SEARCH

SEARCH BY CITATION

Keywords:

  • RNA editing;
  • plant mitochondria;
  • PPR protein;
  • RNA editing factor;
  • lovastatin

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Post-transcriptional RNA editing in flowering plant mitochondria alters several hundred nucleotides from cytidine to uridine, mostly in mRNAs. To characterize the factors involved in RNA editing in plant mitochondria, we initiated a screen for nuclear mutants defective in RNA editing at specific sites. Here we identify the nuclear-encoded gene MEF11, which is involved in RNA editing of the three sites cox3-422, nad4-124 and ccb203-344 in Arabidopsis thaliana. A T-DNA insertion line of this gene was previously characterized as showing enhanced tolerance to the compound lovastatin, an inhibitor of the mevalonate pathway of isoprenoid biosynthesis. The mef11-1 mutant described here shows similar tolerance to lovastatin. Identification of the function of the MEF11 protein in site-specific mitochondrial RNA editing suggests indirect effects of retrograde signalling from mitochondria to the cytoplasm to evoke alteration of the mevalonate pathway. The editing sites cox3-422 and ccb203-344 each alter amino acids that are conserved in the respective proteins, while the nad4-124 site is silent. The single amino acid change in the mef11-1 mutant occurs in the second pentatricopeptide repeat, suggesting that this motif is required for site-specific RNA editing.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

RNA editing in mitochondria of flowering plants changes 400–500 selected cytidines to uridines, mostly in coding regions of mRNAs and some tRNAs (Giegé and Brennicke, 1999; Handa, 2003; Takenaka et al., 2008). At present, the biochemistry of the RNA editing that effects the site-specific deamination of cytidine to uridine has not yet been elucidated. In recent years, in organello and in vitro analyses of mitochondrial RNA editing have identifed several cis-elements required for editing-site recognition in the affected RNA molecules, but no trans-factors (Farréet al., 2001; Neuwirt et al., 2005; van der Merwe et al., 2006). The first nuclear-encoded specificity factor MEF1 (mitochondrial editing factor) involved in recognition and targeting of this activity to three editing sites in plant mitochondria was identified by genetic analysis of an ecotype-specific difference in RNA editing efficiency in Arabidopsis thaliana (Zehrmann et al., 2009).

The mitochondrial editing factor MEF1 and the recently characterized OGR1 factor of rice (Kim et al., 2009) are pentatricopeptide repeat proteins (PPR proteins) of the DYW sub-group. Several of the specific nuclear factors identified as being involved in RNA editing events in plastids are also members of this class of DYW PPR proteins, while others are PPR proteins but do not contain the DYW C-terminus (Kotera et al., 2005; Sasaki et al., 2006; Shikanai, 2006; Okuda et al., 2007, 2008; Chateigner-Boutin et al., 2008; Zhou et al., 2009; Robbins et al., 2009; Yu et al., 2009). In flowering plants, about 450 PPR protein coding genes are found, most of which are predicted to be targeted to plastids and/or mitochondria (Small and Peeters, 2000; Lurin et al., 2004; Schmitz-Linneweber and Small, 2008). The PPR proteins have been sub-grouped according to the type and length of C-terminal extensions beyond the characteristic 4–12 repeats of 35 amino acids. About 140 of the PPR proteins contain the DYW domain as their C-terminal extension; this domain displays a signature characteristic of Zn-containing cytidine deaminases. Consequently, the DYW domain has been proposed to be potentially involved in C to U RNA editing in the two plant organelles (Salone et al., 2007; Rüdinger et al., 2008).

Although deaminating activity of the DYW PPR proteins still needs to be demonstrated, RNA binding properties of the PPR proteins involved in plastid RNA editing have been experimentally verified for the CRR4 protein (Okuda et al., 2006). Of the PPR proteins in plants that have been implicated as active in other post-transcriptional processes in plastids, such as specific intron splicing events, processing of multi-cistronic pre-mRNAs, stabilization of pre-tRNAs etc, direct and specific attachment to the substrate RNA has been shown in several instances (Beick et al., 2008; Williams-Carrier et al., 2008; Schmitz-Linneweber et al., 2005; for review, see Andrés et al., 2007 and Delannoy et al., 2007).

