Arabidopsis lipins mediate eukaryotic pathway of lipid metabolism and cope critically with phosphate starvation

Abstract

Phosphate is an essential nutrient for plant viability. It is well-established that phosphate starvation triggers membrane lipid remodeling, a process that converts significant portion of phospholipids to non-phosphorus-containing galactolipids. This remodeling is mediated by either phospholipase C (PLC) or phospholipase D (PLD) in combination with phosphatidate phosphatase (PAP). Two PLC genes, NPC4 and NPC5, and PLD genes, PLDzeta1 and PLDzeta2, are shown to be involved in the remodeling. However, gene knockout studies show that none of them plays decisive roles in the remodeling. Thus, although this phenomenon is widely observed among plants, the key enzyme(s) responsible for the lipid remodeling in a whole plant body is unknown; therefore, the physiological significance of this conversion process has remained to be elucidated. We herein focused on PAP as a key enzyme for this adaptation, and identified Arabidopsis lipin homologs, AtPAH1 and AtPAH2, that encode the PAPs involved in galactolipid biosynthesis. Double mutant pah1pah2 plants had decreased phosphatidic acid hydrolysis, thus affecting the eukaryotic pathway of galactolipid synthesis. Upon phosphate starvation, pah1pah2 plants were severely impaired in growth and membrane lipid remodeling. These results indicate that PAH1 and PAH2 are the PAP responsible for the eukaryotic pathway of galactolipid synthesis, and the membrane lipid remodeling mediated by these two enzymes is an essential adaptation mechanism to cope with phosphate starvation.

Fig. 1.

Identification of Arabidopsis lipins as functional phosphatidate phosphatases. (A) Complementation of temperature-sensitive phenotypes observed in the Δdpp1Δlpp1Δpah1 triple mutant by Arabidopsis PAH1 or PAH2. (B) Dependency of in vitro PAP activity on Mg²⁺. Data are means ± SD of triplicate assays with independent samples.

Fig. 2.

Characterization of pah1pah2. (A) Relative decrease in PAP activity in the supernatant fraction of leaf crude extract from pah1pah2. Data shown are means ± SD of triplicate measurements. Gray bar, wild-type; black bar, pah1pah2. (B) Relative increase in labeled PA after [³²P]-phosphate labeling of detached pah1pah2 rosette leaves. Data shown are means ± SD of triplicate measurements. Gray bar, wild-type; black bar, pah1pah2. (C) Lipid composition of pah1pah2. Total lipid was extracted from the aerial part of 20-day-old plants and analyzed. Data are means ± SD of triplicate analyses with independently extracted samples. Inset graph shows PA levels. Data are means ± SD of quadruplicate analyses with independently extracted samples. Gray bars, wild type; black bars, pah1pah2. (D) Subcellular localization of PAH1-GFP and PAH2-GFP. Rosette leaves of 35S::PAH1-GFP, pah1pah2 or 35S::PAH2-GFP, pah1pah2 transgenic plants were fractionated into membrane and soluble fraction as described in the Materials and Methods section. Proteins (50 μg each) were run on SDS/PAGE, blotted onto a nitrocellulose membrane and PAH1-GFP/PAH2-GFP detected by monoclonal anti-GFP antibody. Controls used are RuBisCO large subunit (RbcL) for plastid storoma and NADPH-dependent thioredoxin reductase A (NTA) for cytosol (27). Lane 1, crude; lane 2, crude of wild-type; lane 3, pellet; lane 4, supernatant.

Fig. 3.

Impairment of eukaryotic pathway in pah1pah2 by in vivo pulse-chase labeling. (A and B) [¹⁴C]acetate labeling of fatty acids associated with individual lipids in (A) wild-type and (B) pah1pah2 plants were analyzed. Solid arrows indicate the plastid pathway component of MGDG labeling; the broken arrows represent the ER pathway component. Closed diamonds, MGDG; gray squares, PC; open triangles, PE; closed triangles, DGDG; closed circles, PG. (C and D) [¹⁴C]glycerol labeling of glycerol backbone in (C) wild-type and (D) pah1pah2 were analyzed. Closed diamonds, MGDG; gray squares, PC; open triangles, PE+PG; closed triangles, DGDG; closed circles, PI. Experiments were repeated three times, each with similar results. Shown is a representative result.

Fig. 4.

Phenotype of pah1pah2 under phosphate starvation. (A–D) Ten-day-old seedlings were transferred to solid media with or without phosphate for an additional 10 days. Phenotypes of 20-day-old wild-type and pah1pah2 were compared under phosphate-replete (A and B) or phosphate-starved (C and D) conditions. (A and C), wild-type; (B and D), pah1pah2. Scale bars, 1 cm. (E) Root growth in phosphate-starved pah1pah2. Seeds were germinated on vertically maintained solid media plates with or without phosphate (n = 8). (F) Lipid composition of pah1pah2 vegetative shoots under phosphate starvation. The 20-day-old plants that suffered 10-day phosphate starvation were used for lipid analysis. Data are means ± SD of triplicate experiments with independently extracted samples. Inset graph shows PA levels in vegetative shoots. Data are means ± SD of quadruplicate analyses with independently extracted samples. Gray bars, wild-type; black bars, pah1pah2.

Fig. 5.

A proposed model for membrane lipid remodeling during phosphate starvation. The pathway in red arrows shows substrate supply to the MGD1-mediated DGDG production, whereas that in blue arrows indicates pathway to the MGD2/3-mediated DGDG production.