and a minor form “B” (mol wt 600 to 800 X 103) (10). Ho et al. reported similar results at the same time (11). /3-Galactosidases have been purified to varying degrees from rabbit brain (12), calf intestine (3), human liver (13, 14), and bovine testis (8). These enzymes have been defined usually by their activity on low molecular weight chromogenic and fluorogenic &galactosides. Reports of the activities of these preparations with natural substrates such as lactose (3, 4, 12), Gal-Cer (6, 13), Lac-Cer (12, 13), GMI-ganglioside and its asialo-derivative G,l (8, 10, 11, 13), and various glycoproteins (8) have differed.

Thus far no studies have appeared in which a human P-galactosidase has been purified and studied with respect to its kinetics, specificity and structure. This information is necessary for a better understanding of those mutations which affect GM1 p- galactosidase activities in humans, the GMM-gangliosidoses (15, 16). We present here the purification and kinetic, structural, and immunochemical studies of the major human liver GM1 ganglioside fl-galactosidase.

METHODS

Con A-Sepharose (8 mg of Con A per ml of resin), Sephadex G-150, and G-100 and Sepharose 4B and 6B were from.Pharmacia. DEAE-cellulose (DE52) was from Reeve Aneel. ECTEOLA-cellulose (Cellex E)‘was from Bio-Rad. SubstFates were purchased from the following: Gal pl+3 arabinose and lactulose from Pfan- stiehl, p-aminophenylthio+n-galactoside from Cycle; Lac-Cer from Miles-Yeda; chromogenic and fluorogenic substrates from Sigma or Koch-Light. Gal-Cer and ceramide both were isolated from human white matter (17). Glu-Cer was isolated from the spleen of a patient with Gaucher’s disease (17). Gnrz was isolated from the brain of a patient with Tay-Sachs’ disease (18). L-W

Lecithin (egg yolk, type III-E) was from Sigma. Galactose de- hydrogenase was from Boehringer. Freund’s adjuvants were from Difco. Other chemicals were from Sigma or Schwarz-Mann.

Gal-Hyl was the generous gift of Dr. L-W. Cunningham, De- partment of Biochemistry, Vanderbilt University, Nashville, Tenn. [Gal-aH]Lac-Cer and [GaZ-l~C]monogalactosyldipalmitin were the generous gifts of Dr. D. A. Wenger, B. F. Stolinksy Labo- ratories, University of Colorado. Prof. W. M. Watkins of the Lister Institute, London kindly gave us the N-acetyllactosamine.

Gal 81-3 GalNAc (314 (NeuAc a2+3) Gal 81+4 Glu Bl-tl’ceramide; Grit, GalNAc fl14 (NeuAc or2+3) Gal pl+4 Glu pl+1’ ceramide: GA,. Gal al-+3 GalNAc al+4 Gal 81+4 Glu Bl-+l’ ceramide: Gal-Cer; Gal al-l’ ceramide; Lac-Cer, Gal pl+4 Glu pl+l’ (N: stearoyl) dihydrosphingosine; Con A, concanavalin A; Con A- Sepharose, Concanavalin A bound covalently to Sepharose (aga- rose); Gal-Hyl, galactosyl pl hydroxylysine; Gal-Hyl peptides, Gal-Hyl containing peptides prepared from collagen as described in Table III, Footnote f.

7969

by guest on June 1, 2020http://w

ww

.jbc.org/D

ownloaded from

oprn manager datamatics

http://www.jbc.org/

7970

Liquid scintillation counting and radioscans of thin layer chromatography plates were carried out as described previously (10). All of the operations were performed at 1-4” unless otherwise stated. Protein was measured by the method of Lowry et al. (19) following base digestion in the case of homogenates. Bovine serum albumin was used as the protein standard. Sodium chloride concentrations in column effluents were determined by comparison with conductivities of standard sodium chloride solutions. Solutions were concentrated by ultrafiltration using UM-10 mem- branes (Amicon) at 50 psi.

Liver tissues, obtained at autopsy from adult male patients who died from trauma were stored at -20” for up to 6 months before use. These appeared disease-free on gross inspection. No altera- tion in the specific activity of 4-methylumbelliferyl /3-galactosidase occurred during this time.

Buffer I was 500 mM sodium citrate-10 mM sodium chloride-l mM EDTA, pH 4.00. Buffer II was 10 mM sodium phosphate-10 mM sodium chloride-l mM EDTA, pH 7.00.

Enzyme Assays-4-Methylumbelliferyl fl-galactosidase and neutral p-glucosidase were assayed as reported previously (20). 4-Methylumbelliferyl a-L-arabinosidase was assayed in the same way. Chromogenic p-galactosidase activity was measured in the same assay system as that for 4-methylumbelliferyl p-galactosidase except for the use of a spectrophotometric rather than fluorometric measurement. Absorbances were determined at the following wavelengths for the aglycones shown: phenyl, 285 nm; o- and m-nitrophenyl, 420 nm; p-nitrophenyl, 400 nm; l-napthyl, 322 nm; and 2-napthyl, 330 nm. Extinction coefficients for these aglycones were determined under the conditions of assay. For disaccharide substrates, Gal-Hyl and Gal-Hyl peptides, incubations were carried out as for 4-methylumbelliferyl &galactosidase (20). Galactose release was measured as follows: 0.5 ml of 0.25 M

