Abstract
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.
Similar content being viewed by others
Introduction
The conserved structure of chloroplast genomes (plastomes) and the stoichiometric dominance of chloroplast (cpDNA) over nuclear and mitochondrial (mtDNA) in total plant nucleic acid preparations have led to a tremendous increase of available complete plastome sequences with the advent of Next Generation Sequencing (NGS) technologies. The number of completed plant mitochondrial genome sequences (mitogenomes) is much lower in contrast, most notably in vascular plants (tracheophytes). This is largely due to the much more variable and complex mitogenome structures in tracheophytes1,2,3,4,5, which hitherto left complete mitogenome assemblies altogether missing for the large clade of leptosporangiate ferns among the monilophytes.
Complex, recombining mitogenomes clustered with repeat sequences and affected by the lateral invasion of chloroplast DNA, or even by horizontal gene transfer (HGT) from other species, have arisen independently in the four large tracheophyte clades: the angiosperms, the gymnosperms, the monilophytes and the lycophytes. The lycophytes as the evolutionary oldest of the four clades of extant vascular plants reflect this most clearly: The mitogenome in the club moss Phlegmariurus squarrosus is a circular DNA with a rich gene complement and even retaining several ancestral gene syntenies of the circular mtDNAs in bryophytes6. The mtDNAs of the quillwort Isoetes engelmannii or the spike moss Selaginella moellendorffii, in contrast, are strongly depauperated in gene content and heavily affected by recombination leading to complex and coexisting arrangements of coding islands embedded between repeated sequences7,8. Similarly, moderately compact and gene-rich circular mitogenomes are present in the gymnosperms Cycas taitungensis9 and Ginkgo biloba10. In contrast, the mitogenomes of Taxus cuspidata11 and Welwitschia mirabilis10 are reduced in gene content and this even despite size increase to nearly 1000 Kbp in the latter case. As reflected from a collection of mitochondrial scaffolds and chromosomal assemblies, respectively, yet much larger mitogenomes are present in the conifer I clade in Picea species12,13 or Larix sibirica14.
Among angiosperms, the mtDNA of the magnoliid Liriodendron tulipifera represents an ancestral, gene rich state15. Other flowering plant species, however, have complex mtDNAs of enormous sizes exceeding 1000 Kbp, are fragmented into multiple co-existing mitochondrial chromosomes, affected by massive lateral gene transfer (LGT) from the cpDNA or by horizontal gene transfer (HGT) from other species or display combinations of those features to variable degrees. The large mitogenomes exceeding 1 Mbp in the cucumber family16,17, the diverse multipartite mtDNAs in the genus Silene18 or the mtDNA of the isolated flowering plant Amborella trichopoda that is heavily affected by HGT from diverse species19 are prime examples.
Among monilophytes (ferns sensu lato including the horsetails), the second largest group of tracheophytes behind the angiosperms, the situation is less clear since only the two complete mitogenome sequences of Ophioglossum californicum (adder’s tongue) and the whisk fern Psilotum nudum have been determined20. Although the two taxa represent sister clades among the eusporangiate ferns, a paraphyletic grade at the base of extant fern taxa, their mitogenomes reflect differences indicating a dynamic evolution of mtDNAs also among the monilophytes. Aside from slight differences in gene and intron complement, the O. californicum mtDNA maps as a single circular mtDNA whereas two separate circular chromosomes exist in P. nudum20.
Most of extant fern diversity with more than 10,000 species resides in the leptosporangiate ferns, however. Studies on selected mitochondrial loci among ferns revealed interesting dynamics of group II introns in their mtDNAs21,22. The leptosporangiate fern family of Pteridaceae (Polypodiales) proved to be particularly interesting in a study of mitochondrial group II intron gain, loss and coevolution scenarios22, also with respect to the concomitant evolution of C-to-U and reverse U-to-C RNA editing, which is abundantly present in the endosymbiotic organelles of leptosporangiate ferns23,24. The Pteridaceae represent a large fern family comprising some 1150 species in 45 genera, placed into at least five sub-groups and potential sub-families: the cryptogrammoid, the adiantoid-vittarioid, the cheilanthoid-hemionitidoid, the ceratopteroid-parkerioid, and the pteroid ferns. Habitats occupied by these sub-groups are equally diverse, as they range from terrestrial, including epipetric and epiphytic, to even aquatic lifestyles.
We chose Haplopteris ensiformis among the epiphytic, vittarioid shoestring ferns for a detailed analysis of its two organelle genomes and transcriptomes. We found a typical, conserved chloroplast genome structure in H. ensiformis but identified a highly unorthodox mitogenome characterized by numerous active and inactive repeat sequences and a massive insertion of chloroplast DNA. Transcript maturation of the comparatively rich mitochondrial gene complement, lacking only ccm genes for cytochrome maturation, involves splicing of 24 group II and four group I introns and abundant C-to-U and U-to-C RNA editing at nearly 2000 sites. The most surprising novelties of molecular evolution in a plant mitogenome include a unique trans-splicing group I intron in the large ribosomal rRNA and, most notably, extended stretches of bacterial, Rickettsia-like DNA in the Haplopteris mtDNA.
