We have developed a genetic approach to isolate cloned cDNA sequences that determine expression of cell surface oligosaccharide structures and their cognate glycosyltransferases. A cDNA library was constructed in a mammalian expression vector by using mRNA from a murine cell line known to express a UDPgalactose:,B-D-galactosyl-1,4-N-acetyl-D-glucosaminide a-1,3-galactosyltransferase [(al-3)GT; EC 2.4.1.151]. This library was transfected into COS-1 cells, which lack expression of (al-3)GT. Transfected cells containing functional (al-3)GT cDNAs were detected and isolated with a lectin that recognizes the surface-expressed glycoconjugate product of the (al-3)GT enzyme. One cloned (al-3)GT cDNA was rescued from lectin-positive transfected cells. This cDNA contains a single long open reading frame that predicts a 394-amino-acid protein. No significant primary structure similarities were identified between this protein and other known sequences. However, the protein predicts a type II transmembrane topology similar to two other mammalian glycosyltransferases. This topology places the large COOH-terminal domain within the Golgi lumen; this domain was shown to be catalytically active when expressed in COS-1 cells as a portion of a secreted protein A fusion peptide. Biochemical analysis confirmed that this enzyme catalyzes a transglycosylation reaction between UDP-Gal and Gal(fi1-4)GlcNAc to form Gal(al-3)Gal(Ji1-4)GlcNAc. This cloning approach may be generally applicable to the isolation ofcDNAs encoding other mammalian glycosyltransferases.

The FX locus encodes an essential enzyme in the de novo pathway of GDP-fucose biosynthesis. Mice homozygous for a targeted mutation of the FX gene manifest a host of pleiotropic abnormalities including a lethal phenotype that is almost completely penetrant in heterozygous intercrosses on a mixed genetic background. Here we have investigated genetic suppression of FX-mediated lethality. Reduced recovery of heterozygous mice was observed while backcrossing the null FX allele to C57BL/6J (B6), but was less dramatic in an outcross to CASA/Rk and absent in an outcross to 129S1/SvImJ, indicating that genetic background modifies survival of FX+/) progeny. Substantial strain-specific differences in pre-and postnatal survival of FX)/) progeny were also detected in heterozygous crosses of C57BL/6J congenic, 129S1B6F1, and B6CASAF1 mice. Specifically, intrauterine survival of FX)/) mice was greatly increased during a heterozygous intercross on a uniform C57BL/6J genetic background compared with survival on a hybrid genetic background consisting of a mixture of C57BL/6J and 129S2/SvPas. In addition, statistically significant clustering of FX)/) progeny into litters and specific breeding cages was noted during a B6CASAF1 FX+/) intercross, suggesting a rare mechanism for modifier gene action in which parentally expressed genes define the phenotype, in this case the survival potential, of mutant offspring. Our results disclose that lethality in FX mutant mice is determined by one or more strainspecific modifier loci.

No abstract