Inactivation of Organellar Glutamyl- and Seryl-tRNA Synthetases Leads to Developmental Arrest of Chloroplasts and Mitochondria in Higher Plants*

Aminoacyl-tRNA synthetases (ARSs) are key enzymes involved in protein translation, and both cytosolic and organellar forms are present in the genomes of eukaryotes. In this study, we investigated cellular effects of depletion of organellar forms of ARS using virus-induced gene silencing (VIGS) in Nicotiana benthamiana. VIGS of NbERS and NbSRS, which encode organellar GluRS and SerRS, respectively, resulted in a severe leaf-yellowing phenotype. The NbERS and NbSRS genes were ubiquitously expressed in plant tissues, and induced in response to light. Green fluorescent protein (GFP) fusion proteins of the full-length glutamyl-tRNA synthetase (ERS) and seryl-tRNA synthetase (SRS) of Arabidopsis and GFP fusions to the N-terminal extension of these proteins were all dualtargeted to chloroplasts and mitochondria. At the cell level, depletion of NbERS and NbSRS resulted in dramatically reduced numbers of chloroplasts with reduced sizes and chlorophyll content. The numbers and/or physiology of mitochondria were also severely affected. The abnormal chloroplasts lacked most of the thylakoid membranes and appeared to be degenerating, whereas some of them showed doublet morphology, indicating defective chloroplast division. Pulse-field gel electrophoresis analyses demonstrated that chloroplast DNA in subgenomic sizes is the predominant form in the abnormal chloroplasts. Interestingly, despite severe abnormalities in chloroplasts and mitochondria, expression of many nuclear genes encoding chloroplastor mitochondria-targeted proteins, and chlorophyll biosynthesis genes remained unchanged in the ERS and SRS VIGS lines. This is the first report to analyze the effect of ARS disruption on organelle development in plants.

Aminoacyl-tRNA synthetases (ARSs) 5 play a critical role in protein synthesis by catalyzing the addition of amino acids to their cognate tRNAs (1,2). The specificity of aminoacyl-tRNA synthesis in pairing the appropriate tRNAs and amino acids is a key determinant in faithful transmission of genetic information. Despite the crucial function of ARSs in protein synthesis, there are only limited reports on function and regulation of the ARS genes and proteins in plants (3). Protein synthesis in plants takes place in the cytosol, mitochondria, and chloroplasts, and these compartments do not require full sets of unique ARSs encoded by separate nuclear genes. In general, plant ARSs are classified into two groups based on their substrate specificity: the cytosolic enzymes that most efficiently aminoacylate plant or yeast cytosolic tRNAs, and the organellar enzymes that aminoacylate organelle or Escherichia coli tRNAs (4).
Recent research has uncovered an extensive degree of sharing of ARS isoforms between compartments, between cytosol and mitochondria or between plastids and mitochondria (5). For example, the cytosolic and mitochondrial leucyl-tRNA synthetases of Phaseolus vulgaris aminoacylate cytosolic and mitochondrial tRNAs Leu with equal efficiency but do not aminoacylate chloroplast and E. coli tRNAs Leu (6). In addition, AlaRS, ValRS, and ThrRS of Arabidopsis are dual targeted to both the cytosol and mitochondria, and the same gene encodes both the mitochondrial and the cytosolic enzyme in each case (7,8). Dual targeting to mitochondria and chloroplasts has also been observed for HisRS (9), MetRS (10,11), AsnRS (10), CysRS (10), and one of the GlyRS enzymes (12) in Arabidopsis.
Phenotypes of several mutants with defects in ARS gene structure or expression have been characterized in Arabidopsis. A tDNA insertion mutant of the cytosolic and mitochondrial AlaRS showed an embryonic lethal phenotype (13). Embryo development of this mutant was arrested at the globular stage with altered patterns of cell division and differentiation. Altered expression of a cytosolic ValRS in plant tissues also caused severe defects in early embryogenesis in the twn2 mutant in Arabidopsis (14). In this mutant, development of apical cells is arrested at a very early stage, and the basal cells proliferate abnormally, giving rise to multiple embryos. Taken together, these results show that cytosolic ARS activity is indispensable in plant cell viability and development. To date, there has been only one report regarding the function of an organelle-specific ARS (15). The insertion mutation of the organellar GlyRS gene was lethal, arresting embryo growth between the globular and heart stages of embryo development in Arabidopsis. An N-terminal fragment of the GlyRS protein was able to direct a marker protein into both chloroplasts and mitochondria (12). Thus, normal organellar ARS function is obviously essential for plant embryogenesis. However, there has been no report directly demonstrating how loss of organellar ARS function affects the development and physiology of chloroplasts and mitochondria, and how it in turn affects plant cells in which the organelles reside.
