RNA editing in higher plant plastids changes single cytidine residues to uridine through an unknown mechanism. In order to investigate the relation of editing to physiological processes and to other steps in plastid gene expression, we have tested the sensitivity of chloroplast RNA editing to heat shock and antibiotics. We show that heat shock conditions as well as treatment of plants with prokaryotic translational inhibitors can inhibit plastid RNA editing. Surprisingly, this inhibitory effect is confined to a limited number of plastid editing sites suggesting that some site-specific factor(s) but none of the general components of the plastid RNA editing machinery are compromised. Contrary to previous expectations, our results provide evidence for a role of plastid translation in RNA editing.
Plant organellar transcripts can be subject to post-transcriptional pyrimidine-to-pyrimidine conversions referred to as RNA editing. In higher plant plastids, editing involves exclusively cytidine-touridine transitions at highly specific sites (1–4). The number of editing sites in the plastid compartment is rather low and was estimated to amount to ∼25 sites in the entire 140 kb maize chloroplast genome (5). This low editing frequency markedly contrasts the >1000 editing sites estimated to be encoded in higher plant mitochondrial genomes (6). Otherwise, the two plant organellar RNA editing systems show many similar features and may have originated from common evolutionary roots (3).
Editing in chloroplast transcripts is highly specific and usually very efficient. At most sites, the mRNA population shows a virtually complete transition from C to U. A major question has been how to explain the extraordinarily high specificity of the plastid RNA editing machinery in selecting individual cytosine residues for modification. A number of in vivo studies employing chloroplast transformation have demonstrated that, at the mRNA level, sequences immediately flanking the editing site serve as important determinants for the molecular recognition of plastid editing sites (7–10). However, these sequence requirements in cis are unlikely to be uniform with respect to their size and their distance from the editing site (8). In addition to the involvement of cis-acting elements, there is compelling evidence that site-specific trans-acting factors also participate in plastid RNA editing site recognition. Introduction of supernumerary copies of the tobacco psbL editing site into the plastid genome by chloroplast transformation has revealed that the editing activity for this particular site is depletable while editing at other sites remained unaffected (7). The participation of site-specific trans-acting factors in the editing reactions in chloroplasts was also demonstrated by the rescue of a tobacco plastid RNA editing mutant (created by introducing a heterologous editing site taken from the spinach chloroplast genome which does not undergo editing in tobacco chloroplasts). Editing at the heterologous site could be restored by fusing mutant cells with spinach protoplasts suggesting that a site-specific trans-acting factor of extraplastidic origin is required for the editing of this site (11,12). The molecular identity of such trans-acting specificity factors for plastid RNA editing has not yet been determined. Also, none of the general (enzymatic?) components of the chloroplast RNA editing machinery have been identified to date.
To gain some information about the chemical nature of the molecules involved in plastid RNA editing (RNA or protein factors?) and their subcellular origin, we have investigated the temperature sensitivity of the editing reaction as well as its sensitivity to prokaryotic translational inhibitors. We show here that elevated temperature and inhibition of chloroplast translation by antibiotics selectively block RNA editing at a small number of sites thus providing a first evidence for a role of the plastid translational apparatus in RNA editing.
Materials and Methods
Surface-sterilized tobacco (Nicotiana tabacum cv. Petit Havana) or barley seeds were germinated on agar-solidified MS medium (13) containing 30 g/l sucrose. Root material was harvested from plants grown in a sterile environment on the same medium. For analysis of heat shock-treated plants, germinating seeds were transferred from 25 to 37°C and kept at this temperature for ∼1 week. For experiments with translational inhibitors, surface-sterilized seeds were germinated on spectinomycin- or streptomycin-containing medium (500 mg antibiotic/l).
