The role of lysophosphatidic acid acyltransferase 1 in reproductive growth of Arabidopsis thaliana

Abstract

Lysophosphatidic acid acyltransferase1 (LPAT1) catalyzes the second step of de novo glycerolipid biosynthesis in chloroplasts. However, the embryonic-lethal phenotype of the knockout mutant suggested an unknown role for LPAT1 in non-photosynthetic reproductive organs. Reciprocal genetic crossing of the lpat1-1 heterozygous line suggested a female gametophytic defect of the lpat1-1 knockout mutant. By suppressing LPAT1 specifically during seed development, we showed that LPAT1 suppression affected silique growth and seed production. Glycerolipid analysis of the LPAT1 knockdown lines revealed a pronounced decrease of phosphatidylcholine (PC) content in mature siliques along with an altered polyunsaturation level of the polar glycerolipids. In seeds, the acyl composition of triacylglycerol (TAG) was altered albeit not the content. These results indicate that plastidic LPAT1 plays an important role in reproductive growth and extraplastidic glycerolipid metabolism involving PC and TAG.

Introduction

Lysophosphatidic acid acyltransferase (LPAT) catalyzes the conversion of lysophosphatidic acid to phosphatidic acid (PA). Since the product PA is a common precursor for the biosynthesis of different classes of glycerolipids, including phospholipids, galactolipids, and triacylglycerol (TAG), LPAT catalyzes an important reaction step in de novo glycerolipid biosynthesis. The LPAT activity is present in the plastids and endoplasmic reticulum (ER), two major sites of glycerolipid biosynthesis in plants. In Arabidopsis, the LPAT family comprises five isoforms (LPAT1, LPAT2, LPAT3, LPAT4, and LPAT5), and only LPAT1 is localized at the plastids (Kim and Huang, 2004; Yu et al., 2004) while the others are at the ER (Kim et al., 2005; Angkawijaya et al., 2019; Barroga and Nakamura, 2022). We previously showed that LPAT2 is the major ER-localized LPAT isoform for extraplastidic glycerolipid biosynthesis (Barroga and Nakamura, 2022). Indeed, partial suppression of LPAT2 expression causes a pleiotropic effect on plant growth (Barroga and Nakamura, 2022), and a gene knockout study showed gametophytic lethality (Kim et al., 2005). Since LPAT1 is the only plastid-localized isoform, its activity is assumed to be crucial for the plastid-localized biosynthesis of glycerolipids, monogalactosyldiacylglycerol (MGDG), digalactosyldiacylglycerol (DGDG), sulfoquinovosyldiacylglycerol (SQDG), and phosphatidylglycerol (PG). Gene knockout studies of LPAT1 showed that the homozygous mutant cannot be isolated due to lethal developmental arrest at embryogenesis (Kim and Huang, 2004; Yu et al., 2004). While this lethal phenotype of the knockout mutant implies an essential role for LPAT1 in plastid-localized glycerolipid biosynthesis, no embryonic-lethal phenotype is reported in the mutant of ACT1/ATS1, which catalyzes a reaction immediately before LPAT1. Suppression of LPAT1 in the genetic background of non-lethal ats1-1 showed a dwarfed growth phenotype, suggesting that available T-DNA mutants of ACT1/ATS1 are all leaky (Xu et al., 2006). Although a specific role for plastidic LPAT1 in the reproductive processes has been postulated due to the embryonic-lethal phenotype of the knockout mutant (Kim and Huang, 2004; Yu et al., 2004), the unavailability of a homozygous knockout mutant hampered further characterization of the function of LPAT1 during reproductive growth.

Here, we created transgenic knockdown lines of LPAT1 that conditionally suppress LPAT1 expression during seed development using the activity of a seed-specific promoter that drives expression of artificial miRNA (amiRNA) targeted to LPAT1. These non-lethal suppression lines showed gametophytic defects and aborted seed production. Lipid analysis revealed that extraplastidic PC content was the most affected lipid classes in the suppression lines, accompanied by changes of seed TAG quality. We suggest that plastidic LPAT1 is crucial for the reproductive growth and is involved in extraplastidic glycerolipid metabolism.