Tris-chloride, pH 8.0, was added, the tubes were immersed in a boiling water.bath for 2 min and then cooled in an ice bath. One- half milliliter of 5 mM NAD and 50 rre Der ml of nalactose dehvdro- genase in 0.25 M Tris-chloride, pH 8.0, were idded followed by incubation at 37” for 15 min. NADH production was then measured fluorometrically as for 4-methylumbelliferone. Asialofetuin fl-galactosidase assays were carried out as described previously using [Gal-3H]asialofetuin (21). GM~-ganglioside fl-galactosidase was measured as described previously (lo), except that for some assays, the DEAE-cellulose microcolumn was reduced in size to 0.9 x 2.0 cm, the volume of 10 mM galactose washes was reduced to two of 0.6 ml and “Riafluor” (New England Nuclear) was used for scintillation counting. These changes reduced the chemi- luminescence of the samples but made no quantitative difference to the assay apart from a slight improvement in counting effi- ciency. Gal-Cer fi-galactosidase, Lac-Cer p-galactosidase, and monogalactosyldipalmitin p-galactosidase were measured by the method of Wenger (22). Appropriate dilutions were made in the buffer used for &say containing-1 mg per ml of bovine serum albumin and 1 mM EDTA as well as 10 mM NaCl for the Gal-Cer. Lac-Cer, and monogalactosyldipalmitin substrates. When the assay buffer contained detergent, this was omitted from the dilu- tion buffer. Assay times were usually 15 min, except for the Gal- Cer. Lac-Cer. and monogalactosvldipalmitin substrates, in which

months) were brought to pH 4.6 by slowly adding Buffer I with stirring. The solution was stirred for 30 &in and then centrifuged at 50.000 X B for 30 min. The suDernatant was dialvzed against three’changei each of 4 liters of B&er II for 24 hours”(total&aly- sis time) and centrifuged again at 50,000 X g for 3d min. The suDernatant was aDDlied directlv to DEAE-cellulose as described in-Fig. 1.

-_

Pooled fractions from the DEAE-cellulose column shown in Fig. 1 were concentrated to 5.0 ml and applied to a Sephadex G-150 column as described in Fig. 2. Pooled fl-galactosidase A activity from the Sephadex G-150 chromatography was concentrated to 4.8 ml and applied to ECTEOLA-cellulose (“ultrapurified” as described by the manufacturer) as in Fig. 3. The fl-galactosidase activity was pooled as in Fig. 3 and concentrated to 1.2 ml which was dialyzed overnight against 4 liters of Buffer II. Aliquots of 5 to 50 ~1 were then frozen at -20” in tightly sealed capillary tubes prior to use. The enzyme did not lose activity in this state for at least 1 month.

Pooled P-galactosidase B activity from Sephadex G-150 chromatography (Fig. 2) was concentrated to 5.0 ml and applied to a column of Sepharose 6B as described previously for liver supernatant (lo), except that the elution buffer included 10 mM NaCl. The pooled fractions of B enzyme were concentrated to 1.0 ml and stored at -20” as for the A enzyme.

Polyacrylamide gel electrophoresis was carried out in the anionic svstem of Williams and Reisfeld. modified bv the inclusion of 1 mM NaCl in the stacking gel buffer (23). To detect any ensy- matic activity on the gels, they were sliced into 0.12~cm sections and assayed for 15 min with 0.15 ml of the 4-methylumbelliferyl p-galactosidase assay solution as described previously (20). Pro- tein on the gels was detected by staining with Coomasie blue.

Immunological Methods-Male Norwegian white rabbits weigh- ing about 2 kg were given injections of 66 rg of 17,000-fold purified &galactosidase A (33 pg in each footpad) in complete Freund’s adjuvant on Day 0. On day 10, 21 pg in incomplete Freund’s adjuvant was injected intramuscularly. Control antiserum was prepared in the same way except that fl-galactosidase A was replaced by an equivalent volume of Buffer II. Preimmunization sera were also collected. Antisera used in these studies were obtained on Day 30. Double immunodiffusion was performed at 1-4” on agar plates (Hyland 085.073C) and read after 40 hours.

Zdentijkation of Reaction Products-To identify the lipid reaction products of the action of &galactosidase A on GYI, 87 ng of enzyme were incubated for 15 min in 100 ~1 of the standard assay system (10). A control lacking enzyme was also incubated. Then, 0.4 ml of chloroform-methanol-2.5 N ammonia (4:1:0.05) and 0.25 ml of water were added to each tube. Samples were mixed and centrifuged at 800 X g for 5 min. The lower phase was dried, redissolved in 50 ~1 of chloroform-methanol (2:l) and 20 ~1 were applied to a silica thin layer chromatography plate (Brink- mann 5763) alongside standards of GM~, GM~, and sodium taurocholate. The thin layer chromatography plate was developed in n-propyl alcohol-water (7:3) and visualized with an orcinol spray

(24).

RESULTS case they were 2 hours. ipart fr-om column effluents, the linearity of assays was established with respect to both the time of incuba- Pur$cation-The purification of @-galactosidase A is described

tion and the auantitv of m-otein added. Controls were always under “Methods” and Figs. 1 to 3 and is summarized in Table I. I ^

included to account for any apparent nonenzymatic hydrolysis. Puri$cation-(Certain details are included in the legends of

Unless otherwise stated,;11 of the results in this report, except

Figs. 1 to 3.) fi-Galactosidase refers to 4-methylumbelliferyl those of Table II, were obtained either with the most highly

fl-galactosidase unless otherwise stated. Livers used for the puri- purified preparation of Table I, or of a second preparation of

fication had fl-galactosidase-specific activities in supernatants enzyme purified by the same method which had a specific activ- (prepared by Method B of Ref. 10) of 0.005 to 0.014 pmol per mg ity of 45.0 I.tmol per mg per min. Although /3-galactosidases A per min. Each liver was examined for the presence of any electrophoretic variants by starch gel electrophoresis (10); none were

and B co-elute from DEAE-cellulose, they are separable on

found. Preparat.ion of liver supernatant from two livers and Con DEAE-Sephadex A-25 by salt gradient elution with P-galacto-

A-Sepharose chromatography were carried out as before (21), sidase B being bound more tightly than @-galactosidase A.8

except that both the wash buffer and the elution buffer for chroma- When P-galactosidase A was subjected to gel filtration on Seph- tography contained 0.5 M NaC1. It was helpful to layer approximately 4 cm of Sepharose 4B on top of the Con A-Sepharose bed.

arose 6B-or Sephadex G-100 or-starch gel electrophoresis (iO),

The layer of lipid which formed on the Sepharose 4B as the super- no @-galactosidase B activity could be detected. Purified /3-

natant was run through was removed periodically to prevent lipid galactosidase A had the same electrophoretic mobility on starch

from clogging the column. gel as the A band in the initial liver supernatant. The final The pooled fractions of p-galactosidase from the Con A-Sepha-

rose column (which were stable frozen at -20” for at least 3 a A. G. W. Norden and J. S. O’Brien, unpublished experiments.


