Cell surface ␣(1,3)-and ␣(1,4)-fucosylated oligosaccharides have received a substantial amount of attention because some are thought to be essential to the initiation of immune cell adhesion to vascular endothelium during the inflammatory process (reviewed in Refs. 1-5 and 6 -15). This is perhaps best characterized by the "rolling" type of adhesion mediated by interactions between sialyl Lewis x (sLe x )-bearing glycoconjugates on leukocytes and P-and E-selectin expressed by activated vascular endothelium (5). Animal intervention studies indicate that such fucosylated oligosaccharide molecules can act as anti-inflammatory agents by inhibiting leukocyte-endothelial cell interactions (15-18). Furthermore, absence of such molecules on the leukocytes of individuals with the rare human leukocyte adhesion II syndrome is associated with profound defects in leukocyte-endothelial cell adhesion and with nonpyogenic infections (19,20). These observations suggest that compounds capable of specifically inhibiting leukocyte sialyl Lewis x expression might represent candidates for anti-inflammatory pharmacologic agents.The ␣(1,3)-fucosyltransferases (␣(1,3)-Fuc-Ts) 1 represent a target for such inhibitory agents, since synthesis and expression of the sLe x molecule and related ␣(1,3)-and ␣(1,4)-fucosylated oligosaccharides are controlled, in large measure, by ␣(1,3)Fuc-Ts (5). These enzymes catalyze the attachment of L-fucose in ␣ anomeric linkage to one or more distinct oligosaccharide precursors. Biochemical and molecular cloning studies indicate that the human genome encodes at least five distinct ␣(1,3)-Fuc-Ts (9, 21-28; reviewed in Refs. 5, 29, and 30). Each enzyme can react with one or more structurally distinct oligosaccharides and can thereby generate a unique spectra of cell surface ␣(1,3)-fucosylated oligosaccharide products. In turn, these molecules may exhibit distinct biologic functions, including those involving selectin-dependent cell adhesion.Although the primary sequences are known for several ␣(1,3)-Fuc-Ts (9, 21-28), the structural determinants within these enzymes that dictate their different substrate specificities remain undefined. The purpose of this work is to identify such protein sequence(s) and to provide a conceptual background for work to design or identify molecules that might inhibit ␣(1,3)-Fuc-Ts by interacting with acceptor substrate binding sites.In this study, we chose to explore three human ␣(1,3)-Fuc-Ts (Fuc-TIII, Fuc-TV, and Fuc-TVI) with informative structural and catalytic properties. These enzymes share approximately 85% overall amino acid sequence identity (24,25). The COOH termini of these enzymes maintain nearly identical amino acid sequences, whereas their NH 2 -terminal regions are punctuated by foci of nonidentical amino acids; we term these regions "hypervariable" segments. These three enzymes maintain shared and distinct acceptor substrate specificities. In partic-