An RNA recognition motif-containing protein is required for plastid RNA editing in Arabidopsis and maize

Identification of a Protein Carrying Truncated RIP Domains.

A blastp search for homologs to the RIP1 protein returned the 10 previously reported members of the RIP family (19, 20) as well as a unique RIP family member, encoded by the gene At3g20930 (Fig. 1A). This protein was not previously described as either a RIP or MORF protein (19, 20). We used the MEME software (22) to identify four highly significant motifs in the RIP family (Fig. 1B). The RIP block can be defined by the following series: motif 1-gap-motif 2-motif 3-motif 4 (Fig. 1A). The distal motif 4 found in RIP7 is below the threshold of e-10 (P value = 9.6 e-6). Most RIP proteins possess a complete RIP block (Fig. 1C). The unique member of the RIP family encoded by At3g20930 exhibits a duplication of truncated RIP blocks; the first block, from amino acid 89 to amino acid 147, contains a degenerate motif 1 plus motifs 2 and 3, and the second RIP block from amino acid 163 to amino acid 250 contains motifs 1–3. Both RIP blocks in At3g20930 lack motif 4 (Fig. 1C).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

The protein encoded by At3g20930 belongs to the RIP family and contains a pair of truncated RIP domains. (A) Alignment of the conserved regions in the Arabidopsis RIP proteins (RIP1 to RIP10), the DAG protein from Antirrhinum majus (GenBank CAA65064), and the protein encoded by At3g20930 was performed using T-Coffee Version_9.03, and displayed using GeneDoc with the conserved residue shading mode and similarity groups enabled. Overlaid on the aligned sequences are the four motifs detected by the MEME software version_4.9.0. (B) The RIP domain contains four motifs uncovered by MEME. The settings were 6 aa < width < 100 aa, maximum number of motifs to find was four. All four motifs are highly significant (E-values: 7.8e-208, 5.5e-111, 3.7e-86, 4.2e-52). Each motif is given its sequence logo showing the likelihood of residue at each position. (C) The combined motif diagrams are shown for each of the RIP protein, the DAG protein, and the protein encoded by At3g20930. The height of a motif is truncated when its P value is > 1e-10, for example, motif 3 for RIP7.

At3g20930 Protein Contains an RNA Recognition Motif at Its C Terminus and Belongs to a Clade of RRM Proteins.

A motif search with Motif Scan (http://myhits.isb-sib.ch/cgi-bin/motif_scan) identified the presence of a RRM at the C terminus of the protein. The RRM domain is ∼80-aa long and contains two short consensus sequences, RNP1 (octamer) and RNP2 (hexamer), which are characteristic of RRMs (Fig. 2A).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

The RRM domain found at the C terminus of ORRM1 shows most similarity to the RRMs from glycine-rich (GR), mitochondrial (m), and small (S), RBPs. (A) RRMs from the protein encoded by At3g20930, the GR-RBPs, the mRBPs, and the S-RBPs were aligned by T-Coffee version_9.03, and displayed using GeneDoc with the conserved residue shading mode and similarity groups enabled. Depending on the database of protein domains searched, prosite or pfam, the RRM motif was located at position 282–360 with a E-value of 1.3e-11, or at position 284–350 with an E-value of 2.2e-24, respectively. Overlaid on the aligned sequences are the four motifs detected by the MEME software version_4.9.0. Location on the left of each protein refers to the subcellular location predicted by Predotar or TargetP, P, plastid; M, mitochondrion. In parenthesis preceding the name of the protein is given the annotation (m) mitochondrial (24) or (s) small (23). (B) Combined P values and block motifs computed by the MEME software indicated that the RRMs fromORRM1, the GR-RBPs, the mRBPs, and the S-RBPs belong to the same family defined by motifs a, b, c, and d. The product encoded by At5g46480, a RRM-containing protein (motifs a and c only) does not belong to this family.

