Rickettsial DNA and a trans-splicing rRNA group I intron in the unorthodox mitogenome of the fern Haplopteris ensiformis.

Plant mitochondrial genomes can be complex owing to highly recombinant structures, lack of gene syntenies, heavy RNA editing and invasion of chloroplast, nuclear or even foreign DNA by horizontal gene transfer (HGT). Leptosporangiate ferns remained the last major plant clade without an assembled mitogenome, likely owing to a demanding combination of the above. We here present both organelle genomes now for Haplopteris ensiformis. More than 1,400 events of C-to-U RNA editing and over 500 events of reverse U-to-C edits affect its organelle transcriptomes. The Haplopteris mtDNA is gene-rich, lacking only the ccm gene suite present in ancestral land plant mitogenomes, but is highly unorthodox, indicating extraordinary recombinogenic activity. Although eleven group II introns known in disrupted trans-splicing states in seed plants exist in conventional cis-arrangements, a particularly complex structure is found for the mitochondrial rrnL gene, which is split into two parts needing reassembly on RNA level by a trans-splicing group I intron. Aside from ca. 80 chloroplast DNA inserts that complicated the mitogenome assembly, the Haplopteris mtDNA features as an idiosyncrasy 30 variably degenerated protein coding regions from Rickettiales bacteria indicative of heavy bacterial HGT on top of tRNA genes of chlamydial origin.

The dual-targeted RNA editing factor AEF1 is universally conserved among angiosperms and reveals only minor adaptations upon loss of its chloroplast or its mitochondrial target.

KEY MESSAGE: Upon loss of either its chloroplast or mitochondrial target, a uniquely dual-targeted factor for C-to-U RNA editing in angiosperms reveals low evidence for improved molecular adaptation to its remaining target. RNA-binding pentatricopeptide repeat (PPR) proteins specifically recognize target sites for C-to-U RNA editing in the transcriptomes of plant chloroplasts and mitochondria. Among more than 80 PPR-type editing factors that have meantime been characterized, AEF1 (or MPR25) is a special case given its dual targeting to both organelles and addressing an essential mitochondrial (nad5eU1580SL) and an essential chloroplast (atpFeU92SL) RNA editing site in parallel in Arabidopsis. Here, we explored the angiosperm-wide conservation of AEF1 and its two organelle targets. Despite numerous independent losses of the chloroplast editing site by C-to-T conversion and at least four such conversions at the mitochondrial target site in other taxa, AEF1 remains consistently conserved in more than 120 sampled angiosperm genomes. Not a single case of simultaneous loss of the chloroplast and mitochondrial editing target or of AEF1 disintegration or loss could be identified, contrasting previous findings for editing factors targeted to only one organelle. Like in most RNA editing factors, the PPR array of AEF1 reveals potential for conceptually "improved fits" to its targets according to the current PPR-RNA binding code. Surprisingly, we observe only minor evidence for adaptation to the mitochondrial target also after deep losses of the chloroplast target among Asterales, Caryophyllales and Poales or, vice versa, for the remaining chloroplast target after a deep loss of the mitochondrial target among Malvales. The evolutionary observations support the notion that PPR-RNA mismatches may be essential for proper function of editing factors.

Frequent chloroplast RNA editing in early-branching flowering plants: pilot studies on angiosperm-wide coexistence of editing sites and their nuclear specificity factors.

