A moss pentatricopeptide repeat protein binds to the 3′ end of plastid clpP pre-mRNA and assists with mRNA maturation


M. Sugita, Center for Gene Research, Nagoya University, Nagoya 464-8602, Japan
Fax: +81 52 789 3080
Tel: +81 52 789 3080
E-mail: sugita@gene.nagoya-u.ac.jp


Pentatricopeptide repeat (PPR) proteins constitute a large family in land plants and are required for various post-transcriptional steps associated with RNA in plant organelles. The moss Physcomitrella patens PPR protein, PpPPR_38, is a nuclear-encoded chloroplast protein and was previously shown to be involved in the maturation step of chloroplast clpP pre-mRNA. To understand precisely the molecular function of PpPPR_38, we prepared recombinant PpPPR_38 protein and characterized it in maturation steps of clpP pre-mRNA. In vitro RNA-binding assays showed that the recombinant protein strongly bound to the clpP-5′-rps12 intergenic region, which is highly AU-rich and includes an inverted repeat sequence potentially forming a stem-loop structure. Digestion of the bound RNA region by RNase V1 was significantly accelerated by the addition of the recombinant protein. This strongly suggests that the binding of PpPPR_38 facilitates the formation of a stable stem-loop structure. An in vitro degradation assay using chloroplast lysates gave rise to the possibility that the stable stem-loop structure formed by PpPPR_38 contributes the correct intergenic RNA cleavage and protection of mature clpP mRNA against 3′ to 5′ exoribonuclease. Because an RNA-binding assay also showed weak binding to the clpP first exon–intron region, PpPPR_38 is likely to be related to the splicing of clpP pre-mRNA. Taking together all of the above findings, we conclude that PpPPR_38 is necessary for several steps in the clpP mRNA maturation process.


electrophoretic mobility shift assay


intergenic region


inverted repeat


pentatricopeptide repeat


recombinant PpPPR_38 protein


recombinant thioredoxin protein


single-stranded DNA


Chloroplast gene expression is regulated at both the transcriptional and post-transcriptional levels by numerous nuclear-encoded protein factors, many of which have been identified by genetics and biochemical studies [1–5]. Several of these regulatory proteins are members of the pentatricopeptide repeat (PPR) protein family, and are characterized by tandem repeats of a degenerated 35-amino acid motif [6,7]. A large PPR protein family exists, particularly in the mosses to flowering plants [8,9]. Most PPR proteins have a predicted target peptide sequence to the plastids or mitochondria [7] and were reported to be involved in embryogenesis [10,11], plastid biogenesis [12–14], photosynthesis [15,16] and cytoplasmic male sterility [17–20]. The accumulated data provide a context in which PPR proteins are required for almost all of the steps associated with RNA, from transcription to translation in plant organelles [21].

PPR protein has a PPR motif consisting of a pair of α-helices [6] and belongs to members of the helical repeat protein family. The helical repeat protein families consist of repeated motifs with an α-helix that are involved in binding to nucleotides [22]. Several PPR proteins were demonstrated to function as RNA-binding proteins to specific target RNAs. For example, the Arabidopsis HCF152 binds in vitro to the intergenic region (IGR) of psbHpetB pre-mRNA and the 3′ terminal region of the petB intron mRNA [23]. Maize CRP1 was shown by in vivo immunological analysis to associate with the IGR of the plastid cemA-petA and ndhE-psaC pre-mRNAs [24]. Maize PPR4 also was shown to associate with plastid rps12 pre-mRNA [13]. The Arabidopsis CRR4 specifically bound to 36 nucleotides surrounding the editing site of ndhD-1 [25]. Rice PPR protein Rf1 binds to the IGR of atp6-orf79 transcripts [20,26]. Because PPR motifs themselves have a binding property to RNA [23], PPR proteins are generally accepted to act as sequence-specific RNA-binding proteins. However, the available information is poor regarding the details of PPR protein function at the molecular level. The molecular functions of only a few PPR proteins have been characterized in detail, and many questions remain unanswered, such as how their specificity of action is achieved and which transcripts are the targets for each particular PPR protein.

