Knockdown of the plastid-encoded acetyl-CoA carboxylase gene uncovers functions in metabolism and development

Abstract

De novo fatty acid biosynthesis in plants relies on a prokaryotic-type acetyl-CoA carboxylase (ACCase) that resides in the plastid compartment. The enzyme is composed of four subunits, one of which is encoded in the plastid genome, whereas the other three subunits are encoded by nuclear genes. The plastid gene (accD) encodes the β-carboxyltransferase subunit of ACCase and is essential for cell viability. To facilitate the functional analysis of accD, we pursued a transplastomic knockdown strategy in tobacco (Nicotiana tabacum). By introducing point mutations into the translational start codon of accD, we obtained stable transplastomic lines with altered ACCase activity. Replacement of the standard initiator codon AUG with UUG strongly reduced AccD expression, whereas replacement with GUG had no detectable effects. AccD knockdown mutants displayed reduced ACCase activity, which resulted in changes in the levels of many but not all species of cellular lipids. Limiting fatty acid availability caused a wide range of macroscopic, microscopic, and biochemical phenotypes, including impaired chloroplast division, reduced seed set, and altered storage metabolism. Finally, while the mutants displayed reduced growth under photoautotrophic conditions, they showed exaggerated growth under heterotrophic conditions, thus uncovering an unexpected antagonistic role of AccD activity in autotrophic and heterotrophic growth.

Analysis of the only plastid genome-encoded fatty acid biosynthesis gene reveals functions in plastid division and seed development, and antagonistic roles in autotrophic and heterotrophic growth.

Introduction

Present-day plastid (chloroplast) genomes are much reduced compared to the genomes of their cyanobacterial ancestors. After establishment of the endosymbiotic relationship, thousands of cyanobacterial genes were lost or functionally transferred to the nuclear genome of the host cell (Bock and Timmis, 2008; Archibald, 2015). Nearly all of the approximately 130 genes that have been retained in the plastid genome of vascular plants function in either photosynthesis or gene expression. Plastid genes related to photosynthesis encode subunits of the two photosystems (photosystems I and II), the cytochrome b₆f complex, the ATP synthase, and the NADH dehydrogenase-like complex, as well as the large subunit of the CO₂-fixing enzyme Rubisco. Plastid genes encoding components of the gene expression machinery, also referred to as genetic system genes (Shimada and Sugiura, 1991), include genes for a complete set of tRNAs (Alkatib et al., 2012), four genes encoding the core subunits of a bacterial-type RNA polymerase (Igloi and Kössel, 1992), all ribosomal RNA (rRNA) genes and some ribosomal protein genes (Tiller and Bock, 2014), as well as a single gene encoding a component of the proteolytic machinery (clpP; Gray et al., 1990).

One of the very few exceptional plastid genes that are unrelated to photosynthesis or gene expression is accD. It encodes a subunit of the acetyl-CoA carboxylase (ACCase) complex that catalyzes de novo fatty acid synthesis. The AccD protein is the only component of plant lipid metabolism that is encoded in the plastid DNA. It has been hypothesized that the accD gene is the only essential gene in the plastid genome (Kode et al., 2005) and that the functions of the gene expression machinery, as well as of clpP (encoding the proteolytic subunit of the ATP-dependent caseinolytic protease) and the two conserved open reading frames ycf1 (hypothetical chloroplast reading frame 1) and ycf2 (hypothetical chloroplast reading frame 2; Drescher et al., 2000), are solely necessary for the expression, regulation, and function of the essential AccD subunit of the plastid ACCase (de Vries et al., 2015). Indeed, analyses of gene expression in nongreen plastid types (chromoplasts, amyloplasts) have suggested that the main, if not the only, task of the gene expression machinery in nonphotosynthetic plastids may be to synthesize the AccD protein (Kahlau and Bock, 2008; Valkov et al., 2009; Caroca et al., 2013).

Fatty acid biosynthesis in plants represents a prime example of how metabolic pathways have been reorganized after integration of the cyanobacterial endosymbiont into the host cell. Whereas in other eukaryotes, the pathway of de novo fatty acid synthesis is localized in the cytosol, it resides in the plastid compartment of plants (Sasaki et al., 1995; Chapman and Ohlrogge, 2012). In addition to the AccD gene product (the β-carboxyltransferase subunit; β-CT), the acetyl-CoA carboxylase complex contains three subunits that are encoded in the nuclear genome (Sasaki et al., 1995; Salie and Thelen, 2016): the biotin carboxyl carrier protein (BCCP; encoded by the CAC1 gene, with CAC standing for chloroplast acetyl-CoA carboxylase), the biotin carboxylase (BC; encoded by CAC2), and the α-carboxyltransferase (α-CT; encoded by CAC3). ACCase catalyzes the first committed step in fatty acid biosynthesis: the carboxylation of acetyl-CoA to malonyl-CoA. Malonyl-CoA synthesis occurs in two reaction steps. The ATP-dependent carboxylation of biotin with bicarbonate is followed by the transfer of the carboxyl group to acetyl-CoA (Ohlrogge and Browse, 1995). Subsequently, the fatty acid synthase complex catalyzes a series of reactions that produce fatty acids up to a chain length of 18 carbon atoms by the iterative addition of two carbon atoms per cycle, using the malonyl-CoA produced by ACCase as the carbon donor of the chain elongation reactions (Ohlrogge and Browse, 1995). After export from the plastid, further elongation and desaturation of the fatty acids can occur in the endoplasmic reticulum (Li et al., 2016). Chain elongation outside of plastids requires a cytosolic form of ACCase, which is a single large polypeptide harboring all three enzymatic activities (BC, BCCP, and CT; Ohlrogge and Browse, 1995). A related multifunctional enzyme exists in the plastids of grasses (Poaceae), where it has replaced the heteromeric ACCase complex. Consequently, the plastid genomes of grasses lack the accD gene, a feature that, interestingly, correlates with nonessentiality of plastid protein biosynthesis in grasses (Hess et al., 1994; Ahlert et al., 2003).

Given the central role of the plastid ACCase in lipid metabolism, it is unsurprising that accD gene expression and ACCase activity are highly regulated. High expression levels of both nucleus-encoded and plastid-encoded ACCase subunits are detected in growing seedlings (which have a high demand for membrane lipids to accommodate cellular growth) and in seeds when oil reserves are deposited (Ke et al., 2000). Overexpression of accD in tobacco plastids results in increased levels of the three nucleus-encoded ACCase subunits, suggesting that availability of the AccD protein limits biogenesis of the enzyme complex and controls holoenzyme accumulation (Madoka et al., 2002). In addition, the ACCase complex is regulated at the level of enzyme activity. The synthesis of one molecule of palmitic acid (C16:0) requires 14 molecules of NADPH and 7 molecules of ATP, which are produced by the light reactions of photosynthesis. In vitro assays demonstrate that the stimulation of fatty acid synthesis by light is mediated by activation of the plastid ACCase enzyme (Sasaki et al., 1997). The catalytic activity of the enzyme is maximal at pH 8 and 2–5 mM Mg²⁺, conditions that are similar to those in the chloroplast stroma during illumination. Moreover, the ACCase is redox-activated, presumably via the plastid ferredoxin–thioredoxin system (Sasaki et al., 1997).