Results and discussion
Our choice of the shoestring fern H. ensiformis as a leptosporangiate fern for complete assembly of both organelle genomes was based on pronounced variability in mitochondrial RNA editing and intron (co-)evolution in the monilophyte family Pteridaceae and the sub-family Vittarioideae in particular22. We used next-generation sequencing (NGS) data to assemble the chloroplast (cpDNA) and mitochondrial genomes (mtDNA) of H. ensiformis, accompanied by RNA-seq transcriptome analyses to study RNA processing with a special focus on intron splicing and RNA editing. As expected, chloroplast DNA reads dominated in the NGS data with read coverages of ca. 1800–4200 and allowed the straightforward assembly of the H. ensiformis cpDNA. The mitochondrial DNA, in contrast, revealed overall lower and much more variable read coverages and its assembly was highly complicated by a multitude of repeats, long intergenic stretches and the insertions of foreign DNA and required multiple independent PCR amplifications for verification and complete assembly.
The H. ensiformis plastome and its well-resolved RNA editome
The H. ensiformis chloroplast DNA (148,805 bp) reveals a typical conserved circular plastome structure with a large (80,986 bp LSC) and a small (20,773 bp SSC) single copy region separated by a pair of inverted repeats (IRs of 23,523 bp each). The chloroplast genome carries 116 genes widely conserved in other land plants, 85 of which encode proteins, including the recently characterized ycf94 gene, 4 rRNAs and 27 tRNAs (Fig. 1a). Likewise, the H. ensiformis cpDNA contains a set of 20 conserved introns. Our accompanying transcriptome analysis confirmed functional splicing for all of them.
a Haplopteris ensiformis reveals a typical plant circular plastome structure consisting of a large (LSC) and a small (SSC) single‐copy region separated by a pair of inverted repeats (IR) and an expectedly conserved, ancestral gene and intron complement. The genome map was created using OGDRAW90. Gene categories are indicated in the legend. Numbers in parentheses indicate the amount of C-to-U (blue) and U-to-C (red) RNA editing for the respective genes. Creations of start or stop codons by C-to-U editing are indicated by symbols “>” and “*” and the removal of stop codons by U-to-C editing is indicated by the exclamation marks, respectively. b The cpDNA of Haplopteris elongata (accession MH173086) features two “morffo” elements (“mobile ORFs in fern organelles”) in the IR region between rrn5 and the 3’-part of the trans-splicing rps12 gene. Recognizable sequence homologs of morffo2 can presently only be identified in Cyclosorus interruptus (accession MN599066, Thelypteridaceae) and Histiopteris incisa (accession MH319942, Dennstaedtiaceae) and a homolog of morffo1 (orange) can presently only be found in the distant fern Hymenophyllum holochilum (accession MH265124, Hymenophyllales). Only the upstream part of morffo2 (378 bp) is present in the H. ensiformis plastome, while a cpDNA insert in its mitogenome contains an extended region of 628 bp.
One striking structural difference concerns the presence of two mobile ORFs in fern organelles, morffo elements25, in the chloroplast genome of the related species Haplopteris elongata (Fig. 1b). Our data suggest their secondary loss in the now determined cpDNA of H. ensiformis rather than an independent gain in H. elongata as we will discuss below in the context of the numerous cpDNA inserts that we identified in the now determined H. ensiformis mitogenome.
Modern RNA-Seq technologies allow for detection also of low-rate RNA editing sites with reasonable precision and any numbers of reported edits for an organelle transcriptome should nowadays ideally be accompanied by threshold criteria for their detection. Likewise, instead of the frequently used terms “complete” or “partial” editing, the percentage of detected base conversions by RNA editing should be given for the respective editing sites. The generally high coverage of RNA-Seq reads for the chloroplast transcriptome (mostly above 1000× and here reaching coverages of up to 250,000× for the psbE gene in the case of the H. ensiformis cpDNA) allows for a detailed evaluation of RNA editing events with high precision, allowing for three digits behind the decimal point. Using stringent criteria for DNA and RNA read qualities (see “Methods”), we could identify 443 sites of chloroplast RNA editing covered by at least 100 RNA reads and RNA editing frequencies of at least 1.0% (Supplementary Data 1).
We here use our previous nomenclature proposal for unequivocal labeling of RNA editing events26, indicating the affected locus, the nucleotide resulting from C-to-U or reverse U-to-C editing (eU or eC), the position and, for the majority of edits within protein coding regions, the codon meaning before and after the edit (Supplementary Data 1). The careful analysis of RNA editing events contributes crucially to the identification of functional genes in the organelles or the dismissal of others as pseudogenes, notably in plant species like H. ensiformis featuring abundant C-to-U and reverse U-to-C RNA editing at the same time. A case in the point is the small reading frame ycf94/orf51 of hitherto unknown function between rps16 and matK, for which we here find even higher rates of RNA editing at two important sites (ycf94eU2TM and ycf94eU50PL) of 59% and 74%, respectively, than reported previously for other species27,28. Vice versa, we consider rps16 a degenerating pseudogene in H. ensiformis as we could not confirm the expected removals of stop codons by reverse U-to-C editing.
The range of observed editing efficiencies extends from 99.1% for the codon sense-changing editing event psbBeU116SL down to 1.0% for edits accDeU657SS (silent), petBi6g2eU345 (intron) or edit rps2eU349R*, which unexpectedly introduces an early stop codon in the rps2 coding sequence. Such edits at low frequency are likely collateral effects owing to lacking specificity of the chloroplast RNA editing machinery. The same holds true for most codon sense-changing edits that are unexpected (as they do not restore conserved codon identities) and likewise show only inefficient editing with low frequency (Supplementary Data 1).