In Arabidopsis, 196 RRM-containing proteins were previously identified through an in silico search for the RRM motif (23). A blast search using the RRM domain of the protein encoded by At3g20930 revealed that this domain was more closely related to the RRM found in two distinct families described by Lorkovic and Barta (23), the glycine-rich RNA-binding proteins (GR-RBP) and the small RNA-binding proteins (S-RBP). A common feature of these two protein families is their similar domain organization with one N-terminal RRM and a C-terminal extension. GR-RBPs are represented by eight members, but 15 proteins were annotated as S-RBP (23). At the time of the Lorkovic and Barta’s (23) report, Vermel et al. (24) identified by biochemical means a family of mitochondrial-specific RRM-containing proteins that they named mitochondrial RNA-binding proteins (mRBP). The 11 mRBPs belong to either the GR-RBP or the S-RBP family. Fig. 2A illustrates the strong similarity between the RRM domains found in the At3g20930 product, the GR-RBPs, and the mRBPs. To verify the similarity between the RRM domains found in the At3g20930-encoded product, the GR-RBPs, and the mRBPs, we aligned them with the RRM of a protein encoded by At5g46840, which does not belong to any of these subfamilies (Fig. 2A).

When we used the MEME software with the number of motifs set at 4, and width greater than 5 but less than 20 amino acids, four motifs were identified in this set of RRMs that can define the unique subfamily related to the RRM present in the protein containing two RIP motifs. Two of the four motifs, motif “a” and motif “c,” are found in the RRM domains of the unique subfamily but also in the RRM domain of the protein encoded by At5g46840 (Fig. 2B). In contrast, motif “b” and motif “d” are specific to the RRM domains found in the product encoded by the RIP-family protein At3g20930, as well as in the GR-RBPs and the mRBPs (Fig. 2B). All of the proteins shown in Fig. 2A are predicted to be located in plastids or mitochondria. We have therefore named the group of organelle-targeted proteins containing the four motifs described in Fig. 2 the ORRM family. The protein encoded by At3g20930 is hereafter designated as ORRM1.

A Pfam domain search using At3g20930 (ORRM1) as a query identified 642 RRM containing regions in the Arabidopsis thaliana genome. As examples, putative RNA-binding proteins, poly(A)-binding ribonucleoproteins, splicing-factors, U2 small nuclear ribonucleoproteins, Arabidopsis-mei2-like proteins, and chloroplast ribonucleoproteins were retrieved. Several clearly identifiable clusters in the phylogenetic tree can be distinguished, suggesting that RRMs can be classified according to groups of proteins of the same function, such as U2 small nuclear ribonucleoproteins, poly(A)-binding proteins, splicing factors, and chloroplast RNA-binding proteins (CP) (Fig. 3).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Phylogenetic tree based on the amino acid sequences of the RRM motifs in RRM-containing proteins (84 amino acids considered). The tree was inferred using the Neighbor-Joining method, and evolutionary distances were computed using the Poisson-correction method. The scale bar corresponds to 0.1 substitutions per site. CP, chloroplast ribonucleoprotein; PABP, poly(A) binding protein; snRP, small nuclear ribonucleoprotein. In the ORRM1 clade, the figure in parenthesis is the annotation (s) for small given in Lorkovic and Barta (23), or (m) for mitochondrial given in Vermel et al. (24).

ORRM1 appears to form a monophyletic group with members of the glycine-rich RNA binding proteins (Fig. 3). The topology of the tree constructed by the unweighted pair-group method using arithmetic averages (UPGMA), maximum-pasimony, maximum-likelihood, and minimum-evolution methods was not different from that of the Neighbor-Joining tree presented here, supporting a consistent grouping of the proteins.

T-DNA Insertional Mutant in ORRM1 Exhibits Severe Defects in Plastid Editing.

