The Functions of Chloroplast Glutamyl-tRNA in Translation and Tetrapyrrole Biosynthesis^{1  [OPEN]}

Abstract

The chloroplast glutamyl-tRNA (tRNA^Glu) is unique in that it has two entirely different functions. In addition to acting in translation, it serves as the substrate of glutamyl-tRNA reductase (GluTR), the enzyme catalyzing the committed step in the tetrapyrrole biosynthetic pathway. How the tRNA^Glu pool is distributed between the two pathways and whether tRNA^Glu allocation limits tetrapyrrole biosynthesis and/or protein biosynthesis remains poorly understood. We generated a series of transplastomic tobacco (Nicotiana tabacum) plants to alter tRNA^Glu expression levels and introduced a point mutation into the plastid trnE gene, which has been reported to uncouple protein biosynthesis from tetrapyrrole biosynthesis in chloroplasts of the protist Euglena gracilis. We show that, rather than comparable uncoupling of the two pathways, the trnE mutation is lethal in tobacco because it inhibits tRNA processing, thus preventing translation of Glu codons. Ectopic expression of the mutated trnE gene uncovered an unexpected inhibition of glutamyl-tRNA reductase by immature tRNA^Glu. We further demonstrate that whereas overexpression of tRNA^Glu does not affect tetrapyrrole biosynthesis, reduction of GluTR activity through inhibition by tRNA^Glu precursors causes tetrapyrrole synthesis to become limiting in early plant development when active photosystem biogenesis provokes a high demand for de novo chlorophyll biosynthesis. Taken together, our findings provide insight into the roles of tRNA^Glu at the intersection of protein biosynthesis and tetrapyrrole biosynthesis.

The two DNA-containing organelles of plant cells, plastids (chloroplasts) and mitochondria, contain their own protein synthesis machinery. In agreement with their endosymbiotic origin from bacteria, both organelles possess bacterial-type 70S ribosomes consisting of a 30S and a 50S subunit (Tiller and Bock, 2014; Sun and Zerges, 2015; Bieri et al., 2017; Zoschke and Bock, 2018; Waltz et al., 2019). Plastids also encode a complete set of tRNAs that is sufficient to decode all 64 triplets of the genetic code (Alkatib et al., 2012b; Cognat et al., 2013). By contrast, plant mitochondria do not encode a full tRNA set in their genome and depend on the import of some tRNA species from the cytosol (Salinas et al., 2006; Duchêne et al., 2009; Vinogradova et al., 2009), a pathway that is likely absent from plastids (Rogalski et al., 2008a; Alkatib et al., 2012a).

The chloroplast glutamyl-tRNA (tRNA^Glu) is unique in that it has a second essential function. In addition to acting in plastid translation, it is also required for tetrapyrrole biosynthesis. Tetrapyrroles are macrocyclic molecules characterized by four pyrrole rings that are connected by methine bridges. Plants contain four classes of tetrapyrroles (heme, chlorophyll, siroheme, and phytochromobilin) that differ in conjugation state, side chains, and/or chelated ion. The universal precursor for the biosynthesis of all tetrapyrroles is 5-aminolevulinic acid (ALA). There are two alternative pathways for ALA synthesis, the C₄ pathways (or Shemin pathway) and the C₅ pathway. The C₄ pathway exists in animals, fungi, and certain groups of bacteria. It initiates with the condensation of succinyl-CoA and Gly, a reaction that is catalyzed by ALA synthetase, to form ALA. In eukaryotes harboring the C₄ pathway, ALA synthetase typically is localized in the mitochondrial compartment. The C₅ pathway depends on tRNA^Glu and exists in plants, archaea, and some groups of bacteria. In plants, the pathway is localized in plastids and utilizes charged plastid-encoded tRNA^Glu as substrate, from which the enzyme glutamyl-tRNA reductase (GluTR) forms Glu-1-semialdehyde in the committed step of the tetrapyrrole biosynthetic pathway. Glu-1-semialdehyde then undergoes an isomerization reaction catalyzed by the enzyme Glu-1-semialdehyde aminotransferase (GSAT) to form ALA (Grimm, 1998; Brzezowski et al., 2015; Wang and Grimm, 2015). Thus, tRNA^Glu can enter one of two pathways in the chloroplast, namely protein biosynthesis or tetrapyrrole synthesis.

GluTR catalyzes the rate-limiting step in tetrapyrrole biosynthesis and is a highly regulated enzyme (Richter et al., 2019). Tight regulation is essential, because many products and intermediates of the pathway are highly toxic when produced in excess and/or present as free compounds. For example, free chlorophylls and their precursors are extremely phototoxic. ALA synthesis from tRNA^Glu is regulated at multiple levels, including (1) biochemical feedback regulation by pathway products, especially heme; (2) redox control of GluTR activity; (3) interaction with a dedicated regulator protein, the GluTR-binding protein (GBP); (4) protein turnover via recognition of the GluTR N terminus by the chloroplast Clp protease; and (5) control of enzyme activity through interaction with the protein FLUORESCENT IN BLUE LIGHT (FLU; Meskauskiene et al., 2001; Richter et al., 2019).

A C-to-U point mutation at nucleotide position 56 of the mature tRNA^Glu (C56U) in plastids of the photosynthetic protist Euglena gracilis is reported to uncouple translation from tetrapyrrole biosynthesis. The mutant strain of E. gracilis that harbored the corresponding C56T mutation in the plastid gene for tRNA^Glu (trnE gene) was orange in color and incapable of chlorophyll biosynthesis, but was able to perform protein biosynthesis in plastids. Further investigation revealed that the point mutation in trnE likely impairs tRNA^Glu substrate recognition by GluTR (Stange-Thomann et al., 1994).