We previously characterized PPR531-11, a member of the moss Physcomitrella patens PPR protein family [27]; thereafter, PPR531-11 was renamed PpPPR_38 [9]. PpPPR_38 is required for two RNA maturation steps: RNA intergenic cleavage between clpP and 5′-rps12 and RNA splicing of clpP pre-mRNA [27]. In the present study, we show that the recombinant PpPPR_38 (rPPR) protein binds specifically to the 5′ region containing the first exon–intron of clpP and to the IGR of clpP-5′-rps12 pre-mRNA. Moreover, RNA-binding and degradation assays strongly suggest that the binding of PpPPR_38 to the IGR may promote the formation of a stable secondary structure resistant to 3′–5′ exoribonuclease digestion. On the basis of our analyses, we propose a model of PpPPR_38 function in the processing of clpP pre-mRNA and the stability of the mature mRNA.


PpPPR_38 binds to the 5′ and 3′ regions of clpP pre-mRNA

A previous study suggested that PpPPR_38 associates with clpP pre-mRNA and plays an important role in the clpP mRNA maturation step [27]. To investigate this possibility, we performed an in vitro RNA-binding assay using rPPR (68 kDa) and a series of RNA probes. rPPR lacks the N-terminal 55 amino acid residues of PpPPR_38 and is a fusion protein with thioredoxin that is devoid of RNA-binding property. As a control protein, a recombinant thioredoxin protein was also prepared (rTrx; 16 kDa) (Fig. 1A). The expressed recombinant proteins were purified as soluble protein. We then performed an electrophoretic mobility shift assay (EMSA) using five different RNA probes (RNA-a to RNA-e) ranging from 347 to 397 nucleotides in length (Fig. 1B). The 32P-labeled RNA probes were incubated with rPPR or rTrx and RNA–protein complexes were detected as a shifted band in the gel. Clear shifted bands were detected when using RNA-a and -e (Fig. 1C, lanes 2 and 10) but were not detected for RNA-b, -c and -d (Fig. 1C, lanes 4, 6 and 8). Such shifted bands were not detected when rTrx was incubated with RNA-a or -e (Fig. 1C, lanes 11 and 12). RNA-a extends from the translated region of exon 1 (E1) to the 5′ half of the first intron of clpP and RNA-e extends from exon 3 (E3) to the IGR between clpP and 5′-rps12. This finding indicates that PpPPR_38 binds in vitro to the 5′ and 3′ regions of clpP pre-mRNA.

Figure 1.

 EMSA with the recombinant PpPPR_38 protein and target RNAs. (A) Schematic structure of recombinant proteins, PpPPR_38 (rPPR) and thioredoxin (rTrx). V5 and 6xHis indicate V5 epitope and six histidine residues, respectively. The purified proteins were separated by SDS-PAGE and stained with Coomassie brilliant blue. Open and closed arrowheads indicate rPPR (68 kDa) and rTrx (16 kDa), respectively. (B) RNA-a (370 nucleotides), b (356 nucleotides), c (347 nucleotides), d (390 nucleotides) and e (397 nucleotides) were used for EMSA, and their locations are indicated under the gene map. Open boxes (E1–E3) indicate the translated regions of exons 1–3 of the clpP gene. The gray box shows 5′-rps12. (C) EMSA was performed in the presence (+) or absence (−) of recombinant protein (50 nm; rPPR or rTrx) as described in the Experimental procedures. The positions of the protein–RNA complex are indicated by open arrowheads.

PpPPR_38 preferably binds to the IGR between clpP and 5 ′-rps12

To compare the relative binding affinity of PpPPR_38 to RNA-a and -e, we carried out a filter binding assay. For this assay, the protein bound-RNA probe is trapped on the nitrocellulose membrane but the free-RNA will pass through the membrane. After RNA-a or -e was incubated with different amounts of rPPR, incubation mixtures were applied to the membrane and free-RNA was washed out. As shown in Fig. 2, RNA-e bound efficiently to the membrane when increasing the amounts of rPPR (Fig. 2B, panel e), whereas RNA-a was not bound on the membrane (Fig. 2B, panel a). This suggests that the binding affinity of PpPPR_38 to RNA-e is stronger than that to RNA-a. To further investigate the region of RNA recognized by PpPPR_38, a filter binding assay was performed using shorter RNA probes, RNA-IGR (124 nucleotides) corresponding to the IGR between clpP and 5′-rps12, and RNA-ex3 (137 nucleotides) located on clpP exon 3 (Fig. 2A). The assay showed that RNA-IGR was trapped on the membrane but that RNA-ex3 was not (Fig. 2B, panels IGR and ex3). The dissociation constant (Kd) values were 14.8 nm for RNA-e and 13.4 nm for RNA-IGR (Fig. 2B), indicating that rPPR has the same binding affinity to RNA-e and -IGR.

