Enzymatic synthesis of a blood group B-related pentaglycosylceramide by an alpha-galactosyltransferase from rabbit bone marrow.

Abstract An α-galactosyltransferase that catalyzes the biosynthesis in vitro of blood group B-specific pentaglycosylceramide from UDP-[14C]galactose and O-β-galactosyl (1→3)-O-β-(N-acetyl) glucosaminosyl (1←3)-O-β-galactosyl (1→4)-O-β-glucosyl (1 →1) ceramide was detected in homogenates of adult rabbit bone marrow. The enzyme was present in the supernatant solution obtained after homogenization of rabbit bone marrow in the presence of 0.6% Triton X-100 and centrifugation at 100,000 x g. The enzyme was precipitated from the supernatant solution at 0 to 45% saturation with ammonium sulfate. The 14C product of the reaction was isolated, purified, and analyzed for its constituents. It inhibited the hemagglutination reaction of human B-type cells and rabbit erythrocytes against human anti-B serum. The terminal [14C]galactose was cleaved 85% by the action of fig α-galactosidase. The Km values were 1.67 x 10-3 m and 1.43 x 10-4 m for the tetraglycosylceramide (LacTet-cer) and UDP-galactose, respectively.

The enzyme was present in the supernatant solution obtained after homogenization of rabbit bone marrow in the presence of 0.6% Triton X-100 and centrifugation at 100,000 X g. The enzyme was precipitated from the supernatant solution at 0 to 45% saturation with ammonium sulfate.
The 14C product of the reaction was isolated, purified, and analyzed for its constituents. It inhibited the hemagglutination reaction of human B-type cells and rabbit erythrocytes against human anti-B serum.
The terminal [14C]galactose was cleaved 85 % by the action of fig or-galactosidase.
The K, values were 1.67 X lop3 M and 1.43 X 10e4 M for the tetraglycosylceramide (LacTet-cer) and UDPgalactose, respectively.
The chemical structure of blood group-specific glycoprot~eins isolat.cd from cyst fluid and stomach mucosal linings has been elucidat,ed by Morgan (2), Watkins (3), and Kabat (4). On the other hand, tho blood group-specific substances isolated from erythrocyte membranes appeared to be glycosphingolipids, and the structures of those isoantigens were described by Yamakan-a (5), Hakomori (6), and Kosciclak et al. (7). The specificity of blood group substances is determined by the specific oligosaccharide chains attached to the glycoprotcin or ceramide moiet,y. Eto el al.  isolated from rabbit crythrocytcs. The structures of the Le"-and Leb-specific glycosphingolipids isolat,cd from human adenocarcinoma tissues were described by Hakomori et al. (9,10). We have reported previously (I 1, 12) the biosynthesis in 1 3 vitro of the core tctraglycosylceramidc (Gal2 GlcSAc -Gal2 Glc 2 ceramide) from a triglycosylceramide cow taining N-acetylglucosamine.
The present studies are corn cerned n-ith the transfer of galactose to the tetraglycosylccramide t,o form a blood group-specific pentaglycosylceramide. The reaction, catalyzed by an oc-galactosyltransferase also obt,ained from rabbit bone marrow, is as follows. Corporation) ; hexokinase, ga,lactose dehydrogenasr, and glucose-6-P dehydrogenasc (Sigma) ; erythrocytcs and bone marrow of mature rabbit (Pel-Freez Hiologicals, Rogers, Arkansas); human anti-A, and anti-13 blood grouping serum (Hyland Division Travcnol Laboratories, Inc., Costa Mesa, California, and Lederle Laboratories Div.). Glycosphingolipid acceptor was isolated from erythrocytcs by our previously published met'hod (12). Other substances were either commercial preparat,ions or gifts.

Methods
Analytical Procedwcs-Protein was det8ermincd by the method of Lowry et al. (13). The 1% product of the enzymatic reaction and the acceptor glycosphingolipid were methanolyzed by the procedure of Kishimoto and Radin (14). Sphingosine was determined by t,he spcctrophotometric method of Lautcr and Trams lTO0 (15) ; glucose, by a coupled enzymatic reaction with the use of hexokinase and glucose-6-P dehydrogenase (16); and galactose, by the galactosc dehydrogenase procedure (17). Glucosamine was quantitatively determined by a modified  and subscqucntly identified by gas-liquid chromatography (19) on a Packard model 7300 instrument.