To identify further factors involved in the RNA editing process in plant mitochondria, we have initiated a two-pronged genetic approach. The first approach, exploitation of ecotype-specific variants, led to identification of the MEF1 encoding gene (Zehrmann et al., 2008, 2009). The second approach, direct identification of mutant plants impaired in RNA editing in a population of EMS mutants in Arabidopsis thaliana (Takenaka and Brennicke, 2009), is used here to identify the nuclear-encoded gene required for RNA editing at three distinct sites in three different mitochondrial mRNAs, sites cox3-422, nad4-124 and ccb203-344.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Identification of RNA editing factor MEF11 as the loi1 gene product

To screen for mutants deficient in RNA editing at specific sites in plant mitochondria, we recently developed a multiplexed SNaPshot approach (Takenaka and Brennicke, 2009). This assay was used to directly identify mutant plants impaired in RNA editing in a population of EMS mutants in the Columbia (Col) ecotype genetic background of Arabidopsis thaliana. One of the RNA editing defects manifest as in no detectable editing at site cox3-422 (Figure 1). This mutant editing phenotype was assigned to an individual plant in the screen. Interestingly, another RNA editing defect, at site nad4-124 (Figure 1), is seen in the same plant, suggesting that the underlying mutation in a presumably nuclear-encoded gene is responsible for reduced or absent editing at these two distinct sites in two different mitochondrial mRNAs.

image

Figure 1.  RNA editing is not detectable at two specific sites in the mitochondria of mef11-1 mutant plants. Comparison of the cDNA SNaPshot analysis of two RNA editing sites each in the mitochondrial cox3 and nad4 mRNAs between wild-type Arabidopsis thaliana (WT) and the mef11-1 mutant plants shows that the mutant has lost the ability of C[RIGHTWARDS ARROW]U editing at sites nad4-124 and cox3-422. In contrast, two other sites in the same mRNAs, nad4-166 and cox3-314, are correctly edited in wild-type and mutant plants. In the cDNA strand analysed, the detected A nucleotide (green trace) corresponds to the edited U, and the observed G nucleotide (blue trace) is derived from an unedited C.

Download figure to PowerPoint

To map and identify the disrupted gene encoding mitochondrial RNA editing factor 11 (MEF11), the Arabidopsis thaliana mutant plant (mef11-1) in the Col genetic background was crossed with wild-type Landsberg erecta (Ler) plants, and the F1 generation was selfed. Of the F2 offspring, 96 individual plants were analysed for the reduced RNA editing phenotype to identify homozygous mutant individuals. This screen was performed by analytical RT-PCR assays, for which primers were designed which allow edited and unedited mRNAs to be distinguished (Figure 2).

image

Figure 2.  RT-PCR assay with T- and C-specific primers to differentiate between edited and unedited mRNAs in a screen of the F2 generation from the cross between mef11-1 and Ler plants. The top part shows the primer design. In addition to the 3′ G or A at the edited or unedited nucleotide, a T·C or G·A mismatch is introduced two nucleotides upstream to enhance the importance of base pairing at the editing site. The lower part shows a sample gel image of RT-PCR reactions using these primers on RNA preparations from four individual F2 plants. The plant on the left (81) without detectable editing is homozygous for the mutation in the Col background (mef11-1/mef11-1), the two plants in the center (82 and 83) are heterozygous individuals (mef11-1/Ler), and the plant on the right (84) is homozygous for the wild-type Ler allele (Ler/Ler). The differential C and T signals in the gel image indicate the presence (+) or absence (−) of editing at this site (cox3-422).

Download figure to PowerPoint

The 20 homozygous plants identified were tested for recombination events using a set of nuclear Ler/Col SNP markers, which identified chromosome 4 as the site of the mutated gene locus. To map the gene further, all 76 editing-competent plants were screened to identify individuals in which crossing-over events recombined the wild-type Ler region containing the MEF11 locus with left or right mef11-1-derived Col sequences in one chromosome, with the other chromosome showing all markers of the Col background derived from the mef11-1 allele. This detailed mapping of the homozygous mutant and heterozygous editing-positive plants narrowed the location of the affected gene to a window of about 1.0 Mb.

The 76 plants with intact editing at the two sites were investigated for their genetic constitution in the genomic region of the mef11-1 locus. Both homozygous wild-type plants (Ler/Ler) as well as heterozygous plants (mef11-1/Ler) show complete editing at the cox3-422 site, confirming the recessive nature of the mutant mef11-1 allele. The dominant trait of the wild-type MEF11 allele is demonstrated by the full editing competence of the heterozygous F2 plants.

We then searched the annotated genes in the 1.0 Mb region for likely candidate genes. The PPR proteins are the most probable editing site-specific co-factors as such proteins have been found to be involved in several editing events in plastids (Kotera et al., 2005; Okuda et al., 2006; Chateigner-Boutin et al., 2008) and in plant mitochondria (Kim et al., 2009; Zehrmann et al., 2009). In the genomic 1.0 Mb window on chromosome 4, six genes are predicted to encode PPR proteins: At4g14050, At4g14170, At4g14190, At4g14820, At4g14850 and At4g15720. These six genes were sequenced in wild-type Col and Ler plants and in the mef11-1 mutant line. Five of these coding sequences were identical between wild-type Col and the mutant. Only the gene located at At4g14850 showed a single nucleotide difference between wild-type Col and the mutant, suggesting that this altered nucleotide is the cause of the reduced editing phenotype. This nucleotide alteration is a C[RIGHTWARDS ARROW]T transition, a typical EMS-induced mutation.

