OTP70 is a pentatricopeptide repeat protein of the E subgroup involved in splicing of the plastid transcript rpoC1

Authors

  • Anne-Laure Chateigner-Boutin,

    1. Australian Research Council Centre of Excellence in Plant Energy Biology, University of Western Australia, Crawley 6009 WA, Australia
    Search for more papers by this author
    • Present address: UR1268 Biopolymères, Interactions, Assemblages, INRA, F-44316 Nantes, France.

  • Catherine Colas des Francs-Small,

    1. Australian Research Council Centre of Excellence in Plant Energy Biology, University of Western Australia, Crawley 6009 WA, Australia
    Search for more papers by this author
  • Etienne Delannoy,

    1. Australian Research Council Centre of Excellence in Plant Energy Biology, University of Western Australia, Crawley 6009 WA, Australia
    Search for more papers by this author
    • Present address: URGV (Unité de Recherche en Génomique Végétale), UMR INRA 1165 – Université UEVE (Evry), 2, rue Gaston Crémieux – CP 5708, 91057 Évry Cedex, France.

  • Sabine Kahlau,

    1. Australian Research Council Centre of Excellence in Plant Energy Biology, University of Western Australia, Crawley 6009 WA, Australia
    Search for more papers by this author
  • Sandra K. Tanz,

    1. Australian Research Council Centre of Excellence in Plant Energy Biology, University of Western Australia, Crawley 6009 WA, Australia
    2. Centre of Excellence in Computational Systems Biology, University of Western Australia, Crawley 6009 WA, Australia
    Search for more papers by this author
  • Andéol Falcon de Longevialle,

    1. Australian Research Council Centre of Excellence in Plant Energy Biology, University of Western Australia, Crawley 6009 WA, Australia
    Search for more papers by this author
    • Present address: URGV (Unité de Recherche en Génomique Végétale), UMR INRA 1165 – Université UEVE (Evry), 2, rue Gaston Crémieux – CP 5708, 91057 Évry Cedex, France.

  • Sota Fujii,

    1. Centre of Excellence in Computational Systems Biology, University of Western Australia, Crawley 6009 WA, Australia
    Search for more papers by this author
  • Ian Small

    Corresponding author
    1. Australian Research Council Centre of Excellence in Plant Energy Biology, University of Western Australia, Crawley 6009 WA, Australia
    2. Centre of Excellence in Computational Systems Biology, University of Western Australia, Crawley 6009 WA, Australia
    Search for more papers by this author

(fax +61 8 6488 4401; e-mail ian.small@uwa.edu.au).

Summary

Over 20 proteins of the pentatricopeptide repeat (PPR) family have been demonstrated to be involved in RNA editing in plant mitochondria and chloroplasts. All of these editing factors contain a so-called ‘E’ domain that has been shown to be essential for editing to occur. The presumption has been that this domain recruits the (unknown) editing enzyme to the RNA. In this report, we show that not all putative E-class PPR proteins are directly involved in RNA editing. Disruption of the OTP70 gene leads to a strong defect in splicing of the plastid transcript rpoC1, leading to a virescent phenotype. The mutant has a chloroplast transcript pattern characteristic of a reduction in plastid-encoded RNA polymerase activity. The E domain of OTP70 is not required for splicing, and can be deleted or replaced by the E domain from the known editing factor CRR4 without loss of rpoC1 splicing. Furthermore, the E domain of OTP70 is incapable of inducing RNA editing when fused to the RNA binding domain of CRR4. We conclude that the truncated E domain of OTP70 is no longer functional in RNA editing, and that the protein has acquired a new function in promoting RNA splicing.

Introduction

Mitochondria and plastids are compartments within the cell that possess their own genome, a remnant of their bacterial origin. These organelles carry out essential processes for plant cells, including respiration and photosynthesis. Only a few organelle proteins are encoded by their genome, the majority being synthesized from cytosolic transcripts and post-translationally imported. Protein complexes involved in organelle functions generally consist of both organelle and nucleus-encoded subunits. Tight regulation between nuclear and organelle gene expression is thus needed to ensure the correct stoichiometry of these subunits.

Organelle genes are transcribed by two types of RNA polymerase: nuclear-encoded polymerases (NEPs) related to bacteriophage polymerases with three members in Arabidopsis (RPOTP, RPOTM and RPOTMP) or the plastid-encoded polymerase (PEP) encoded by the plastid genes rpoA, rpoB, rpoC1 and rpoC2. Whereas mitochondrial genes are only transcribed by RPOTM or RPOTMP, plastid genes are potentially transcribed either by RPOTP or PEP, as transcripts of all genes were detected in PEP mutants, such as tobacco plants lacking rpoB (Allison et al., 1996) or the Arabidopsis PEP-deficient mutants ptac2, clb19 and ys1 (Pfalz et al., 2006; Chateigner-Boutin et al., 2008; Zhou et al., 2009). In these mutants, the expression of photosynthesis-related genes usually preferentially transcribed by PEP is drastically reduced, whereas transcripts usually preferentially produced by RPOTP are overexpressed (Hajdukiewicz et al., 1997).

