Single nucleotides in higher plant organellar mRNAs are subject to post‐transcriptional alterations by RNA editing, typically resulting in changes of the encoded protein sequence. Although some information has been acquired on the general features of the editing processes in both plastids and plant mitochondria, the mechanisms and factors involved in the selective recognition of the nucleotide to be edited are still unknown. To gain a better understanding of how an editing site is specifically selected by the organellar RNA editing machinery, we have attempted to rescue a previously generated tobacco plastid editing mutant. Using an interspecific protoplast fusion approach, we were able to restore RNA editing activity for a specific site in the psbF transcript that otherwise remained unedited. Our results suggest (i) that site‐specific trans‐acting factors mediate chloroplast editing site recognition and (ii) that these factors are of extraplastidic origin.
Gene expression in a variety of genetic systems is affected by a post‐transcriptional nucleotide modification mechanism termed RNA editing (Benne, 1996). In plants, RNA editing occurs in two cellular compartments, mitochondria and plastids (Covello and Gray, 1989; Gualberto et al., 1989; Hiesel et al., 1989; Hoch et al., 1991; Kudla et al., 1992). In both organelles, editing involves mainly cytidine→uridine conversions (Bonnard et al., 1992; Kössel et al., 1993; Schuster and Brennicke, 1994; Maier et al., 1996), with few examples of reverse events (Gualberto et al., 1990; Schuster et al., 1990; Yoshinaga et al., 1996). The selective modification of only very few nucleotides within an mRNA poses the intriguing question of how these residues are recognized specifically by the organellar editing machinery. In the absence of efficient in vitro editing systems (Araya et al., 1991; Yu and Schuster, 1995), evidence for the site specificity of the editing reaction has come from two sets of plastid transformation experiments. Firstly, a spinach editing site transferred into the tobacco plastid genome remained unmodified in the transgenic tobacco organelles (Bock et al., 1994), suggesting that an essential site‐specific (cis‐ or trans‐acting) component was missing. Secondly, introduction of additional copies of the tobacco psbL editing site revealed that the editing activity for this site is depletable (Chaudhuri et al., 1995; Chaudhuri and Maliga, 1996).
We have used the above‐mentioned plastid RNA editing mutant resulting from the interspecific editing site transfer (Bock et al., 1994) to investigate the nature and cellular origin of specificity factors for plastid editing site selection. This mutant carries a modified version of the psbF gene containing the spinach psbF editing site in an internal segment of 35 bp complete sequence identity (Figure 1). Editing of this site changes a genomically encoded serine codon into a phenylalanine codon (Bock et al., 1993) which is evolutionarily conserved in the psbF‐encoded proteins (β‐subunit of cytochrome b559) of all photosynthetically active organisms. The psbF editing site is not present in tobacco wild‐type plants since here the correct phenylalanine codon is already specified at the DNA level (Figure 1). We have shown previously that replacement of part of the tobacco psbF gene with the homologous region from spinach results in a mutant phenotype owing to the absence of any detectable psbF editing activity in the transgenic plastids (Bock et al., 1994).
Three possible reasons for the lack of editing of the spinach site in tobacco can be envisaged: (i) the 35 bp region of the spinach psbF gene used to replace the homologous tobacco sequence is not sufficient to direct editing, i.e. the missing factor is a cis‐acting sequence element outside this 35 bp segment; (ii) a plastid‐encoded trans‐acting specificity factor for psbF editing is missing in tobacco chloroplasts; and (iii) a nuclear‐encoded specificity factor for psbF editing is present and imported into the chloroplast compartment in spinach but not in tobacco. To distinguish between these three possibilities, we set out to test whether combination of the mutant chloroplasts with a spinach nucleus would restore psbF editing activity.
