The moss pentatricopeptide repeat protein with a DYW domain is responsible for RNA editing of mitochondrial ccmFc transcript


For correspondence (fax +81 52 789 3080; e-mail


In most land plants RNA editing frequently occurs in many organelle transcripts, but little is known about the molecular mechanisms of the organelle RNA editing process. In this study, we have characterized the Physcomitrella patens PpPPR_71 gene that is required for RNA editing of the ccmFc transcript. This transcript harbors two RNA editing sites, ccmF-1 and ccmF-2, that are separated by 18 nucleotides. Complementary DNA sequence analysis of ccmFc suggested that RNA editing at the ccmF-1 site occurred before ccmF-2 editing. RNA editing of the ccmF-2 downstream site was specifically impaired by disruption of the PpPPR_71 gene that encodes a polypeptide with 17 pentatricopeptide repeat motifs and a C-terminal DYW domain. The recombinant PpPPR_71 protein expressed in Escherichia coli specifically bound to the 46-nucleotide sequence containing the ccmF-2 editing site. The binding affinity of the recombinant PpPPR_71 was strongest when using the edited RNA at ccmF-1. In addition, the DYW domain also binds to the surrounding ccmF-2 editing site. We conclude that PpPPR_71 is an RNA-binding protein that acts as a site recognition factor in mitochondrial RNA editing.


In the most plant organelles, RNA editing frequently occurs in many transcripts. More than 400 RNA editing sites, mostly involving cytidine (C) to uridine (U) transitions, have been identified in the mitochondria (Giegé and Brennicke, 1999; Notsu et al., 2002; Handa, 2003) and 26 to 44 sites in the plastids of seed plants (Kahlau et al., 2006; Zeng et al., 2007). However, little is known about the molecular mechanisms of RNA editing in plant organelles. CRR4 was first identified as a trans-acting factor responsible for RNA editing in Arabidopsis thaliana plastids (Kotera et al., 2005). CRR4 is a member of the pentatricopeptide repeat (PPR) proteins that are characterized by tandem repeats of a degenerate 35-amino-acid motif (Small and Peeters, 2000). The PPR family is one of the large protein families that contain more than 450 members in seed plants (O’Toole et al., 2008; Schmitz-Linneweber and Small, 2008). Most PPR proteins are predicted to localize to either mitochondria or plastids (Lurin et al., 2004).

Several PPR proteins have been demonstrated to be involved in site-specific cleavage, splicing, stability, RNA editing, or translation of targeted organelle transcripts (Delannoy et al., 2007; Saha et al., 2007; Schmitz-Linneweber and Small, 2008). The PPR proteins are structurally divided into P and PLS subfamilies. The P subfamily is composed of canonical PPR motifs only and the PLS subfamily consists of canonical PPR motifs and their variants. The latter is further classified into PLS, E, and DYW subclasses based on characteristic C-terminal motifs (Lurin et al., 2004). The Arabidopsis CRR4, CRR21, and OTP80 are involved in RNA editing for the ndhD-1, ndhD-2, and rpl23 sites, respectively (Kotera et al., 2005; Okuda et al., 2007; Hammani et al., 2009). The Arabidopsis CLB19 was recently shown to be required for editing of two distinct plastid transcripts, rpoA and clpP (Chateigner-Boutin et al., 2008). These three PPR proteins belong to the E subclass PPR protein.

The DYW domain contains a conserved region, which includes invariant residues that match the active site, C/HxExx…xPCxxC, of cytidine deaminases from bacteria, plants, animals, and yeast (Salone et al., 2007). Neither RNA editing nor DYW domains could be identified in algae and one clade of liverworts. There is a correlation between the presence of nuclear DYW genes and organelle RNA editing among embryophytes. Therefore, a hypothesis was proposed in which the DYW domains are responsible for RNA editing in plant organelles and catalyze RNA editing (Salone et al., 2007). Thereafter, Arabidopsis PPR proteins with the C-terminal DYW domain were shown to be required for RNA editing in plastids. CRR22 and CRR28 are responsible for RNA editing of some specific sties of ndhB and ndhD transcripts (Okuda et al., 2009). Five OTP proteins are required for eight editing sites of ndhB, D, F, G, rps12, rps14, and psbZ transcripts (Hammani et al., 2009; Okuda et al., 2010) and YS1 is involved in RNA editing of rpoB (Zhou et al., 2009). Editing of psbF and accD transcripts also requires DYW subclass PPR proteins (Cai et al., 2009; Robbins et al., 2009; Yu et al., 2009). Mitochondrial PPR proteins with a DYW domain have recently been shown to be required for RNA editing at multiple sites in the mitochondrial transcripts in Arabidopsis (Tang et al., 2009; Verbitskiy et al., 2009; Zehrmann et al., 2009) and rice (Kim et al., 2009).