List of oligonucleotides
The following synthetic oligonucleotides were used in this study (I = inosine):
apA1: 5′ CTCCTAGATAAGCTTCACG 3′;
apA2: 5′ GAGCCTCTTCAACAAGGGC 3′;
apF1: 5′ TTTCTTTGGGCCACTGGCC 3′;
apF2: 5′ CCCCTCCACGCAGTTCTTC 3′;
ndi: 5′ CTTGCGGCGCGTATTGTTTGTTGA 3′;
na1: 5′ ACTATTAAATCTTGGGTGG 3′;
ndt1: 5′ GTAAGAGATGTGAATCCGCCTGT 3′;
ndt2: 5′ TGAGTACAGACGTTTCTTTCC 3′;
am3: 5′ GAAGAAAGAAAATTAAGGAACC 3′;
pet1: 5′ TGTATTTCCGGAGTATGAG 3′;
pet2: 5′ CTCTATAAAGGCCCAGAAA 3′;
pet3: 5′ TGTGGGACAATCTACTTTG 3′;
Y: 5′ CCTTCCCTATTCATTGCGGGTTGG 3′;
7652: 5′ CCGAATGAGCTAAGAGAATCTT 3′;
7355: 5′ GACTATAGATCGAACCTATCC 3′;
B5: 5′ ATAATATCAGATTGGGGAGG 3′;
B15: 5′ TGGAATATACCGAATTGTGATC 3′;
B23: 5′ TTGGTIAGATACAATTTGAAGGA 3′;
B24: 5′ TTTIIGTCAATGGATTIGTTTGATC 3′;
nb0: 5′ TGCTTCTCTTCGATGGAAG 3′;
nb2: 5′ GGCGGAACAGATCTACTAATTC 3′;
nb4: 5′ TAATCCTGCATAATCTCGAATG 3′;
nb10: 5′ CAGCTACTCTAGGGGGAATG 3′;
nb11: 5′ TTCATGCTTGTTTGAGTAATAGC 3′;
P12: 5′ TTTTTCTAGAGGTCTAATGAGGCTACTAGG 3′;
P13: 5′ TTTTGGATCCTATACATTTGTGTATTGATA 3′;
P14: 5′ TTTTTCTAGAATATACCAAGAAAGATGTAC 3′;
P16: 5′ TTTTTCTAGACGCTCATATTCATTACCGTA 3′;
P19: 5′ TTTTGGATCCTTTGATGAGAAATAAAAAGG 3′.
Isolation of nucleic acids
Total plant nucleic acids were extracted according to a rapid miniprep procedure described by Doyle and Doyle (14). Total cellular RNA was isolated using the TRIzol reagent (Gibco/BRL). For cDNA synthesis, an aliquot of the RNA preparation was treated with RNase-free DNase I (Boehringer Mannheim).
For cDNA synthesis, total cellular RNA samples were treated with DNase I (Boehringer Mannheim) to remove contaminating DNA molecules. Reverse transcription reactions were primed with random hexanucleotide primers for 10 min at room temperature. The elongation reaction was performed with Moloney murine leukemia virus RNase H-free reverse transcriptase (Gibco/BRL) at 42°C following the manufacturer's instructions.
Polymerase chain reaction (PCR)
DNA and cDNA templates were amplified according to standard protocols (45 s at 94°C, 1.5 min at 55–58°C, 1.5 min at 72°C; 30 or 40 cycles) using a Perkin Elmer GeneAmp 9600 thermocycler. The following primer combinations were employed for amplification of chloroplast DNA and cDNA sequences (see List of oligonucleotides):
E 0, E I: P19/nb0;
E II-E VII: P14/nb4;
E VIII: P16/nb2;
Direct sequence determination of the PCR-amplified cDNA population or amplified DNA samples was performed by a modified chain termination method as described by Bachmann et al. (15). The following synthetic oligonucleotides were used as sequencing primers (see List of oligonucleotides):
E 0: nb0;
E I: nb10;
E II, E III: P12, P13;
E IV, E V: P16;
E VI, EVII: nb11;
E VIII: nb2;
E I: B15;
E II: B5.