In this study, we investigated cellular functions of organellar ARSs by generating reduced-expression mutants using virus-induced gene silencing (VIGS), thereby circumventing the embryonic-lethal phenotypes associated with the complete loss of ARS function. VIGS of NbERS and NbSRS, encoding the Nicotiana benthamiana organellar forms of GluRS and SerRS, respectively, severely affected the numbers, morphology, and physiology of both chloroplasts and mitochondria. These data are consistent with the import of the Arabidopsis orthologs of NbERS and NbSRS into both chloroplasts and mitochondria and suggest that NbERS and NbSRS are responsible for the GluRS and SerRS activity required for translation in both organelles.

EXPERIMENTAL PROCEDURES
Virus-induced Gene Silencing-Various cDNA fragments of NbERS and NbSRS were PCR-amplified and cloned into the pTV00 vector containing a part of the TRV genome (16) using BamHI and ApaI sites. VIGS was carried out as described (16 -18). The fourth leaf above the infiltrated leaf was used for RT-PCR and cytological analyses.
Measurement of Chlorophyll Content-The fourth leaf above the infiltrated leaf was collected from the VIGS plants and boiled in 95% ethanol at 80°C for 30 min to extract chlorophyll. Chlorophyll concentration per unit fresh weight was calculated as described before (20).
Analysis of Starch Content-Leaves were harvested from the VIGS lines at 20 days after infiltration, and bleached in 80% (v/v) ethanol. After rinsing with double-distilled water, the leaves were stained with Lugol's iodine staining reagent (Sigma) and briefly destained with water.
Histochemical Analyses-Tissue sectioning and microscopy were carried out as described previously (17), using the fourth leaf above the infiltrated leaf from the VIGS lines.
Confocal Laser Scanning Microscopy-For TMRM (tetramethylrhodamine methyl ester, Molecular Probes) staining of mitochondria, TMRM was added into the leaf protoplasts at the final concentration of 200 nM. After incubation for 1-2 min at 25°C, protoplasts were transferred to wells on microscope slides and examined by confocal microscopy (Carl Zeiss LSM 510) with a BP560-615 (543 nm excitation, 560 -615 nm emission) optical filter to visualize the red fluorescent probe. Quantitative images were captured, and data were analyzed using the LSM 510 software (version 2.8). Chlorophyll autofluorescence was observed using an LP650 (excitation 488 nm, emission 650 nm) optical filter. MitoTracker Green FM (Molecular Probes) staining of protoplasts and 4Ј,6-diamidino-2-phenylindole staining of chloroplasts were carried out, as described previously (18).
PFGE Analysis-Protoplasts were prepared from the fourth to sixth leaves above the infiltrated leaf of the VIGS lines and counted. Roughly equal numbers of protoplasts from the VIGS lines were embedded, and pulse-field gel electrophoresis was carried out as described (18,21). The rbcL cDNA was used as a probe.

Virus-induced Gene Silencing of the NbERS and NbSRS Caused Severe
Leaf-yellowing Phenotypes-Functional genomics has been carried out in N. benthamiana using TRV-based virus-induced gene silencing (VIGS) to assess functions of signaling genes and genes that likely cause embryo or seedling lethality when their expression is suppressed. VIGS is based on the phenomenon that gene expression is suppressed in a sequence-specific manner by infection with viral vectors carrying host genes (22) and has been proved to be a powerful tool to analyze embryoor seedling-lethal genes (17,18). Using a VIGS screen, we found that gene silencing of the N. benthamiana NbERS and NbSRS genes, encoding isoforms of GluRS and SerRS, respectively, causes severe leaf yellowing and abnormal leaf morphology, while maintaining almost normal plant growth.
