Biosynthesis in Vitro of a Blood Group B-active Fucose-containing Hexaglycosylceramide from Neolactopentaosylceramide in Bovine Spleen*

A solubilized a-fucosyltransferase activity has been isolated from a bovine spleen Golgi-rich membrane fraction. The enzyme transfers fucose from GDP-P-L-fucose to a blood group B-active pentaglycosylceramide acceptor (Gal(al-3)Gal(/?l-4)GlcNAc(/31-3)Gal(~l- 4)Glc-ceramide) isolated from rabbit erythrocytes. the membranes with 0.2% (final sodium taurodeoxycholate detergent produced maximal recovery (90%) of activity. A cationic detergent, G-3634-A, is required for optimal activity and the enzyme does not require addition of exogenous metal ion for activation. The purified 14C-labeled product of the reaction migrated with human blood group B-active hexaglycosylceramide on Silica Gel G thin layer plates.

The glycoconjugate constituents of animal cell membranes with ABO(H) and Lewis blood group specificities are characterized by the presence of fucose residues as nonreducing terminals. A family of enzymes called fucosyltransferases transfer L-fucose residues from GDP-L-fucose into specific glycoconjugate acceptor molecules. Blood group glycoprotein: fucosyltransferases have been detected in milk (l), gastrointestinal mucosa (2)(3)(4), submaxillary glands (3), human serum (5, 6), and bone marrow (7). However, very little work has been published on the biosynthesis in vitro of blood groupactive glycosphingolipids.

Lactotriaosylceramide(GlcNAc(/31-3)Gal(/3l-4)Glc-Cer;
* This work was supported by United States Public Health Service Grants NS-09541 and CA-14764. This is the 10th paper in a series dealing with biosynthesis in vitro of blood group-related glycosphingolipids. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Part of this work was submitted to the University of Notre Dame in partial fulfiUment of the requirements for the degree of Doctor of Philosophy in 1979. Present address, Department of Biochemistry, University of Iowa College of Medicine, Iowa City, IA 52242.
$ Recipient of a grant-in-aid from Miles Laboratories (Elkhart, IN). To whom correspondence should be addressed.