Organelle transcripts are mono- or polycistronic. Long polycistronic transcripts are processed by endonucleolytic cleavage and 5′- and 3′-end maturation before they can be translated. In addition, many transcripts undergo RNA splicing and/or RNA editing. RNA splicing is required for many organellar transcripts in plants, although the ancestral prokaryotic genes were intronless. Introns are either cis-spliced (both exons are present in the same primary transcript) or trans-spliced (when exons are encoded in different primary transcripts). Some mitochondrial transcripts such as Arabidopsis mitochondrial nad5 or plastid rps12 require both types of splicing events (Falcon de Longevialle et al., 2010).

RNA editing is a special step in plant organelle gene expression, which modifies the transcript sequence such that it differs from that encoded in the organelle genome. Nucleotides are altered at specific sites in the transcripts, named editing sites. In angiosperms, editing consists mainly of cytidine to uridine (C → U) substitutions in mRNA. In Arabidopsis thaliana mitochondria 526 editing sites have been reported (Giege and Brennicke, 1999; Bentolila et al., 2005, 2008; Falcon de Longevialle et al., 2007; Zehrmann et al., 2009), whereas in chloroplast 34 cytidines are targets of editing (Chateigner-Boutin and Small, 2007). Editing is an essential process, as some mutations that lead to a loss of the ability to edit one or a few sites lead seedling lethality (e.g. clb19; Chateigner-Boutin et al., 2008).

All steps of organelle gene expression involve members of a numerous family of RNA binding proteins, the pentatricopeptide repeat (PPR) proteins. These proteins are defined by degenerate motifs of 35 amino acids repeated in tandem (Small and Peeters, 2000). Plant genomes encode 400–600 PPR proteins that, with a few exceptions, are found in mitochondria or plastids (Lurin et al., 2004; O’Toole et al., 2008). Many mutants disrupted in members of the PPR family have been isolated by forward or reverse genetics. Contrary to many multigene families, there appears to be little redundancy in the molecular functions of PPR proteins (reviewed in Schmitz-Linneweber and Small, 2008). This is thought to result from the high degree of sequence specificity in their RNA binding activities (Delannoy et al., 2007).

PPR proteins can be divided into four subclasses based on their C-terminal domains and the presence of longer (L) or shorter (S) variant PPR motifs within the tandem arrays of the canonical PPR motifs (O’Toole et al., 2008). The most numerous class (P class) consists uniquely of classical PPR motifs, and has been implicated in splicing, the determination of 5′ or 3′ termini of organellar RNAs, and transcript stability in land plants and in other eukaryotes. Some act as fertility restorer genes for cytoplasmic male sterility (reviewed in Schmitz-Linneweber and Small, 2008).

The other three subclasses of PPR proteins are only found in land plants (Salone et al., 2007; O’Toole et al., 2008). The E (for ‘extended’) and DYW (named after the conserved last three amino acids of the domain, Asp-Tyr-Trp) subclasses include additional C-terminal domains. So far all E-class PPR proteins the functions of which are known have been implicated in RNA editing (Kotera et al., 2005; Okuda et al., 2007; Chateigner-Boutin et al., 2008; Takenaka and Brennicke, 2009; Takenaka et al., 2010). Many DYW-class PPR proteins have also been identified as RNA editing factors (Zhou et al., 2009; Cai et al., 2009; Hammani et al., 2009; Okuda et al., 2009, 2010; Robbins et al., 2009; Tang et al., 2010; Zehrmann et al., 2009; Tasaki et al., 2010; Verbitskiy et al., 2010a,b), but one is associated with RNA cleavage (Hashimoto et al., 2003).

As most PPR proteins contain only PPR motifs, the PPR motif was proposed to be the domain of RNA binding. Actual binding of individual PPR proteins to their target RNA has been experimentally verified for several of them (reviewed in Delannoy et al., 2007). Cis-elements in the target RNA have been experimentally defined for the P-class PPR10 (Pfalz et al., 2009), the E-class PPR CRR4 (Okuda et al., 2006) and the DYW-class PPR PpPPR_71 (Tasaki et al., 2010). The mechanism of interaction and specificity between PPR proteins and RNA remains poorly understood in the absence of structural data.

The roles of the additional C-terminal domains are being investigated. The E domains of two plastid PPR proteins were swapped with no impact on the editing of their target (Okuda et al., 2007), proving that this domain is not involved in determining target specificity. Deletion assays showed that the E domain is, however, essential for editing (Okuda et al., 2007, 2009). It was proposed to be a protein–protein interaction domain that recruits the still unidentified editing enzyme.

In this report we have analyzed an E-class PPR gene by reverse genetics. The organelle transcript processing 70 mutant (otp70) shows a typical PEP-deficient phenotype and altered editing of the rpoC1 transcript. However, we show that the molecular function of OTP70 is more probably involved in splicing of rpoC1 transcripts than in editing: the editing phenotype is an indirect effect of the altered splicing patterns. Furthermore, the E domain of OTP70 is not required for its function in either splicing or editing.