Among 103 PPR genes in the moss Physcomitrella patens, 10 encode the DYW-subclass PPR proteins while E subclass PPR proteins are absent (O’Toole et al., 2008). This raises the question of whether PPR-DYW proteins are responsible for RNA editing in mosses. In P. patens, RNA editing is a rarely occurring event as only two editing sites have been identified in plastids (Sugiura et al., 2003; Miyata and Sugita, 2004) and 11 editing sites in mitochondria (Rüdinger et al., 2009).

The P. patens gametophyte is haploid and is dominant in the life cycle, making it possible to study the phenotype of knockouts directly after transformation for PPR gene targeting (Hattori et al., 2007). In this study, we have shown that the moss PpPPR_71 protein with the DYW domain is required for RNA editing of the ccmF-2 site in the mitochondrial ccmFc transcript encoding a polypeptide involved in cytochrome c maturation.


Disruption of the PpPPR_71 gene

To investigate the function of the DYW-containing PPR proteins, we disrupted and characterized a PpPPR_71 gene (gene model name, estExt_gwp_gw1.C_480142; protein ID, 181369; Rensing et al., 2008) that encodes a polypeptide of 833 amino acids with 17 PPR motifs and the C-terminal DYW domain (O’Toole et al., 2008). Polyethylene glycol-mediated transformation gave 16 geneticin G418-resistant mosses. Among these we isolated one gene-disrupted line, 6-11 (Figure 1a). Biolistic transformation by particle bombardment yielded nine G418-resistant mosses, and one gene-disrupted line 7–9 was selected (Figure 1a). Probing with a PpPPR_71 gene fragment (426 bp) detected the predicted 1.5-kb SstI DNA fragment signal by genomic Southern analysis in the wild-type moss (Figure 1b). By contrast, if the 2.1-kb nptII cassette was inserted in the targeted region, a 3.6-kb signal could be detected in the disruptants. However, two fragments of 5.1 and 3.6 kb were detected (Figure 1b). Polymerase chain reaction (PCR) analysis using pairs of P1 and P2 or P3 and P4 gave expected amplified DNA fragments of 2.9 kb or 1.0 kb, respectively (Figure 1c). When P3 and P5 were used for PCR, a 2.4-kb fragment was amplified. These results indicate that the nptII cassettes were inserted in tandem into the targeted PpPPR_71 gene (Figure 1a). Such a multiple insertion of the introduced sequence often occurs in P. patens (Kamisugi et al., 2006). Furthermore, the absence of PpPPR_71 transcript in the 6–11 and 7–9 transgenic lines was verified by reverse transcription (RT)-PCR (Figure 1d). A 542-bp band was detected in wild-type moss but not in the 6–11 and 7–9 transgenic lines. By contrast, actin gene transcript was detected in both wild-type and PpPPR_71 disruptant mosses. This result clearly indicated that PpPPR_71 transcripts were absent from the 6–11 and 7–9 transgenic mosses, which are probably PpPPR_71 deficient. Thus, the 6–11 and 7–9 transgenic lines are the PpPPR_71 gene disruptant. The disruptants displayed significantly poor growth of the protonemata (Figure 1e).

Figure 1.

