|Home | About | Journals | Submit | Contact Us | Français|
The entire mitochondrial genome of rapeseed (Brassica napus L.) was sequenced and compared with that of Arabidopsis thaliana. The 221 853 bp genome contains 34 protein-coding genes, three rRNA genes and 17 tRNA genes. This gene content is almost identical to that of Arabidopsis. However the rps14 gene, which is a pseudo-gene in Arabidopsis, is intact in rapeseed. On the other hand, five tRNA genes are missing in rapeseed compared to Arabidopsis, although the set of mitochondrially encoded tRNA species is identical in the two Cruciferae. RNA editing events were systematically investigated on the basis of the sequence of the rapeseed mitochondrial genome. A total of 427 C to U conversions were identified in ORFs, which is nearly identical to the number in Arabidopsis (441 sites). The gene sequences and intron structures are mostly conserved (more than 99% similarity for protein-coding regions); however, only 358 editing sites (83% of total editings) are shared by rapeseed and Arabidopsis. Non-coding regions are mostly divergent between the two plants. One-third (about 78.7 kb) and two-thirds (about 223.8 kb) of the rapeseed and Arabidopsis mitochondrial genomes, respectively, cannot be aligned with each other and most of these regions do not show any homology to sequences registered in the DNA databases. The results of the comparative analysis between the rapeseed and Arabidopsis mitochondrial genomes suggest that higher plant mitochondria are extremely conservative with respect to coding sequences and somewhat conservative with respect to RNA editing, but that non-coding parts of plant mitochondrial DNA are extraordinarily dynamic with respect to structural changes, sequence acquisition and/or sequence loss.
Mitochondria are semiautonomous organelles whose universally recognized function is to produce cellular ATP by the process of oxidative phosphorylation. This mitochondrial function is conserved in all eukaryotic cells, i.e. animal, fungi and plant cells, but nevertheless, the mitochondrial genomes of higher plants exhibit a number of unique features compared to their counterparts in animals or fungi. Higher plant mitochondrial genomes are not only larger in size, but they also contain structural rearrangements due to homologous intra- or inter-molecular recombination events. Moreover, specific modes of gene expression (e.g. cis- and trans-splicing, RNA editing, etc.) complicate the analysis of the information encoded by plant mitochondrial genomes.
The sizes of mitochondrial genomes vary widely even among higher plant species. With 208 kb, Brassica hirta has the smallest mitochondrial genome in higher plants (1), while the mitochondrial genome of muskmelon is estimated to be over 2400 kb (2). This size variation can occur relatively rapidly in evolution as exemplified within the cucumber family, where the mitochondrial genome size varies by more than 6-fold (2). Similarly, the plant mitochondrial genome organization is known to be very dynamic. It would be of great interest to understand such rapid evolutionary changes in size and structure, and their consequences with respect to gene content. To date, the complete mitochondrial genome sequences in higher plants have been determined in two dicot plants, Arabidopsis thaliana and sugar beet (3,4) and in one monocot, rice (5). Although significant structural differences are observed among the mitochondrial genomes of these three plant species, most of the genes are conserved aside from the occasional transfer of a mitochondrial gene to the nuclear genome in one angiosperm clade or the other. However, these species are only distantly related to each other, and not suitable for investigating rapid evolutionary changes over the entire mitochondrial genome.
In this study, rapeseed (Brassica napus L.) was chosen for the analysis of the tempo and mode of evolutionary changes for several reasons. First, rapeseed belongs to Cruciferae, the same family as Arabidopsis. It occupies a good phylogenetic position for making comparisons with the already sequenced Arabidopsis mitochondrial genome. Second, physical mapping revealed that the mitochondrial genome in rapeseed is about 220 kb in size (6), which is only two-thirds of that in Arabidopsis (367 kb) (3). This implies that these two plants can be used to study the evolution of genome size. Third, it is well known that Brassica species, including rapeseed, have the smallest mitochondrial genomes among higher plants. Investigation of the rapeseed mitochondrial genome can thus reveal the minimum sequence requirement for the mitochondrial genome of higher plants.
Here the complete nucleotide sequence and RNA editing content of the mitochondrial genome in rapeseed (Brassica napus L.) are presented and compared with those of A.thaliana.
Mitochondria were isolated from green leaves of 8-week-old rapeseed plants (cv. Wester). Mitochondrial DNA (mtDNA) and mitochondrial RNA (mtRNA) were prepared using procedures described previously (7).