Materials and methods

Plant materials and growth conditions

Arabidopsis thaliana Columbia-0 ecotype (wild type; WT) was used in this work. Mutant lpat1-1 (SALK_073445) (Kim and Huang, 2004) seeds were obtained from the Nottingham Arabidopsis Stock Centre (NASC). The plants were germinated and grown on half-strength Murashige and Skoog (MS) medium containing 1% sucrose and 0.6% agarose (Murashige and Skoog, 1962) under 16 h light/8 h dark photoperiodic condition with white light illumination (75 μmol m^–2 s^–1) at 22 °C for the first week then subsequently grown on soil.

Isolation of the T-DNA mutant

Homozygous plants of the T-DNA mutant lpat1-1 were isolated by PCR-based genotyping with gene-specific primers and T-DNA-specific primers as listed in Supplementary Table S1. The primers used for lpat1-1 were YN749/ISY029. The positions of the T-DNA insertion were confirmed by DNA sequencing.

Vector construction and plant transformation

ProLPAT1:LPAT1

The genomic sequence of LPAT1 (3020 bp) was amplified from Arabidopsis Col-0 DNA using primers NB108/NB106. Phusion High-Fidelity DNA Polymerase (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) was used to amplify the PCR product, which was gel purified and cloned into an entry vector pENTR/D-TOPO plasmid (Invitrogen, Thermo Fisher Scientific) to construct pNB038 (pENTR:ProLPAT1:LPAT1).

ProLPAT1:LPAT1-GUS

To fuse the β-glucuronidase (GUS) reporter protein driven by the LPAT1 endogenous promoter, an SfoI site was inserted before the stop codon of pNB038 by using PCR-based site-directed mutagenesis (Sawano and Miyawaki, 2000) with primer NB115 to create pNB039 (pENTR:ProLPAT1:LPAT1-SfoI). Then, pAA040 (ProLPAT5:LPAT5-GUS) (Angkawijaya et al., 2019) was digested with SfoI and the GUS fragment was ligated into the SfoI site of pNB039 to create pNB042 (pENTR:ProLPAT1:LPAT1-GUS). Next, pNB042 was recombined into a pBGW destination vector (Karimi et al., 2002) using Gateway LR Clonase II (Invitrogen, Thermo Fisher Scientific) to create pNB044 (pBGW:ProLPAT1:LPAT1-GUS). Finally, pNB044 was transduced into the lpat1-1/+ heterozygous plants (SALK_073445) by Agrobacterium-mediated transformation. Twenty-four T₁ lines were screened, and lines #3 and #20 were selected for GUS staining analysis.

ProLPAT1:LPAT1-Ven

To express the LPAT1–Ven fusion protein driven by the LPAT1 endogenous promoter, a triple repeat of Venus fluorescent protein was digested from pAA011 (ProLPAT3:LPAT3-Ven) by SfoI and ligated into the SfoI site of pNB039 to create pNB040 (pENTR:ProLPAT1:LPAT1-Ven). Then, pNB040 was recombined to a pBGW destination vector using Gateway LR Clonase II to create pNB041 (pBGW:ProLPAT1:LPAT1-Ven). Finally, pNB041 was transduced into lpat1-1/+ by Agrobacterium-mediated transformation. Twenty-four T₁ lines were screened, and line #6 was selected for confocal microscope observations.

Cloning of the β-phaseolin promoter

To clone the β-phaseolin promoter (ProPhas), total DNA was extracted from common bean (Phaseolus vulgaris) according to the method of Maciel et al. (2001). Then, the β-phaseolin gene was amplified by PCR using Phusion High-Fidelity DNA Polymerase (Invitrogen, Thermo Fisher Scientific) employing primers NB094/NB096. The amplified PCR product was gel purified and cloned into an entry vector pENTR-TOPO plasmid (Invitrogen, Thermo Fisher Scientific) to construct pNB030 (pENTR_ProPhas).