Since the trnE genes from E. gracilis and seed plant plastids share a high degree of homology, we hypothesized that the C56U mutation might have similar effects in plants and also uncouple plastid translation from tetrapyrrole biosynthesis. An important difference between plants and E. gracilis is that E. gracilis possesses both the C₄ (Shemin pathway) and the C₅ pathway for ALA synthesis. Consequently, even if the C₅ pathway is dysfunctional in E. gracilis, the cells should still be able to synthesize some tetrapyrroles (especially heme, which is essential for cell survival) via the alternative C₄ pathway. By contrast, this is unlikely to be the case with plants, which have only the C₅ pathway for synthesis of all tetrapyrroles.

The aim of this study was to assess the role of tRNA^Glu in plastid translation and tetrapyrrole biosynthesis in seed plants. How the tRNA^Glu pool is distributed between the two pathways and whether tRNA^Glu allocation limits tetrapyrrole biosynthesis and/or protein biosynthesis in plastids is currently not known. We therefore used plastid transformation in the model plant tobacco (Nicotiana tabacum) to (1) alter tRNA^Glu expression levels, and (2) introduce the C56T mutation in an attempt to uncouple plastid translation from tetrapyrrole biosynthesis. We report that the C56T mutation, when introduced into the endogenous trnE gene in the tobacco chloroplast genome, is lethal. We also show that ectopic expression of a mutated trnE gene copy results in a pigment-deficient phenotype that, surprisingly, is due to impaired tRNA maturation. Importantly, whereas accumulation of immature tRNA^Glu has no detectable effect on plastid translation, ALA synthesis rate is strongly diminished. Our findings indicate that immature tRNA^Glu inhibits GluTR at the level of enzyme activity. By contrast, overexpression of the wild-type tRNA^Glu in chloroplasts affects neither tetrapyrrole biosynthesis nor protein biosynthesis, suggesting that tetrapyrrole biosynthesis in plants is not controlled at the level of tRNA^Glu provision.

RESULTS

Introduction of the C56T Mutation and Generation of Transplastomic Tobacco Lines with Altered Expression Levels of tRNA^Glu

To dissect the functions of tRNA^Glu in chloroplast translation and tetrapyrrole biosynthesis, four constructs for stable transformation of the tobacco plastid genome were designed (Fig. 1). Vector pEndWt (Fig. 1A) was constructed to test whether overexpression of tRNA^Glu is possible and how this would affect plastid translation and tetrapyrrole biosynthesis. To this end, the strongest known plastid promoter, specifically the ribosomal RNA (rRNA) operon promoter (Nt Prrn), was placed upstream of the endogenous trnE promoter (Hanaoka et al., 2003). Transformation vector pEndMut is similar to pEndWt but replaces the endogenous copy of trnE in the chloroplast genome with the mutated version harboring the C56T point mutation. Considering that (1) the point mutation abolishes the C₅ pathway in E. gracilis, and (2) plants do not have the alternative (Shemin) pathway for heme synthesis, this mutation could potentially lead to inviable plants. We therefore designed two additional constructs that left the endogenous trnE gene unchanged but introduced an additional gene copy into a neutral insertion site within a distant region of the plastid genome (Fig. 1B; Ruf et al., 2001; Wurbs et al., 2007). Vector pEctWt introduces an additional wild-type copy of trnE, whereas vector pEctMut introduces an additional trnE gene copy that carries the C56T point mutation (Fig. 1B).

Using biolistic transformation and selection for spectinomycin resistance conferred by the chimeric aadA gene (Svab and Maliga, 1993), putative transplastomic lines were obtained with all four constructs. The transplastomic status of the lines was preliminarily confirmed by PCR assays amplifying the aadA transgene, and positive lines were subjected to additional rounds of regeneration and selection to enrich the transplastome and isolate homoplasmic lines (Maliga, 2004; Bock, 2015). Transplastomic lines will subsequently be referred to as Nt-EndWt, Nt-EndMut, Nt-EctWt, and Nt-EctMut lines, respectively, with “End” indicating that the endogenous trnE gene copy was manipulated and “Ect” indicating that an ectopic copy of the trnE gene was introduced (Fig. 1).

Lethality of the C56T Mutation in Tobacco

After three consecutive rounds of regeneration under stringent antibiotic selection, the homoplasmic status of the transplastomic lines was assessed by DNA gel blot analyses (Fig. 2). To this end, total DNA was extracted from transplastomic plants and digested with the restriction enzymes BamHI (for Nt-EndWt and Nt-EndMut) or BglII (for Nt-EctWt and Nt-EctMut) for analysis by restriction fragment length polymorphism (RFLP). After separation of DNA fragments by agarose gel electrophoresis, the blotted digestion products were detected by hybridization using either a trnE-trnD-trnY locus-specific probe (expected fragment size was 5.8 kb in the transplastomic Nt-EndWt and Nt-EndMut lines and 4.5 kb in the wild type; Fig. 2A) or a psaB-specific probe (expected fragment size was 5.3 kb in the transplastomic Nt-EctWt and Nt-EctMut lines and 3.5 kb in the wild type; Fig. 2B). Virtual absence of the wild-type-size fragment from the transplastomic samples was taken as preliminary evidence of homoplasmy. Several independent homoplasmic Nt-EctWt, Nt-EctMut, and Nt-EndWt lines were identified, transferred to the greenhouse, and grown to maturity to obtain seeds.