BACKGROUND: RNA editing by cytidine-to-uridine conversions is an essential step of RNA maturation in plant organelles. Some 30-50 sites of C-to-U RNA editing exist in chloroplasts of flowering plant models like Arabidopsis, rice or tobacco. We now predicted significantly more RNA editing in chloroplasts of early-branching angiosperm genera like Amborella, Calycanthus, Ceratophyllum, Chloranthus, Illicium, Liriodendron, Magnolia, Nuphar and Zingiber. Nuclear-encoded RNA-binding pentatricopeptide repeat (PPR) proteins are key editing factors expected to coevolve with their cognate RNA editing sites in the organelles. RESULTS: With an extensive chloroplast transcriptome study we identified 138 sites of RNA editing in Amborella trichopoda, approximately the 3- to 4-fold of cp editing in Arabidopsis thaliana or Oryza sativa. Selected cDNA studies in the other early-branching flowering plant taxa furthermore reveal a high diversity of early angiosperm RNA editomes. Many of the now identified editing sites in Amborella have orthologues in ferns, lycophytes or hornworts. We investigated the evolution of CRR28 and RARE1, two known Arabidopsis RNA editing factors responsible for cp editing events ndhBeU467PL, ndhDeU878SL and accDeU794SL, respectively, all of which we now found conserved in Amborella. In a phylogenetically wide sampling of 65 angiosperm genomes we find evidence for only one single loss of CRR28 in chickpea but several independent losses of RARE1, perfectly congruent with the presence of their cognate editing sites in the respective cpDNAs. CONCLUSION: Chloroplast RNA editing is much more abundant in early-branching than in widely investigated model flowering plants. RNA editing specificity factors can be traced back for more than 120 million years of angiosperm evolution and show highly divergent patterns of evolutionary losses, matching the presence of their target editing events.

Horsetails are the sister group to all other monilophytes and Marattiales are sister to leptosporangiate ferns.

The "Monilophyte" clade comprising ferns, horsetails and whisk ferns receives unequivocal support from molecular data as the sister clade to seed plants. However, the branching order of its earliest emerging lineages, the Equisetales (horsetails), the Marattiales, the Ophioglossales/Psilotales and the large group of leptosporangiate ferns has remained dubious. We investigated the mitochondrial nad2 and rpl2 genes as two new, intron-containing loci for a wide sampling of taxa. We found that both group II introns - nad2i542g2 and rpl2i846g2 - are universally present among monilophytes. Both introns have orthologues in seed plants where nad2i542g2 has evolved into a trans-arrangement. In contrast and despite substantial size extensions to more than 5kb in Psilotum, nad2i542g2 remains cis-arranged in the monilophytes. For phylogenetic analyses, we filled taxonomic gaps in previously investigated mitochondrial (atp1, nad5) and chloroplast (atpA, atpB, matK, rbcL, rps4) loci and created a 9-gene matrix that also included the new mitochondrial nad2 and rpl2 loci. We extended the taxon sampling with two taxa each for all land plant outgroups (liverworts, mosses, hornworts, lycophytes and seed plants) to minimize the risk of phylogenetic artefacts. We ultimately obtained a well-supported molecular phylogeny placing Marattiales as sister to leptosporangiate ferns and horsetails as sister to all remaining monilophytes. In addition, an indel in an exon of the here introduced rpl2 locus independently supports the placement of horsetails. We conclude that under dense taxon sampling, phylogenetic information from a prudent choice of loci is currently superior to character-rich phylogenomic approaches at low taxon sampling. As here shown the selective choice of loci and taxa enabled us to resolve the long-enigmatic diversifications of the earliest monilophyte lineages.

Horizontal gene transfer of chlamydial-like tRNA genes into early vascular plant mitochondria.

Mitochondrial genomes of lycophytes are surprisingly diverse, including strikingly different transfer RNA (tRNA) gene complements: No mitochondrial tRNA genes are present in the spikemoss Selaginella moellendorffii, whereas 26 tRNAs are encoded in the chondrome of the clubmoss Huperzia squarrosa. Reinvestigating the latter we found that trnL(gag) and trnS(gga) had never before been identified in any other land plant mitochondrial DNA. Sensitive sequence comparisons showed these two tRNAs as well as trnN(guu) and trnS(gcu) to be very similar to their respective counterparts in chlamydial bacteria. We identified homologs of these chlamydial-type tRNAs also in other lycophyte, fern, and gymnosperm DNAs, suggesting horizontal gene transfer (HGT) into mitochondria in the early vascular plant stem lineages. These findings extend plant mitochondrial HGT to affect individual tRNA genes, to include bacterial donors, and suggest that Chlamydiae on top of their recently proposed key role in primary chloroplast establishment may also have participated in early tracheophyte genome evolution.

