The Arabidopsis thaliana protein GOLGI-LOCALIZED NUCLEOTIDE SUGAR TRANSPORTER (GONST1) has been previously identified as a GDP-D-mannose transporter. It has been hypothesized that GONST1 provides precursors for the synthesis of cell wall polysaccharides, such as glucomannan. Here, we show that in vitro GONST1 can transport all four plant GDP-sugars. However, gonst1 mutants have no reduction in glucomannan quantity and show no detectable alterations in other cell wall polysaccharides. By contrast, we show that a class of glycosylated sphingolipids (glycosylinositol phosphoceramides [GIPCs]) contains Man and that this mannosylation is affected in gonst1. GONST1 therefore is a Golgi GDP-sugar transporter that specifically supplies GDP-Man to the Golgi lumen for GIPC synthesis. gonst1 plants have a dwarfed phenotype and a constitutive hypersensitive response with elevated salicylic acid levels. This suggests an unexpected role for GIPC sugar decorations in sphingolipid function and plant defense signaling. Additionally, we discuss these data in the context of substrate channeling within the Golgi.

No abstract

Polysaccharides containing beta-1,4-mannosyl residues (mannans) are abundant in the lignified secondary cell walls of gymnosperms, and are also found as major seed storage polysaccharides in some plants, such as legume species. Although they have been found in a variety of angiosperm tissues, little is known about their presence and tissue localisation in the model angiosperm, Arabidopsis thaliana (L.) Heynh. In this study, antibodies that specifically recognised mannans in competitive ELISA experiments were raised in rabbits. Using these antibodies, we showed that Golgi-rich vesicles derived from Arabidopsis callus were able to synthesise mannan polysaccharides in vitro. Immunofluorescence light microscopy and immunogold electron microscopy of Arabidopsis inflorescence stem sections revealed that the mannan polysaccharide epitopes were localised in the thickened secondary cell walls of xylem elements, xylem parenchyma and interfascicular fibres. Similarly, mannan epitopes were present in the xylem of the leaf vascular bundles. Surprisingly, the thickened epidermal cell walls of both leaves and stems also contained abundant mannan epitopes. Low levels were observed in most other cell types examined. Thus, mannans are widespread in Arabidopsis tissues, and may be of particular significance in both lignified and non-lignified thickened cell walls. Polysaccharide analysis using carbohydrate gel electrophoresis (PACE) of cell wall preparations digested with a specific mannanase showed that there is glucomannan in inflorescence stems. The findings show that Arabidopsis can be used as a model plant in studies of the synthesis and functions of mannans.

Transport of nucleotide-sugars across the Golgi membrane is required for the lumenal synthesis of a variety of essential cell surface components, and is mediated by nucleotide sugar transporters (NSTs) which are members of the large drug/metabolite superfamily of transporters. Despite the importance of these proteins in plants, so far only two have been described, GONST1 and AtUTr1 from Arabidopsis thaliana. In this work, our aim was to identify further Golgi nucleotide-sugar transporters from Arabidopsis. On the basis of their sequence similarity to GONST1, we found four additional proteins, which we named GONST2, 3, 4 and 5. These putative NSTs were grouped into three clades: GONST2 with GONST1; GONST3 with GONST4; and GONST5 with six further uncharacterized proteins. Transient expression in tobacco cells of a member of each clade, fused to the Green Fluorescent Protein (GFP), suggested that all these putative NSTs are localised in the Golgi. To obtain evidence for nucleotide sugar transport activity, we expressed these proteins, together with the previously characterised GONST1, in a GDP-mannose transport-defective yeast mutant (vrg4-2). We tested the transformants for rescue of two phenotypes associated with this mutation: sensitivity to hygromycin B and reduced glycosylation of extracellular chitinase. GONST1 and GONST2 complemented both phenotypes, indicating that GONST2, like the previously characterized GONST1, is a GDP-mannose transporter. GONST3, 4 and 5 also rescued the antibiotic sensitivity, but not the chitinase glycosylation defect, suggesting that they can also transport GDP-mannose across the yeast Golgi membrane but with a lower efficiency. RT-PCR and analysis of Affymetrix data revealed partially overlapping patterns of expression of GONST1-5 in a variety of organs. Because of the differences in ability to rescue the vrg4 - 2 phenotype, and the different expression patterns in plant organs, we speculate that GONST1 and GONST2 are both GDP-mannose transporters, whereas GONST3, GONST4 and GONST5 may transport other nucleotide-sugars in planta.