Based on the physical map of rapeseed mitochondria (6) and available sequences of rapeseed mitochondrial genes, 28 oligonucleotides were synthesized. Using mitochondrial DNA as a template, long PCR was carried out using these nucleotides as primers to amplify 14 parts of the mitochondrial genome. The amplified fragments were completely digested with BamHI, EcoRI and HindIII, or were partially digested with Sau3AI. A recombinant DNA library containing BamHI, EcoRI, HindIII and Sau3AI fragments in pBluescript II SK+ vector (Stratagene, USA) provided templates for sequencing. PCR-derived amplification products were also used to close gaps between restriction fragments. Each base of the rapeseed mtDNA was covered at least three times (6.8 times on average). The DNA sequencer used was a Li-COR 4200L2 (Li-COR, USA).
Sequences were evaluated and assembled using GENETYX-MAC ver11.2.3 (Software Development, Japan). A database search was done at the National Center for Biotechnology Information (NCBI) using the BLAST network service (http://www.ncbi.nlm.nih.gov/BLAST/; 8). A tRNA gene search was carried out with the tRNA scan-SE service (http://www.genetics.wustl.edu/eddy/tRNAscan-SE/; 9). The Pairwise BLAST program on our local server was used for the comparison between the whole genome sequences of rapeseed and the mitochondrial genomes of other plants.
Mitochondrial RNA was reverse transcribed with random hexamers using Superscript II reverse transcriptase (Invitrogen, USA). PCR amplification was done with primer pairs specific to each open reading frame (ORF), and the resultant PCR products were sequenced directly. Fluorescent charts for cDNA and genomic sequences were compared to determine RNA editing sites.
The rapeseed mtDNA sequence forms a circle of 221 853 bp (accession number: AP006444) (Fig. (Fig.1),1), two-thirds the size of the Arabidopsis mtDNA (366 924 bp) (3). The overall G + C content of rapeseed mtDNA (45.2%) is comparable to that of Arabidopsis mtDNA (44.8%). Homology searches using the BLASTN and BLASTX programs and a tRNA gene search using the tRNA scan-SE program detected a total of 54 genes in the rapeseed mitochondrial genome: 34 known protein-coding genes, three ribosomal RNA genes (rrn5, rrn18, rrn26), and 17 tRNA genes (Tables (Tables11 and and2),2), which altogether account for 17.4% (38 662 bp) of the genome. The positions of these genes in the rapeseed mitochondrial genome are illustrated in Figure Figure1.1. The content of protein-coding genes is completely identical between rapeseed and Arabidopsis with the exception of a gene for ribosomal protein S14 (rps14) (Table (Table1).1). Due to a stop codon and a deleted nucleotide, the rps14 gene is a pseudo-gene in Arabidopsis (10) but an intact ORF in rapeseed. In rapeseed and Arabidopsis, ccmFN genes are divided into two reading frames (ccmFN1 and ccmFN2) (3,11), a known peculiarity of the mitochondrial genomes of these two Cruciferae species, which distinguishes them from other plants (5). Seventeen tRNA genes specifying 15 species of amino acids were identified in the rapeseed mitochondrial genome (Table (Table2).2). Of these, 11 tRNAs are of mitochondrial origin and six are considered to be of plastid origin. Arabidopsis mtDNA encodes 22 tRNA genes, five more than rapeseed mtDNA. However, in both plant species the genes for at least six tRNAs (for amino acids G, A, V, L, T and R) are absent from the mitochondrial genome and these tRNAs have to be imported from the nucleus. The DNA sequences of the genes described above are highly conserved between rapeseed and Arabidopsis. Most of the protein-coding genes show more than 99% sequence similarity; the lowest similarity (97.3%) is found between rps12 genes and the highest (99.9%) between rpl16 genes.
In addition to previously identified genes, 45 ORFs larger than 100 codons in size were annotated in the rapeseed mitochondrial genome. None of these ORFs could be assigned a function based on sequence similarity at either the nucleotide or protein level. Moreover, none of these rapeseed ORFs has a counterpart in the Arabidopsis mitochondrial genome, in which 85 ORFs larger than 100 codons were described, although limited similarity is observed between some ORFs of rapeseed and Arabidopsis. Also, most ORFs do not show any significant homology to other genes found in the nuclear or plastid genome, although some have similarity to nuclear transposon or plastid sequences. However, all of these ORFs of possible transposon or plastid origin seem to be non-functional because of their fragmented and/or truncated nature.
