Anionic lipids facilitate membrane development and protochlorophyllide biosynthesis in etioplasts

Abstract

Dark-germinated angiosperm seedlings develop chloroplast precursors called etioplasts in cotyledon cells. Etioplasts develop lattice membrane structures called prolamellar bodies (PLBs), where the chlorophyll intermediate protochlorophyllide (Pchlide) forms a ternary complex with NADPH and light-dependent NADPH:Pchlide oxidoreductase (LPOR). The lipid bilayers of etioplast membranes are mainly composed of galactolipids, which play important roles in membrane-associated processes in etioplasts. Although etioplast membranes also contain 2 anionic lipids, phosphatidylglycerol (PG) and sulfoquinovosyldiacylglycerol (SQDG), their roles are unknown. To determine the roles of PG and SQDG in etioplast development, we characterized etiolated Arabidopsis (Arabidopsis thaliana) mutants deficient in PG and SQDG biosynthesis. A partial deficiency in PG biosynthesis loosened the lattice structure of PLBs and impaired the insertion of Mg²⁺ into protoporphyrin IX, leading to a substantial decrease in Pchlide content. Although a complete lack of SQDG biosynthesis did not notably affect PLB formation and Pchlide biosynthesis, lack of SQDG in addition to partial PG deficiency strongly impaired these processes. These results suggested that PG is required for PLB formation and Pchlide biosynthesis, whereas SQDG plays an auxiliary role in these processes. Notably, PG deficiency and lack of SQDG oppositely affected the dynamics of LPOR complexes after photoconversion, suggesting different involvements of PG and SQDG in LPOR complex organization. Our data demonstrate pleiotropic roles of anionic lipids in etioplast development.

Negatively charged phospholipid and sulfur-containing glycolipids facilitate internal membrane formation and synthesis of chlorophyll intermediates in chloroplast precursors of dark-grown seedlings.

Introduction

In plants, chlorophyll (Chl) is synthesized within plastids in coordination with the development of photosynthetic machinery (Wang and Grimm 2021). Chl biosynthesis begins with the formation of 5-aminolevulinic acid (ALA) from glutamyl-tRNA^Gul, which is the rate-limiting step for the regulation of the entire pathway (Tanaka et al. 2011). ALA is then converted to protoporphyrin IX (Proto IX), the last common intermediate of heme and Chl biosynthesis, through multiple enzymatic steps. For Chl biosynthesis, Mg²⁺ is inserted into Proto IX by Mg-chelatase to yield Mg-Proto IX, which is further converted to Mg-Proto IX monomethylester (Mg-Proto IX ME) by S-adenosyl-L-Met:Mg-Proto IX methyltransferase (MgMT) and to protochlorophyllide (Pchlide) by Mg-Proto IX ME cyclase (MgCY). Pchlide is then reduced and phytylated, resulting in Chls.

In cyanobacteria and plants, there are 2 types of Pchlide oxidoreductase (POR) for converting Pchlide to chlorophyllide (Chlide), the immediate precursor of Chl (Masuda 2008). Whereas dark-operative POR can reduce Pchlide to Chlide in the absence of light, light-dependent POR (LPOR) absolutely requires light for its reaction. Angiosperms have only LPOR, so they accumulate Pchlide in plastids called etioplasts when germinated in the dark (Solymosi and Schoefs 2010). Etioplasts are chloroplast precursors containing 3D lattice membrane structures called prolamellar bodies (PLBs) and flattened lamellar prothylakoids (PTs). Pchlide in etioplasts forms ternary complexes with LPOR and NADPH, most of which further oligomerize and cover the membrane tubules of PLBs in a helical fashion (Floris and Kühlbrandt 2021; Nguyen et al. 2021). The Pchlide bound by LPOR at the active site (photoactive Pchlide) is immediately converted to Chlide with flash illumination, while the rest of the Pchlide (nonphotoactive Pchlide) remains unchanged (Schoefs and Franck 2003). The nonphotoactive and photoactive Pchlides can be distinguished by their different fluorescence spectra at 77 K with emission peaks at around 633 and 655 nm, respectively (Schoefs 2001). The photoconversion of Pchlide by LPOR in the oligomers results in the oligomeric Chlide–LPOR–NADP⁺ complexes, which emit a fluorescence band peaking around 690 nm at 77 K (Boddi et al. 1990). Then, NADP⁺ is replaced by NADPH, and the oligomeric complexes are dissociated into dimeric forms (Aziz Ouazzani Chahdi et al. 1998). These processes are observed as a gradual shift (Shibata shift) of emission peaks from 690 to 680 nm at 77 K (Shibata 1957; Smeller et al. 2003; Solymosi et al. 2007). In addition, some oligomeric Chlide–LPOR–NADP⁺ complexes exchange Chlide for Pchlide prior to the replacement of NADP⁺ to rapidly regenerate the oligomeric Pchlide–LPOR–NADPH complexes (Franck et al. 1999).

When dark-germinated seedlings are continuously illuminated, PLBs are transformed to the thylakoid membrane and etioplasts differentiate to chloroplasts (Kowalewska et al. 2016). Although PLBs greatly differ from the thylakoid membrane in morphology, the lipid compositions of these membranes are similar, with 2 galactolipids monogalactosyldiacylglycerol (MGDG) and digalactosyldiacylglycerol (DGDG) accounting for ∼80 mol% of total membrane lipids in both membranes (Fujii et al. 2019). In addition, sulfoquinovosyldiacylglycerol (SQDG) and phosphatidylglycerol (PG), which are anionic lipids with a negative charge in their polar head groups, make up the remaining 20 mol%. In the thylakoid membrane of chloroplasts, these lipids provide the fluid matrix for the photosynthetic complexes and prevent free diffusion of protons and other ions across the membrane. In addition, a number of lipid molecules are embedded in the structure of photosynthetic complexes including PSI and PSII and would play essential roles in the photochemical and electron transport reactions (Yoshihara and Kobayashi 2022). Particularly, the necessity of PG for photosynthesis is evident as loss of the PG in the thylakoid membrane severely disrupts photosynthetic reactions in both cyanobacteria and plants (Kobayashi et al. 2016).