Generation of interspecific hybrid cells by fusion of spinach and mutant tobacco protoplasts
The transplastomic tobacco mutant Nt‐pRB8‐S6 carries the spinach psbF editing site and the surrounding sequence motif from −16 to +18 (Bock et al., 1994). This segment replaces the homologous tobacco psbF sequences which naturally do not contain the RNA editing site (Figure 1). We have shown recently that the tobacco plastid RNA editing machinery is incapable of editing the heterologous spinach site, resulting in a single amino acid substitution within the psbF gene product (β‐subunit of cytochrome b559). In homoplasmic plants, this lack of editing is associated with a mutant (i.e. photosynthetically deficient) phenotype (Bock et al., 1994). The phenotype is stable under all environmental conditions tested, indicating that neither spontaneous restoration of psbF editing nor elimination of the editing site by a genomic C→T mutation occurs (R.Bock, unpublished results).
The plastid psbF gene is part of the plastid psbE operon containing the psbE, psbF, psbL and psbJ reading frames. Theoretically, sequences outside the introduced −16 to +18 fragment could be required for editing of the psbF site and thus could be the cause of the lack of editing in the transplastomic tobacco line. If this were the case, the missing editing factor in tobacco plastids would be a distant cis‐acting element. Alternatively, a trans‐acting specificity factor could participate in the psbF editing site recognition. Evolutionary loss of the psbF editing site in tobacco would most likely be accompanied by the loss of the site‐specific trans‐factor in the absence of selective pressure. In view of the extremely high degree of sequence similarity of the spinach and tobacco psbE operons (Carrillo et al., 1986; Bock et al., 1993), we considered a missing distant cis‐acting factor a less likely explanation for the lack of psbF editing in the Nt‐pRB8‐S6 tobacco line. Moreover, recent data indicate that in most cases the essential cis‐elements for plastid RNA editing site recognition reside in close proximity to the editing site (Bock et al., 1996; Chaudhuri and Maliga, 1996) and thus are likely to be contained in the 35 bp segment of complete sequence identity. We speculated, therefore, that a trans‐acting specificity factor for psbF editing is present in spinach but absent from tobacco plastids. We favored the idea that this factor is provided by the nucleus rather than encoded by the plastid genome for two reasons: (i) plastid RNA editing was shown to be entirely independent of chloroplast translation (and thus of plastid‐encoded proteins; Zeltz et al., 1993) and (ii) we have been unable to detect guide RNA‐like molecules in plastids (Bock and Maliga, 1995; our unpublished results from low stringency hybridization analyses) which, by analogy to the kinetoplast editing system (Simpson et al., 1993; Benne, 1996; Kable et al., 1996), could serve as trans‐acting specificity factors also for chloroplast editing. If our nuclear‐encoded specificity factor hypothesis were correct, combination of the mutant tobacco chloroplasts with a spinach nucleus should restore psbF editing activity.
Interspecific combinations of a given cytoplasm with a different nuclear background can be achieved by using cell fusion technologies. We attempted, therefore, to generate hybrid cells resulting from spinach and Nt‐pRB8‐S6 tobacco protoplasts. To facilitate microscopic selection of heterologous fusion products and to exclude unfused protoplasts as well as homologous tobacco–tobacco and spinach–spinach fusion products, we used sterile etiolated spinach seedlings and prepared protoplasts from their bright yellow cotyledons (Figure 2A and C). The presence of undifferentiated plastids (etioplasts) in the spinach protoplasts is unlikely to interfere with editing of the psbF site. We have shown previously that psbF editing is equally efficient in spinach chloroplasts and etioplasts (Bock et al., 1993).