By contrast, Nt-EndMut plants, in which replacement of the endogenous trnE with the mutated version was attempted, were clearly not homoplasmic (Fig. 2A). Stable heteroplasmy in the presence of selection is typically associated with genetic changes that inactivate an essential gene function in the plastid genome (Drescher et al., 2000; Shikanai et al., 2001; Kode et al., 2005; Rogalski et al., 2006). Moreover, in addition to the expected bands for the wild-type genome and the transformed genome, all lines displayed bands that presumably originated from undesired recombination events (Fig. 2A). Such unexpected hybridization products are usually due to recombination between duplicated expression elements that drive transgene expression as well as expression of the endogenous gene from which they were taken (Svab and Maliga, 1993; Rogalski et al., 2006; Gray et al., 2009; Li et al., 2011). After multiple rounds of plant regeneration in the presence of spectinomycin, a few homoplasmic lines were obtained. When these lines were tested for the presence of the C56T point mutation in the trnE gene by sequencing of amplified PCR products, all of them were found to lack the mutation, whereas the mutation was readily detectable in the heteroplasmic lines. Previous research had shown that deleterious mutations that are present in a heteroplasmic state are frequently eliminated by gene conversion (Khakhlova and Bock, 2006). Occurrence of gene conversion in conjunction with segregation to homoplasmy strongly suggests that the C56T mutation is lethal and remains heteroplasmic until the mutation is eliminated through gene conversion with a wild-type trnE copy.

Heteroplasmic Nt-EndMut plants displayed a characteristic leaf-loss phenotype in that the leaves showed various deformations and often lacked parts of the leaf blade (Fig. 3A). Very similar phenotypes were described previously for tobacco mutants with induced inactivation of plastid translation (Ahlert et al., 2003) and transplastomic plants in which essential components of the plastid gene expression machinery, including essential tRNA genes (Rogalski et al., 2008a; Alkatib et al., 2012a, 2012b), were targeted by reverse genetics (Rogalski et al., 2008b; Fleischmann et al., 2011). In these plants, segregation to homoplasmy leads to the loss of plastid gene expression, which in turn results in cell death. When this occurs early in leaf development, it can cause the loss of an entire cell lineage, which then results in misshapen leaves with large sectors of the leaf blade missing.

Lethality of the C56T mutation in tobacco suggests that it is incompatible with tRNA^Glu function in translation or tetrapyrrole biosynthesis (or both). Since the Nt-EndMut transplastomic plants were genetically unstable, accumulated undesired recombination products, and tended to lose the C56T point mutation in trnE, they were excluded from further analyses.

Phenotypes of Transplastomic Lines Overexpressing Wild-Type or Mutant trnE Alleles

When grown in soil under standard greenhouse conditions, transplastomic Nt-EctWt and Nt-EndWt plants were indistinguishable from wild-type tobacco plants (Fig. 3B). By contrast, transplastomic Nt-EctMut plants were severely retarded in growth (by >2 weeks) compared to the wild type (Fig. 3B). Moreover, leaves of Nt-EctMut plants that developed at early stages of plant growth displayed a pale or variegated phenotype. Interestingly, this phenotype was absent from leaves that developed at later stages of plant growth (Fig. 3B, bottom). The degree of chlorosis and the number of leaves showing the pale or variegated phenotype was quite variable, even among plants from the same transplastomic line. This suggests that, similar to the case for previously described variegation mutants (Miura et al., 2007; Wang et al., 2018), a physiological threshold effect may be involved in the development of the phenotype.

Seeds were readily obtained from all three sets of transplastomic lines. In order to ultimately confirm homoplasmy of the lines and assess the phenotype at the seedling stage under heterotrophic conditions, seeds were surface sterilized and germinated on synthetic (Suc-containing) culture media with or without spectinomycin (Fig. 4). Whereas wild-type seedlings were clearly sensitive to spectinomycin (i.e. they bleached out and ceased to grow), the progenies of all transplastomic lines showed uniform resistance to the antibiotic. Lack of segregation in the F1 generation is consistent with the uniparental inheritance of the plastid genome in tobacco (Greiner et al., 2015) and provides strong genetic evidence of homoplasmy of the transplastome. Nt-EctMut seedlings were pale green, but their phenotype was clearly distinguishable from spectinomycin-sensitive wild-type seedlings, which were completely white and showed arrested shoot and root growth. Also, the pale green appearance of Nt-EctMut seedlings was independent of the presence of the antibiotic (Fig. 4), confirming that these plants were also homoplasmic.

Expression of tRNA^Glu in Transplastomic Tobacco Plants

All constructs introduced into the plastid genome share the strong constitutive Nt Prrn promoter upstream of trnE (Fig. 1). To determine whether the presence of this additional promoter (in the Nt-EndWt lines) and/or the ectopic expression of an extra copy of the trnE gene (in the Nt-EctWt and Nt-EctMut lines) results in overexpression of tRNA^Glu in transplastomic chloroplasts, northern blot experiments were conducted with total RNA samples extracted from two independent transplastomic lines for each of the constructs. A radiolabeled oligonucleotide derived from the sequence of the mature tRNA^Glu was used as the hybridization probe for detection of trnE-specific transcripts (Fig. 5).

Interestingly, provision of the endogenous trnE with an additional Nt Prrn promoter (Nt-EndWt plants) did not lead to detectable overaccumulation of tRNA^Glu compared to the wild type (Nt-Wt; Fig. 5). By contrast, insertion of an additional copy of the trnE gene into the plastid genome resulted in a strong increase in tRNA^Glu abundance in transplastomic Nt-EctWt plants (Fig. 5). Surprisingly, transplastomic Nt-EctMut plants that ectopically express an additional trnE gene carrying the C56T mutation showed an additional hybridizing band in the northern blot analysis. This band was substantially larger (∼150–200 nucleotides) than that of the mature tRNA^Glu (76 nucleotides, including the 3′ CCA end that is added enzymatically after 3′ processing by ribonuclease [RNase] Z). This band was absent from the wild type and all other transplastomic lines (which only accumulated the mature tRNA) and could represent incompletely processed tRNA^Glu molecules (Fig. 5). This observation raised the interesting possibility that the C at position 56 of trnE plays a crucial role in the maturation of pre-tRNA^Glu to produce functional tRNA^Glu. The nucleotide at position 56, although part of the loop region of the TΨC arm, is known to form a Watson-Crick base pair with the nucleotide at position 19. This tertiary interaction plays an important role in tRNA elbow formation and is believed to be involved in the recognition of the tRNA by several protein factors, including the 5′ and 3′ processing enzymes RNase P and RNase Z (Zhang and Ferré-D’Amaré, 2016).