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7971

0.6

r

0.5%

5 0.4 w

s

0.3 E

0.2

0.1

0.0

1

0 500 IWO

EFFLUENT VOLUME (ml1

FIG. 1. DEAE-cellulose chromatography of human liver /3- galactosidase. Dialyzed supernatant (220 ml) from the pH 4.6 precipitation step was applied to a DEAE-cellulose column (2.5 X 20 cm) pre-equilibrated with Buffer II. After all of the supernatant had been applied, Buffer II alone was run on (indicated by begin wash). At the point marked begin NaCl gradient, a gradient of NaCl in Buffer II was begun. Tubes of @-galactosidaee shown by the horizontal bar were pooled. The flow rate was 1.3 ml per min and 5.6-ml fractions were collected. 4-Methylumbelli- feryl (4.JfU) @-galactosidase, O-O; protein (-J&O), O-O; NaCl gradient, A-A.

I I I I 1 150 200 250 300

1 0.4

0.3 z s u

z

0.2 I2 L

E

0.1

0.0

EFFLUENT VOLUME (ml)

FIG. 2. Sephadex G-150 chromatography of human livero-galactosidases. Five milliliters of the concentrated pool of activity from DEAE-cellulose was applied to a Sephadex G-159 bed (2.5 X 98 cm) pre-equilibrated with Buffer II and eluted with the same buffer. b-Galactosidases A and B were pooled separately as shown by the horizontal bars. The flow rate was 0.4 ml per min and 2.7-ml fractions were collected. 4-Methylumbelliferyl (4MU) fl-galactosidase, 0-O ; protein (A,so), A-A.

specific activity of fi-galactosidase B, following Sepharose 6B chromatography, was 5.4 pmol per mg per min. No attempt was made to assess its purity or to purify it further. Prepara- tions of /3-galactosidase B were initially free of A activity following Sepharose 6B chromatography, but a small amount of A

0 I I I I 0 100 200 300 400

EFFLUENT VOLUME (ml)

4

200

150 ZI E

ii

100 :

s ”

3 0

50 =

0

FIG. 3. ECTEOLA-cellulose chromatography of human liver fl-galactosidase A. The concentrated p-galactosidase A activity (4.8 ml) from Sephadex G-150 was applied to a ECTEOLA-cellulose bed (0.9 X 20 cm) pre-equilibrated with Buffer II and washed with the same buffer up to the point, indicated by begin NaCl gradient. A gradient of NaCl in Buffer II was applied. Tubes shown by the horizontal bar were pooled. The flow rate was 0.2 ml per min and fractions of 3.0 ml were collected. 4-Methylumbelliferyl (4MU) fl-galactosidase, 0-0, NaCl gradient, O-O.

activity was detectable by starch gel electrophoresis after storage for 2 days at l-4’.

Polyacrylamide gel electrophoresis of 25 to 50 pg of fl-galactosidase A demonstrated a single band of protein which coincided with a single band of enzymatic activity. In one preparation, two barely detectable bands of protein of slower mobility than P-galactosidase A were found when 50 /Ig of protein were applied to the gel. Gel electrophoresis of the denatured, reduced, and carboxymethylated enzyme in the presence of sodium dodecyl sulfate revealed only a single band (Fig. 4). Alcian blue staining by the method of Wardi and Michos (25) on 25+g or 50-pg samples under nondenaturing conditions failed to disclose coinci- dent carbohydrate. Double immunodiffusion of P-galactosidase A or liver supernatant (prepared by Method B of Ref. 10) and anti-p-galactosidase A antiserum showed a single precipitin band (Fig. 5).

As a further criterion of purity, &galactosidase A (2.1 mg in 4.5 ml of Buffer II), from preparations different from those in Tables I and II, was applied to a column of DEAE-cellulose (of the same dimensions and using the same elution conditions as those of the ECTEOLA-cellulose column of Fig. 3). The enzyme eluted as a single peak with constant specific activity. Thus, the mean specific activity of five tubes comprising at least 90% of the total activity was 25.7 (range 24.3 to 26.4) pmol per &o per min. The pooled tubes had a specific activity of 46.3 pm01 per mg per min.

Molecular Weight-Gel filtration was used to estimate the molecular weight of native /3-galactosidase A as described previously for liver supernatant (lo), except that the elution buffer contained 10 mM NaCl and the sample was 10 pg of /3-galactosidase A dissolved in 2 ml of 5 mg per ml of bovine serum albumin in Buffer II. The elution volume corresponded to a molecular weight of 65,000 to 75,000. In separate experiments using the partially purified fl-galactosidases following the DEAE-cellulose step (Table I), no difference in the elution volumes, or the pro-


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7972

Fraction

Homogenate .................. Supernatant .................. Concanavalin A-Sepharose. ... pH 4.6 Supernatant ........... DEAE-cellulose. .............. Sephadex G-1505 .............. ECTEOLA-cellulose ..........

TABLE I

PuriJication of P-galactosidase A

VOhlW Activity Protein

ml jmol/min +w

9,270 a33 315,000 g,ooo 774 152,000

218 444 1,920 220 315 352

5.0 159 64 4.8 96 10.0 1.2 44.5 1.0

Specific activity Purification RlXOVWy

pmol/mg/min -fda %

0.0026 1 100 0.0051 2.0 93 0.23 89 54 0.89 342 38 2.5 970 19 9.5 3,660 12

44.5 17,100 5.4

0 p-Galactosidases A and B are separated on this column. Activities prior to this step therefore refer to a mixture of the A and B enzymes. The proportion of the two is estimated to be 85% A and 15% B from gel filtration of the supernatant on Sepharose 6B as described previously (10).