ProPhas:GUS

To express the GUS reporter protein by the β-phaseolin promoter, pNB030 was recombined into pGWB633 (Nakamura et al., 2010), a Gateway destination vector containing the GUS reporter protein without a promoter, using Gateway LR Clonase II to create pNB045 (pGWB_ProPhas:GUS). Subsequently, pNB045 was transduced into Arabidopsis WT (Col-0) plants using Agrobacterium-mediated transformation. Twenty-four lines were screened, and lines #2 and #6 were selected for GUS staining assay.

ProPhas:amiLPAT1-3 and ProPhas:amiLPAT1-4

To suppress the expression of LPAT1 during seed development, amiRNA was designed to specifically target LPAT1 which is driven by the seed-specific ProPhas. First, two amiRNA sequences were designed (5'-TAAATGCTTATGGACGCCCAA-3' and 5'-TTATTGGAACTACTGCAGCTC-3') using Web microRNA designer (http://wmd3.weigelworld.org) to create amiLPAT1-3 and amiLPAT1-4, respectively, from the designed sequences. These two amiRNA sequences were assembled using pRS300 as a template plasmid. To express the amiRNA sequences by ProPhas, the XbaI site was added to pNB030 using PCR site-directed mutagenesis with primer NB104 to create pNB031 (pENTR_ProPhas-XbaI). The two assembled amiRNA precursor fragments were ligated into XhoI and XbaI sites of pNB031 to create pNB032 (pENTR_ProPhas:amiLPAT1-3) and pNB033 (pENTR_ProPhas:amiLPAT1-4). Lastly, pNB032 and pNB033 were recombined into a pBGW destination vector to create pNB034 (pBGW_ProPhas:amiLPAT1-3) and pNB035 (pBGW_ProPhas:amiLPAT1-4), which were transduced into Arabidopsis WT (Col-0) plants using Agrobacterium-mediated transformation. Twenty-four plants were screened, and two representative lines were selected: ProPhas:amiLPAT1-3 (#15, #24) and ProPhas:amiLPAT1-4 (#9, #13).

Lipid analysis

Total lipid was extracted from 24-day-old rosette leaves and mature siliques as described (Bligh and Dyer, 1959; Nakamura et al., 2014a). For the polar glycerolipid fraction, the total lipid extract was separated on a two-dimensional silica gel TLC plate with a solvent system of chloroform:methanol:aqueous ammonia=120:80:8 (v/v) for the first development, followed by the second development with the solvent system of chloroform:methanol:acetic acid:water=170:20:15:3 (v/v). As for the TAG fraction, the total lipid was separated on one-dimensional TLC plates with the solvent system of hexane:diethylether:acetic acid=160:40:4 (v/v).

To visualize the lipid spots, we sprayed the TLC plate with primuline solution. The spots were scraped off, hydrolyzed, and methylesterified with HCl-methanol in the presence of pentadecanoic acid (C15:0) as an internal standard. The fatty acid methyl esters of each lipid were quantified using a gas chromatograph (GC-2030; Shimadzu) equipped with a ULBON HR-SS-10 column (Shinwa Chemical Industries). All lipid data are the mean ±SD of at least three biologically independent samples.