In summary, despite their large size difference, the mtDNAs of the two species share nearly the same set of functional genes (protein genes, rRNA genes and tRNA genes). The additional ORFs are not shared between these two closely related plants, which means that these ORFs could not code for additional important protein information and may instead, although it seems unlikely, fulfil a species-specific function.
In this study, a 2427 bp sequence was found to be present as a direct repeat in the rapeseed mitochondrial genome, as reported previously (6). This sequence included the first exon, the intron, and part of the second exon of the cox2 gene (Fig. (Fig.1).1). Due to this duplication, two copies of cox2 genes exist in rapeseed, although these copies diverge from each other at the point 55 bp upstream of the stop codon. One copy (cox2-1 in Fig. Fig.1)1) is homologous to other plant mitochondrial cox2 genes, but the other (cox2-2 in Fig. Fig.1)1) has an extension that shows no homology to any other sequences. The presence of these 2427 bp repeats implies a tricircular structure for the rapeseed mitochondrial genome, as postulated before (6). Via these 2427 bp repeats, the rapeseed mitochondrial genome could recombine into two subgenomic circles (124 908 and 96 945 bp circles).
Large repeat sequences are also found in the Arabidopsis and sugar beet mitochondrial genomes (6.5 and 4.2 kb, and 6.2 kb, respectively; 3,4), and have been suggested to be active in intramolecular recombination. However, the sequences of these repeats (atp6 and orf139 regions in Arabidopsis, rrn26 region in sugar beet, and cox2 region in rapeseed) are completely different from each other. Intramolecular recombination to generate the complexity of the plant mitochondrial genome is a common feature of higher plant mitochondrial genomes (12), except for one known example, B.hirta (1). However, the DNA sequences involved in the intramolecular recombination are highly species-specific, and different sequences act as recombination repeats even within related species such as rapeseed and Arabidopsis.
Using the Pairwise BLAST software, the whole sequence of the rapeseed mitochondrial genome was compared with those of the Arabidopsis mitochondrial genome (3), sugar beet mitochondrial genome (4), Arabidopsis plastid genome (13) and rapeseed mitochondrial linear plasmid (14). The E (expectation) value threshold was set at 1e–10 for these comparative analyses among rapeseed and other genomes.
When the mitochondrial genome of rapeseed is compared to that of Arabidopsis, the sequences shared by these two related species total 143 126 bp, which represents 64.5% of the rapeseed genome. On the other hand, the sequences shared between rapeseed and sugar beet total 64 972 bp, representing 29.3% of the rapeseed genome. It has been reported that 78 057 bp of the mtDNA sequences are shared between Arabidopsis and sugar beet (4). These values are in good agreement with the phylogenetic relationship among these three plant species: rapeseed and Arabidopsis belong to the same family, Cruciferae, while sugar beet is a member of another family, Chenopodiaceae. However, one-third (about 78.7 kb) and two-thirds (about 223.8 kb) of the rapeseed and Arabidopsis mitochondrial genomes, respectively, were not aligned with each other. These non-homologous sequences are scattered through the rapeseed mitochondrial genome, and a total of 31 such regions longer than 1000 bp were identified. These data suggest that after divergence of these two species, in addition to the multiple recombination events, the mitochondrial genomes of both plants were very quickly modified by sequence acquisition and/or sequence loss encompassing up to two-thirds of the genome in the case of Arabidopsis. The large size variation between these two plants might be due partly to such rapid processes of sequence acquisition and/or sequence loss.
A BLAST search was carried out for the 31 non-homologous regions longer than 1 kb (a total of 51 517 bp) to explore the origins of these sequences. Of these sequences, 13.2% showed similarity to sequences in other plant mitochondrial genomes (not to Arabidopsis sequences) and 5.9% appeared to be of plastid origin. The fraction of sequences showing similarity to nuclear genome sequences was only 0.3%. The remaining sequences (80.6%) did not show any similarity to the registered sequences in the DNA sequence databases. Identifying the origin of these non-homologous sequences is of great importance and interest for understanding the evolution of higher plant mitochondrial genomes.
The mtDNA sequences shared by three dicot plants, rapeseed, Arabidopsis and sugar beet, totaled 61 870 bp, which corresponds to 27.9% of the rapeseed genome. This value is comparable to the sum total (68 343 bp) of coding sequences (38 662 bp) and cis-splicing introns (29 681 bp) of the rapeseed mitochondrial genome. These data suggest that the gene-coding sequences may be highly conserved among the mitochondrial genomes of higher plants, but other regions, i.e. non-coding sequences, may have diverged very rapidly during evolution, as they can no longer be aligned with each other.