Attempts to study the function of the plastid accD gene in detail by generating accD knock-out plants have failed due to the essentiality of the gene. Chloroplast transformation experiments in tobacco have resulted in persistent heteroplasmy, indicating that the presence of wild-type copies of accD is indispensable for cell viability (Kode et al., 2005). The aim of this study was to develop a transplastomic strategy that facilitates the comprehensive analysis of the functions of accD in plant metabolism and development by reverse genetics. We report that stable transplastomic accD knockdown lines can be obtained by mutating the translation initiation codon from AUG to UUG. Analysis of the knockdown mutants revealed important roles of the AccD protein and plastid fatty acid biosynthesis in chloroplast division and seed development, and an unexpected antagonistic impact on autotrophic versus heterotrophic growth.

Results

Generation of transplastomic tobacco lines with mutated accD start codons

Gene expression in chloroplasts is mainly controlled at the post-transcriptional level, especially at the level of translation (Eberhard et al., 2002; Kahlau and Bock, 2008; Zoschke and Bock, 2018). The efficiency of translation depends on ribosome recruitment to the 5′-untranslated region (5′-UTR) and the efficiency of start codon recognition (Scharff et al., 2017). Consequently, translational efficiency can be manipulated in an mRNA-specific manner by introducing appropriate mutations into the translational start codon or the ribosome-binding site in the 5′-UTR, also referred to as the Shine–Dalgarno sequence (Majeran et al., 2000; Rott et al., 2011; Moreno et al., 2017).

Given that accD is an essential gene that cannot be knocked out (Kode et al., 2005), we tested the possibility to obtain stable accD knockdown plants that display analyzable mutant phenotypes. To this end, we changed the initiator codon of accD from the canonical AUG triplet to GUG or UUG in a cloned plastid DNA fragment from tobacco (Figure 1). Both GUG and UUG are rarely used as start codons for translation initiation in plastids (Hirose et al., 1999; Kuroda et al., 2007), and are generally assumed to be less efficiently recognized than AUG (although recognition efficiency of non-AUG start codons can be strongly dependent on the sequence context; Kozak, 2005). Thus, although GUG is structurally more similar to AUG than UUG (in that, it has a purine in the first codon position) and also occurs more frequently than UUG in both bacterial and organellar genomes, it cannot necessarily be assumed that an AUG-to-GUG mutation will result in a milder reduction of translation than an AUG-to-UUG exchange.

Figure 1.

Knockdown of the accD gene in the tobacco plastid genome. A, Physical maps of the region in the tobacco plastid genome (ptDNA) containing the accD gene and the modified region in transplastomic lines that harbor the selectable marker gene aadA and a point mutation in the translation initiation codon of accD (indicated as TTG and GTG, respectively). The selectable marker gene was targeted to the intergenic region between rbcL and accD. Genes above the line are transcribed from left to right, and genes below the line are transcribed in the opposite direction. Expression of the aadA marker was driven by the plastid ribosomal RNA operon promoter (Prrn) and the 3′-UTR from the plastid psbA gene (TpsbA). The expected sizes of DNA fragments in RFLP analyses with the restriction enzyme EcoRI are indicated above both maps. The accD-derived hybridization probe used in RFLP and RNA gel blot analyses is represented as a black bar. B, Southern blot analysis of transplastomic tobacco lines. Total DNA samples were digested with EcoRI and hybridized to the accD-specific probe indicated in (A). Two independently generated transplastomic lines per construct are shown. The faint band of wild-type size observed in all transplastomic lines most likely represents promiscuous plastid DNA in the nucleus (see text for details). Wt: wild-type Nicotiana tabacum. C, Inheritance assay to confirm homoplasmy of the transplastomic lines. T1 seeds of transplastomic lines harboring the mutated accD start codon were germinated on synthetic medium in the presence of spectinomycin (+spec). In contrast to the white (spectinomycin-sensitive) seedlings emerging from wild-type seeds, all transplastomic lines give rise to a uniform population of antibiotic-resistant seedlings, strongly suggesting homoplasmic presence of the aadA cassette.

To facilitate selection of plastid-transformed (transplastomic) lines, a chimeric aadA expression cassette (conferring resistance to the aminoglycoside-type antibiotic spectinomycin; Svab and Maliga, 1993) was inserted into the intergenic region between accD and the upstream rbcL gene (encoding the large subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase; Figure 1A). To avoid undesired transcriptional upregulation of accD due to read-through transcription from the strong plastid ribosomal RNA operon promoter (P_rrn) driving the aadA marker gene, the aadA cassette was integrated in antisense orientation relative to accD (Figure 1A). In this way, three vectors for chloroplast transformation were produced that only differ in the identity of the accD initiator codon: vector paadA-accD harboring the canonical ATG codon, vector pGTG-accD containing a GTG accD start codon, and vector pTTG-accD carrying a TTG start codon (Figure 1A).

The three vectors were used to stably transform tobacco plastids using the biolistic protocol and selection for spectinomycin resistance conferred by the chimeric aadA gene (Svab and Maliga, 1993). Transplastomic lines were obtained with all three constructs and passed through additional rounds of regeneration under antibiotic selection to enrich the transformed plastid genome and isolate homoplasmic lines (Svab and Maliga, 1993; Bock, 2001). In each regeneration round, samples were taken to test for the presence of the point mutation in the initiator codon by PCR amplification and DNA sequencing. While all primary transplastomic lines generated with vectors pGTG-accD and pTTG-accD (subsequently referred to as GTG-accD and TTG-accD lines) carried a detectable signal for the mutated nucleotide in the sequence chromatograms, most lines lost this signal in subsequent regeneration rounds. Out of 30 transplastomic TTG-accD lines tested, only two had the mutated start codon (TTG) after three regeneration rounds, whereas all others had reverted to the native (ATG) start codon (and thus became identical to the aadA-accD control plants). Similarly, from 30 tested GTG-accD lines, only three retained the mutated start codon after the third-regeneration cycle. Loss of the point mutation during the heteroplasmic phase (when wild-type and transformed plastid genomes are simultaneously present) is due to gene conversion between the two genome types, which is a highly active mutation-eliminating mechanism in chloroplasts (Khakhlova and Bock, 2006).

After three regeneration rounds, homoplasmy of transplastomic lines was assessed by restriction fragment length polymorphism (RFLP) analysis using southern blotting (Figure 1B). All aadA-accD, GTG-accD, and TTG-accD lines showed a strongly hybridizing fragment of the expected size (∼5.5 kb) that corresponded to insertion of the aadA cassette by homologous recombination. In addition, all lines showed a faint hybridizing band corresponding to the size of the wild-type fragment (∼4.3 kb). Weak wild-type-like hybridization signals that persist after multiple rounds of selection and regeneration are typically due to plastid DNA sequences that have integrated into the nuclear genome (the so-called promiscuous DNA; Timmis and Scott, 1983; Hager et al., 1999; Ruf et al., 2000; Bock, 2017).

To ultimately confirm the homoplasmic state of the transplastomic lines, seeds were harvested and germinated on medium-containing spectinomycin (Figure 1C). The absence of antibiotic-sensitive seedlings from the progenies of all lines strongly suggests that the plants were homoplasmic. To verify that the mutated start codon had been retained, the region spanning the accD start codon was amplified by PCR with DNA extracted from pools of ∼20 T1 seedlings as template, and the products were sequenced. This analysis revealed the presence of the desired start codon in all transplastomic lines in a homoplasmic state (Supplemental Figure S1).