The importance of galactolipids in etioplasts was revealed by the analyses of Arabidopsis (Arabidopsis thaliana) mutants deficient in MGDG or DGDG biosynthesis. A 36% decrease in MGDG content impaired Pchlide biosynthesis and the formation and oligomerization of the Pchlide–LPOR–NADPH ternary complex, while only slightly affecting the lattice structure of PLBs (Fujii et al. 2017). This is consistent with the data showing that MGDG mediates the oligomerization of the ternary complex in vitro (Gabruk et al. 2017). Similarly, an 80% decrease in DGDG content impaired the Pchlide biosynthesis and the formation of the ternary complex, although it did not affect the oligomerization of the ternary complex (Fujii et al. 2018). Moreover, the DGDG deficiency slowed the regeneration of the Pchlide–LPOR–NADPH complex from the Chlide–LPOR–NADP⁺ complex and severely disrupted the lattice structure of PLBs and the development of PTs. These data indicate a wide involvement of galactolipids in the membrane-associated processes during etioplast development. By contrast, roles of PG and SQDG in Pchlide accumulation, organization of the pigment–LPOR complexes, and internal membrane biogenesis during etioplast development remain unknown.

To address how PG and SQDG are involved in the processes of etioplast development in dark-germinated A. thaliana, we characterized mutants deficient in PG and SQDG biosynthesis. Phosphatidylglycerophosphate synthase 1 (PGP1) is the enzyme required for PG biosynthesis in plastids and mitochondria. The pgp1-1 mutant, which carries a single amino acid substitution (P170S) in PGP1, has an 80% reduction in PGP1 activity and a 30% reduction in total PG content from the wild-type level, resulting in decreased Chl content and photosynthetic activity in leaves of light-grown plants (Xu et al. 2002). UDP-sulfoquinovose synthase (SQD1) and SQDG synthase (SQD2) are essential for SQDG biosynthesis, and knockout mutants of SQD1 (sqd1) or SQD2 (sqd2-2) with T-DNA insertions completely lack SQDG (Okazaki et al. 2009, 2013). Although both sqd1 and sqd2-2 grew similar to wild type under nutrient-sufficient conditions, these mutants showed decreased photosynthetic activity when the PG content was decreased under phosphate-starved conditions (Yoshihara et al. 2021). Furthermore, the sqd1 pgp1-1 and sqd2-2 pgp1-1 double mutants showed stronger impairments in photosynthesis and growth than the parental single mutants, indicating that PG and SQDG substitute each other (Yoshihara et al. 2021), as they share common features of anionic lipids (Yu and Benning 2003). Of note, indistinguishable phenotypes between the sqd1 pgp1-1 and sqd2-2 pgp1-1 double mutants indicate that another anionic glycolipid glucuronosyldiacylglycerol (GlcADG), whose synthesis requires SQD2 but not SQD1, does not compensate for the loss of PG in A. thaliana (Okazaki et al. 2013; Yoshihara et al. 2021).

In this study, we investigated the membrane structures of etioplasts, the biosynthesis and accumulation of Pchlide, and the functions of the Pchlide–LPOR–NADPH complex in pgp1-1, sqd1, and the sqd1 pgp1-1 double mutants. Some data were also obtained from sqd2-2 and the sqd2-2 pgp1-1 double mutant to corroborate the role of SQDG in etioplasts.

Results

Changes in lipid composition in etiolated seedlings of the anionic lipid mutants

To reveal how the pgp1-1 and sqd1 mutations affect the lipid metabolism in etiolated seedlings, we determined the lipid composition in 4-d-old seedlings of wild type, sqd1, pgp1-1, and sqd1 pgp1-1 grown under continuous darkness (Fig. 1). In sqd1 and sqd1 pgp1-1 mutants, no SQDG was detected, consistent with the previous report that the knockout mutation of SQD1 causes complete loss of SQDG (Okazaki et al. 2009). Although the mean values of the proportion of PG were ∼25% lower in pgp1-1 and sqd1 pgp1-1 and 36% higher in sqd1 than that in wild type, these differences were not statistically significant. When the total anionic lipid content (SQDG + PG) was compared between wild type and the mutants, only the sqd1 pgp1-1 double mutant showed a significant decrease (Supplemental Fig. S1A). Meanwhile, the proportion of both MGDG and DGDG was not significantly changed in all mutant lines (Fig. 1). In addition, there was no significant difference in the total amount of plastidic lipids (MGDG + DGDG + SQDG + PG) between the wild type and the mutants (Supplemental Fig. S1B).

Figure 1.

Proportion of major plastidic lipids in total membrane glycerolipids in 4-d-old etiolated seedlings of wild-type (WT) and anionic lipid mutants. Data are means ± Se from 3 independent experiments. The lowercase letters indicate statistically significant differences between each line (P < 0.05, 1-way ANOVA and Tukey's post hoc honestly significant difference test). MGDG, monogalactosyldiacylglycerol; DGDG, digalactosyldiacylglycerol; SQDG, sulfoquinovosyldiacylglycerol; PG, phosphatidylglycerol; ND, not detected.

In pgp1-1 and sqd1 pgp1-1, fatty acid compositions of galactolipids but not anionic lipids were significantly changed when compared with those in wild type. In both mutants, proportions of 16:3 in MGDG and 16:0 in DGDG were increased with decreased proportions of 18:3 in both lipids (Supplemental Fig. S2). Although some minor changes in fatty acid composition were observed in the mixture of other phospholipids (phosphatidylcholine + phosphatidylethanolamine + phosphatidylinositol) between each plant, the patterns of the changes were greatly different from those in galactolipids.

Disordered membrane structures of etioplasts in anionic lipid mutants

To assess whether anionic lipids contribute to the membrane organization in etioplasts, we observed the ultrastructure of etioplasts in cotyledons of wild-type and the anionic lipid mutants (Fig. 2). The wild-type etioplasts developed highly regular lattice membrane structures of PLBs and long PTs connected to the PLBs (Fig. 2, A to D). The membrane structures of the sqd1 etioplasts were indistinguishable from those of wild type. By contrast, the PLB network was loosened in the pgp1-1 etioplasts, which resulted in decreased membrane area in PLBs when compared with those in wild type and sqd1 (Fig. 2E). Disorder of the PLB membrane structure was further enhanced in the etioplasts of the sqd1 pgp1-1 double mutant. In the double mutant, some etioplasts showed loose aggregates of filamentous membrane tubules in stromal region, although some showed electron-dense aggregates of small membranes. In addition, the sqd1 pgp1-1 double mutant frequently showed etioplast sections that had almost no PLB and PT membranes (Fig. 2F), which were rarely observed in wild-type and single mutants. Meanwhile, the double mutant showed no significant differences in the circularity (Fig. 2G) and size of etioplasts (Fig. 2H) compared with the wild-type and the single mutants. We also observed ultrastructures of etioplasts in sqd2-2 and sqd2-2 pgp1-1 mutants (Supplemental Fig. S3). The sqd2-2 pgp1-1 etioplasts showed disordered PLB structures similar to sqd1 pgp1-1, whereas the sqd2-2 etioplasts were comparable with those in wild type and sqd1.