 Structure of the PpPPR_71 gene and the plasmid pET-2, and verification of the targeted disruption of PpPPR_71.
(a) The translated PpPPR_71 gene is represented by a white box and the untranslated regions by gray boxes. The nptII gene cassette was inserted into the StuI site. The fragment sizes after SspI digestion of genomic DNA for Southern analyses are shown. Primers and the expected fragment sizes for PCR analysis are also shown.
(b) Total cellular DNA from wild-type (WT) and disrupted (6–11, 7–9) mosses was digested with SspI, and hybridized with the DNA probe.
(c) The PCR analysis using indicated pairs of primers showed that the predicted 2.9, 1.0, and 2.4 kb fragments were amplified derived from the transgenic lines (6–11, 7–9).
(d) Reverse transcription (RT)-PCR for detection of PpPPR_71 transcripts using primers P6 and P7, and total cellular RNA from WT and the disruptants (6–11, 7–9). RT-PCR for actin transcripts was used as a control.
(e) The protonemal colony of the disruptants and the WT grown for 18 days on a BCD medium plate without geneticin G418. Scale bar = 10 mm.

PpPPR_71 protein is localized in the mitochondria

The Predotar and TargetP programs (Emanuelsson et al., 2000; Small et al., 2004) predict that PpPPR_71 is localized in the mitochondria. To confirm this prediction we introduced pET-3 encoding the predicted transit peptide of 95 N-terminal residues of PpPPR_71 fused to green fluorescent protein (GFP) into protonemal cells by particle bombardment. One day after bombardment, GFP fluorescence was observed in the mitochondria-like particles but not in the chloroplasts (Figure 2, 71-GFP). Their subcellular localization looks similar to that of the mitochondrial γATPase–GFP fusion protein (Figure 2, mt-GFP). This indicates that PpPPR_71 is probably a mitochondrial PPR protein.

Figure 2.

 Subcellular localization of GFP fused to the transit peptide of PpPPR_71 protein.
The plasmid pET-3 encoding the N-terminal 95 residues of PpPPR_71 fused to GFP (71-GFP) was transformed into the Physcomitrella patens protonemata. The localization of GFP and chloroplast pigments (chlorophyll) in the transformed cells was detected by fluorescent microscopy. The construct for a mitochondrial γATPase–GFP fusion protein (mt-GFP) was shown as a control for mitochondrial localized protein. Merged images are shown in the bottom panels (Merged). Scale bars = 20 μm.

PpPPR_71 is involved in RNA editing of ccmFc transcript

Prior to investigating the possibility that PpPPR_71 protein is responsible for RNA editing in mitochondria, we analyzed RNA editing in the mitochondria. For this analysis we predicted 27 putative editing sites in Ppatens mitochondria using Gclust Server ( with the Mt23 dataset (Terasawa et al., 2007) (Table S1 in Supporting Information). To identify RNA editing sites experimentally, we prepared and sequenced cDNAs derived from 4-day-old protonemal RNA. Among 27 predicted sites, we identified 11 C-to-U editing sites in 9 gene transcripts by comparison of the mitochondrial gene and cDNA sequences (Table S1). The number of RNA editing sites is identical to that recently reported by Rüdinger et al. (2009).

We then examined RNA editing for 11 sites in the PpPPR_71 disruptants. As shown in Figure 3, RNA editing of the ccmF-2 site of ccmFc transcript was specifically impaired in the disruptants but the ccmF-1 site and the other editing sites were not affected. To examine whether the RNA processing of ccmFc transcripts is impaired in the PpPPR_71 disruptants we performed Northern blot hybridization analysis. The ccmFc transcript profiles were no different in the wild type and the disruptants (Figure 4). This indicates that the fact that there is no RNA editing at the ccmF-2 site in the disruptants is not due to a secondary effect of the ccmFc RNA processing defect.

Figure 3.

 RNA editing defects in the PpPPR_71 disruptants.
Specific genes and cDNAs of Physcomitrella patens were amplified and sequenced directly. The editing sites are marked by arrowheads and the ccmF-2 editing sites are also marked by arrows.

Figure 4.

 Transcript profiles of ccmFc gene in the wild type (WT) and PpPPR_71 disruptants.
Total cellular RNA was extracted from the WT and disruptants (6–11, 7–9) protonemata and subjected to northern blot analysis using a ccmFc-specific probe (364 bp). Ethidium bromide-stained gel was used as a loading control. The positions of RNA size markers are indicated on the left.