Most of the detected non-silent RNA editing sites in the chloroplast coding sequences, however, confirm expectations for restoring conserved amino acid positions very well and are efficiently edited and, vice versa, we find low frequencies of editing nearly exclusively in silent or non-coding positions like 5’- or 3’-UTRs (Supplementary Data 1). However, exceptions exist: Prime example for efficient RNA editing events in non-coding regions, which could have been missed altogether in typical RT-PCR-based studies focusing on coding sequences, are cytidine-to-uridine conversions petBi6g2eU478 and rps12i346g2eU80 in the respective introns within petB and rps12 of 94.0% and 90.2%, respectively. Conversely, we observe strikingly low efficiencies in many cases of reverse U-to-C editing removing stop codons, for example rpoC2eC232*Q reconstituting a glutamine codon in the rpoC2 reading frame is edited with only 17.0% efficiency in the transcript population.
Ten start codons and one stop codon are created by C-to-U editing and altogether 26 stop codons are removed by reverse U-to-C RNA editing in the chloroplast gene transcripts (Fig. 1 and Supplementary Data 1). We here use the accD gene as a somewhat less conserved protein coding region as an example for discussion (Supplementary Fig. 1). Codon-changing edits confirm predictions very well with editing frequencies between 61.0% for accDeC730SP and 94.1% for accDeU779SL. Unpredictable edits in the UTRs or silent position accDeU657SS are edited much less efficiently except for accDeU-1 right upstream of the start codon created by editing (Supplementary Fig. 1b). An intriguing case is editing site accDeU625HY inefficiently edited to 6.4%, for which a histidine or tyrosine is found variably in other taxa. An additional reverse edit is predicted for position accDeC580FL but remained unconfirmed. Note that in most fern cpDNA database entries, just two RNA editing sites are arbitrarily postulated to create an intact accD reading frame with a start codon edit in position 2 and to remove a stop codon in position 772.
Given that the organelle transcriptomes of H. ensiformis, and especially the mitochondrial transcriptome (see below), proved to be new examples for abundant C-to-U and U-to-C editing, we here use the opportunity to introduce a nomenclature amendment addressing the complex issue of multiple editings in individual codons. We suggest indicating the individual and cumulative effects on codon meaning after and before a pipe symbol (|), respectively, and additional silent edits by an underline symbol (Supplementary Fig. 2). As an example, we consider edits atpBeU1381PL|PS and atpBeU1382PL|PL changing a proline (P) codon identity in the atpB transcript. The first position edit takes place with 97.9% efficiency to expectedly reconstitute a conserved serine (S) codon whereas the unexpected additional second position edit causes a change into a leucine (L) codon, although with only 1.7% efficiency. Yet more complex is the example of a CCC proline codon in the rpoB gene edited with different frequencies in all three positions. Edit rpoBeU662PF|PL in the second codon position causes the expected change towards a conserved leucine codon with 66.3% efficiency. However, unexpected edit rpoBeU661PF|PS in the first codon position with only 1.2% efficiency would, considered alone, cause a change to serine. In combination with the edit in the second position the codon is changed to a phenylalanine (F) codon (Supplementary Data 1 and Supplementary Fig. 2). Moreover, these two non-silent edits are accompanied by a third position edit with 36% efficiency (rpoBeU663PP_FF), which is silent for any of the four possible codon identities (P, L, S, or F).
Assembly of the highly complex H. ensiformis mitogenome
The assembly of the H. ensiformis mitogenome turned out to be very demanding owing to a combination of several factors, which we will address in separate paragraphs below. The mtDNA sequence reads were not only ca. 10-fold less abundant than those of the cpDNA (on average ca. 150×) but also much more variable in coverage (ca. 70×–600×). We relied on parallel transcriptome analysis to verify authentic native mitochondrial genes characterized by RNA editing. Evident mtDNA contig assemblies ran into numerous repeated sequences, generally represented with higher read coverage or into extended insertions of laterally transferred chloroplast DNA fragments, a typical feature of vascular plant mitogenomes. Finally, we found numerous surprising similarities with Rickettsia-type bacterial genomes indicating multiple horizontal gene transfer (HGT) beyond the previously identified chlamydial tRNA genes in early branching tracheophytes29.
Multiple recombination breakpoints allow for a huge spectrum of alternative, and likely co-existing, mitogenome arrangements with variable stoichiometries. For clarity, we chose to assemble twelve mtDNA contigs (A to L, ranging in sizes from 2646 to 77,705 bp) comprising the full mitochondrial sequence complement into nine circular mtDNA chromosomes (chr1 to chr9) as separate GenBank database accessions (OM867545-OM867553). These nine circular chromosomes, however, likely represent only a substoichiometric minority of the truly existing mtDNA molecules owing to the numerous recombination breakpoints (Figs. 2 and 3). In total, 32 recombination breakpoints could be identified. These recombination breakpoints are labeled with “r,” the respective contig and consecutive numbers (rA1–rL2). We here display chromosome 5 comprising contigs A–C (Fig. 2) and chromosomes 1–4 and 6–9, variably integrating contigs D–K into the former (Fig. 3). Among these, contig A features prominently as it is flanked at its ends by repeat R9170 in inverted orientations, the largest repeat sequence that we identified in the H. ensiformis mitogenome (Fig. 2). Repeat R9170 carries the upstream part of the ribosomal rRNA gene cluster (Fig. 2), hence distantly reminding of the typical IRs in chloroplast genomes (see Fig. 1).