Figure 2.

Structure analysis of cotyledon etioplasts in 4-d-old etiolated seedlings of wild-type (WT) and anionic lipid mutants. Ultrastructures of cells A), etioplasts B and C), and internal membranes of etioplasts D) in WT, sqd1, pgp1-1, and sqd1 pgp1-1. The images in A) to C) are high-resolution montages constructed from multiple images. In D), 2 different types of internal membrane structures are shown for the sqd1 pgp1-1 etioplasts. Bars = 2 μm in A) 500 nm in B and C) and 100 nm in D). E) Proportion of membrane area in PLB (n = 15). F) Proportion of etioplasts lacking internal membranes (n = 60). G) Etioplast circularity (n = 60). H) Area of etioplasts (n = 60). In E G, and H) dots show the raw data, the width of the violin plots represents the density of the data, and horizontal bars indicate median values. The lowercase letters indicate statistically significant differences between each line (P < 0.05, 1-way ANOVA, and Tukey's post hoc honestly significant difference test). Asterisks (***) in F) indicate a significant difference from the WT (P < 0.001, Pearson’s χ² test). n/d, not determined; ND, not detected; PLB, prolamellar body.

Because the LPOR protein is involved in the development of PLBs (Sperling et al. 1998; Franck et al. 2000), we determined LPOR levels in etiolated wild-type and the anionic lipid mutants by using polyclonal anti-LPOR antibodies that react with all Arabidopsis LPOR isoforms (Masuda et al. 2003). The sqd1 pgp1-1 and sqd2-2 pgp1-1 double mutants showed ∼60% decreases in the LPOR level compared with the wild type (Fig. 3). By contrast, no significant change in LPOR levels was detected in the sqd1, sqd2-2, and pgp1-1 mutants.

Figure 3.

LPOR levels in 4-d-old etiolated seedlings of wild-type (WT) and anionic lipid mutants. A) Immunodetection of total LPOR proteins (∼37 kDa) in 4-d-old etiolated seedlings of WT and anionic lipid mutants. Coomassie brilliant blue staining of the same protein samples is shown as a loading control. B) Protein levels of LPOR in the anionic lipid mutants relative to the WT level. Data are means ± Se from 4 independent experiments. Asterisks indicate statistically significant differences from the WT level (P < 0.05, Student's t-test).

Decreased Pchlide content in etiolated seedlings of anionic lipid mutants

To determine the role of PG and SQDG in Pchlide accumulation, we measured Pchlide content in 4-d-old etiolated wild-type and the anionic lipid mutants (Fig. 4). Total Pchlide content in sqd1 and sqd2-2 was similar to and slightly higher than that in the wild type, respectively. By contrast, total Pchlide content decreased by 41% in pgp1-1 and by ∼70% in the sqd1 pgp1-1 and sqd2-2 pgp1-1 mutants compared with the wild-type level. The amounts of Pchlide remaining after a 0.7 ms flash, which corresponds to nonphotoactive Pchlide, also decreased by 52% in pgp1-1 and by ∼70% in both double mutants from the wild-type level. As a result, the ratio of nonphotoactive Pchlide to total Pchlide was not notably changed between wild-type and these mutants (Fig. 4B).

Figure 4.

Pigment content in 4-d-old etiolated seedlings of wild-type (WT) and anionic lipid mutants. A) Content of total and nonphotoactive protochlorophyllides (Pchlides), which were extracted before and after flash treatment, respectively. Data are means ± Se from 16 independent experiments. The lowercase letters (a to d for total Pchlide and aʹ and bʹ for nonphotoactive Pchlide) indicate statistically significant differences between each line (P < 0.05, 1-way ANOVA, and Tukey's post hoc honestly significant difference test). B) Proportion of nonphotoactive Pchlide to total Pchlide. Data are represented as means ± Se from 16 independent experiments. No significant difference was observed among all plants (P > 0.05, Welch's t-test). C) Total carotenoid content. Data are means ± Se from 5 independent experiments. The lowercase letters indicate statistically significant differences between each line (P < 0.05, 1-way ANOVA, and Tukey's post hoc honestly significant difference test). D) Size of 4-d-old etiolated cotyledons. Data are represented as means ± Se from 196 cotyledons. The horizontal line in each box represents the median value of the distribution. The top and bottom of each box represent the upper and lower quartiles, respectively. The whiskers represent the range. The lowercase letters indicate statistically significant differences between each line (P < 0.05, 1-way ANOVA Tukey's post hoc honestly significant difference test).

We also determined total carotenoid levels in etiolated seedlings. The sqd1 pgp1-1 and sqd2-2 pgp1-1 mutants but not sqd1, sqd2-2, and pgp1-1 single mutants showed slight reductions in carotenoid content compared with the wild type (Fig. 4C). By contrast, the cotyledon size of the double mutants was comparable with that of wild type (Fig. 4D), suggesting that the seedling growth was not impaired in the mutants.

Impairment of the Pchlide biosynthesis pathway in anionic lipid mutants

To examine which step of the Pchlide biosynthesis pathway in etioplasts is impaired in anionic lipid mutants, we treated etiolated seedlings with ALA in darkness for 24 h to bypass the rate-limiting step of the Pchlide biosynthesis pathway and determined the accumulation levels of Pchlide and the porphyrin intermediates (Proto IX, Mg-Proto IX, and Mg-Proto IX ME). In wild type and sqd1, the ALA treatment did not cause remarkable accumulation of all 3 porphyrin intermediates (Fig. 5A). By contrast, pgp1-1 and sqd1 pgp1-1 showed strong accumulation of Proto IX without noticeable increases in Mg-Proto IX and Mg-Proto IX ME content, which suggests that the conversion of Proto IX to Mg-Proto IX is particularly impaired in these mutants. Furthermore, the sqd1 pgp1-1 mutant showed lower Pchlide accumulation even in the presence of ALA. The sqd2-2 pgp1-1 mutant also showed specific accumulation of Proto IX along with impaired accumulation of Pchlide in the presence of ALA (Supplemental Fig. S4).

