Abstract

The mtDNA of Cycas taitungensis is a circular molecule of 414,903 bp, making it 2- to 6-fold larger than the known mtDNAs of charophytes and bryophytes, but similar to the average of 7 elucidated angiosperm mtDNAs. It is characterized by abundant RNA editing sites (1,084), more than twice the number found in the angiosperm mtDNAs. The A + T content of Cycas mtDNA is 53.1%, the lowest among known land plants. About 5% of the Cycas mtDNA is composed of a novel family of mobile elements, which we designated as “Bpu sequences.” They share a consensus sequence of 36 bp with 2 terminal direct repeats (AAGG) and a recognition site for the Bpu 10I restriction endonuclease (CCTGAAGC). Comparison of the Cycas mtDNA with other plant mtDNAs revealed many new insights into the biology and evolution of land plant mtDNAs. For example, the noncoding sequences in mtDNAs have drastically expanded as land plants have evolved, with abrupt increases appearing in the bryophytes, and then in the seed plants. As a result, the genomic organizations of seed plant mtDNAs are much less compact than in other plants. Also, the Cycas mtDNA appears to have been exempted from the frequent gene loss observed in angiosperm mtDNAs. Similar to the angiosperms, the 3 Cycas genes nad1, nad2, and nad5 are disrupted by 5 group II intron squences, which have brought the genes into trans-splicing arrangements. The evolutionary origin and invasion/duplication mechanism of the Bpu sequences in Cycas mtDNA are hypothesized and discussed.

Introduction

Presently, the complete mitochondrial genomes (mtDNAs) of land plants are only known for 1 liverwort (Marchantia polymorpha, Oda et al. 1992), 1 moss (Physcomitrella patents, Terasawa et al. 2007), and 7 angiosperms, including 4 eudicots (Arabidopsis thaliana, Unseld et al. 1997); sugar beet (Beta vulgaris, Kubo et al. 2000); rapeseed (Brassica napus, Handa 2003); Tobacco (Nicotiana tabacum, Sugiyama et al. 2005), and 3 monocot crops (rice [Oryza sativa], Notsu et al. 2002) maize (Zea mays subsp. mays, Clifton et al. 2004); wheat (Triticum aestivum, Ogihara et al. 2005) (table 1). The evolution of mtDNAs is more complex than that of plastid genomes (cpDNAs) because mtDNAs have horizontal gene transfer, uptake of plastid and nuclear DNA sequences, extreme size and rate variations, rapid gene rearrangements, and more trans-splicing exons and RNA editing sites (Bergthorsson et al. 2004; Knoop 2004; Sugiyama et al. 2005; Wang et al. 2007). However, to date, no comparative mtDNA study has been performed in the 1,000+ species of gymnosperms, which are considered to be “inextricably related to the origins and early divergence of seed plants” (Chaw et al. 1997). In an effort to expand our understanding of mtDNA organization and evolution, we herein determined and examined the first gymnosperm mtDNA from Taitung cycad (Cycas taitungensis), an endangered species restricted to the mountain valleys of eastern Taiwan.

Table 1

MtDNAs of the 11 Plants Examined in This Study

Classification Scientific Name Common Name mtDNA Accession Number Reference 
Algae 
    Charophyta 
 Chara vulgaris Green algae AY267353 Turmel et al. (2003) 
Land plants 
    Bryophyta 
 Physcomitrella patens Moss NC_007945 Terasawa et al. (2007) 
 Marchantia polymorpha Liverwort NC_001660 Oda et al. (1992) 
Seed plants 
    Gymnosperm 
 Cycas taitungensis Taitung cycad BABI01000001 Present study 
    Angiosperm 
    Monocots Triticum aestivum Wheat NC_007579 Ogihara et al. (2005) 
Oryza sativa sp. Japonica Rice BA000029 Notsu et al. (2002) 
Zea mays Maize NC_007982 Clifton et al. (2004) 
    Dicots Beta vulgaris L. Sugar beet AP000396-7 Kubo et al. (2000) 
Nicotiana tabacum Tobacco NC_006581 Sugiyama et al. (2005) 
Brassica napus Oilseed rape AP006444 Handa (2003) 
Arabidopsis thaliana Mouse-ear cress NC_001284 Unseld et al. (1997) 
Classification Scientific Name Common Name mtDNA Accession Number Reference 
Algae 
    Charophyta 
 Chara vulgaris Green algae AY267353 Turmel et al. (2003) 
Land plants 
    Bryophyta 
 Physcomitrella patens Moss NC_007945 Terasawa et al. (2007) 
 Marchantia polymorpha Liverwort NC_001660 Oda et al. (1992) 
Seed plants 
    Gymnosperm 
 Cycas taitungensis Taitung cycad BABI01000001 Present study 
    Angiosperm 
    Monocots Triticum aestivum Wheat NC_007579 Ogihara et al. (2005) 
Oryza sativa sp. Japonica Rice BA000029 Notsu et al. (2002) 
Zea mays Maize NC_007982 Clifton et al. (2004) 
    Dicots Beta vulgaris L. Sugar beet AP000396-7 Kubo et al. (2000) 
Nicotiana tabacum Tobacco NC_006581 Sugiyama et al. (2005) 
Brassica napus Oilseed rape AP006444 Handa (2003) 
Arabidopsis thaliana Mouse-ear cress NC_001284 Unseld et al. (1997) 

Cycads (Cycadales) appeared in the Pennsylvanian era, approximately 300 million years ago (MYA) and dominated the Mesozoic forests along with conifers and ginkgos. The extant cycads include 2 or 3 families with some 300 species in 10 genera (Chaw et al. 2005). They are largely confined to the tropics and subtropics in both the Old and New Worlds, except for the genus Cycas (the only genus in Cycadaceae), which has the widest distribution as far north as Japan (Jones 2002). Cycads are regarded as being closely linked with spore-producing ferns because their young leaves are circinate, their trunks lack axillary buds but show unique girdling leaf traces and dichotomous branching (vs. the axillary branching seen in other seed plants), their pollen tubes possess multiciliate sperms, and their ovules are borne on the margins of leaf-like megasporophylls (Stevenson 1990). However, recent phylogenetic analysis and comparative chloroplast genomics (e.g., Chaw et al. 1997; Wu et al. 2007) suggested that the cycads and Ginkgo are sister groups, reflecting the above characters should be regarded as plesiomorphic rather than apormorphic.

Here we report the complete mtDNA sequence of Cycas taitungensis and its surprising organization. Notably, it contains abundant short interspersed repetitive elements and RNA editing sites. Compared with the known mtDNAs from other land plants, that of Cycas has a particularly low A + T content, fewer gene losses, and no large repeats >2 kb. Moreover, we describe a novel family of mobile elements, herein termed Bpu sequences/elements, more than 500 variants of which are distributed across the Cycas mtDNA. Our evolutionary analysis further reveals that, within seed plants, the mtDNA of Cycas shows higher substitution rates for protein-coding genes than in other known plants. The evolutionary origin and invasion/duplication mechanisms in Cycas mtDNA are hypothesized and discussed.

Materials and Methods

Determination of the Complete mtDNA Sequence of C. taitungensis

Young leaves (less than 10-day-old) were collected from an 8-year-old C. taitungensis tree grown in the greenhouse of the Academia Sinica. Intact mitochondria were isolated using the method described by Kadowaki et al. (1996). Dnase I was used to digest any nuclear genome (nrDNA) contaminants. mtDNA was isolated according to a cetyltrimethylammonium bromide-based protocol (Stewart and Via 1993), sheared into random fragments of ∼2–3 kb with a Hydroshear device (Genomic Solutions Inc., Ann Arbor, MI), and then directly cloned into the EcoRV site of the pBluescriptSK vector to generate a shotgun library. Shotgun clones were sequenced as previously described (Wang et al. 2007), except that each nucleotide had about 10× coverage. Gaps were filled with specific primers designed based on the sequenced clones. The complete Cycas mtDNA sequence has been deposited in DNA Data Bank of Japan, under NCBI accession number BABI01000001.