Protoplast fusion was induced by DC pulses following dielectrophoretic protoplast alignment in an AC field. Since the fusion does not lead to immediate mixing of the cytoplasms of the fusion partners, tobacco–spinach hybrid protoplasts can easily be identified by their partially green and partially yellow cytoplasms as well as by their chimeric internal structure (Figure 2E and F). Microscopic selection was carried out using a modified computer‐based, stepmotor‐driven microdevice (Koop and Schweiger, 1985; Schweiger et al., 1987; Koop and Spangenberg, 1989; Spangenberg and Koop, 1992). Individual heterologous fusion products were identified and taken up by a microcapillary (connected to a nanoliter pump) in a volume of 10 nl. Computer‐directed micromotors then transferred the capillary to a microdroplet in a collection dish (2 μl culture medium overlaid with mineral oil) where the hybrid cell was released. This way, 15–20 fusion products per experiment were selected and subsequently subjected to microdroplet culture in the dark. The hybrid cells did not undergo cell divisions, which is most likely due to incompatibility of the nuclei and/or cytoplasms of the only distantly related species tobacco and spinach. However, when kept in suitable cultivation media, the fusion products maintained physiological activity for a few days.
Analysis of psbF editing in spinach–tobacco hybrid cells
To facilitate the analysis of psbF editing in as few as 15 fusion products, we have optimized the RNA isolation, cDNA synthesis and PCR procedures resulting in reproducible amplification of psbF cDNA from five tobacco protoplasts. Since the hybrid cells contain both spinach and mutant tobacco plastids, the spinach chloroplast psbF cDNA molecules needed to be selectively excluded from amplification. This was made feasible by choosing a PCR primer derived from a variable region in the downstream psbL/J intergenic spacer (Shinozaki et al., 1986; Bock et al., 1993). psbF cDNA can be amplified with primers located outside the coding region since the plastid psbE, psbF, psbL and psbJ genes are part of a tetracistronic transcription unit which does not undergo post‐transcriptional processing into monocistronic mRNAs (Haley and Bogorad, 1990; Willey and Gray, 1990).
As expected, PCR with the tobacco‐specific primer resulted in cDNA amplification products for tobacco and for the hybrid cells, but not for spinach (Figure 3). In order to verify that indeed cDNA and no contaminating DNA was amplified, we first sequenced the downstream psbL portion of the PCR products. The psbL initiation codon has been shown to be created by RNA editing in both tobacco and spinach (Kudla et al., 1991; Bock et al., 1993). The presence of a virtually completely edited AUG initiation codon indicated that exclusively cDNA was amplified (data not shown).
We then tested the editing status of the psbF site in fusion products cultured for 2 days. Primer extension analysis revealed, in addition to a signal for the unedited nucleotide, also a significant peak for the edited nucleotide (Figure 4A), indicating that combination of the mutant tobacco plastids with a spinach nucleus may have indeed resulted in restoration of editing activity. Direct sequencing of the amplified cDNAs confirmed the presence of a mixed cDNA population in the cultured hybrid cells: besides molecules with a C, psbF cDNAs with a T at the editing position are also detected (Figure 4B). Quantitation of editing efficiency (Bock et al., 1996) using a PhosphorImager yielded an approximate proportion of 25% edited mRNAs.
To provide direct proof for the existence of edited transcript molecules in the hybrid cell plastids, we have analyzed individual cDNA clones. Sequencing of 20 clones resulted in the detection of four clones reflecting edited psbF mRNA sequences (Figure 4C). The presence of all diagnostic sequence deviations both upstream and downstream of the editing site (Figures 1 and 4C) confirmed that the edited mRNA molecules originate from the chimeric psbF gene of the mutant tobacco plastids and not from co‐amplified spinach psbF cDNA molecules. This as well as the presence of completely unedited tobacco and fully edited spinach psbF cDNA populations in an unfused control sample also excludes template switching by the Taq polymerase as a possible cause of the amplification of edited cDNA molecules.
No significant restoration of editing could be observed in the fusion products immediately after fusion, strongly suggesting that a steady flow of a trans‐acting specificity factor into the mutant plastids is the cause of the observed psbF editing in the hybrid cells. To examine the time course of the restoration more precisely, we analyzed fusion products after 0, 1, 2 and 5 days incubation and measured the psbF editing efficiencies. The detected amount of edited mRNA was ∼10% after 1 day of microdroplet culture, ∼25% after 2 days and ∼15% after 5 days [values determined as described previously (Bock et al., 1996) and rounded to full 5 or 10%]. The observed decrease from day 2 to day 5 is likely to reflect the loss of viability of the hybrid cells as caused by the nucleo‐cytoplasmic incompatibility.