This gene MEF11 at At4g14850 had been previously found to play a role in physiological plant growth reactions to the compound lovastatin (Kobayashi et al., 2007a). Knockout mutant lines of this gene confer enhanced tolerance to lovastatin, and this gene had accordingly been termed LOVASTATIN INSENSITIVE 1 (LOI1). Here we use the name MEF11 for locus At4g14850 as it better describes the primary function of the encoded protein moiety, but indicate LOI1 in parentheses when referring to the context of this previous work.

The mef11-1 mutant shows slightly increased tolerance against lovastatin

To test the physiological consequences of the mef11-1 EMS mutant on plant growth, seeds of wild-type-Col and mef11-1 mutants were germinated, and seedlings were allowed to grow either in the presence or absence of lovastatin in the growth medium (Figure 3). In the absence of lovastatin, comparison of the root lengths showed similar growth patterns between the wild-type and mutant plants, with the mutants slightly lagging in their development. In the presence of lovastatin, root elongation was inhibited more profoundly in the wild-type plants than in the mutant plants. These observations suggest that, like the loi1 mutants, the mef11-1 mutant plants tolerate lovastatin better than the wild-type plants. The reduced root length may be a direct effect of inhibition of the mevalonate (MVA) pathway or may be indirectly caused by a reduction of cytokinin concentrations (Hartig and Beck, 2005). In the mef11-1 mutant, tolerance against lovastatin is less pronounced than in the tagged insertion mutant plants (loi1-1; Kobayashi et al., 2007a). The latter are therefore more suitable for further physiological analysis.

image

Figure 3.  Mutation of the MEF11 gene confers slightly increased tolerance towards lovastatin. Plants homozygous for the mef11-1 mutant allele developed slightly more slowly than wild-type (WT) seedlings, but otherwise were phenotypically normal. The presence of lovastatin (+) reduces root length growth in wild-type seedlings of Arabidopsis thaliana more effectively than in the mef11-1 mutant plants (Kobayashi et al., 2007a). This observation suggests a retrograde connection between the mitochondrial RNA editing effect and the whole-cell response of increased MVA pathway activity. Whether the lovastatin-induced reduced development of root length is a direct effect of inhibition of the MVA pathway or is indirectly caused by a resulting reduction of cytokinin concentrations is at present unclear (Hartig and Beck, 2005).

Download figure to PowerPoint

Analysis of T-DNA insertion line mef11-2 of the At4g14850 gene

Although no RNA editing at the two affected sites cox3-422 and nad4-124 was detected in the mutant mef11-1 by SNaPshot or RT-PCR sequence analysis (Figure 1), a very low level of residual RNA editing activity cannot be excluded. To characterize the phenotype of the RNA editing pattern in a potential gene knockout, we analysed a T-DNA insertion line of the MEF11 locus at At4g14850 from the SALK collection, which was designated mef11-2 (Figure 4). Like the mef11-1 EMS mutant, plants of the homozygous SALK line mef11-2 (SALK_131734) showed the same phenotype of RNA editing deficiency at the cox3-422 and nad4-124 sites (Figure 4), but showed normal growth under standard greenhouse conditions.

image

Figure 4.  The T-DNA insertion in the MEF11 gene in a SALK line abolishes RNA editing at the same sites as the mef11-1 mutation. The top part shows the schematic structure of the MEF11 PPR protein encoded by locus At4g14850. The amino acid sequence between the C-terminal P repeat and the DYW domain is similar to the E domains usually encoded at this position, but is not predicted to be a genuine E domain. This region is therefore designated ‘E-like’. The asterisk indicates the location of the single nucleotide mutation in the mef11-1 mutant, and the resulting amino acid change is indicated above. The T-DNA insertion site of the mef11-2 mutant (homozygous SALK line SALK_131734) is indicated by the black triangle. The bottom part shows the cDNA sequence traces for two editing sites each in the cox3 and nad4 mRNAs from the SALK mef11-2 mutant. As in the mef11-1 mutant, the sites cox3-422 and nad4-124 are not edited, while the nearby sites cox3-413 and nad4-158 are fully edited from C to T. Color traces are C, blue, T, red, G, black, A, green.

Download figure to PowerPoint

In the heterozygous state, both the mef11-1 EMS mutant and the SALK line mef11-2 show the same wild-type phenotype of complete RNA editing at the cox3-422 and nad4-124 sites, thus both mutations are recessive. This suggests that neither the truncated MEF11-2 protein in the T-DNA insertion line (if it is stably synthesized at all) nor the altered MEF11-1 protein compete sufficiently with the wild-type version of the MEF11 protein encoded by the wild-type allele to influence the outcome of editing at these target sites.

These observations further confirm that editing at both these mitochondrial RNA sites requires an intact MEF11 gene, and show that the MEF11 gene does indeed encode a specific mitochondrial RNA editing factor. To further verify this, we tested the rescue of RNA editing by an intact MEF11 gene in the mef11-1 mutant.