Figure 5.

Altered protochlorophyllide (Pchlide) metabolism in anionic lipid mutants compared with wild type (WT). Accumulation levels of porphyrin pigments in 4-d-old etiolated seedlings treated with A) or without B) 10 mm 5-aminolevulinic acid for 24 h in the dark. Data are represented as means ± Se from 4 A) and 3 B) independent experiments. The lowercase letters indicate statistically significant differences between each line (P < 0.05, 1-way ANOVA, and Tukey's post hoc honestly significant difference test). ND, not detected; Proto IX, protoporphyrin IX; Mg-Proto IX ME, Mg-Proto IX monomethylester.

Then, we tested whether the pgp1-1 and sqd1 pgp1-1 mutants accumulate the porphyrin intermediates even in the absence of exogenous ALA. As in the wild type, only negligible amount of Proto IX was accumulated in both pgp1-1 and sqd1 pgp1-1, with Mg-Proto IX and Mg-Proto IX ME being undetectable (Fig. 5B).

To ascertain whether the impaired conversion of Proto IX to Mg-Proto IX in pgp1-1 and sqd1 pgp1-1 is caused at the transcription level, we investigated mRNA levels of CHLD, CHLH, and CHLI1 genes in etiolated seedlings (Fig. 6A). CHLD and CHLH encode the D and the H subunits of Mg-chelatase, respectively, whereas CHLI1 encodes the major isoform of the I subunit. We also examined the mRNA level of GENOMES UNCOUPLED 4 (GUN4), which encodes the GUN4 protein required for Mg-chelatase activity (Tanaka et al. 2011). The sqd1 and pgp1-1 mutants showed no significant difference in these gene expression levels compared with wild type. By contrast, the sqd1 pgp1-1 mutant showed lower mRNA levels of CHLH and GUN4 than the wild type. To test whether genes involved in Pchlide biosynthesis are globally downregulated in sqd1 pgp1-1, we examined the mRNA levels of HEMA1 encoding the major isoform of glutamyl-tRNA reductase, CHLM encoding MgMT, CHL27 encoding a membrane-bound subunit of MgCY, and PORA and PORB encoding major LPOR isoforms in etiolated seedlings (Fig. 6B). Among these tetrapyrrole biosynthesis genes, HEMA1 and PORA were significantly decreased in sqd1 pgp1-1. Although pgp1-1 also showed a decreasing trend in mRNA levels of these genes, the differences were not significant compared with the wild-type levels.

Figure 6.

Reverse transcription quantitative PCR analysis of the mRNA expression of genes involved in the tetrapyrrole biosynthesis pathway. The mRNA levels of A) genes involved in the enzymatic step of Mg insertion into protoporphyrin IX and B) other tetrapyrrole biosynthesis genes are presented as fold differences from the wild-type (WT) level after normalizing to the control genes ACTIN8 and UBIQUITIN11. A) Data are means ± Se from 9 (WT and sqd1), 7 (pgp1-1), and 6 (sqd1 pgp1-1) independent experiments. B) Data are means ± Se from 3 independent experiments for all plants. Asterisks indicate statistically significant differences from the WT level (P < 0.05, Student's t-test). GUN4, GENOMES UNCOUPLED 4.

Plastidic anionic lipids affect the behavior of Chlide–LPOR complexes after photoconversion

To examine how the deficiency of anionic lipids affects the activity and the behavior of the LPOR-pigment complexes, we examined in situ low temperature fluorescence spectra from etiolated cotyledons of wild-type and anionic lipid mutants. At 77 K, 2 emission bands peaking at 634 and 657 nm, which were attributed to nonphotoactive and photoactive Pchlide, respectively, were observed in wild-type and all anionic lipid mutants (Fig. 7A). With a 0.7-ms flash treatment, the emission band around 657 nm disappeared in all plants and a new band emerged around 695 nm, which originates from Chlide in oligomeric Chlide–LPOR–NADP⁺ complexes (Fig. 7B). The data demonstrate that the instantaneous photoconversion of Pchlide to Chlide by LPOR efficiently occurred in the anionic lipid mutants as well as wild type.

Figure 7.

In situ 77 K fluorescence spectra in 4-d-old etiolated seedlings of the wild-type (WT) and anionic lipid mutants. Fluorescence spectra in cotyledons frozen in liquid nitrogen before A) and immediately after B) the flash treatment. Fluorescence spectra in cotyledons frozen after dark incubation for 2.5 min C), 5 min D), and 20 min E) after the flash treatment. Averaged data from 3 independent experiments are shown. Vertical lines are on 655, 680, and 695 nm. F) Overview of the Shibata shift after the flash treatment. Data are represented as means ± Se from 3 independent experiments.

After photoconversion, a gradual shift (Shibata shift) of the Chlide fluorescence was observed in etiolated cotyledons during dark incubation, which reflects the dissociation process of the oligomeric Chlide–LPOR–NADP⁺ complex (Shibata 1957; Smeller et al. 2003; Solymosi et al. 2007). At 2.5 min after the flash, wild type and sqd1 showed a shoulder band near 680 nm in addition to the peak near 695 nm (Fig. 7C). By contrast, pgp1-1 and sqd1 pgp1-1 showed a strong fluorescence around 680 nm already at 2.5 min after the flash, suggesting a faster shift of the Chlide fluorescence in these mutants. At 5 min after the flash, sqd1 still showed the Chlide fluorescence peaking near 690 nm, although the wild type showed a peak at 683 nm (Fig. 7D). In pgp1-1 and sqd1 pgp1-1, Chlide fluorescence peaks were observed at 681 and 678 nm, respectively, at 5 min. At 20 min after the flash, all plants showed Chlide fluorescence peaks around 680 nm (Fig. 7E). The quantification of the peak wavelengths of Chlide fluorescence demonstrates that the shift proceeded most rapidly in sqd1 pgp1-1, followed by pgp1-1, wild type, and sqd1 (Fig. 7F). The sqd2-2 and the sqd2-2 pgp1-1 mutants also showed slower and faster shifts similar to sqd1 and sqd1 pgp1-1, respectively (Supplemental Fig. S5), confirming the effects of SQDG deficiency on the dissociation process of the Chlide–LPOR oligomers.