The full-length NbERS cDNA was obtained by 5Ј-rapid amplification of cDNA ends. The full-length NbSRS cDNA was obtained by PCR based on the sequence in the TIGR data base (accession number TC8134). NbERS and NbSRS encode polypeptides of 569 and 506 amino acids, respectively, corresponding to theoretical molecular masses of 63410.40 and 56266.37 Da. The NbERS and NbSRS genes exhibited high sequence similarity to the corresponding genes of prokaryotes, particularly cyanobacterial species, indicating that they encode an organellar isoform of GluRS and SerRS. The Arabidopsis data base contains one sequence encoding an organellar GluRS (GenBank TM accession number NP_201210), designated AtERS, and a sequence for an organellar SerRS (AAL16281), designated AtSRS. When the amino acid sequences of the organellar ERS and SRS genes from plants were aligned with the corresponding genes from prokaryotes, a notable difference between the plant proteins and the corresponding bacterial proteins is that plant proteins contain a long extension at their N terminus, which has features of chloroplast and mitochondria targeting signals (23). VIGS with these constructs all resulted in a similar yellowing phenotype in newly emerged leaves, whereas overall plant growth was normal ( Fig.  1, A and B). During flower formation, the sepals became yellow in the ERS plants (Fig. 1A). However, except that, flower development was normal in both ERS and SRS VIGS lines (data not shown). In contrast, VIGS of NbIRS, which encodes an organellar form of IleRS resulted in only a mild leaf-yellowing phenotype, when a 1.5-kb cDNA fragment was used for VIGS (Fig. 1C). As a control, we also carried out VIGS of c-IRS, which encodes a cytosolic form of IleRS, and it resulted in arrested plant growth and severe developmental abnormalities in newly emerged leaves, distinct from the VIGS phenotypes of the organellar ARSs (Fig. 1D).

Suppression of the NbERS and NbSRS Transcripts by VIGS-To
Effects of gene silencing on the amounts of endogenous NbERS and NbSRS mRNAs were examined by semiquantitative RT-PCR (Fig. 1, E and F). RT-PCR using the ERS-A primers (indicated in Fig. 1A) produced significantly reduced amounts of PCR products in the yellow sectors of the leaves from the TRV:ERS(2) line, indicating that the endogenous level of the NbERS transcripts is greatly reduced in these plants (Fig. 1E). The same primers detected high levels of viral genomic transcripts containing the N-terminal region of NbERS in the TRV: ERS(1) and TRV:ERS(3) lines. Similarly, RT-PCR using the ERS-B prim-  Fig. 1A) produced significantly reduced amounts of PCR products in the yellow sectors of leaves from the TRV:ERS(1) line, whereas the same primers detected high levels of viral genomic transcripts containing the C-terminal region of NbERS in the TRV:ERS(2) and TRV:ERS (3) lines. In all of the ERS VIGS lines, the transcript level of NbSRS, NbIRS, and actin remained constant (Fig. 1E).

ers (indicated in
Silencing of NbSRS was also examined by semiquantitative RT-PCR (Fig. 1F). RT-PCR using the SRS-A and SRS-B primers produced significantly reduced amounts of PCR products in yellow sectors of leaves from the TRV:SRS(3) and TRV:SRS(2) lines, respectively, indicating that the endogenous level of the NbSRS transcripts is significantly reduced in those plants. In all of the SRS VIGS lines, the transcript levels of NbERS, NbIRS, and actin stayed constant (Fig. 1F). These results clearly show that VIGS using the NbERS and NbSRS cDNA fragments specifically suppresses expression of these genes. The transcripts levels of NbERS and NbSRS in TRV were identical to those in wild-type, indicating that TRV infection did not affect the expression of NbERS and NbSRS (supplementary Fig. S1A). Although VIGS of NbIRS caused only mild phenotypes compared with silencing of NbERS and NbSRS, the degree of gene silencing of NbIRS in its VIGS lines was similar to that of NbERS and NbSRS, at least at the mRNA level (supplementary Fig. S1B). These data suggest that each organellar ARS may have a distinct threshold level required for normal organelle function.