RNA extraction and reverse transcription–PCR

Total RNA was isolated from 24-day-old rosette leaves and young siliques using the SV total RNA isolation system (Promega, Madison, WI, USA). For mature siliques, a pre-treatment was performed before RNA extraction as described (Meng and Feldman, 2010). Briefly, 100 mg of fresh mature seeds were ground with liquid N₂ to a fine powder, immediately suspended in 1 ml of extraction buffer [100 mM Tris–HCl (pH 9.5), 150 mM NaCl, 1% (w/v) sarkosyl (Sigma, L5125), and 5 mM DTT] and the mixture was transferred to a 2 ml microtube. The homogenate was vortexed for 5 min at room temperature and spun down. The supernatant was collected and mixed with 500 μl of chloroform into a new 2 ml microtube. Then, 500 μl of water-saturated acidic phenol (pH 4.3, Sigma, P4682) was added, mixed by vortexing for 2 min, and centrifuged at 11 000 g for 15 min at room temperature. The upper phase of solution was transferred to a new microtube and mixed with 90 μl of 3 M sodium acetate and 600 μl of 2-propanol by inverting the microtube, and centrifuged at 11 000 g for 10 min at room temperature. The pellet was washed with 1 ml of 75% (v/v) ethanol. The RNA remaining in the washed pellet was further extracted by using the SV total RNA isolation system (Promega). cDNA was synthesized using the SuperScript III first-strand synthesis kit (Invitrogen). Quantitative reverse transcription–PCR (RT–qPCR) was performed as previously described (Lin et al., 2015). Briefly, the analysis was performed using the 7500 Real-Time PCR System (Applied Biosystems) with the comparative threshold cycle method to determine relative gene expression. The expression of ACTIN2 (KK129/KK130) was used as an internal control (Barroga and Nakamura, 2022). Data are the mean ±SD from three biological replicates, with three technical replicates. The primer sets for RT–qPCR were reported previously (Nakamura et al., 2014b). The oligonucleotide sequences of primer sets for RT–qPCR are given in Supplementary Table S1.

β-Glucuronidase staining

GUS expression of ProLPAT1:LPAT1-GUS was observed at various growth stages and in different organs, as described (Nguyen and Nakamura, 2023). Briefly, fresh samples were harvested then immediately immersed in pre-cooled 90% (v/v) acetone for 15 min, followed by immersion in GUS staining solution. After incubation in the dark at 37 °C for 3–4 h, staining was stopped by replacing the solution with 70% ethanol. For pigmented tissues, pigments were removed by immersing the tissues in ethanol:acetic acid=6:1 (v/v). The images were obtained using a stereo microscope (Stemi 508, Zeiss, Germany).

Confocal laser scanning microscopy

For observation of 3×Venus fluorescence in the embryos of ProLPAT1:LPAT1-Ven lpat1-1/– transgenic plants at different developmental stages, embryo samples were cleared using clearance solution [chloral hydrate:glycerol:water=8:2:1 (v/v)] prior to confocal microscope observation. Fluorescence of 3×Venus was observed under a microscope (Leica TCS SP8 WLL FALCON; Wetzlar, Germany) equipped with HC PL APO ×1/0.4 CS2, HC PL APO ×20/0.75 CS2, HC PL APO ×40/1,1 W CORR CS2. Images were captured by use of the Leica TCS SP8 with filters for Venus (514 nm laser, 520–555 nm band-pass) and for chlorophyll autofluorescence (488 nm laser, 650 nm long-pass).

Microscopy analysis of seeds and seedlings

For imaging the shoot, 22-day-old seedlings of WT, ProPhas:LPAT1-3 WT (#15; #24), and ProPhas:LPAT1-4 (#9; #13) were used. Images of the shoot were acquired with a Lumix DC-TZ90 digital camera. Shoot area was measured using Python script. The final values were obtained from 11 images in each genotype.

For the seed observation, mature seeds of WT and ProPhas:LPAT1 knockdown lines were observed by LSM 880 Airyscan FAST (Zeiss, Oberkochen, Germany) equipped with Plan-Apochromat ×10/0.45-NA, Plan-Apochromat ×20/0.8-NA, and differential interference contrast (DIC) optics. For the seed area analysis, images were acquired with a stereo microscope (Stemi 508, Zeiss, Germany) and measured using ImageJ software (https://imagej.net/imaging/particle-analysis).