PpPPR_71 is a trans-acting factor required for RNA editing of ccmF-2

If PpPPR_71 were a bona fide trans-acting factor for RNA editing of the ccmF-2 site, PpPPR_71 could bind to the RNA carrying a ccmF-2 editing site. To investigate this possibility, we performed an electrophoretic mobility shift assay (EMSA) using the recombinant PpPPR_71 and the putative target RNA. The recombinant PpPPR_71 protein without the transit peptide (r-71FULL) was expressed in E. coli as a fusion protein with thioredoxin at its N-terminus (Figure 5a). We also prepared the recombinant DYW domain (r-71DYW). In our preliminary experiment, we noticed that the DYW domain possesses a weak RNA-binding property. Therefore, we carried out RNA-binding assays using the recombinant DYW domain for this study to compare binding properties of the full-length PpPPR_71 and the DYW only.

Figure 5.

 Detection of binding of the recombinant PpPPR_71 to the target RNA.
(a) Expression and purification of the recombinant PpPPR_71 and DYW proteins. The purified recombinant proteins (r-71FULL and r-71DYW) were separated by SDS-PAGE and stained with Coomassie brilliant blue. Molecular mass standards are indicated on the left of each gel.
(b) Schematic representation of the RNA probe used for the experiments.
(c) An electrophoretic mobility shift assay (EMSA) was carried out as described in Experimental Procedures. The binding of r-71FULL or r-71DYW to the RNA probe was examined in lanes 3 to 9. Binding of the r-Trx to the RNA probe was examined in lane 2. The concentration of recombinant proteins is indicated above each lane. The positions of the protein–RNA complex and free RNA are indicated by white and black arrowheads, respectively.
(d) The EMSAs using cold ccmF RNA as a competitor RNA were carried out as above. The concentration of competitor RNA is indicated above each lane. Non-labeled RNA for competition was pre-incubated with the recombinant PpPPR_71 (50 nm) and yeast tRNA (0.5 nm) for 10 min before the labeled RNA probe (0.05 nm) was added.
(e) The EMSA using cold cox3L RNA as a competitor RNA was carried out as above.

The recombinant PpPPR_71 was used for the analysis of RNA-binding properties. The target RNA (402-nucleotide (nt) ccmF RNA) includes 258 nt of upstream and 143 nt of downstream sequences surrounding the ccmF-2 editing site (Figure 5b). The RNA was 32P-labeled and incubated with the recombinant protein. The protein–RNA complex was detected as shifted bands that migrated more slowly than free RNA probe in the gel. As shown in Figure 5(c), the retarded bands were detected upon addition of 5 nm of r-71FULL (right panel, lane 6) or 50 nm of r-71DYW (left panel, lane 8). Thus, the binding affinity of r-71FULL to the RNA probe was much stronger than that of r-71DYW. To eliminate the possibility that thioredoxin binds to the RNA probe, the RNA was incubated with recombinant thioredoxin (r-Trx). Even a 2000-fold amount of r-Trx (100 nm) added to the RNA (0.05 nm) did not induce any shift band (Figure 5c, lanes 2), indicating that the RNA-binding activity depends on PpPPR_71 and its DYW domain.

The formation of RNA–protein complexes was inhibited by the addition of 10 times excess amounts of cold ccmF RNA in r-71FULL (Figure 5d, right panel). In contrast, the retarded RNA appeared in the presence of excess amounts of cold unrelated RNA, cox3L RNA (Figure 5e). This indicated that PpPPR_71 and its DYW specifically bind to the target RNA.

PpPPR_71 preferably binds to the edited RNA rather than the unedited RNA at the ccmF-1 site

The two editing sites of ccmFc mRNA are separated by only 18 nt and therefore RNA editing at each site could be influenced by the other. To investigate this possibility we examined the RNA editing status of ccmFc transcripts from the wild-type moss protonemata. Among 108 ccmFc cDNA clones, 89 cDNAs (82.4%) were fully edited, 10 (9.3%) were unedited for both sites, and 9 were edited at the ccmF-1 site but not at the ccmF-2 site. However, we could not isolate cDNA edited at ccmF-2 only. This suggests that RNA editing for ccmF-2 may depend on RNA editing of ccmF-1.