Phenotypes of transplastomic lines with mutations in the accD initiator codon

When grown under standard growth conditions in the greenhouse, homoplasmic GTG-accD plants were phenotypically inconspicuous and indistinguishable from wild-type plants and aadA-accD control plants (Figure 2). In contrast, the TTG-accD plants displayed a clear mutant phenotype. At the seedling stage, the leaves were light green to yellow, and the leaf shape was narrower than in the control plants (Figure 2A).

Figure 2.

Phenotype of accD mutants raised under standard growth conditions. A, Phenotype of plants four weeks after sowing. Note the delay in growth, the pigment deficiency, and the narrow leaf blades of the TTG-accD mutant. B, Phenotype seven weeks after sowing. Although growth retardation is still observed in the TTG-accD plants, leaf size and pigmentation are now nearly indistinguishable from the wild-type and the other transplastomic mutants. The onset of flowering was delayed by approximately one week in the TTG-accD mutant. Scale bars: 5 cm.

Interestingly, at later developmental stages, both the pigmentation phenotype and the leaf shape phenotype disappeared, and the TTG-accD plants now resembled wild-type plants, although their growth was still substantially retarded (Figure 2B). The onset of flowering was delayed by approximately one week compared to the other plant lines. These observations preliminarily suggested that accD knockdown in TTG-accD plants makes fatty acid biosynthetic capacity limiting for growth during early plant development, while at later developmental stages, fatty acid biosynthesis can catch up, possibly by accumulating ACCase over time (and not turning it over).

Analysis of accD transcript accumulation in transplastomic tobacco mutants

To examine the relationship between accD gene expression and the mutant phenotype of TTG-accD plants, northern blot analyses were performed. To this end, leaf material was harvested from both young and mature plants, to account for the dependence of the mutant phenotype on the developmental stage (Figure 2). Material from young plants was collected when the fifth leaf was 3- to 4-cm long, and material from mature plants was harvested from the youngest fully expanded leaf at the onset of flowering. Northern blot analyses with an accD-specific probe revealed that comparable amounts of accD mRNA accumulated in wild-type plants, aadA-accD control plants, and GTG-accD plants (Figure 3). In contrast, accD transcript levels were strongly elevated in the TTG-accD plants. Increased transcript levels were seen in both young and mature plants (Figure 3, A and B), indicating that they are unrelated to the pigment deficiency (that occurred only in young plants; Figure 2, A and B). Instead, they may represent a regulatory response to ACCase deficiency. However, as the factors involved in transcriptional regulation of accD are not known, this hypothesis currently cannot be tested experimentally.

Figure 3.

Northern blot analysis of accD mRNA accumulation in transplastomic plants. A, Analysis of leaf samples from young plants (cf. Figure 2A). Samples of 7-µg total cellular RNA was loaded, gel electrophoretically separated under denaturing conditions, blotted and hybridized to the accD-specific probe as shown in Figure 1A. The ethidium bromide-stained gel prior to blotting is shown as a control for equal loading below the blot. The two large rRNAs of the cytosolic 80S ribosome (25S rRNA and 18S rRNA) are indicated (wt: wild type). The right side shows a schematic transcript map with the major mRNA species detected and their relative quantification (with the intensity of the strongest band in the blot set to 1). The 1.6-kb band represents the monocistronic accD mRNA. Additional prominent transcript species of ∼3 and 6.7 kb were detected in all lines. They result from cotranscription of accD with the downstream psaI operon (Shinozaki et al. 1986; Hajdukiewicz et al. 1997). B, Analysis of leaf samples from mature plants (cf. Figure 2B).

All plant lines showed a complex pattern of accD transcripts, with the major RNA species having a size of ∼3 kb and less prominent transcript species of >6.0 and ∼1.6 kb (Figure 3). This complex transcript pattern has been observed in previous studies (Meurer et al., 1996; Hajdukiewicz et al., 1997) and is due to accD residing in a large cluster of genes that all have the same transcriptional orientation. In addition to accD, the cluster comprises the genes rbcL, psaI (encoding a small subunit of photosystem I; Schöttler et al., 2017), ycf4 (hypothetical chloroplast reading frame 4), ycf10 (hypothetical chloroplast reading frame 10) and petA (encoding apocytochrome f), and polycistronic transcripts spanning the entire gene cluster have been detected in Arabidopsis (Arabidopsis thaliana; Walter et al., 2010). Interestingly, all accD transcript species detected in our northern blot experiments accumulated to higher levels in the TTG-accD mutants (Figure 3).

Reduced accD translation in transplastomic mutants

Our analyses of accD mRNA accumulation (Figure 3) suggested that if the TTG-accD mutants are indeed ACCase-deficient, the cause of this deficiency must lie downstream of transcription and RNA stability. Given that the translational start codon was mutated, it seemed reasonable to assume that, although TTG-accD mutant plants accumulate more accD transcripts, these transcripts are only poorly translated. To test this assumption, polysome association analyses were conducted. In these assays, polysomes (complexes of mRNAs with translating ribosomes) are isolated and separated by sucrose density-gradient fractionation (Barkan, 1993; Barkan, 1998). The more the ribosomes are associated with a given mRNA molecule, the deeper the complex will migrate into the density gradient. Subsequent gradient fractionation and RNA isolation from the fractions facilitate the assessment of mRNA loading with ribosomes by RNA gel blot analysis (Barkan, 1993; Rogalski et al., 2008a,b).

To identify the gradient fractions that contain free mRNAs and those that contain polysomal complexes, a puromycin-containing gradient was analyzed. Puromycin releases translating ribosomes from mRNAs by causing premature chain termination. Analysis of the fractionated puromycin-containing gradient revealed that gradient fractions 2–4 contained free accD mRNAs (Figure 4). Comparison to the untreated wild-type sample (and the aadA-accD control plants) identified fractions 5–10 as fractions containing accD mRNAs associated with ribosomes. The polysomes showed a pronounced peak in fractions 8 and 9 (Figure 4). Interestingly, the polysome profile of the TTG-accD mutant was strongly altered. The bulk of the accD transcripts accumulated in fractions 3–5, fractions that contain untranslated mRNAs (fractions 3 and 4), or mRNAs that are only sparsely covered with ribosomes (fraction 5; Figure 4). These data suggest that accD mRNAs with a UUG initiator codon are not efficiently translated and indicate that the TTG-accD mutant plants are indeed ACCase-deficient due to knockdown of accD at the level of protein biosynthesis.

Figure 4.

Polysome association of the accD mRNA in transplastomic tobacco plants. The translation rate of accD in the TTG-accD mutant compared to the wild-type and the aadA control line (aadA-accD) was evaluated by analyzing the ribosome loading of the transcripts. Polysomes were separated in sucrose gradients. Fractions were taken and numbered from the top (1; lowest density) to the bottom (10; highest density). RNA was purified from each fraction and equal aliquots were loaded onto 1.5% denaturing agarose gels for electrophoretic separation. After blotting, membranes were hybridized with a radiolabeled accD gene-specific probe. A puromycin-treated sample served as a control and identified fractions 5–10 as containing actively translated mRNA. Fractions 3 and 4 (underlined) contain the bulk of the untranslated accD transcripts. The major fractions containing translated accD mRNAs are in bold and underlined. Note that the TTG-accD mutant lacks the peak of polysome-associated accD transcripts in fractions 8 and 9 (that is present in the wild-type and the aadA control line) and, instead, shows an extended peak in the region containing untranslated and lowly translated mRNAs (fractions 3–5). The ethidium bromide-stained agarose gels prior to blotting are shown below each blot to confirm RNA integrity and similar ribosome (rRNA) distribution across the fractions. Relative quantification of the major transcript species is shown at the right (cf. Figure 3), with the intensity of the strongest band in each blot set to 1.