Results

The lethal phenotype of lpat1-1/– is associated with a female gametophytic defect

The lpat1-1/– mutant is not viable due to an embryonic developmental defect (Kim and Huang, 2004; Yu et al., 2004). To further investigate the reproductive developmental defect of lpat1-1/–, a reciprocal genetic crossing was performed and segregation of the mutant allele was examined by PCR-based genotyping. Crossing of WT pollen grains to the pistil of lpat1-1/+ produced a total of 48 offspring—31 WT and 17 lpat1-1/+, giving a ratio of WT and lpat1-1/+ of 1.8:1 instead of 1:1 (Table 1). Conversely, crossing pollen grains of lpat1-1/+ with the pistil of the WT gave a nearly 1:1 ratio of WT and lpat1-1/+ (23:25 out of the total of 48). These results demonstrate that the transmission of the lpat1-1 allele from the female to male gametophyte is significantly lower than that from the male to female, suggesting that the lpat1-1 mutant may have a partial female gametophyte defect.

Expression of LPAT1 in reproductive tissues

Among five LPAT isoforms in Arabidopsis, only LPAT1 and LPAT2 are essential for the reproductive processes due to the embryonic- and gametophytic-lethal phenotypes of the knockout mutants (Kim and Huang, 2004; Yu et al., 2004; Kim et al., 2005) We compared the expression pattern of the five LPAT isoforms in the reproductive organs using the Arabidopsis eFP browser. As shown in Fig. 1A, LPAT1, LPAT2, and LPAT5 showed preferential expression in stigmas and ovaries as well as developing embryos, compared with LPAT3 which is pollen specific. LPAT4 did not show preferential expression pattern in these organs. Next, to examine the tissue expression pattern of LPAT1 in reproductive organs, we created an LPAT1–GUS translational reporter construct driven by its native promoter in the lpat1-1/– genetic background (ProLPAT1:LPAT1-GUS lpat1-1/–). Histochemical GUS staining assay showed that LPAT1–GUS was highly expressed in stigma (Fig. 1C), anther filament (Fig. 1D), and anthers of young floral buds (Fig. 1E), but no clear GUS stain was observed in siliques (Fig. 1F). A magnified view of an anther showed that pollen grains exhibited an intense LPAT1–GUS stain (Fig. 1G). These tissue expression patterns were consistently observed in another transgenic ProLPAT1:LPAT1-GUS lpat1-1/– line (Supplementary Fig. S1).

To observe the expression pattern of LPAT1 during embryogenesis, we created an LPAT1–Ven fluorescent reporter system (ProLPAT1:LPAT1-Ven lpat1-1/–) since the GUS signal was not clearly observed (Fig. 1F). During embryo development, LPAT1–Ven was uniformly expressed throughout the embryo, with slightly higher intensity in the peripheral part in the heart stage and tips of cotyledons in the torpedo stage (Fig. 2A). In mature embryos, LPAT1–Ven signal co-localized with the periphery of chlorophyll autofluorescence (Fig. 2B), which suggests that LPAT1 is localized at the chloroplast envelope during embryo development. Thus, LPAT1 is expressed in reproductive organs including pollen grains and developing embryos.

Targeted suppression of LPAT1 in developing seeds produced short siliques with aborted seeds

To examine the specific role of LPAT1 during seed development, we created transgenic plants that express amiRNA targeted to LPAT1 driven by a seed-specific β-phaseolin promoter (ProPhas) (Sengupta-Gopalan et al., 1985; Karchi et al., 1994; Chandrasekharan et al., 2003). We tested the seed-specific expression pattern of ProPhas by histochemical GUS staining of the plants harboring the ProPhas:GUS reporter construct, which showed intense GUS staining in the embryos (Supplementary Fig. S2A) and germinating seedlings (Supplementary Fig. S2B), but not the subsequent vegetative growth stage (Supplementary Fig. S2C–E). Using the ProPhas promoter, we expressed two different amiRNA sequences (amiLPAT1-3 and amiLPAT1-4) that are specifically designed to target LPAT1. In both transgenic constructs, the majority of transgenic plants screened from 24 independent T₁ lines showed shorter siliques with aborted seeds as represented in ProPhas:amiLPAT1-3 #15, #24, and ProPhas:amiLPAT1-4 #9, #13 lines (Fig. 3A). Consistent with the seed-specific expression of ProPhas:GUS, no significant morphological phenotype was observed in vegetative growth (Supplementary Fig. S3A, B). In siliques of the transgenic plants, quantitative measurement showed a significant decrease in silique length and increase in the percentage of aborted seeds compared with the WT, except for ProPhas:amiLPAT1-4 #9 showing the weakest phenotypic effect (Fig. 3B). This observation indicates that these phenotypes were caused by the suppression of LPAT1 during seed development. To confirm whether the LPAT1 transcript level was suppressed in siliques, we performed RT–qPCR using the cDNA of rosette leaves, and young and mature siliques (Fig. 3C). In rosette leaves, these four transgenic lines did not show a significant reduction; while, the transcript levels of LPAT1 were decreased in mature siliques. Although #9 and #13 showed a similar decrease in the transcript levels, #13 showed a much stronger silique phenotype than #9, indicating a somewhat unclear correlation between the transcript level and phenotype expression between these two lines. Thus, the transgenic lines specifically suppressed LPAT1 in mature siliques, and seed-specific suppression of LPAT1 is sufficient to produce the reproductive defect.