Complementation of the mef11-1 mutant

The connection between the MEF11 gene and the competence for RNA editing at the cox3-422 and nad4-124 sites was assayed by exploring the ability of the wild-type Col MEF11 gene to complement the mef11-1 mutant. The wild-type Col MEF11 gene was cloned under the control of the CaMV 35S promoter, and the plasmid DNA was transfected into mef11-1 mutant protoplasts (Yoo et al., 2007; Zehrmann et al., 2009). The MEF11 wild-type gene did indeed restore the ability for RNA editing at both editing sites in the transfected mutant protoplasts (Figure 5).

image

Figure 5.  Complementation of protoplasts from the EMS mutant line mef11-1 with the Col MEF11 gene and the mef11-3 and mef11-4 mutated sequences. The top line shows the amino acid sequence differences in the MEF11 protein between the three ecotypes Col, Ler and C24. Neither of the amino acids altered in comparison to the Col sequence detectably affects RNA editing. The schemes below indicate the locations of the mef11-3 and mef11-4 mutations. The resulting amino acid changes are given underneath. An editing site-specific cDNA analysis for cox3-422 (left) reveals edited as well as unedited products after complementation with the mutant mef11-3 and only unedited products in the control transfection with the gene for the unrelated editing factor mef1. The sequence traces compare the effects of protoplast complementation assays of the mef11-1 mutant with the Col gene and the mef11-3 and mef11-4 mutant genes. All these variants of the MEF11 Col gene restore RNA editing at both the cox3-422 and nad4-124 sites. As a control, transfection with the MEF1 gene was assayed. This gene is involved in editing at other mitochondrial editing sites and does not complement the mef11-1 mutant protoplasts.

Download figure to PowerPoint

Although not the full level of 100% editing as seen in Col wild-type plants was recovered, the level of successful editing of approximately 50% observed when mutant plants were complemented with the wild-type MEF11 gene versus undetectable editing in control transfections with only a GFP sequence in the vector indicate that the wild-type MEF11 gene does recover the ability for RNA editing at these sites. Further control transfections with the gene for a different mitochondrial RNA editing factor, MEF1, which is involved in editing at three other sites (Zehrmann et al., 2009), confirmed that these factors are specific for their cognate editing sites and that MEF1 cannot substitute for MEF11.

The recovered lower level of editing in comparison to the wild-type Col plants (100% editing) may be due to not all protoplasts being transfected successfully, incomplete assembly of the introduced factor with the putative additional endogenous RNA editing components, or the time for which the protoplasts are incubated. The incompletely recovered RNA editing may also be caused by inherent properties of the protoplasts in connection with the editing activity. In these mef11-1 mutant protoplasts, competition between the non-functional MEF11-1 protein and the wild-type MEF11 synthesized from the transfected plasmid is unlikely as heterozygous mef11-1 mutant plants show normal levels of complete RNA editing. Any competition between the mutant and wild-type MEF11 proteins would be expected to lower the editing efficiency in these heterozygous plants, which is not observed. These results show that a functional MEF11 gene is indeed required for RNA editing at both the cox3-422 and nad4-124 sites.

Target sequences at the RNA editing sites share scattered nucleotide identities

The nucleotide sequence motifs required in cis for correct targeting of the sites cox3-422 and nad4-124 have not yet been determined directly. However, in vitro and in organello investigations have delineated the cis-elements in the RNA context for several other editing sites in mitochondria and plastids as approximately 20–25 nucleotides upstream (5′) and 1–3 nucleotides downstream (3′) of the edited C (Chaudhuri and Maliga, 1996; Farréet al., 2001; Miyamoto et al., 2002; Hegeman et al., 2005; Neuwirt et al., 2005; van der Merwe et al., 2006), and a similar size for the recognition sequences of these cox3 and nad4 sites may be expected. Within this 28 nucleotide window between −25 and +3 relative to the edited C, nine scattered nucleotides are shared between the genomic nucleotide sequences around the two RNA editing sites targeted by the MEF11 gene product, cox3-422 and nad4-124 (Figure 6a, top). Interestingly, other RNA editing events upstream of the edited C in each of these two presumed specific recognition regions raise the number of shared nucleotide identities from 9 to 11.

