Glycosylinositol phosphorylceramides from Rosa cell cultures are boron-bridged in the plasma membrane and form complexes with rhamnogalacturonan II

Abstract

Boron (B) is essential for plant cell-wall structure and membrane functions. Compared with its role in cross-linking the pectic domain rhamnogalacturonan II (RG-II), little information is known about the biological role of B in membranes. Here, we investigated the involvement of glycosylinositol phosphorylceramides (GIPCs), major components of lipid rafts, in the membrane requirement for B. Using thin-layer chromatography and mass spectrometry, we first characterized GIPCs from Rosa cell culture. The major GIPC has one hexose residue, one hexuronic acid residue, inositol phosphate, and a ceramide moiety with a C18 trihydroxylated mono-unsaturated long-chain base and a C24 monohydroxylated saturated fatty acid. Disrupting B bridging (by B starvation in vivo or by treatment with cold dilute HCl or with excess borate in vitro) enhanced the GIPCs' extractability. As RG-II is the main B-binding site in plants, we investigated whether it could form a B-centred complex with GIPCs. Using high-voltage paper electrophoresis, we showed that addition of GIPCs decreased the electrophoretic mobility of radiolabelled RG-II, suggesting formation of a GIPC-B-RG-II complex. Last, using polyacrylamide gel electrophoresis, we showed that added GIPCs facilitate RG-II dimerization in vitro. We conclude that B plays a structural role in the plasma membrane. The disruption of membrane components by high borate may account for the phytotoxicity of excess B. Moreover, the in-vitro formation of a GIPC-B-RG-II complex gives the first molecular explanation of the wall-membrane attachment sites observed in vivo. Finally, our results suggest a role for GIPCs in the RG-II dimerization process.

Keywords: Rosa sp; boron; cell wall; cross-linking; glycosylinositol phosphoceramides; glycosylinositol phosphorylceramides; lipid rafts; plasma membrane; rhamnogalacturonan II.

Figure 1

Mass spectrometric and thin-layer chromatographic analysis of Rosa glycosylinositol phosphorylceramides (GIPCs). (a) ESI-MS analysis of GIPC extract from Rosa cell culture. The spectrum was acquired in the negative ion mode. Abbreviations: Hex, hexose residue (probably α-mannose); HexA, hexuronic acid residue (probably α-glucuronic acid); Pent, pentose residue; Ins, myo-inositol; P, phosphate; Cer, phytoceramide. Inset: proposed structure of the predominant GIPC species; phytoceramide moiety in grey box. (b) ESI-MS/MS (collision-induced dissociation spectrum) analysis of the predominant Hex-HexA-Ins-P-Cer peak seen in (a) as the [M-2H]²⁻ ion at m/z 630. Nitrogen was used as collision gas in a Q-TRAP instrument, with the collision energy set to −40 eV. The standard nomenclature for glycolipid fragmentation has been applied (Costello and Vath, 1990; Levery et al., 2001). Inset: proposed identity of the ion at m/z = 438.4, indicating an h24:0 ceramide moiety. (c, d) Thin-layer chromatography (TLC) of GIPC extract. Lipids were chromatographed in CHCl₃/CH₃OH/4 m NH₄OH (9:7:2, by vol.) with 0.2 m ammonium acetate (Kaul and Lester, 1978) and located by orcinol reagent (c) or periodic acid–Schiff staining (d). Lipid bands are labelled: 1, Hex-HexA-Ins-P-Cer; 2, (Hex)₂-HexA-Ins-P-Cer; 3, Pent-(Hex)₂-HexA-Ins-P-Cer.