Expression Patterns of NbERS and NbSRS-To examine tissue-specific expression patterns of NbERS and NbSRS, semiquantitative RT-PCR analyses were carried out using the NbERS-and NbSRS-specific primer sets. RT-PCR products of NbERS and NbSRS were detected in all of the tissues examined, including roots, stems, mature leaves, young leaves, flower buds, and open flowers in N. benthamiana, with somewhat higher levels of the transcripts in young tissues such as young leaves and flower buds ( Fig. 2A). The light-dependent gene expression of NbERS and NbSRS was examined by semiquantitative RT-PCR (Fig.  2B). N. benthamiana seedlings were grown for 7 days on MS media under normal light conditions (16 h light/8 h dark), under dark conditions, or under dark conditions followed by a transfer to light for 1 h. The NbERS and NbSRS mRNA levels were lower in dark-grown seed-lings than light-grown seedlings, but exposure to light for 1 h after the dark-grown period increased the NbERS and NbSRS mRNA levels to that of light-grown seedlings (Fig. 2B). As a positive control, NbGyrB expression in response to light was also monitored. NbGyrB encodes subunit B of DNA gyrase, which plays a critical role in chloroplast DNA metabolism, and is induced by light (18). These results indicate that expression of NbERS and NbSRS is stimulated by light.
Dual Targeting of AtERS and AtSRS to Chloroplasts and Mitochondria-We examined the subcellular localization of AtERS and AtSRS (Fig. 3). Both AtERS and AtSRS contain a long extension at their N terminus, which contains features of dual targeting signals to chloroplasts and mitochondria (23). We generated fusion proteins in which the N-terminal regions or the full-length proteins of AtERS and AtSRS were fused to GFP. For AtERS, the N-terminal sequence of 70 aa (Met-1 to Gly-70, M1G70) and the full-length sequence of 570 aa (Met-1 to Thr-570, M1T570) were used. For AtSRS, the N-terminal sequence of 90 aa (Met-1 to Ala-90, M1A90) and the full-length sequence of 514 aa (Met-1 to Lys-514, M1K514) were used. These DNA constructs encoding different forms of AtERS-GFP or AtSRS-GFP fusion proteins under the control of the CaMV35S promoter were introduced into protoplasts isolated from Arabidopsis seedlings. To track mitochondria, a DNA construct encoding an F 1 ATPase-␥-RFP fusion protein of the mitochondrial F 1 ATPase-␥ subunit (19), and red fluorescent protein (RFP) was cotransformed into the protoplasts. After incubation at 25°C for 24 h, expression of the introduced genes was examined by confocal laser scanning microscopy with different filters to capture images of GFP, RFP, and autofluorescence of chlorophyll.
For both AtERS(M1G70)-GFP-and AtERS(M1T570)-GFP-transformed protoplasts, the green fluorescent signal completely overlapped with both autofluorescence of chloroplasts and the red fluorescent signal of F 1 ATPase-␥-RFP, indicating that the AtERS protein was targeted to both chloroplasts and mitochondria (Fig. 3A). Similarly, both AtSRS(M1A90)-GFP-and AtSRS(M1K514)-GFP-transformed protoplasts exhibited overlapping fluorescent signals from GFP, RFP, and  chlorophyll autofluorescence, indicating dual targeting of AtSRS (Fig.  3B). These results demonstrate that the N-terminal extension of AtERS and AtSRS contains a signal for transport of the protein to both chloroplasts and mitochondria.
Numbers of Chloroplasts and Mitochondria-Protoplasts were generated from yellow sectors of the leaves from the TRV:ERS and TRV: SRS lines, and examined by confocal laser scanning microscopy (Fig.  4A). As controls, protoplasts from the TRV were also observed. In both TRV:ERS and TRV:SRS, almost all of the protoplasts, mostly originated from leaf mesophyll cells, exhibited drastically reduced numbers of chloroplasts. The average number of chloroplasts in the TRV:ERS and TRV:SRS lines was about 27 and 35% of the TRV control, respectively (Fig. 4B). Chlorophyll content of the leaves was consistently much lower in the VIGS lines than the TRV controls (Fig. 4, C and G). Furthermore, chloroplasts from the TRV:ERS and TRV:SRS lines were significantly smaller than TRV controls (Fig. 4D) and measured ϳ31 and 55% of the average control diameter, respectively. Some of the chloroplasts in the TRV:ERS and TRV:SRS lines were dumbbell-shaped, indicating disrupted chloroplast division (results not shown). The chloroplast numbers and diameter, chlorophyll autofluorescence, and total chlorophyll contents in TRV were very similar to those in wild-type (supplementary Fig. 2, A-D).