Sequence Data Analysis

Annotation of Protein-Coding, rRNA, and tRNA Genes

A database search was carried out by using the National Center for Biotechnology Information's (NCBI) Web-based Blast service (http://www.ncbi.nlm.nih.gov/BLAST/), and genes with e-values smaller than 0.001 were selected. The exact gene and exon boundaries were determined by alignment of homologous genes from available and annotated plant mtDNAs (table 1). Multiple sequence alignments were performed using the MAP2 Web service (http://deepc2.psi.iastate.edu/aat/map2/map2.html). Alignments of both nucleotide sequences and the translated amino acid sequences of the protein-coding genes were manually inspected. The tRNA genes were annotated using the tRNAscan-SE program (Lowe and Eddy 1997).

ORF Finding and Intron Identification

Identification of open reading frames (ORFs) was performed with the Web-based NCBI ORF Finder (http://www.ncbi.nlm.nih.gov/projects/gorf/). The standard genetic code was applied. The intron types were identified on the basis of their sequences and secondary structures (Michel and Ferat 1995). The predicted introns were also verified by alignment of their sequences with orthologous sequences previously elucidated from other species.

RNA Editing Sites

Putative RNA editing sites in protein-coding genes were predicted using the PREP-mt Web-based program (http://www.prep-mt.net/) (Mower 2005). To achieve a balanced trade-off between the number of false positive and false negative sites, the cutoff score (C-value) was set to 0.6 as suggested by the author. In addition, the top 4 genes with highest number of editing sites were verified by reverse transcription–polymerase chain reaction (RT–PCR) with gene-specific primers (supplementary table S7, Supplementary Material online).

Analysis of Repeats

Identification of repeats was carried out using the REPuter Web-based interface (http://bibiserv.techfak.uni-bielefeld.de/reputer/) (Kurtz et al. 2001). Both sequence directions (forward and reverse complement) were searched. The number of maximum computed repeats was set to 5,000. Overlapped repeating sequences were manually removed from each result. Information on tandem repeats was obtained using the tandem repeats finder (http://tandem.bu.edu/trf/trf.html) (Benson 1999). Searches for repeated regions and tandem repeats were carried out on all elucidated plant mtDNAs.

Phylogenetic Analysis

The 22 protein-coding genes common to the 11 sampled mtDNAs (atp1, atp4, atp6, atp8, atp9, ccmB, cob, cox1, cox2, cox3, nad1, nad2, nad3, nad4, nad4L, nad5, nad6, nad9, mttB, rps3, rps4, and rps12) were extracted for phylogenetic analysis. The sequences were separately aligned and then concatenated. Gaps and stop codons were removed manually. Divergence of nucleotide sequence between each pair of taxa was estimated in terms of the numbers of substitutions per synonymous (Ks) or nonsynonymous (Ka) site, based on the Pamilo–Bianchi–Li method implemented in MEGA 3.0 program (Kumar et al. 2004). The Neighbor-Joining trees reconstructed with the Ka values and Ks values were rooted at Chara. The number of bootstrap replicates was set to 500. All phylogenetic analyses and tree reconstructions were performed using MEGA 3.0.

Results and Discussion

Evolution of mtDNA Organization in Land Plants

Characteristics of Cycas mtDNA and Insights into the Evolution of Land Plant mtDNAs

The complete mtDNA of C. taitungensis is a circular molecule of 414,903 bp (fig. 1). Table 2, which compares the main features of mtDNAs from a charophyte (Chara vulgaris), 2 bryophytes (Marchantia and Physcomitrella), and 7 angiosperms, shows that the Cycas mtDNA is about 6-, 2.2-, and 4.0-fold larger than those of Chara, Marchantia, and Physcomitrella, respectively, but does not significantly differ from the average (414 ± 102 kb) of the previously elucidated angiosperm mtDNAs (P = 0.502). The A + T content of the Cycas mtDNA is 53.1%, the lowest among known algae and land plants. As further shown in table 2, the total numbers of protein- and tRNA-coding genes decrease from charophytes (39 and 26, respectively) to seed plants (29–40 and 17–27, respectively) (detailed in supplementary table S2, Supplementary Material online). In contrast, the numbers of rRNA gene species remain the same in all lineages, with the exception of obvious gene duplications in the Poaceae (grass family) and Beta lineages, in which some tRNA genes are also duplicated.

Table 2

Comparison of Features among 11 Elucidated Land Plant Mitochondrial Genomes

 Taxon 
Features Chara Marchantia Physcomitrella Cycas Triticum Oryza Zea Beta Nicotiana Brassica Arabidopsis 
Size (bp) 67,737 186,609 105,340 414,903 452,528 490,520 569,630 368,799 430,597 221,853 366,924 
A + T content (%) 59.1 57.6 59.4 53.1 55.7 56.2 56.1 56.1 55.0 54.8 55.2 
Gene numbera (protein-coding/tRNA/rRNA) 39/26/3 {39/26/3} 41/27/3 {41/29/3} 39/24/3 {39/24/3} 39/22/3 {39/26/3} 33/17/3 {35/25/8} 35/18/3 {40/27/6} 32/17/3b {34/22/5b} 29/18/32 {9/22/5} 36/21/3 {39/22/4} 33/17/3 {34/17/3} 32/17/3 {33/22/3} 
Coding sequences (%)a 90.7 20.3 37.0 10.1 8.6 11.1 6.2 10.3 9.9 17.3 10.6 
Repeat sequences (%)c 3.2 10.1 11.6 15.1 10.1 28.8 11.4 12.5 10.8 5.5 7.8 
Tandem repeat sequences (%)d 0.22 0.49 0.14 4.97 0.42 0.09 0.35 0.71 0.05 0.36 0.34 
Introns 
    Group I 14 
    Group II 
    Cis-spliced 13 25 25 20 17 17 15 14 17 18 18 
    Trans-spliced 
Cp-derived sequences (bp) —e —e —e 18,113 13,455 22,593 25,132 —e 9,942 7,950 3,958 
(%) —e —e —e 4.4 3.0 6.3 4.4 2.1 2.5 3.6 1.1 
RNA editing sites —e 1,084f —e 491 —e 370g —e 427 441 
 Taxon 
Features Chara Marchantia Physcomitrella Cycas Triticum Oryza Zea Beta Nicotiana Brassica Arabidopsis 
Size (bp) 67,737 186,609 105,340 414,903 452,528 490,520 569,630 368,799 430,597 221,853 366,924 
A + T content (%) 59.1 57.6 59.4 53.1 55.7 56.2 56.1 56.1 55.0 54.8 55.2 
Gene numbera (protein-coding/tRNA/rRNA) 39/26/3 {39/26/3} 41/27/3 {41/29/3} 39/24/3 {39/24/3} 39/22/3 {39/26/3} 33/17/3 {35/25/8} 35/18/3 {40/27/6} 32/17/3b {34/22/5b} 29/18/32 {9/22/5} 36/21/3 {39/22/4} 33/17/3 {34/17/3} 32/17/3 {33/22/3} 
Coding sequences (%)a 90.7 20.3 37.0 10.1 8.6 11.1 6.2 10.3 9.9 17.3 10.6 
Repeat sequences (%)c 3.2 10.1 11.6 15.1 10.1 28.8 11.4 12.5 10.8 5.5 7.8 
Tandem repeat sequences (%)d 0.22 0.49 0.14 4.97 0.42 0.09 0.35 0.71 0.05 0.36 0.34 
Introns 
    Group I 14 
    Group II 
    Cis-spliced 13 25 25 20 17 17 15 14 17 18 18 
    Trans-spliced 
Cp-derived sequences (bp) —e —e —e 18,113 13,455 22,593 25,132 —e 9,942 7,950 3,958 
(%) —e —e —e 4.4 3.0 6.3 4.4 2.1 2.5 3.6 1.1 
RNA editing sites —e 1,084f —e 491 —e 370g —e 427 441 
a