tRNA^Glu Processing in Transplastomic Nt-EctMut Plants

To test the idea that the C-to-U substitution at position 56 in tRNA^Glu results in impaired pre-tRNA processing, northern blot analyses were performed with hybridization probes specific to the 5′ leader and 3′ trailer sequences of the pre-tRNA^Glu. Bands were detected exclusively in the Nt-EctMut plants (Fig. 6, A and B) using both probes, suggesting that (1) tRNA processing in the wild-type and the transplastomic Nt-EndWt and Nt-EctWt plants is very efficient, since precursors and processing intermediates are undetectable; and (2) the C56U exchange affects maturation of tRNA^Glu in transplastomic Nt-EctMut plants at both the 5′ and 3′ ends. The size of the incompletely processed tRNA^Glu molecules was in the same range as the larger band detected in northern blots with the trnE probe (Fig. 5), although the leader and trailer probes also detected additional presumptive precursor RNAs of even larger size but much lower abundance (Fig. 6, A and B).

To ultimately confirm the 5′ and 3′ processing defects in the Nt-EctMut plants, RNA circularization experiments were undertaken. To this end, the regions corresponding to the bands seen in the RNA gel blot experiments were excised from urea-containing denaturing polyacrylamide gels, self-ligated, and reverse transcribed into complementary DNA. Subsequent PCR amplification of the junction of the ligated termini precisely identified the 5′ and 3′ ends. When this assay was performed with RNA samples from the wild type, the Nt-EndWt line, and the Nt-EctWt line, only fully processed tRNA^Glu molecules were detected (Fig. 6C). Interestingly, none of the processed tRNA^Glu molecules in the Nt-EctMut plants carried the point mutation, indicating that the faithfully processed tRNA molecules originate exclusively from the resident trnE gene copy in the plastid genome. Conversely, all incompletely processed tRNA^Glu molecules carried the point mutation (Fig. 6C; Supplemental Fig. S1). Whereas all immature tRNA molecules had 3′ extensions, only relatively few (2 of 20 sequenced clones) carried 5′ leader sequences (Fig. 6C; Supplemental Fig. S1). This finding potentially indicates that the C56U mutation largely blocks precursor tRNA processing by the 3′ processing enzyme RNase Z (Stern et al., 2010; Stoppel and Meurer, 2012), whereas 5′ processing by the chloroplast form of RNase P (PRORP; Gutmann et al., 2012; Zhou et al., 2015) is less severely affected.

Ectopic Expression of Mutated tRNA^Glu Affects Neither Aminoacylation of Mature tRNA^Glu Nor Processing of Other Plastid tRNAs

Having uncovered processing defects in the mutated tRNA^Glu (tRNA^Glu_C56U), we next considered the possibility that accumulation of immature tRNA^Glu molecules interferes with tRNA aminoacylation in the Nt-EctMut plants, for example, by inhibiting the corresponding aminoacyl-tRNA synthetase. We therefore determined the tRNA aminoacylation state by acidic urea PAGE of acylated versus deacylated tRNAs. These experiments revealed that the aminoacylation state of tRNA^Glu was unaffected in all transplastomic lines, including the Nt-EctMut lines ectopically expressing the mutant tRNA^Glu (Fig. 7A).

Our sequencing analyses had indicated that the Nt-EctMut lines efficiently processed the tRNA^Glu molecules transcribed from the wild-type trnE gene copy, whereas processing of the mutated tRNA^Glu_C56U was selectively impaired (Fig. 6C; Supplemental Fig. S1). This finding argues against the possibility that the mutated tRNA interferes with the general activity of PRORP and/or RNase Z in the chloroplast. To further verify this assumption, the processing status of two other chloroplast tRNAs, namely tRNA^Phe and tRNA^Arg, was investigated by northern blot analysis. As expected, only mature tRNAs could be detected in all transplastomic lines investigated (Fig. 7B), confirming that accumulation of immature tRNA^Glu_C56U does not generally interfere with the activity of PRORP and/or RNase Z in the Nt-EctMut lines.

Accumulation of Immature tRNA^Glu_C56U Does Not Affect Chloroplast Protein Biosynthesis

Our analyses of tRNA processing and aminoacylation suggested that processes upstream of tRNA^Glu utilization in chloroplast translation and ALA synthesis are unlikely to be causally responsible for the mutant phenotype observed in Nt-EctMut plants. Therefore, we subsequently examined whether expression of the mutated trnE gene negatively affects chloroplast protein biosynthesis.

Polysome analyses were performed to determine whether the pale-leaf phenotype of Nt-EctMut plants (Fig. 3B) was the result of interference of tRNA^Glu_C56U with the translation by chloroplast ribosomes. Recent work has suggested that the abundance of plastid-encoded chlorophyll-binding proteins is adjusted to chlorophyll availability at the level of proteolysis rather than de novo synthesis (Zoschke et al., 2017). Consequently, any observable effect on plastid translation rates should be a direct consequence of the interaction of mutated tRNA^Glu molecules with translating ribosomes rather than a secondary effect from the pigment deficiency of Nt-EctMut plants. Northern blot analysis of RNAs extracted from fractionated polysome gradients revealed unaltered ribosome association of the mRNAs of two key genes involved in photosynthesis: rbcL (encoding the large subunit of Rubisco) and psbD (encoding the D2 protein of the PSII reaction center; Fig. 8). These results indicate that accumulation of the mutated tRNA^Glu_C56U does not affect protein synthesis rates in the chloroplast.