Siliques of ProPhas:amiLPAT1 plants affected phosphatidylcholine contents and polyunsaturation levels

To assess the effect of LPAT1 suppression by the β-phaseolin promoter, we analyzed the polar glycerolipid contents in mature siliques of the WT and four lines of ProPhas:amiLPAT1 plants (ProPhas:amiLPAT1-3 #15, #24 and ProPhas:amiLPAT1-4 #9, #13). A consistent decrease in PC content was observed among the four lines (Fig. 4A), which was also observed in terms of mol% (Fig. 4B). Although LPAT1 is assumed to be involved in the plastid-localized glycerolipid biosynthesis, the amount of plastid lipids such as MGDG, DGDG, PG, and SQDG did not show a consistent decrease in these four lines. To confirm whether seed PC content was decreased, we analyzed the amount of PC in dry seeds of the WT and four lines of ProPhas:amiLPAT1 plants. As shown in Supplementary Fig. S4, these four suppression lines showed a significant decrease in PC contents.

We next investigated the fatty acid composition of the glycerolipid classes analyzed in Fig. 4. The composition of linoleic acid (18:2) was consistently decreased in the extraplastidic phospholipid classes (PC, PE, and PI) with the concomitant increase in 18:3 (Fig. 5). MGDG showed a slight but consistent decrease in 18:3 but not in 16:3. Since a subtle GUS stain was observed in leaves of the transgenic plant harboring ProPhas:GUS (Supplementary Fig. S2), we examined the possible leaky effect of LPAT1 suppression in leaves by analyzing the polar glycerolipid contents. As shown in Supplementary Fig. S5, MGDG levels demonstrated a slight but consistent decrease among the four lines, although the rosette leaves did not display any pale-green phenotype (Supplementary Fig. S3A). This suggests that a slight expression of the β-phaseolin promoter in leaves suppresses LPAT1 and thus decreases MGDG levels due to its commitment to plastid-localized glycerolipid biosynthesis, while in siliques LPAT1 suppression primarily affects the extraplastidic lipid class PC.

Seed-specific suppression of LPAT1 altered triacylglycerol composition and seed size

Since PC is involved in TAG production in developing seeds, we investigated the phenotype of mature seeds in the four lines of ProPhas:amiLPAT1 transgenic plants. The morphology of mature seeds appeared normal (Fig. 6A) except that the two lines with a stronger silique phenotype (ProPhas:amiLPAT1-3 #15 and ProPhas:amiLPAT1-4 #13; Fig. 3B) showed slightly larger seed size. The quantitative measurement of seed weight (Fig. 6B) and size (Fig. 6C) indicated that these two lines are significantly heavier and larger compared with the WT. To test whether TAG content is altered in these seeds, we analyzed TAG in the four lines of ProPhas:amiLPAT1 plants. No consistent increase was observed in TAG content among the four lines (Fig. 6D); however, the analysis of fatty acid composition in TAG showed a consistent decrease in 18:3 (Fig. 6E). Thus, in developing seeds, LPAT1 is involved in PC contents and the acyl quality of the accumulated TAG.

