Experimental determination of proline hydroxylation and hydroxyproline arabinogalactosylation motifs in secretory proteins

Many secretory and several vacuolar proteins in higher plants contain hydroxylated proline residues. In many cases, hydroxyprolines in proteins are glycosylated with either arabinogalactan or oligoarabinose. We have previously shown that a sporamin precursor is O-glycosylated at the hydroxylated proline 36 residue with an arabinogalactan-type glycan when this protein is expressed in tobacco BY-2 cells (Matsuoka et al., 1995). Taking advantage of the fact that this is the only site of proline hydroxylation and glycosylation in sporamin, we analyzed the amino acid requirement for proline hydroxylation and arabinogalactosylation. We expressed several deletion constructs and many amino acid substitution mutants in tobacco cells and analyzed glycosylation and proline hydroxylation of the expressed sporamins. Hydroxylation of a proline residue requires the five amino acid sequence [AVSTG]-Pro-[AVSTGA]-[GAVPSTC]-[APS or acidic] (where Pro is the modification site) and glycosylation of hydroxyproline (Hyp) requires the seven amino acid sequence [not basic]-[not T]-[neither P, T, nor amide]-Hyp-[neither amide nor P]-[not amide]-[APST], although charged amino acids at the -2 position and basic amide residues at the +1 position relative to the modification site seem to inhibit the elongation of the arabinogalactan side chain. Based on the combination of these two requirements, we concluded that the sequence motif for efficient arabinogalactosylation, including the elongation of the glycan side chain, is [not basic]-[not T]-[AVSG]-Pro-[AVST]-[GAVPSTC]-[APS].