FIG. 4 (left). Polyacrylamide gel of human liver p-galactosidase A in the presence of sodium dodecyl sulfate, stained with Coomasie blue. fi-Galactosidase A (25 pg) was denatured, reduced, and carboxymethylated and subjected to electrophoresis in a 5yo polyacrylamide gel by the method of Weber et al. (26). The cathode and origin are uppermost. The gels were 7 cm long prior to staining.

FIG. 5 (right). Double immunodiffusion (Ouchterlony) plate of &galactosidase A and liver supernatant with anti-&galactosidase A antiserum (center well). The procedure is described under “Methods.” Contents of the wells were: 1, p-galactosidase A (20 ng); 8, fl-galactosidase A (50 ng); 5, liver supernatant diluted to 0.044 pm01 per ml per min of 4-methylumbelliferyl p-galactosidase activity; 4, fi-galactosidase A (20 ng); 6, liver supernatant, 0.44 pmol per ml per min of 4-methylumbelliferyl fl-galactosidase activity. Liver supernatant was prepared by Method B of Ref. 10 except for the use of a 1:l (w/v) homogenate. Dilutions were in 10 mM sodium phosphate-10 mM NaCl, pH 7.0. Each well contained 5 ~1. No precipitin bands were seen either when control or preimmunization sera were used or when &galactosidase A was replaced by 10 mM sodium phosphate-10 mM NaCl, pH 7.0.

portions, of the A or B enzymes could be found at the following N&l concentrations: 5 mM, 10 mM, 100 mM, and 500 mM or 10 mM NaCl containing 1 mM dithiothreitol.

Gel electrophoresis of denatured P-galactosidase A was carried out to detect the presence of protein subunits (Fig. 4). The enzyme was either denatured in guanidine, reduced, and carboxymethylated (26), or heated with sodium dodecyl sulfate and mercaptoethanol prior to electrophoresis (26). Migration relative to bromphenol blue was the same whether denaturation was accompanied by carboxymethylation (0.59 and 0.60 for two samples) or was not (0.60 and 0.61 for two samples). When Buffer II alone was carried through either of the denaturation protocols and subjected to electrophoresis, no staining of the gels was observed. Migration relative to polypeptides of known molecular weights (Fig. 6) corresponded to a molecular weight of 72,000. This is in good agreement with that found above by gel filtration of native P-galactosidase A and also that found previously for A activity in liver supernatant (lo), (mol wt 69,000 to 70,000). These results suggest that P-galactosidase A con- sists of a single polypeptide chain of molecular weight 72,000.

Comparison with Other @-Galactosidases-Previous work by us (10) and Ho et al. (11) has indicated the identity of 4-methylumbelliferyl fl-galactosidase A and GM r &galactosidase A. Our present results further confirm this finding, showing that both activities co-purify (Table II). In contrast, fl-galactosidase activities against Gal-Cer, Lac-Cer, and monogalactosyldipalmitin do not co-purify with G,r &galactosidase, indicating that these activities are due to other enzymes.

Gnrl @-Galadosidase Activity-Using the purified enzyme, activity was linear with respect to time and quantity of the enzyme protein over the ranges shown in Fig. 7. Galactose was the only radioactive product of the reaction (Fig. 8). Examination of the lipid products of the reaction by thin layer chromatography showed GM2 to be the only nonradioactive product.

Michaelis-Menten Constants-A study of the apparent Mi- chaelis-Menten parameters of purified /3-galactosidase A with 18 substrates indicated that the structural requirements for the aglycone are not stringent (Table III). However, the enzyme did not cleave lactulose, Gal-Hyl, or Gal-Hyl peptides.

Competition between G&-Ganglioside and .&Methylumbellijeryl /3-Gulactoside-The hydrolysis of an equimolar mixture of GM~- ganglioside and 4-methylumbelliferyl @-galactoside was studied over the range 0.0125 to 0.2 mM (individual substrate concentrations). Mixtures were assayed for the production of [aHlgalac- tose and 4-methylumbelliferone. The rate of formation of the


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

90

60

70 l- I ” 60 ci 3 2 50

$

Y 40

s I x

‘: 0 30

0.4 0.6 0.6 I .o MIGRATION OF PROTEIN

MIGRATION OF BROMOPHENOL BLUE

FIG. 6 (left). Comparison of the mobilities of denatured, reduced, and carboxymethylated fl-galactosidase A and molecular weight standards on polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. Five per cent acrylamide gels were used as described by Weber et al. (26). Standards (purchased from Sigma Chemical Co.) were: phosphorylase a from rabbit muscle; bovine serum albumin; ovalbumin (grade V), and bovine pancreatic cu-chymotrypsinogen A. Standards were heated in sodium dodecyl sulfate and mercaptoethanol prior to electrophoresis (26). About 10 pg of each were applied to a gel.

TABLE II

Specijk activities of B-galactosidases during puri$cation

Purification shown here is not the same one as that in Table I.

,%Galactoside substrates

Fraction

I I

4-metbyl- GMI-gan- umbelliferyl glioside

Jmwl/mg/min

Supernatant.......... 0.0084 0.0011 Concanavalin A-Seph-

arose............... 0.28 0.032 ECTEOLA-cellulose.. 39.5 6.4

-

monogalac- tosyldi-

palmitin

nmol/nzg/min

0.046 0.012 0.0046

1.15 0.31 0.064 13.3 0.458 0.66

sum of these products was analyzed as described in Table III and gave a Vm,, of 20.3 f 2.9 pmol per mg per min. The the- oretical value, using the treatment of Dixon and Webb (32) for competition of two substrates for the same enzyme and the data of Table III, is 22.8 pmol per mg per min. The close agreement of these two numbers strongly suggests that the same catalytic site(s) hydrolyze both substrates.