In the rapeseed mitochondrial genome, 14 stretches of plastid-like sequences ranging from 43 to 2181 bp were identified. These sequences totaled 7950 bp, comprising 3.6% of the mitochondrial genome. In Arabidopsis, about 1% of the mitochondrial genome contains sequences imported from the plastid. In rapeseed, nine of the 14 plastid sequences found contain tRNA genes, seven of which are well conserved between the two plants. The fragments of rbcL and rpoB are also conserved in both plant mitochondria. Thus, these sequences may have been transferred from the plastid genome to the mitochondrial genome before the divergence of rapeseed and Arabidopsis. However, the rapeseed mitochondrial genome contains pieces of the psaA and ycf2 genes, which are not found in Arabidopsis mtDNA. On the other hand, the pieces of the psbD and ndhB genes in the Arabidopsis mitochondrial genome are not present in rapeseed mtDNA. These results also suggest that functional sequences may be highly conserved, but sequence acquisition and/or sequence loss may be very frequent for non-functional gene pieces. Indeed, plastid tRNA genes in the mitochondrial genomes are intact and most likely functional, but protein-coding gene pieces seem to have no function in the mitochondrial genome of either plant.
It has been reported that a rapeseed mitochondrial linear plasmid had no sequence similarity to the mitochondrial genome (14). However, a 19 bp sequence, CTCTYCTTTCAGTYGAGTT, was found to be common to these two DNA molecules. This sequence is located at three positions in the upstream regions of ORF2, 3 and 4 of the linear plasmid but is scattered at 41 positions in the rapeseed mitochondrial genome. Since there are no known functions (promoter, binding, etc.) for this sequence motif, it is unclear what role it has in rapeseed mitochondria. Interestingly, this motif is found 22 times in the Arabidopsis mitochondrial genome, although there have been no reports showing the presence of linear plasmids in Arabidopsis mitochondria. On the other hand, the sugar beet mitochondrial genome contains no such sequence motif. This sequence may be related to the co-evolution of the mitochondrial linear plasmid and mitochondrial genome specific to the Cruciferae species.
By comparing sequences surrounding transcription start sites of plant mitochondria, the conserved nonanucleotide motif (CNM) has been identified in dicotyledonous plant species (15). Since no experimental information about the 5′ termini of primary transcripts is available for rapeseed, identification of potential promoter regions was carried out by screening of the rapeseed mitochondrial genome sequence with the CNM motif. In this search, only six CNM sequence variations were used, following the criteria of Dombrowski et al. (16).
Twenty-nine CNM motifs were found in the rapeseed mitochondrial genome. The number of potential promoters was reduced to 21 by specifying other characteristics of promoter sequences of dicot plants: the presence of an AT-box upstream of CNM and at least one purine at nucleotide positions +3 and +4. Fifteen promoters are located within a 1 kb region upstream of known genes and ORFs. These potential promoters may promote the transcription of the coding information in vivo. For instance, two of these potential promoters are found in the upstream region of the 18S rRNA genes (rrn18). In Arabidopsis, the transcription initiation site upstream of rrn18 was experimentally investigated by primer extension analysis, and the CNM was found to reside at the transcription initiation site (17). Potential promoter sequences of rapeseed rrn18 were identified at the same position in the homologous region of Arabidopsis rrn18, except that the rapeseed sequence is duplicated. In total, seven potential promoters upstream of identified genes, including atp8 [formerly named orfB; (18)], atp9, rrn26 and some tRNA genes, are located at the same positions as in the homologous regions of Arabidopsis. These genes may be transcribed in the same manner as in Arabidopsis using the same promoters. However, the four potential promoters for rps7, rpl2 and two tRNA genes are either absent or present at a different position in the corresponding regions of the Arabidopsis mitochondrial genome. For these genes, there may be different transcriptional controls between rapeseed and Arabidopsis.
Some of the genes without a CNM in their 5′ UTR show only limited similarity to the corresponding regions of the Arabidopsis mitochondrial genome. For example, the rapeseed nad6 and atp1 5′ UTR sequences diverged 2 and 53 bp upstream of the initiation codon, respectively, compared with the Arabidopsis sequences. The transcription of the nad6 and atp1 genes is likely to be initiated from different non-CNM promoters in the two plants. The actual number of promoters in the rapeseed mitochondrial genome should be investigated experimentally.