A survey of PPR proteins identifies DYW domains like those of land plant RNA editing factors in diverse eukaryotes.

The pentatricopeptide repeat modules of PPR proteins are key to their sequence-specific binding to RNAs. Gene families encoding PPR proteins are greatly expanded in land plants where hundreds of them participate in RNA maturation, mainly in mitochondria and chloroplasts. Many plant PPR proteins contain additional carboxyterminal domains and have been identified as essential factors for specific events of C-to-U RNA editing, which is abundant in the two endosymbiotic plant organelles. Among those carboxyterminal domain additions to plant PPR proteins, the so-called DYW domain is particularly interesting given its similarity to cytidine deaminases. The frequency of organelle C-to-U RNA editing and the diversity of DYW-type PPR proteins correlate well in plants and both were recently identified outside of land plants, in the protist Naegleria gruberi. Here we present a systematic survey of PPR protein genes and report on the identification of additional DYW-type PPR proteins in the protists Acanthamoeba castellanii, Malawimonas jakobiformis, and Physarum polycephalum. Moreover, DYW domains were also found in basal branches of multi-cellular lineages outside of land plants, including the alga Nitella flexilis and the rotifers Adineta ricciae and Philodina roseola. Intriguingly, the well-characterized and curious patterns of mitochondrial RNA editing in the slime mold Physarum also include examples of C-to-U changes. Finally, we identify candidate sites for mitochondrial RNA editing in Malawimonas, further supporting a link between DYW-type PPR proteins and C-to-U editing, which may have remained hitherto unnoticed in additional eukaryote lineages.

Plant-type mitochondrial RNA editing in the protist Naegleria gruberi.

RNA editing converts hundreds of cytidines into uridines in plant mitochondrial and chloroplast transcripts. Recognition of the RNA editing sites in the organelle transcriptomes requires numerous specific, nuclear-encoded RNA-binding pentatricopeptide repeat (PPR) proteins with characteristic carboxy-terminal protein domain extensions (E/DYW) previously thought to be unique to plants. However, a small gene family of such plant-like PPR proteins of the DYW-type was recently discovered in the genome of the protist Naegleria gruberi. This raised the possibility that plant-like RNA editing may occur in this amoeboflagellate. Accordingly, we have investigated the mitochondrial transcriptome of Naegleria gruberi and here report on identification of two sites of C-to-U RNA editing in the cox1 gene and in the cox3 gene, both of which reconstitute amino acid codon identities highly conserved in evolution. An estimated 1.5 billion years of evolution separate the heterolobosean protist Naegleria from the plant lineage. The new findings either suggest horizontal gene transfer of RNA editing factors or that plant-type RNA editing is evolutionarily much more ancestral than previously thought and yet to be discovered in many other ancient eukaryotic lineages.

A unique transcriptome: 1782 positions of RNA editing alter 1406 codon identities in mitochondrial mRNAs of the lycophyte Isoetes engelmannii.

The analysis of the mitochondrial DNA of Isoetes engelmannii as a first representative of the lycophytes recently revealed very small introns and indications for extremely frequent RNA editing. To analyze functionality of intron splicing and the extent of RNA editing in I. engelmannii, we performed a comprehensive analysis of its mitochondrial transcriptome. All 30 groups I and II introns were found to be correctly removed, showing that intron size reduction does not impede splicing. We find that mRNA editing affects 1782 sites, which lead to a total of 1406 changes in codon meanings. This includes the removal of stop codons from 23 of the 25 mitochondrial protein encoding genes. Comprehensive sequence analysis of multiple cDNAs per locus allowed classification of partially edited sites as either inefficiently edited but relevant or as non-specifically edited at mostly low frequencies. Abundant RNA editing was also found to affect tRNAs in hitherto unseen frequency, taking place at 41 positions in tRNA-precursors, including the first identification of U-to-C exchanges in two tRNA species. We finally investigated the four group II introns of the nad7 gene and could identify 27 sites of editing, most of which improve base pairing for proper secondary structure formation.