We obtained an Arabidopsis mutant from the Arabidopsis Biological Resource Center stock collection and verified that it was homozygous for a T-DNA insertion in the first exon of ORRM1 (Fig. 4A) (SALK_072648, designated here as orrm1). The homozygous mutant did not show any phenotypic defect when grown under growth room conditions (Fig. 4B). No expression of ormm1 was detected by RT-PCR (Fig. 4C). We examined the organelle transcriptome of the mutant for editing defects because other proteins carrying RIP domains have been shown to be editing factors (19, 20). We analyzed the plastid RNA editing extent with a new methodology based on RNA-seq. Briefly, total RNA is isolated from leaves and RT-PCR products corresponding to known organelle genes are obtained by using gene-specific primers. The products are mixed in equimolar ratio, sheared, and used as templates to produce an Illumina TruSeq library. This RNA-seq analysis demonstrated that ORRM1 is a plastid editing factor; 12 among 34 plastid sites exhibit a severe reduction of editing extent in the mutant relative to the wild-type (Fig. 4D).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

A T-DNA insertional mutant in ORRM1 is severely impaired in plastid editing. (A) Schematic representation of the model gene for ORRM1 with exons represented as squares and introns as lines, the T-DNA insertion is shown as a triangle in the first exon. The primers used for the RT-PCR are indicated by arrows. (B) The homozygous mutant plant (Left) does not show any phenotypic defect compared with a Columbia wild-type plant (Right) when grown under growth-room conditions. (C) No expression of ORRM1 is detectable by RT-PCR after 45 cycles in the orrm1 mutant, although expression is readily observed in wild-type. (D) Thirteen plastid sites show a severe reduction of editing extent (∆ORRM1) > 90% in the ORRM1 T-DNA insertional mutant. Because the orrm1 homozygous mutant line was in a Columbia background, five wild-type siblings of other rip family insertional mutants that were in a Columbia background served as positive controls.

In addition to the 12 sites where editing is reduced by 90% or more, 9 plastid sites exhibit a reduction of editing extent between 10 and 90% in the mutant. Thus, 62% (21 of 34) of the plastid-editing sites are under the control of ORRM1 (Table S1). We confirmed the gene-specific RT-PCR results on the orrm1 mutant by performing RNA-seq on total plastid RNA. For this purpose, total RNA was extracted from chloroplasts purified from mutant and the wild-type and reverse-transcribed using random hexamers. The number of reads per chloroplast gene ranged from ∼20 to ∼5,000 (Table S1). The numbers of reads are much higher in the gene-specific RT-PCR– generated Illumina library, with averages of ∼7,000 and ∼11,000 for the wild-type and the mutant, respectively (Table S1). Despite the difference in depth coverage between the gene-specific and total plastid RNA Illumina libraries, the reductions of editing extent in the mutant are highly consistent between the two assays (Fig. S1 and Table S1).

To verify the suitability of the RNA-seq method for assaying RNA editing, we performed poisoned primer extension (PPE) on a selection of transcripts and compared the results to the gene-specific RNA-seq and total plastid RNA-seq data (Fig. 5). In the PPE assay, cDNA serves as a template for an extension reaction in the presence of a dideoxy G. When dideoxy G is incorporated, the extension stops at the first unedited C that is encountered by the enzyme. The products of edited vs. unedited transcripts will differ in size; the amount of each product can be accurately monitored on gels. As an example, we show the PPE data for three C targets of editing in the ndhD transcript, one has an editing extent that is unaffected by the orrm1 mutation (Fig. 5A), but the other two exhibit almost complete loss of editing (Fig. 5 B and C). We performed PPE assays on six additional transcripts (Fig. S2) and demonstrate the consistency of the two assays by graphing the RNA-seq editing extent data against the data from the PPE assay (Fig. 5D).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

PPE assay validates the editing extents derived from RNA-seq. (A–C) Acrylamide gels separate the PPE products obtained from samples used in this study. E, edited; U, unedited. The name of the site assayed is given above each gel. The quantification of editing extent derived from the measure of the band’s intensity is represented by a bar below each lane of the acrylamide gels (blue diagonal background). By way of comparison, the editing extent derived from RNA-seq is represented by a magenta bar. RNA-seq was performed on gene specific cDNAs for orrm1#1, wt-1, wt-2, wt-3, wt-4, and wt-5, and on total plastid RNA for orrm1 and wt. orrm1#2 was not analyzed by RNA-seq. (D) The correlation between the editing extent values derived from PPE assay and RNA-seq verifies RNA-seq is a sound method to determine editing extent. The correlation was calculated by plotting 72 points (8 samples × 9 PPE gels).