Glycosyltransjerase Activity-In order to determine whether the enzyme possessed transferase activity, as has been found for a bovine testis &galactosidase (8), glucose at 0.25 M was added to an enzyme preparation in the presence of p-nitrophenyl- P-galactoside (5 mM). The release of p-nitrophenol was stimulated by 147& Using similar conditions with bovine testis P-galactosldase, Jourdian and Distler (8) found a stimulation of 109%. However, when the natural substrate GM1-ganglioside was used, glucose inhibited rather than stimulated the reaction

0

IO

6

0 I 2 3 4 5 6 7

/3-GALACTOSI DASE PROTEIN (nanograms)

FIG. 7 (right). a, effect of time of incubation on galactose release from GM I-ganglioside by fl-galactosidase A. p-Galactosidase A, 0.43 ng, was incubated for the times shown in the assay system described under “Methods.” b, effect of enzyme protein incubated on galactose release from Gnl by @-galactosidase A. Incu- bations were for 15 min in the assay system described under “Methods.” All of the incubations were in duplicate.

14

12

IO

I CL 8 Q I

‘: 0 6 -

4

0 i

I I I I I I I I I

0 2 4 6 8 IO I2 14 I6

DISTANCE ON TLC PLATE (cm)

FIG. 8. Examination of the radioactive products of reaction of p-galactosidase A and Gxl-ganglioside. Gnl was incubated either with p-galactosidase A (123 ng) (0 ) or without (0 ) for 15 min in the standard GM~ &galactosidase A assay system. The reactions were then cooled in ice and 10 pl were applied to a cellulose thin layer chromatography plate (Eastman 6064) which was developed in n-propyl alcohol-water (7:3) and then radioscanned as described previously (10). Recovery of radioactivity from each sample exceeded 95%.


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7974

TABLE III TABLE IV

Michaelis-Menten parameters of @-galactosidase A substrates

Conformity of the kinetic data to the Michaelis-Menten equa- tion was first established by the method of Lineweaver and Burk (27). Values shown here were then calculated with the FOR- TRAN computer program of Cleland (28). All K, values are apparent K, values.

Effect, of substrate analogs and thiol reagents on the GMl-ganglioside p-galactosidase activity of @-galactosidase A

Control activity was that found for 6.7 ng of p-galactosidase A in the standard GM~ @-galactosidase assay system. When p-chloromercuribenzoate and mercuric chloride were included, the enzyme was diluted without EDTA. This gives 930/, of control activity in the absence of either thiol reagent.

Unless otherwise stated all compounds examined were present at a final concentration of 1.0 mM. GM,-ganglioside was present

Km zt S.E.

Chromogenic and fluorogenic p-galactosides

?lLM

4-methylumbelliferyl-. 0.27 f 0.01 4-methylumbelliferyl-a. 0.23 f 0.03 phenyl-. . 2.6 f 0.3 o-nitrophenyl- . 1.03 f 0.09 m-nitrophenyl- 1.26 f 0.09 p-nitrophenyl-. 1.24 f 0.15 1-naphthylb. > 0.85 2-naphthylb. > 0.85

4-methylumbelliferyl a-L-arabinoside 2.8 f 0.1

p-nitrophenyl-p-n-fucosidec.. 77.6 f 12.2 Gal p1+4 Glc (lactose) 7.8 & 0.4 Gal 01-14 GlcNAc (N-acetyllac-

tosamine). __.___..__..__..._. 3.5 f 0.3 Gal 0143 arabinose.. 7.7 f 0.6 Gal ~144 fructose (1actulose)d.. Gal pl Hyle. Gal pl Hyl peptidesf. GMl-ganglioside 0.077 f 0.00: Asialofetuing.. 0.23 f 0.02

vmx f S.E.

61.0 f 1.2 65.4 f 5.6 15.0 f 0.9 43.9 f 1.8 65.7 f 2.5 69.8 f 4.3

>38.0 >53.6

13.9 f 0.3 86.7 f 10.2 26.3 f 0.9

30.0 f 1.3 2.8 f 0.3

<O.l <O.l <O.l

8.6 f 0.6 41.7 f 1.4

a 4-methylumbelliferyl o-gal substrate using the same buffer system as for GM~ b-galactosidase. This buffer includes 14 mM sodium taurocholate.

* The solubility limit of these substrates was about 1 mM. Measurements were made up to 0.85 mM. The reaction was first order with respect to substrate concentration up to this concentration; it is assumed that this is because the K, is much higher than 0.85 mM.

c The relatively high standard error of these measurements is due to all data having been collected below the K, of this substrate. The highest substrate concentration which could be examined, due to limited solubility was 42 mM.

d Substrate concentration range examined was 0.3 to 20.0 mM. c Substrate concentration range examined was 0.06 to 2.0 mM. 1 Peptides were prepared by the collagenase and pronase di-

gestion of calf skin acid-soluble collagen (Calbiochem) by the method of Spiro (29). Glycopeptides were isolated and deglu- cosylated (30) by incubation with 0.1 N sulfuric acid at 100” for 28 hours. The substrate concentration range examined was 0.1 to 5.0 rnM.

Q K, and V,,, refer to galactose equivalents. It is assumed that there are 12.4 mol of galactose per mol of asialofetuin as reported by Spiro (31) for fetuin.

(Table IV). Analysis of the products of the reaction with /3- galactosidase A, [GaZ-aH]GM1, and 0.25 M glucose present revealed that galactose release was inhibited by 55yo of the level found in the absence of glucose. Approximately 10% of the radioactivity released appeared as compounds migrating slower than galactose. None of these compounds co-migrated with lactose and they had mobilities relative to lactose of 0.06, 0.67, and 1.35. Further characterization of these products was not attempted.

Inhibitors of GM1 /%Galactosida.se A-A number of galactose analogs, glycosides, and lipids inhibited the activity of GM1

at 0.2 mM.