The Presence of Precursor tRNA^Glu_C56U Strongly Reduces ALA Synthesis Rates

Having excluded tRNA metabolism, aminoacylation, and translation as possible causes of the mutant phenotype of Nt-EctMut lines, the possibility remained that the accumulation of mutated and incompletely processed tRNA^Glu molecules inhibits tetrapyrrole biosynthesis. Such an inhibitory effect could come from binding of immature tRNA^Glu_C56U to GluTR, the enzyme that uses tRNA^Glu as substrate for ALA synthesis. As no mature tRNA^Glu_C56U accumulates and 3′ processing is completely blocked (Fig. 6C; Supplemental Fig. S1), the tRNA^Glu_C56U can neither undergo 3′ CCA addition nor be charged with Glu. Thus, it seems clear that tRNA^Glu_C56U cannot serve as substrate to form Glu-1-semialdehyde in the committed step of tetrapyrrole biosynthesis. However, if the mutated and immature tRNA molecules would nonetheless bind to GluTR, they could either block the activity of the enzyme or, alternatively, destabilize the enzyme and condemn it to rapid degradation.

To investigate these possibilities, ALA synthesis rates were measured in wild-type, Nt-EctWt, and Nt-EctMut plants. Although the transplastomic Nt-EctWt plants strongly overexpress tRNA^Glu (Fig. 5), tRNA overaccumulation did not result in a significant stimulation of ALA synthesis (Fig. 9A), indicating that tRNA^Glu availability does not limit tetrapyrrole biosynthesis. By contrast, ALA synthesis was strongly reduced in Nt-EctMut plants, with synthesis rates reaching only ∼20% of those in wild-type and Nt-EctWt plants (Fig. 9A). This finding raises the interesting possibility that the ectopically expressed tRNA^Glu_C56U inhibits ALA synthesis by GluTR.

To examine whether tRNA^Glu_C56U affects GluTR accumulation (e.g. by forming a dead-end complex that triggers proteolytic degradation of the enzyme), immunoblot analysis with GluTR-specific antibodies was performed. These experiments revealed that GluTR protein levels were very similar in all investigated plant lines (Fig. 9B), indicating that tRNA^Glu_C56U does not exert its inhibitory action at the level of protein stability. Instead, this finding suggests that inhibition occurs at the level of enzyme activity, presumably by the immature and uncharged tRNA^Glu_C56U molecules forming a stable dead-end complex with GluTR.

Finally, we examined accumulation of two of the end products of tetrapyrrole biosynthesis, chlorophyll a and chlorophyll b, in the transplastomic lines (Supplemental Fig. S2). We particularly wanted to know whether the green leaves appearing later in the development of Nt-EctMut plants (Fig. 3B) show full recovery at the level of chlorophyll accumulation. This was indeed the case for both chlorophyll a and chlorophyll b (Supplemental Fig. S2), suggesting that rapid growth and high demand for chlorophylls early in development cause the pale-leaf phenotype, whereas the reduced demands later in development can be fully satisfied by the reduced synthesis capacity in Nt-EctMut plants. Also, none of the other transplastomic lines displayed any alteration in chlorophyll accumulation (Supplemental Fig. S2), consistent with our finding that ALA synthesis rates are unaffected by overexpression of tRNA^Glu (Fig. 9A).

DISCUSSION

In the course of this work, we generated a set of transplastomic tobacco plants to alter the expression of chloroplast tRNA^Glu and introduced a point mutation in the trnE gene, the latter of which had been described in a spontaneous mutant of E. gracilis and reported to uncouple plastid translation from tetrapyrrole biosynthesis (Stange-Thomann et al., 1994). The C56T mutation in the trnE gene of the Euglena mutant led to reduced accumulation of tRNA^Glu, but the mutant tRNA appeared to be faithfully aminoacylated and therefore was suggested to participate in plastid protein biosynthesis. By contrast, the mutant tRNA^Glu did not support ALA synthesis by the C₅ pathway (Stange-Thomann et al., 1994).

Our work reported here in the seed plant tobacco did not confirm many of the findings reported for the Euglena mutant. Most importantly, when introduced into the trnE gene of tobacco plastids, the C56T mutation was incompatible with cell survival. Transplastomic lines remained heteroplasmic and displayed the malformed leaf phenotype that is typical of (heteroplasmic) loss-of-function mutants of essential genes in the plastid genome (Figs. 2A and 3A). Notably, this phenotype has been observed for all genes encoding essential components of the plastid translational machinery when they are targeted by reverse genetics, including ribosomal proteins and tRNAs (Ahlert et al., 2003; Alkatib et al., 2012b; Tiller and Bock, 2014; Zoschke and Bock, 2018). Our analysis of tRNA^Glu_C56U expression in transplastomic plants revealed defective tRNA processing and a complete absence of correctly processed tRNA molecules that could be aminoacylated and serve as substrates in translation elongation (Fig. 6C; Supplemental Fig. S1). This finding suggests that the C56T mutation represents a trnE loss-of-function allele, thus providing a straightforward explanation for the phenotype of the Nt-EndMut plants (Fig. 3A). However, since the C₅ pathway is likely also essential in seed plants (Czarnecki et al., 2011a; Richter et al., 2019), the reason why Nt-EndMut plants remain heteroplasmic and segregation to homoplasmy causes cell death and leaf malformation could be twofold.