Mitochondria in leaf protoplasts isolated from the TRV control, TRV: ERS, and TRV:SRS lines were examined by TMRM and MitoTracker Green FM (MG) fluorescent probes (Fig. 4, A, E, and F). TMRM is a lipophilic cation that accumulates in mitochondria in proportion to the mitochondrial membrane potential (24), and a drop in the membrane potential leads to a decrease in fluorescence. The average TMRM fluorescence of protoplasts from TRV:ERS and TRV:SRS leaves was ϳ9 and 25% of TRV controls, respectively (Fig. 4E). This decrease in fluorescence could be due to either reduced mitochondrial numbers or altered membrane potential of mitochondria in the protoplasts. MitoTracker Green FM (MG) accumulates in mitochondria regardless of the mitochondrial membrane potential, and is widely used to determine mitochondrial mass (25). The MG fluorescence of the protoplasts from TRV: ERS and TRV:SRS leaves, ϳ8 and 11%, respectively, was significantly lower than that of the TRV control (Fig. 4, A and F). This indicates that mitochondrial numbers and/or mass are reduced in the ERS and SRS VIGS lines. The TMRM and MG fluorescence of mitochondria in TRV were very similar to those in wild-type (supplementary Fig. S2, E and F).
Ultrastructural Analysis of Chloroplasts-Transmission electron microscopy of transverse leaf sections of the TRV:ERS and TRV:SRS VIGS lines showed a large difference in the chloroplast number and morphology compared with TRV controls (Fig. 5), whereas the leaf cell structure was relatively normal (results not shown). In TRV controls, mesophyll cell chloroplasts have a well developed thylakoid membrane system and large starch granules (Fig. 5, A, D, and G). Yellow sectors of the leaves from the TRV:ERS and TRV:SRS lines exhibited severely reduced chloroplast numbers, and morphological abnormalities were apparent in the chloroplasts (Fig. 5, B, E, C, and F). The abnormal chloroplasts were small and irregular in shape, and contained few starch grains (Fig. 5, E and F). The thylakoid membranes were either poorly developed or in the process of disintegration and exhibited radical alterations in arrangement and distribution (Fig. 5, H and I). Interestingly, some of the chloroplasts in the TRV:ERS and TRV:SRS lines exhibited a dumbbell-shaped morphology with a central constriction, mimicking chloroplasts undergoing division, but with the two daughter chloroplasts differing in size (Fig. 5F). Those organelles may represent chloroplasts arrested during division due to abnormal conditions within the organelle. Consistent with the lack of starch accumulation in the abnor-mal chloroplasts (Fig. 5, B and C), starch contents in the yellowing leaves from the TRV:ERS and TRV:SRS lines were much lower than that of TRV leaves during the light period, based on iodine staining (Fig. 5J). As expected, the yellow sectors of the leaves from the TRV:ERS and TRV: SRS lines perfectly overlapped with the regions that were not stained with iodine (Fig. 5J, results not shown). In contrast to the leaf-yellowing phenotype of the organellar ARSs, VIGS of the cytosolic form of IleRS (c-IRS) resulted in arrested plant growth and abnormal leaf development (Fig. 1D). Transverse leaf sections revealed that the TRV control leaves had the typical leaf structure of dicotyledonous plants (Fig. 5K). In the TRV:c-IRS leaves, the cell numbers decreased, and the cell sizes increased in every layer, which was accompanied by an irregular cell morphology, whereas the typical dorsoventral organization of the palisade and mesophyll cells was mostly maintained (Fig. 5L). Transmission electron microscopy showed that chloroplasts of the mesophyll cells in the TRV:c-IRS lines contained large starch granules (Fig. 5M), a dramatic contrast to the lack of starch accumulation in the TRV:ERS and TRV:SRS chloroplasts (Fig. 5, B and C). Thus, depletion of the organellar or the cytosolic form of ARS resulted in completely different cellular and organellar phenotypes in plants.