Discussion

PG biosynthesis affects various processes of etioplast development

The pgp1-1 mutant carries a point mutation in the PGP1 gene and has an 80% reduction in PGP1 activity, which results in a 30% reduction in PG content in light-grown leaves (Xu et al. 2002). The small difference in the PG content between pgp1-1 and wild type can be partly explained by the activity of PGP2, which is the minor isoform of phosphatidylglycerophosphate synthase localized to the endoplasmic reticulum (ER; Tanoue et al. 2014), in addition to the remaining PGP1 activity in this mutant. Although the mean value of PG content was 23% lower in etiolated pgp1-1 seedlings than the wild type, the difference was not statistically significant (Fig. 1). The data indicate that the effect of the pgp1-1 mutation on total PG content was small or negligible in etiolated seedlings. By contrast, the pgp1-1 mutation had strong impacts on the multiple processes in etioplasts including formation of the regular PLB structure (Fig. 2), Pchlide biosynthesis (Figs. 4 and 5), and LPOR-pigment dynamics (Fig. 7). Therefore, not the change in total PG content in the cell but the local disruption of the PG biosynthesis in etioplasts would strongly affect these processes in pgp1-1. Although PGP1 is targeted to mitochondria in addition to plastids (Babiychuk et al. 2003), it is unlikely that the disruption of mitochondrial PG biosynthesis in pgp1-1 strongly affects the processes in etioplasts, because even the knockout mutants of PGP1, which have severe disruptions in chloroplast development with a seedling lethal phenotype, have functional mitochondria in leaves (Hagio et al. 2002; Babiychuk et al. 2003).

Previous studies revealed that deficiencies of galactolipids also disrupt the PLB structure, Pchlide biosynthesis, and LPOR-pigment dynamics (Fujii et al. 2017, 2018). In pgp1-1, the contents of galactolipids (MGDG + DGDG), anionic lipids (SQDG + PG), and total plastidic lipids (MGDG + DGDG + SQDG + PG) were comparable with those in wild type (Fig. 1 and Supplemental Fig. S1), so the pgp1-1 mutation affected etioplast development without changing the global lipid levels in etioplasts. In contrast, the pgp1-1 mutation significantly changed the fatty acid compositions of MGDG and DGDG, although it did not strongly affect those of SQDG, PG, and other phospholipids (Supplemental Fig. S2). It is unknown how the pgp1-1 mutation specifically affects the fatty acid compositions of MGDG and DGDG. In plant cells, galactolipids are synthesized via 2 distinct pathways; the plastid pathway completed within plastids and the ER pathway via phospholipid biosynthesis in the ER (Li-Beisson et al. 2013). Because 16:3-containing MGDG, which is exclusively synthesized via the plastid pathway, was increased by the pgp1-1 mutation, the plastid pathway may be enhanced by the decreased PGP1 activity or consequent impairments in etioplast development. We cannot exclude that the changes in fatty acid compositions of galactolipids by pgp1-1 caused etioplast disruptions. In fact, the fatty acid composition of membrane lipids, particularly the proportion of unsaturated fatty acids, strongly affects the fluidity of the lipid bilayer (Mikami and Murata 2003). In pgp1-1 and sqd1 pgp1-1, DGDG showed a decreased ratio of polyunsaturated fatty acids (mainly 18:3) to saturated fatty acids (mainly 16:0), whereas MGDG of these mutants maintained the high proportion of polyunsaturated fatty acids with an increased 16:3 complementary to a decreased 18:3. Therefore, the altered fatty acid composition in galactolipids, particularly the enhanced saturation of fatty acids in DGDG, might affect the membrane structures of the pgp1-1 and sqd1 pgp1-1 etioplasts. Nevertheless, the sqd1 pgp1-1 double mutant showed stronger disruptions in etioplast development than pgp1-1 despite the similar fatty acid compositions of galactolipids between these mutants. Thus, we assume that the lack of anionic lipids but not the altered fatty acid composition in galactolipids would be the main cause of the perturbed etioplast development in pgp1-1 and the double mutants.

In contrast to the strong effects of the partial limitation of PG biosynthesis on the etioplast development, complete lack of SQDG did not notably change the PLB structure (Fig. 2 and Supplemental Fig. S3) and Pchlide biosynthesis (Figs. 4 and 5). The data indicate that SQDG is not essential for these processes in Arabidopsis. In sqd1, total anionic lipid content was maintained to the wild-type level (Supplemental Fig. S1A), so PG may function to compensate the loss of SQDG in the sqd1 etioplasts. Of note, the additional mutation of sqd1 in pgp1-1, which resulted in a significant decrease in total anionic lipid content, strongly enhanced the defects in PLB formation and Pchlide biosynthesis, suggesting the importance of SQDG when the PG biosynthesis is limited. We confirmed that the sqd2-2 mutation affected the etioplast functions similar to sqd1, suggesting that GlcADG, which is abolished in sqd2-2 but not in sqd1, has no important function in etioplasts at least under nutrient-sufficient conditions. As suggested in the process of chloroplast development (Yu and Benning 2003; Yoshihara et al. 2021), SQDG in etioplasts would play a role in maintaining total anionic lipid constant, whereas PG has specific functions that cannot be complemented by SQDG even during etioplast development.

Anionic lipids are essential for the internal membrane formation in etioplasts

The PLB lattice structure was disrupted by the pgp1-1 mutation. Nguyen et al. (2021) reported that Arabidopsis LPORs require PG and MGDG in addition to Pchlide and NADPH to form in vitro helical tubes that fluorometrically resemble isolated PLBs. Because LPOR is a direct membrane-binding and membrane-remodeling enzyme, an interaction between LPOR and PG may be important for PLBs to form the regular lattice.