To specify the binding site of PpPPR_71, we performed EMSA using 46-nt RNA probes, RNA1 to RNA3, that include 40 nt of the upstream sequence and 5 nt of the downstream sequence surrounding the ccmF-2 editing site. The sequences of the three RNA probes differed by the C or U residue at the ccmF-1 and ccmF-2 editing sites (Figure 6a). RNA1, RNA2 and RNA3 are unedited, partially edited and fully edited RNA, respectively. The affinity of r-71FULL to RNA2 was much higher than that to RNA1 and RNA3, while r-71DYW showed a similar binding affinity to the three RNA probes. This result indicates that the full-length PpPPR_71 preferably binds to RNA2 carrying the edited U at the ccmF-1 site and the unedited C at the ccmF-2 (Figure 6b, right panel). Thus, the binding activity of PpPPR_71 probably depends on the RNA editing status of the target RNA.

Figure 6.

 Detection of binding of the recombinant PpPPR_71 to 46-nt RNAs.
(a) Schematic representation of RNA probes used for the experiments. RNA2 and RNA3 indicate only the nucleotides at ccmF-1 and ccmF-2 editing sites.
(b) The binding of the r-71FULL or r-71DYW to the RNA probe was examined by electrophoretic mobility shift assay (EMSA). The binding of the r-Trx to the RNA probe was also examined. The concentration of recombinant proteins is indicated above each lane. The positions of the protein–RNA complex and free RNA are indicated by white and black arrowheads, respectively.

To further estimate the relative binding affinities of PpPPR_71 and DYW for the three RNA probes, EMSA was performed using 32P-labeled RNA2 for detection and a non-labeled (cold) RNA probe as competitor RNA. The addition of increasing amounts of competitor RNA resulted in the disappearance of the shifted band when using both r-71DYW and r-71FULL (Figure 7a). The binding of r-71DYW (100 nm) to labeled RNA2 (0.05 nm) disappeared when using 1.0 nm cold RNA1 and RNA2, but binding was observed when using 1.0 nm cold RNA3 and the shift band disappeared in the presence of 5.0 nm cold RNA3. This result indicated that the binding affinity of r-71DYW to RNA3 is lower than that to RNA1 and RNA2. When cold cox3S RNA (192 nt), including an RNA editing site (Table S1) was added, the shift band did not disappear. The concentration of competitor RNA at which formation of r-71DYW and labeled RNA2 complex was inhibited by 50% were 112 pm cold RNA1, 96 pm cold RNA2, and 653 pm cold RNA3, respectively (Figure 7b, left panel). This indicates that r-71DYW specifically binds to the surrounding unedited ccmF-2 site of ccmFc RNA, rather than to the edited ccmF-2 site.

Figure 7.

 Electrophoretic mobility shift assay (EMSA) with the recombinant PpPPR_71 and the target RNA2 in the presence of cold competitor RNA.
(a) An EMSA using competitor RNA was carried out as described in Experimental Procedures. The concentrations of competitor RNAs (Cold RNA1, 2, 3, or cox3S RNA) are indicated above each lane. Non-labeled RNAs for competition, which were added in the concentrations indicated above each lane, were pre-incubated with the recombinant r-71DYW (100 nm) or r-71FULL (50 nm) for 10 min before the labeled RNA2 probe (0.05 nm) was added. The positions of the protein–RNA complex and free RNA are indicated by white and black arrowheads, respectively.
(b) The graphs show the relative levels of shifted bands at the indicated concentrations of cold competitor RNA.

The shift band of r-71FULL and labeled RNA2 was observed by the addition of 10 nm cold RNA1 and RNA3, but disappeared with 50 nm cold RNA1 and RNA3. In contrast, the shift band completely disappeared for 10 nm cold RNA2. The concentration of competitor RNA at 50% inhibition of formation of r-71FULL and labeled RNA2 complex was 10 nm cold RNA1, 2.7 nm cold RNA2, and 10 nm cold RNA3 (Figure 7b, right panel). This indicates that the binding affinity of RNA2 against r-71FULL is higher than that of RNA1 and RNA3. Cold cox3S RNA did not compete in the formation of RNA2–protein complex, indicating that r-71FULL preferably binds to RNA2 carrying an edited ccmF-1 site and an unedited ccmF-2 site.