The lack of psbF editing in a transplastomic tobacco mutant resulting from the interspecific transfer of the spinach psbF editing site into the tobacco plastid genome (Bock et al., 1994) can be explained by three hypotheses: (i) lack of an essential distant cis‐acting sequence element located outside the transferred psbF segment; (ii) absence of a plastid‐encoded trans‐acting specificity factor for psbF editing; or (iii) lack of a site‐specific trans‐factor provided by the spinach nucleus. In the course of this work, we have experimentally tested the hypothesis that a nuclear‐encoded trans‐acting factor confers the site specificity of the psbF editing reaction. We have shown that psbF editing can be partially restored by combining the mutant Nt‐pRB8‐S6 chloroplasts with a spinach nucleo‐cytoplasmic background. This finding provides direct evidence for the involvement of site‐specific trans‐factors in plastid RNA editing site recognition.
We have considered the possibility that this trans‐factor is synthesized in the spinach plastids and introduced into Nt‐pRB8‐S6 plastids by recombination between spinach and tobacco plastids. However, recombination between chloroplasts is known to be an extremely rare event that may occur only accidentally (Medgyesy et al., 1985) and is even undetectable under non‐selective conditions (Kutzelnigg and Stubbe, 1974; Chiu and Sears, 1985). This is most probably due to a very low frequency of chloroplast fusion (Birky, 1995) and contrasts with the situation in plant mitochondria where organelle fusion and genome recombination appear to be regular events accompanying cell fusions (Akagi et al., 1995). The possibility of recombination between chloroplasts after cell fusion was studied extensively, and attempts to identify chloroplast recombinants have mostly failed (e.g. Fluhr et al., 1984). Applying strong selection for plastid recombination, Maliga and co‐workers were able to isolate a single chloroplast recombinant from a large number of heterologous tobacco protoplast fusions (Medgyesy et al., 1985). This recombinant line was analyzed thoroughly and found to contain multiple sites of recombination between the two parental genomes (Maliga and Fejes, 1985), suggesting that the rate‐limiting step of chloroplast recombination is the extremely low probability of chloroplast fusion events (Maliga et al., 1987). The chloroplast recombination frequency was at best 10−6 [calculated according to Maliga et al. (1987) for the average number of 100 chloroplasts per tobacco cell (Chen et al., 1977)], which by no means can account for any significant restoration of psbF editing in our protoplast fusion experiments.
We conclude, therefore, that the spinach nucleus is the source of the trans‐acting specificity factor that restored psbF editing in the Nt‐pRB8‐S6 plastids. This markedly contrasts the editing site selection mechanism in a different, evolutionarily unrelated editing system, the kinetoplasts of trypanosomes (Benne et al., 1986). In this system, the specificity factors are not imported into the organelle. Instead, they are encoded as small RNA molecules (termed guide RNAs) by the kinetoplast genome itself (Simpson et al., 1993; Benne, 1996).
It was shown recently that the sequences immediately surrounding the nucleotide to be edited act as important cis‐guiding determinants in editing site selection (Bock et al., 1996; Chaudhuri and Maliga, 1996). Thus it seems reasonable to assume that the trans‐acting site‐specific factor shown to be involved in psbF editing is the molecule specifically binding to these cis‐acting sequence elements. The molecular interaction of RNA substrate and specificity factor may then facilitate binding of the (hypothetical) general component(s) of the editing machinery which may carry the catalytic cytidine‐modifying activity (e.g. a nuclear‐encoded cytidine deaminase).