The sqd1 pgp1-1 double mutant showed impaired formation of PTs in addition to the severely disrupted PLB structures (Fig. 2). It was reported that a mutation (dgd1) in DGDG SYNTHASE 1, which caused an 80% decrease in DGDG content, also impaired the development of PTs in addition to disordering the PLB lattice (Fujii et al. 2018). Because DGDG is the second most abundant lipid in etioplasts, the substantial decrease in this lipid greatly alters the ratio of the nonbilayer-forming MGDG to bilayer-forming lipids (DGDG + SQDG + PG; Jouhet 2013) and may also cause a deficiency of lipid molecules to extend the lamellar membranes in etioplasts. By contrast, the sqd1 pgp1-1 double mutations did not largely affect the abundance of total plastidic lipids (Supplemental Fig. S1B) and the ratio of the nonbilayer-to-bilayer-forming lipids in etioplasts (Fig. 1). Thus, not a lack of total bilayer lipids but specific loss of the anionic lipids would perturb PT formation in sqd1 pgp1-1 etioplasts. Another possibility is that the decreased abundance of LPORs in the double mutant results in the decreased PT formation in etioplasts (Fig. 3). However, considering that antisense RNA inhibition of LPOR expression in A. thaliana did not impair the formation PTs (Franck et al. 2000), the effect of decreased abundance of LPORs in sqd1 pgp1-1 on the PT formation would be marginal.

Anionic lipids are essential for the Pchlide biosynthesis pathway

In the pgp1-1 mutant, Pchlide content significantly decreased, which was further enhanced in the sqd1 pgp1-1 and sqd2-2 pgp1-1 double mutants (Fig. 4). ALA feeding experiments showed specific accumulation of Proto IX in pgp1-1 and the double mutants (Fig. 5A and Supplemental Fig. S4), indicating that the Mg insertion into Proto IX by Mg-chelatase was specifically impaired by lack of anionic lipids. This result contrasts with the reports that the deficiencies of MGDG and DGDG both caused high accumulation of Mg-Proto IX with ALA feeding to etiolated seedlings (Fujii et al. 2017, 2018). Thus, anionic lipids play a crucial role in the Mg insertion step by Mg-chelatase, whereas galactolipids are required for the Mg-Proto IX metabolism by MgMT. Of note, deficiencies of galactolipids also caused the accumulation of Proto IX and Mg-Proto IX ME with ALA feeding, suggesting a broad effect of galactolipids on the membrane-associated steps of the Pchlide biosynthesis pathway (Fujii et al. 2019). By contrast, anionic lipids may be more specifically involved in the Mg insertion step.

Because the mRNA expression of the genes involved in the Mg insertion step was not significantly suppressed in pgp1-1 (Fig. 6A), transcriptional regulation would be irrelevant to this impairment. Nevertheless, we cannot exclude the possibility that the enzymes involved in this step are not sufficiently accumulated or recruited to the etioplast membranes, in addition to the possibility that the Mg-chelatase requires anionic lipids to exert its activity. Because the CHLH protein, the core subunit of the Mg-chelatase, is undetectable in etiolated seedlings (Stephenson and Terry 2008), it is difficult to address the stability and localization of the functional Mg-chelatase complex in etioplasts. Considering that CHLH becomes detectable during the greening process of the seedlings (Stephenson and Terry 2008), characterization of Mg-chelatase proteins in developing chloroplasts will provide important information about the effects of anionic lipids on the stability and/or localization of Mg-chelatase. It is worth mentioning that decreased Chl biosynthesis by loss of PG biosynthesis was also observed in the cyanobacterium Synechocystis sp. PCC 6803 (Kopečná et al. 2015). Although the requirement of PG for the Mg insertion step remains unknown in cyanobacteria, these data demonstrate the importance of PG for the Chl biosynthesis pathway across cyanobacteria and plants.

High accumulation of Proto IX in pgp1-1 and sqd1 pgp1-1 by ALA treatment suggests that the biosynthetic pathway from ALA to Proto IX is functional in this mutant (Fig. 5A). Meanwhile, in the absence of ALA, neither pgp1-1 nor sqd1 pgp1-1 showed excess accumulation of Proto IX and other porphyrin intermediates despite the impaired Pchlide biosynthesis (Fig. 5B). These data imply that, in pgp1-1 and sqd1 pgp1-1, Pchlide biosynthesis is suppressed at the pathway preceding the ALA formation, as similarly observed in MGDG and DGDG-deficient A. thaliana mutants (Fujii et al. 2017, 2018).

In sqd1 pgp1-1, mRNA levels of CHLH and GUN4 significantly decreased from the wild-type levels (Fig. 6A). Therefore, the decreased expression of CHLH and GUN4 in the double mutant may explain the stronger impairment of the Mg-chelation step even compared with the pgp1-1 mutant. In sqd1 pgp1-1, mRNA levels of HEMA1 and PORA also significantly decreased (Fig. 6B). Thus, in the double mutant, the severe disruption of membrane development in etioplasts may attenuate the expression of some tetrapyrrole biosynthesis genes, although the regulatory mechanism is unknown. In sqd1 pgp1-1, the decreased expression of HEMA1 and PORA might decrease Pchlide biosynthesis and LPOR protein accumulation, respectively, at least partially.

Disruptions of PG biosynthesis and SQDG biosynthesis oppositely affect the dynamics of Chlide–LPOR complex after photoconversion

In pgp1-1 and sqd1 pgp1-1, the ratio of photoactive Pchlide to total Pchlide was similar to that in the wild type (Fig. 4B). Moreover, in these mutants, all photoactive Pchlide was instantaneously converted to Chlide with a flash (Fig. 7B). Although PG was shown to increase the affinity of LPORs toward NADPH in vitro (Gabruk et al. 2017), our data indicate that PG and SQDG are not required for the formation of photoactive Pchlide–LPOR complexes in vivo or that the decreased amounts of PG in pgp1-1 and sqd1 pgp1-1 are still sufficient for the photoactive complex formation. In addition, unlike the MGDG-deficient A. thaliana plants (Fujii et al. 2017), pgp1-1 and sqd1 pgp1-1 showed no remarkable fluorescence emission from the dimeric forms of the photoactive complex around 645 nm at 77 K (Fig. 7A), suggesting that the partial deficiency of PG and complete loss of SQDG do not affect the oligomerization of the photoactive complexes.