Chloroplast DNA Structure-Isolated protoplasts from TRV, TRV: ERS, and TRV:SRS lines were stained with DAPI and squashed on a microslide to visualize individual chloroplasts and chloroplast nucleoids (Fig. 6A). The TRV control exhibited faint staining of chloroplast nucleoids. Chloroplast nucleoids are dispersed as small particles in the stroma mostly associated with thylakoids, and the nucleoid numbers and chloroplast DNA (cp-DNA) content vary during chloroplast development (26,27). Recent studies have revealed that the amount of cp-DNA per chloroplast decreases as the chloroplasts develop, and reaches undetectable levels in mature leaves long before the onset of senescence in Arabidopsis and maize (28,29). Compared with the TRV control, chloroplasts of TRV:ERS and TRV:SRS lines had increased DAPI fluorescence suggesting increased chloroplast DNA content, whereas the chloroplast size was much reduced.
We used PFGE to estimate the size of cp-DNA molecules in leaf chloroplasts isolated from TRV, TRV:ERS, and TRV:SRS VIGS lines (Fig. 6B). Equal numbers of leaf protoplasts prepared from each VIGS line were embedded and subjected to PFGE. Previously, Lilly et al. (21) demonstrated that both linear and circular DNA fibers with one to four copies of the chloroplast genome are present in tobacco chloroplasts, with monomers being the predominant structure. In the TRV control, hybridization of the rbcL gene probe to DNA blots from PFGE consistently revealed a band corresponding to the monomeric cp-DNA (Fig.  6B). Faint hybridization signals below the monomeric band (indicated by the red arrow) are likely to represent cp-DNA molecules of subgenomic size or sheared cp-DNA. In the TRV:ERS line, a distinct monomeric cp-DNA band was not detected and the subgenomic hybridization signal was prominent. In the TRV:SRS line, the monomeric band was visible, but the hybridization signal from the subgenomic cp-DNA was more prominent than control cp-DNA. The small cp-DNA molecules abundant in these lines could be a cleavage product of replication intermediates. In contrast, cp-DNA molecules containing multiple chloroplast genomes were not observed in these VIGS lines. Thus, compromised translation activity in the chloroplasts appeared to induce accumulation of cp-DNA representing partial genome units.
Expression of the Nuclear Genes for the Chloroplast-or Mitochondriatargeted Proteins, and the Chlorophyll Biosynthetic Enzymes-Chloroplast development involves coordinated expression of both plastidic and nuclear genes, and subsequent translocation of nuclear-encoded proteins into developing chloroplasts (30). It has been proposed that

Organellar Aminoacyl-tRNA Synthetases in Plants
signals from chloroplasts, known as plastid factors, regulate nuclear gene expression (31). The effects of depletion of the organellar ARSs on transcription of various nuclear genes were examined by semiquantitative RT-PCR using RNA from wild-type and TRV leaves, and yellow sectors of the TRV:ERS and TRV:SRS leaves (Fig. 7). The mRNA levels of nuclear genes encoding chloroplast-targeted proteins such as RbcS (Rubisco small subunit), Lhcb (chlorophyll a/b-binding protein), and FtsZ (chloroplast division protein) in the TRV:ERS and TRV:SRS lines were very similar to those in the TRV control. Furthermore, transcript levels of nuclear genes that encode enzymes in the chlorophyll biosynthesis pathway, such as ChlD, ChlH, and CHS (chalcone synthase), in the TRV:ERS and TRV:SRS lines were also similar to those of TRV (Fig.  7). As controls, expression of SGT1 and SKP1, signaling genes in plant defense (32)(33)(34), and the actin gene remained constant. Thus, many nuclear genes for chloroplast function were expressed at normal levels, regardless of the grossly abnormal status of chloroplasts in the cell, indicating disrupted communication between the chloroplasts and the nucleus. This result indicates that the chloroplast translation system may be involved in biosynthesis of the plastid factor.
