Metabolic interactions between the Lands cycle and the Kennedy pathway of glycerolipid synthesis in Arabidopsis developing seeds

Abstract

It has been widely accepted that the primary function of the Lands cycle is to provide a route for acyl remodeling to modify fatty acid (FA) composition of phospholipids derived from the Kennedy pathway. Lysophosphatidylcholine acyltransferase (LPCAT) is an evolutionarily conserved key enzyme in the Lands cycle. In this study, we provide direct evidence that the Arabidopsis thaliana LPCATs, LPCAT1 and LPCAT2, participate in the Lands cycle in developing seeds. In spite of a substantially reduced initial rate of nascent FA incorporation into phosphatidylcholine (PC), the PC level in the double mutant lpcat1 lpcat2-2 remained unchanged. LPCAT deficiency triggered a compensatory response of de novo PC synthesis and a concomitant acceleration of PC turnover that were attributable at least in part to PC deacylation. Acyl-CoA profile analysis revealed complicated metabolic alterations rather than merely reduced acyl group shuffling from PC in the mutant. Shifts in FA stereo-specific distribution in triacylglycerol of the mutant seed suggested a preferential retention of saturated acyl chains at the stereospecific numbering (sn)-1 position from PC and likely a channeling of lysophosphatidic acid, derived from PC, into the Kennedy pathway. Our study thus illustrates an intricate relationship between the Lands cycle and the Kennedy pathway.

Figure 1.

LPCATs Are Expressed in Various Tissues. Promoter:GUS fusion analysis of LPCAT1 ([A] to [F]) and LPCAT2 ([G] to [L]) reveals distinct patterns of spatial expression. Tissues in (A) to (E) and (G) to (K) from the T2 generation and tissues in (F) and (L) from the T1 generation were analyzed. The LPCAT1 promoter was active in the seedling (A), rosette leaves (B), flower, particularly in sepals (white arrow in [C]) and stigma (black arrow in [C]), the chalazal endosperm (white arrowheads in [E] and [F]) from the globular stage (E) to the mature cotyledon stage (F), and the vascular bundle in the silique coat (white arrow in [F]). The LPCAT2 promoter was active in the seedling (G), rosette leaves (H), and anthers of flowers (black arrow in [I]), particularly pollen grains (black arrows in [J]). During seed development, the LPCAT2 promoter activity was maintained in both endosperm and embryo (white arrow in [K] and inset in [L]) from the globular stage (K) to the mature cotyledon stage (L). Bars = 1 mm (A) and (G), 1 cm in (B) and (H), 200 μm in (C), (D), (F), (I), (J), and (L), and 20 μm in (E), (K), and inset in (L).

Figure 2.

Alterations in Seed FA Composition Correlate with Transcript Abundance of LPCATs. (A) and (C) RT-PCR analysis revealed the deficiency of full-length transcripts of LPCAT1 and LPCAT2 in homozygous lpcat1 lpcat2-2 double mutant (A) and accumulated transcripts of both LPCAT1 and LPCAT2 in their respective overexpressed transgenic lines (C). 18S rRNA was used as an internal positive control. OE1-3 and OE1-6 represent two lines overexpressing LPCAT1; OE2-8 and OE2-9 represent two lines overexpressing LPCAT2. Control indicates the line transformed with an empty vector. WT, the wild type. (B) and (D) FA composition in mature seeds of lpcat1 lpcat2-2 (B) and overexpression lines (D). Minor FAs were omitted. Data represent mean and sd of at least four independent biological replicates. Asterisks indicate significant difference at P < 0.05 compared with the wild type based on Student’s t test.

Figure 3.

LPCAT Enzyme Activity and the Operation of the Lands Cycle Are Compromised in lpcat1 lpcat2-2 Developing Seeds. (A) LPCAT enzyme assay from microsomal preparations of developing siliques. Data represent mean and sd of three independent biological replicates. (B) The mol % of steady state LPC from developing seeds determined by ESI-MS/MS. Values are means and sd of five independent replicates. (C) The distributions of fatty acyl moieties and backbone, which were double labeled with [¹⁴C]acetate and [³H]glycerol, in PC (black) and DAG (white) species. Labeled PC and DAG species were transmethylated so that ¹⁴C-labeled fatty acyl group and ³H-labeled backbone were counted separately. The comparison of the ratio of ¹⁴C/³H between PC and DAG in both the wild type and lpcat1 lpcat2-2 after 5 min labeling is shown. Data represent mean and sd of two independent biological replicates. The absolute radioactivity is shown in Supplemental Table 5 online. (D) Stereochemical distribution of [¹⁴C]acyl chains in PC labeled with [¹⁴C]acetate reveals asymmetric distributions between sn-1 (black) and sn-2 (white) positions in the wild type and relatively even patterns in lpcat1 lpcat2-2. Three time points 5, 15, and 45 min, were monitored. Data represent mean and sd of two independent biological replicates. The absolute radioactivity is shown in Supplemental Table 6 online.

Figure 4.