Analysis of ACCase activity in accD knockdown plants

To determine the consequences of reduced accD translation in the TTG-accD mutant plants, two approaches were pursued: (1) immunodetection of the AccD protein with specific antibodies and (2) establishment of an enzyme activity assay for ACCase. Although immunodetection of AccD has been reported (Sasaki et al., 1993), we were unable to reliably detect and (semi)quantitatively assess AccD protein levels in our transplastomic lines. We, therefore, attempted to establish an ACCase activity assay for tobacco plants, which, in view of the complex regulation of ACCase at the level of enzyme activity (see Introduction section), offers the additional advantage of providing information on the relevance of any reduction in AccD expression level to the activity of fatty acid biosynthesis in plastids.

We adapted a previously developed spectrophotometric assay for the measurement of ACCase activity (Kroeger et al., 2011) that couples acetyl-CoA carboxylation to the NADPH-dependent reduction of malonyl-CoA. The latter reaction is catalyzed by the malonyl-CoA reductase (MCR) from the photosynthetic thermophilic nonsulfur bacterium Chloroflexus aurantiacus. MCR was recombinantly produced in Escherichia coli and used to set up suitable assay conditions with tobacco leaf extracts (see Methods section).

In view of the developmental phenotype of the TTG-accD mutants, young plants and mature plants were measured under the optimized assay conditions (Figure 5). While young wild-type plants, aadA-accD control plants and GTG-accD mutant plants had comparable ACCase activity, enzyme activity in TTG-accD mutant plants was undetectable (Figure 5). Interestingly, in mature TTG-accD plants, ACCase activity became detectable (Figure 5), consistent with the disappearance of the pigment deficiency (Figure 2) and the suspected increasing accumulation of ACCase over time.

Figure 5.

Enzyme activity assays to determine ACCase activity in transplastomic plants. ACCase activity was measured in young wild-type plants, aadA-accD control plants, and the GTG-accD and TTG-accD start codon mutants. Plants were grown under standard light conditions (100–150 μE m⁻² s⁻¹, 16 h light/8 h dark, 25°C/20°C) or under high light conditions at 800 μE m⁻² s⁻¹. Four true leaves were harvested (and pooled) from 3-week-old plants (wild-type, aadA-accD, and GTG-accD) or 3.5-week-old TTG-accD plants (to compensate for their slight growth delay and adjust all plants to the same developmental stage; see Figure 2). For comparison, mature plants grown under standard light conditions were also measured. To this end, leaf number 5 (from the bottom) was harvested from 9-week-old wild-type plants or 10-week-old TTG-accD plants (representing the same developmental stage). The data represent average values ±SD (n =3; two plants per replicate). Letters indicate samples that were not significantly different (p<0.05; one-way ANOVA with Holm–Sidak post hoc test).

Altered chloroplast morphology and ultrastructure in TTG-accD plants

To characterize the mutant phenotype of the TTG-accD plants in more detail, microscopic investigations were undertaken (Figure 6). Analysis of mesophyll cells revealed that chloroplast morphology in pale leaves of young TTG-accD plants was severely altered (Figure 6A). Chloroplasts were much reduced in number and strongly enlarged, suggesting that ACCase deficiency causes defects in chloroplast division. Since the accD-TTG mutant showed recovery from pigment deficiency at later developmental stages, chloroplast morphology was also analyzed in mature plants. Mesophyll cells of mature plants contained normal sized plastids (Figure 6B), indicating that the aberrant chloroplast morphology is restricted to early plant development, when ACCase deficiency is most severe (Figure 5).

Figure 6.

Analysis of chloroplast morphology in accD-deficient mutants by differential interference contrast (DIC) microscopy. Tissue samples were taken from green or yellow leaf regions of TTG-accD plants (Figure 2) and from control plants. Mesophyll cells were prepared for microscopy (Wu and Xue 2010) and analyzed by DIC microscopy. A, Images taken 30 d after sowing (DAS). Samples were taken from yellow patches of TTG-accD leaves (Figure 2A) and the corresponding areas of wild-type and aadA-accD leaves. Note the abnormal chloroplasts in the TTG-accD mutant. Chloroplasts are reduced in number and much larger than the chloroplasts in the control plants. B, Images taken 50 DAS. At this stage, TTG-accD mutant plants have fully recovered and restored wild-type-like leaf pigmentation (cf. Figure 2B). Normal chloroplast size and shape are also restored. Scale bar: 20 µm.

To test whether restricted supply of malonyl-CoA to fatty acid biosynthesis also results in aberrant internal membrane structure and organization, thylakoid ultrastructure was investigated by electron microscopy. Indeed, chloroplasts of young TTG-accD plants displayed severely altered thylakoid organization, with grana stacks largely lacking and the stroma lamellae having no discernable thylakoid lumen (Figure 7).

Figure 7.

Transmission electron microscopic images of chloroplasts in the TTG-accD mutant. A, Overview of the chloroplasts inside a single cell. The greatly enlarged chloroplasts in the TTG-accD mutant are clearly visible. Scale bars: 5 µm. B, Images taken at higher magnification to reveal the organization of the thylakoids. The right side represents a zoom into the boxed region indicated in the left images. Scale bars: 1 µm.

Lipid analysis of accD-TTG mutant plants

Having confirmed that ACCase activity is strongly reduced in our transplastomic TTG-accD plants, we next wanted to determine the resulting changes in lipid accumulation. Since the mutant phenotype is evident only during early stages of development (Figure 2), leaf material from young plants was harvested for lipid analysis. Lipid profiling (lipidomics) was performed by UPLC/MS (Burgos et al., 2011; Giavalisco et al., 2011; Armarego-Marriott et al., 2019) for the wild-type, the aadA-accD control plants, and the TTG-accD mutant plants (Figure 8).

Figure 8.

Analysis of polar lipids in leaves of wild-type tobacco plants and transplastomic mutants. Lipids were measured using a UPLC/MS-based lipidomic platform (Burgos et al. 2011). Leaves of wild-type plants and the aadA-accD and TTG-accD transplastomic lines were analyzed one month after sowing. To reduce the influence of differences in development, the samples were collected when the fifth leaf had reached a length of 3–4 cm. The x-axis indicates the lipid species and the desaturation level. For example, 34:6 corresponds to acyl chains of 16C and 18C each harboring three double bounds. The y-axis gives normalized intensities. A, MGDG. B, DGDG. C, Sulquinovosyldiacylglycerol (SQDG). D, PC. Note that MGDG, DGDG, and SQDG are lipid classes that are exclusively present in plastid membranes, whereas PC occurs mainly in extraplastidic membranes. Bars heights in (A–D) represent the median value, and the upper and lower limits of the error bars correspond to the first and third quartile, respectively. Six independent wild-type, aadA-accD, and TTG-accD plants were used for the measurements. Asterisks indicate significant differences (Student’s t test; p<0.01).