In this study, RNA editing at the ccmF-2 site of the ccmFc transcript was completely inhibited in the PpPPR_71 disruptants displaying severe growth retardation. RNA editing at this site results in a change from serine (uCu) to phenylalanine (uUu) at the 41st residue of CcmFc. This suggests that phenylalanine at position 41 is critical for the function of CcmFc, which is involved in cytochrome c maturation (Giegéet al., 2004).

It has been found that RNA editing of the ccmF-2 site also occurs in A. thaliana (Giegé and Brennicke, 1999). A BLAST search using PpPPR_71 as a query hit yielded At1g11290 (CRR22), At4g18750, At3g57430 (OTP84), At2g27610, and At2g01510, all of which contained a C-terminal DYW domain. The former three are predicted to be plastid-localized, while At2g27610 (868 amino acids with 18 PPR motifs) and At2g01510 (825 amino acids with 10 PPR motifs) are predicted to be localized in the mitochondria. It would be interesting to know whether At2g27610 or At2g01510 are involved in RNA editing of the ccmF-2 site in Arabidopsis.

PpPPR_71 is a trans-acting factor required for RNA editing at the ccmF-2 site

PpPPR_71 is responsible for RNA editing of the ccmF-2 site of the ccmFc transcript in the mitochondria (Figure 3). Interestingly, ccmF-1 and ccmF-2 editing sites are separated by only 18 nt and therefore neighboring sites could influence each other. However, although cauliflower atp9 RNA is edited at two sites that are separated by 30 nt, both of these sites were shown by an in vitro RNA editing assay to be edited independently (Van der Merwe et al., 2006). Similarly, ndhD-3 and ndhD-5 sites of Arabidopsis plastid RNA are separated by 8 nt, and edited by a specific trans-acting factor, CRR28 and CRR22, respectively (Okuda et al., 2009). Mutations of the CRR28 gene caused a loss of RNA editing for ndhD-3 but not for ndhD-5. Thus, RNA editing of the downstream site occurs independently of whether the upstream site is edited or not. By contrast, our cDNA sequence analysis suggested that RNA editing of ccmF-2 depends on that of ccmF-1 and occurs after RNA editing at the ccmF-1 site. Alternatively, the result of cDNA sequence analysis can be explained, namely that editing at the ccmF-1 site is simply faster or more efficient than editing at the ccmF-2 site.

In general, C-to-U RNA editing in plant organelles requires a cis-acting element for site recognition. For plant mitochondria, about 15–40 nt upstream and 6 nt or none downstream are necessary for in vitro RNA editing in pea lysate and in organello RNA editing in wheat (Farréet al., 2001; Takenaka et al., 2004; Neuwirt et al., 2005). Our electrophoretic mobility shift assay results (Figures 6 and 7) clearly indicated that PpPPR_71 binds specifically to the 46-nt RNA sequence surrounding the ccmF-2 editing site. We therefore conclude that PpPPR_71 is a trans-acting factor required for RNA editing of the ccmF-2 site.

PpPPR_71 may preferably bind to the edited ccmF-1 transcript

PpPPR_71 binds strongly to the RNA2 probe, which carries edited U at ccmF-1 and unedited C at ccmF-2. In contrast, PpPPR_71 binds weakly to RNA1 (unedited) and RNA3 (fully edited). Thus, the RNA status at the ccmF-1 site is somewhat involved in PpPPR_71 binding. Our significant finding is that the DYW domain itself has an RNA-binding activity and binds to the surrounding ccmF-2 editing site. Taken together with the in vivo RNA editing status of ccmFc transcripts, we propose a model in which PpPPR_71 preferentially binds to the cis-acting element carrying U at ccmF-1 and then acts as a trans-acting factor for ccmF-2 RNA editing after completion of ccmF-1 editing. The PPR motifs and the DYW domain cooperatively bind to the cis-acting element for RNA editing. If a defect of the ccmF-1 recognition factor also abolishes or strongly inhibits the editing of ccmF-2, our model will be supported. One of the remaining nine PPR-DYW proteins of P. patens is possibly involved in RNA editing at the ccmF-1 site. We are currently screening the knockout mutants of moss PPR-DYW genes.