The Initial Incorporation Rates of Newly Synthesized FAs Are Affected in the lpcat1 lpcat2-2 Developing Seeds. (A) to (F) The incorporations of [¹⁴C]acetate are presented in DPM/mg fresh weight. Predominant incorporation of ¹⁴C into PC is shown for the wild type (WT; [A] and [B]), and closer incorporation rates into PC and DAG are shown for lpcat1 lpcat2-2 ([C] and [D]). Two- to 45-min time course ([A] and [C]); 2- to 15-min course ([B] and [D]). Results in (E) and (F) represent relative levels of ¹⁴C in PC (E) and DAG (F) in lpcat1 lpcat2-2 compared with the wild type. Data represent mean and sd of two biological replicates. (G) and (H) The incorporations of [¹⁴C]-18:1 during the time course of 5, 15, and 45 min are presented in the wild type (G) and lpcat1 lpcat2-2 (H) in DPM/mg fresh weight. (I) and (J) Relative levels of ¹⁴C in PC (I) and DAG (J) in lpcat1 lpcat2-2 compared with the wild type. Data represent mean and sd of three biological replicates.

Figure 5.

LPCAT Deficiency Affects the Flux into de Novo Glycerolipid Synthesis. [¹⁴C]Glycerol backbone feeding during the time course of 2, 5, and 15 min. Presented as DPM/mg fresh weight, the kinetics of labeled PC and DAG are shown in the wild type (WT) (A) and lpcat1 lpcat2-2 (B). Comparison of labeled PA between the wild type and lpcat1 lpcat2-2 is presented in (C). Values are mean and sd of two biological replicates. Note: These values included small proportions of [¹⁴C]glycerol incorporated into acyl chains. Our analysis confirmed that the relative levels of PC, DAG, and TAG in the wild type and lpcat1 lpcat2-2 were not affected by labels in the acyl groups (see Supplemental Figure 5C online).

Figure 6.

Incorporation of [¹⁴C]Choline into PC and LPC in the lpcat1 lpcat2-2 Developing Seeds. [¹⁴C]Choline feeding for 1 and 2 h shows that lpcat1 lpcat2-2 seeds contain higher ¹⁴C in PC (A) and LPC (B) compared with the wild type (WT). Data are presented as DPM/mg fresh weight. Values are mean and sd of three independent replicates.

Figure 7.

Accelerated PC Turnover in lpcat1 lpcat2-2. (A) After 1-h pulse labeling with [¹⁴C]choline, the chase was conducted in nonlabeled medium. The labels in PC at 0, 1, 4, and 8 h were determined. Values are mean and sd of two independent replicates. WT, the wild type. (B) The chase was conducted after a 45-min pulse labeling with [¹⁴C]-18:1. The labels in PC were determined at 0, 1, 2, and 4 h. Data are representative of two independent experiments, and another replicate is shown in Supplemental Figure 8 online.

Figure 8.

Genes Associated with PC Synthesis and Breakdown Are Upregulated in lpcat1 lpcat2-2. The relative expression level of representative genes from the microarray data was verified by qRT-PCR. The transcript levels were normalized first to those of TUBULIN3 and then further normalized to the levels in the wild type (WT). Data represent mean and sd of five biological replicates. Asterisks indicate significant difference at P < 0.05 compared with the wild type based on Student’s t test.

Figure 9.

LPCAT Deficiency Impacts the Composition of the Acyl-CoA Pool. (A) Comparison of mol % of the acyl CoA esters of the wild type (WT) and lpcat1 lpcat2-2 developing seeds reveals shifts in the composition of fatty acyl moieties. Data represent mean and sd of three independent replicates. Asterisks indicate significant difference at P < 0.05 compared with the wild type based on Student’s t test. (B) Representative chromatographic traces of the wild type and lpcat1 lpcat2-2 are shown. ISTD, internal standard.

Figure 10.

Working Model of the Interplay between the Kennedy Pathway and the Lands Cycle. This model summarizes the major steps of the Kennedy pathway and the Lands cycle. The Kennedy pathway encompasses TAG synthesis steps mediated by GPAT, LPAAT, PAP, and diacylglycerol acyltransferase, as well as the reaction steps of de novo PC synthesis. The Lands cycle, depicted by blue arrows, involves the deacylation of de novo–synthesized PC by PLA₂ or the LPCAT reverse reaction and the reacylation of LPC to PC through LPCAT. Several alternative reactions, including PDCT, PLD + PAP, and the reversed reaction of AAPT, mediate PC-DAG interconversion. TAG synthesis can also be accomplished via PDAT. In the wild type, as shown in green arrows and lines, major flux of newly synthesized fatty acyl-CoAs (pool A, largely 18:1) is channeled through the sn-2 position of LPC by LPCAT. The FAs originating from PC deacylation mix with the nascent acyl-CoA to form pool B. In the lpcat1 lpcat2-2 mutant, compromised reacylation of 18:1 to LPC is indicated with a stop sign, while both the Kennedy pathway and PC deacylation route are enhanced. The dashed line in red denotes a potential route of LPC to LPA conversion catalyzed by an unidentified enzyme in plants.