Role of Peptide Sequence and Neighboring Residue Glycosylation on the Substrate Specificity of the Uridine 5’-Diphosphate-R-N-acetylgalactosamine:Polypeptide N-acetylgalactosaminyl Transferases T1 and T2: Kinetic Modeling of the Porcine and Canine Submaxillary Gland Mucin Tandem Repeats † Thomas A. Gerken,* Chhavy Tep, and Jason Rarick W. A. Bernbaum Center for Cystic Fibrosis Research, Departments of Pediatrics and Biochemistry, Case Western ReserVe UniVersity School of Medicine, CleVeland, Ohio 44106 ReceiVed April 23, 2004; ReVised Manuscript ReceiVed June 1, 2004 ABSTRACT:A large family of uridine 5′-diphosphate (UDP)-R-N-acetylgalactosamine (GalNAc): polypeptide N-acetylgalactosaminyl transferases (ppGalNAc Ts) initiates mucin-type O-glycan biosynthesis at serine and threonine. The peptide substrate specificities of individual family members are not well characterized or understood, leaving an inability to rationally predict or comprehend sites of O-glycosylation. Recently, a kinetic modeling approach demonstrated neighboring residue glycosylation as a major factor modulating the O-glycosylation of the porcine submaxillary gland mucin 81 residue tandem repeat by ppGalNAc T1 and T2 [Gerken et al. (2002) J. Biol. Chem. 277, 49850-49862]. To confirm the general applicability of this model and its parameters, the ppGalNAc T1 and T2 glycosylation kinetics of the 80+ residue tandem repeat from the canine submaxillary gland mucin was obtained and characterized. To reproduce the glycosylation patterns of both mucins (comprising 50+ serine/threonine residues), specific effects of neighboring peptide sequence, in addition to the previously described effects of neighboring residue glycosylation, were required of the model. Differences in specificity of the two transferases were defined by their sensitivities to neighboring proline and nonglycosylated hydroxyamino acid residues, from which a ppGalNAc T2 motif was identified. Importantly, the model can approximate the previously reported ppGalNAc T2 glycosylation kinetics of the IgA1 hinge domain peptide [Iwasaki, et al. (2003) J. Biol. Chem. 278, 5613-5621], further validating both the approach and the ppGalNAc T2 positional weighting parameters. The characterization of ppGalNAc transferase specificity by this approach may prove useful for the search for isoform-specific substrates, the creation of isoform-specific inhibitors, and the prediction of mucin-type O-glycosylation sites. O-Glycosylated mucin-like domains serve important struc- tural and biological roles in many secreted and membrane- associated glycoproteins. For example, biological process such as the protection of epithelial cell surfaces, cellular adhesion, cellular protein targeting, the immune and inflam- matory responses, and the immune evasion of tumor cells are modulated by glycoproteins containing mucin-like do- mains, which are required for their biological properties (for example, refs 1-7). These domains typically contain 20- 30% hydroxyamino acids, serine and threonine, and are commonly composed of heavily O-glycosylated polypeptide tandem repeats. The mucin-type O-linked glycans, which typically range from one to more than 10 carbohydrate residues in length, are attached to serine or threonine via R-N-acetylgalactosamine (GalNAc) and impart an extended polypeptide conformation (8-10). The unique placement of specific mucin-type O-glycan structures in the peptide sequence in some instances is required for full biological activity (2, 11-16) and perhaps even for development (17, 18). Recently we have demonstrated that the extent of substitution and distribution of O-glycan structures on the porcine submaxillary gland mucin (PSM) 1 81 residue tandem repeat varies in a reproducible manner along the peptide sequence (19). Together, these findings suggest that the glycosyltransferases involved in the initial steps of mucin- type O-glycan biosynthesis are uniquely sensitive to features of the peptide sequence that are presently not fully under- stood. In the Golgi, the transfer of GalNAc to the peptide core is performed by a family of uridine 5′-diphosphate (UDP)- GalNAc:polypeptide N-acetylgalactosaminyl transferases † Supported by the National Institutes of Health, National Cancer Institute, Grant RO1-CA-78834, and by the Cystic Fibrosis Foundation. * Corresponding Author. Mailing address: Department of Pediatrics, Case Western Reserve University, School of Medicine, BRB, Cleveland, OH 44106-4948. Telephone: 216-368-4556. Fax: 216-368-4223. E-mail: txg2@cwru.edu. 1 Abbreviations: ppGalNAc T, UDP-GalNAc:polypeptide R- GalNAc transferase; CSM, canine submaxillary gland mucin; PSM, porcine submaxillary gland mucin; r 2 , least-squares correlation coef- ficient; SD, standard deviation; WOHn and WOGn , positional weighting coefficients for the presence of a nonglycosylated or glycosylated serine/ threonine residue at position n (see eqs 2 and 3 of ref 46); f(OG+OH), glycosylation state rate constant multiplier function (see eq 4 of ref 46); g(Pro,Glu,Arg), peptide sequence rate constant multiplier function defined in Experimental Procedures and eq 2; F(Xaa) n, residue specific positional rate factor defined in Experimental Procedures and eq 3. 9888 Biochemistry 2004, 43, 9888-9900 10.1021/bi049178e CCC: $27.50 © 2004 American Chemical Society Published on Web 07/08/2004