Before translation, mitochondrial messages in higher plant mitochondria may undergo RNA splicing and/or RNA editing, two post-transcriptional processing events that may critically alter the coding message.
In the rapeseed mitochondrial genome, 24 introns ranging in size between 839 and 3448 bp were identified. All of these introns belong exclusively to group II, as in other higher plants. Nineteen introns were cis-spliced sequences, and comprised 29 681 bp or 13.4% of the rapeseed mitochondrial genome. A total of five trans-spliced introns were found in the three nad genes, nad1, nad2, nad5. The intron numbers and positions are completely identical to those of Arabidopsis mitochondrial genome (3). Intron structures are also highly conserved between these two species.
RNA editing leads to change of a C to a U residue, sometimes creating an initiation or termination codon, but more often creating an internal codon with strong functional relevance. The editotype of Arabidopsis, that is, all editing sites found in the systematic transcriptome analysis of Arabidopsis, has been determined (19) and shown to consist of 441 sites scattered throughout the mitochondrial genome. There are 34 known protein-coding genes in rapeseed mitochondria as stated above. In total, 24 genes and nine ORFs longer than 150 amino acids were subjected to the editing analysis in this study. The remaining 10 genes had already been analyzed for RNA editing (Table (Table3).3). In addition, editing of orf222, the homologue of orf224 associated with Polima male sterile cytoplasm, was detected (20).
In summary, 427 C to U RNA editing modifications were detected in genes and ORFs of rapeseed mitochondrial transcripts (Table (Table3).3). None of nine unidentified ORFs examined had editing sites in their transcripts, except for one partial editing event in the orf188 transcript. The total number of editing sites (427 sites) in rapeseed is nearly identical to that (441 sites) in Arabidopsis (19). However, the number of editing sites shared by both plant mitochondria was only 358, which correspond to 83.8% and 81.2% of the total editing sites in rapeseed and Arabidopsis transcripts, respectively (Table (Table3).3). These percentages seem to be quite low considering that the average similarity in the primary DNA sequences between rapeseed and Arabidopsis is 99.2% for protein coding regions. This means that the tempo of diversification of RNA editing is more rapid than that of coding information.
Two RNA editing sites are located within the overlap region of rps3 and rpl16. The second of these two edits creates a stop codon 21 amino acids downstream from the rpl16 ATG codon. This RNA editing event is also observed in the Petunia and Arabidopsis rpl16 transcripts (21,22), which suggests that rpl16 gene might be a pseudo-gene. However, no ATG codon is encoded in the rpl16 reading frames of Oenothera and Marchantia mitochondria, and a valine GTG codon is postulated to be used for translation initiation (23,24). This GTG codon is conserved in rapeseed mitochondria, 15 amino acids downstream from the stop codon introduced by the editing event. If GTG is used for translation initiation, RNA editing would have no effect on the translation of the rpl16 gene.
The rapeseed rps14 gene transcript is not edited. The rps14 gene is one example of the gene-transfer events from the mitochondrion to the nucleus often found in plants. In fact, rps14 is located in the mitochondrial genomes of broad bean, Oenothera and pea (25–27). In contrast rps14 has been translocated to the nucleus in Arabidopsis, maize and rice (28–30). Two RNA editing events were observed in Oenothera rps14 transcripts (26). The first editing event altered a codon from TCC (Ser) to TTC (Phe), and the second one modified a CCT (Pro) triplet to TCT (Ser). Arabidopsis Ψrps14 transcripts have retained the first editing site in the Oenothera transcript (26). Alignment of the deduced amino acid sequences from rapeseed and other plant rps14 genes allowed us to determine that the second editing event in Oenothera occurred at a non-conserved position (Fig. (Fig.2).2). However the first RNA editing event, leading to the change of a Ser residue to a conserved Phe, seems to be essential for producing mature RPS14 protein (Fig. (Fig.2).2). It is possible that the rapeseed mitochondrial copy of rps14 might be non-functional, although it is an intact reading frame and is transcribed. Further work is needed to examine whether a functional rps14 gene exists in the rapeseed nuclear genome.
