Quantitative trait locus mapping identifies REME2, a PPR-DYW protein required for editing of specific C targets in Arabidopsis mitochondria

Abstract

Targeted RNA editing by C-to-U alteration occurs at hundreds of sites in the mitochondrial transcriptome of flowering plants. By using natural variation and positional cloning on a population of Arabidopsis recombinant inbred lines between the ecotypes Col and Ler, we found that two of these occurrences involve the Arabidopsis PPR-DYW protein REME2 (Required for Efficiency of Mitochondrial Editing2). The analysis of a knockdown mutant along with silenced tissues confirms the specificity of REME2 for both sites located in mitochondrial ribosomal protein genes (rps3-1534 and rps4-175). The conservation level of both cis elements is relatively high, as is the amino acid conservation among flowering plants for both genes in that location, underlining the importance of these editing events and REME2.

Keywords: QTL mapping; RPS3; RPS4; plant mitochondria; virus-induced gene silencing.

Figure 1. Mapping strategy used to identify the PPR-DYW protein encoded by At4g15720 as a possible mitochondrial editing factor. (A) Two Arabidopsis accessions, Ler and Col, exhibit a differential editing extent at two mitochondrial sites, rps3-1534 and rps4-175. Poisoned Primer Extension (PPE) gels show a more intense unedited (U) band for Col than for Ler at the two sites assayed. The percentage of editing [edited (E)/ edited (E) + unedited (U)] is given below each lane. (B) QTL mapping identifies an editing QTL that controls the editing of rps3-1534 and rps4-175 on chromosome 4 in the interval between markers 7900K and 9500K. The curves for the likelihood odds ratio (LOD) scores, a statistic used to detect a QTL, are very similar for the two editing traits analyzed; in particular the maximum LOD scores, above the significance threshold of 3, are located in the same interval, 7900K-9500K, for both rps3-1524 and rps4-175. (C) Three recombinant inbred lines (RILs) that have recombined genomes in the QTL interval allow the further localization of the editing factor between 8810K and 9160K. (D) Among the 105 genes lying in the 8810K-9160K interval, only one, At4g15720, encodes a PPR protein which belongs to the DYW subclass. The localization of the T-DNA used in this study is shown on the gene model in the 3′UTR, 85 nucleotides downstream of the stop codon. The modular structure of the predicted protein encoded by At4g15720 with the type of PPR (P, S and L) symbolized by squares, is represented below the gene model. The 190-nucleotide region of At4g15720 RNA targeted by VIGS is delimited by facing arrows. The six amino acids substitutions occurring between REME2 encoded by Col and Ler are indicated (Col → Ler).

Figure 2. Analysis of rps3-1534 and rps4-175 editing extent in a mutant and in tissue silenced for REME2 demonstrate a specific control of these two sites by REME2. A T-DNA insertional mutant in the REME2 3′UTR exhibits a significant reduction of editing extent for rps3-1534 (A) and rps4-175 (B). Upper panels, PPE products separated on an acrylamide gel; E, edited; U, unedited; −/−, T-DNA mutant; +, wild-type. Below, the PPE gels are shown the quantification of editing extent based on the intensity of the E and U bands. Each bar corresponds to the editing extent based on the products of the lane directly above: gray bar, mutant; white bar, wild-type. On the right of the graphs are shown the averages with error bars (s.d.). Silencing of REME2 by VIGS results in a significant decrease of editing extent of rps3-1534 (C) and rps4-175 (D) in silenced plants. Upper, PPE gel; E, edited; U, unedited; control, uninoculated plants; silenced, REME2-silenced plants. As for (A and B), the graphs below the PPE gel show the quantification of editing extent corresponding to each lane: gray bar, silenced; white bar, control. Analysis of sites not controlled by REME2 does not reveal any reduction of their editing extent in reme2 or in the REME2-silenced plant. (E) PPE gel shows a similar editing extent in reme2 and the wild-type for ccb206-406. (F) RT-PCR bulk sequencing demonstrates the specificity of REME2 on rps4-175 editing extent as other sites on the same transcript do not exhibit any difference between the wild-type (+) and reme2 mutant (−/−) or between the control and the silenced plant.

Figure 3. REME2 is targeted to mitochondria. The left panel shows the subcellular localization in onion epidermal cells of REME2 fused to GFP while the right panel is the mitoCherry fluorescent marker localized to the mitochondria. The middle panel is a merged figure showing co-localization of both fluorescent markers.

Figure 4. Analysis of the putative cis elements of the RNA targets recognized by REME2. (A) Nucleotide alignment of the putative cis element (-20 +5) of the two editing sites under the control of REME2, and those of sites under the control of three other PPR trans-factors. The target C for editing is underlined. Conserved nucleotides are highlighted in gray. The names of editing sites (gene-position) are given on the right of each sequence, along with the corresponding PPR trans-factors. (B) Predicted RNA nucleotides binding to REME2 according to the recognition code. The first row shows the modular structure of REME2 in relation to the nature of the PPR motifs, S, P or L. Row 6 indicates the identity of the residues found at position 6 of each PPR motif, while row 1’ shows the residue found at position 1 of the next PPR motif. The expected nucleotides binding to REME2 are predicted based on the identities of the residues at positions 6 and 1’ in PPR motifs S and P only. The targeted C for editing is shown in magenta and nucleotides in the rp4-175 putative cis element that do not fit the recognition model are highlighted by a magenta box.

Figure 5. Alignment of the E+ motifs found in organelle editing trans factors. The amino-acid substitution between the Col and Ler REME2 proteins is circled. All known PPR protein editing factors containing an E+ domain are aligned.

Figure 6. Amino acid conservation between flowering species for the region around rps3-1534 and rps4-175. The amino acid C or S for rps3-1534 and rps4-175, respectively, resulting from editing of C targets by REME2, are highlighted in gray. The corresponding amino acid from species where editing is not necessary because the T is genomically encoded are not highlighted. An arrow indicates the position of the edited C in both transcripts for the Arabidopsis DNA sequence displayed above the amino acid alignment.