Results

Isolation of a pigment-defective mutant by reverse genetics

We are screening a collection of T-DNA insertion lines disrupted in genes encoding members of the PPR protein family (Falcon de Longevialle et al., 2007, 2008; Chateigner-Boutin et al., 2008; Hammani et al., 2009) for defects in organelle gene expression. One line, Salk_090845, gave pale-yellow seedlings, indicative of a plastid defect (Figure 1). The insertion in this line was mapped to the locus At4g25270. Sequencing of the flanking region confirmed the insertion in At4g25270 at nucleotide 12938168. To confirm that the pigment deficiency was caused by the insertion in At4g25270, a wild-type copy of the gene was introduced into the mutant, and resulted in the restoration of normal green pigmentation in young seedlings (Figure 1). Salk_090845 was named organelle transcript processing 70-1 (otp70-1). Analysis of transcripts in the mutant and complemented lines confirmed that otp70-1 is likely to be a null mutant (Figure S1).

Figure 1.

 Visible phenotype of otp70 mutants.
Homozygous otp70-1 seedlings (a) show a marked growth delay with respect to wild type (Wt), and exhibit short, rounded leaves and distinctly paler leaves. The cotyledons are yellow, but the true leaves turn pale green. Mature otp70 plants eventually produce flowers and fertile seeds. Mutant otp70 plants complemented with the OTP70 gene are indistinguishable from the wild type (b). The growth phenotypes of otp70 mutants are intermediate between those of the plastid-encoded polymerase (PEP)-defective mutants clb19 and ys1 (c).

A second line, Salk_013177, with an insertion described as in the ‘promoter’ of At4g25270, was analyzed. The insertion was verified and homozygous plants were obtained, but they did not show a pigment defect, so this line was not analyzed further.

OTP70 is an E-class PPR protein targeted to plastids

The gene disrupted in otp70-1 encodes an E-subclass PPR protein. It is intron-less like many other PPR genes, encodes 10 PPR motifs and an E domain (Figure 2). OTP70 includes a relatively long N-terminal region with no predicted PPR motifs. Part of this sequence is predicted by targetp (Emanuelsson et al., 2000) to encode a plastid signal peptide with a cleavage site located after 43 amino acids. A subcellular localization assay conducted in Arabidopsis cells transformed with OTP70 fused to the fluorescent protein GFP confirmed the predicted plastid localization (Figure S2).

Figure 2.

 Gene model and T-DNA insertion in OTP70 (At4g25270).
Two insertion lines were grown and genotyped: the T-DNA insertion in Salk_013177 is upstream of the At4g25270 locus, and homozygous plants do not show an obvious phenotype; the T-DNA insertion in Salk_090845 (otp70-1) is in the third pentatricopeptide (PPR) motif of the gene. The various motifs are identified and labeled as in Lurin et al. (2004).

OTP70 is not defective in editing the 34 known plastid sites

As all of the characterized E-class PPR proteins so far are involved in editing, we screened the 34 editing sites known in Arabidopsis plastid transcripts using a high-resolution melting (HRM) assay (Chateigner-Boutin and Small, 2007): all of the sites were found to be edited (Figure S3). This screen is not quantitative and thus does not exclude alterations in the level of editing at any site, but does exclude that editing is absent at any of the known sites.

OTP70 is a PEP-defective mutant

The phenotype of otp70-1 was further characterized to help in determining the molecular defect responsible for the lack of pigment. Seed germination was normal on soil, but seedling cotyledons were pale yellow (Figure 1). Seedling growth was delayed compared with wild type, but even in long-day conditions there was no obvious seedling lethality. True leaves developed and greening eventually occurred (Figure 1). Homozygous otp70-1 plants developed slowly, but eventually flowered and gave fertile seeds. We compared the otp70-1 phenotype with the phenotype of other pigment-defective mutants with disrupted PPR genes: clb19 (Chateigner-Boutin et al., 2008), ptac2 (Pfalz et al., 2006), otp51 (Falcon de Longevialle et al., 2008) and ys1 (Zhou et al., 2009) (Figure 1c and data not shown). clb19, ptac2 and otp51 mutants can germinate on soil, but die shortly afterwards. When sown on medium supplemented with sucrose they give white seedlings. In reduced light conditions, clb19 but not otp51 develops pale yellow-green leaves, and can be transferred to soil at a later stage to give mature plants that flower and produce fertile seeds. ys1 can germinate and grow on soil and give yellow seedlings that eventually turn green. The otp70 phenotype appears less severe than that of clb19, but more severe than that of ys1 (Figure 1c). It shares with these two mutants the fact that young seedlings are pigment-defective, but older plants are green.