Pseudogenes and unique ORFs (such as ORF222 reported in Handa 2003) were excluded. Upper: duplicate genes were counted only once. Lower within curly brackets: all duplicate genes were included.

b

The plasmid-localized tRNA-Trp was included (Clifton et al. 2004).

c

The REPuter program (http://bibiserv.techfak.uni-bielefeld.de/reputer) was used to obtain the estimates. All duplicate copies of repeats are included. Overlapped sequences were counted only once.

d

The tandem repeats finder program (Benson 1999) was used to obtain the estimates.

e

Data were not available from original papers.

f

The PREP-mt program (Mower 2005) was used to obtain the estimates. The cutoff value of the reported score was set to 0.6.

g

370 C-to-U editing sites were identified in 28 gene/ORF transcripts of sugar beet (Kubo et al. 2000).

FIG. 1.—

(A) Gene map of the Cycas taitungensis mtDNA. Genes are color coded into 12 groups according to their biological functions. Genes on the outside and inside of the 2 circles are transcribed clockwise and counterclockwise, respectively. Predicted mtpts are indicated in deep green between the 2 circles. Intron-containing genes are indicated by asterisks and marked with exon numbers. Degrees on the circular genome correspond to the linear maps in (B). (B) Lengths and distribution of repeat sequences and single Bpu sequences across the entire Cycas mtDNA. Upper: linear presentation of (A). Middle: the position of each repeat sequence (in blue) mapped onto the genome. Lower: the locations and lengths of tandem repeats detected using the tandem repeats finder. Single Bpu sequences (including variants) and their tandem repeats are shown in red, other tandem repeats are shown in black.

FIG. 1.—

(A) Gene map of the Cycas taitungensis mtDNA. Genes are color coded into 12 groups according to their biological functions. Genes on the outside and inside of the 2 circles are transcribed clockwise and counterclockwise, respectively. Predicted mtpts are indicated in deep green between the 2 circles. Intron-containing genes are indicated by asterisks and marked with exon numbers. Degrees on the circular genome correspond to the linear maps in (B). (B) Lengths and distribution of repeat sequences and single Bpu sequences across the entire Cycas mtDNA. Upper: linear presentation of (A). Middle: the position of each repeat sequence (in blue) mapped onto the genome. Lower: the locations and lengths of tandem repeats detected using the tandem repeats finder. Single Bpu sequences (including variants) and their tandem repeats are shown in red, other tandem repeats are shown in black.

Table 2 also shows that noncoding sequences (spacers, introns, and pseudogenes) account for 89.9% of the Cycas mtDNA sequence, consistent with the proportions found in other angiosperm mtDNAs (89.4 ± 3.1%). As land plants evolved from charophycean green algae (Chara, 9.3%), the closest living relatives of land plants (Karol et al. 2001), the noncoding sequences have drastically expanded in the mtDNA, showing abrupt increases in the bryophytes and then in the seed plants. As a result, the genomic organizations of mtDNAs are much less compact in seed plants when compared with lower plants.

Repeated sequences that present in the genome as multiple copies comprise approximately 15.1% of the Cycas mtDNA (table 2). The repeats, very few of which are over 2-kb long, are evenly distributed across the genome and mainly occur in the noncoding regions, including the intergenic spacers and introns (fig. 1B). As shown in table 2, most mtDNAs of land plants (except for rice) have 2–5 times more repeated sequences than the mtDNA of Chara, whereas more than a quarter of the rice mtDNA consists of repeated sequences. Among the sampled plant mtDNAs, that of Cycas contains the highest percentage of tandem repeats (4.97%; fig. 2A, detailed in supplementary table S3 [Supplementary Material online]); these include a novel family of mobile elements, the Bpu sequences/elements (see below), which are mainly found in 3- or 4-copy arrays (fig. 2B). Ogihara et al. (2005) noted the presence of many repeats in wheat mtDNA and hypothesized that alternative physical structures may be adopted by wheat mtDNA. Therefore, it seems logical to hypothesize that alternative circular mtDNAs in various recombinant forms might coexist in Cycas cells in vivo.

FIG. 2.—

Comparison of copy numbers among various tandem repeats in the mtDNAs of Cycas and 10 other land plants. (A) Percentage of total genome comprised of tandem repeat sequences. (B) The copy numbers of tandem repeat sequences in each genome (x axis) are separated into 5 groups (y axis) and their histograms (z axis) are shown. The single-copy Bpu sequences of Cycas taitungensis described in figure 1 are not included in the counts. The copy numbers may not be integers because the boundaries of tandem repeats were determined by a probabilistic method described by Benson (1999).

FIG. 2.—

Comparison of copy numbers among various tandem repeats in the mtDNAs of Cycas and 10 other land plants. (A) Percentage of total genome comprised of tandem repeat sequences. (B) The copy numbers of tandem repeat sequences in each genome (x axis) are separated into 5 groups (y axis) and their histograms (z axis) are shown. The single-copy Bpu sequences of Cycas taitungensis described in figure 1 are not included in the counts. The copy numbers may not be integers because the boundaries of tandem repeats were determined by a probabilistic method described by Benson (1999).

No group I intron was detected in the mtDNA of Cycas or the other 7 previously elucidated angiosperm mtDNAs (table 2). This observation is consistent with Knoop's (2004) speculation that group I introns were lost from the common ancestor of hornworts and tracheophytes. Although Cho et al. (1998) demonstrated that a group I intron located in the cox1 gene is widespread among 48 angiosperm genera, the examined angiosperms represented an exceptionally patchy phylogenetic distribution and did not include any of our sampled 7 angiosperms. Furthermore, the authors estimated that the intron invaded the cox1 gene by cross-species horizontal transfer over 1,000 times during angiosperm evolution and “is of entirely recent occurrence.”

In contrast, we found 20–25 group II introns in the mtDNAs of the examined land plants (table 2), nearly double the number found in those of their sister group, the charophytes (13). Similar to the angiosperms, the 3 genes nad1, nad2, and nad5 in Cycas are disrupted by 5 group II intron sequences, which have brought the genes into trans-splicing arrangements. Studies of Malek et al. (1997) and Malek and Knoop (1998) concluded that trans-spliced group II introns had evolved from formerly cis-spliced introns before the emergence of hornworts. Most recently, the evolving date was discovered to be even before the emergence of mosses (Groth-Malonek et al. 2005).