We also surveyed the mitochondrial transcriptome of the orrm1 mutant with gene-specific primers; none of the 574 mitochondrial sites assayed showed a significant difference in editing extent between the mutant and the wild-type. Thus, ORRM1 is an editing factor that is specific to plastids.

RNA Editing Defects Are Detected in orrm1 Chloroplast Transcripts That Do Not Differ in Abundance from Wild-Type Transcripts.

Although our gene-specific RNA-seq method does not provide any information on transcript abundance, our RNA-seq experiments using total chloroplast RNA do allow us to quantify relative abundance of transcripts from different genes. The number of total plastid RNA reads corresponding to each plastid transcript in the orrm1 and wild-type total plastid RNA data exhibit little variation (Table S1), indicating that changes in RNA abundance do not explain the effect of the mutation on editing at specific C targets. To verify that abundance of transcripts carrying Cs affected in editing extent do not vary greatly between the mutant and wild-type, we performed RNA blots with total chloroplast RNA from orrm1 and wild-type plants (Fig. S3). We used three probes corresponding to the matK, ndhB, and ndhD genes, whose transcripts carry C targets with reduced editing in the orrm1 mutant. These blots demonstrate that there is no difference in the complexity of the RNA profile or abundance of particular transcript species between wild-type and orrm1 (Fig. S3). In addition, different Cs located on the same transcript sometimes vary greatly in their editing extent in the orrm1 mutant. Both ndhB and ndhD carry a C target that is unaffected in the orrm1 mutant, but the other Cs in the ndhB and ndhD exhibit major reduction in editing efficiency in the mutant (Fig. 5 and Fig S2).

RRM Domain Can Rescue the Editing Defect in orrm1 Protoplasts.

We determined whether the T-DNA insertional mutation could be complemented by transient expression of ORRM1 under the control of a 35S promoter in mutant protoplasts. PPE assay demonstrated that mutant protoplasts transfected with the construct carrying the full-length ORRM1 exhibited significant increase in editing extent (Fig. 6A, lane F). The extent of editing for ndhB-872 and rps12-i58 in transfected mutant protoplasts, 43% and 27% respectively, is sufficient to observe a very distinct edited product band on the PPE gel compared with untransfected protoplasts (Fig. 6A, lane NT). As expected in a transient expression assay, the level of editing extent of ndhB-872 and rps12-i58 in the transfected protoplasts does not reach the level observed in the wild-type plant, which is 95% and 40%, respectively (Table S1). Seventy-six percent (16 of 21) of the plastid sites assayed that showed a reduction or a lack of editing extent in the mutant exhibited a significant increase of their editing extent in the transfected protoplasts with the full-length ORRM1 (Fig. 6B).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

ORRM1 is able to complement orrm1 protoplasts with its RRM domain, not its RIP domain. (A) PPE products from not transfected protoplasts (NT), protoplasts transfected with a construct encoding a full-length (F), the N-terminal portion (N) that contains the RIP-RIP domain, or the C-terminal portion (C) that contains the RRM domain. Above each gel is given the name of the editing site (gene-position), E, edited; U, unedited; P, primer. The presence of the edited bands is only observed in protoplasts transfected with the full-length construct or with a construct encoding the RRM. (B) Twenty-three plastid sites showing a decrease in the orrm1 mutant were assayed for complementation in the transfected mutant protoplasts. Among the 16 sites complemented by the full-length construct, 15 also exhibit a complementation by the construct encoding the RRM domain of ORRM1.