In spite of incubation over several days, only a partially edited psbF mRNA pool was detected in the generated spinach–tobacco hybrid cells. The incomplete restoration of psbF editing may be due to limiting amounts of the trans‐factor (i.e. competition between the spinach and tobacco chloroplasts for the factor provided by the spinach nucleus) or, alternatively, could be caused by an inefficient uptake of the spinach nuclear factor by the tobacco chloroplast import machinery. Interestingly, recent studies have suggested that the introduction of supernumerary editing sites (and the concomitant increase in the number of substrate molecules) can exhaust the capacity of the plastid editing apparatus (Chaudhuri et al., 1995; Chaudhuri and Maliga, 1996). Similarly, usage in both spinach and tobacco plastids may lead to a depletion of the psbF‐specific trans‐factor synthesized by the spinach nucleus only.
How specific are such trans‐factors for plastid RNA editing? Besides a strict site specificity, it also seems possible that one trans‐factor mediates the editing of more than one site. Certain plastid editing sites were found to share common sequence elements (Maier et al., 1992; Freyer et al., 1995), raising the possibility that they are recognized by one and the same trans‐acting factor(s). Our data suggest that the evolutionary elimination of the psbF editing site in tobacco (by a genomic C→T mutation) led to a subsequent loss of the corresponding trans‐acting specificity factor owing to the lack of selective pressure. Thus it appears unlikely that this factor was required for the editing of other substrates in tobacco plastids. We propose, therefore, that at least this trans‐factor is strictly specific for the psbF site. Given the high number of plastid RNA editing sites (Maier et al., 1995) and the at least two orders of magnitude higher number of plant mitochondrial editing sites, we consider it unlikely that these site‐specific trans‐acting factors have the property of an editing enzyme. We rather view them as small molecules whose interaction with the editing site allows for binding of the catalytic component of the editing machinery. Future investigations will be aimed at testing this model and determining the molecular identity of the factors involved. The knowledge that site‐specific trans‐acting editing factors exist and are encoded by nuclear genes, as well as the availability of mutant phenotypes linked to plastid RNA editing, will now allow us to devise suitable genetic screens and to employ the powerful tools of plant nuclear genetics in order to identify these factors.
Materials and methods
Tobacco mutant and wild‐type plants (Nicotiana tabacum cv Petit Havana) were grown under sterile conditions on agar‐solidified MS medium (Murashige and Skoog, 1962) containing 30 g/l sucrose. The tobacco editing mutant was generated by chloroplast transformation (Svab et al., 1990; Svab and Maliga, 1993; Kanevski and Maliga, 1994) as described previously (Bock et al., 1994). Etiolated spinach seedlings were obtained by germinating surface‐sterilized seeds in the dark on sucrose‐containing MS medium.
Protoplast isolation procedures
Protoplasts were prepared from cotyledons of sterile etiolated spinach plants (20 pairs) and from a single leaf of the homoplasmic transformed tobacco plastid RNA editing mutant (Bock et al., 1994). Protoplast isolation was carried out by incubation of tissue samples overnight in the dark at 25°C in PIN medium (Koop et al., 1996) with 0.5% cellulase RS, 0.5% pektolyase Y23, 0.5% macerozyme R‐10, 500 μg/ml spectinomycin and 100 μg/ml carbenicillin. Purified protoplasts were kept in 0.5 M mannitol/1 mM CaCl2 and transferred to 0.45 M mannitol/1 mM CaCl2 immediately before fusion (Eigel et al., 1991).