Gene Content and Evolution of Gene Loss in the mtDNAs of Land Plants

Our phylogenetic analysis using either nonsynonymous (Ka) or synonymous (Ks) substitutions of 22 mitochondrial protein-coding genes shared by the 11 studied plants generated identical tree topologies. Figure 3A shows the Neighbor-Joining trees reconstructed using the Ka and Ks values, respectively, with Chara being designated as the outgroup. The topologies of these 2 trees strongly indicate that, excluding the bryophytes Physcomitrella and Marchantia, the seed plants (including Cycas and the 7 angiosperms) and the angiosperms form nested monophyletic clades. Within the angiosperms, the monocots and eudicots constitute 2 distinct subclades. Noting that sisterhood relationship of Physcomitrella and Marchantia has to be treated with caution because the sampled seed plants share a long branch. In addition, a recent multigene analysis (Qiu et al. 2006) based on dense taxon sampling suggested that hornwort (including Marchantia) diverged before moss (including Physcomitrella). The Ka- and Ks-derived branch lengths leading to Cycas are nearly equal (fig. 3A), whereas the Ka-based branch lengths for the other species are strikingly shorter than the corresponding Ks-based branches. Statistical analysis using a Z-test further indicated that Ka value for the Cycas branch is higher than the average Ka of the 7 studied angiosperms (P < 0.05, Z-test). The elevated Ka in Cycas suggests that a rapid evolution/divergence may have occurred in some of the protein-coding genes of the Cycas mtDNA. This result is consistent with the observation that some genes in the Cycas mtDNA contain abundant RNA editing sites (see below).

FIG. 3.—

Phylogenies inferred from concatenated data from 22 protein-coding genes (see supplementary table S2, Supplementary Material online) common to the sequenced mtDNAs from 10 land plants and Chara (the outgroup). Nodes received 100% bootstrap replicates unless indicated. Branch lengths are drawn to scale except as otherwise noted. (A) Two superimposed Neighbor-Joining trees based on the Ka and Ks values, respectively. (B) Scenarios of gene losses (open bar) and splits (hatch triangle) along the single maximum parsimony tree (9,857 steps). The total numbers of protein-coding genes in each mtDNA species are given within parentheses.

FIG. 3.—

Phylogenies inferred from concatenated data from 22 protein-coding genes (see supplementary table S2, Supplementary Material online) common to the sequenced mtDNAs from 10 land plants and Chara (the outgroup). Nodes received 100% bootstrap replicates unless indicated. Branch lengths are drawn to scale except as otherwise noted. (A) Two superimposed Neighbor-Joining trees based on the Ka and Ks values, respectively. (B) Scenarios of gene losses (open bar) and splits (hatch triangle) along the single maximum parsimony tree (9,857 steps). The total numbers of protein-coding genes in each mtDNA species are given within parentheses.

A total of 39 protein-coding genes were identified in the Cycas mtDNA, which is the highest gene number identified to date among the studied seed plant mtDNAs (fig. 3B). When the distributions of conserved genes in the mtDNA from the 11 sampled species (supplementary table S2, Supplementary Material online) are mapped to their respective branches in a maximum parsimony tree (fig. 3B; based on the data used in fig. 3A), it is possible then to estimate the time of loss of a particular gene. As shown in figure 3B, our analysis indicates that there have been at least 31 independent events of gene loss from all the land plant mtDNAs elucidated to date.

When a gene is missing from the mtDNA of a given species, it is generally believed that the original copy has been transferred to nucleus, where it functions through cytosolic protein synthesis followed by transit peptide–assisted import back to the mitochondria (Adams and Palmer 2003; Knoop 2004). Frequent gene losses, especially of the ribosomal protein genes (30 of 34), appear to have occurred after the divergence of the angiosperm lineages approximately 150 MYA (Chaw et al. 2004). However, only 1 gene loss was observed in the Cycas lineage after it branched off from the common ancestor of angiosperms, approximately 300 MYA. Adams and Palmer (2003) suggested that angiosperm mtDNAs have experienced a recent evolutionary surge of loss and/or transfer of genes (primarily those encoding ribosomal proteins) to the nucleus. Our data give additional support to their contention but suggest that the mtDNA of Cycas appears to have been excluded from this surge. Our results suggest that the Cycas mtDNA tends to evolutionarily maintain its gene diversity and/or enjoys less gene transfer than other angiosperm mtDNAs (Selosse et al. 2001; Adams et al. 2002; Adams and Palmer 2003). Coincidentally, among the some 40 published cpDNAs of seed plants, that of Cycas also undergoes the least gene loss (Wu et al. 2007, supplementary fig. 1 [Supplementary Material online]). Future work will be required to examine why the genomes of these 2 Cycas organelles appear to have frozen after the cycads divergence from angiosperms, especially in comparison to mtDNAs from other major gymnosperm clades, such as Ginkgo, gnetophytes, and pines.

A Novel Family of Short Interspersed Mitochondrial Elements

Numerous Short Interspersed Mitochondrial Elements, Termed Bpu Sequences, Are Present in the Cycas mtDNA

Sequence analysis revealed that numerous copies of a 36-nt repeat, herein designated as a “Bpu sequence/element,” are interspersed throughout the Cycas mtDNA (fig. 1B). Figure 4A shows the characteristic sequences of 500 Bpu elements having 0 to 4 mismatches. If up to a 7-nt mismatch to the dominant type is allowed, the total copy number of Bpu sequences increases to 512. The Bpu sequences feature 2 conserved terminal direct repeats (AAGG) and the recognition sequence for the restriction endonuclease, Bpu10I (CCTGAAGC; nt 15–21). These repeat elements/sequences do not appear to have coding potential.

FIG. 4.—

Bpu sequences and examples of their insertion loci. Target sites and terminal repeats are underlined. (A) Upper: a dominant sequence based on 500 Bpu sequences (see text). Numbers along the abscissa indicate each nucleotide position within the prototype Bpu sequence. Recognition sites for the Bpu10I restriction enzyme are marked with asterisks. The ordinate scales each by the total bits of information multiplied by its relative occurrence at that position (Wasserman and Sandelin 2004). Lower: identities of Bpu sequences that differ from the dominant type by 0–4 bp. (B) Secondary structure of the dominant Bpu sequence. (C)–(F): lower cases, periods, and dashes denote mismatches, identical bps, and deletions, respectively, compared with the uppermost sequence. (C) Partial alignment of mt-rrn18 sequences from Cycas and Ginkgo. A Bpu sequence and a reversely complementary Bpu sequence are present in Ginkgo and Cycas, respectively. (D) Partial alignment of mtpt-atpB sequences extracted from Cycas, Oryza, Arabidopsis, and Nicotiana. Note that 2 Bpu sequence insertions are detected in Cycas but not in the other plants. (E) Partial alignment of mtpt-rbcL sequences extracted from Oryza, Arabidopsis, Nicotiana, and Cycas. The sequence from Cycas contains a Bpu-like sequence with a 5-bp insertion (shaded box) versus the dominant type. (F) Partial alignment of the mtpt-atpE and chloroplast atpE sequences of Cycas. Two reversely complementary Bpu sequences are found in the mtDNA. Note that the 5′ Bpu sequence is partly degenerated at its 3′ end.

FIG. 4.—

Bpu sequences and examples of their insertion loci. Target sites and terminal repeats are underlined. (A) Upper: a dominant sequence based on 500 Bpu sequences (see text). Numbers along the abscissa indicate each nucleotide position within the prototype Bpu sequence. Recognition sites for the Bpu10I restriction enzyme are marked with asterisks. The ordinate scales each by the total bits of information multiplied by its relative occurrence at that position (Wasserman and Sandelin 2004). Lower: identities of Bpu sequences that differ from the dominant type by 0–4 bp. (B) Secondary structure of the dominant Bpu sequence. (C)–(F): lower cases, periods, and dashes denote mismatches, identical bps, and deletions, respectively, compared with the uppermost sequence. (C) Partial alignment of mt-rrn18 sequences from Cycas and Ginkgo. A Bpu sequence and a reversely complementary Bpu sequence are present in Ginkgo and Cycas, respectively. (D) Partial alignment of mtpt-atpB sequences extracted from Cycas, Oryza, Arabidopsis, and Nicotiana. Note that 2 Bpu sequence insertions are detected in Cycas but not in the other plants. (E) Partial alignment of mtpt-rbcL sequences extracted from Oryza, Arabidopsis, Nicotiana, and Cycas. The sequence from Cycas contains a Bpu-like sequence with a 5-bp insertion (shaded box) versus the dominant type. (F) Partial alignment of the mtpt-atpE and chloroplast atpE sequences of Cycas. Two reversely complementary Bpu sequences are found in the mtDNA. Note that the 5′ Bpu sequence is partly degenerated at its 3′ end.