Polar lipids are the most abundant lipid species occurring in membranes. The main components of plastid membranes are galactoglycerolipids, with monogalactosyldiacylglycerol (MGDG) and digalactosyldiacylglycerol (DGDG) accounting for ∼50% and ∼25% of chloroplast lipids, respectively (Block et al., 1983). The two major MGDG species are 34:6 MGDG (sn-1 18:3/sn-2 16:3; Figure 8A) and 36:6 MGDG (sn-1 18:3/sn-2 18:3), representing approximately 95% of the total MGDG pool. The only detectable change in the TTG-accD mutant was a decrease in 36:6 MGDG.

The DGDG pool of tobacco leaves is mainly composed of 34:3 DGDG, 34:6 DGDG and 36:6 DGDG, together accounting for ∼90% of the total DGDG content (Figure 8B). TTG-accD mutant plants showed a reduction in 34:6 and 36:6 DGDG species (Figure 8B). A third class of galactolipids, sulfoquinovosyldiacylglycerol, has 34:3 and 36:6 as major species, both of which were significantly reduced in the TTG-accD mutants (Figure 8C).

As a lipid species occurring mainly outside of the chloroplast, phosphatidylcholine (PC) was analyzed. Interestingly, we found that 36:4, 36:5, and 36:6 PC were significantly more abundant in the TTG-accD mutant plants than in the wild-type and the aadA-accD control plants (Figure 8, D), while PC species containing 34 carbon atoms were not significantly altered.

When converted to triacylglycerols (TAGs), lipids can serve as energy-rich storage compounds. To determine whether reduced ACCase activity impacts the metabolism of neutral storage lipids in TTG-accD mutant plants, TAG accumulation was measured (Figure 9). These analyses revealed a general decrease in TAGs in the mutant plants. The two most prominent classes of TAGs were those containing 52 or 54 carbon atoms, and all of the abundant lipid species within these classes were markedly reduced in TTG-accD plants (Figure 9). The strongest reductions were seen in the 54:7, 54:8, and 54:9 TAGs. This indicates that, upon ACCase knockdown, the provision of long acyl chains, particularly 18C fatty acids, becomes limiting in lipid synthesis.

Figure 9.

Analysis of TAGs in leaves. Samples were prepared from leaves of young tobacco plants (wild-type, aadA-accD, and TTG-accD plants) harvested one month after sowing. Measurements were performed using a UPLC/MS-based lipidomic platform (Burgos et al. 2011). The x-axis indicates the lipid species and the desaturation level, and the y-axis gives normalized intensities. The most abundant TAG species was 54:8 (54 carbon atoms and 8 double bounds), corresponding to a TAG that is likely composed of two 18:3 and one 18:2 acyl chains. Bars heights represent the median value, and the upper and lower limits of the error bars correspond to the first and third quartile, respectively. Six independent wild type, aadA-accD, and TTG-accD plants were used for the measurements. Asterisks indicate significant differences (Student’s t test; p<0.01).

Analysis of accD mutants under stress conditions

ACCase activity is tightly regulated in response to environmental conditions, especially light. High activity of photosynthetic electron transfer increases the pH and the Mg²⁺ concentration in the chloroplast stroma, conditions that activate ACCase (Sasaki et al., 1997), presumably to increase utilization of ATP and NADPH. Also, de novo fatty acid biosynthesis is important for the correct assembly of the photosynthetic apparatus under cold stress (Takami et al., 2010). We, therefore, decided to investigate the influence of high light and low temperature on the phenotype of the TTG-accD and GTG-accD mutants (Figure 10).

Figure 10.

Phenotype of accD knockdown plants under high light and chilling stress. A, Phenotype under high-light treatment. Seeds were germinated under standard light conditions and transferred to 800 µE m⁻² s⁻¹ after emergence of the cotyledons (i.e. after 10 d). Note the severe growth retardation and delay in development of the TTG-accD lines (two independent transplastomic lines shown; cf. Figure 1B). Scale bars: 5 cm. B, Phenotype under chilling stress. After 5 weeks of growth in standard conditions, plants were transferred to 10°C. Photographs were taken after six weeks of chilling stress. Note the strong growth retardation and pigment deficiency in the TTG-accD mutant. Scale bars: 5 cm.

When grown under high light, the TTG-accD mutant plants displayed a more severe phenotype than under standard conditions (Figures 2 and 10, A). Growth of the TTG-accD plants was delayed, and plants reached the onset of flowering 2 weeks later than the wild type and the GTG-accD plants. To associate the light-dependent phenotype with ACCase activity, enzyme activity measurements were conducted. ACCase activity was undetectably low in young TTG-accD plants grown under high light (Figure 5). Although it recovered with time, the activity in mature plants was still substantially lower than in the wild-type. Interestingly, young GTG-accD plants also showed a small reduction in ACCase activity when grown in high light (Figure 5), suggesting that when the plants are forced to grow at maximum speed, ACCase activity can also become limiting in the GTG-accD mutant.

As lipids and membrane fluidity play important roles in cold acclimation (Wada et al., 1994; Moon et al., 1995), and de novo fatty acid synthesis is also required for faithful photosystem assembly at low temperatures (Takami et al., 2010), we also tested whether the differences in lipid content and composition in the TTG-accD mutant affect growth at low temperatures. Exposure of plants to chilling stress at 10°C for 6 weeks caused severe growth retardation in TTG-accD plants (Figure 10B). In addition, pigment loss occurred in those leaves that developed under chilling stress (i.e. the young leaves; Figure 10B), whereas leaves that had matured prior to the exposure to 10°C remained green. This phenotype suggests that chloroplast biogenesis in the cold is impaired, when the fatty acid synthesis capacity is low.

ACCase deficiency impacts seed production and seed storage metabolism

The production of TAGs (oil) was strongly reduced in leaves of the TTG-accD mutant plants (Figure 9). In tobacco, oil is a major storage compound of seeds, accounting for ∼20% of the dry weight. We, therefore, wanted to determine the consequences of reduced ACCase activity on seed development (Figure 11). Comparison of capsules revealed a marked size reduction in the TTG-accD mutant. While the length of the capsules was largely unaffected, the capsules of TTG-accD plants were much thinner than those of the control plants (Figure 11A).

Figure 11.

Seed production and seed development in accD knockdown plants. A, Comparison of capsule size between the wild-type, the TTG-accD mutant, and the aadA-accD control plants. Scale bars: 1 cm. B, Comparison of 100 seed weights. 100 seeds were randomly picked and weighed. There were no statistically significant differences between the analyzed lines (p<0.05; Student’s t test). C, Seed yields per 10 capsules. Seeds from 10 randomly chosen capsules were collected and weighed. *p<0.01 (Student’s t test). D, Total seed yields. Seeds from all capsules produced by individual plants were collected and weighed. Values plotted in (B–D) are average ±sd of four independent plants of the TTG-accD and aadA-accD transplastomic mutants and of three wild-type plants, respectively. *p<0.01 (Student’s t test). E, Analysis of the seed set in immature capsules. Capsules were harvested ∼10 d after pollination and dissected for observation under a stereomicroscope. Scale bars: 1 cm (left images) and 3 mm (right images).