Discussion

The present study investigated the seed-specific role of LPAT1 in reproductive development and glycerolipid biosynthesis. In Arabidopsis, the de novo glycerolipid biosynthesis pathway (or Kennedy pathway) exists in both the plastid and ER, and LPAT1 is assumed to be responsible for the plastid-localized lysophosphatidic acid acyltransferase activity for the second reaction of the Kennedy pathway. Glycerolipids synthesized in the plastids are important for photosynthesis because mutants defective in synthesizing MGDG or PG showed a pale phenotype with a significant photosynthetic defect (Hagio et al., 2000; Kobayashi et al., 2007). However, the knockout mutant of LPAT1 caused embryonic lethality (Kim and Huang, 2004; Yu et al., 2004), which hampered the in-depth functional study of the knockout mutant plants. Here, we used an approach to suppress LPAT1 specifically in seeds and demonstrated that plastid-localized LPAT1 is required for reproductive growth and involved in the synthesis of extraplastidic lipids, PC and TAG. The results suggest the importance of the plastid-localized glycerolipid biosynthesis pathway in the synthesis of extraplastidic lipids.

Role of LPAT1 in reproductive development

The result of reciprocal genetic crossing indicated reduced penetration of the lpat1-1 mutant allele from the female to male gametophyte, but not vice versa (Table 1). This result suggests that mutation of LPAT1 may cause a partial defect in the female gametophyte. Indeed, ProPhas:amiLPAT1 lines produced a shorter silique with a number of aborted seeds, indicating reduced gametophyte viability (Fig. 3A). A similar reduction in silique length was observed by the suppression of ACT1/ATS1, which catalyzes the reaction one step before LPAT1 (Xu et al., 2006). Although the morphology and viability of this mutant line have not been reported, the shorter silique length is indicative of a gametophytic defect. Thus, the function of LPAT1 may be crucial for the female gametophyte and the plastidic Kennedy pathway may be committed to gametogenesis.

The similarity in the embryonic lethal phenotypes observed in the siliques of lpat1-1/+ (Kim and Huang, 2004; Yu et al., 2004), mgd1-2/+ (Kobayashi et al., 2007), and tgd1-1/+ (Xu et al., 2005) suggests that galactolipid biosynthesis may be essential for photosynthesis during embryogenesis. Embryogenesis is arrested at the heart stage in both mutants, so this appears to be a critical developmental stage when galactolipid biosynthesis is essential. Interestingly, such a gametophytic phenotype has not been observed in the mutants of DGDG synthases as the double mutant dgd1dgd2 still retains seed germination viability despite dwarfed vegetative growth (Kelly et al., 2003). Although the residual DGDG synthesis by the processive galactosyltransferase SENSITIVE TO FREEZING 2 (SFR2) (Moellering et al., 2010) may account for the embryonic viability, it is possible that MGDG and DGDG may play a distinct role during embryogenesis.