It is currently unclear why processed and aminoacylated tRNA^Glu accumulated in the Euglena mutant. All sequenced Euglena strains contain a single copy of the trnE gene in their chloroplast genome. The tRNA processing machinery is highly conserved in that 5′ processing is conducted by RNase P and 3′ processing is performed by RNase Z in nearly all organisms. Also, the tertiary interaction nucleotide C₅₆ is involved in its role in tRNA elbow formation, and its involvement in the interaction with the tRNA processing enzymes appears well conserved (Zhang and Ferré-D’Amaré, 2016). It therefore would be surprising if the same mutation completely blocks tRNA maturation in tobacco plastids but allows for correct processing in Euglena plastids. In this context, it should be noted that the chloroplast genome of the Euglena mutant was not characterized beyond PCR amplification of the known genomic trnE locus. Thus, the presence of a second wild type-like copy of the trnE gene that provides the functional tRNA^Glu identified in the aminoacylation assays (Stange-Thomann et al., 1994), along with an uncharacterized genetic defect that causes the chlorophyll deficiency, cannot be ruled out. Such gene duplication could be the result of a genomic rearrangement that went undetected in the studies conducted with the mutant. It is noteworthy in this regard that gene duplications in Euglena plastids are more stable than in seed-plant plastids. This is suggested, for example, by the presence of the rRNA operon as tandem repeats (Hallick et al., 1993), and it may be due to the absence (or very low activity) of homologous recombination (Doetsch et al., 2001). Alternatively, it is possible that the correctly processed tRNA^Glu is encoded in the nuclear genome of Euglena and imported into plastids. However, whereas tRNA import is common in plant mitochondria (Duchêne et al., 2009), no direct evidence for the presence of it in plastids has been obtained to date, although some indirect evidence comes from the existence of highly reduced plastid genomes that do not encode a complete set of tRNAs (Morden et al., 1991; Delannoy et al., 2011). In view of all this, a thorough reinvestigation of the Euglena mutant, if still available, would be very worthwhile.

In this study, we also tested two different approaches toward overexpression of the chloroplast tRNA^Glu. Since tRNA^Glu serves as substrate of the rate-limiting initial reaction of tetrapyrrole biosynthesis (Brzezowski et al., 2015; Richter et al., 2019), it was conceivable that tRNA^Glu provision limits flux through the pathway. In fact, expression of tRNA^Glu has been shown to be developmentally regulated (Hanaoka et al., 2005). Moreover, tRNA^Glu was proposed to feedback-regulate chloroplast transcription by directly binding to one of the two RNA polymerases in plastids (Hanaoka et al., 2005). The two overexpression strategies we pursued produced transplastomic plants that were indistinguishable from wild-type plants (Fig. 3B). Whereas insertion of an additional strong promoter upstream of the trnE gene in its native genomic location did not result in appreciable overaccumulation of tRNA^Glu, integration of an extra gene copy into a distant genomic location resulted in strong overexpression of the tRNA (Fig. 5). However, this did not result in a measurable effect on ALA synthesis rates (and chlorophyll accumulation) in the transplastomic Nt-EctWt plants (Fig. 9A; Supplemental Fig. S2), suggesting that tRNA^Glu does not limit tetrapyrrole biosynthesis, at least not under standard growth conditions.

Ectopic expression of tRNA^Glu_C56U produced plants that exhibit a strong pigment-deficient phenotype at the seedling stage (Fig. 3B). Interestingly, as the plants matured, the mutant phenotype became less severe and the leaves appearing later in development became progressively greener until they finally were nearly indistinguishable from wild-type leaves (Fig. 3B, lower, young Nt-EctMut leaves). Nt-EctMut plants showed unaltered expression of the endogenous trnE gene and accumulated wild-type levels of fully processed and faithfully aminoacylated tRNA^Glu (Figs. 5 and 7A). In addition, the plants accumulated the mutant tRNA^Glu_C56U to a similar level; however, this tRNA remained unprocessed and therefore cannot be aminoacylated. These results suggest that the pigment-deficient mutant phenotype of the Nt-EctMut plants is the result of a dominant-negative effect exerted by the ectopic expression of tRNA^Glu_C56U. We showed that GluTR, the enzyme that utilizes tRNA^Glu to synthesize ALA, accumulates to unaltered levels in Nt-EctMut plants. However, ALA synthesis rates were drastically reduced in the transplastomic plants (Fig. 9), indicating that the immature tRNA^Glu_C56U molecules inhibit the enzyme, presumably by binding to it and forming a dead-end complex.

Our data suggest that inhibition of GluTR activity to ∼20% of the wild-type levels causes ALA synthesis to limit biogenesis of the photosynthetic apparatus. This bottleneck is particularly severe in young developing plants that have high rates of photosystem biogenesis (Hojka et al., 2014; Schöttler and Tóth, 2014; Armarego-Marriott et al., 2019) and therefore a high demand for chlorophylls (Fig. 3B). By contrast, when growth decelerates later in development, the low ALA synthesis rates are sufficient to satisfy the needs of the chloroplast, resulting in a striking developmental gradient in the leaf phenotypes. Interestingly, the pale leaves that developed under ALA limitation do not recover and never regreen (Fig. 3B). This is likely due to loss of the photosystem assembly capacity that is known to occur as leaves mature (Schöttler et al., 2011; Hojka et al., 2014; Schöttler and Tóth, 2014).

The charging of tRNA^Glu with the amino acid Glu occurs by the two-step reaction typical of aminoacyl-tRNA synthetases. l-Glu is first activated by ATP to form glutamyl-AMP and then transferred to the acceptor end (i.e. the 3′ CCA end) of the tRNA^Glu molecule. The aminoacylation of mature tRNA^Glu was virtually complete in all transplastomic lines generated in this study (Fig. 7A), suggesting that even strong overexpression of the tRNA does not exceed the capacity of the plastid tRNA^Glu synthetase. Consistent with this finding, chloroplast translation was unaffected in all lines (Fig. 8). As unprocessed and uncharged tRNAs are not taken into the ribosome, it is unsurprising that accumulation of immature tRNA^Glu_C56U does not interfere with translation elongation.

In summary, our work here shows that the C56U substitution in chloroplast tRNA^Glu, unlike that previously reported for Euglena (Stange-Thomann et al., 1994), does not uncouple plastid protein biosynthesis from tetrapyrrole biosynthesis in tobacco. Instead, the C56T substitution represents a lethal mutation. It inhibits tRNA processing and in this way prevents translation of Glu codons. Moreover, our work has uncovered an unexpected inhibitory activity of unprocessed tRNA^Glu on GluTR. Importantly, whereas overexpression of tRNA^Glu does not stimulate ALA biosynthesis, reduction of GluTR activity through inhibition by tRNA^Glu precursors makes ALA synthesis limiting to an extent that chlorophyll biosynthesis in early plant development cannot satisfy the demands of photosystem biogenesis.