Identification oj Sugar Al&to1
Acetates by Gas-Liquid Claromatograplzy-Glycosphingolipids (100 pg to 2 mg) were dissolved in 1 ml of a ethanol-water-concentrated HCl mixture (1: 1: 1, v/v/v)' and heated in a special hydrolysis t,ube (14) for 4 hours at 100". The mixt,ure was dried under Nz and rehydrolyzed with 1 N HCl for 1 hour at 100". The hydrolysate was extracted with 5 volumes of ether and the lower, aqueous layer was dried under nitrogen.
The residue was dissolved in 2.0 ml of 1 N NI&OI-I and reduced by the addition of 20 mg of sodium borohydridc for 1 hour at room temperature (20). The mixture was extracted with 5 volumes of ethyl acet.ate and the pH of the lower, aqueous layer was adjusted to 3.0 with glacial acetic acid. The water layer was evaporated to dryness under nitrogen six times with 5 ml of methanol and finally dried over P205 and KOH in a vacuum desiccator for 12 to 16 hours. Then 1 ml of acetic anhydride was added to the residue, which was heated at 120" for 3 hours; excess acetic anhydride was removed under nitrogen. The mixture of alditol acetat.es was partitioned in the lower, chloroform layer (chloroform-water, 2: 1, v/v) anddried under nitrogen. The residue was dissolved in 50 t.o 100 ~1 of analytical grade acetone or chloroform; 0.5 to 1.0 ~1 of the solution was injected into a coiled glass column (4 feet x 2 mm) of 2.7% ECNSS-M coated on Gas-ChromZ and developed at' 180" for 15 min, then programmed up t'o 210" for 16 min at a rate of temperature increase of 15" per min. By this method the molar ratio of glucitol to galactitol (1:2) was determined in the substrate tetraglycosylceramide.
Preparation of Enzy?rre-The oc-galactoayltransferase was detected in homogenates of rabbit bone marrow.
All steps in the preparation of enzyme were conducted between 0" and 5". Rabbit bone marrow (20 to 25 g, purchased frozen) was homogenized with 3 volumes of 0.32 M sucrose containing 0.001 M EDTA and 0.014 N mercaptoethanol, at $1 7.0, in a Potter-Elvehjem homogenizer with a loosely fit.ting Teflon pestle. The homogenate was filtcrcd through a double layer of cheesecloth to remove fat and t,hcn was centrifuged in a Spinco L2-6513 centrifuge for 90 min at 100,000 X g. 'I'hc pellet was rehomogenized with 2 volumes of sucrose-EDTA-mcrcaptoethanol containing Triton X-100 (6 mg per ml) and then was centrifuged for 2 hours at 100,000 X g. The enzyme activity was precipitated by 0 to 45$& 1 J-L. Chien and S. Besn, unpublished experiments.  (13)). The mixtures were incubated for 1 hour at 37" and the reaction was st,opped by adding 1.0 pmolc of EDTA (pf1 7.0). The incubation mixture was applied ont,o Whatman T\l'o. 31111 paper and subjected to high voltage electrophoresis in 1.057; sodium tetraborate buffer, pH 9.1, at 60 I-olts per cm for 40 to 45 min. Under these conditions, the radioactive glycosphingo~ lipid product remained at the origin of the clec:tl,oplloretogram, while the free [%]galactose, [14C]galact'ose-l-I', and UDP-[14C]-galacOose migrated 10 to 20 cm toward t,he anode. After the paper had dried, the cathode portion of the elcctrophoretogram was developed in t'he ascendin, m fashion with chloroform-methanol-water (60 : 35 : 8). The 14C product migrated 6 t,o 9 cm from the origin.
The appropriate areas of each of the chromatograms were quantitatively determined by liquid scintillnt,ion techniques with a Packard scintillation counter, model 3375.