Figure 2.

 Filter binding assay with the recombinant PpPPR_38 protein and target RNAs. (A) The positions of RNA probes are indicated. RNA-a and RNA-e are the same as those used in Fig. 1. RNA-ex3 (137 nucleotides) and RNA-IGR (124 nucleotides) represent the clpP exon 3 (E3) and the IGR, respectively. (B) Filter binding assays. 32P-labeled RNAs (0.5 fmol) were slot-blotted on the nylon membranes (Input RNA). Recombinant protein rPPR (0.1–50 nm) or rTrx (50 nm) were mixed with 32P-labeled RNAs (0.5 fmol) and then passed through the membrane. The trapped 32P-labeled RNA was detected by autoradiography. The graph shows the binding curves of bound RNA probe versus the rPPR concentration for each RNA probe. The ratio of bound RNAs to input RNAs was calculated from the results of filter binding assays and spotted on the graph by the indicated symbols. Approximating curves were drawn for the result of RNA-e (solid line) and RNA-IGR (broken line). Dissociation constants (Kd) were calculated from the concentration of rPPR that is at a 0.5 ratio. (C) Nonlabeled RNAs for competition, which were added in the concentrations indicated above each lane, were mixed with rPPR (final concentration 15 nm) and yeast tRNA (2.5 fmol) before the labeled RNA probe (0.5 fmol) was added. Quantitative data of competition binding assays are also shown under the autoradiographic data.

To confirm that the binding of rPPR to RNA-IGR region is sequence-specific, we performed the assay using nonlabeled RNA-IGR and -ex3 as competitor RNAs (Fig. 2C). The addition of a 50-fold amount of nonlabeled RNA-ex3 had no effect on rPPR-RNA binding (Fig. 2C, panel ex3). By contrast, the addition of a five-fold and a more than ten-fold amount of nonlabeled RNA-IGR resulted in a slightly and significantly reduced amount of bound RNA probe, respectively (Fig. 2C, panel IGR). These results indicate that PpPPR_38 specifically binds to the IGR between clpP and 5-rps12.

PpPPR_38 is an RNA-binding protein that recognizes AU-rich sequence

RNA-IGR is highly AU-rich (43% adenosines and 47% uridines). This suggests that, in general, PpPPR_38 has a binding property to AU-rich sequences. To confirm this, we performed a filter binding assay using RNA homopolymers or calf thymus single-stranded DNA (ssDNA) as competitors (Fig. 3). The addition of a 50-fold excess amount of poly(A) and poly(U) resulted in the significant reduction of binding of rPPR to the RNA-IGR probe (Fig. 3, lanes 3 and 4). By contrast, the addition of poly(G), poly(C) or ssDNA had no effect on the binding (Fig. 3, lanes 5–7). This result indicates that PpPPR_38 has a strong binding affinity to AU-rich sequences.

Figure 3.

 Assay of RNA binding property of the recombinant PpPPR_38 protein. rPPR was preincubated with a 50-fold amount of competitor RNA homopolymers poly(A), poly(U), poly(G) or poly(C) or competitor ssDNA before the addition of 32P-labeled RNA-IGR (0.5 fmol). The binding was quantified by an imager, using binding in the absence of competitor as 100.

PpPPR_38 stabilizes a stem-loop structure at 3′ UTR of clpP mRNA

On the basis of the position of the cleavage site, we prepared and used further three RNA probes, RNA-f (199 nucleotides), RNA-g (111 nucleotides) and RNA-h (77 nucleotides) for the filter binding assay (Fig. 4A). As shown in Fig. 4B, rPPR specifically bound to the RNA-f and -g but not to RNA-h and -ex3. This indicates that PpPPR_38 binds to the upstream region from the cleavage site.

Figure 4.