Haplopteris ensiformis mtDNA contigs A (hatched), B (gray), and C (dotted) can be connected into a circular chromosome of 217,669 bp. Recombination breakpoints are numbered for each contig and preceded by a small “r” with radial lines in the circular map indicating transitions into other contigs, allowing for numerous alternative mtDNA arrangements. Eight further circular chromosomes, as listed on top, connect sequences of chromosome V with nine further mtDNA contigs D-L as shown in Fig. 3. Contig gene maps were created using the SnapGene Viewer software. Native mitochondrial protein coding and tRNA gene sequence are given in lighter and darker blue, respectively, and introns are indicated with black arrows. Numbers next to genes indicate C-to-U (blue) and U-to-C RNA edits (red) with additional symbols indicating removal (!) or creation of stop (*) or start (>) codons, respectively. Introns are indicated with stippled lines and additional black arrows and their standardized labels. Ribosomal RNA genes (here on contig A) are shown in red with the rrnL gene featuring a complex gene structure requiring trans-splicing via the disrupted group I intron rrnLi825g1 and PSX labels indicate pseudogene fragments shown in gray. The peculiar case of the trans-spliced group I intron rrnLi825g1 is highlighted in yellow. Genes for tRNAs of chloroplast or bacterial origin are indicated in green or purple, respectively. For clarity, no other chloroplast or bacterial DNA insertions are shown here. The latter are listed together with annotated features in Supplementary Data 3.
The Haplopteris ensiformis mtDNA contigs D-L are linked to recombination points in chromosome 5 (Fig. 2) or within themselves, creating further and/or alternative mtDNA arrangements. Recombination endpoints are labeled and the display of contigs and labels for genes and RNA editing events is as in Fig. 2. Possible, circular chromosomal structures chr1 to chr4 and chr6 to chr9 are shown.
Repeats and recombination in the H. ensiformis mitogenome
Other than by the many chloroplast DNA inserts that we will discuss below, the assembly of the H. ensiformis mtDNA was complicated by numerous repeated sequences of different sizes. Ancestral mitochondrial gene syntenies are widely eradicated in H. ensiformis owing to numerous recombination events—the rps19-rps3-rpl16 gene continuity on contig F is one remaining exception (Fig. 3). While single-copy sequence contigs with mitochondrial genes had average NGS DNA read coverages around 150-fold (albeit with a broad distribution), identical repeat sequences mostly had coverages exceeding 300-fold. For clarity and discussion, we have labeled and annotated repeats with “R” followed by the number of nucleotides, also in the corresponding GenBank accessions.
We carefully checked on recombinational activity across repeats with a template-switch-avoiding tsa-PCR strategy and examples are shown in Fig. 4. Repeat R596 (Fig. 4a) is a particularly intriguing example as it is not intergenic but shared as an identical sequence present in domains I of group II introns nad5i1242g2 and rrnLi833g2, located on contigs H and F, respectively (Fig. 3). R596 had a coverage of ca. 500×, whereas the flanking single-copy contigs with nad5 and rrnL coding sequences had coverages of ca. 150× and 350×, respectively (Fig. 4a), indicating a non-equilibrium and different stoichiometries of the nad5 and rrnL gene copies. Naturally, active recombination would cause dysfunctional chimeric nad5/rrnL genes in this case. Exploring potentially active recombination across R596, we confirmed the gene continuities for nad5 and rrnL but found only very weak evidence for active recombination potentially giving rise to the two reciprocal arrangements (Fig. 4a).
Recombinations across repeats (in orange) R596 (A), repeats R203 and R 513 (B) and R180 (C) was investigated by tsa-PCR (”template-switch-avoiding”) strategies. a R596 is identically present in domains IV of group II introns nad5i1242g2 and rrnLi833g2. Average read coverage of the flanking single copy regions were ca. 150× for nad5 (arrangement A-C) and ca. 350× for rrnL (arrangement B-D), apparently adding up to ca. 500× for R596. PCR products are obtained for the expected gene continuities (AC, BD) with only minor evidence for reciprocal exchanges (AD, BC). b All combinations of flanking sequences (AC-BD) are identified for recombination across R203 whereas a clear bias is seen for recombination across R513 where a product for primer combination B-C remains undetected. c One copy of repeat R180 is located in intron nad2i709g2, another one downstream of the nad9 gene. The region between R180 and nad9 contains additional repeats R137 and R259 and all consecutive recombination products can be found whereas there is only very weak evidence for recombination across the R180 copy located in nad2i709g2. The marker lane of the electrophoresis gel covering R180 has been rearranged as indicated by the dotted line. The images of electrophoretic gels used in this figure are deposited as supplementary figures 5a and 5b.
However, similar tsa-PCR approaches run across other (intergenic) repeats indicated (stronger) active recombination across the respective repeated sequence resulting in more than only two co-existing, or at least strongly dominating, conformations of the flanking single-copy sequences. Examples are shown for R203 and R513 (Fig. 4b). Whereas all combinations of flanking sequences appear to co-exist for R203, the arrangements A-C and B-D appear to strongly dominate for R513 while arrangement A-D is stoichiometrically under-represented and arrangement B-C even remains undetected (Fig. 4b).