GlcNAc(/31-3)Gal(/31-4)Glc-Cer) using Golgi-rich membrane
fraction has been achieved previously (14). Employing human serum as the source of fucosyltransferase, the biosynthesis of an H-active glycolipid has also been achieved (15). Fucosyltransferases involved in the biosynthesis in vitro of types H and B glycolipids have been detected in a membrane preparation isolated from a human neuroblastoma-derived clonal cell line, IMR-32 (16). However, structures of these biosynthesized products have yet to be determined. Treatment of a bovine spleen Golgi-rich membrane (14) fraction ( Fig. 1) with 0.2% sodium taurodeoxycholate produced maximum recovery of a-L-fucosyltransferase activity in the 100,000 X g supernatant fraction. A 10-fold increase in the specific activity of the enzyme was obtained relative to the total homogenate. Endogenous acceptor activity of the system was almost completely eliminated (1-2 nmol/mg of protein in 45 min) after solubilization with taurodeoxycholate.
The radioactive hexaglycosylceramide was isolated from a 20-fold increased incubation mixture using GDP-P-L-['~C]~~cose (2.3 X lo6 cpm/pmol) and nLcOsesCer (0.5 pmol) as substrates in the a-fucosyltransferase-catalyzed reaction. After incubation at 37 "C for 4 h, the mixture was spotted on Whatman No. 3 " paper and labeled products were resolved by descending chromatography in 1% borate buffer as described under "Enzyme Assay."2 The areas of the chromatograms containing radioactive material were eluted with chloroform/methanol/water (60:30:6, v/v/v) and the extract (140,oOO cpm) was applied to a 5-g Unisil column (1.5 X 5 cm). The radioactive product was eluted from the Unisil column as a single peak with chloroform/methanol (3:7, v/v; 93% recovery). The purified product migrated with standard human blood group B-active hexaglycosylceramide on Silica Gel G thin layer plates (Fig. 4, Lane I ). In a separate experiment, Substrate competition experiments (14) suggest that the biosyntheses in vitro of the Hand B-active glycosphingolipids are probably catalyzed by the same enzyme. The activity of glycolipid:(al-2)fucosyltransferase, along with other studies with oligosaccharides, led to the conclusion that the structural determinants of oligoglycosyl substrates reside in the N-acetylglucosamine moiety. It is therefore not surprising that the activities with both substrates (nLcOse4Cer and nLcOse5Cer) are almost identical. These results suggest an alternative pathway (14, 49) to the accepted biosynthetic schemes for blood group substances (oligosaccharides, glycolipids, or glycoproteins), and it is proposed here, along with our previous studies (8,13, E), that an (al-2)fucosyl unit is not essential for further chain elongation. However, the presence of a fucosyl unit conferring H specificity on the carbohydrate chains of the precursor substance necessary for the functioning of blood group A or B gene-controlled glycosyltransferases (50, 51) cannot be ruled out at the present time.
The solubilized fucosyltransferase is similar to the ody other solubilized mammalian fucosyltransferases studied thus far, from mouse brain (28) and HeLa cells (52), in that addition of exogenous metal ion is not required for enzyme activation ( Table I). The human milk cy fucosyltransferases (27, 53) purified by affinity chromatography on GDP-hexanolamineagarose likewise do not have metal ion requirement. The bovine spleen enzyme transfers fucose to terminal galactose residues attached to penultimate N-acetylglucosamine, whereas the brain and HeLa cell enzymes require penultimate N-acetylgalactosamine residues. However, the a-fucosyltransferase activities isolated from bovine spleen and rat small intestinal mucosa (29) both are inhibited competitively by GDP, the product of the reaction. In addition to GDP, GMP and GTP also inhibit these enzyme activities. Whether this inhibition is due to the attachment of any specific groups of the guanosine moiety at the GDP-fucose binding site has not been established unambiguously. The radioactive hexaglycosylceramide product of the reaction is most probably a mixture of Gal-[14C-Fuc]Gal-GlcNAc-Gal-Glc-Cer and Gal-Gal-[14C-F~~]GlcNAc-Gal-Glc-Cer, since two bands of similar RF values (Fig. 4) have been observed on a radioautogram. The first reaction product co-migrated with human B-type glycosphingolipid, and the second product migrated in the region of fucose-containing hexaglycosylceramides, but slightly below the B-active glycolipid, resembling the chromatographic mobility of glycolipids in which fucosyl residues are attached to N-acetylglucosamine (45). Treatment of the 14C-labeled product with a-L-fucosidases such as Charonia Eampas and Venus mercenaria cleaved 40% and 80%, respectively. However, reaction conditions might have prevented complete hydrolysis.
Lactosylceramide is a poor acceptor with this enzyme preparation. Perhaps the taurodeoxycholate extract contains another specific fucosyltransferase that catalyzes the transfer of L-fucose to either position C-3, C-4, or C-6 of the N-acetylglucosamine residue.
The presence of glycolipid:(al-3)fucosyltransferase activity in human serum has been reported by Pacuszka and Koscielak (15). The occurrence of two specific g1ycoprotein:fucosyltransferases in human serum and in rat liver microsomal membranes has been investigated by Schachter and his co-workers (6, 54, 55). Since assay conditions can be better controlled with a solubilized enzyme system, our present report should be useful for further purification of specific fucosyltransferases and studies of the biosynthetic mechanisms of various blood group-active glycosphingolipids. Recently, Hill and co-workers (56, 57) reported the purification and substrate specificity of an H blood group pgalactoside a-fucosyltransferase from porcine submaxillary glands. Unlike bovine spleen (cy1-2)fucosyltransferase, this porcine a-fucosyltransferase shows a preference for acceptors with nonreducing terminal galactose in a Pl+ 3 linkage to an N-acetylhexosamine (GlcNAc or GalNAc). However, these differences in the substrate specificity between the two (al-2)fucosyltransferases, isolated from two different species, might be due to the presence of two different proteins. The followlng materials were Obtalned as gift samples: H-cyst glycopra-prepared according to the method of Hakomori 1421 with the use of a plastic template. well 1 contamed nonradioactive rabbit erythrocyte nkOsegCer lGallol-3~GalI81-4~GlcNAcI~1-3~GallE1-4~Glc-Cer;100u91l0 "11; well 2 contained 14C-labeled enzymatic Product 13000 CPmIi well 3 con-Microimmunodiffusion Reaction -A double microdiffusion plate was taind 14C-labeled enzymatic product 13000 cpml and h u k n B-active hexaglycosylceranide lGalll-3~[Fuclal-2~1-Gal-GlcNRc-Gal-Glc-C~~: 100 ug1100 "11 and well a contamed 14C-labeled enzymatic product and rabbLt erythrocyte nkOse5Cer. The blood group B-specific lectin isolated from Euonyms enropeus seeds by the method of Kabat 1431 was placed to diffuse for 5 days at 4-6'. The precipitin lines were visualized in the center well 130 ug Of protein110 u l l . The plate was allowed with 0.2% Amido black [ Fig. 5A1, followed by extensive washing with 2% acetic acid, and drled at 37' in an incubator. The stained plate was exposed to an X-ray fzlm for 6 months (Fig. 5 Table I) did not change the rate Of reaction. indicatmg that addition of exogenous metal ion IS not requued for activation of the Solubilized Ol-f"cOBylt=an.feTa8e.

Biosynthesis in
In Contrast, the aetzvities of the Crude homogenate [ Fig. 1) or the particulate system IBSWGS) were somewhat dependent on the Concentration of exqenous Mg2+. These results are in agreement with those Of mouse bran mlcroeomal 1281 and rat small-intestinal 1291 fucosyltrmsferases Solubxlired with Triton X-100. The O-L-fUC08yltTansfe=aBe activity of the particulate system was also inhibited by purine nucleotides (Fig. 21. Of the nucleotides tested, GTP I1 m M l was the m s t effective inhLbrtor of the enzyme 10.6% activity relative to the Controll. In addition, GMP I1 m) and GDP I1 mM1 were very effective inhibitors 11.7% and 7% relative actinty, respectively).  Table 111. The most active acceptors llactosamlne and dealallzed ol-glycoprotein) contamed penultimate N-acetylglucosamine residues. If a fucose residue LS already present on the N-acetylglucosamine molety. as Is the case for lacto-N-fucopentaose 111, then perhaps the inner fucose residue hinders the transfer Of an additional fucose to the termlnal galactose 114). In addition, desiallzed fetuin lGall81-4lGlcNAc-Rl accepted fucose I" the presence of thls enzyme preparatlon. H-cyst glycoprotein was the least actLYe acceptor, as expected. ~l n c e fucose 2s already present on the galactose moiety. Llttle 114CIfucose was Incorporated Into lacto-N-bloSe I1 and IaCtO-N-tr~OSe I1 In comparison with lactosamlne.
Olzgosaccharrde Acce tor Speciflclty -The inCorPOZatlOn Of The double precipltln 13n. 6 Present may have been due to 8 . e u r~ eus G e l P-10 colum by Petryniak. Pereira, and Kabat 143).

Biosynthesis in
Vitro of Blood Group BI-active Glycolipid The complete incubatlon mixture contained the following co-nents Table I11 Oligosaccharide acceptor Bpeoificity studies with the bovine apleen Colgitaurcdeoxysholate supernatant Conditions were the same atho-described in Table I