Compound

D-Glucose

250 mM 25 mM

2.5 mM. _. _. _. n-Galactose.............. n-Galactitol.. n-Galacturonic acid. . n-Galactono-r-lactone.. Gal pl-+3 arabinose . Isopropylthio-&n-galac-

pyranoside. p-Aminophenylthio-fl-n-

galactopyranoside.. n-Gala&al (1,2 dideoxy-

n-lyxo-hex-l-enopyran- ose)....................

hntrol activity

%

54 87 97 48 96

100 27 82

60

16

37

Compound

Gal-Cer Glu-Cer Lac-Cer Ceramide. . . Lecithin. . . Oleic acid. Cholesterol. . . Gng-ganglioside. p-Chloromercuriben.

zoate (0.2 mM). Mercuric chloride

(0.2 rnM).

Dithiothreitol. ,

:ontrcJ ctivity

%

63 84 85 75 35

102 94 34

<l

16 91

fl-galactosidase A (Table IV). Inhibit,ion by lipids did not necessarily relate to a structural resemblance to the galactosyl moiety; L-a-lecithin and GMlz-ganglioside were the most potent lipid inhibitors.

Zmmunochemical Studies--In view of the simultaneous loss of GM1 P-galactosidase A and B activity in the GMl-gangliosidoses (lo), disorders which are apparently single gene mutations, it was of interest to determine whether the two enzymes were structurally similar. Double immunodiffusion employing p- galactosidase B and anti-P-galactosidase A antiserum under various conditions of temperature and antigen to antiserum ratio did not yield visible precipitin bands. However, we found that anti-P-galactosidase A antiserum quantitatively precipitated GM1 /%galactosidase B activity either as the partially purified enzyme or from liver supernatant (Table V). This implies that /3-galactosidases A and B share antigenic deter- minants. Gal-Cer &galactosidase, Lac-Cer &galactosidase, and neutral &glucosidase (nonspecific @-galactosidase) were not precipitated by anti-A antiserum, indicating that they are probably different enzymes.

DISCUSSION

Reports on the substrate specificities of P-galactosidases which act on synthetic substrates, disaccharides, and galactolipids have presented a confusing picture. In part, this may be due to the deductions made from the use of impure preparations. The preparation of highly purified /3-galactosidase A has allowed more precise characterization of the specificity of this enzyme.

The following physical and kinetic properties of the purified and crude &galactosidase A (10) showed no significant differ- ences: electrophoretic mobility on starch gel; apparent molecular weight on gel filtration; and apparent K, with either GMI-gan-


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7975

TABLE V

Immunoprecipitation of P-galactosidases and neutral @-glucosidase (“nonspecijk” P-galactosidase)

A volume of the enzyme source containing 0.05 pmol per min of total 4-methylumbelliferyl D-galactosidase, activity was made up to 0.15 ml with 10 mM sodium phosphate-10 mM sodium chloride, pH 7.0. Either 0.5 ml of anti-b-galactosidase A antiserum or 0.5 ml of control antiserum was added. The solution was left at 2” for 16 hours and then centrifuged at 50,000 X g for 30 min. The supernatants were then assayed for the activities shown as described under “Methods.” Figures give the percentage of enzymatic activity which was not precipitated after incubation with anti-p-galactosidase A antiserum compared to control antiserum.

&Galactosidase A. . 3.1 0.7 &Galactosidase B. . <0.5 0.7 Liver supernatantb. <0.5 13.7 112 86 88

a 4MU represents 4-methylumbelliferyl. * Supernatant prepared by Method B of Ref. 10.

glioside (0.077 mM in both cases) or 4-methylumbelliferyl fl- galactoside (0.25 mM crude versus 0.27 mM purified). Using these criteria, the properties of the enzyme appear to be un- altered during purification.

In addition to the hydrolysis of Ghll-ganglioside, P-galactosidase A possesses substantial activity against N-acetyllactosamine and asialofetuin, substrates which possess saccharide structures common to many glycoproteins (33). This suggests a major role for the enzyme in glycoprotein degradation. Ge- netic evidence supports this contention. In the GhlI-gangliosidoses, in which both Ghll @-galactosldases A and B are severely deficient (lo), oligosaccharides and glycopeptides, which have recently been characterized (34-37)) accumulate in addition to GM1 and G,l. The structure of these storage products re- sembles partially degraded skeletal keratosulfate (36) and partially degraded erythrocyte stromal glycoproteins (37). The oligosaccharide structure of asialofetuin is similar to a known major MN blood group active stromal glycoprotein (33) and appears to be a model compound for studying the action of p- galactosidase A on such substrates.

Jungalwala and Robins (12) isolated a @-galactosidase from rabbit brain which hydrolyzed o- and p-nitrophenyl @-galactoside and lactose but not Lac-Cer; possible activity with GhlI-ganglioside was not reported. Intestinal /3-galactosidases have been studied by Alpers (3) and Asp and Dahlqvist (4). A lysosomal fi-galactosidase was found which hydrolyzed nitrophenyl @-galactosides and lactose and was sensitive to p-chloromercuribenzoate (4); this seems similar to the enzyme studied here. A bovine testis P-galactosidase purified by Jourdian and Distler (8) hydrolyzed chromogenic /3-galactosides, lactose, GM1, and a wide range of other P-galactosides. This preparation differed from ours in that a substantial stimulation of activity by glucose was found with the p-nitrophenyl &galactoside substrate; our preparation possesses about a 20-fold higher specific activity than theirs as measured with p-nitrophenyl P-galactoside. The low level of transferase activity of the purified human liver enzyme con- trasts with the significant transferase activity of the enzyme from bovine testis.

The inactivity of P-galactosidase A with Gal-Hyl and Gal-Hyl peptides implies that this enzyme is not involved in the degradation of the carbohydrate-protein linkage region of collagen. To

our knowledge, no fl-galactosidase has been described which is active with this linkage region.