Protoplast fusion and sorting of hybrid cells
For protoplast fusion, a pair of stainless steel electrodes was inserted into the well of an agarose layer (1% agarose, 0.45 M mannitol, 1 mM CaCl2). The well was produced by insertion of an appropriate mold into a melted agarose layer and was suitable to take up the electrodes (electrode gap 2.5×0.3×0.6 cm) after solidification of the agarose. A 0.4 ml mixture of spinach and tobacco protoplasts (∼104 protoplasts each) were aligned dielectrophoretically in an AC field (60 V/cm, 1 MHz) and subsequently fused by a few DC pulses of 1 kV/cm (50 μs pulse duration). A modified computer‐based, stepmotor‐driven microdevice (Koop and Schweiger, 1985; Schweiger et al., 1987; Koop and Spangenberg, 1989; Spangenberg and Koop, 1992) was used for microscopic selection of hybrid cells. Individual heterologous fusion products were taken up in a volume of 10 nl by a glass microcapillary and subjected to microdroplet culture in the dark [2 μl PCN culture medium (Koop et al., 1996) with 500 μg/ml spectinomycin overlaid with mineral oil].
Isolation of nucleic acids
Fifteen hybrid cells were cultured in a microdroplet in the dark at 25°C for 1–5 days. RNA was isolated using 100 μl of TRIZOL reagent (Gibco BRL) and adding glycogen (Boehringer Mannheim) as carrier prior to precipitation. The following controls were included: (i) 15 tobacco protoplasts, (ii) 15 spinach protoplasts, (iii) 15 tobacco + 15 spinach protoplasts, unfused and (iv) 15 hybrid cells not subjected to microdroplet culture. Contaminating DNA was removed by digestion with RNase‐free DNase I (Boehringer Mannheim), followed by phenol/chloroform treatment and ethanol precipitation after glycogen addition.
DNA was isolated from protoplast samples according to a rapid miniprep procedure described by Doyle and Doyle (1990).
cDNA synthesis and PCR
Reverse transcription was primed with random hexanucleotide primers (1 ng per reaction) and elongated with Moloney murine leukemia virus RNase H‐free reverse transcriptase (Gibco/BRL) in a total volume of 10 μl for 45 min. Five μl of the cDNA reaction were amplified in a hot start PCR with the primer pair P7355 (5′ GACTATAGATCGAACCTATCC 3′) / P7652 (5′ CCGAATGAGCTAAGAGAATCTT 3′) according to standard protocols (45 s at 94°C, 1.5 min at 58 °C, 1 min at 72°C; 30 or 40 cycles). This primer pair is specific for tobacco psbE operon‐derived cDNA and excludes spinach cDNA from amplification (Figure 3).
Primer extension analysis
Primer extension analyses were carried out using 50 ng of gel‐purified PCR product in a 10 μl reaction (30 s at 94°C, 30 s at 58°C, 10 s at 72°C; six cycles) with 2 ng of 5′ fluorescence‐labeled primer P529 (5′ GAACTGCATTGCTGATATTG 3′), 10 M dATP and 10 M dCTP. Extension products were separated in denaturing polyacrylamide gels and detected by the ALF system (Pharmacia).
Cloning and sequencing techniques
DNA and cDNA amplification products were purified by electrophoresis on 2% agarose gels and subsequent extraction from gel slices using the QIAEX II kit (QIAGEN). For analysis of individual cDNA clones, amplified cDNA was digested with Sau3AI and cloned into a Bluescript vector (Stratagene) cut with EcoRV and BamHI.
Plasmid DNA was sequenced by cycle sequencing reactions using the Sequitherm kit (Biozyme). Direct sequencing of purified PCR products was performed by a modified chain termination method described by Bachmann et al. (1990). Oligonucleotide P177 (5′ CCCCAGTAGAGACTGGTACG 3′) served as sequencing primer. Quantitation of RNA editing efficiencies was carried out as described earlier (Bock et al., 1996).
The authors are indebted to Drs Pal Maliga and Zora Svab for making available ptDNA clones and plastid transformation vectors. Excellent technical assistance by Mrs Marita Hermann is gratefully acknowledged. This research was supported by the Sonderforschungsbereich 184 (H.‐U.K.), the Human Frontiers Science Program Organization (RG‐437/94M), the Wissenschaftliche Gesellschaft, Freiburg and by grants from the Deutsche Forschungsgemeinschaft (R.B.).
- Copyright © 1997 European Molecular Biology Organization