Total cellular RNA was electrophoresed on formaldehydecontaining 1% agarose gels and transferred onto Hybond N+ membranes (Amersham) using standard protocols (16). A [α-32P]dATP-labelled ndhB-specific probe was generated by random priming (Boehringer Mannheim) following the instructions of the manufacturer. Hybridizations were carried out at 65°C in Rapid Hybridization Buffer (Amersham).
Inhibition of plastid RNA editing at two tobacco ndhB sites under heat shock conditions
Site-specific trans-acting factors have been implicated in the RNA editing reaction in higher plant plastids (7,12). However, the chemical nature of these specificity factors is currently unknown. By analogy to the RNA editing system in trypanosomatid mitochondria (17), small RNA molecules termed guide RNAs could serve as trans-acting factors imposing site specificity onto the plastid RNA editing machinery by hybridizing to the mRNA and thereby mediating editing site recognition. If annealing of a small RNA molecule to the mRNA were necessary to allow for editing site recognition in plastids, this first step in the editing reaction should be temperature sensitive and become less efficient at temperatures exceeding the hybridization temperature of the guide RNA and the mRNA template. We therefore set out to test whether elevated temperatures would have an effect on the editing efficiency at chloroplast RNA editing sites.
We transferred tobacco seeds upon germination from standard growth conditions (24–26°C) to 37°C. The seedlings continued to grow at 37°C; the hypocotyl elongated and the apical cotyledon hook opened. However, the plants displayed a pale green phenotype indicating that the heat shock conditions exerted a negative effect on the synthesis or assembly of the photsynthetic protein complexes. After maintenance at 37°C for 7–10 days, the hypocotyls and cotyledons were harvested and used for RNA extraction and cDNA synthesis. We first amplified cDNAs derived from the plastid ndhB mRNA since this transcript was previously shown to be highly edited in tobacco and maize (18,19). In addition, it offers the advantage of harboring a group II intron thus allowing for distinguishing cDNA amplification products from the corresponding DNA products by their different sizes. Amplification of cDNA samples from heat shock-treated plants and a control sample (kept at 25°C) using a ndhB-specific primer pair (with one primer in each exon) gave rise to products of identical size reflecting correctly spliced transcripts (data not shown). This suggests that plant maintenance at the elevated temperature did not significantly inhibit excision of the ndhB intron. We then sequenced directly the amplified cDNA populations using a primer suitable to determine the sequence in the region of ndhB editing sites II and III (8). Whereas sites II and III were virtually completely edited in the 25°C control sample, a major portion of the mRNA population turned out to remain unedited in the sample incubated at 37°C (Fig. 1).
In order to test whether plastid editing in general is compromised by elevated temperatures, two other sites in the ndhB transcript were analyzed. The editing sites IV and V (8) are located in the same exon of ndhB ∼100 nucleotides downstream of sites II and III. The amplification products analyzed for site II and III editing also cover sites IV and V thus allowing their direct use for the analysis of editing at the downstream sites IV and V. Interestingly, editing was found to be normal for both site IV and site V in the 37°C sample (Table 1) indicating that heat shock treatment does not exert a negative effect on editing at these two sites. Selective impairment of editing at sites II and III but not sites IV and V may suggest that some site-specific editing factor(s) involved in site II and III editing are either not synthesized or not functional under conditions of elevated temperature.
We next wanted to determine the time course of recovery of the ndhB site II and III editing activities. We therefore transferred heat shock-treated plants back to standard growth conditions (25°C) and took samples at daily intervals. Analysis of the corresponding ndhB cDNA populations revealed that, after only 1 day of incubation at 25°C, editing at both site II and site III was virtually completely restored (data not shown). This fast recovery of the editing activity may indicate that sufficient quantities of the sites II- and III-specific editing factors become rapidly available after lowering the temperature to 25°C.
