Abnormal glycosphingolipid mannosylation triggers salicylic acid-mediated responses in Arabidopsis

Abstract

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.

Figure 1.

Substrate Specificity of GONST1. GONST1 was expressed in yeast and the total membrane fraction isolated. Proteoliposomes were prepared and preloaded with 10 mM of the nucleotide sugar. The uptake was initiated by adding [¹⁴C]-labeled GMP (corrected for background, calculated from yeast membranes lacking GONST1 expression). Data are presented compared with [¹⁴C]GMP/GMP uptake, which was set to 100%. n = 3 ± sd. Letters indicate significant differences between genotypes (one-way analysis of variance and Tukey’s honestly significant difference test, P < 0.05).

Figure 2.

Phenotype of gonst1. (A) Top row: 15-d-old, agar-grown wild type (WT) and gonst1. gonst1-1 has a Ws background ecotype, whereas gonst1-2 is Col-0. Bar = 1 cm. The inset shows a 15-d-old gonst1-2 leaf, displaying spontaneous lesions. Bottom row: 5-week-old wild-type and gonst1 soil-grown plants. Bar = 2.5 cm. (B) and (C) Fresh weights of the rosettes (B) and percentage of leaves displaying hypersensitive lesions (C) of 15-d-old, agar-grown gonst1-1, gonst1-2, and gonst1-3, along with their respective wild types. n = 21 to 28 individual plants grown simultaneously; ±se; asterisk indicates significant difference from the wild type (Student’s t test, P < 0.05). HR, hypersensitive response. (D) Neutral monosaccharide composition of aerial AIR from 6-week-old wild-type Ws, gonst1-1, gonst2-1, and gonst1-1 gonst2-1 plants. AIR was hydrolyzed with 2 M TFA and analyzed by HPAEC-PAD. n = 2, ±sd. (E) Immunofluorescent labeling of stem sections. Sections were labeled with an antibody specific to mannan, and cellulose was visualized with Calcofluor White. Bar = 100 µm. (F) PACE fingerprint of mannan in stem cell wall. Oligosaccharides released from AIR by treatment with mannanases were derivatized with 8-aminonapthalene-1,3,6-trisulphonic acid and visualized by PACE. Konjac glucomannan (KGM) treated the same way is shown for comparison, along with (Man)_1-6 oligosaccharide standards. A representative gel from multiple experiments is shown.

Figure 3.

Further Characterization of gonst1. (A) XyG structure. Leaf AIR was digested with xyloendoglucanase, and the resulting oligosaccharides analyzed by matrix-assisted laser desorption/ionization time of flight mass spectrometry. XyG structure of gonst1-2 appears like wild-type (WT) Col-0, whereas gonst1-2 mur1-1 resembles the Fuc biosynthetic mutant mur1-1, in which the oligosaccharides with Fuc mass 1393 (mass-to-charge ratio [m/z]) and 1555 (m/z) [M+Na]⁺ are absent. Instead, the oligo with l-Gal (XXJG) of 1409.30 (m/z) [M+Na]⁺ is increased. (B) RG-II dimerization. AIR was digested with endopolygalacturonase, soluble material separated by size exclusion chromatography, and the ratio of RGII dimerization was measured by refractive index. n = 2, ±sd. (C) AsA content of leaves; n = 3, +sd. For (B) and (C), no significant difference was seen between the mutants and their respective wild type (Student’s t test, P < 0.05).

Figure 4.

GIPC and Sphingolipidomic Analysis of the Wild Type and gonst1. (A) GIPC headgroup hexose composition. GIPCs were isolated from the wild type (WT) and gonst1 callus total membrane preparations. The GIPCs were analyzed by MS and categorized according to the number of hexoses in the sugar moiety (Hex_n). MRM was performed on a triple-quad mass spectrometer (QqQ-MS) for expected m/z (see Supplemental Data Set 1 online for details). Relative quantities were determined based on the peak areas of detected GIPCs, n = 3, ± sd. (B) Monosaccharide composition of the isolated GIPC headgroups. Sugars were hydrolyzed with 2 M TFA and analyzed by HPAEC-PAD. Fuc, Rha, Xyl, and GalA were undetectable in these samples. n = 3, ±sd. (C) Summary of sphingolipidomic analysis from whole callus extracts, as performed according to the method of Markham and Jaworski (2007); n = 3, ±sd (for detailed information, see Supplemental Figure 4 and Supplemental Data Set 1 online). Asterisks mark values significantly different from the wild type (Student’s t test; P < 0.05). dw, dry weight.

Figure 5.

gonst1 Exhibits a Constitutive Hypersensitive Response. (A) Cellular H₂O₂ was detected using DAB. Leaves of 15-d-old, agar-grown plants were analyzed. Controls (leaves incubated without DAB) are shown alongside. Representative images are shown of three biological replicates. WT, the wild type. (B) and (C) Total (B) and free (C) SA content of 15-d-old, agar-grown leaves. n = 5, ±sd. Asterisks mark values significantly different from the wild type (Student’s t test; P < 0.05). FW, fresh weight.

Figure 6.

The gonst1 Phenotype Can Be Partially Rescued by the Suppression of Salicylic Acid. (A) gonst1-2 was crossed with a plant expressing the bacterial salicylate hydroxylase gene (NahG) under the control of the cauliflower mosaic virus 35S promoter or with ics1, which has a lesion in an SA biosynthetic gene. Plants shown are as follows: top panel, 15 d old, agar grown, bar = 1 cm; bottom panel, 5 weeks old, soil grown, bar = 2.5 cm. (B) Cellular H₂O₂ detected using DAB staining of 15-d-old, agar-grown leaves. Representative images of three biological replicates are shown. Bar = 1 mm. (C) Total SA content of 15-d-old, agar-grown leaves (n = 3 to 5, ±sd). FW, fresh weight. (D) and (E) Number of necrotic lesions per plant (D) and fresh weight of 15-d-old, agar-grown plants (E). n = 10 to 23, ±se. WT, the wild type. (F) Stem height. Plants were transferred from agar to soil after 15 d of growth, and stem height was measured approximately every 7 d. Twelve plants per genotype were grown (two per cell), dispersed in a randomized pattern over three trays. Data shown are the mean ± se. Letters indicate significant differences between genotypes (repeated-measures analysis of variance and Tukey’s honestly significant difference test, P < 0.05).