 Binding affinity of PpPPR_38 to the IGR between clpP and 5′-rps12. (A) The RNA probes used for filter binding assay. RNA-ex3 are the same as those used in Fig. 2A. RNA-f (199 nucleotides), RNA-g (111 nucleotides) and RNA-h (79 nucleotides) reside in the IGR between clpP exon 3 (E3) and 5′-rps12. The arrowhead indicates the intergenic cleavage site. (B) Filter binding assays of rPPR. Input RNA (0.5 fmol) was slot-blotted on the membrane. Filter binding assays were performed in the absence of protein (no protein) or in the presence of 50 nm recombinant protein (rPPR or rTrx). (C) Predicted secondary structure of 3′ end of clpP mRNA. The arrowhead shows the intergenic cleavage site. (D) RNase V1 digestion assay for RNA-g. RNA-g was incubated with RNase V1 in the absence (lane 2) or presence (lanes 3–6) of rPPR. RNA-g was mixed with rPPR but not RNase V1 (lane 6), whereas it was mixed with rTrx and RNase V1 (lane 7). After RNase V1 digestion, RNA-g was analyzed by nondenaturing PAGE. The RNA level was quantified by an imager, using digestion in the absence of rPPR and RNase V1 as 1.0 (lane 1).

In the 3′ UTR of clpP mRNA, there is an inverted repeat (IR) sequence that form a potential stem-loop structure (Fig. 4C), which is composed of 71 nucleotides, representing a relatively unstable structure (ΔG = −8.8). To test the configuration of the stem-loop structure, RNA-g was incubated with RNase V1, which is known to digest only double-stranded RNAs [28]. Some digestion (i.e. 50–60%) takes place in the absence (Fig. 4D, lane 2) but much more in the presence of rPPR (lanes 3–5). Such an enhancement of RNA digestion did not occur when rTrx was added (lane 7). This suggests that an incomplete stem-loop structure is formed in the absence of PpPPR_38 and is weakly digested by RNase V1, whereas PpPPR_38 binds to the incomplete stem-loop structure and makes it a stable double-stranded RNA structure.

PpPPR_38 is possibly involved in the restraint of 3′ to 5′ degradation of clpP mRNA

A stem-loop structure in the 3′ UTR of plastid mRNA is required for the protection of mRNA against 3′ to 5′ exoribonuclease [29,30]. To examine whether the conformation of stable stem-loop structure in the clpP 3′ UTR is required for RNA stability, we performed an in vitro degradation assay using chloroplast lysate and exogenously added RNA probe. RNA-IR(+) (309 nucleotides) contains the clpP exon 3 and 3′ UTR, and RNA-IR(−) (222 nucleotides) consists of the clpP exon 3 only (Fig. 5A). RNA-IR(+) was incubated in the chloroplast lysate from the wild-type or PpPPR_38 disruptant moss, and monitored for its digestion levels (Fig. 5B). Both chloroplast lysates digested approximately 95% of RNA-IR(+) within 30 min (Fig. 5B, lanes 3 and 5), and also digested RNA-IR(−) efficiently (Fig. 5C, lanes 2 and 3). When rPPR was mixed with RNA-IR(+), and then incubated in the chloroplast lysate from the PpPPR_38 disruptant, degradation of the RNA probe was inhibited (Fig. 5B, lanes 6–8). Degradation was strongly inhibited in the presence of 2 nm rPPR (Fig. 5B, lane 8). By contrast, this inhibition was not observed for RNA-IR(−) (Fig. 5C, lane 4). This suggests that PpPPR_38 binds to the 3 end of clpP and stabilizes its stem-loop structure.

Figure 5.

 Effect of PpPPR_38 on in vitro RNA digestion in the chloroplast lysate. (A) The regions of RNA-IR(+) and RNA-IR(−) are indicated. RNA-IR(+) contains a clpP 3′ UTR sequence and RNA-IR(−) does not. (B) In vitro RNA digestion assays for RNA-IR(+). RNA-IR(+) was incubated in the chloroplast (Cp) lysate from the wild-type moss (W, lanes 2 and 3) or PpPPR_38 disruptant (M, lanes 4 and 5) in the absence (lanes 1 to 5) or presence (lanes 6 to 9) of rPPR. (C) In vitro RNA digestion assays for RNA-IR(−). RNA-IR(−) was incubated in the Cp lysate from the wild-type moss (lane 2) or PpPPR_38 disruptant (lanes 3 and 4) in the absence or presence of rPPR.