image

Figure 6.  Comparison of nucleotide identities in the presumed recognition sequences around the cox3-422 and nad4-124 editing sites. (a) The top alignment shows the nucleotides from −25 to +3 around the two editing sites identified in the SNaPshot mutant screen as affected in the mef11-1 mutant. Shared nucleotide identities are highlighted in grey. Upstream editing events (bold U) in each of the two mRNAs increase the number of shared identical nucleotides from 9 to 11. The alignment below the horizontal line shows the results of an in silico search for annotated editing sites with similar upstream sequences in the mitochondrial transcriptome. Three editing sites share 6-8 nucleotides with both bona fide MEF11 target sites (black shading); in the nad4L-95 site, an upstream editing event (bold) also increases the number of shared nucleotides. However, this site is not affected in the mutants, but site ccb203-344 is targeted by MEF11 (Figure 7). The bottom alignments show the extensive stretch of shared nucleotides (shaded grey) between sites ccb203-344 and cox3-422 and the nucleotides shared between sites ccb203-344 and nad4-124. (b) Comparison of nucleotide and amino acid alignments shows that the site of the cox3-422 editing event is conserved (asterisk above the bold C) in dicots and also in wheat. Surprisingly, this site is not edited in rice even though the surrounding nucleotides are identical to the other plants aligned here. Interestingly, an upstream editing site (bold C) is also altered in dicots and wheat, but not edited in rice. The amino acid comparison shows that the unedited sites in rice lead to radical amino acid differences (boxed). The bottom part shows that the ccb203-344 editing event (asterisk above the bold C) is conserved in dicot and monocot plants, including rice. The amino acid leucine (L) introduced into the CCB203 protein by the ccb203-344 editing event (asterisk) is conserved between flowering plants by analogous editing events. All sequences are shown 5′[RIGHTWARDS ARROW]3′ or from the N- to the C-terminus, from left to right.

Download figure to PowerPoint

Searching the entire mitochondrial genome of Arabidopsis in silico, no other full match with these 9-11 nucleotides was found. Several instances where 8 or 9 out of the 11 nucleotides shared were seen at the correct distance from a C in other mitochondrial mRNAs; however, these Cs are not edited (data not shown). Of the annotated editing sites, only nad4L-95, nad7-316 and ccb203-344 show 7 or 8 shared nucleotides (Figure 6a). To determine whether these similarities are sufficient to target editing, we investigated their status in mutant mef11-1. Editing sites nad4L-95 and nad7-316 are not affected by the mutation in the MEF11 gene, but editing at site ccb203-344 is abolished in the mutant line (Figure 7). Surprisingly, in the SALK line mef11-2, approximately 60% editing is detected at this site. To corroborate the connection between MEF11 and editing site ccb203-344, we tested complementation of the editing defect in protoplasts from the mef11-1 mutant (Figure 7, bottom). The rescue of editing at this site in the mef11-1 mutant confirms that the MEF11 protein is also required for this editing event.

image

Figure 7.  A third RNA editing site target is identified by in silico analysis. The top line shows the MEF11 protein with the location of the mef11-1 SNP mutation and the T-DNA insertion site in the mef11-2 SALK line. Editing at the ccb203-344 site is not detectable in the mef11-1 mutant, and is decreased to approximately 60% in the mef11-2 T-DNA insertion line. The nearby editing site ccb203-356 is not affected and is edited in both mutants. Protoplast complementation of the mef11-1 mutant with the MEF11 gene and the two in vitro generated mutants mef11-3 and mef11-4 recovered similar levels of RNA editing at site ccb203-344. Control transfection with the gene for the unrelated MEF1 editing factor did not regenerate the ability to edit this site.

Download figure to PowerPoint

These observations suggest that the eight nucleotides in addition to the edited C shared by the three bona fide MEF11 targets or their alternative combinations of nucleotides (Figure 6a, bottom) may be sufficient to specify these unique sites in the plant mitochondrial transcriptome. Only five nucleotides are shared by the three target editing sites of the previously identified editing factor MEF1 (Zehrmann et al., 2009), which are not sufficient to guide editing by MEF1, and additional features defining the specificity of the interaction are required in those RNA targets.

The two editing sites in the plastid transcriptome targeted by the CLB19 PPR protein likewise show little sequence conservation in their vicinity (Chateigner-Boutin et al., 2008). In addition, a thorough biochemical analysis of two RNA editing sites in tobacco suggests that one protein is involved in binding at both sites even though they share very little conserved sequence (Kobayashi et al., 2007b).

In contrast, the eight or more nucleotides shared between the three sites targeted by the MEF11 editing factor identified here may be sufficient as a specific binding signal. The increased number of shared nucleotides in the pairwise comparison between the three sites is especially intriguing (Figure 6a), as this may suggest that different domains of the MEF11 protein could be involved in binding these extended motifs, provided that MEF11 is indeed responsible for the specific interaction with the RNA. This was addressed experimentally through in vitro or in vivo analyses of mutations.

Differential effect of mutations in the MEF11 sequence on the RNA editing activity

The MEF11 gene product is a PPR DYW protein and contains fourteen 35 amino acid repeats followed by an E-like domain, and terminates with a DYW domain (Figure 4). The ability of the N-terminal amino acids to direct import of the MEF11 protein into mitochondria has been investigated previously by fusion of the 39 N-terminal amino acids to the GFP sequence (Kobayashi et al., 2007a).