Because Bpu elements are extremely short in their lengths and terminal repeats and contain 2 terminal direct repeats rather than the inverted repeats found in plant miniature inverted-repeat transposable elements (MITEs), they are here classified as short interspersed mitochondrial elements (SIMEs). This distinguishes them from the MITEs and short interspersed nuclear elements (SINEs, e.g., Alu DNA repeats in the genomes of primates) that have been extensively reported from the genomes of plants and animals (Feschotte et al. 2002). Based on the similarity between SIMEs and SINEs, however, we can hypothesize that Bpu elements are likely to be transposed through RNA without a requirement for reverse transcriptase.

The Bpu10I recognition site characteristic of the Bpu elements is highly conserved except at the very last base pair, that is, the 21st bp of Bpu sequence (fig. 4A, lower chart). In contrast, the 5′ terminal direct repeat and its downstream 6 nt (nt 5–9) tend to be more variable than the 3′ terminal direct repeat. Enigmatically, that elements can form a secondary structure (fig. 4B) with a predicted free energy of −12.5 (kcal/mole), as calculated by the MFOLD program (http://frontend.bioinfo.rpi.edu/zukerm/comments/FAQs.html). The significance of this secondary structure remains to be elucidated. Comparison of the Bpu element insertion sites in the Cycas mtDNA with corresponding sites in the mtDNAs of other species reveals that the target sites for transpositions of Bpu elements are identical (within 1–2 mismatched base pair) to the 5′ terminal direct repeat (fig. 4C−F).

Bpu Sequences Distinguish Cycas from Other Cycads and Seed Plants

Bpu sequences are present exclusively in the noncoding regions of the Cycas mtDNA, that is, within the introns and intergenic spacers, with 1 exception: a Bpu sequence is present within the coding region of rrn 18, which codes the 18S rRNA (fig. 4C). Among the many mitochodrial rrn18 (mt-rrn18) genes available in GenBank, only those of Cycas (including Cycas revoluta, GenBank accession number: AB029356) and Ginkgo biloba (the only living species of the order Ginkgoales) have Bpu elements (fig. 4C). The Bpu insertion sites of the 2 Cycas taxa are orthologous (data not shown), whereas that of Ginkgo is different, indicating that the insertions of Bpu elements in rrn18 did not occur in the common ancestor of cycads and ginkgo. In sequenced mtDNAs, the oldest cpDNA-derived sequences (also termed mtpt) cluster, trnV(uac)-trnM(cau)-atpE-atpB-rbcL, is reported to have existed since the common ancestor of seed plants (Wang et al. 2007; fig. 3). Bpu elements are present in the Cycas mtpt-atpB (fig. 4D) and mtpt-rbcL (fig. 4E) sequences but not in the corresponding sequences of the 7 angiosperm mtDNAs elucidated to date (Wang et al. 2007), suggesting that these Bpu insertions took place after the split of the Cycas lineage from the other seed plants.

We further examined the occurrence of Bpu elements in 1 of the other 2 cycad families, Zamiaceae (including 1 species from each of Dioon, Macrozamia, and Zamia) by sequencing the exons 1 and 2 of their nad2 genes. Whereas the nad2.i1 gene of Cycas mtDNA contains 5 Bpu elements, such elements are absent from the nad2 genes of the 3 sampled Zamiaceae genera. However, we found 1, 2, and 6 Bpu elements in mitochondrial nad5.i1 of Cycaspanzhihuaensis (GenBank accession number: AF43425) and mitochondrial nad1.i1 and cp rps19-rpl16 spacers/intron of C. revoluta (GenBank accession number: AY354955, AY345867), respectively. These data seem to suggest that no Bpu sequence has successfully invaded the mtDNAs of Cycadales genera other than Cycas.

A Bpu-like sequence with a 1-bp insertion and 89% similarity to the dominant type in Cycas mtDNA was retrieved from the coding sequence of the mitochondrial rps11 gene from the core eudicot, Weigela hortensis. Alignment of this Bpu-like sequence reveals that its predicted Bpu10I endonuclease recognition site differs from that of Cycas by 2 bp (gCTGAGt). Because this Bpu element-like sequence does not interrupt the reading frame of rps11 and because the mitochondrial rps11 gene of Weigela lacks a target duplication site, we do not consider that this Bpu-like element shares a common origin with the Bpu sequences of Cycas. We further postulate that Bpu elements are likely absent from or very rare in angiosperms.

Surprisingly, the cpDNA of Cycas also contains 2 Bpu elements and each of them locates in the petN-psbM and psbA-trnK spacers, respectively (see Wu et al. 2007). Additionally, we also identified 1 Bpu sequence in the atpB-rbcL spacer of the cpDNA from Pinus luchuensis (GenBank accession number: DQ196799). However, no Bpu sequence has been detected in the cpDNAs of the 3 other Pinaceae genera and 3 gnetophyte orders we have sequenced to date (Chaw SM, Wu CS, Lai YT, Wang YN, Lin CP, Liu SM, unpublished data). Collectively, the available evidence inclines us to believe that the Bpu sequences have proliferated specifically in the mtDNA of Cycas (or the Cycadaceae). The sporadic occurrence of Bpu elements in the cpDNAs of Cycas and Pinus suggests that they are likely derived from nonhomologous recombination with DNA fragments that leaked out of mitochondria in the former case and via lateral transfer in the latter case.

The Tai Sequence and Its Association with Bpu Sequences

The second intron of the rps3 gene (rps3.i2) is only found in the mtDNAs of Cycas (Regina et al. 2005), including those studied in the present work. Regina et al. (2005) suggested that rps3.i2 is a group II intron that was independently gained in the gymnosperms likely at or just after the divergence of the angiosperms. Moreover, the authors reported a high similarity between a partial segment of the Cycas rps3.i2 and orf760 of the Chara mtDNA (Turmel et al. 2003). Orf760 harbors functional domains for a maturase and a reverse transcriptase, prompting Regina et al. (2005) to propose that the Cycas rps3.i2 gene originally encoded a maturase and a reverse transcriptase but evolved over time into a partially degenerated ORF. Here, we further report that rps3.i2 of the Cycas mtDNA contains a 900-bp fragment comprising an array of 4 Bpu elements with the Bpu elements on each end lacking 1 terminal repeat, followed by a 440-bp fragment (designated as “Tai” sequence), a Ψorf760 sequence, and 1 perfect Bpu sequence (fig. 5A). Most intriguingly, additional Tai sequences are scattered throughout the Cycas mtDNA; they occur more densely in longer spacers than in shorter ones (fig. 5B) and are generally found in close association with repeated Bpu sequences. These findings lend extra and robust support to the proposal of Regina et al. (2005).

FIG. 5.—

Genomic organization of the second rps3 intron (rps3.i2) and association of Tai sequences with Bpu sequences. (A) Organization of rps3.i2. The segment between ψorf760 and the 4 tandem Bpu sequences is designated as a Tai sequence (gray). Arrowheads indicate a Bpu sequence lacking the 2 terminal repeats (banded bars) and its complementary sequence. (B) Upper: association of Tai sequences with Bpu tandem repeats across the genome (abscissa) and their respective lengths (ordinate). Lower: examples of 3 Tai variants (boxed) showing that the sequences are variously degenerate and truncated as compared with the typical Tai shown in (A) (broken line indicates mismatching sequences). Thin arrows indicate the orientations of homologous segments between Tai variants and the typical Tai.