Similar transfection experiments were performed with constructs encoding either the N-terminal portion of ORRM1 that contains the duplicated RIP-RIP region (Fig. 6A, lane N) or the C-terminal portion, which carries the RRM domain (Fig. 6A, lane C). Of the two truncated constructs tested, only the construct encoding the RRM domain was able to complement the editing defect of the mutant (Fig. 6A, lane C). Among the 16 sites partially complemented by the full-length ORRM1, 15 showed a significant increase of editing extent in the mutant protoplasts upon transfection with the construct encoding the RRM domain (Fig. 6B). At three sites, the full-length construct was able to complement the editing defect more efficiently than did the RRM construct. Among these sites, ndhD-674 is the only one for which no effect of the RRM construct was observed (Fig. 6B). The RRM construct more efficiently complemented five sites than did the full-length construct; among these sites, rps12-i58 shows almost twice the amount of edited transcripts in protoplasts transfected with the RRM construct than with the full-length construct (Fig. 6A). The increase of editing extent in transfected protoplasts was below the significance threshold for two sites, ndhB-746 and rps14-149. An absence of effect from the transfection with either the full-length or the RRM construct was observed in only three of the assayed sites (Fig. 6B).

Maize ORRM1 Ortholog Is Required for the Editing of Both Orthologous and Maize-Specific Sites.

Maize mutants with Mu transposon insertions in the ortholog of ORRM1 (Zm-orrm1) were recovered during the identification of causal mutations in a large collection of nonphotosynthetic mutants (http://pml.uoregon.edu/photosyntheticml.html). The Zm-orrm1 mutants originally came to our attention because of their unusual spectrum of protein deficiencies (see below), which could not easily be explained by defects in known chloroplast biogenesis genes. Therefore, the mutants were selected for gene identification with a high-throughput method for sequencing Mu insertion sites (25) (Materials and Methods). Two alleles were identified, both with an insertion in the first exon (Fig. 7A). Complementation crosses yielded heteroallelic mutant progeny (Zm-orrm1-1/Zm-orrm1-2) displaying a pale green phenotype (Fig. 7B). The mutant progeny of complementation crosses were used for the molecular analyses summarized below, as phenotypes expressed in this material must result from disruption of the Zm-orrm1 gene. These heteroallelic mutants will be referred to hereafter as Zm-orrm1 mutants.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 7.

The maize orthologous gene to ORRM1 encodes a plastid editing factor. (A) Gene model of Zm-ORRM1 with exons (squares) and introns (lines). The two independent Mu insertions are shown as triangles in the first exon, with the 9-bp target-site duplication shown in capital letters. The insertions in exon 1 are at the indicated position with respect to the start codon. (B) Zm-orrm1-1/1–2 complementation cross progeny mutant plant is a photosynthetic mutant exhibiting a pale green phenotype (Left) compared with the wild-type (Right). (C) Immunoblot analysis of photosynthetic enzyme accumulation in Zm-orrm1-1/Zm-orrm1-2 mutants. An immunoblot of total leaf extracts (5 µg or the indicated dilutions) was probed with antibodies to the indicated proteins. The same blot stained with Ponceau S is shown below; the band corresponding to the large subunit of Rubisco (RbcL) is marked. The Zm-orrm1 mutants are siblings derived from the same complementation cross (genotype Zm-orrm1-1/Zm-orrm1-2). hcf7 is a previously described mutant with a global decrease in chloroplast translation (65). (D) Bulk-sequencing electrophoretograms of RT-PCR products from Zmorrm1 and a wild-type sibling at four plastid sites show a total loss of editing in Zm-orrm1 as no edited peak (T) is detectable. (E) PPE assay reveals a residual editing extent in Zm-orrm1 mutant plant for ndhB-586 and rpoB-545, 7% and 14%, respectively.

Defects in the major photosynthetic enzyme complexes were profiled in Zm-orrm1 mutants by quantifying one core subunit of each complex: PetD of the cytochrome b₆f complex, PsaD of photosystem I, PsbA of photosystem II, RbcL of Rubisco, and AtpB of the plastid ATP synthase (Fig. 7C). The accumulation of these proteins is known to parallel that of other closely associated subunits in the same complex. Zm-orrm1 mutants have a severe deficiency for PetD (<<10% of normal levels), a moderate deficiency for PsaD (roughly 25% of normal levels), and mild, somewhat variable reductions in RbcL, AtpB, and PsbA (40–80% of normal levels). In light of the editing defects observed in Arabidopsis orrm1 mutants, a reasonable hypothesis was that these protein deficiencies result from defects in chloroplast RNA editing.