We also observed that expression of nuclear genes encoding the mitochondrial pyruvate dehydrogenase E1␣ (PDHE1␣) and malate dehydrogenase (MDH) remained unaffected by impaired mitochondrial function in the leaf cells of the ERS and SRS VIGS plants (Fig. 7). The mechanism of communication between mitochondria and the nucleus has been well documented in yeast and animal systems (35,36). However, the regulation of nuclear-encoded mitochondrial proteins in plants is largely unknown. The results in this study suggest that a defec-  . Protoplasts were isolated from the fourth leaf above the infiltrated leaf in TRV, TRV:ERS, and TRV:SRS VIGS lines. After DAPI staining, the protoplasts were gently squashed under a microslide to visualize individual chloroplasts. Fluorescent signals of DAPI and chlorophyll autofluorescence were examined using a confocal laser scanning microscope. B, pulse-field gel electrophoresis of cp-DNA was carried out, as described previously (21). The DNA blots were hybridized with a radiolabeled rbcL gene probe. The black arrow indicates the band representing monomeric molecules of cp-DNA, and the red arrow indicates subgenomic sized or sheared cp-DNA. DNA size markers are indicated in kilobases. FIGURE 7. Expression of the nuclear genes coding for chloroplast-or mitochondriatargeted proteins, and chlorophyll biosynthetic enzymes. Semiquantitative RT-PCR analysis was carried out with total RNA from wild-type, TRV, and two different TRV:ERS and TRV:SRS lines. As controls, the transcript levels of SGT, SKP, and actin were examined. tive mitochondrial translation system may inhibit the signaling of mitochondrial dysfunction to the nucleus.

DISCUSSION
In this study, we show that depletion of the organellar forms of glutamyl-tRNA synthetase (NbERS) and seryl-tRNA synthetase (NbSRS) effectively eliminated both chloroplasts and mitochondria, indicating that these proteins are responsible for the GluRS and SerRS activity required for translation in chloroplasts and mitochondria. In higher plants, cytosolic tRNAs are all nucleus-encoded, and chloroplastic tRNAs are all chloroplast-encoded (37). However, the mitochondrial genome specifies fewer than the minimally required tRNAs, and as a result, one-third to one-half of the mitochondrial tRNAs are nuclearencoded and imported from the cytosol (38,39). The number of imported tRNAs and their identities vary among plant species, and in Arabidopsis, these include tRNA Ala , tRNA Thr , tRNA Val , and tRNA Gly (38,39). To complicate matters, some of the mitochondria-encoded tRNA genes, such as trnW in Arabidopsis (40), and trnP and trnR in rice (41), are either untranscribed or processed inefficiently. In Arabidopsis, the cytosolic tRNA Trp is imported to mitochondria to compensate for the lack of expression of the trnW gene (40). Correlated with import of the cytosolic tRNAs into mitochondria, the cytosolic forms of some ARSs, such as AlaRS, ThrRS, ValRS, and GlyRS in Arabidopsis, are imported into the mitochondria to catalyze aminoacylation of their cognate nucleus-encoded tRNAs (38). For Arabidopsis AlaRS, ThrRS, ValRS, and GlyRS, a single gene encodes both the cytosolic and mitochondrial enzymes by using two in-frame translation initiation codons (7,8). However, further studies revealed that the mitochondria-targeted cytosolic GlyRS is inactive in the mitochondrial fraction, and the organellar GlyRS is likely to glycylate both the organelle-encoded tRNA Gly and the imported tRNA Gly (12). Thus, to understand the complete picture of mitochondrial import of individual tRNAs and ARS enzymes in plants, further investigation is required.
The mitochondrial genome of Nicotiana tabacum, a close relative of N. benthamiana, was recently sequenced (42) and found to contain a single chloroplast-derived gene for tRNA Glu (trnE-UUC) and three native genes for tRNA Ser (trnS-GCU, trnS-UGA, and trnS-GGA). These genes have also been found in the mitochondrial genome of Arabidopsis, rapeseed, wheat, rice, and sugar beet (39,(43)(44)(45)(46). In these plants, and most likely in N. benthamiana as well, the mitochondria-encoded tRNAs are sufficient to read all the codons of glutamic acid and serine by wobble base pairing for protein synthesis in mitochondria. Consistently, the cytosolic forms of Arabidopsis GluRS and SerRS do not have the conserved mitochondria-targeting signal or the two different translation initiation codons at the N terminus (data not shown). Furthermore, expression of the maize organellar SerRS complemented the phenotype of a yeast mutant with a defective mitochondrial SerRS, and the purified organellar SerRS protein aminoacylated the maize chloroplast and mitochondrial tRNA Ser with equal efficiency (47). These results, together with our current results, suggest that the organellar forms of GluRS and SerRS are responsible for aminoacylation of tRNAs Glu and tRNAs Ser in the mitochondria of higher plants.