MATERIALS AND METHODS

Plant Material and Growth Conditions

Sterile tobacco (Nicotiana tabacum ‘Petit Havana’) plants were grown on Murashige and Skoog (MS) medium (Murashige and Skoog, 1962) containing 3% (w/v) Suc. The growth chambers had a light intensity of 50 µE m⁻² s⁻¹ and a photoperiod of 16 h light at 24°C and 8 h dark at 22°C. Soil-grown plants were raised in growth chambers with a light intensity of 350 µE m⁻² s⁻¹ and a photoperiod of 16 h light at 24°C and 8 h dark at 22°C.

Construction of Plastid Transformation Vectors

For construction of transformation vector pNt-EctWt (Fig. 1), PCR was performed to amplify (1) the Nt Prrn promoter using primer pair oSAA57_trnE_F3/oSAA58_ trnE_R3, (2) the trnE gene from the tobacco plastid genome and its promoter using primer pair oSAA59_trnE_F4/oSAA65_ trnE_R8, and (3) the 3′ untranslated region from the plastid rbcL gene of Chlamydomonas reinhardtii (Cr TrbcL; Lu et al., 2017) with primer pair oSAA64_trnE_F7/oSAA66_trnE_R7 (Supplemental Table S1). Overlap extension PCR was performed to generate the Nt Prrn-trnE-Cr TrbcL fusion product. The resulting amplification product was digested with the restriction enzymes SacI and HindIII and cloned into a modified pRB96 vector (Ruf et al., 2001; Wurbs et al., 2007) cut with SacI and HindIII.

For construction of transformation vector pNt-EctMut, site-directed mutagenesis was performed on the fusion product generated for construction of pNt-EctWt to introduce the desired point mutation into trnE. To this end, PCRs were performed with the Nt Prrn-trnE-Cr TrbcL fusion product as template using the primer pairs oSAA57_trnE_ F3/oSAA62_trnE_ R5 and oSAA61_trnE_ F5/oSAA66 _trnE_R7 (Supplemental Table S1). Overlap extension PCR was performed to produce the Nt Prrn-trnE_C56U-Cr TrbcL fusion product. This amplification product was cloned in the pRB96-derived plastid transformation vector as described above.

For construction of vectors pNt-EndWt and pNt-EndMut, an intermediate vector (pSAA26) was generated. To this end, PCR was performed to amplify (1) an ∼1-kb region upstream of trnE using tobacco total DNA as template and the primer pair oSAA53_trnE_F1 (introducing a SacI site)/oSAA54_ trnE_R1, and (2) an aadA expression cassette (Fig. 1A) with the primer pair oSAA55_trnE_F2/oSAA56_trnE_R2 (introducing a HindIII site; Supplemental Table S1). Overlap extension PCR was performed to generate a fusion product, which was subsequently digested with SacI and HindIII and cloned into plasmid pBS KS(+) cut with SacI and HindIII.

For construction of pNt-EndWt, PCRs were performed to amplify (1) the Nt Prrn sequence with primer pair oSAA57_trnE_F3 (introducing a HindIII site)/oSAA58_ trnE_R3, and (2) the trnE-trnY-trnD region of the tobacco plastid genome and an ∼500-bp region downstream with primer pair oSAA59_trnE_ F4/oSAA60_trnE_R4 (introducing an XhoI site; Supplemental Table S1). Overlap extension PCR was performed to obtain a fusion product, which was then digested with HindIII and XhoI and cloned into pSAA26 cut with HindIII and XhoI. pNt-EndMut was constructed via site-directed mutagenesis performed on the fusion product produced for construction of pNt-EndWt to introduce the desired point mutation into the trnE gene sequence. To this end, PCRs were performed with the fusion product as template and the primer pairs oSAA57_trnE_F3 (introducing a HindIII site)/oSAA62_ trnE_R5 and oSAA61_trnE_F5/oSAA60_ trnE_R4 (introducing an XhoI site; Supplemental Table S1). Overlap extension PCR yielded a new fusion product, which was digested with HindIII and XhoI and cloned into pSAA26 cut with HindIII and XhoI.

Plastid Transformation of Tobacco and Isolation of Transplastomic Lines

Biolistic transformation of plastids was conducted according to published protocols (Svab and Maliga, 1993). Briefly, young tobacco leaves from plants grown under sterile conditions were bombarded with plasmid DNA-coated gold particles using the DuPont PDS-1000/He biolistic gun (Bio-Rad). After bombardment, leaves were cut into small pieces ∼5 × 5 mm in size, placed onto MS-based plant regeneration medium containing 500 mg L⁻¹ spectinomycin, and incubated for 2 to 3 months. Primary spectinomycin-resistant shoots or calli were subjected to two to three additional rounds of regeneration in the presence of the antibiotic to enrich the transformed plastid genome and isolate homoplasmic transplastomic lines (Bock, 2001). Finally, regenerated shoots were rooted and propagated on hormone-free MS medium containing 500 mg L⁻¹ spectinomycin, then transferred to soil and grown to maturity under standard greenhouse conditions with a photoperiod of 16 h light at 25°C and 8 h darkness at 20°C.

Seed Assays

To test for maternal inheritance of the engineered plastid genomes, T1 generation seeds from the transplastomic lines were surface-sterilized and sown on MS medium containing 500 mg L⁻¹ spectinomycin. The presence of a homogeneous antibiotic-resistance progeny confirmed homoplasmy of the transformed chloroplast genome (Bock, 2001).