Organellar RNA editing and plant-specific extensions of pentatricopeptide repeat proteins in jungermanniid but not in marchantiid liverworts.

The pyrimidine exchange type of RNA editing in land plant (embryophyte) organelles has largely remained an enigma with respect to its biochemical mechanisms, the underlying specificities, and its raison d'être. Apparently arising with the earliest embryophytes, RNA editing is conspicuously absent in one clade of liverworts, the complex thalloid Marchantiidae. Several lines of evidence suggest that the large gene family of organelle-targeted RNA-binding pentatricopeptide repeat (PPR) proteins plays a fundamental role in the sequence-specific editing of organelle transcripts. We here describe the identification of PPR protein genes with plant-specific carboxyterminal (C-terminal) sequence signatures (E, E+, and DYW domains) in ferns, lycopodiophytes, mosses, hornworts, and jungermanniid liverworts, one subclass of the basal most clade of embryophytes, on DNA and cDNA level. In contrast, we were unable to identify these genes in a wide sampling of marchantiid liverworts (including the phylogenetic basal genus Blasia)--taxa for which no RNA editing is observed in the organelle transcripts. On the other hand, we found significant diversity of this type of PPR proteins also in Haplomitrium, a genus with an extremely high rate of RNA editing and a phylogenetic placement basal to all other liverworts. Although the presence of modularly extended PPR proteins correlates well with organelle RNA editing, the now apparent complete loss of an entire gene family from one clade of embryophytes, the marchantiid liverworts, remains puzzling.

A hypothesis on the identification of the editing enzyme in plant organelles.

RNA editing in plant organelles is an enigmatic process leading to conversion of cytidines into uridines. Editing specificity is determined by proteins; both those known so far are pentatricopeptide repeat (PPR) proteins. The enzyme catalysing RNA editing in plants is still totally unknown. We propose that the DYW domain found in many higher plant PPR proteins is the missing catalytic domain. This hypothesis is based on two compelling observations: (i) the DYW domain contains invariant residues that match the active site of cytidine deaminases; (ii) the phylogenetic distribution of the DYW domain is strictly correlated with RNA editing.

Evolution of a pseudogene: exclusive survival of a functional mitochondrial nad7 gene supports Haplomitrium as the earliest liverwort lineage and proposes a secondary loss of RNA editing in Marchantiidae.

Gene transfer from the mitochondrion into the nucleus is a corollary of the endosymbiont hypothesis. The frequent and independent transfer of genes for mitochondrial ribosomal proteins is well documented with many examples in angiosperms, whereas transfer of genes for components of the respiratory chain is a rarity. A notable exception is the nad7 gene, encoding subunit 7 of complex I, in the liverwort Marchantia polymorpha, which resides as a full-length, intron-carrying and transcribed, but nonspliced pseudogene in the chondriome, whereas its functional counterpart is nuclear encoded. To elucidate the patterns of pseudogene degeneration, we have investigated the mitochondrial nad7 locus in 12 other liverworts of broad phylogenetic distribution. We find that the mitochondrial nad7 gene is nonfunctional in 11 of them. However, the modes of pseudogene degeneration vary: whereas point mutations, accompanied by single-nucleotide indels, predominantly introduce stop codons into the reading frame in marchantiid liverworts, larger indels introduce frameshifts in the simple thalloid and leafy jungermanniid taxa. Most notably, however, the mitochondrial nad7 reading frame appears to be intact in the isolated liverwort genus Haplomitrium. Its functional expression is shown by cDNA analysis identifying typical RNA-editing events to reconstitute conserved codon identities and also confirming functional splicing of the 2 liverwort-specific group II introns. We interpret our results 1) to indicate the presence of a functional mitochondrial nad7 gene in the earliest land plants and strongly supporting a basal placement of Haplomitrium among the liverworts, 2) to indicate different modes of pseudogene degeneration and chondriome evolution in the later branching liverwort clades, 3) to suggest a surprisingly long maintenance of a nonfunctional gene in the presumed oldest group of land plants, and 4) to support the model of a secondary loss of RNA-editing activity in marchantiid liverworts.