While there were no significant differences between the plant lines in the 100 seed weights (Figure 11B), a strong difference was found when seed production per 10 capsules was measured. Capsules of TTG-accD plants contained fewer seeds, and their seed content was reduced to ∼30% of that of wild-type capsules (Figure 11C). Consistent with this finding, the total seed yield per plant was also strongly reduced (by ∼50%) in the TTG-accD mutant (Figure 11D). These data suggest that the main effect of ACCase deficiency on seed development is a reduction in seed number rather than a reduction in seed weight. The reduced seed set was already evident early in fruit development. In immature capsules of TTG-accD plants, large areas of the columella were free of seeds or only loosely covered with seeds (Figure 11E). This observation suggests that, when faced with a limited capacity to synthesize storage compounds, the plant invests in fewer seeds (rather than producing smaller seeds), thus likely optimizing reproductive success.

Finally, we analyzed the metabolic composition of the seeds to assess the impact of reduced ACCase activity on oil deposition and accumulation of other storage compounds (Table 1). Seeds of TTG-accD plants indeed had a reduced oil content, while their protein content was largely unaltered. The reduction in oil (by approximately one-third; Table 1) was accompanied by a pronounced increase in carbohydrates. While soluble sugars were only mildly (and not statistically significantly) increased, the starch content of TTG-accD seeds was more than ten-fold higher than in the control lines. Thus, to a certain extent, reduced seed oil synthesis can be compensated by increased starch deposition.

Table 1.

Storage compounds in seeds of wild-type tobacco plants, aadA-accD control plants, and the TTG-accD knockdown mutants.

	Oil (%)	Protein (%)	Sucrose (µmol/g FW)	Glucose (µmol/g FW)	Starch (µmol/g FW)
Wild-type	18.69 ± 1.30a	30.22 ± 2.14a	28.40 ± 2.28a	5.33 ± 1.39a	0.73 ± 0.48a
aadA-accD	18.74 ± 1.89a	30.20 ± 3.22a	29.47 ± 1.45a	6.03 ± 0.70a	0.70 ± 0.35a
TTG-accD	12.18 ± 2.01b	31.57 ± 2.82a	33.10 ± 3.96a	8.68 ± 2.78a	8.40 ± 1.86b

The data represent average values ± SD (n = 3). Identical superscript letters indicate values that are not significantly different from each other (p < 0.01; one-way Analysis of variance (ANOVA) with Tukey’s post hoc test).

ACCase deficiency confers accelerated plant growth under heterotrophic conditions

When TTG-accD seeds were germinated on standard synthetic medium [Murashige and Skoog {MS} medium; Murashige and Skoog, 1962 supplemented with 2% sucrose], we noticed that, unexpectedly, the seedlings grew substantially faster than wild-type seedlings and seedlings of the aadA-accD control line (Figure 12). This was clearly not an effect of earlier germination (Figure 12A), but due to accelerated postgermination growth (Figure 12B).

Figure 12.

Exogenously supplied metabolizable sugars trigger exaggerated growth of the TTG-accD mutant. Surface-sterilized seeds were geminated on synthetic medium containing 58 or 234 mM of sucrose, glucose, or sorbitol and, as a control, on medium not supplemented with sugar (0 mM sugar). A, Germinating seeds photographed 9 d after sowing. B, Seedling phenotypes 4 weeks after sowing.

To further explore this surprising phenomenon and establish a direct link to metabolism, we replaced the sucrose in the medium with sorbitol, a nonmetabolizable sugar. This did not result in growth differences between the three plant lines (Figure 12), excluding an osmotic effect, suggesting that utilization of the sugar in metabolism is required for the exaggerated growth of TTG-accD seedlings.

Since sucrose inhibits gene expression at various levels (e.g. Chan and Yu, 1998; Wiese et al., 2004), we considered the possibility that the TTG-accD seedlings are less sensitive to sucrose suppression. Low photosynthetic activity of the seedlings (Figure 2A) could lead to increased utilization of the sugar in the medium and, in this way, potentially relieve repressive effects of sucrose on gene expression. However, this is unlikely to be the case, because transplastomic photosynthesis mutants are also more strongly dependent on the sucrose in the medium, but yet do not show enhanced growth on sucrose-containing synthetic medium (Krech et al., 2012, 2013;). Likewise, transplastomic mutants with general defects in plastid gene expression do not have a growth advantage over wild-type seedlings (Rogalski et al., 2008a,b; Alkatib et al., 2012; Ehrnthaler et al., 2014). To provide additional evidence against a role of sucrose suppression, we tested growth on glucose as an alternative metabolizable sugar. Glucose also conferred enhanced growth of TTG-accD seedlings (Figure 12), further supporting the conclusion that feeding reduced carbon into primary metabolism is the cause of the exaggerated growth phenotype of the mutant.

Finally, to directly assess the possible role of sucrose suppression, we supplied sucrose in concentrations that are known to be inhibitory. When seeds were sown on medium with 8% sucrose, a clear delay in germination was observed (Figure 12A). However, this delay occurred uniformly in all plant lines, again arguing against substantial differences in sensitivity to sucrose suppression. Importantly, the growth advantage of the TTG-accD seedlings persisted even at 8% sucrose. Surprisingly, whereas the TTG-accD seedlings showed yellow leaves in 2% sucrose, they displayed normally green pigmentation in the presence of 8% sucrose (Figure 12B). The exaggerated growth phenotype also persisted when sucrose was replaced by high glucose (at the same molar concentration; Figure 12).

Discussion

The plastid accD gene is special in many ways. It is the only plastid genome-encoded gene whose product is involved in a metabolic pathway other than photosynthesis. The gene also shows an unusually high nucleotide substitution rate (Kahlau et al., 2006; Greiner et al., 2008) and has been suggested to be under strong positive selection (Rockenbach et al., 2016). In addition to its function in lipid metabolism, accD has recently been implicated in chloroplast inheritance and competition between plastid genomes (Sobanski et al., 2019).

Essentiality of the gene (Kode et al., 2005) and its location in the plastid genome has previously prevented genetic analysis of accD function. In the course of this work, we developed a transplastomic reverse genetic approach that enabled us to generate stable knockdown mutants for accD in tobacco. By changing the accD initiator codon from AUG to UUG, accD expression could be sufficiently reduced to obtain a discernable mutant phenotype while retaining plant viability (Figures 1 and 2; Supplemental Figure S1). Interestingly, the severity of the mutant phenotype strongly depended on the development of the plant, in that young seedlings were more severely affected by AccD deficiency than older plants (Figure 2). Enzyme activity measurements suggest that this phenotypic recovery with age is due to accumulation of ACCase over time (Figure 5), presumably by a reduction in ACCase turnover in response to AccD deficiency.

Although in some material of our accD knockdown plants ACCase activity was undetectably low, this merely reflects the limited sensitivity of the only currently available enzyme activity assay (Figure 5; Kroeger et al., 2011). Genetic work in both tobacco (Kode et al., 2005) and Arabidopsis (Parker et al., 2014; Parker et al., 2016) has clearly established the essentiality of the plastid aacD gene and the inability of cytosolic (homomeric) ACCase enzymes to compensate for the loss of the plastid enzyme. The continued production of low levels of AccD protein even in young TTG-accD plants (where enzyme activity was below detection limit; Figure 5) is also supported by our polysome loading analyses that revealed residual translation activity of accD transcripts (Figure 4). The latter is consistent with previous studies showing low-level translation from UUG initiation codons in tobacco plastids (Rott et al., 2011; Moreno et al., 2017).