As an additional complication, some recombination points at the end of repeats were in in very close proximity resulting in complex combined repeats. We here show results for R180, one copy of which is in group II intron nad2i709g2 of the nad2 gene (Fig. 4c). In another location, R180 is flanked on one side by consecutive repeats R137 and R259 resulting in alternative pathways to three different endpoints (D, E and F in Fig. 4c). Endpoints D and E result in intact atp4 and atp6 gene copies, respectively, whereas the alternative arrangement A-F creates pseudogene fragments for both genes downstream of nad9. Only very weak evidence is found for recombination of nad2i709g2 across R180 into any of the three alternative arrangements B-D, B-E and B-F or for the reciprocal arrangement A-C (Fig. 4c).
Other than identical sequence repeats in distant locations, the mitogenome of Haplopteris carries copies of variably diverging sequences, sometimes closely spaced. Examples are an imperfect 39 bp IR embedding the cox1 gene or an inverted sequence repeat of 1.1 Kbp embedding the nad9 gene with (only) 70% sequence identity between the two copies. Direct repeats of an 85 bp sequence are located behind the nad1 gene overlapped by a repeated tridecanucleotide motif (CCTCTACTGAGGG) at their ends.
The mitochondrial gene complement in H. ensiformis
Despite its highly complex structure, the H. ensiformis mitogenome has a rich gene complement (Table 1). All expected genes for subunits of the respiratory chain complexes I–V (nad, sdh, cob, cox, and atp genes) are present, including sdh3 and sdh4 encoding subunits of complex II. Likewise, there is a surprisingly large set of genes for ribosomal proteins, given that many are absent even in the mtDNAs of more ancestral lineages. Notable is the retention of rps7, which is lacking in all available hornwort and lycophyte mtDNAs but retained in H. ensiformis. Missing from the Haplopteris mtDNA is only the suite of ccm genes (ccmB, ccmC, ccmF) for cytochrome c maturation. Given their absence also in Ophioglossum californicum but their conservation in Psilotum nudum (Table 1) among the eusporangiate ferns, this evidently reflects a further independent loss of the ccm gene suite among ferns along with other phylogenetic deep losses in the lycophytes and hornworts.
The tRNA gene complement of the H. ensiformis mtDNA is particularly interesting owing to several tRNA genes of xenologous origin. Chloroplast-derived gene copies are present for trnF-GAA, trnMe-CAU, trnN-GUU, trnP-UGG and trnS-GGA (Table 1). Moreover, the H. ensiformis mtDNA also carries chlamydial-type tRNAs trnN-GUU (Fig. 3) and trnR-UCG (Figs. 2 and 3) described previously20,29. This results in xenologous genes coexisting with their native counterparts for trnF-GAA and trnP-UGG. Genes for native tRNAs are lost for trnL-CAA, trnMe-CAU, trnN-GUU, trnR-ACG. The chlamydial-type trnR-UCG is remarkable since it exists in three slightly differing copies. Similarly notable is the presence of a trnL-UAG in Haplopteris that is absent in the eusporangiate ferns (Table 1).
RNA editing in H. ensiformis mitochondria
Given the complex mitogenome structure, the parallel transcriptome and RNA editing analysis was fundamental to identify functional mitochondrial genes in the H. ensiformis mtDNA. All protein-coding genes were found to be affected by RNA editing. Altogether, we identified 1618 events of mitochondrial RNA editing: 1091 of the C-to-U type and 527 edits of the reverse U-to-C type (Supplementary Data 2). The abundance of RNA editing is highly biased among genes with the cox1 mRNA being affected by 145 RNA editing sites in contrast to the atp1 mRNA with only three edits, respectively.
While the very high coverage of RNA-Seq reads in chloroplasts allowed for determination of editing frequencies with high precision, the more than threefold abundance of edits in mitochondria allowed for a better statistical classification of edits categorized by their location and effect (Supplementary Data 2). Among the total of 1618 mitochondrial edits, 1171 introduce codon changes and of the latter more than 900 are strongly predicted and more than 100 others moderately or weakly expected. We here illustrate the prediction of editing sites by PREPACT30 with the example of atp9 finding a perfect match between identified edits and expectations and for atp6 with only minor deviations from the expected editing pattern (Supplementary Fig. 3). Notably, the strongly expected RNA editing events have an average editing frequency of 83%, much higher than only 53% on average for non-predicted changes of codon identities (Supplementary Data 2). Yet lower RNA editing efficiencies are observed for silent sites or the ones in non-coding regions, e.g. only 29% on average in 3’-UTRs (Supplementary Data 2). A notable exception from efficient editing of coding regions is the matR maturase encoded in the terminal nad1 intron nad1i728g2 (Supplementary Data 2 and Table 1). Whereas identification of a start codon for this conserved, and only, mitochondrial group II intron maturase in flowering plants has been puzzling, we now find matR in H. ensiformis continuous with the upstream nad1 reading frame, accordingly, to be labeled mat-nad1i728g2c following a recent nomenclature proposal28. Although only lowly edited, the numerous, and expected, events of RNA editing reconstituting conserved codons and including 15 stop codon removals confirm the functional role of mat-nad1i728g2c (Supplementary Data 2).