Indeed, when all 27 known edited nucleotides in maize chloroplasts were assayed by bulk sequencing of RT-PCR products, multiple RNA editing defects were detected in the Zm-orrm1 mutant (Fig. S4). Four sites, ndhB-586, petB-668, rpoB-545, and ycf3-185 exhibited a complete loss of editing (Fig. 7D). Only five plastid sites did not show any reduction of editing extent in the Zm-orrm1 mutant (Fig. S4 and Table S2). Maize and Arabidopsis plastid transcripts share seven common editing sites so we more precisely assayed the editing extent of these sites in the Zm-orrm1 mutant by a PPE assay using fluorescent primers designed for Arabidopsis (Fig. 7E and Fig S5). The PPE assay is more sensitive than the bulk-sequencing assay, and indicates residual editing of the ndhB-586 and rpoB-545 sites in the maize mutant (Fig. 7E).

These RNA editing defects correlate well with the protein deficiencies in Zm-orrm1 mutants. The failure to edit petB-668 could account for the severe PetD deficiency, as this editing event is essential for the accumulation of the cytochrome b₆f complex (26). The reduction in ycf3-185 and ycf3-44 editing is likely to account for the loss of PsaD, as Ycf3 is required for the assembly and accumulation of photosystem I (27, 28). The minor loss of other proteins likely results from compromised chloroplast transcription and translation, possibly because of defects in the editing of rpoB, rpl20, and rps8 (encoding subunits of the plastid RNA polymerase, large ribosomal subunit, and small ribosomal subunit, respectively).

The maize and Arabidopsis orthologous ORRM1 genes evidently play similar roles in editing of the C targets that are shared between the two species. Sites such as ndhB-586 exhibit a pronounced reduction of editing extent in both Arabidopsis and maize mutants (Tables S1 and S2). In contrast, the editing extents of ndhB-1481 and rps14-80 exhibit little or no change in either the Arabidopsis or maize mutants (Tables S1 and S2).

The similarity in editing function between the maize and Arabidopsis orthologs was further analyzed by transfecting Arabidopsis mutant protoplasts with the Zm-ORRM1 under the control of either the 35S or the CMV immediate-early promoter. The CMV promoter is used in mammalian expression systems; however, apparently it is able to drive the expression of Zm-ORRM1 in Arabidopsis protoplasts (Fig. S6A). Unexpectedly, the editing extent of six plastid sites was significantly higher in mutant protoplasts transfected with the CMV construct than with the 35S promoter-containing construct (Fig. S6A). In fact, the editing of rpoA-200 was partially complemented by the CMV maize construct but not with the 35S construct carrying the maize gene (Fig. S6A). Expression of Zm-ORRM1 from the CMV promoter may be higher than the 35S promoter in these experiments. The majority of the plastid sites experienced a significant increase of editing extent in transfected mutant protoplasts independently of the promoter used in the construct (Fig. S6A).

Taken together, the similarity in the editing defects in the maize and Arabidopsis mutants and the transfection experiments with the Zm-ORRM1 strongly support a similar function in editing for the maize and Arabidopsis genes. The maize and Arabidopsis RRM domains are highly conserved (Fig. S6B), which likely explains the cross-species complementation of function.

Recombinant At-ORRM1 Binds Near Several ORRM1-Dependent Editing Sites in Vitro.