Inhibition of editing by translational inhibitors
What could be the cause of the drastically reduced editing at ndhB sites II and III under heat shock conditions? As discussed above, one possibility is that elevated temperatures could prevent annealing of the site-specific factor(s) to the substrate mRNA. However, incubation of plants at high temperatures is also known to result in the loss of functional chloroplast ribosomes (20). Thus, the reduced editing at the ndhB sites II and III could also be a direct consequence of the absence of chloroplast translation. To distinguish between these two possibilities, tobacco seedlings were grown in the presence of prokaryotic translational inhibitors known to selectively block protein synthesis in the plastid compartment (21). Tobacco seeds still germinate normally in the presence of the aminoglycoside antibiotic spectinomycin at very high concentrations (500 mg/l) expected to lead to a nearly complete block of plastid translation. The seedlings are completely white (22) owing to the lack of chloroplast differentiation. Analysis of ndhB editing in these plants revealed an even more dramatic effect on editing at sites II and III. Site II was edited only to a very small extent and site III editing turned out to be completely abolished by spectinomycin (Fig. 1). As a control, again sites IV and V were analyzed. As observed for the heat shock-treated samples, site V was also completely edited in plants grown in the presence of spectinomycin. We observed, however, a small reduction of editing at site IV in the spectinomycin-treated but not in the heat shock-treated plants (Table 1).
Inhibition of editing at ndhB sites II and III not only by heat shock treatment but also by spectinomycin strongly suggests that the observed effects are due to the absence of plastid translation. Complete loss of site III editing (and nearly complete loss of site II editing) by spectinomycin treatment as compared with incomplete inhibition by incubation at 37°C may reflect a certain leakiness of the heat shock system, i.e. survival of a small population of plastid ribosomes resulting in some residual translational activity. This conclusion is also supported by the light green phenotype of the heat shock-treated plants as compared with completely white seedlings grown in the presence of the antibiotic.
RNA editing in the rpoB transcript of barley plastids was reported earlier to be independent of plastid translation (23). In order to test if inhibition of plastid translation also compromises ndhB editing in species other than tobacco we chose barley as a monocotyledonous plant and analyzed ndhB editing under heat shock conditions as well as after treatment with prokaryotic translational inhibitors. As numerous other graminaceous plants (24), barley was found to be endogenously resistant to spectinomycin, most probably due to the presence of a spectinomycin-resistant plastid 16S rRNA allele. We therefore tested streptomycin as an alternative antibiotic and bleaching of seedlings germinated on streptomycin-containing medium indicated sensitivity of the barley plastid translational apparatus to this drug. Analysis of ndhB site II and III editing revealed that maintenance of barley plants at 37°C nearly completely abolished editing at both sites (Fig. 2). In contrast with the results obtained with tobacco, the inhibitory effect of the antibiotic was smaller than the effect of the heat shock treatment (Fig. 2), most probably due to the delayed action of the drug as caused by its slow uptake by the large barley caryopses. The less severe effect of the antibiotic than that of the elevated temperature is in agreement with the observed incomplete bleaching of the seedlings by streptomycin.
ndhB transcript patterns are not affected by heat shock and antibiotic treatment
Theoretically, the absence of chloroplast translation could exert a negative effect on transcript processing and the inhibited ndhB site II and III editing could be a secondary consequence of a defect in one of the other RNA processing steps operating on plastid transcripts. For example, lack of processing of precursor transcripts into mature mRNAs can lead to a drastically reduced editing efficiency as previously observed for a rice mitochondrial transcript (25). To exclude this possibility, we analyzed the ndhB transcript patterns upon inhibited plastid translation as compared with a control sample. A major transcript species of ∼1.5 kb corresponding to fully processed monocistronic ndhB mRNA (32) was detectable in all samples suggesting that the absence of chloroplast translation does not block RNA processing (Fig. 3). This is consistent with the idea that all plastid RNA processing (and degrading) enzymes are probably encoded by nuclear genes. The hybridization detects in addition to the major 1.5 kb species several higher molecular weight mRNAs (Fig. 3) which most likely reflect precursor transcripts and partially processed mRNAs. Their relative abundance is significantly higher in the 37°C and spectinomycin-treated samples than in the 25°C control. However, by no means can this increase account for a complete loss of editing as observed for ndhB editing site III in antibiotic-treated plants.