The accD knockdown plants displayed a pronounced light-sensitive phenotype and enhanced sensitivity to chilling stress (Figures 2 and 10). Both high light and chilling stress represent conditions that are known to lead to increased production of reactive oxygen species, which, in turn, trigger lipid peroxidation (op den Camp et al., 2003; Sonoike, 2011; Ramel et al., 2013). As increased lipid damage leads to enhanced lipid turnover, it increases the demand for de novo synthesized fatty acids, thus likely explaining the stress sensitivity of the mutant plants.

Lipidomic analysis of the mutants revealed reduced accumulation of many species of polar and nonpolar lipids (Figures 8 and 9). A remarkable exception was the PCs, a class of polar lipids that largely occur outside of plastids. 36C PCs were increased in the TTG-accD mutant plants (Figure 8D), suggesting that, under conditions of limited fatty acid provision, plants prioritize the accumulation of (extraplastidic) membrane fatty acids over chloroplast lipids and storage lipids.

Analysis of the AccD-deficient mutant plants implicated the plastid ACCase in several additional biological processes. The TTG-accD plants showed severely disturbed chloroplast division (Figure 6). Impaired chloroplast division was also observed in an Arabidopsis mutant related to lipid metabolism: a T-DNA insertion mutant in the gene for ketoacyl-ACP synthase I, an enzyme that elongates ACP-bound acyl species (Wu and Xue, 2010). The connection between accD function and chloroplast division is particularly interesting in light of the recent finding that accD is involved in plastid competition in evening primroses (Oenothera), a genus that displays biparental chloroplast inheritance (Sobanski et al., 2019). If AccD activity determines the rate of chloroplast division, it is conceivable that accD alleles that encode more active enzymes confer faster division rates and, in this way, make the plastid genomes carrying such strong alleles more competitive.

AccD deficiency also had a profound effect on seed development and storage compound metabolism (Figure 11 and Table 1). Previous work had shown that overexpression of accD does not lead to altered fatty acid content of seeds, although total seed production was reported to increase (Madoka et al., 2002). In contrast, our TTG-accD knockdown plants showed reduced seed oil content, which interestingly was accompanied by a massive increase in starch (Table 1). This finding suggests that oil deficiency can be partially compensated by increased deposition of carbohydrates as storage compounds.

The most striking developmental phenotype seen in the mutant plants was a strong reduction in seed production (Figure 11). While seed weight was largely unaltered (Figure 11B), seed number (both per capsule and per plant) was much lower than in the wild type. This observation indicates that seed production is tightly coupled to the capacity of storage metabolism. If not enough storage compounds can be provided, the plant responds by investing in fewer seeds rather than producing seeds equipped with insufficient reserves in the endosperm (Figure 11). Conceivably, this is the optimal strategy to ensure reproductive success under adverse conditions.

Another developmental phenotype associated with AccD deficiency was the altered leaf shape in TTG-accD mutant plants (Figure 2A). The leaf blades of TTG-accD plants were narrower than in the wild-type, a phenotype that was previously observed in mutants with defects in plastid translation (Tiller and Bock, 2014). The occurrence of the same phenotype upon ACCase deficiency raises the interesting possibility that the activity of fatty acid biosynthesis in the plastid affects leaf development, and that the altered leaf morphology of chloroplast translation mutants (Tiller and Bock, 2014) is due to low accD expression.

A very surprising observation was that AccD plays antagonistic roles in autotrophic and heterotrophic growth. Whereas the negative impact on autotrophic growth (Figures 2 and 10) was to be expected, the stimulatory effect on heterotrophic growth of TTG-accD knockdown plants (Figure 12) defies a straightforward explanation. Given that fatty acid biosynthesis and, secondarily, photosynthetic activity are reduced in the mutant, it seems unlikely that the activity of primary metabolism directly triggers the exaggerated growth under heterotrophic conditions. Instead, low flux of reduced carbon into fatty acid synthesis, together with the availability of the metabolizable sugars that are taken up, may trigger the utilization of excess reduced carbon for growth. This could be mediated, for example, by signaling pathways that normally restrict cell proliferation and growth. Future research will be directed toward testing candidate pathways that could link low AccD activity to accelerated growth, including phytohormone signaling pathways (Achard et al., 2006) and the target of rapamycin pathway, a central regulator of metabolism and growth (Caldana et al., 2013).

Methods

Plant material and growth conditions

Aseptic tobacco (Nicotiana tabacum) cv Petit Havana plants for plastid transformation experiments were raised from seeds on MS medium (Murashige and Skoog, 1962) supplemented with 30 g/L sucrose under a light intensity of ∼50 µE m⁻² s⁻¹. Regenerated transplastomic plants were grown on the same medium and under identical lighting conditions. After they had rooted, plantlets were transferred to soil and grown to maturity under standard greenhouse conditions.

For physiological and biochemical analyses, plants were raised from seeds and grown in controlled environment chambers at 150 µE m⁻² s⁻¹, under a diurnal cycle of 16 h light and 8 h dark, at 22°C. For high light treatment, plants were germinated under standard conditions and then transferred to high light conditions (800 µE m⁻² s⁻¹) immediately after cotyledon emergence. For cold stress treatments, seeds were germinated in standard conditions and transferred to low temperature (10°C) 5 weeks after germination.

Construction of plastid transformation vectors

Plastid transformation vectors for mutagenesis of the accD start codon were constructed by cloning the accD region from the tobacco plastid genome as a 3.65-kb EcoRI (nucleotide position 61,990)/PstI (nucleotide position 58,332) fragment into vector pUC19 digested with the same pair of restriction enzymes. To remove undesired restriction sites from the polylinker region, the resulting plasmid was digested with EcoRI and EcoRV, treated with the Klenow fragment of E. coli DNA polymerase I, and religated. The resulting vector was then subjected to site-directed mutagenesis to modify the annotated accD start codon (Caroca et al., 2013) using the QuickChange II SDM kit (Stratagene) in combination with the synthetic oligonucleotide pairs listed in Supplemental Table S1. It should be noted that the accD start codon is incorrectly annotated in the NCBI database. The correct start codon was identified and experimentally confirmed in a transplastomic study (Caroca et al., 2013).

Mutated clones were identified by PCR and DNA sequencing (Supplemental Table S1) and subsequently digested with the restriction enzyme Bsu36I and treated with the Klenow fragment of E. coli DNA polymerase I to generate blunt ends. A chimeric aadA cassette (Svab and Maliga, 1993) was then inserted as Ecl136II/DraI fragment into the intergenic spacer between the accD and rbcL genes (Figure 1A). Clones harboring the aadA marker gene in the opposite transcriptional orientation relative to the accD gene were selected and designated pGTG-accD and pTTG-accD, respectively. As a control, a vector containing the unmutated accD start codon (ATG) and the aadA inserted into the same location and transcriptional orientation was produced (vector paadA-accD; Figure 1A).