Role of LPAT1 in glycerolipid metabolism

Since plastid glycerolipids are tightly associated with the photosynthetic function, the role of LPAT1 in glycerolipid metabolism in photosynthetic tissues has been discussed (Kim et al., 2010). Although the lpat1-1/– mutant shows an embryonic-lethal phenotype, how LPAT1 suppression affects glycerolipid metabolism during the reproductive process remains elusive. We showed that seed-specific suppression of LPAT1 altered PC but not plastid glycerolipids (Fig. 2B). This effect is distinct from the suppression of LPAT2, a major extraplastidic isoform of LPAT involved in de novo glycerolipid biosynthesis (Barroga and Nakamura, 2022), which decreased PC content in mature flowers but not siliques. LPAT1 is localized at the chloroplast envelope (Yu et al., 2004; Dubots et al 2012), so the substrate and products of LPAT1 appear to be distant from the ER-localized glycerolipid biosynthesis pathway. However, a recent report showed that a chloroplastic PA phosphatase LPP epsilon 1 (LPPɛ1) contributes to ER glycerolipid biosynthesis (Nguyen and Nakamura, 2023), implying a possible route of providing plastidic glycerolipid precursors to the ER. Suppression of chloroplast-localized LPPɛ1 decreased the PC content (Nguyen and Nakamura, 2023). Although the mechanism by which plastidic LPAT1 and LPPɛ1 affect extraplastidic PC content awaits further investigation, our observation suggests that in siliques, PA produced by LPAT1 at the stromal side of the inner envelope may be relocated to the outer envelope and serves as a substrate of LPPɛ1, which converts PA to DAG as the precursor of ER-localized glycerolipid biosynthesis. Of note, the tgd1-1 mutant has an increased PC content and contains 16:3, an acyl species exclusively produced in the chloroplasts (Xu et al., 2006; Yang et al., 2017), which indicates that a part of PC biosynthesis uses a chloroplast-derived precursor. Since the predominant lipid metabolic flux in the developing seeds is toward ER-localized TAG biosynthesis, it is possible that the plastid-localized glycerolipid biosynthesis pathway may have more commitment to the extraplastidic pathway in the developing seeds, which forms an additional organ-specific metabolic pathway that utilizes the plastidic pool of acyl lipids for lipid biosynthesis in the ER. It should be noted that the decreases of polyunsaturated fatty acids in the suppression lines (18:3 in MGDG and TAG; 18:2 in PC and PE) were statistically significant but small, which could be due either to the leaky effect of LPAT1 suppression or the limited contribution of this pathway to the overall acyl composition of glycerolipids. How this putative new route contributes to the overall acyl composition remains to be investigated.

In conclusion, our data indicate that LPAT1 plays a role in reproductive development and ER-localized glycerolipid biosynthesis in siliques and seeds. Further investigation is anticipated to uncover the commitment of the plastid glycerolipid biosynthesis pathway in non-photosynthetic organs.

Supplementary data

The following supplementary data are available at JXB online.

Fig. S1. Histochemical GUS staining observation of ProLPAT1:LPAT1-GUS lpat1-1/– line #20.

Fig. S2. Histochemical GUS staining observation of the transgenic Arabidopsis plants harboring the ProPhas:GUS reporter construct.

Fig. S3. Vegetative growth of ProPhas:amiLPAT1 knockdown lines.

Fig. S4. Phosphatidylcholine (PC) content in dry seeds of the wild type (WT) and transgenic plants ProPhas:amiLPAT1-3 WT (#15 and #24) and ProPhas:amiLPAT1-4 WT (#9 and #13).

Fig. S5. Polar glycerolipid profiles in 24-day-old rosette leaves of the wild type (WT), ProPhas:amiLPAT1-3 WT (#15 and #24), and ProPhas:amiLPAT1-4 WT (#9 and #13) transgenic plants.

Table S1. List of oligonucleotide sequences used in this study.

Abbreviations:

Abbreviations:
 
• DGDG

  digalactosyldiacylglycerol

 
• LPAT

  lysophosphatidic acid acyltransferase

 
• MGDG

  monogalactosyldiacylglycerol

 
• PA

  phosphatidic acid

 
• PC

  phosphatidylcholine

 
• PE

  phosphatidylethanolamine

 
• PG

  phosphatidylglycerol

 
• PI

  phosphatidylinositol

 
• SQDG

  sulfoquinovosyldiacylglycerol

 
• TAG

  triacylglycerol.

Authors contributions

NB and VN: performed experiments, data analysis, and edited the manuscript; YN: conceptualization, supervision of the experiments, and drafted/finalized the manuscript. All authors checked the manuscript and approved the contents.

Conflict of interest

The authors declare that they have no conflicts of interest.

Funding

This work was supported by GteX Program Japan Grant Number JPMJGX23B0.

Data availability

All data supporting the findings of this study are available within the paper, within its supplementary data published online, and from the corresponding author, Yuki Nakamura, upon request.

References

Author notes