It was recently shown that ndhB is transcribed by both plastid RNA polymerases: the plastid genome-encoded Escherichia coli-like enzyme and the nuclear-encoded phage polymerase-like enzyme (26). Absence of chloroplast translation is expected to result in a reduced activity of the plastid-encoded polymerase which most likely accounts for the somewhat lower RNA accumulation levels observed in the heat shock- and spectinomycin-treated samples as well as for the differences in relative transcript abundances.
ndhB editing in the absence of chloroplast development
The efficiency of plastid RNA editing can be subject to tissue-specific and developmental stage-specific changes (27–29). Thus, it could be possible that the impairment of editing at the two ndhB sites is due to the absence of chloroplast differentiation as caused by the inhibition of plastid translation. To exclude this possibility, we analyzed the efficiency of ndhB editing in non-green tissue. ndhB cDNAs derived from tobacco root RNA were amplified and their editing status was determined by direct sequencing. We found ndhB site III (which remained completely unedited upon treatment with spectinomycin) as well as site IV (showing reduced editing in the presence of spectinomycin) fully edited in non-green root plastids indicating that chloroplast differentiation is not required for editing at these sites. Editing at site II (which was partially inhibited by spectinomycin and heat shock treatment), however, was reduced to ∼50% in root plastids (Fig. 4) suggesting that the absence of chloroplast differentiation could account for the incomplete editing of ndhB site II upon inhibition of plastid translation. This finding indicates that only ndhB site II is subjected to plastid differentiation-related changes in editing efficiency and the reduced editing at this site in the absence of plastid translation may thus, at least to a certain extent, be due to the inhibition of chloroplast development. At site III, however, where editing was completely blocked by spectinomycin, the efficiency of editing seems to be independent of chloroplast differentiation.
Sensitivity of editing at other sites to heat shock and antibiotic treatment
In order to test whether the inhibition of editing in the absence of plastid translation is restricted to the analyzed ndhB editing sites, we systematically tested all known editing sites encoded in the tobacco chloroplast genome and determined the editing efficiencies under heat shock conditions as well as upon inhibition of protein synthesis by spectinomycin. Editing at most of the sites was still virtually complete in the absence of translation (Table 1). None of the other editing sites encoded in the tobacco chloroplast genome was found to be completely blocked by the loss of plastid translation as observed for ndhB editing site III upon plant treatment with spectinomycin. We, however, observed a decrease in editing efficiency for some sites: two other sites in the ndhB transcript (sites 0 and I; 19), the ndhD initiation codon (30) and two sites in the tetracistronic psbE/F/L/J transcript (2; J. Kudla and R. Bock, unpublished; Table 1). As in the case of ndhB sites II and III, the inhibitory effect of spectinomycin was more severe than that of the elevated temperature, most likely reflecting a less efficient translation inhibition by heat shock treatment than by the action of the antibiotic. As for sites II, III and IV, we also tested the ndhB editing sites 0 and I for whether or not the reduced editing efficiency can be attributed to the lack of chloroplast development. We found both sites to undergo complete editing in root tissue (data not shown) demonstrating that they are efficiently edited also in the absence of photosynthesis and chloroplast differentiation.
In the course of this work, we have identified heat shock and plant treatment with prokaryotic translational inhibitors as conditions that inhibit editing at a limited number of plastid RNA editing sites. Block of chloroplast translation by spectinomycin leads to a complete loss of RNA editing at ndhB editing site III and reduces editing at some other editing sites encoded in the chloroplast genome (Table 1). Our results establish for the first time, and in contrast with earlier belief (23), a role of plastid translation in the editing of specific sites.