As PPR proteins are involved in RNA processing, we analyzed the levels of plastid transcripts in otp70 mutants using a quantitative RT-PCR assay (Chateigner-Boutin et al., 2008; Zhou et al., 2009; Okuda et al., 2009). As shown in Figure 3, many transcripts are differentially represented in the wild type and in otp70. The levels of transcripts known to be preferentially transcribed by the PEP (rbcL, and most of the pet, psb and psa transcripts) are reduced in otp70, whereas transcripts known to be transcribed by the NEP (accD, rps genes, rpl genes, rpoA and rpoB) are increased in otp70. The transcript profile is similar to that of the known PEP-defective mutants clb19 and ptac2 (Figure 3). A hierarchical clustering of the plastid transcript levels of several organelle transcript processing mutants groups otp70 with PEP-defective lines (Figure S4). An obvious difference in these patterns can, however, be observed concerning the transcript rpoC1, which is increased over wild-type levels in clb19 and ptac2, but is apparently reduced below wild-type levels in otp70.

Figure 3.

 Plastid transcript levels in otp70 seedlings compared with those of clb19 and ptac2.
Genome-wide quantitative RT-PCR measurements were made of chloroplast transcripts from young leaves of the three mutants (shown here as the log2 ratios with respect to the levels in the wild type). A major difference is seen for rpoC1 (boxed); black, otp70; mid-gray, clb19; pale gray, ptac2. The values are means of three biological replicates (error bars indicate SDs). The primers used for each transcript are listed in Table S1.

otp70 is defective in the splicing of rpoC1

rpoC1 is preferentially transcribed by the NEP, as are the other genes encoding subunits of the PEP: rpoA, rpoB and rpoC2. As shown in Figure 3, rpoC1 transcripts are apparently reduced in otp70, although transcripts of the other rpo genes are increased. Northern blots were performed and hybridized with an antisense RNA probe complementary to part of rpoC1, revealing that rpoC1 is in fact over-expressed in otp70 compared with wild type, as in the PEP mutant clb19, but that the size of the major transcript is larger in otp70 compared with Wt and clb19, matching the expected size of the unspliced transcript (Figure 4).

Figure 4.

rpoC1 shows a splicing defect in otp70.
(a) RNA from young seedlings of Col-0 (Wt), otp70 and clb19 run on a gel, and blotted and stained with methylene blue as an indicator of equivalent loading.
(b) Membrane hybridized with a probe complementary to part of the second exon of rpoC1 (indicated in c), showing overexpression of rpoC1 in otp70 and clb19. The rpoC1 transcript in otp70 migrates slower than in the Wt and in clb19. The intron in rpoC1 is 791 nt.
(c) Structure of the rpoC1 gene indicating the two exons as gray boxes, the probe used in (b) and the primers used in (d). The location of the rpoC1(21806) editing site is shown by C.
(d) RT-PCR of rpoC1 transcripts from Wt, otp70 and complemented otp70 plants. The expected size of the product from spliced transcripts is 447 bp, and that from unspliced transcripts is 1238 bp. Molecular weight markers (M) are shown to the left, and the amplification product from total DNA (gDNA) is given in the centre. The right side of the panel shows a lack of amplification products from the RNA templates used for reverse transcription, indicating that the 1238-bp product from the cDNA sample is from unspliced transcript, and is not contaminating genomic DNA; 0, control with no template.

RT-PCR using primers flanking the intron (Figure 4c) confirmed a splicing defect in otp70 seedlings, and this defect is complemented in mutants transformed with a wild-type copy of OTP70 (Figure 4d). The primers used for the quantitative RT-PCR (Figure 3) also flank the intron, so the defect in splicing explains the discrepancy between the northern hybridization and the quantitative RT-PCR results.

To confirm that the splicing defect was specific to rpoC1, we quantitatively analyzed the splicing of all protein-encoding genes in plastids by quantitative RT-PCR. We compared the results with the splicing of transcripts in the PEP mutant clb19. As shown in Figure 5, PEP mutants such as clb19 exhibit characteristic splicing defects, especially in PEP-transcribed genes such as ndhA, petB and petD. otp70 showed the same defects as clb19, but in addition, a strong and specific defect in the splicing of rpoC1.

Figure 5.

 Quantitative analysis of plastid transcript splicing in otp70 and clb19 compared with wild type (WT).
Quantitative RT-PCR was carried out using primers flanking each plastid intron, and also with primer pairs placed across splice sites on unspliced transcripts. The ratio of the quantities of each pair of amplification products indicates the extent of splicing for each intron. The histogram shows log2 ratios of the splicing ratios in the mutants compared with the WT (i.e. defects in splicing correspond to negative values). Splicing defects are similar in otp70 (black) and clb19 (gray), with the exception of the rpoC1 intron (boxed), the splicing of which is specifically affected in otp70. The values are means of three biological replicates (error bars indicate SDs).

All the previously characterized E-class PPR proteins are editing factors, and there are known examples of correlation between editing and splicing (Castandet et al., 2010). Therefore, we carefully screened the rpoC1 transcript (exons and intron) in Col-0 (wild type) for previously unknown editing sites that could be defective in otp70, and could thereby prevent correct splicing. We sequenced RT-PCR products and carried out exhaustive HRM analysis. Only the editing site rpoC1(21806) that we identified using the same HRM scanning method in a previous work was detected (Chateigner-Boutin and Small, 2007).