The numbers of chloroplasts are much reduced in affected cells of the ERS and SRS VIGS lines. Furthermore, the chloroplasts were significantly smaller than those of the TRV control (approximately one-thirtieth and one-sixth the control volume, for ERS and SRS, respectively), but had a higher cp-DNA content. In higher plants, the population size of mature chloroplasts and photosynthetic competence are achieved by the post-mitotic division of young chloroplasts in the expanding mesophyll cells (48,49). At early stages, the proplastids in mitotic cells undergo a series of divisions that are correlated with cell division, ensuring that the young post-mitotic cells contain at least 15 proplastids (48,49). During cell expansion, the young chloroplasts expand and begin to divide when they have grown to ϳ50% of their final volume. The reduced sizes and numbers of chloroplasts in the ERS and SRS VIGS lines indicate defective proplastid/chloroplast expansion and division. The chloroplasts in the ERS and SRS VIGS lines also appeared to accumulate a large number of small-sized cp-DNA molecules representing partial genome units (Fig. 6B). The abnormal cp-DNA structure as well as other defects in biogenesis due to suppression of plastid translation likely inhibited further growth and differentiation of chloroplasts and suppressed normal division, which eventually led to degeneration of the organelle.
The ERS, SRS, and IRS VIGS lines each showed a different degree of phenotypic severity. Suppression of NbERS resulted in the most severe abnormality among the three genes, as shown by chloroplast numbers and leaf chlorophyll content, whereas gene silencing of NbIRS caused only mild leaf yellowing. In the NbERS, NbSRS, and NbIRS VIGS lines, the degree of silencing of each individual mRNA did not seem to differ significantly ( Fig. 1 and Supplementary Fig. S1). A possible reason for the difference in phenotypes may be differences in the threshold level of each ARS required for minimal translation activity to maintain normal chloroplast function. Threshold differences may reflect differences in biochemical characteristics of individual ARS proteins or substrate availability for each ARS within the chloroplasts. Recently, in addition to the role in the translation process, ARSs have been shown to have other cellular functions, including RNA processing and trafficking, apoptosis, rRNA synthesis, angiogenesis, and inflammation (50,51). In plants, it has been reported that the plastid-encoded tRNA Glu mediates the switch in RNA polymerase usage from nuclear-encoded plastid RNA polymerase to plastid-encoded plastid RNA polymerase during chloroplast biogenesis, by directly binding to nuclear-encoded plastid RNA polymerase and inhibiting the transcriptional activity of nuclear-encoded plastid RNA polymerase (52). Thus, there is a possibility that the phenotypic differences we observed in this study are caused by other functions of the organellar ARSs, yet unidentified.
Biogenesis of plastids is tightly coupled to temporal and spatial stages of plant development and involves coordinated expression of plastidic genes and nuclear genes encoding chloroplast proteins, which requires precise communication between the nucleus and plastids. Signals from the plastid, termed "plastid factors," have been postulated to provide the nuclear genome with information about the developmental status of the plastid (31). These signals were initially identified by their capacity to affect transcription of subsets of nuclear genes. For example, impaired chloroplast function or development by herbicides or mutations resulted in reduced expression of the nuclear genes Lhcb (chlorophyll a/b-binding protein) and RbcS (Rubisco small subunit) (53)(54)(55). Recent evidence suggests that at least two different plastid factors operate in plants: redox signals (56) and intermediates in chlorophyll biosynthesis (57,58). Recently, Ahlert et al. (59) used cell-line specific recombination-based knockouts of chloroplast translation to demonstrate that plastid protein synthesis is essential for plant development and cell division. Furthermore, the nuclear-encoded plastid ribosomal proteins L35 and S9 were shown to be critical for the very early stages of embryo development in maize (60,61).
These results indicate that the chloroplast gene(s) and gene product(s) critically influence processes outside the chloroplast, in addition to the well established metabolic signals between the chloroplast and nucleus. In this study, we observed that the defective chloroplasts that result from depletion of GluRS and SerRS did not influence the expres-sion of diverse nuclear genes encoding chloroplast-targeted photosynthetic proteins and chlorophyll biosynthesis enzymes. Thus, in the affected cells, the nucleus did not seem to sense the functional status of the chloroplasts. We speculate that a block of protein synthesis within the plastids resulting from a lack of GluRS and SerRS activity may interfere with the biosynthetic pathway of the plastid factors. Genetic analyses of plastid-encoded genes may identify the pathway and the signaling mechanism that links the plastids to the nucleus.