Particularly remarkable is that 44% of the reverse U-to-C edits (233 of 530) serve to re-convert stop codons into arginine or glutamine codons, an important issue to distinguish functional from dysfunctional pseudogenes. For example, within 20 codons upstream of rps3i249g2, six stop codons are removed from the rps3 coding sequence, including one within the intron binding site (Supplementary Data 2).
A yet more dramatic example is the rpl6-rps13-rps11 cluster with 20 genomic stop codons in a short region (Supplementary Fig. 4). In fact, the rps11 gene at DNA level initially appeared to be an amino-terminally truncated pseudogene but turned out to have a proper start codon created by C-to-U editing and a total of seven stops removed by reverse U-to-C editing within the first 20 codons of its reading frame. We wished to test how different RT-PCR-based approaches would perform in comparison to the RNA-Seq approach to detect RNA editing sites. Towards that end, we used three different strategies for cDNA synthesis using either random hexamers or specific primers targeting the 3’-end of rps11 in edited or non-edited versions. Here, we made use of two edits in the 3’-UTR closely behind the rps11eU466Q* stop codon editing. Sequencing of an internal amplicon revealed that many editing events were not confirmed in the cDNA sample primed with random hexamers with better performance by the specific primers and notably the one for the edited version of the 3’-end.
A striking bias concerns silent edits leaving codon identities unchanged. We observed only 32 silent U-to-C edits but the nine-fold amount (282) of silent C-to-U edits. Interestingly, silent C-to-U edits are frequently found to neighbor non-silent sites (NESIs) as has previously been observed for the huge editome in the Selaginella uncinata chloroplast31.
Mitochondrial introns in H. ensiformis include a trans-splicing group I intron
The H. ensiformis mitogenome shows notable differences to the intron complements in the two eusporangiate fern taxa (Table 1). Ancient group II introns nad1i477g2, nad1i669g2, nad5i1477g2 and nad7i917g2 are lost from the H. ensiformis mtDNA. Vice versa, group I intron cox1i395g1 and group II introns cox2i373g2 and rps14i114g2 in the Haplopteris mitogenome lack counterparts in Ophioglossum and Psilotum (Table 1). The group II introns and cox1i395g1 are evidently of ancient origin in the land plant lineage. The latter has previously been identified in liverworts but also in the leptosporangiate ferns and in the horsetail Equisetum arvense32,33.
A striking example documenting the H. ensiformis mitogenome complexity concerns the ribosomal rRNA cluster with a peculiar arrangement featuring a disrupted rrnL gene (Figs. 2 and 3). Maturation of the large ribosomal RNA requires trans-splicing of a broken group I intron, rrnLi825g1. Secondary structure modeling suggests base-pairing interaction of the two intron parts in the disrupted domain P9.0/P9.1 (Fig. 5). Intriguingly, rrnLi825g1 has a distant positional ortholog as a cis-splicing homolog in the charophyte alga Chara vulgaris34. Despite overall only weak similarity, conserved regions include intron domains P7 and P8 known to contribute to the reactive core of group I introns. Three further introns are located in the downstream part of the H. ensiformis rrnL gene. Group II intron rrnLi833g2 is located only 8 bp downstream of the rrnLi825g1 3’-splice site and has distant homologs in liverwort mtDNAs. The two downstream group I introns rrnLi1897g1 and rrnLi1928g1 have hitherto only been identified serendipitously in rrnL gene samplings of other Polypodiales species. Notably, none of the four introns are present in the eusporangiate ferns, which feature continuous rrnL genes.
a Maturation of rrnL includes trans-splicing of the disrupted group I intron rrnLi825g1. Three additional introns are removed from the downstream part of rrnL. Group II intron rrnLi833g2 has a homolog conserved in the mtDNAs of liverworts and the downstream group I introns rrnLi1897g1 and rrnLi1928g1 are conserved in other Polypodiales species. Splice sites of rrnLi825g1 and rrnLii833g2 (black and gray triangles, respectively) frame the tiny second rrnL exon of only eight nucleotides. b Secondary structure model of the disrupted, trans-splicing group I intron rrnLi825g1. The figure was generated using the VARNA software91. Regions for base-pairing between the upstream and downstream parts of rrnLi825g1 are found in the group I intron secondary structure P9.1. Intron rrnLi825g1 intron has a positional ortholog in the sweet-water alga Chara vulgaris. Green and yellow shading indicates identical nucleotides and transitions, respectively.
Laterally transferred chloroplast DNA fragments in the Haplopteris mitogenome
Among other issues, the assembly of the H. ensiformis mitogenome was much complicated by the fact that it contains ca. 80 inserts of chloroplast DNA of variable sizes and with variable degrees of sequence conservation. Similar to the repeat sequences, we annotated these cpDNA inserts with numbers indicating their sizes in base pairs preceded by “cp” (Supplementary Data 3). While the separate chloroplast sequence inserts may have originated from fragmentation after insertion of large stretches of cpDNA, their different degree of sequence conservation rather argue for independent transfer events and likely document independent cpDNA insertions at different time points in evolution (Fig. 6). As an example, an array of seven likely independently acquired cpDNA inserts is present in the intergenic region between nad5 and sdh4 (Fig. 6a). This region includes cp4165, the largest continuous stretch of promiscuous cpDNA derived from the ndhH-ndhA-ndhI-ndhG-ndhE region, sharing 93% sequence identity with the native chloroplast counterpart. This stretch is directly flanked by cp1271 derived from the chloroplast IR region encoding the rrnL gene and sharing even 99% of identical nucleotides, likely indicating a yet more recent inter-organellar migration into the mitogenome. Chloroplast insert cp1039 downstream of rps4 (Fig. 6b) is another example for a likely recently acquired lateral sequence transfer. In a phylogenetic analysis, it branches next to the homologous sequence from the atpA-atpF region of the now determined H. ensiformis cpDNA as sister to the counterpart in H. elongata, even despite the generally high sequence drift among Pteridaceae (Fig. 6b).