The E domain of OTP70 is not required for the complementation of the pigment defect

We transformed otp70-1 mutants with modified OTP70 genes coding for a truncated version of OTP70 lacking an E domain and a chimaeric OTP70, in which the E domain was exchanged with the E domain of CRR4 (Figure 6a). In both cases, transformants grew normally and resembled the wild type and plants complemented with the unmodified OTP70 gene (Figure 6b). Therefore, only the PPR part of OTP70 is required for the complementation of the pigment defect, and the E domain of CRR4 does not prevent the complementation. As would be expected from the restoration of wild-type pigment and growth phenotypes, the complemented plants have much reduced levels of unspliced rpoC1 transcripts (Figure 6c), partly reflecting the re-establishment of near wild-type levels of rpoC1 splicing (Figure 6d), and probably also partly reflecting the re-establishment of near wild-type levels of PEP and NEP activities.

Figure 6.

 Macroscopic and molecular phenotype of complemented otp70 mutants.
(a) Constructs expressing full-length OTP70 (transformants FL1, FL2 etc.), OTP70 lacking its E domain (transformants P1, P2 etc.) or OTP70 with its E domain replaced by the E domain of CRR4 (transformants E1, E2 etc.) were used to transform otp70.
(b) All three constructs gave rise to apparently wild-type plants lacking the growth and pigment defects of otp70.
(c) Levels of unspliced rpoC1 transcripts were restored to near normal levels in the transformants. Values are normalized against the wild type (Wt), and are means of three technical replicates; errors are SDs.
(d) Ratios of unspliced to spliced rpoC1 transcripts were restored to near normal levels in the transformants. Values are normalized against the wild type (Wt), and are means of three technical replicates; errors are SDs.

Editing of rpoC1 is influenced by splicing

We verified the level of rpoC1(21806) editing in the various mutant and complemented lines, analyzing unspliced and spliced transcripts separately. The editing of unspliced rpoC1 transcripts in the wild type is relatively low, but the same transcripts in otp70 are almost completely edited (Figure 7). Lines complemented with the wild-type protein show editing at near wild-type levels, but the other complemented lines have levels of editing intermediate between that of the wild type and otp70 plants. The results are consistent with editing being determined by the kinetics of splicing: the longer each rpoC1 transcript takes to be spliced, the more likely it is to be edited before splicing occurs (Figure 7). The inverse relationship between rpoC1 editing and the presence of OTP70 make it extremely unlikely that OTP70 plays any direct role in the editing of rpoC1.

Figure 7.

 Editing of rpoC1 is increased in otp70.
Editing of the rpoC1(21806) site in RNA from leaves was quantified using poisoned primer extension using ddGTP to stop extension at the first C encountered. The primers used to amplify reverse-transcribed RNA to form the template were specific for unspliced transcripts. Data for spliced transcripts are given in Figure S4. C indicates the band obtained from unedited template; U indicates the band from the edited template. H2O, no template; edited, cloned edited control template; unedited, cloned unedited control template; other lanes are labeled as in Figure 6.

Spliced rpoC1 transcripts are generally more edited than the unspliced transcripts (Figure S5).

The E domain of OTP70 is not functional in editing

To confirm that the E domain of OTP70 cannot function in editing, we swapped the E domain of CRR4 with the E domain of OTP70. CRR4 is an E-class editing factor required for the editing of ndhD(117161). crr4 mutants were transformed with a wild-type CRR4 gene or with chimaeric CRR4/OTP70. Editing of the target site of CRR4 was monitored in the transformed plants. Editing was restored only with the wild-type CRR4, confirming that the E domain of OTP70 is probably not functional for editing (Figure 8).

Figure 8.

 The E domain of OTP70 cannot replace that of CRR4.
(a) Constructs expressing full-length CRR4 or CRR4 with its E domain replaced by the E domain of OTP70 (E70) were used to transform crr4.
(b) Poisoned primer extension gel quantifying editing of the ndhD(117161) site. Edited, cloned edited control template; unedited, cloned unedited control template; Wt, wild-type; crr4::CRR4, crr4 mutant expressing a wild-type CRR4 transgene (results from two independent transformants are shown); crr4::E70, crr4 mutant expressing the E70 transgene (results from two independent transformants are shown).

The E domain of CRR4 is not sufficient to turn OTP70 into an editing factor

We sequenced RT-PCR products corresponding to the unspliced rpoC1 transcript in the lines transformed with OTP70 fused to the E domain of CRR4 with the idea that adding a functional E domain to OTP70 could turn it into an editing factor, thus perhaps revealing an OTP70 binding site. However, only the known editing site rpoC1(21806) was detected.