After the photoconversion of Pchlide to Chlide, low-temperature fluorescence from oligomeric Chlide-LPOR complexes around 690 nm shifts to shorter wavelength (∼680 nm) during dark incubation, which reflects the dissociation of the oligomeric complexes and the replacement of NADP⁺ with NADPH in the complex (Smeller et al. 2003; Solymosi et al. 2007). As compared with wild type, pgp1-1 and sqd1 showed faster and slower transitions of Chlide fluorescence, respectively (Fig. 7F), suggesting that the partial disruption of PG biosynthesis accelerates the dissociation of the oligomeric Chlide-LPOR complexes, whereas the complete loss of SQDG delays it. Surprisingly, the sqd1 pgp1-1 double mutant showed even a faster shift of Chlide fluorescence than the pgp1-1 single mutant. A compensatory increase in PG content in response to the loss of SQDG has been widely observed in plants (Yu and Benning 2003), algae (Sato et al. 1995), and cyanobacteria (Güler et al. 1996; Aoki et al. 2004; Endo et al. 2016). In fact, the mean value of the relative PG content was 1.4-fold higher in sqd1 than wild type, although the difference was not statistically significant (Fig. 1). We hypothesize that PG has an effect to prevent the dissociation of the oligomeric Chlide–LPOR complex, and thus a compensatory increase of the PG content in sqd1 may stabilize the oligomeric Chlide-LPOR complexes, whereas the partial PG deficiency in pgp1-1 may destabilize them. This assumption is consistent with the rapid dissociation of the Chlide-LPOR oligomers in sqd1 pgp1-1, whose PG content was not increased in response to the loss of SQDG. Because the PLB lattice structure was substantially loosened in pgp1-1 and sqd1 pgp1-1 (Fig. 2), the disorder of the membrane structure might affect the behavior of the Chlide–LPOR complexes. However, the dgd1 mutant also showed disordered PLB lattice, although this mutant had retarded dissociation of the Chlide–LPOR oligomers (Fujii et al. 2018). Therefore, not the global change in the PLB structure but the loss of the specific interaction between anionic lipids and Chlide-LPOR complexes may affect the behavior of the Chlide–LPOR oligomers during the Shibata shift.

Materials and methods

Plant materials and growth conditions

The wild type (Col-0), sqd1 (Okazaki et al. 2009), sqd2-2 (Okazaki et al. 2013), pgp1-1 (Xu et al. 2002), sqd1 pgp1-1, and sqd2-2 pgp1-1 (Yoshihara et al. 2021) were the Columbia accession of Arabidopsis (A. thaliana).

Surface-sterilized seeds on half strength MS medium (adjusted to pH 5.7 with KOH) containing 1% (w/v) sucrose solidified with 0.8% (w/v) agar were cold-treated at 4 °C for 3 or 4 d in the dark. All plants were grown at 23 °C in a growth chamber. The cold-treated seeds were illuminated with white light (∼90 μmol photons m⁻² s⁻¹) for 4 h to synchronize germination and then germinated in the dark. Etiolated seedlings were sampled under dim green light unless stated otherwise.

Lipid analysis

Total lipids were extracted from seedlings crushed in liquid nitrogen according to Bligh and Dyer (1959) and were separated by 1D thin layer chromatography as described (Sato et al. 2020) with a modified solvent system of acetone:toluene:methanol:water (40:15:5:2, v/v/v/v). After visualization with 0.01% (w/v) primuline in 80% (v/v) acetone under UV light, MGDG, DGDG, SQDG, PG, and a mixture of other phospholipids (phosphatidylcholine, phosphatidylethanolamine, and phosphatidylinositol) were scraped from silica gel plates. Fatty acids in each lipid fraction were methyl esterified by incubation in 1 M HCl in methanol at 85 °C for 2 h and quantified by gas chromatography (GC-2014; Shimadzu; http://www.shimadzu.com/) with myristic acid as an internal standard.

Transmission electron microscopy analysis

Fixation of etiolated cotyledons, sectioning of the fixed samples, staining of the ultrathin sections with uranyl acetate and lead citrate, and observation of the samples by transmission electron microscopy (TEM) were performed as described previously (Fujii et al. 2017). To construct montages of high-resolution images, TEM pictures of an ultrathin section were captured and combined by an auto montage system provided in JEM-1400 (JEOL). Quantitative analysis of etioplast ultrastructure was performed with the ImageJ variant software Fiji (Schindelin et al. 2012). The proportion of membrane area in PLBs was determined by measuring the total membrane area in a 0.09 μm² area of a PLB. To calculate the frequency of internal membrane-lacking etioplasts in ultrathin sections, we defined etioplasts with a membrane area <2% of the total area as “internal membrane-lacking etioplasts” and counted the number of them out of 60 etioplasts for each line. The circularity index of etioplasts was calculated as follows: 4 × ᴨ × (area of etioplast)/(perimeter of etioplast)². Violin plots of the results were produced using the ggplot2 package in R program (Wickham 2009).

Immunoblot analysis

Extraction, quantification, and separation of total proteins from etiolated seedlings as well as the detection and quantification of total LPOR proteins were performed as described previously (Fujii et al. 2017), with some modifications. Total protein content was determined using a protein assay reagent (XL-Bradford, Pharma Foods International, Kyoto, Japan), and 10 μg of total proteins were subjected to SDS-PAGE. The anti-LPOR antibody, which react with all Arabidopsis LPOR isoforms (Masuda et al. 2003), was used as a primary antibody. Goat antirabbit IgG antibody conjugated with horseradish peroxidase (SA00001-2, ProteinTech) was secondarily reacted with the anti-LPOR antibody and detected using a chemiluminescence reagent (Immobilon; Merck Millipore) and an imager (LuminoGraph I; ATTO). For the loading control, total proteins separated by SDS-PAGE were stained with Coomassie brilliant blue.

Determination of pigment content

To determine the steady-state levels of Pchilde and carotenoids in 4-d-old etiolated seedlings, pigments were extracted in 80% (v/v) acetone and analyzed as described (Fujii et al. 2017). In brief, Pchlide content was determined by measuring the fluorescence of the extract at 634 nm under 433-nm excitation with a spectrofluorometer (RF-5300PC; Shimadzu) and a Pchlide standard of known concentration. The amount of nonphotoactive Pchlide was determined by illuminating intact seedlings with a 0.7-ms single flash from an electronic flash equipment (PZ42X, Sunpak) before pigment extraction. Carotenoid content was determined with the V-730 BIO spectrophotometer (JASCO) according to the following formula: 5.05 × absorbance at 470 nm (μg carotenoids mL⁻¹; Lichtenthaler 1987).

