NH2-terminal amino acid sequence obtained from a UDP-GalNAc:polypeptide N-acetylgalactosaminyl-transferase (GalNAc-transferase) isolated from bovine colostrum was used for the construction of synthetic oligonucleotide primers. Subsequent polymerase chain reaction and library screenings of a bovine intestine cDNA library produced seven positive clones. The largest clone had a 2294-base pair insert that contained an open reading frame coding for a protein composed of 559 amino acids with a predicted polypeptide molecular mass of 64,173 Da. The cloned molecule has no significant sequence homology to previously reported cloned glycosyltransferases, but appears to have a similar domain structure. It is a type II membrane protein with a 23-amino acid putative transmembrane region starting 8 amino acids from the NH2 terminus. The transmembrane segment of the molecule is immediately followed by a sequence rich in proline residues. The molecule contains three consensus sequences for N-linked glycosylation and five predicted sites for O-glycosylation. Northern blot analysis of poly(A+) mRNA isolated from Madin-Darby bovine kidney cells, bovine mammary tissue, and eight human tissues demonstrated the expression of two transcripts differing in size by approximately 1 kilobase. The cloned DNA was expressed in insect cells using a baculovirus vector. This resulted in an almost 100-fold increase in GalNAc-transferase activity in lysates prepared from cells infected with virus containing the GalNAc-transferase gene compared to cells infected with virus containing DNA coding for an unrelated molecule or uninfected cells. Immunoprecipitation from lysates prepared from infected cells labeled in vivo with [35S] methionine showed a large increase in the recovery of an approximately 67-kDa protein.