Discussion

otp70 is a PEP-deficient mutant

We describe here a pigment-defective (virescent) mutant that exhibits a strong reduction in RNAs transcribed by PEP as a result of a strong defect in rpoC1 splicing, thus presumably decreasing the level of available RPOC1 protein, which is the beta’ subunit of the PEP complex. The phenotype of the mutant is intermediate between that of two other Arabidopsis mutants defective in the expression of rpoA (clb19) or rpoB (ys1). No complete knock-outs of PEP have been obtained in Arabidopsis, contrary to Nicotiana tabacum (tobacco), where they can be constructed by the deliberate deletion of the plastid genes. Tobacco PEP mutants, where either rpoA or rpoB were deleted, show similar defects in transcript pattern and accumulation to the Arabidopsis mutants. Contrary to the Arabidopsis mutants, these defects do not revert in older plants, this difference probably reflects a residual low level of PEP activity in the Arabidopsis mutants rather than any species-specific roles for the PEP enzyme. Post-germinative growth and greening are very demanding for PEP activity, not only for the transcription of protein-coding genes required for photosynthesis, but also to produce the tRNAGlu required to make aminolevulinic acid (ALA), a precursor for chlorophyll synthesis (Zhou et al., 2009).

OTP70 is involved in splicing

The primary defect in otp70 mutants is a near-total lack of splicing of rpoC1 transcripts in young seedlings. OTP70 thus joins a rapidly lengthening list of protein factors implicated in RNA splicing in land plant organelles (reviewed in Falcon de Longevialle et al., 2010). In angiosperms, one plastid-encoded and fourteen nuclearly encoded plastid splicing factors have been identified, each involved in splicing a different subset of introns. The plastid-encoded maturase K binds seven introns in tobacco (in the trnV, trnI, trnA, trnK, rpl2, rps12 and atpF transcripts; Zoschke et al., 2010). Other plastid RNA-binding proteins implicated in splicing include chloroplast RNA splicing and ribosome maturation (CRM) proteins (Barkan et al., 2007), plant-specific RNA recognition (PORR) proteins (Kroeger et al., 2009) and PPR proteins (Arabidopsis OTP43, OTP51 and ABO5, and maize PPR4 and PPR5) (Schmitz-Linneweber et al., 2006; Falcon de Longevialle et al., 2007, 2008; Beick et al., 2008; Koprivova et al., 2010; Liu et al., 2010).

In some cases, splicing in mutants lacking these proteins is totally impaired, such as for the splicing of intron 2 of ycf3 in Arabidopsis otp51 mutants (Falcon de Longevialle et al., 2008); for others, the defect is only partial. The mechanisms of the action of PPR proteins in splicing are poorly understood. PPR4 and OTP51 contain additional RNA-binding domains that could have a role in the splicing process, the RRM domain and the LAGLIDAGD endonuclease domain, respectively (Schmitz-Linneweber et al., 2006; Falcon de Longevialle et al., 2008), but most PPR proteins implicated in splicing are indistinguishable from those involved in other transcript processing activities. None of the protein complexes in which splicing factors have been identified (Kroeger et al., 2009) have been shown to include PPR proteins, and therefore it is likely that PPR splicing factors contact the RNA directly to influence splicing by stabilizing the unspliced precursor, inducing folding into a splicing-competent state, or by recruiting the ‘spliceosome’ (Falcon de Longevialle et al., 2010).

CAF1 and CRS2 are also involved in the splicing of the rpoC1 intron and intron 1 of clpP (Asakura and Barkan, 2006). The described macroscopic phenotype of seedlings with a T-DNA insertion disrupting CAF1 resembles that of otp70 and clb19 mutants, respectively defective in rpoC1 splicing and clpP (and rpoA) editing. otp70 mutants are only slightly defective in the splicing of clpP, suggesting it is unlikely to be an obligate component of any complex containing CAF1 or CRS2. Most of the splicing defects in otp70 (besides the strong defect in rpoC1) are likely to be secondary effects of the reduction in PEP activity, as they are seen in other PEP-deficient mutants.

OTP70 has a truncated E domain and does not function as an RNA editing factor

Previously described E-class PPR proteins have all been directly implicated in RNA editing. Our results clearly show that OTP70 is not functional in editing, and that its E domain is not capable of inducing editing, even when transferred to another known editing factor. Instead, OTP70 is required for the efficient splicing of rpoC1, and this function does not require its E domain.

The E domain of OTP70 contains 65 amino acids. When it is compared with the E domain of other E-type PPR proteins for which a molecular function is known, it is obvious that its E domain is truncated (Figure 9). MEF9, MEF18, MEF19 and MEF20 are mitochondrial E-class PPR proteins with shorter than typical E domains, but all are functional in the editing of mitochondrial transcripts (Takenaka, 2010; Takenaka et al., 2010). The size difference between OTP70 and the MEF E domains is approximately 10 amino acids. These 10 amino acids may be essential for the editing function of the E domain, which has been proposed to recruit the editing activity (Chateigner-Boutin and Small, 2010). In the Arabidopsis nuclear genome, there are other E-type PPR genes with a truncated E domain, but only one, At2g39620, has a shorter E domain than OTP70.

Figure 9.

 Alignment of E domains from known editing factors, OTP70 and putative orthologs.
Putative orthologs of OTP70 were identified from genome sequences at Phytozome (http://www.phytozome.net). E domains were identified using hmmsearch, and displayed with jalview (Waterhouse et al., 2009). The sequences have been sorted by length, and the depth of shading indicates the degree of amino acid identity.