Like the mef11-2 SALK line investigated here, the loi1-1 and loi1-2 T-DNA activation tagging mutants analysed for their lovastatin resistance (Kobayashi et al., 2007a) had disrupted reading frames as a result of the insertions (Figure 3). These disruptions destroy the open reading frame of the gene and thus presumably create null alleles. In contrast, the mef11-1 EMS mutant line differs from the Col-wild-type MEF11 gene by a single nucleotide alteration, and thus a single amino acid alteration of Leu to Phe at residue 48. This change is located in the second PPR repeat, suggesting that this motif is essential for function (Figure 4).

To investigate the suitability of the protoplast transfection system for functional analysis of the PPR repeats, the E-like and/or the DYW domains, several point mutations were generated by cloning PCR products from the wild-type Col MEF11 sequence using an error-prone Taq polymerase enzyme. The mutant sequences selected, mef11-3 and mef11-4, were transfected into mef11-1 protoplasts, and the levels of RNA editing recovered were analysed (Figure 5). Gene variant mef11-3 contains a single mutation in the fourth PPR repeat, which changes amino acid 125 from Val to Thr, and was able to restore RNA editing in the protoplasts to levels comparable with the introduced wild-type Col MEF11 gene (Figure 5). The mef11-4 gene variant with single nucleotide variations in PPR repeats 8 and 12 and in the E-like domain showed similar levels of recovered RNA editing in the transfected protoplasts (Figure 5). This result suggests that the functional effect of targeted mutations can be conveniently and rather rapidly investigated using the protoplast transfection system.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The MEF11 gene was originally identified as encoding a potential regulatory factor of isoprenoid biosynthesis in Arabidopsis thaliana (LOI1; Kobayashi et al., 2007a), but has been found to encode the MEF11 factor of RNA editing in plant mitochondria. The three RNA editing sites for which an intact MEF11 gene is required differ in their effect on the encoded mitochondrial proteins. While the nad4-124 site is silent and thus does not alter the NAD4 protein, the cox3-422 site changes a proline residue acid to leucine at amino acid 141 in COX3 (Figure 6b). All dicot plants investigated for this editing event have the leucine amino acid in this position, as predicted by the cDNA sequences either after or without an RNA editing event. However, the monocot rice COX3 protein is apparently fully functional with the proline amino acid at this position (Figure 6b). It is therefore unclear what effect the proline residue has on the COX3 function, given that the leucine is evolutionarily conserved in dicot plants. Proline is known to introduce a bend in the polypeptide chain, but the effect on the function of the COX3 subunit and how this may connect to the cytosolic MVA pathway needs to be investigated.

Similarly, the ccb203-344 editing event changes a proline residue to leucine at position 115 in CCB203. All flowering plants have leucine in this position in the protein (Figure 6b), suggesting that this conserved residue is important for proper function. As suggested for COX3, an impaired function of the CCB203 protein might have physiological consequences for the MVA pathway, in this case through altered efficiency of the assembly of the haem moieties on the outside of the inner mitochondrial membrane. The less efficient cytochrome b activity that results may trigger a signal for an increased demand for ubiquinone through the MVA pathway.

Functional implications of the MEF11 mutations

The MEF11 protein has been shown to bind to ssRNA homopolymers (LOI1; Kobayashi et al., 2007a). However, so far no C to U deaminase activity has been shown for any of the PPR proteins involved in RNA editing at specific sites in plastids or mitochondria, and it is generally assumed that additional proteins are required. Therefore, the point mutation in mef11-1 in the second PPR repeat (Leu to Phe), which abolishes detectable in vivo editing, may affect either binding to the RNA or the interaction with other proteins. Alternatively, the overall stability of the MEF11 protein may be compromised and its rapid degradation may be responsible for the lack of editing at these sites. The one altered nucleotide in the fourth PPR domain in mef11-3 and the three altered amino acids in the mutant mef11-4 do not seem to interfere with either protein stability, RNA binding or connection to other editing proteins, as the resulting protein variants appear to fully complement the non-functional mutant mef11-1 (Figure 5).

The observed strong preference for binding to homopolymers of rG over all other nucleotide identities (Kobayashi et al., 2007a) is not reflected in the primary sequence around the affected editing sites (Figure 6). The G content is not particularly increased, and the sequence specificity may instead reside in the order of the nucleotides, rendering the assays with the homopolynucleotides not directly informative for the in vivo situation and site recognition. Nevertheless, the confirmed binding of the MEF11 (LOI1) protein to the ssRNA suggests that the protein makes direct contact with the RNA to be edited in vivo.

Which specific nucleotide or motif is the crucial element for binding of the MEF11 protein to the target editing sites remains to be investigated in detail. The observation that the nucleotide identities shared between the three affected editing sites are located mostly between positions −4 and −25 relative to the targeted C agrees with previous in vitro findings, which had shown that nucleotide identities between the edited nucleotide and nucleotide −5 are not crucial for efficient editing (Neuwirt et al., 2005; van der Merwe et al., 2006; Takenaka et al., 2008).