The present study does not exclude the possibility that GM1 /3-galactosidase A possesses some Lac-Cer &galact,osidase activity; nonetheless this activity is at most 0.4% of that with GM1- ganglioside (Table II). Because the bulk of the two activities neither co-purify (Table II) nor co-precipitate with anti-& galactosidase A antiserum (Table V), they are probably the properties of two different enzymes. Gal-Cer P-galactosidase behaved similarly. The finding that all three P-galactosidases bind specifically to Con A implies that they are glycoproteins (21), as found for a number of lysosomal hydrolases (38, 39). In view of this, our inability to detect carbohydrate in the purified enzyme on polyacrylamide gels with Alcian blue staining suggests that the method used was too insensitive.

The failure of Suzuki and Suzuki (40) to find a stimulation of GM1 fi-galactosidase activity by ionic detergents is surprising and is in contrast to the findings of Ho et al. (11) and our own previous report (10). We reported previously at least a 100.fold stimulation of GM1 P-galactosidase activity by anionic and cationic detergents. The specific activities Suzuki and Suzuki report for the enzyme are consequently more than 2 orders of magnitude lower. For this reason, their conclusions relating to the properties of GNIl P-galactosidase should be treated cau- tiously. Miyatake and Suzuki (41), however, have reported that sodium taurocholate was required for the activity of rat brain GM1 /3-galactosidase.

The finding that fi-galactosidase A hydrolyzed 4-methylumbelliferyl a-L-arabinoside and p-nitrophenyl-P-n-fucoside in addition to @-n-galactosides is similar to that of the Escherichiu coli fl-galactosidase (1). Substitution of Cc or the C& hydroxyl group of galactose by hydrogen does not prevent hydrolysis, although the apparent K, increases. The recent reports that both 4-methylumbelliferyl fi-n-fucosidase (42) and cr-L-arabinosidase activities (16) are severely deficient in the GMl-gangliosidoses are in keeping with these properties of the enzyme. It has recently been reported that the only form of fl-n-fucosidase that could be detected in pig kidney hydrolyzed p-nitrophenyl- @-n-galactoside but not lactose (43).

Of the eight lipids examined, every one, except oleic acid, inhibited GM1 &galactosidase activity. Inhibition occurred despite minimal structural resemblance of some of the lipids examined (lecithin, cholesterol, ceramide) to GMll-ganglioside. The simplest explanation is that these lipids change the structure of the micelles which are critical for optimal activity (10). Some of the effects of inhibitors reported by Ho et al. (11) are rather different to those found here, but because these workers used a different assay system and cruder preparations of enzymes, a comparison with their work is difficult.

The molecular weight of the native enzyme was very similar to that found under the conditions which strongly favor subunit dissociation. This result suggests that there is only a single polypeptide chain in @-galactosidase A.

A structural relationship between P-galactosidases A and B is suggested by the following findings. Both enzyme activities are concurrently absent in the GMI-gangliosidoses (10). Both forms hydrolyze 4-methylumbelliferyl fi-galactoside, GMI-ganglioside, and asialofetuim3 fi-Galactosidase B is quantitatively precipitated by anti-P-galactosidase A antiserum. The failure to find precipitin bands with the B enzyme in gel diffusion experiments might be accounted for by severely restricted diffusion of the B enzyme, reflecting its high molecular weight.

Many of the pleiotropic effects of the mutation(s) of the GMl-


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

7976

gangliosidoses are clarified by the substrate specificity of @- galactosidase A. The accumulation of gangliosides, oligosaccharides derived from glycoproteins, and glycopeptides is to be expected. Lac-Cer &galactosidase and Gal-Cer P-galactosidase are not expected to be primary enzyme deficiencies in these disorders. Suzuki and Suzuki have, however, recently reported a deficiency of Lac-Cer /3-galactosidase activity in the GMM-gangliosidoses (44). In contrast, Wenger et al. (45) have found normal levels of Lac-Cer /3-galactosidase in GM I-gangliosidosis type I and have recently confirmed this finding in a further six GM1-gangliosidosis livers.4 In human cerebral gray matter, Brady et al. (46) have reported increased Lac-Cer /!-galactosidase. As expected from the negligible hydrolysis of Gal-Cer by @-galactosldase A, Gal-Cer /3-galactosidase activity has been found to be normal or elevated in the GM1-gangliosidoses (44,45).

The antigenic cross-reactivity of P-galactosidases A and B suggests that they possess polypeptide sequences in common and is consistent with the deficiency of both of these enzymes in the GMl-gangliosidoses. Using the antibody prepared here, it will be of interest to see whether there is enzymatically inactive mate- rial in these disorders which retains cross-reactivity with fi- galactosidase A.

Acknowledgments-We would like to thank Drs. Audre Lo- Verde, Arnold Miller, Russell Frost, and Jack Alhadeff for their stimulating advice throughout this work. Ms. Wendy Sawyer worked with us in the initial stages of this project. Haynes W. Sheppard Jr. helped to immunize and bleed rabbits. Drs. W. M. Watkins, L. W. Cunningham, and D. A. Wenger are thanked for their generous gifts of substrates.

REFERENCES

1. WALLENFELS, K., AND MALHOTRA, 0. P. (1960) in The Enzymes (BOYER, P. D., LARDY, H., AND MYRBACK, K., eds) Vol. 4, pp. 409-430, Academic Press, New York

2. WALLENFELS, K., AND WEIL, R. (1972) in The Enzymes (BOYER, P. D., ed) .Vol. 7, pp. 617-663, Academic Press, New York

3. ALPERS. D. H. (1969) J. Biol. Chem. 241, 1238-1246 4. ASP, N:-G., AND DA~LQVIST, A. (1972) Anal. Biochem. 47,

527-538 5. GATT, S. (1967) Biochim. Biophys. Acta 137, 192-195 6. HAJRA, A. K., BOWEN, D. M., KISHIMOTO, Y., AND RADIN,