An ortholog of OTP70 is present in cereals, although there is no intron in rpoC1

A global phylogenetic analysis of PPR genes conducted in Arabidopsis and rice revealed that PPR genes are unusually well-conserved in terms of gene number, with few losses or gains in either lineage (O’Toole et al., 2008). This conservation has been useful: the lack of identifiable orthologs of the Arabidopsis PPR protein RARE1 in rice was a criteria for its selection as a candidate for the editing factor recognizing the Arabidopsis site within the coding sequence of accD, which is not edited in rice (Robbins et al., 2009).

The prevalence of E-class proteins amongst identified editing factors and the previous lack of E-class proteins amongst identified splicing factors suggest that OTP70 is derived from the truncation of an ancestral editing factor. We examined other sequenced plant genomes for potential orthologs to see whether any would include extant examples of this putative ancestral editing factor. Potential orthologues were identified from several plant species (Figure 9), but all possess truncated E domains, and are unlikely to be editing factors.

The presence of a potential ortholog of OTP70 in the rice genome is unexpected, because there is no intron in rice rpoC1 (Shimada et al., 1990). The rice ortholog of OTP70 may not have the same function or the same target(s) in rice. Such apparent alterations in target are predicted for editing factors. Many (21) C targets of editing in Arabidopsis are already encoded as a T in the rice genome (Robbins et al., 2009). CLB19 is required for editing of sites in Arabidopsis rpoA and clpP transcripts that do not exist in rice, yet there is a clear candidate for a rice ‘ortholog’ (O’Toole et al., 2008; Robbins et al., 2009). Hence, despite the conservation of PPR genes implied by phylogenetic trees (O’Toole et al., 2008), it is becoming obvious that many of these apparent orthologs do not have strictly conserved functions.

RNA editing can be strongly influenced by other RNA-processing activities

As well as the decrease in splicing, otp70 mutants show a striking difference in the editing of rpoC1 transcripts. It is difficult to distinguish cause and effect in these processes. One interpretation could be that OTP70 directly reduces the editing of rpoC1, perhaps by competing with an unidentified editing factor required for recognizing the rpoC1(21806) site, and that this promotes splicing via an effect on RNA conformation. Alternatively, the changes in RNA processing may indirectly alter the level of RNA editing. The editing site in rpoC1 transcripts is sufficiently close to the intron (57 nucleotides) that splicing may affect the binding of the hypothesized editing specificity factor that recognizes the site. The near-complete editing of rpoC1 transcripts in otp70 mutants suggests that the unspliced transcripts might be preferentially edited: rapid splicing in the wild type may prevent rpoC1 transcripts from being fully edited. A third possibility is that editing is simply governed by the half-life of the transcripts, and that rpoC1 transcripts turn over much more slowly in otp70 mutants, and therefore have more time to be edited. Whatever the correct explanation, it is clear from these results that even a ten-fold difference in editing at a specific site between wild type and mutant cannot be used to infer that the missing factor is directly involved in editing. Considering the relatively low levels of editing of rpoC1 transcripts in wild-type seedlings, there is the possibility of two forms of RpoC1 being translated in Arabidopsis leaves.

Experimental procedures

Primers

Unless otherwise stated, the primers used in this work are listed in Table S1.

Plant material

Arabidopsis thaliana ecotypes Columbia (Col-0) and Landsberg erecta (Ler) were used in this work. SALK lines were obtained from the ABRC Stock through The Arabidopsis Information Resource (http://www.arabidopsis.org). Plants were genotyped by PCR for homozygous lines, and the insertion position was confirmed by sequencing with a T-DNA left border primer. otp70-1 is from the line SALK_090845 in Col-0, and contains an insertion in chromosome 2 at position 12 938 168. SALK_013177 carries an insertion in chromosome 2 at position 12 938 934. A homozygous crr4 line (GT-5-38860) from the JIC SM collection in the Ler background was obtained from the John Innes Centre, UK (Sundaresan et al., 1995).

Plants were grown on soil or on half-strength Gamborg medium supplemented with 1% sucrose under long-day conditions (16-h light/8-h dark) at 22°C.

Analysis of RNA editing

A poisoned primer extension of RT-PCR products was performed as previously described (Chateigner-Boutin and Small, 2007). High-resolution melting analysis of amplicons was performed as previously described (Chateigner-Boutin and Small, 2007), using the primers listed in Okuda et al. (2009) and for rpoC1 unspliced transcript using the RT-PCR primers listed in Table S1.

For bulk sequencing of RT-PCR products, total RNA was isolated from 3-day-old seedlings using an RNeasy Plant Mini kit (Qiagen, http://www.qiagen.com) and DNase treated with a DNA-free kit (Ambion, http://www.ambion.com). Reverse transcription was carried out using a Superscript III kit (Invitrogen, http://www.invitrogen.com) from 3 μg of DNA-free RNA from random hexamers. The cDNA was amplified by PCR with the primers listed in Table S1, which were also used for the sequencing reaction. Purification of the RT-PCR products and sequencing were performed by Macrogen Inc. (http://www.macrogen.com). rpoC1 RT-PCR products were also cloned in pGEM-T Easy vector (Promega, http://www.promega.com), and 10 individual clones were sequenced to check for editing.