PpPPR_38 is an RNA-binding protein that is targeted to clpP mRNA

The results obtained in the present study clearly show that the rPPR protein binds specifically to the 5′ and 3′ regions in clpP pre-mRNA. Notably, the binding of rPPR to the 3′ UTR of clpP mRNA was detected by filter binding assays (Fig. 2). This indicates that PpPPR_38 is likely to have a higher affinity to an AU-rich sequence in the clpP 3′ UTR, as supported by the competition experiments using cold RNA homopolymers (Fig. 3). Accordingly, PpPPR_38 would be characterized as an AU-rich sequence binding protein. High binding affinity to poly(A) was observed for PPR protein HCF152 [23]. Four tested Arabidopsis PPR proteins were shown to preferentially bind poly(G) homopolymer [7]. A mitochondrial PPR protein LOVASTATIN INSENSITIVE 1 also exhibited a poly(G)-binding property [31]. Thus, the binding properties to RNA homopolymers differ among the PPR proteins. This would be the result of differences in the number and arrangement of multiple PPR motifs.

The binding sequences of some PPR proteins were assigned. Maize CRP1 protein interacts with the 69 nucleotide regions of single-stranded RNA encompassing the conserved 7-mer and 11-mer motifs and the intervening 51 nucleotides in the psaC and petA 5′ UTRs [24]. Similarly, maize PPR10 associates with the conserved 25 nucleotides sequences in the IGRs of psaJ-rpl33 and atpI-atpH [32]. Arabidopsis CRR4 specifically binds to the 36 nucleotide sequence surrounding the editing site of ndhD-1 [25]. These PPR proteins might recognize certain primary sequences in the target RNA. By contrast, PpPPR_38 is likely to recognize and bind to a potential stem-loop structure in the clpP 3′ UTR. This is confirmed by the assay using RNase V1 that specifically cleaves base-paired nucleotides [28]. The target RNA for PpPPR_38 (RNA-g probe) was digested by RNase V1 to some extent (40% of input RNA) without rPPR protein, indicating that this RNA indeed forms a secondary or tertiary structure with some double-stranded regions. Of great interest, digestion of RNA-g by RNase V1 was significantly accelerated in the presence of rPPR (Fig. 4D). The simplest explanation for this result is that incomplete or unstable double-stranded RNA might be altered to a structure having stable and longer double-stranded regions. Such a structurally altered RNA exhibits more sensitivity to RNase V1. However, with respect to the binding affinity, some PPR proteins are reported to preferentially bind to single-stranded RNA [7,33]. It is thus possible that PpPPR_38 recognizes single-stranded sequences both upstream and downstream of the stem-loop structure, thereby stabilizing it. This would be consistent with the preserved access of RNase V1 to the stem. In a future study, it would be interesting to investigate whether the length and sequence of the stem-loop structure are important for the interaction with PpPPR_38, or whether substitutions within the stem-loop are tolerated, provided that the sequences surrounding it are conserved. This has been observed for a stem-loop structure located at the 5′ end of the rbcL mRNA and required for its stability in the green alga Chlamydomonas reinhardtii [34].

PpPPR_38 alters the conformation of stem-loop structure and stabilizes mRNA

The second intriguing observation (Fig. 5) provides a possible function of PpPPR_38 in RNA stability. Once PpPPR_38 binds to an IR region in the clpP 3′ end and forms stable RNA–PPR complexes, the bound RNA itself becomes resistant to nucleases, most likely 3′–5′ exoribonucleases (Fig. 5B). This occurs in a PpPPR_38 protein dose-dependent manner. This is consistent with the well accepted context that many plastid mRNAs have an IR region at their 3′ end that potentially forms a stem-loop structure associated with RNA stability or degradation [35–37]. A stem-loop structure in the 3′ end of plastid mRNA is required for the protection of mRNA against 3′–5′ exoribonucleases [29,30,37]. Chloroplast lysate contains various RNase activities and thereby the exogenously added 3′ IR-less RNAs are rapidly degraded in the lysate [38,39]. In the present study, we showed that recombinant PpPPR_38 protein also recognizes and binds to a stem-loop structure at the 3′ end of clpP mRNA. Probably, PPR proteins are an important determinant for the stability of mRNA because chloroplast RNA-binding proteins act as stabilizer of mRNAs. Recently, Pfalz et al. [32] have shown that the site-specific binding of PPR10 defines and stabilizes 5′ and 3′ mRNA termini in chloroplasts. Thus, moss PpPPR_38 may also have a similar function.

