Biosynthesis of Lewis fucolipid antigens in human colorectal carcinoma cells

Purified glycolipids were tested for their ability to serve as acceptors of [ 14 C]fucose from GDP‐[ 14 C]fucose as catalyzed by cell‐free extracts and purified membrane fractions of human colorectal carcinoma cells, SW1116, cultured in serum‐free medium. Purified lactotetraosyl ceramide (Galβ1→3GlcNAcβ1→3Galβ1→4Glc‐Cer or LcOse 4 Cer) and H‐1 glycolipid (Fucα1→2Galβ1→3GlcNAcβ1→3Galβ1→4Glc‐Cer or IV 2 FucαLcOse 4 Cer) stimulated incorporation of radioactivity into lipid‐soluble glycolipid at a rate greater than ten times that of Le a glycolipid [Galβ1→3(Fucα1→4)GlcNAcβ1→3Galβ1→4Glc‐Cer or III 4 FucαLcOse 4 Cer]. The enzymatic activities in crude and purified membrane fractions were optimized for substrate concentrations (glycolipid and GDP‐fucose), detergent requirement (taurocholate), pH, time and protein. The radioactive product of H‐1 fucosylation migrated as discrete and distinct bands on high‐performance thin‐layer chromatograms (HPTLC). Evidence for their identity with Le b fucolipid described previously [Fucα1→2Galβ1→3(Fucα1→4)GlcNAcβ1→3Galβ1→4Glc‐Cer or III 4 IV 2 (Fucα) LcOse 4 Cer] is presented. The radioactive product of LcOse 4 Cer fucosylation was mainly Le a fucolipid as determined by co‐migration with authentic Le a fucolipid in three HPTLC systems as native and acetylated derivatives. Our results also indicated a low level of H‐1 and Le b glycolipid synthesis from LcOse 4 Cer. On the basis of the optima, linearity for time, and enzyme‐limiting conditions, we obtained a 12–19‐fold purification of the LcOse 4 Cer and H‐1 fucosyl transferase acceptor activities in three peaks of a sucrose gradient. The peak with the highest specific activity (peak 3) was highest in density and in Na + , K + , ATPase specific activity, although NADH–cytochrome‐ c reductase and UDP‐GalNac transferase were also present in peak 3. The apparent K m values of LcOse 4 Cer acceptor activity and H‐1 acceptor activity in peak 3 were significantly different ( p < 0.01) by statistical tests, 2.4 μM and 0.5 μM, respectively. These apparent K m values were much lower (10 3 ×) and the pH optima were lower (4.8–5.3), than the corresponding properties reported for the α1→3/α1→4 fucosyl transferase purified from human milk. Our results suggest a role for the non‐glycosidic moieties of the acceptors and/or the tissue‐specific or primitive expression of these fucosyl transferase activities.