Analysis of splicing

cDNA prepared as described above was amplified with the primers chloro39F and chloro41R that flank the rpoC1 intron. qRT-PCR analysis of plastid gene splicing was performed as previously described (Falcon de Longevialle et al., 2008).

Analysis of plastid transcript abundance

qRT-PCR analysis of plastid transcript abundance was performed as previously described (Chateigner-Boutin et al., 2008).

Northern blot

Total RNA was isolated using an RNeasy Plant Mini kit (Qiagen). Fifteen micrograms of RNA was loaded on formaldehyde agarose gels and transferred onto Hybond N+ nylon membranes (Amersham Biosciences, now part of GE Healthcare, http://www.gelifesciences.com). RNA integrity, loading and transfer were verified by staining the membrane with methylene blue. Hybridizations were performed under high-stringency conditions using antisense RNA probes internally labeled with biotinylated cytidine. The probes correspond to part of the coding sequence of rpoC1 (RT-PCR with primers chloro39F and 21861R) amplified from cDNA of Col-0. The probes were synthesized by cloning RT-PCR products in pGEM-T Easy vector (Promega). Clones with inserts in antisense orientation were amplified by PCR using the forward primer (21861R) and M13/pUC reverse primer. The PCR product served as a template for in vitro transcription using biotinylated CTP with SP6 polymerase following the manufacturer’s recommendations (Maxiscript kit; Ambion). Pre-hybridization of the membrane was carried out for 1 h in hybridization buffer (5× SSC, 50% v/v formamide, 0.5% SDS and 100 μg ml−1 heparin) at 65°C. Hybridization with RNA probes was carried out in the same buffer overnight at 65°C, followed by three 15-min washes at 22°C in 1× SSC/0.5% SDS and two washes at 60°C in 0.1× SSC/0.1% SDS for 20 min and 1 h, respectively. Signal detection was performed using the Chemiluminescent Nucleic Acid Detection Module (Pierce, http://www.piercenet.com) and read in an ImageQuant-RT ECL (Amersham Biosciences, now part of GE Healthcare).

Analysis of targeting via GFP fusions

The first 300 bp of the coding sequence of otp70 was amplified using the Expand High Fidelity PCR system (Roche, http://www.roche.com), with primers containing the attB sites for Gateway® cloning, according to the manufacturer’s instructions. The PCR product was cloned into a pDON207 vector using Gateway BP clonase enzyme mix (Invitrogen), and was sequenced to check PCR accuracy. This entry clone was used to build a construct to express a fusion between the OTP70 targeting sequence and GFP. Cloning, expression and fluorescence microscopy were carried out as previously described (Koprivova et al., 2010).

Sequence alignments

In total, 187 E-class PPRs were identified from the collection of Arabidopsis PPR genes created in our previous study, using the same definition of the E domain (O’Toole et al., 2008).

Constructs and plant transformation

The complete open reading frame (ORF) of OTP70 and CRR4 was amplified using the Expand High Fidelity PCR system (Roche) with primers containing the attB sites for Gateway® cloning, according to the manufacturer’s instructions. The PCR products were cloned into a pDON207 vector using Gateway BP clonase enzyme mix (Invitrogen), and were sequenced to check PCR accuracy.

Different constructs containing modified OTP70 and CRR4 were obtained by PCR. OTP70 lacking the E domain was generated using a reverse primer stopping at the last nucleotide before the E domain, and containing a stop codon in frame. OTP70 fused to the E domain of CRR4 was generated by amplifying the PPR part of OTP70 with a reverse primer that was complementary to the last 15 nucleotides before the start of the E domain of OTP70, and the first 15 nucleotides of the E domain of CRR4. A second PCR was carried out with a forward primer consisting of the 15 nucleotides before the start of the E domain of OTP70, the first 15 nucleotides of the E domain of CRR4 and the reverse primer used for cloning the complete ORF of CRR4. A third amplification was performed with a dilution of the products of the first two PCRs as the template, and using the OTP70 forward and CRR4 reverse primers that were used to clone the fused ORF. CRR4 with the E domain of OTP70 was generated by the same method with appropriate primers (all primer sequences are given in Table S1).

The constructs were cloned into pGWB2 binary vectors using Gateway LR clonase II enzyme mix (Invitrogen), and were introduced into otp70-1 or crr4 mutants via Agrobacterium tumefaciens GV3101. Transformants were obtained by selection on half-strength Gamborg agar plates containing 25 μg ml−1 hygromycin, and confirmed by PCR. Expression of all transgenes was verified by RT-PCR using primers specific to transgene transcripts.

Accession numbers

Sequence data from this article can be found in the Arabidopsis Genome Initiative or Genbank/EMBL databases under the following accession numbers: OTP70 (At4g25270); CRR4 (At2g45350).

Acknowledgements

This research was supported by the Australian Research Council Centre of Excellence grant CE0561495, the WA State Goverment centres of excellence scheme and the University of Western Australia.

Ancillary