ORRM1 includes an RRM domain and so was anticipated to be an RNA-binding protein. However, ORRM1 could potentially act in one of two ways: (i) it could bind RNA nonspecifically, relying on recruitment to specific editing sites by interaction with a PPR specificity factor; or (ii) it could contribute to editing site choice by binding with specificity near its targets. To address these alternatives, we performed RNA binding assays with purified recombinant ORRM1 fused to maltose binding protein (MBP-ORRM1) (Materials and Methods). Gel mobility-shift assays were used to monitor binding to synthetic RNAs mapping between −40 and +19 with respect to the ORRM1-dependent editing sites ndhD-674, accD-1568, and matK-640. Two sequences lacking editing sites were used as negative controls: a 60-mer spanning the Arabidopsis petB 5′ UTR and a 50-mer from the psbH 5′ UTR. MBP-ORRM1 consistently bound with much higher affinity to the ndhD RNA than to the negative controls (Fig. 8 and Fig. S7). MBP-ORRM1 also bound preferentially to the accD RNA in comparison with the petB negative control, albeit with lower affinity than it bound to the ndhD RNA. However, MBP-ORRM1 exhibited only minimal preference for the matK RNA under the binding conditions explored. The boundaries of the cis-element required to specify matK-640 editing is not known, so it remains possible that ORRM1 interacts specifically with RNA outside the assayed region. Taken together, these results support the view that ORRM1 has intrinsic specificity for sequences near at least some of its RNA targets.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 8.

RNA binding activity of recombinant ORRM1. Gel mobility-shift assays were performed with MBP-ORRM1 at the indicated concentrations, together with the radiolabeled RNAs shown below (edited site underlined). The petB sequence is not edited and serves as a negative control. MBP has been shown to lack RNA binding activity under the conditions used here (66). B, bound RNA; U, unbound RNA.

ORRM1 Interacts with an Editing Recognition trans-Factor Through its Duplicated RIP Moiety.

We have previously shown that RIP1 interacts via its RIP moiety with RARE1, a PPR-DYW trans-factor that controls the editing of the plastid editing site accD-794 (19). A yeast two-hybrid (Y2H) test was performed between ORRM1 and a series of PPR-PLS trans-factors known to control the editing extent of plastid sites. We chose the PPR-PLS motif-containing proteins to be tested based on the effect of the ORRM1 mutation on the editing extent of the sites they control. CRR28 is required for the editing of ndhB-467 and ndhD-878 (16) and OTP82 is needed for editing of ndhG-50 and ndhB-836 (17). The editing extent of these sites is severely reduced to more than 90% in the orrm1 mutant (Table S1). In the Y2H analysis we also included RARE1 and OTP84, which are required for the editing of accD-794, ndhF-290, ndhB-1481, and psbZ-50, respectively (29, 30). The editing extent of accD-794, ndhF-290, ndhB-1481, and psbZ-50 in the orrm1 mutant is identical to the wild-type plant (Table S1).

CRR28 and OTP82 interacted with ORRM1 in the Y2H assay, whereas RARE1 and OTP84 did not (Fig. 9A). The lack of interaction between ORRM1 and the latter two PPR-PLS editing trans factors is expected given the absence in the orrm1 mutant of an effect on the editing extent of the sites they control (Fig. 9A and Table S1). By testing constructs with either the RIP or the RRM region of ORRM1, we determined that the RIP region is sufficient to interact with CRR28 and OTP82, whereas the RRM domain is not (Fig. 9B).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 9.

ORRM1 interacts selectively with PPR-PLS recognition trans-factors via its RIP domain in a Y2H assay. (A) Yeast colonies were able to grow in selective media (-histidine) only when ORRM1 fused to the AD and CRR28 or OTP82 fused to the BD were coexpressed into transformed yeast. (B) The interaction between ORRM1 and CRR28 or OTP82 is mediated by the RIP domain of ORRM1 as yeast colonies were able to grow in selective media (-histidine) only when the RIP region of ORRM1 fused to the AD and CRR28 or OTP82 fused to the BD were coexpressed into transformed yeast. No yeast growth was observed when the RIP-AD fusion protein was substituted by the RRM-AD fusion protein. The interaction between the RIP domain of ORRM1 and OTP82 (B, third panel from the top) is weaker than the interaction of the full lengthORRM1 (A, second panel from the top). None of the constructs used in these experiments showed autoactivation for HIS3 reporter.