Figure 2

Influence of boron on glycosylinositol phosphorylceramide (GIPC) extraction. (a) A cloudy layer was observed during butanol/water phase-partitioning of a GIPC-enriched lipid sample extracted with neutral ethanol from Rosa cell cultures that had been grown in the usual B concentration (i, iv). This cloudy layer disappeared in the presence of 0.1 m HCl (iii, vii), 10 mm βMCD (ii), or 6 mm borate buffer, pH 9.2 (vi). The horizontal arrow indicates the slight cloudy layer left in the presence of βMCD (butanol above). In contrast, 6 mm ammonium buffer, pH 9.2 (v), only led to a partial disappearance. (b) TLC of the different phases after butanol/water phase-partitioning of a GIPC-rich lipid extract from Rosa cell cultures grown in media with (B+) or without boron (B−). The lipids had been extracted in 70% ethanol that contained 0.1 m HCl (H+) or lacking acid (H−). BP, butanol phase; CL, cloudy layer; AP, aqueous phase; Suc, sucrose (marker). In lanes 9 and 10, 10 mm βMCD was present during the partitioning step. Lipids labelled on lane 10: bands 1–3, as in Figure1; band 4, (Pent)₂-(Hex)₂-HexA-Ins-P-Cer.

Figure 3

Purification of glycosylinositol phosphorylceramide (GIPC) from Rosa cell culture. (a) GIPC purification scheme adapted from Buré et al. (2011). (b) Thin-layer chromatography (TLC) of the phase A, B and C obtained during the step shaded grey in (a). Suc, sucrose marker.

Figure 4

High-voltage paper electrophoresis of radiolabelled RG-II and the effect of boric acid, Pb²⁺ and glycosylinositol phosphorylceramides (GIPCs). (a) [³H]RG-II was loaded on to the paper after 9 h of pre-treatment of monomeric [³H]RG-II in 250 mm pyridine buffer, pH 4.7, with or without 1 mm H₃BO₃ or 1 mm H₃BO₃ plus 0.5 mm Pb(NO₃)₂. After electrophoresis, the paper was cut into strips, which were assayed for tritium by scintillation-counting. The positions of non-radioactive markers (glucose, galacturonic acid and Orange G) are also shown. (b) Samples were loaded on the paper after 4 h of pre-treatment of monomeric [³H]RG-II (RGIIm) in 50 mm ammonium acetate, pH 4.8, with or without various combinations of 1.2 mm boric acid (B), 0.5 mm Pb(NO₃)₂ (Pb²⁺), GIPCs (purified as described in Figure3), commercial phytoceramide (PC), and 0.1 m hydrochloric acid (HCl). After electrophoresis, papers were read on a LabLogic AR2000 radio-TLC Imaging Scanner. The distance migrated (d) is given relative to dRGIIm (the distance migrated by monomeric RG-II, that is to say approximately 6 cm). Grey bands indicate the positions of monomeric RG-II (right) and the GIPC–RG-II complex (left).

Figure 5

Glycosylinositol phosphorylceramide (GIPC)-induced dimerization of monomeric RG-II. After 4 h of pre-treatment of monomeric RG-II without or with 1.2 mm H₃BO₃ ± GIPC or with 1.2 mm H₃BO₃ and 0.5 mm Pb(NO₃)₂, samples were analysed by polyacrylamide gel electrophoresis and silver stained. Abbreviations as in Figure4. Parts (a) and (b) show the results of two independent experiments.

Figure 6

Co-expression network between genes involved in RG-II and glycosylinositol phosphorylceramide (GIPC) biosynthesis in various species using an edge-weighted force-directed approach, based on data retrieved from ATTED-II and visualised in Cytoscape 2.8 (http://www.cytoscape.org). Circles: white circle, LOH (encoding a ceramide synthase); pale grey, encoding enzymes involved in synthesis of RG-II side-chains; dark grey, encoding enzymes involved in synthesis of the homogalacturonan backbone; black, C4H (encoding cinnamate-4-hydroxylase, involved in lignin synthesis; included as a control). Lines: green lines, Arabidopsis thaliana; blue, Populus trichocarpa; red, Oryza sativa; black, Glycine max; yellow, Zea mays. Solid line, strong co-expression [mutual rank (MR) < 1000]; dotted line, weak co-expression (5000 > MR > 1000); no line, no transcriptomic data available or no co-expression (MR > 5000).