During the systematic analysis of RNA editing sites in rapeseed mitochondria, it was found that ccmC cDNA could not be amplified using the primer pair ccmC-F and ccmC-R, which are located before the start codon and after the stop codon of ccmC, respectively (Fig. (Fig.3A3A for the primer positions and Fig. Fig.3B3B for the PCR products). Two other 3′-primers, ccmC-R1 and ccmC-R2 were also tested (Fig. (Fig.3A).3A). Only one primer pair, ccmC-F and ccmC-R1, could amplify ccmC cDNAs (Fig. (Fig.3B).3B). These results suggest that the ccmC transcripts are interrupted within the 50 bp-sequence before the stop codon. Twenty-five RNA editing sites were identified in the truncated ccmC transcript (ccmC-R1-primed cDNA); however, no editing creating a stop codon has been detected. This situation is also observed in Arabidopsis ccmC transcripts. The end of the ccmC transcript in Arabidopsis was mapped between primers ccmC-R1 and ccmC-R2 (Fig. (Fig.3A)3A) (S.Binder and P.Giegé, personal communications). There are no data demonstrating the presence of a nuclear copy of the ccmC gene or the presence of the CcmC protein itself in rapeseed mitochondria at present. However, the present finding raises the question of the existence of CcmC and its mode of translation. Previously we reported that the structure of the ccmFN gene in rapeseed mitochondria was quite unusual (11). The ccmFN sequence is divided into two parts, which are about 74 kb apart (ccmFN1 and ccmFN2, see Fig. Fig.1).1). These two parts are transcribed and their transcripts are edited in a similar manner as the homologous transcripts of wheat (31). However, no mature transcript covering the whole coding region has been detected, a result which excludes a trans-splicing event. No other copy of this gene was found in either the nuclear or mitochondrial genome. Nevertheless, the protein product of ccmFN was identified in rapeseed mitochondria by western blot analysis. The mode of expression of these CcmFN and CcmC proteins in rapeseed remains mysterious.
The comparative analysis presented here allows a more comprehensive understanding of mitochondrial genome evolution in higher plants. It had previously been noted that plant mitochondrial DNA evolves rapidly in structure, but slowly in sequence (6), which was concluded based on comparisons among Brassica species. Palmer and Herbon found numerous internal rearrangements of mitochondrial genomes by homologous recombination events through short dispersed repeats as the points of crossover. This picture is still true in part. In this study, rapeseed and Arabidopsis were shown to have nearly the same sets of genes for mitochondrial proteins, rRNAs and tRNAs, although the gene orders are different. The intron contents and locations are also completely conserved in the two plants. These results support the idea that internal genome rearrangement is one of the major driving forces for the mitochondrial genome evolution of higher plants.
However, this study showed that large parts of mitochondrial genomes in higher plants are species-specific and show no homology to other plant mitochondrial genomes. One-third and two-thirds of the rapeseed and Arabidopsis mitochondrial genomes, respectively, were not aligned with each other. This means that after divergence of these species, the mitochondrial genomes of both plants were very quickly modified by sequence acquisition and/or sequence loss, accounting for up to two-thirds of the genome in the case of Arabidopsis. At this level, rapid sequence transfer, sequence acquisition and/or sequence loss is thought to have been a more important factor in the evolution than internal genome rearrangement. Parts of these non-alignable sequences originated from plastid or nuclear genomes, but most of them show no homology to registered sequences. It will be necessary to clarify the origin of these sequences and how they were transferred to mitochondria. Furthermore, the mtDNA sequences shared by three dicot plants, rapeseed, Arabidopsis and sugar beet, are limited to the coding sequences and introns. These data suggest that the plant mitochondrial genome is very dynamic and fluid, except for the primary gene information.
Some potential promoter regions are conserved between rapeseed and Arabidopsis, but the upstream regions of some genes are completely divergent from each other. On the other hand, RNA editing sites shared by both plant mitochondria constitute about 80% of the total editing sites. The remaining 20% of editing sites are species-specific. This percentage is quite high considering that the primary DNA sequences for the protein-coding regions are highly conserved between rapeseed and Arabidopsis. These data suggest that the evolutionary speed is higher at the level of gene regulation than at the primary gene sequence level.
An additional notable finding is that the transcripts of the ccmC gene are truncated before the stop codon. Previously we reported the unusual structure and expression of the ccmFN gene in rapeseed mitochondria. Although there are no precise data about the expression of CcmC protein in rapeseed mitochondria so far, the present finding raises another question about expression of Ccm proteins in rapeseed.
The author wishes to thank Ms Kazuko Hashimoto and Kyoko Minagawa for their skilful technical assistance. The author is greatly indebted to Dr Géraldine Bonnard for helpful advice and for help in the revision of the manuscript. Appreciation is expressed to Drs Stefan Binder and Philippe Giége for providing unpublished data.
DDBJ/EMBL/GenBank accession no. AP006444