To analyze Pchlide synthesizing activity, whole 4-d-old etiolated seedlings were submerged in a solution containing 10 mm MES-KOH (pH 5.7) and 5 mm MgCl₂, with or without 10 mm ALA, on the rotary shaker in the dark at 23 °C for 24 h. Then, porphyrin pigments were extracted from the seedlings in N,N-dimethylformamide and quantified by HPLC as described (Fujii et al. 2017). The HPLC system consists of L-6200 and L-6000 pumps (Hitachi), an injector with a 20-μL sample loop (LC-organizer, Hitachi), a L-column2 C8 guard column (5 µm, 4.6 × 10 mm; Chemicals Evaluation and Research Institute), and a reverse-phase C8 column (Symmetry C8 column, 100 , 3.5 µm, 4.6 × 150 mm; Waters) and detected by using an FP-4025 spectrofluorometric detector (JASCO).

Reverse transcription quantitative PCR analysis

Total RNA was extracted by using the NucleoSpin RNA Plant (MACHEREY-NAGEL), followed by genomic DNA digestion and reverse transcription with the use of the ReverTra Ace qPCR RT Master Mix with gDNA Remover kit (TOYOBO). Complementary DNA was amplified by 2-step thermal cycling consisting of an initial denaturation step at 95 °C for 60 s followed by 40 cycles of 15 s at 95 °C and 45 s at 60 °C with the use of the Thunderbird SYBR qPCR Mix (TOYOBO) and 200 nm gene-specific primers (Supplemental Table S1). The thermal cycling and signal detection were performed in duplicate by use of StepOne Real-Time PCR System (Applied Biosystems). The relative abundance of all transcripts amplified was normalized to the means of constitutive expression level of ACTIN8 and UBIQUITIN11 according to Pfaffl (2001).

In situ low temperature fluorescence spectroscopy

Fluorescence emission spectra from Pchlide and Chlide at 77 K were obtained directly from etiolated seedlings in liquid nitrogen under 440-nm excitation with a spectrofluorometer (RF-5300PC; Shimadzu; Fujii et al. 2017).

Statistical analysis

Statistical significance analysis was performed by using the Pearson's χ² test for Fig. 2F, the Welch's t-test for Fig. 4B, and the Student's t-test for Figs. 3B and 6. For multiple comparisons, 1-way ANOVA with post hoc Tukey honestly significant difference test was conducted by using the multcomp R package (Hothorn et al. 2008). The 0.05 level of probability was used as the criterion for significance.

Accession numbers

Sequence data of the genes investigated in this article can be found in The Arabidopsis Information Resource under the following accession numbers: PGP1 (AT2G39290), SQD1 (AT4G33030), SQD2 (AT5G01220), ACT8 (AT1G49240), UBQ11 (AT4G05050), CHLH (AT5G13630), CHLD (AT1G08520), CHLI1 (AT4G18480), GUN4 (AT3G59400), HEMA1 (AT1G58290), CHLM (AT4G25080), CHL27 (AT3G56940), PORA (AT5G54190), and PORB (AT4G27440).

Author contributions

A.Y. performed most experiments, analyzed and interpreted data, and prepared the first draft of the manuscript. Ke.K. and N.N. performed transmission electron microscopic experiments. S.F. supported fluoroscopic experiments and edited the drafts of the manuscript. H.W. provided intellectual support and edited the final draft of the manuscript. Ko.K. designed and conducted experiments, analyzed and interpreted data, and wrote all draft of the manuscript.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Abundance of plastidic lipids in 4-d-old etiolated seedlings of wild-type and anionic lipid mutants.

Supplemental Figure S2. Fatty acid compositions of membrane lipids from 4-d-old etiolated seedlings of wild-type and anionic lipid mutants.

Supplemental Figure S3. Ultrastructures of etioplasts in sqd2-2 and sqd2-2 pgp1-1.

Supplemental Figure S4. Protochlorophyllide metabolism in sqd2-2 and sqd2-2 pgp1-1.

Supplemental Figure S5. In situ 77 K fluorescence spectra in 4-d-old etiolated seedlings of sqd2-2 and sqd2-2 pgp1-1.

Supplemental Table S1. Oligonucleotide primers used for reverse transcription quantitative PCR analysis.

Funding

This work was supported by the Japan Society for the Promotion of Science (KAKENHI nos. 18H03941, 20K06691, 22H05076 to Ko.K.). A part of this research was carried out under the Cooperative Research Project of Research Center for Biomedical Engineering.

Data availability

The data underlying this article are available in the article and in its online supplementary material.

Dive Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

• PGP2 Gramene: AT4G25960

• PGP2 Araport: AT4G25960

• ALA CHEBI: CHEBI:17549

• NADP CHEBI: CHEBI:25523

• DGD1 Gramene: AT3G11670

• DGD1 Araport: AT3G11670

• NADPH CHEBI: CHEBI:16474

• PGP1 Gramene: AT2G39290

• PGP1 Araport: AT2G39290

• SQD1 Gramene: AT4G33030

• SQD1 Araport: AT4G33030

• SQD2 Gramene: AT5G01220

• SQD2 Araport: AT5G01220

• CHLH Gramene: AT5G13630

• CHLH Araport: AT5G13630

• CHLD Gramene: AT1G08520

• CHLD Araport: AT1G08520

• CHLI1 Gramene: AT4G18480

• CHLI1 Araport: AT4G18480

• GUN4 Gramene: AT3G59400

• GUN4 Araport: AT3G59400

• HEMA1 Gramene: AT1G58290

• HEMA1 Araport: AT1G58290

• CHLM Gramene: AT4G25080

• CHLM Araport: AT4G25080

• CHL27 Gramene: AT3G56940

• CHL27 Araport: AT3G56940

• PORA Gramene: AT5G54190

• PORA Araport: AT5G54190

• PORB Gramene: AT4G27440

• PORB Araport: AT4G27440