Selected examples for altogether approximately 80 inserts of chloroplast DNA populating the Haplomitrium ensiformis mitogenome (see Supplementary Data 3). Maximum likelihood trees were constructed with IQ-TREE 92 after automatic model selection of TIM+F+I+G4 or GTR+F+I+G4 and trees were rooted with the Lindsaeaceae family or the Eupolypod II clade, respectively, for cp1039 and cp2126. Bootstrap support is derived from 500 replicates. a The intergenic region between nad5 and sdh4 contains the largest collection of likely independently acquired cpDNA inserts including the largest individual insert cp4165 with 93% similarity to the native chloroplast ndhH-ndhE region. The other inserts share variable sequence identities with the native H. ensiformis cpDNA ranging from 73% for cp686 to 99% for cp1271. Inserts cp364 and cp749 lack evident homologies in the H. ensiformis plastome, but are identified by sequence similarities to cpDNAs in other fern genera like Asplenium or Vittaria, highlighted in red. b Chloroplast DNA insert cp1039 derived from the chloroplast atpA-atpF region is an example for a likely very recently acquired insert as evident from its well-supported sister placement to the newly assembled H. ensiformis cpDNA (Fig. 1) in a phylogeny including the homologous plastome regions from diverse polypod ferns. c Chloroplast insert cp2128 embedded in other cp inserts of variable sequence conservation carries a unique morffo element, identified by sequence similarity only in the cpDNA of Hemionitis subcordata.
At the other end of the spectrum, some cpDNA inserts lack detectable homologous sequences in the Haplopteris chloroplast genomes and could only be recognized by similarities with the cpDNA of other taxa. Inserts cp364 and cp749 (Fig. 6a) are examples, which could only be identified by their similarity to sequences in the IR regions in the cpDNAs of genera like Asplenium or Vittaria. Even more notable are the cases of morffos, the highly variable mobile ORFs in fern organelles25. The chloroplast sequence insert cp2126 is an example along those lines (Fig. 6c), which includes a morffo element that has a top sequence identity of 75% in the cpDNA of Hemionitis subcordata in the distant subfamily Cheilanthoideae.
Yet more striking is the case of morffo2, which is intact in the H. elongata cpDNA but truncated to its 5’-terminal 378 bp in the now determined H. ensiformis chloroplast genome (Fig. 1b). Intriguingly, chloroplast insert cp1712 in the mtDNA (Supplementary Data 3) includes remarkably more of the 5’-end of the truncated morffo2 element (628 bp). Taken together, the slow sequence drift in the mitogenome offers examples allowing for a molecular archeology of former chloroplast DNA sequences that are not present any more in the recent chloroplast genome. In the latter case, it further supports the point for degeneration of the morffo elements in H. ensiformis rather than their independent origin in H. elongata (Fig. 1b).
Rickettsial DNA in the H. ensiformis mitogenome
The chlamydial-type trnN-GUU gene initially identified in the lycophyte P. squarrosus29 and the chlamydial-type trnR-UCG gene found subsequently in the eusporangiate fern mitogenomes20 are now also identified in the H. ensiformis mtDNA, further corroborating the concept of HGT from bacteria into the mitogenomes of early-branching vascular plants. Most strikingly, we now discovered numerous inserts of “Rickettsia-like” sequences in the H. ensiformis mitogenome. Altogether, we identified 30 variably degenerated protein coding regions evidently derived from rickettsial bacteria (see Supplementary Data 3). Similar to the labels for repeats and cpDNA inserts, we annotated the xenologous bacterial DNA inserts indicating their respective extension in base pairs, in this case preceded by an “x” (Supplementary Data 3 and Fig. 7).