Function of PpPPR_38 in clpP mRNA maturation

In our previous study [27], the loss of PpPPR_38 resulted in the abnormal accumulation of the 3.2 kb primary transcript of clpP and 1.7 kb clpP pre-mRNA. These phenomena can be explained as follows. The 3.2 and 1.7 kb transcripts accumulated as a result of the loss or significant reduced site-specific cleavage between clpP and 5′-rps12 translated regions and a significant reduction of the splicing of clpP pre-mRNA, respectively. This strongly suggested that PpPPR_38 has a role in the site-specific RNA cleavage and splicing of clpP pre-mRNA [27]. In a study related to our results, HCF152 was shown to bind to the short IGR between psbH and petB and to the 3′ region of petB group II intron [23]. On the basis of the results obtained from carefully designed experiments, an intriguing model was proposed in which the binding of a HCF152 homodimer in this region somehow stabilizes the splicing products, possibly by folding the RNA into the correct splicing structure [23].

Taking together the results obtained in the present study and those from previous observations [27], a model of PpPPR_38 function in maturing of clpP pre-mRNA can be depicted (Fig. 6). First, PpPPR_38 weakly binds to the clpP first exon–intron region and strongly to the AU-rich region in the clpP 3′ UTR of the primary transcript. Presumably, its binding to the AU-rich region is a critical step and leads to the formation of a certain secondary structure in the IGR of clpP–5′-rps12. Perhaps, such a local conformational change of the target RNA results in the appearance of a target site for a certain endoribonuclease. Then, as a second step, immediately after completion of site-specific endonucleolytic cleavage, PpPPR_38 bound 1.7 kb clpP pre-mRNA proceeds to splice and produce a mature clpP mRNA of 0.6 kb. Thereafter, PpPPR_38 might be involved in the stability of the mature mRNA. PpPPR_38 most likely has an ability to bind the 5′ region of clpP pre-mRNA (Fig. 1), even though its binding was not observed by the filter binding assay (Fig. 2). It is possible that the binding of PpPPR_38 to the 5′ exon is required for the splicing of clpP pre-mRNA. Presumably, PpPPR_38 contributes to this process together with several RNA-binding proteins. PPR proteins are an important determinant for the stability of mRNA because chloroplast RNA-binding proteins act as stabilizers of mRNAs [2,40]. However, at present, it is not known whether PpPPR_38 retains at the 5′ region of clpP mRNA and is required for stabilization of the 5′ terminus of clpP mRNA.

Figure 6.

 A model of PpPPR_38 function in clpP mRNA maturation step. PpPPR_38 strongly binds to the IGR of clpP–5′-rps12. This is a critical step to forming a stem-loop structure. This conformational change will provide a correct cleavage site and leads the site-specific cleavage by an endoribonuclease. After the site-specific cleavage, PpPPR_38 bound clpP pre-mRNA facilitates maturation of the pre-mRNA. Thereafter, PpPPR_38 assists in the stability of the mature mRNA against a 3′–5′ exoribonuclease. PpPPR_38 also weakly binds to the clpP first exon–intron region and is required for the splicing of clpP pre-mRNA.

Experimental procedures

Preparation of the recombinant protein

A DNA fragment (1428 bp) encoding PpPPR_38 lacking N-terminal 55 amino acids was amplified from cDNA#4 [27] by PCR using primers P1 and P2 (Table S1). The amplified DNA was inserted in-frame into pBAD/Thio-TOPO (Invitrogen, Carlsbad, CA, USA) and the recombinant protein was expressed in Escherichia coli LMG194 strain as described previously [41]. The recombinant protein (rPPR) was purified by binding to nickel-nitrilotriacetic acid agarose resin (Qiagen, Valencia, CA, USA) and dialyzed against a buffer E [20 mm Tris–HCl (pH 7.9), 100 mm KCl, 12.5 mm MgCl2, 0.1 mm EDTA, 2 mm dithiothreitol and 17% glycerol]. rTrx was also prepared and purified by the same procedure as described above.

Preparation of RNA probes

DNA templates for preparation of RNA probes were amplified from P. patens chloroplast DNA [42], using the primers listed in Table S1. To prepare RNA probes, in vitro transcription was performed by T7 RNA polymerase (TaKaRa, Otsu, Japan) with 200 ng of the amplified template DNA, 0.5 mm ATP, 0.5 mm GTP, 0.5 mm CTP, 0.025 mm UTP, 1.49 MBq of [α-32P]UTP, 1.25 mm dithiothreitol and 4 U of RNase inhibitor (TaKaRa) at 37 °C for 1 h. Synthesized RNAs were treated with DNase I (TaKaRa) at 37 °C for 15 min and subjected to 6% PAGE containing 7 m urea, and eluted from the gel as described previously [43]. The gel-purified RNAs were suspended in RNase-free water. Nonlabeled RNAs as a competitor were synthesized by the same reaction conditions, except for the addition of 0.5 mm UTP in substitution for [α-32P]UTP.