FIG. 5.—

Genomic organization of the second rps3 intron (rps3.i2) and association of Tai sequences with Bpu sequences. (A) Organization of rps3.i2. The segment between ψorf760 and the 4 tandem Bpu sequences is designated as a Tai sequence (gray). Arrowheads indicate a Bpu sequence lacking the 2 terminal repeats (banded bars) and its complementary sequence. (B) Upper: association of Tai sequences with Bpu tandem repeats across the genome (abscissa) and their respective lengths (ordinate). Lower: examples of 3 Tai variants (boxed) showing that the sequences are variously degenerate and truncated as compared with the typical Tai shown in (A) (broken line indicates mismatching sequences). Thin arrows indicate the orientations of homologous segments between Tai variants and the typical Tai.

In conclusion, we hypothesize that a Tai sequence and an orf760 flanked by a Bpu sequence at each end most likely constitute an ancestral retrotransposon. This hypothesis is founded on 2 observations: 1) Tai sequences are highly associated with Bpu elements and 2) small segments of Tai sequences are vigorously and diversely rearranged, deleted, or truncated as illustrated in the 3 examples shown in figure 5B. These observations also suggest that an ancestral retrotransposon in Cycas mtDNA spread very actively, presumably after Cycas branched off from the 10 other extant cycad genera. Afterward, the offspring or duplicates of the ancestral transposon gradually may have lost their mobility through varying degrees of deletions/truncations from the 3′ region, which encoded both maturase and reverse transcriptase functions.

We further speculate that even after the ancestral retrotransposon lost its transposon function, the remaining Bpu elements might have retained the ability to proliferate or amplify via unequal crossing-over or slipped-strand mispairing because their 2 direct terminal repeats can pair with each other's complimentary strands. Future studies and sequencing of additional mtDNAs from basal Cycas will be required to confirm this hypothesis and may provide additional insight into the evolutionary origins of Bpu sequences and the molecular mechanisms underlying their amplification.

The Fates of CpDNA-Derived Sequences (mtpts) in Cycas mtDNA

Table 2 shows a comparative analysis of mtpts among the 11 studied plant mtDNAs. The total percentage of mtpts in Cycas (4.4%) is relatively high compared with that in dicots (1.1–3.6%) but falls within the range seen among monocots (3.0–6.3%). We previously discovered that the frequency of mtpt transfer is positively correlated with variations in mtDNA size (coefficient value r2 = 0.47) (Wang et al. 2007). Here, we report that the Cycas mtDNA contains 8 protein-coding genes, as well as 2 rRNA and 5 tRNA gene sequences originating from the cpDNA (fig. 1). However, frameshifts and indels within the protein-coding genes suggest that these Cycas mtpts have degenerated and are nonfunctional. In contrast, the 5 cpDNA-derived tRNAs in the Cycas mtDNA are able to fold into standard cloverleaf structures, as shown by tRNAscan-SE analysis (Lowe and Eddy 1997), and are thus likely to be functional. Previously, Sugiyama et al. (2005) used tRNAscan-SE to scan the tobacco mtDNA and concluded that the 6 cp-derived tRNAs shared by angiosperms are functional. Because no mtpt was observed in Chara, Marchantia, or Physcomitrella, Wang et al. (2007) concluded that frequent DNA transfer from cpDNA to mtDNA has taken place no later than in the common ancestor of seed plants, approximately 300 MYA.

Abundant RNA Editing Sites in Cycas mtDNA

Among the land plant DNAs, Cycas has the most predicted RNA editing sites (1,084 sites; supplementary table S4 [Supplementary Material online]). It is commonly believed that RNA editing arose together with the first terrestrial plants (Steinhauser et al. 1999). By using the PREP-mt software (Mower 2005) with the cutoff score set to 0.6, 1,084 sites within the protein-coding genes of the Cycas mtDNA were predicted to be C-to-U RNA editing sites. This is more than double the number of predicted sites within the elucidated mtDNAs of other land plants (table 2). If the cutoff score, which indicates the conservation degree of each editing site compared those found in the other published plant mtDNAs, is set to the most stringent criterion in PREP-mt (i.e., = 1), the number of editing sites decreases to 738, which is still the largest number found among the land plant mtDNAs elucidated to date.

It is believed that RNA editing is essential for functional protein expression as it is required to modify amino acids or generate new start or stop codons (Hoch et al. 1991; Wintz and Hanson 1991; Kotera et al. 2005; Shikanai 2006). For this reason, the large number of RNA editing sites in the Cycas mtDNA may indicate higher complexity at the DNA level and formation of various transcripts through RNA editing, thus potentially reflecting rapid divergence in Cycas.

Although RNA editing sites are known to be sporadically distributed in the genomes of plant organelles/mitochondria from seed plants, the mechanisms underlying this distribution pattern are not yet known (Shikanai 2006; Mulligan et al. 2007). Supplementary table S4 (Supplementary Material online) shows that within the Cycas mtDNA, gene complex I (which includes 9 nad genes), is the most extensively edited gene category. Nad4 and nad5 show the first and second highest number of editing sites, followed by the cox1 gene of complex III. The least edited gene is rps10, which contains only 6 predicted sites even when the cutoff score is set to 0.6. Mulligan et al. (2007) interpreted the sporadic pattern as a result of lineage-specific loss of editing sites through retroconversion, which could remove adjacent editing sites by replacement with the edited sequences. The abundant RNA editing sites in the Cycas mtDNA, however, are mysterious and completely deviate from the sporadic patterns reported in other seed plant mtDNAs. Therefore, the mechanism of RNA editing in Cycas mtDNA may have a different evolutionary history from those of other seed plants.

Previous studies (Steinhauser et al. 1999; Lurin et al. 2004; Shikanai 2006) led to the proposal that variation or multiplication of trans-acting factors (e.g., members of the pentatricopeptide repeat family) may allow rapid increases in the number of editing sites. This could possibly be correlated to the lineage-specific explosion of RNA editing sites observed in Cycas mtDNA. Furthermore, Handa's (2003) conclusion that the evolutionary speed in RNA editing is higher at the level of gene regulation than at the primary gene sequence level appears to provide additional support for our speculation. To improve understanding of the mechanisms involved in abundant RNA editing sites in Cycas mtDNA, critical analysis of mtDNAs from ferns, conifers, and other cycads will be required.

In all plant lineages, RNA editing often alters the amino acid sequence (Adams and Palmer 2003). Supplementary table S5 (Supplementary Material online) summarizes the number of predicted preediting codons and potential edited codons in the Cycas mtDNA. The 2 most frequent edited codons are TCA (Ser) and TCT (Ser), which are predicted to be edited into TTA (Leu) and TTT (Phe), respectively. The least frequently edited codon is CAA (Gln), which would putatively be edited in only 2 cases, wherein the codon's first position would be altered to yield the stop codon, TAA. Supplementary table S6 (Supplementary Material online) further depicts the formation of 4 start (from ACG to ATG) and 7 stop codons (from CGA to TGA) through RNA editing. To verify the accuracy of the editing sites predicted using PREP-mt, partial transcripts of the cox1, atp1, atp6, and ccmB genes were experimentally assayed using RT-PCR. Comparison of our RT-PCR sequences with the editing sites predicted by PREP-mt indicated accuracy values of 97.66%, 99.13%, 97.91%, and 99.25% for cox1, atp1, atp6, and ccmB, respectively.