Diverged function of the DYW domain in organelle gene expression

To date, two distinct functions of the DYW domain have been considered. Firstly, it is proposed that the DYW domain has cytidine deaminase activity (Salone et al., 2007). Although extensive trials were performed to detect cytidine deaminase activity in the recombinant DYW domains (Nakamura and Sugita, 2008; Okuda et al., 2009), the activity remains to be detected. Secondly, the DYW-subclass PPR protein CRR2 is involved in the intergenic RNA cleavage between rps7 and ndhB in Arabidopsis plastids (Hashimoto et al., 2003). The DYW domain was recently demonstrated to have endoribonuclease activity (Nakamura and Sugita, 2008; Okuda et al., 2009). In the present study, we observed that DYW has an ability to bind RNA. If the DYW domain has cytidine deaminase and/or nuclease activities, this domain should contact the RNA somehow. Binding to RNA can be just a part of these enzyme functions.

During the evolution of land plants DYW domains diverged and expanded their functions, e.g. editing, cleavage, or binding to target RNA. Many questions remain unanswered, such as how their specificity of action is achieved and which transcripts are targets for each particular PPR protein.

Experimental procedures

Plant material and culture conditions

Protonemata of the moss P. patens subsp. patens were grown at 25°C on cellophane-overlaid BCDATG or BCD medium under constant light conditions (30 μmol photons m−2 sec−1) as described previously (Nishiyama et al., 2000). The protonemata were collected and spread on fresh medium every 4 days.

Isolation of DNA and RNA

Total cellular DNA was isolated by a cetyltrimethylammonium bromide method (Murray and Thompson, 1980) and treated with RNase A (TaKaRa, to remove residual RNAs. Total cellular RNA was isolated using TRIzol (Invitrogen,, chloroform, and isoamyl alcohol and treated with DNase I (TaKaRa) (Kabeya et al., 2007).

Construction of plasmid and moss transformation

To disrupt the PpPPR_71 gene, the 5′ region (3010 bp) of the gene was amplified from the moss total cellular DNA by PCR with the primers 71KO-F and 71KO-R (Table S2) and cloned into pGEM-T Easy (Promega, to generate a plasmid pET-1. The chimeric nptII gene cassette from pG35KN1 (provided by Chieko Sugita), which has the cauliflower mosaic virus 35S gene promoter, the nptII coding region, and the 35S gene terminator, was excised as a 2095-bp SmaI fragment. This fragment was inserted into the blunted StuI site in the PpPPR_71 gene in pET-1. The resulting plasmid, pET-2, was digested by NotI and introduced into the P. patens protonemal protoplasts by polyethylene glycol-mediated transformation or particle bombardment as described previously (Miwa et al., 2006; Hattori et al., 2007). Transformed mosses were cultured on BCDAT medium containing 50 μg/ml geneticin G418. Gene disruption was confirmed by genomic Southern blot analysis and PCR analysis. A PpPPR_71 gene probe (426 bp) was amplified by PCR with the primers PpPPR_71F and PpPPR_71R (Table S2), and labeled with [α-32P]dCTP using a random primer DNA labeling kit (TaKaRa).

Intracellular localization of PpPPR_71

DNA-free RNA was reverse transcribed to synthesize cDNAs using an oligo(dT) primer. The nucleotide sequence encoding the N-terminal part of PpPPR_71 protein was amplified from cDNA by PCR using primers 71GFP-F and 71GFP-R. The amplified DNA fragment (285 bp) was cloned in-frame into the SmaI site in pKSPGFP9 (provided by Chieko Sugita), that was derived from the sGFP(S65T) gene (Chiu et al., 1996). The resultant chimeric plasmid pET-3 was introduced into the 4-day-old P. patens protonemata by particle bombardment using a helium gun device, GIE-III IDERA (Tanaka Co., as described previously (Miwa et al., 2006). One day after bombardment, the fluorescence of GFP was monitored using a confocal laser scanning microscope LSM510 (Carl Zeiss,

Identification of RNA editing sites

DNA-free RNA was reverse transcribed to synthesize cDNAs using a random hexamer. Complementary DNA was amplified by PCR using 49 primers (Table S2) and subjected to direct sequencing with the DYEnamic ET Terminator Cycle Sequencing Kit (GE Healthcare, and ABI 3100 DNA sequencer (ABI Applied Biosystems, RNA editing sites were analyzed by comparing the sequence of each cDNA to that of genomic DNA (Terasawa et al., 2007). To examine the order of ccmF-1 and ccmF-2 editing in ccmFc transcripts, cDNA was amplified by PCR. The amplified cDNA fragments were subcloned into pGEM-T Easy, and DNA sequencing was performed using M13 universal primers.