Plastid transformation and selection of transplastomic tobacco lines

Young leaves from aseptically grown tobacco plants were bombarded with plasmid DNA-coated 0.6-µm gold particles using a helium-driven biolistic gun (PDS1000He; BioRad) with the Hepta Adaptor setup. Primary transplastomic lines were selected on regeneration medium of plants (RMOP medium; Svab and Maliga, 1993) containing 500 µg/mL spectinomycin. Spontaneous spectinomycin-resistant mutants were identified by their sensitivity to streptomycin (500 µg/mL; Bock, 2001) and eliminated. Confirmed transplastomic lines were tested for the presence of the desired start codon mutations by amplifying and sequencing the region containing the accD initiator codon (for primer sequences, see Supplemental Table S1). Additional regeneration rounds under antibiotic selection were conducted to eliminate residual wild-type copies of the (highly polyploid) plastid genome (Golczyk et al., 2014). As in the heteroplasmic state, point mutations can be lost by gene conversion (Khakhlova and Bock, 2006), the PCR and DNA sequencing tests were repeated after each regeneration round, until homoplasmy had been attained.

Isolation and analysis of nucleic acids

For RFLP analyses by southern blotting, total plant DNA was extracted using a cetyltrimethylammoniumbromide-based method (Doyle and Doyle, 1990). DNA samples were digested with the restriction enzyme EcoRI, separated by gel electrophoresis in 1% (w/v) agarose gels, and blotted onto Hybond N nylon membranes (GE Healthcare).

For RNA gel blot analyses, total cellular RNA was extracted from fresh leaves using the peqGOLD TriFast reagent (Peqlab), electrophoretically separated in denaturing formaldehyde-containing 1.2% agarose gels and blotted onto Hybond N nylon membranes (GE Healthcare). For detection of nucleic acids on membranes, [α-³²P]dCTP-labeled hybridization probes were produced by random priming (with the Multiprime DNA labeling kit; GE Healthcare) using gene-specific PCR products as template. Hybridizations were performed at 65°C using standard protocols.

Polysome analyses were performed as described previously (Tiller et al., 2012). Briefly, polysomal complexes were separated in sucrose density gradients, followed by gradient fractionation and RNA extraction from individual fractions. The distribution of accD mRNAs across the gradients was assayed by RNA gel blot analysis. A puromycin-treated sample was used as a control to identify the polysome-containing gradient fractions.

Quantification of bands in northern blots was performed with ImageJ (https://imagej.nih.gov/ij/). The strongest hybridization signal in each blot was set to 1, followed by calculation of the relative intensities of all other bands.

Microscopy

Preparations of mesophyll cells (Wu and Xue, 2010) were used to analyze chloroplast morphology by differential interference contrast microscopy. To this end, yellow or green leaf sectors were cut into small pieces (∼0.5 × 1 cm), submerged in fixation buffer (3.5% glutaraldehyde, 0.1 M NaHPO₄, pH 7.2), vacuum-infiltrated with the buffer until they sunk to the bottom of the tube, and then incubated for 1 h at 4°C in darkness for tissue fixation. Subsequently, the fixation solution was removed, and the samples were transferred to 0.1 M EDTA (ethylenediaminetetraacetic acid) and incubated in darkness at 4°C overnight. To disrupt the cell walls, the fixed leaf pieces were then incubated at 60°C with vigorous shaking for 2–3 h. Finally, the specimens were mounted in water on microscope slides, and cells were released by gentle tapping on the cover slips. Chloroplasts were imaged with an epi-fluorescence microscope (BX-61, Olympus).

Sample preparation for electron microscopy and microscopic analysis of chloroplast ultrastructure was performed as described previously (Zhou et al., 2015).

ACCase activity assays

ACCase activity was measured in crude leaf extracts. Leaves were frozen in liquid nitrogen and ground in a mortar. Then, 30–40 mg of leaf powder was suspended in 350 µL of ice-cold extraction buffer (100 mM HEPES (N‐(2‐hydroxyethyl)piperazine‐N′‐(2‐ethane sulphonic acid)/NaOH, pH 7.5, 1 mM EDTA, 10% [w/v] glycerol, 5 mM DTT (dithiothreitol), complete protease inhibitor [Roche]) and centrifuged for 10 min at 20,000g at 4°C. The supernatant was collected, and after addition of 15 mg polyvinylpolypyrrolidone (Fluka), briefly vortexed and centrifuged again (as above). Twenty microliter of the supernatant was immediately subjected to the ACCase assay (Kroeger et al., 2011). The assay mixture (100-µL final volume) contained 100 mM MOPS (3-(N-morpholino) propanesulfonic acid)-KOH, pH 7.8, 5 mM MgCl₂, 1 mM DTT, 15 mM NaHCO₃, 2 mM ATP (adenosine triphosphate), 0.4 mM NADPH (nicotinamide adenine dinucleotide phosphate), 3 µg recombinant MCR from Chloroflexus aurantiacus. Assays were performed in the presence or absence of 0.4 mM acetyl-CoA. The oxidation of NADPH was monitored at 340 nm (40°C) using an ELx800 absorbance microplate reader (BioTek).

Lipid analysis

For lipid extraction for LC–MS analysis, frozen leaf material was ground to a fine powder in a grinding mill (Retsch GmbH). Samples of 100-mg ground tissue were suspended in 1 mL of precooled (−15°C) extraction mixture (Giavalisco et al., 2011) by vortexing. Samples were then incubated for 10 min in an orbital shaker at 4°C, followed by 10-min incubation in an ultrasonic bath. Subsequently, 500 µL of water:methanol (3:1) was added and samples were mixed thoroughly. Phase separation was obtained by centrifugation for 10 min at maximum speed in a table-top centrifuge. Then, 500 µL of the upper organic phase was collected and concentrated in a SpeedVac concentrator (Concentrator 5301, Eppendorf). Lipids were measured by UPLC-MS and data were analyzed as described previously (Giavalisco et al., 2011; Armarego-Marriott et al., 2019; Sobanski et al., 2019).

Analysis of seed storage compounds

For seed oil extraction, 100 mg aliquots of frozen homogenized seeds were processed using the same protocol for lipid extraction as described above. After phase separation, the complete volume of the organic upper phase was collected and concentrated in a SpeedVac concentrator (Concentrator 5301, Eppendorf). The concentration step allowed the evaporation of the organic solvent, and the remaining oil was quantified by weight. For isolation of total proteins from seeds, a phenolic extraction method was used (Cahoon et al., 1992) with samples of 50 mg of frozen ground material. The protein concentration of the extracts was measured with the BCA protein assay kit (Pierce). Soluble sugars were extracted with ethanol and enzymatically assayed (Stitt et al., 1989). Starch was determined enzymatically in the insoluble material after ethanolic extraction of soluble sugars (Hendriks et al., 2003).

Accession numbers

accD (accession number Z00044.2, region: 59,798–61,336), AccD (NP_054508), rbcL (accession number Z00044.2, region: 57,418–59,174), psaI (accession number Z00044.2, region: 62,088–62,198), ycf4 (accession number Z00044.2, region: 62,643–63,197), ycf10 (accession number Z00044.2, region: 63,420–64,109), and petA (accession number Z00044.2, region: 64,340–65,302).

Supplemental data

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

Supplemental Figure S1. DNA sequence analysis to confirm presence of mutated accD start codons in transplastomic tobacco plants.

Supplemental Table S1. List of oligonucleotides used in this study.

 

R.C., K.A.H., I.M., A.B., N.T., T.P. and M.G.A. performed the research. All authors designed the research and analyzed the data. R.B. conceived of the study and wrote the paper together with R.C. All authors contributed to the editing and review of the manuscript. R.B. serves as the author for contact.

The authors responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instruction for author (https://academic.oup.com/plphys) is: Ralph Bock (rbock@mpimp-golm.mpg.de).