N. S. (1966) J. Lipid Res. 7, 379-386 7. ARONSON, N. N. (1972) in The Glycoproteins (GOTTSCHALK, A.,

ed) Part B, pp. 1211-1227, Elsevier Publishing Co., Amster-

12. JUNGALWALA, F. B., AND ROBINS, E. (1968) J. Biol. Chem. 243.4258-4266

13. MEISLER, M. (1972) Methods Enzymol. 28,820-824 14. SLOAN, H. R. (1972) Methods Enzymol. 28, 868-874 15. O’BRIEN, J. S. (1972) in The Metabolic Basis sf Inherited Dis-

ease (STANBURY, J. B., WYNGAARDEN, J. B., AND FREDERICK- SON, D. S.. eds) DD. 639-662. McGraw-Hill. New York

16. HOOF; F. V.‘(1973)-<n Lysosomes and Storage’Disease (HERS, H. G., AND HOOF, F. V., eds) pp. 305-321, Academic Press, New York

17. O’BRIEN, J. S., AND SAMPSON, E. L. (1965) J. Lipid Res. 6, 537-544

18. WENGER, D. A., OKADA, S., AND O’BRIEN, J. S. (1972) Arch. Biochem. Biophys. 163, 116-129

19. LOWRY, 0. H.,.R~SEBROUGH, N. J., FARR, A. L., AND RAN- DALL. R. J. (1951) J. Biol. Chem. 193. 265-275

20. Ho, MI W., S&K, j., SCHMIDT, D., VEA~H, M. L., JOHNSON, W., BRADY, R. O., AND O’BRIEN, J. S. (1972) Amer. J. Hum. Genet. 24, 37-45

21. NORDEN, A. G. W., AND O’BRIEN, J. S. (1974) Biochem. Bio- ph y/s. Res. Commun. 66, 193-198

22. WENGER, D. A. (1974) Clin. Chim. Acta, in press 23. WILLIAMS, D. E., AND REISFELD, R. A. (1964) Ann. N. Y. Acad.

Sci. 122, 373-381 24. SVENNERHOLM, L. (1956) J. Neurochem. 1,42-53 25. WARDI. A. H.. AND MICHOS. G. A. (1972) Anal. Biochem. 49.

607-609 ' \ I I

26. WEBER, K., PRINGLE, J. R., AND OSBORN, M. (1972) Methods Enzymol. 26, 3-27

27. LINEWEAVER, H., AND BURK, D. (1934) J. Amer. Chem. Sot. 66, 658-666

28. CLELAND, W. W. (1967) Advan. Enzymol. 29, l-32 29. SPIRO, R. G. (1967) J. Biol. Chem. 242, 1923-1932 30. SPIRO, R. G. (1967) J. Biol. Chem. 242, 4813-4823 31. SPIRO, R. G. (1960) J. Biol. Chem. 236, 2860-2869 32. DIXON, M., AND WEBB, E. C. (1964) in Enzymes pp. 84-87,

Academic Press, New York 33. 34. 35.

36.

37.

38. 39.

40.

41.

SPIRO, R. G. (1973) Advan. Protein Chem. 27, 349-467 SUZUKI. K. (1968) Science 169. 1471-1472 BRUNN~RAB~R, G. G., BERRA; B., AND ZAMBOTTI, V. (1973)

Fed. Eur. Biochem. Sot. Lett. 34. 350-352 TSAY, G. C., AND DAWSON, G. (1973) Biochem. Biophys. Res.

Commun. 62, 759-766 WOLFE, L. S., SENIOR, R. G., AND NG YING KIN, N. M. K.

(1974) J. Biol. Chem. 249. 1828-1838 T&usT&, 0. (1973) Mol. C&. Biochem. 2, 169-177 SRIVASTAVA, S. K., YOSHIDA, A., AWASTHI, Y. C., AND BEUT-

LER, E. (1974) J. Biol. Chem. 249.2049-2053 SUZUKI, Y., AND SUZUKI, K. (1974) J. Biol. Chem. 249, 2098-

2104 MIYATAKE, T., AND SUZUKI, K. (1974) Biochim. Biophys.

Acta 337. 333-342 dam

8. DISTLER, J. J., AND JOURDIAN, G. W. (1973) J. Biol. Chem. 42. HINDMAN, -J., AND COTLIER, E. (1972) Clin. Chem. 18, 971-

975 248,6772-6780

9. SVENNERHOLM, L. (1963) J. Neurochem. 10, 613-623 43. WIEDERSCHAIN, G. Y.-A., AND PROKOPENKOV, A. A. (1973)

10. NORDEN, A. G. W., AND O’BRIEN, J. S. (1973) Arch. Biochem. Arch. Biochem. Biophys. 168, 539-543

Biophys . 169, 383-392 44. SUZUKI, Y., AND SUZUKI, K. (1974) J. Biol. Chem. 249, 2113-

2117 11. Ho, M. W., CHEETHAM, P., AND ROBINSON, D. (1973) Biochem.

J. 136, 351-359 45. WENGER, D. A., SATTLER, M., AND HIATT, W. (1974) Proc.

Nat. Acad. Sci. U. S. A. 71, 854-857

4 Dr. D. A. Wenger, personal communication. 46. BRADY, R. O., O’BRIEN, J. S., BRADLEY, R. M., AND GAL, A. E.

(1970) Biochem. Biophys. Acta 210, 193-195


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

Anthony G. W. Norden, Linda L. Tennant and John S. O'BrienENZYME FROM HUMAN LIVER

-Galactosidase A : PURIFICATION AND STUDIES OF THEβ Ganglioside m1G

1974, 249:7969-7976.J. Biol. Chem.

http://www.jbc.org/content/249/24/7969Access the most updated version of this article at

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here

http://www.jbc.org/content/249/24/7969.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/content/249/24/7969

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;249/24/7969&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/249/24/7969

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=249/24/7969&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/249/24/7969

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/content/249/24/7969.full.html#ref-list-1

http://www.jbc.org/

gm1 ganglioside @-galactosidase a · gm1 ganglioside @-galactosidase a purification and studies of...

Documents