Our RT-PCR analysis additionally identified a U-to-C editing type in atp1; this was missed in the prediction because detection of the U-to-C editing type is limited when using PREP-mt (Mower 2005). Unfortunately, programs for predicting RNA editing sites in nonprotein-coding genes, such as ribosomal RNA and tRNA genes, and that of silent editing sites (which alter the mRNA but do not change the translated amino acid) are not yet available. We are currently examining these additional editing types and the locations of such sites in the Cycas mtDNA.

No Large Repeats in Gymnosperms but Some in Angiosperms

In the mtDNAs of angiosperms, long repeated sequences vary in size from 2,427 bp in rapeseed to 127,600 bp in rice, and they show no homology to each other, implying that repeated sequences were independently acquired during the evolution from Marchantia to angiosperms (Sugiyama et al. 2005). In Cycas mtDNA, we only detected a few long repeated sequences, all less than 1.5 kb in length and mostly composed of Bpu sequences. It has been demonstrated that genes are continuously transferred from the nuclear genome and cpDNA to mtDNA (Knoop 2004). However, the mechanisms underlying the emergence of large repeats are not yet fully understood.

Conclusions

The complete Cycas mtDNA shows a number of unprecedented features that are atypical of the mtDNAs previously elucidated for other land plants, including the lowest A + T content, the highest proportion of tandem repeat sequences (mainly Bpu sequences) and gene number found so far, abundant RNA editing sites, and an exceptionally elevated Ka value for the protein-coding genes. The latter 2 features might be correlated. In comparison with the other known angiosperm organelle genomes, the cpDNA and mtDNA of Cycas have experienced the least gene loss. Peculiarly, the cpDNA of Cycas has the fewest RNA editing sites among land plants (Wu et al. 2007), whereas its mtDNA has the most. On the other hand, some characteristics are shared by mtDNAs of Cycas and angiosperms but not bryophytes; these include common gene transfers from cpDNA, similar ratios of noncoding sequences, lack of group I introns, relatively low numbers of tRNAs, and concurrence of trans-spliced group II introns. Nevertheless, the above-mentioned Cycas-specific features suggest that the Cycas mtDNA has a lineage-specific evolutionary history.

A novel family of SIMEs, designated as Bpu elements/sequences, represents another unique feature of the Cycadaceae lineage. Bpu elements are widely distributed throughout the noncoding regions of Cycas mtDNA (over 500 copies), with only 1 occurrence in a coding region (rrn 18). In the Cycas mtDNA, the highly conserved nature of Bpu sequences and their association with Tai sequences as well as Ψorf760 (located in rps3.i2) suggest that these sequences may have collectively originated from an ancestral retrotransposon. Further investigation into the origin, proliferation, and evolution of Bpu sequences will be desirable to help clarify the peculiar features of Cycas mtDNA.

Supplementary Material

Supplementary tables S1–S7 and figure 1 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).

This work was supported by joint research grants from the Research Center for Biodiversity, Academia Sinica, to S.-M.C. and D.W., and partially by National Science Council grant to S.-M.C. (94-2311-B001-059) and by a grant from the Institute of Information Science, Academia Sinica, to A.C.-C.S. We thank Chiao-Lei Cheng for sequencing the RT-PCR products. We are grateful for the 3 anonymous reviewers, who provided critical comments and valuable suggestions.

References

Adams
KL
Palmer
JD
Evolution of mitochondrial gene content: gene loss and transfer to the nucleus
Mol Phylogenet Evol
 , 
2003
, vol. 
29
 (pg. 
380
-
395
)
Adams
KL
Qiu
YL
Stoutemyer
M
Palmer
JD
Punctuated evolution of mitochondrial gene content: high and variable rates of mitochondrial gene loss and transfer to the nucleus during angiosperm evolution
Proc Natl Acad Sci USA.
 , 
2002
, vol. 
99
 (pg. 
9905
-
9912
)
Benson
G
Tandem repeats finder: a program to analyze DNA sequences
Nucleic Acids Res
 , 
1999
, vol. 
27
 (pg. 
573
-
580
)
Bergthorsson
U
Richardson
AO
Young
GJ
Goertzen
LR
Palmer
JD
Massive horizontal transfer of mitochondrial genes from diverse land plant donors to the basal angiosperm Amborella
Proc Natl Acad Sci USA.
 , 
2004
, vol. 
101
 (pg. 
17747
-
17752
)
Chaw
SM
Chang
CC
Chen
HL
Li
WH
Dating the monocot-dicot divergence and the origin of core eudicots using whole chloroplast genomes
J Mol Evol
 , 
2004
, vol. 
58
 (pg. 
424
-
441
)
Chaw
SM
Walters
TW
Chang
CC
Hu
SH
Chen
SH
A phylogeny of cycads (Cycadales) inferred from chloroplast matK gene, trnK intron, and nuclear rDNA ITS region
Mol Phylogenet Evol.
 , 
2005
, vol. 
37
 (pg. 
214
-
234
)
Chaw
SM
Zharkikh
A
Sung
HM
Lau
TC
Li
WH
Molecular phylogeny of extant gymnosperms and seed plant evolution: analysis of nuclear 18S rRNA sequences
Mol Biol Evol
 , 
1997
, vol. 
14
 (pg. 
56
-
68
)
Cho
Y
Qiu
YL
Kuhlman
P
Palmer
JD
Explosive invasion of plant mitochondria by a group I intron
Proc Natl Acad Sci USA.
 , 
1998
, vol. 
95
 (pg. 
14244
-
14249
)
Clifton
SW
Minx
P
Fauron
CM
, et al.  . 
(13 co-authors)
Sequence and comparative analysis of the maize NB mitochondrial genome
Plant Physiol
 , 
2004
, vol. 
136
 (pg. 
3486
-
3503
)
Feschotte
S
Zhang
X
Wessler
SR
Miniature inverted-repeat transposable elements and their relationship to established DNA transposons
 , 
2002
Washington (DC)
ASM Press
Groth-Malonek
M
Pruchner
D
Grewe
F
Knoop
V
Ancestors of trans-splicing mitochondrial introns support serial sister group relationships of hornworts and mosses with vascular plants
Mol Biol Evol
 , 
2005
, vol. 
22
 (pg. 
117
-
125
)
Handa
H
The complete nucleotide sequence and RNA editing content of the mitochondrial genome of rapeseed (Brassica napus L.): comparative analysis of the mitochondrial genomes of rapeseed and Arabidopsis thaliana
Nucleic Acids Res
 , 
2003
, vol. 
31
 (pg. 
5907
-
5916
)
Hoch
B
Maier
RM
Appel
K
Igloi
GL
Kossel
H
Editing of a chloroplast mRNA by creation of an initiation codon
Nature
 , 
1991
, vol. 
353
 (pg. 
178
-
180
)
Jones
DL
Cycads of the World—ancient plant in today's landscape
 , 
2002
2nd ed
Sydney (Australia)
Reed New Holland
Kadowaki
K
Kubo
N
Ozawa
K
Hirai
A
Targeting presequence acquisition after mitochondrial gene transfer to the nucleus occurs by duplication of existing targeting signals
EMBO J
 , 
1996
, vol. 
15
 (pg. 
6652
-
6661
)
Karol
KG
McCourt
RM
Cimino
MT
Delwiche
CF
The closest living relatives of land plants
Science
 , 
2001
, vol. 
294
 (pg. 
2351
-
2353
)
Knoop
V
The mitochondrial DNA of land plants: peculiarities in phylogenetic perspective
Curr Genet
 , 
2004
, vol. 
46
 (pg. 
123
-
139
)
Kotera
E
Tasaka
M
Shikanai
T
A pentatricopeptide repeat protein is essential for RNA editing in chloroplasts
Nature
 , 
2005
, vol. 
433
 (pg. 
326
-
330
)
Kubo
T
Nishizawa
S
Sugawara
A
Itchoda
N
Estiati
A
Mikami
T
The complete nucleotide sequence of the mitochondrial genome of sugar beet (Beta vulgaris L.) reveals a novel gene for tRNA(Cys)(GCA)
Nucleic Acids Res
 , 
2000
, vol. 
28
 (pg. 
2571
-
2576
)
Kumar
S
Tamura
K
Nei
M
MEGA3: integrated software for molecular evolutionary genetics analysis and sequence alignment
Brief Bioinform
 , 
2004
, vol. 
5
 (pg. 
150
-
163
)
Kurtz
S
Choudhuri
JV
Ohlebusch
E
Schleiermacher
C
Stoye
J
Giegerich
R
REPuter: the manifold applications of repeat analysis on a genomic scale
Nucleic Acids Res
 , 
2001
, vol. 
29
 (pg. 
4633
-
4642
)
Lowe
TM
Eddy
SR
tRNAscan-SE: a program for improved detection of transfer RNA genes in genomic sequence
Nucleic Acids Res
 , 
1997
, vol. 
25
 (pg. 
955
-
964
)
Lurin
C
Andres
C
Aubourg
S
, et al.  . 
19 co-authors
Genome-wide analysis of Arabidopsis pentatricopeptide repeat proteins reveals their essential role in organelle biogenesis
Plant Cell
 , 
2004
, vol. 
16
 (pg. 
2089
-
2103
)
Malek
O
Brennicke
A
Knoop
V
Evolution of trans-splicing plant mitochondrial introns in pre-Permian times
Proc Natl Acad Sci USA.
 , 
1997
, vol. 
94
 (pg. 
553
-
558
)
Malek
O
Knoop
V
Trans-splicing group II introns in plant mitochondria: the complete set of cis-arranged homologs in ferns, fern allies, and a hornwort
RNA
 , 
1998
, vol. 
4
 (pg. 
1599
-
1609
)
Michel
F
Ferat
JL
Structure and activities of group II introns
Annu Rev Biochem
 , 
1995
, vol. 
64
 (pg. 
435
-
461
)
Mower
JP
PREP-Mt: predictive RNA editor for plant mitochondrial genes
BMC Bioinformatics
 , 
2005
, vol. 
6
 pg. 
96
 