It is well established that certain antibiotics can interact directly with RNA molecules (36). We have, therefore, considered the possibility that the inhibition of plastid editing by spectinomycin or streptomycin is due to a direct binding of antibiotic molecules to the editing sites (or to adjacent mRNA sequences acting as cis-elements for editing site recognition). We consider this unlikely since the editing at those sites affected by aminoglycoside antibiotics is also inhibited by heat shock treatment, strongly suggesting that the block of translation is the cause of the inhibited editing reaction.
Amplification of spliced ndhB cDNA as well as identity of the transcript patterns in green versus translationally inhibited tissue (Fig. 3) indicate that the loss in editing activity does not reflect a dependence of these editing events on splicing or other RNA processing steps. We have also considered the possibility that, in the absence of chloroplast translation, ndhB transcripts are extremely unstable, thereby not allowing the editing machinery to act efficiently on these substrates. If this were the case, the difference in half-life time of the ndhB transcripts would have to be several orders of magnitude, since no editing at all was detected at site III in spectinomycin-treated plants whereas the site undergoes complete editing in untreated control samples. Also, this scenario would involve the additional assumption that editing at the affected sites is much slower than at other sites in the very same transcript (e.g. sites V–VIII which are fully edited in heat shock- and antibiotic-treated plastids). We believe this not to be the case since our transcript analyses failed to provide evidence for substantial degradation of ndhB mRNAs in heat shock- and antibiotic-treated plants (Fig. 3).
Analysis of editing in non-green tissue revealed that, except for the reduced editing at site II, the observed effects of inhibited plastid translation on ndhB editing cannot be attributed to the absence of chloroplast development. We therefore propose that plastid translation plays a role in editing of some chloroplast RNA editing sites. It was shown earlier that, in the ribosome-free barley mutant albostrians, the plastid rpoB transcript is fully edited in spite of the mutant's deficiency in plastid translation (23). In tobacco, we also found rpoB editing not affected by blocked plastid translation (Table 1). Inhibition of ndhB site III editing in both tobacco and barley and reduction of editing at a number of other sites, however, suggest that plastid RNA editing is not generally independent of translation.
What could be the role that plastid translation plays in editing of those sites? Formally, two possibilities can be envisaged: (i) plastid DNA-encoded gene products are directly involved in the editing reaction by acting as site-specific trans-acting factor(s) for the molecular recognition of some editing sites or (ii) plastid translation product(s) serve as auxiliary factor(s) mediating or improving the accessibility of the substrate site to the editosome. We favor the latter possibility since it provides a more plausible explanation for the fact that, except for ndhB site III where editing is completely blocked in the absence of translation, editing is incompletely inhibited in the other cases. At present, we can only speculate about which plastid genome-encoded gene product(s) could be involved in the editing of these sites. Proteins with RNA-binding capacity, such as ribosomal proteins, would be attractive candidates. Alternatively, active translation of the transcript by functional chloroplast ribosomes might resolve mRNA secondary structures that otherwise would disturb the accessibility of the editing site to the plastid RNA editing machinery.
In view of the many similarities between the RNA editing processes in chloroplasts and plant mitochondria, it would be interesting to test the plant mitochondrial RNA editing system for its sensitivity to heat shock and translational inhibitors. However, nothing is currently known about the sensitivity of the plant mitochondrial translational apparatus to elevated temperatures or antibiotics. None of the many spontaneous antibiotic resistance mutations characterized thus far appears to reside in the plant mitochondrial genome (31) suggesting that the plant mitochondrial translational apparatus may be resistant to at least some of the classical prokaryotic inhibitors of protein biosynthesis.
This work was supported by grant BO 1482/3-2 from the Deutsche Forschungsgemeinschaft to R.B.