Isolation of Nucleic Acids and Gel Blot Analyses

Leaf tissue snap-frozen in liquid nitrogen was used for the isolation of nucleic acids. For extraction of total cellular DNA, a cetyltrimethylammoniumbromide-based method was used (Doyle and Doyle, 1990). Total plant RNA was extracted using the peqGOLD TriFast reagent (Peqlab). For RFLP analysis, samples of 3 μg total DNA were digested with appropriate restriction enzymes, separated in 1% (w/v) agarose gels by electrophoresis, and transferred onto Hybond nylon membranes (GE Healthcare) by capillary blotting. For northern blot analysis, total cellular RNA was electrophoretically separated in 2% (w/v) agarose gels containing 5% (w/v) formaldehyde, and blotted onto Hybond nylon membranes (GE Healthcare).

Probes for RFLP analysis were amplified by PCR using primers listed in Supplemental Table S1, followed by agarose gel electrophoresis and purification of the PCR products from excised gel slices using the NucleoSpin Extract II kit (Macherey-Nagel). Purified probes were radiolabeled with [α-³²P]dCTP by random priming using the Multiprime DNA labeling system (GE Healthcare). Antisense oligonucleotides (synthesized by Eurofins Genomics and listed in Supplemental Table S1) were used as probes for northern blot analyses. For radioactive labeling, 20 pmol of the oligonucleotide were incubated with 10 units of T4 polynucleotide kinase (New England Biolabs) and 30 µCi of [γ-³²P]ATP for 30 min at 37°C to allow 5′ end labeling. All hybridizations, except those with the trnE 5′ leader and 3′ trailer sequences as probes, were performed at 65°C using standard protocols (Church and Gilbert, 1984). For probes derived from trnE 5′ leader and 3′ trailer sequences, 55°C was used as the hybridization temperature. Signals were analyzed using a Typhoon Trio+ variable mode imager (GE Healthcare).

Circular RT-PCR

Samples of 8 µg total RNA were electrophoretically separated in 8% (w/v) urea-containing polyacrylamide gels. Bands corresponding to the sizes of processed and precursor tRNAs were cut out and treated with the enzyme RppH (New England Biolabs) to remove the pyrophosphate from the 5′ end of triphosphorylated RNAs. Subsequently, the monophosphorylated RNA molecules were circularized using T4 RNA ligase (New England Biolabs), followed by reverse transcription with primer oSAA105_tRNA-E_R that anneals specifically to the trnE gene sequence (Supplemental Table S1). RT-PCR was performed with complementary DNA as template and the primer pair oSAA101_tRNA-E_F/oSAA105_tRNA-E_R, followed by cloning of the amplification products using the TOPO TA Cloning Kit (Invitrogen). The clones obtained were analyzed by colony PCR and DNA sequencing.

tRNA Aminoacylation Assays

Aminoacylation of tRNA^Glu was analyzed as described previously (Zhou et al., 2013). A specific hybridization probe for tRNA^Glu was prepared as described above.

Polysome Analysis

Polysome analyses and puromycin controls were performed essentially as described previously (Barkan, 1998; Rogalski et al., 2008a). RNA was extracted from gradient fractions and the RNA pellet was resuspended in 30 µL RNase-free water. Aliquots of 5 µL per fraction were analyzed by northern blotting.

Determination of ALA Synthesis Rate and Measurement of Chlorophyll Accumulation

ALA synthesis rates were determined according to published protocols (Czarnecki et al., 2011b). Chlorophyll concentrations were measured in frozen tissue of defined fresh weight by extraction with 100% (v/v) acetone following published procedures (Porra et al., 1989; Lu et al., 2017).

Protein Isolation and Immunoblot Analyses

Samples of 100 mg ground frozen leaf tissue were used for extraction of total cellular protein using a phenol-based method (Cahoon et al., 1992) with the minor modification that instead of phenyl-methanesulfonyl fluoride, cOmplete Protease Inhibitor Cocktail (Roche) was added. The protein pellets were dissolved in 1% (w/v) SDS and the protein concentration was determined with the BCA Protein Assay kit (Pierce Instruments). For immunoblot analysis, samples of 20 µg protein were separated by 10% (w/v) SDS-PAGE and blotted onto polyvinylidene difluoride membranes (GE Healthcare). Membranes were treated with blocking buffer (1× Tris-buffered saline, 0.1% [v/v] Tween 20, and 1% [w/v] bovine serum albumin) for 1 h and then incubated with a rabbit polyclonal antiglutamyl-tRNA reductase antibody (Agrisera) for another hour. After incubation with antirabbit secondary antibody (Agrisera), immunodetection was performed with the ECL Prime system (GE Healthcare). For actin detection, the membranes were stripped using Thermo Scientific Restore Plus western blot stripping buffer (Fisher Scientific) according to the manufacturer’s instructions, followed by treatment with blocking buffer for 1 h, incubation in mouse polyclonal antiactin antibody (Sigma) for 1 h, incubation in antimouse secondary antibody (Sigma), and detection using the ECL Prime system (GE Healthcare).

Accession Numbers

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession number NC_001879.2 (Nicotiana tabacum, complete plastid genome).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Determination of the ends of incompletely processed trnE transcripts in the Nt-EctMut mutant.

• Supplemental Figure S2. Quantification of chlorophyll a and b contents in mature transplastomic plants (12-leaf stage) grown under standard greenhouse conditions in a diurnal cycle of 16 h light at 25°C and 8 h darkness at 20°C.

• Supplemental Table S1. List of synthetic oligonucleotides used as PCR primers in this study.

ACKNOWLEDGMENTS

We thank Pierre Endries and Davina Störmer for help with tissue culture and transformation, Dr. Alexander Hertle for microscopy, the Max-Planck-Institut für Molekulare Pflanzenphysiologie GreenTeam for plant cultivation, and Fabio Moratti and Dr. Reimo Zoschke for helpful discussion.

LITERATURE CITED

Author notes