This observation and the successful in silico identification of the third RNA editing target of MEF11 suggest that the nucleotide identities shared between the three sites in the Arabidopsis thaliana mitochondrial transcriptome are indeed involved in site recognition, and may thus be sufficient to specify these three RNA editing sites. Dedicated in vitro mutation assays in the vicinities of the RNA editing sites will be required to define whether these 9–11 shared nucleotides are the recognition motif.

The improvement of the number of shared nucleotide identities in the presumed recognition region of the cox3-422 and nad4-124 RNA editing events by editing at upstream nucleotide positions suggests the intriguing possibility that these upstream editing events may precede site recognition by the MEF11 protein (Figure 6). This potential hierarchy of editing sites may be supported by the observation that, in rice, in which site cox3-422 is not edited, the upstream site cox3-413 is also not edited. To investigate the order of editing events in this region, we analysed the status of the downstream site cox3-422 in mRNAs in which the upstream site cox3-413 either was or was not edited. In the sub-population of mature cox3 transcripts in which the upstream site cox3-413 is edited, the downstream site cox3-422 is fully edited to 100% T in the cDNA sequence (data not shown). In transcripts in which the upstream site cox3-413 is not edited, the downstream site cox3-422 is still edited to approximately 40% T (data not shown). In this latter assay, ‘young’ RNA molecules are analyzed in which no editing has taken place at the upstream site and thus complete editing cannot be expected for the downstream site. This assay addressed the question of whether any RNA editing at the downstream site is possible at all without editing of the upstream site with the resulting lower number of common nucleotide identities. The fact that downstream editing occurred without editing at the upstream site indicates that editing at this upstream site is not canonically required for recognition by the specific editing factor MEF11 even though this editing does improve the number of shared nucleotides.

PPR repeats are important for competence in specific RNA editing

The single amino acid exchange in the second PPR repeat in the mef11-1 mutant shows that this repeat unit is crucial for the specific editing events targeted by the MEF11 protein. If this mutant does not just destabilize the protein, its identification will allow testing of the function of these PPR repeats experimentally, for example by comparing the RNA binding properties of the wild-type and mutant proteins in gel shift assays.

Another potentially interesting line of investigation arises as a result of the observation that the two sites cox3-413 and cox3-422 are not edited in rice, although, like the genomes of poplar and grape, the rice genome does encode an ortholog of MEF11. These proteins share about 60% identical amino acids, and cross-species tests of the functional equivalence of these proteins may yield information on the function of the various domains in the MEF11 protein.

The surprising observation that the ccb203-344 editing event is decreased but still occurs in the T-DNA insertion line mef11-2 while the other two sites appear completely blocked (Figure 7) requires explanation. One possibility could be that another factor can also target this site and perform editing when MEF11 is mutated. However, in the mutant mef11-1, no trace of editing at site ccb203-344 is detected. To explain this, it is necessary, for example, to invoke an altered binding specificity by which the MEF11-1 protein blocks any access to the editing event at ccb203-344. However, this would not allow complementation of the mef11-1 mutant, which is clearly documented. A more likely explanation may be that the T-DNA insertion in mef11-2 does not completely block transcription and translation of the MEF11-2 protein, but only alters its expression and/or its function by altering the C-terminus of the DYW region. Processing of the transcripts does not appear to be affected, as analysis of the mef11-2 cDNA by RT-PCR reveals that the mRNA is correctly spliced and transcribed into the T-DNA insert (not shown). This suggests that the T-DNA insertion is more likely to act at the MEF11-2 protein level. The T-DNA insertion removes 26 amino acids from the C-terminus and adds 11 different amino acids instead (data not shown). This change may still allow some editing at site ccb203-344, but not at the other two sites. The precise explanation for this observation remains to be determined.

Functional requirement for MEF11-mediated RNA editing in plant mitochondria

Based on its being evolutionarily conserved in all dicot plants, the RNA editing event at nucleotide cox3-422 is expected to be required for a functional COX3 protein. The cox3 gene is one of the best conserved protein coding genes in the mitochondrial genomes of such diverse organisms as humans, plants, fungi and protozoa.

However, the absence of this editing event in rice suggests that, in this monocot, complex IV is functionally fully competent even with the amino acid in the COX3 subunit that is specified by the unedited genomic sequence (Figure 6b). On the other hand, conserved upstream editing events in the cox3 mRNA are likewise not annotated in the databases for rice, and only one editing site further downstream is recorded in this open reading frame in this plant. Either these editing events were missed in the analysis in rice or they are not necessary for COX3 function in rice. Altered protein–protein interactions in complex IV may be responsible for this altered requirement in the monocot rice.

Alternatively, or in addition, the RNA editing event at ccb203-344 may cause a disturbance in assembly and transport of the haem group. A compromised supply of haem may also influence the efficiency of the respiratory chain, for example through cytochrome b, and may thus trigger a signal towards the MVA pathway system.