Northern blot analysis

Twenty micrograms of total cellular RNA was loaded onto a 1% agarose gel containing formaldehyde, and then transferred to Hybond N+ nylon membrane (Amersham Biosciences, The membrane was hybridized and washed at 65°C (Hattori et al., 2007). A ccmFc-specific DNA probe (364 bp) was amplified by PCR using appropriate primers (Table S2). The DNA probe was labeled with [α-32P]dCTP using a random primer DNA labeling kit (TaKaRa).

Production of recombinant proteins

The DNA sequences (324 and 2379 bp) encoding the C-terminal DYW domain and the full size of PpPPR_71 protein without putative transit peptide, 40 amino acids, were amplified by PCR from genomic DNA using the primers 71-R and 71DYW-F or 71FULL-F. The amplified DNA fragments were inserted in-frame into pBAD/Thio-TOPO vector (Invitrogen) and sequenced for verification. The recombinant proteins were expressed in E. coli TOP10 as a fusion protein with thioredoxin at its N-terminus, and a V5 epitope and six histidine residues at the C-terminus. They were purified by binding to Ni-NTA Agarose (Qiagen, and dialyzed against a solution [20 mm 2-amino-2-(hydroxymethyl)-1,3-propanediol (TRIS)-HCl, pH 7.9, 60 mm KCl, 12.5 mm MgCl2, 0.1 mm EDTA, 2 mm dithiothreitol, and 17% glycerol].

Preparation of RNA probes

DNA sequences were amplified by PCR with the primers ccmF-F3 and ccmF-R2 for preparation of 402-nt ccmF RNA, the primers cox3-F2 and cox3-R2 for 418-nt cox3L RNA probe, or primers cox3-F3 and cox3-R for 192-nt cox3S RNA. The amplified DNA fragments were used as templates for in vitro transcription using T7 RNA polymerase (TaKaRa) to produce [α-32P]-CTP labeled RNA probe or cold RNA as a competitor. For this purpose, the T7 promoter sequence (ATGTAATACGACTCACTATAGGGG) was attached to the 5′-end of the forward primers ccmF-F3, cox3-F2, or cox3-F3. Forty-six-nucleotide RNA1, RNA2, and RNA3 were synthesized as synthetic oligonucleotides. Each synthetic oligonucleotide was 5′ end labeled by incubating with T4 polynucleotide kinase (TaKaRa) and [γ-32P]-ATP at 37°C for 1 h and extracted by ethanol precipitation (Sambrook et al., 1989).

Electrophoretic mobility shift assay

The assay was essentially performed according to Hattori and Sugita (2009). A binding reaction was carried out by mixing the various amounts of r-71DYW or r-71FULL with 32P-labeled RNA probe (0.05 nm) in a total volume of 20 μl of a solution containing 4 mm TRIS-HCl (pH 7.9), 7 mm MgCl2, 2.5 mm dithiothreitol, 25 mm KCl, 4.3% glycerol, 0.025 mm EDTA, and 0.5 nm yeast tRNA. Non-labeled competitor RNAs were pre-incubated with the protein for 10 min, and then the labeled RNA was added. The reaction mixture was incubated for 15 min at 25°C. The samples were then loaded onto 6% non-denaturing polyacrylamide gel and electrophoresed at 4°C in TRIS–borate–EDTA (TBE) buffer (Sambrook et al., 1989). The gel was dried and exposed to an imaging plate overnight, then visualized by a multibioimager STORM 820 (Amersham Biosciences).


We thank Yoshiaki Aoki for construction of plasmids used in this study. We are also grateful to Chieko Sugita for giving us for pG35KN1 and pKSPGFP9, Yasunori Machida and Yasushi Yoshioka for kindly instruction of a confocal microscopy. This work was supported by a Grant-in-Aid for Scientific Research (C) from the Japan Society for the Promotion of Science (JSPS) KAKENHI (19570157).