Mulligan
RM
Chang
KL
Chou
CC
Computational analysis of RNA editing sites in plant mitochondrial genomes reveals similar information content and a sporadic distribution of editing sites
Mol Biol Evol
 , 
2007
, vol. 
24
 (pg. 
1971
-
1981
)
Notsu
Y
Masood
S
Nishikawa
T
Kubo
N
Akiduki
G
Nakazono
M
Hirai
A
Kadowaki
K
The complete sequence of the rice (Oryza sativa L.) mitochondrial genome: frequent DNA sequence acquisition and loss during the evolution of flowering plants
Mol Genet Genomics
 , 
2002
, vol. 
268
 (pg. 
434
-
445
)
Oda
K
Kohchi
T
Ohyama
K
Mitochondrial DNA of Marchantia polymorpha as a single circular form with no incorporation of foreign DNA
Biosci Biotechnol Biochem
 , 
1992
, vol. 
56
 (pg. 
132
-
135
)
Ogihara
Y
Yamazaki
Y
Murai
K
, et al.  . 
(14 co-authors)
Structural dynamics of cereal mitochondrial genomes as revealed by complete nucleotide sequencing of the wheat mitochondrial genome
Nucleic Acids Res
 , 
2005
, vol. 
33
 (pg. 
6235
-
6250
)
Qiu
YL
Li
L
Wang
B
, et al.  , 
(13 co-authors)
The deepest divergences in land plants inferred from phylogenomic evidence
Proc Natl Acad Sci USA
 , 
2006
, vol. 
103
 (pg. 
15511
-
15516
)
Regina
TM
Picardi
E
Lopez
L
Pesole
G
Quagliariello
C
A novel additional group II intron distinguishes the mitochondrial rps3 gene in gymnosperms
J Mol Evol
 , 
2005
, vol. 
60
 (pg. 
196
-
206
)
Selosse
M
Albert
B
Godelle
B
Reducing the genome size of organelles favours gene transfer to the nucleus
Trends Ecol Evol
 , 
2001
, vol. 
16
 (pg. 
135
-
141
)
Shikanai
T
RNA editing in plant organelles: machinery, physiological function and evolution
Cell Mol Life Sci
 , 
2006
, vol. 
63
 (pg. 
698
-
708
)
Steinhauser
S
Beckert
S
Capesius
I
Malek
O
Knoop
V
Plant mitochondrial RNA editing
J Mol Evol
 , 
1999
, vol. 
48
 (pg. 
303
-
312
)
Stevenson
DW
Morphology and systematics of the Cycadales
Mem N Y Bot Gard
 , 
1990
, vol. 
57
 (pg. 
8
-
15
)
Stewart
CN
Jr
Via
LE
A rapid CTAB DNA isolation technique useful for RAPD fingerprinting and other PCR applications
Biotechniques
 , 
1993
, vol. 
14
 (pg. 
748
-
750
)
Sugiyama
Y
Watase
Y
Nagase
M
Makita
N
Yagura
S
Hirai
A
Sugiura
M
The complete nucleotide sequence and multipartite organization of the tobacco mitochondrial genome: comparative analysis of mitochondrial genomes in higher plants
Mol Genet Genomics
 , 
2005
, vol. 
272
 (pg. 
603
-
615
)
Terasawa
K
Odahara
M
Kabeya
Y
Kikugawa
T
Sekine
Y
Fujiwara
M
Sato
N
The mitochondrial genome of the moss Physcomitrella patens sheds new light on mitochondrial evolution in land plants
Mol Biol Evol
 , 
2007
, vol. 
24
 (pg. 
699
-
709
)
Turmel
M
Otis
C
Lemieux
C
The mitochondrial genome of Chara vulgaris: insights into the mitochondrial DNA architecture of the last common ancestor of green algae and land plants
Plant Cell
 , 
2003
, vol. 
15
 (pg. 
1888
-
1903
)
Unseld
M
Marienfeld
JR
Brandt
P
Brennicke
A
The mitochondrial genome of Arabidopsis thaliana contains 57 genes in 366,924 nucleotides
Nat Genet
 , 
1997
, vol. 
15
 (pg. 
57
-
61
)
Wang
D
Wu
YW
Chun-Chieh Shih
A
Wu
CS
Wang
YN
Chaw
SM
Transfer of chloroplast genomic DNA to mitochondrial genome occurred at least 300 MYA
Mol Biol Evol
 , 
2007
, vol. 
24
 (pg. 
2040
-
2048
)
Wasserman
WW
Sandelin
A
Applied bioinformatics for the identification of regulatory elements
Nat Rev Genet
 , 
2004
, vol. 
5
 (pg. 
276
-
287
)
Wintz
H
Hanson
MR
A termination codon is created by RNA editing in the petunia atp9 transcript
Curr Genet
 , 
1991
, vol. 
19
 (pg. 
61
-
64
)
Wu
CS
Wang
YN
Liu
SM
Chaw
SM
Chloroplast genome (cpDNA) of Cycas taitungensis and 56 cp protein-coding genes of Gnetum parvifolium: insights into cpDNA evolution and phylogeny of extant seed plants
Mol Biol Evol
 , 
2007
, vol. 
24
 (pg. 
1366
-
1379
)

Author notes

William Martin, Associate Editor
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.