The observation of slightly slowed growth and the modified tolerance against lovastatin shows that these editing events do have functional consequences for mitochondrial functions. The phenotypic consequences connect these RNA editing events in plant mitochondria with their physiological and biochemical performance. The MEF11 mutant proteins may thus provide interesting material to study mitochondrial functions in plants and the integration of this organelle into the cellular metabolism.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Plant material and preparation of nucleic acids

Arabidopsis thaliana seeds of the Landsberg erecta (Ler) ecotype were a kind gift from J. Forner and S. Binder (Molekulare Botanik, Universität Ulm, Germany). The EMS mutant population of Arabidopsis thaliana ecotype Col was obtained from Lehle Seeds (http://www.arabidopsis.com). The methods used for growth of the Arabidopsis thaliana plants and preparation of DNA or RNA from the leaves were as described previously (Takenaka and Brennicke, 2007). For tests of lovastatin tolerance, seed were sown onto 1% agar in a Murashige–Skoog medium in large Petri dishes. Germination and growth took place under 16 h light/8 h dark cycles in growth chambers. Dishes were kept in almost vertical positions to allow the roots to develop and extend. After 2–4 weeks, plant and root development were documented.

SNaPshot assays and mutant analysis

The EMS mutant library was screened by multiplexed single-base extension (Takenaka and Brennicke, 2009) for plants with altered RNA editing at specific sites. Plants were first analysed in pools of ten, from which deviant plants were recovered. In the identified individual plants, the compromised RNA editing phenotype was verified by cDNA sequence analysis of the status of the investigated editing site. In the mef11-1 mutant and the plants of ecotype Col, the six candidate PPR genes in the 1.0 Mb window on chromosome 4 were investigated by sequence analysis of the relevant PCR products. Most sequences were obtained commercially from 4base lab (http://www.baselab.de) or Macrogen (http://www.macrogen.com).

Analysis of RNA editing sites

Specific cDNA fragments were generated by RT-PCR amplification using established protocols (Takenaka and Brennicke, 2007). The cDNA sequences were compared for C[RIGHTWARDS ARROW]T differences resulting from RNA editing. For initial rapid screening of large numbers of samples, RT-PCR was initiated from primers specific for either the edited or the unedited nucleotide (Figure 2).

Screening assays with C- or T-specific RT-PCR

The 96 F2 generation plants obtained from the cross between a mef11-1 mutant plant (with a Col background) and Ler were screened for homozygous mutants by RT-PCR. The downstream primers were designed with a single nucleotide mismatch and either a C- or T-specific 3′ terminal nucleotide (Figure 2). The mismatch yielded optimal effects when positioned three nucleotides upstream of the 3′ end. Generally, the products obtained after PCR on a previously generated gene-specific longer cDNA strand yielded predominantly reciprocal products depending on the RNA editing status of this site (Figure 2). Although temperatures were carefully adjusted to the correct melting temperatures, non-specific products were repeatedly observed, so some RNA preparations and assays had to be repeated. Generally, however, as visible in the gel in Figure 2, determination of whether or not RNA editing had occurred was possible even in the presence of cross-contaminating background signals.

Protoplast complementation assays

Protoplasts were prepared from 3–4-week-old plantlets as described by Yoo et al. (2007). Transfected genes, including GFP as control and the mutant or wild-type Col MEF11 reading frames, were expressed from the 35S promoter in the cloning site of vector pSMGFP4. The efficiency of transfection was monitored based on signals from separately introduced or co-transfected GFP genes in the cytoplasm. Typically, GFP fluorescence was detected in more than 80% of the transfected protoplasts. Total RNA was isolated after 20–24 h incubation at room temperature. Sequences of cDNAs were determined after RT-PCR with specific primers. RNA editing levels were estimated based on the relative heights of the various nucleotide peaks in the sequence analyses (Zehrmann et al., 2009).

Generation of MEF11 mutant clones

To induce a very low level of single nucleotide alterations in the MEF11 open reading frame, amplification by PCR between two cloning primers was performed using GoTaq polymerase (Promega, http://www.promega.com/). For faithful PCR and cloning, the enzyme Phusion (Finnzymes, http://www.finnzymes.com) was employed. The resulting fragments of the entire open reading frame were cloned into the in planta expression vector pSMGFP4 with a translational stop codon after the MEF11 reading frame. Individual Escherichia coli clones were sequenced, and mutants with altered amino acids were selected for transfection into protoplasts.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We thank Dagmar Pruchner for excellent experimental help. We are very grateful to Stefan Binder, Joachim Forner, Angela Hölzle and Christian Jonietz (all at Molekulare Botanik, Universität Ulm, Germany) for their gifts of seeds, markers and other materials. This work was supported by grants from the Deutsche Forschungsgemeinschaft to M.T. and A.B.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References