Requiremenk
for Enzyme dctiuhy-As shown in Table I  Substitution of cithcr Co*+, Cd"+, %II", or IV?+ for 1\In2f rcduccd the rate to 15 to 25c/1, and c'u"", I"+', Ca2+, 01'   UDP-glucose, UDP-N-acetylglucosamine, or UDP-A-acetylgalactosamine ( <3y0). The rate of the reaction was changed no more than 5% by the addition of nonradioactive galactose, which indicated that free galactose was not incorporated into the pentaglycosylceramide. The effect of varying the concentration of UDP-[Wlgalactose on the rate of the reaction is shown in Fig. 1. The calculated K', value for UDPm['4C]galactose is 0.14 InM.
Effect of Sub&ate Concentration and Substrate Specijicities-The effect of varying the conccnt,ration of tetraglycosylceramide (LacTet-cer) on the rate of the reaction is shown in Fig. 2. The calculated K, value is 1.67 IIIM.
The resu1t.s of studies with different potential lipid acceptors are shown in Table II. Asialoganglioside, Tay-Sachs ganglioside (GM2),3 lactosylceramide, and ceramide (containing 2-hydroxystearic acid) were completely inactive under these assay conditions.
The significance of the high activity of the LacTri-ceramide is discussed in the following section. The acceptor specificity of the cr-galactosyltransferase was test.ed with a number of different oligosaccharides.
2'.Fucosyllactose and lacto-N-fucopentaoses I and II were poor acceptors.
The significance of these results is considered under "Discussion." Substrate Competition-We previously reported the presence of a /%galactosyltransferase activity in rabbit bone marrow (12) that catalyzes the transfer of galactose from IJDP-galactose t.o the triglycosylceramide (G~cNAc ---) Gal + Glc 4 ceramide) to form tetraglycosylceramide.
It remains to be determined whether the ol-galactosyltransferase activity which catalyzes the transfer of galactose from UDP-galactose to the tetraglycosylceramide (Gal -+ GlcNAc + Gal + Glc 4 ceramide) is the result of a single t'ransferase with dual specificities or is due to the presence of two different galactosyltransferases.
To obtain definite information on this quest,ion, competition experiments were conducted as shown in Table III Action of Galacfosidases--'4C products of the two reactions (LacTri-cer and LacTet-cer) were isolated, purified, and t'reated with fi-and Lu-galactosidases as shown in Table IV   in methanol for 1 hour and t'hen dialyzed overnight against directly for r4C! and for its constituents after methanolysjs, distilled water.
The mixtures were heated for 10 to 15 min at 60" and cooled to room temperature before addition of antiserum and cells. In the absence of any glycosphingolipid the positive hemagglutination reaction was tested upon addition of 10 ~1 of 6% cell suspension (human B-type erythrocytes or rabbit erythrocytes) to 5 ~1 of anti-B human serum (purchased from Hyland Laboratories, Costa Mesa, Calif.). Inhibition of the hemagglutination reaction was not observed with human 1Mype red cells until the concentration of nonradioactive pentaglycosylceramide (or, 14C product) reached 3 pg/O.O25 ml level (Table V). The hemagglutination inhibition reaction was also observed with rabbit erythrocytes in the presence of nonradioactive pentaglycosylceramide (minimum concentration 5 pg/O.O25 ml) and W product (6 pg/O.O25 ml). IV Action of (Y-and P-galactosidases on 14C products For treatment with fig a-galactosidase, incubation mixtures contained t,he following components in final volumes of 0.06 ml: '"C products, 0.5 to 0.6 nmole; sodium taurocholate, 25 pg; and 0.2 unit of a-galactosidase (0.025 pmole of p-nitrophenyl-a-ngalactopyranoside was hydrolyzed per unit of enzyme per 15 min). After 4 hours at 37", t,he incubation mixtures were assayed by high voltage paper electrophoresis as described in the text. The [Wlgalactose migrated from the origin 10 to 12 cm. The appropriate areas of each of the chromatograms were quantitatively determined by liquid scintillation techniques. For keatment wit,h testicular o-galactosidase, incubation mixtures contained the following components in final volumes of 0.05 ml: 1% products, 0.2 to 0.5 nmole; Triton X-100, 500 pg; citrate-phosphate buffer, pH 4.3, 5.0 pmoles: and 2.7 units of enzyme (1 unit of the enzyme hydrolyzed 1.0 nmole of p-nitrophenyl-p-n-galactopyranoside per min; specific activity, 3390 units per mg of protein). After 4 hours at 37", the incubation mixtures were assayed by the method described above.
Further studies of the structure of the enzymatic product are in progress.
Recently Hearn et al. (34) have reported the isolation of a UDP-galactose : glycoprotein oc-galactosyltransferase from ovarian cyst fluid which transfers galactose only to oligosaccharides in which L-fucose is attached to the terminal fl-galactose by an (~-1 + 2 linkage. The present studies show (Table II) that the most active acceptors are those cont.aining a P-galactopyranosyl residue at the nonreducing end without n-fucose attached to it, such as N-acetyllactosamine, la&o-A-tetraose, lactose, and Each reaction is catalyzed by a specific glycosyltransferase and requires the corresponding sugar nucleotide.