RNA binding assay

For the EMSA, 32P-labeled RNA (0.1 fmol) was denatured in 19 μL of a reaction buffer 1 [2.8 mm Tris-HCl (pH 6.8), 17.2 mm KCl, 4.9 mm MgCl2, 0.02 mm EDTA, 2.3 mm dithiothreitol, 2.5 fmol of yeast RNA (Ambion, Austin, TX, USA) and 2.9% glycerol] at 95 °C for 30 s, and then cooled down to 55 °C. One microliter of recombinant protein (1 μm) was added to the denatured RNA mixture and incubated at 27 °C for 15 min. Two microliters of 80% glycerol was added to the incubation mixture, and subjected to 5% PAGE with 7 m urea in TBE buffer [43]. Free-RNA and protein-bound RNA were detected by Storm 820 (GE Healthcare, Milwaukee, WI, USA).

The filter binding assay was carried out as described previously [44]. 32P-labeled RNA (0.5 fmol) was mixed with or without a competitor in 19 μL of a reaction buffer 2 [50 mm Tris–HCl (pH 6.8), 150 mm KCl, 10 mm MgCl2, 5 mm dithiothreitol and 1 fmol of yeast RNA] and denatured as described above. One microliter of recombinant protein (1 μm) was added to the denatured RNA mixture, and gently agitated at 27 °C for 15 min. The mixture was applied to a nitrocellulose membrane (Protran BA 85; Schleicher & Schuell BioScience GmbH, Dassel, Germany; 0.45 μm), which was pre-soaked with wash buffer [50 mm Tris-HCl (pH 6.8), 150 mm KCl and 10 mm MgCl2]. After vacuum filtration, the membrane was washed twice with wash buffer. 32P-labeled RNA trapped on the membrane was detected by Storm 820. RNA homopolymers, poly(A), poly(U), poly(C), poly(G) and calf thymus ssDNA were purchased from Sigma-Aldrich (St Louis, MO, USA) and were used as competitors.

RNase V1 digestion assay

32P-labeled RNA (0.5 fmol) was performed in 8 μL of Structure buffer attached to RNase V1 (Ambion), containing 1 fmol·μL−1 yeast RNA. After heat denaturing, 1 μL of the recombinant protein or buffer E (i.e. the same as above) was mixed, and they were incubated at 27 °C for 10 min. 1 × 10−3 U of RNase V1 was added to the mixtures and reacted at 37 °C for 30 min. The reaction was stopped with precipitation/inactivation buffer attached to RNase V1, and the RNAs were precipitated by centrifugation. The resultant RNAs were subjected to 8% PAGE containing 7 m urea. The predicted secondary structure of RNA was calculated by mfold, version 3.2 (http://frontend.bioinfo.rpi.edu/applications/mfold/cgi-bin/rna-form1.cgi).

Preparation of chloroplast lysate and in vitro digestion assay

Physcomitrella patens chloroplasts were prepared from 1 g of fresh weight 4-day-old protonema [42] and were suspended in 200 μL of lysis buffer [50 mm Tris-HCl (pH 8.0), 100 mm KCl, and 10 mm MgCl2] with pipetting. After centrifugation at 15 000 g for 15 min, the supernatant was dialyzed against buffer E at 4 °C overnight. The dialyzed supernatant (150 ng total protein·μL−1) was used as chloroplast lysate for the assay.

32P-labeled RNA-g (0.5 fmol) was used for an in vitro digestion assay with 8 μL of reaction buffer 3 [50 mm Tris–HCl (pH 6.8), 50 mm KCl and 10 mm MgCl2]. After heat denaturing, the recombinant protein (0.5–2 nm) was added and incubated at 27 °C for 10 min. The chloroplast lysate (1 μL) was added and incubated at 37 °C for 10 or 30 min. The RNAs were purified by phenol and ethanol precipitation, and subjected to 8% PAGE containing 7 m urea.


We thank Takahiro Nakamura for useful discussions and advice on the RNA binding assays. Mitsuru Hattori was supported by the Japan Society for the Promotion of Science (JSPS) Research Fellowship for Young Scientists (17-7839). 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).