Numerous inserts of bacterial, mostly rickettsial, protein-coding sequences were found to be integrated into the Haplopteris ensiformis mitogenome. Xenologous bacterial inserts were annotated to indicate their sizes in base pairs, preceded by “x” (see Supplementary Data 3), here showing examples for x625 (a), x888 and x1623 (b) and x1920 (c). Average read coverages are shown on top and PCR amplicons used to verify the mitogenome assemblies are indicated, with PCR results exemplarily shown for x625. a Bacterial insert x625 is located next to cpDNA insert cp416 between cox2 and the downstream part of rrnL. Two overlapping PCRs confirm the mitogenome assembly with linkages into both genes for two independent samples from distantly grown Haplopteris ensiformis plant isolations but failed to find products for a Vittaria lineata sample growing near H. ensiformis isolate 1. PCR products of expected sizes (framed with stippled boxes) were cut out and sequenced and confirmed sequence identity with the mitogenome assembly. The graph on top shows a continuity for the average read coverages of ca. 170× for cox2, cp416 and x625 with an increase to ca. 440× for the downstream rrnL region. Insert x625 carries the central region for DNA recombination protein RmuC with the upstream part located on insert x1623. b Xenologous inserts x888 and x1623 are located between nad1 and cpDNA insert cp232. Coding regions for HscA, an FeS-protein assembly chaperone and for RmuC borne on x1623 (see Supplementary Data 3 and Supplementary Data 4) are truncated, but highly conserved in primary sequence as exemplarily shown in the phylogenetic analysis for RmuC documenting a close association with Caedimonas varicaedens. Severely degenerated coding sequences for OMBB, an outer membrane beta-barrel domain containing protein borne on x888 are highly degenerated and do not allow a clear affiliation with a specific Rickettsiales bacterium. c Xenologous insert x1920 represents a continuous stretch of Rickettsia DNA with top similarities to four coding sequences in Caedimonas spp. Coding sequences (see Supplementary Data 4) are full-length for ribosome maturation factor RimM and the tRNA guanine-methyltransferase TrmD but amino-terminally truncated for MfdD, a transcription-repair-coupling factor and carboxy-terminally truncated for RpL19 encoding protein 19 of the large ribosomal subunit, respectively. The marker lane of the electrophoresis gel covering x625 PCR 2 has been rearranged as indicated by the dotted line. The original images of electrophoretic gels used in this figure are deposited as Supplementary Fig. 5c, d.
We carefully verified the surprising observation of Rickettsia-like DNA inserts in the H. ensiformis mitogenome by PCRs anchoring in the flanking mtDNA regions and consistently corroborated the mitogenome assemblies, as we here exemplarily show for x625 representing a central coding region for the bacterial DNA recombination protein RmuC (Fig. 7a). Towards that end we used independent DNA preparations from two H. ensiformis isolates from separate locations and, vice versa, included material from Vittaria lineata, a closely related species that grows next to one of the H. ensiformis populations in the Botanic Garden Bonn. PCR products were consistently obtained for the two independent H. ensiformis isolates, but not for the V. lineata sample (Fig. 7a). Sequencing of the PCR products fully confirmed the mitogenome assembly, interspersed by Rickettsial-like DNA insertions. Moreover, none of the bacterial insert sequences pointed to contamination by living bacteria as they mostly revealed characteristic degeneration of the protein coding genes to pseudogenes. Finally, we observed continuities of the average read coverages continuing from flanking mitochondrial sequences into the bacterial DNA inserts (Fig. 7). We do not assume functional expression of the xenologous bacterial genes given that we could detect only negligible RNA coverage for some of the regions carrying bacterial DNA inserts.
After insertion into the Haplopteris mtDNA, the Rickettsia-type coding sequences degenerated by recombination and sequence drift. Some protein sequence similarities, however, remain astonishingly high, likely indicating very recent horizontal gene transfers (Supplementary Data 4 and 5). Specifically, top similarities were often observed with the corresponding sequences from Caedimonas varicaedens35, followed by slightly lower similarities with homologous loci in Cand. Paracaedimonas acanthamoebae, other Caedibacter spp. or Cand. Nucleicultrix amoebiphila. These species belong to the family Holosporaceae among Rickettsiales sensu lato. The family Holosporaceae is alternatively considered to be a separate order of its own, the Holosporales.
As in the case of the chloroplast sequence inserts, the variable degrees of sequence degeneration suggest independent events of horizontal transfers at different time points in evolution. Alternatively, the insertion of larger xenologous genomic regions followed by later fragmentation in the mitogenome followed by different degrees of sequence degeneration may be possible. One evident example is the bacterial HscA-RmuC region (Fig. 7b) located between the nad1 gene and cp232 (Fig. 7b, see Supplementary Data 3). The HscA coding region is disrupted by an insert in the coding sequence and the downstream RmuC gene is truncated with its central region located as x625 between the cox2 gene and the rrnL-rrnS gene cluster, directly flanking cpDNA insert cp416 (Fig. 7a). Despite these discontinuities, the degree of amino acid sequence conservation reaching up to 94% identity is astonishing and phylogenetic analysis allows a close affiliation of the xenologous RmuC region in the Haplopteris mitogenome with Caedimonas varicaedens among the Holosporales (Fig. 7). Another xenologous insert x888 (Fig. 7b) carrying parts of the coding region for OMBB, an outer membrane beta-barrel domain containing protein, is much more degenerated and has much lower similarities to top-scoring hits with less clearly defined Rickettsiales bacteria (33% identity, 50% similarity).
Whereas most of the xenologous bacterial DNA inserts show fragmentation and considerable degeneration of protein-coding regions, there are also striking counterexamples like x1850, an insert of ca. 1.8 Kbp, containing the serS-surE-nlpD region located between cp763 and cp1456 (Supplementary Data 3). The coding region of the upstream serine tRNA ligase SerS is N-terminally truncated and carries two stop codons but the reading frame of the downstream coding region for the 5’/3’-nucleotidase SurE is perfectly conserved and shares even 97% sequence similarity with its Caedimonas varicaedens counterpart (Supplementary Fig. 2). In this case a similarity of 81% can even be identified at the nucleotide sequence level (Supplementary Fig. 3), clearly suggesting the bacterial donor to be closely related to C. varicaedens. Intriguingly, the end of this bacterial sequence insert contributes to repeat R295.
Examples for particularly large stretches of Holosporaceae DNA inserts include x2170 derived from the bacterial XpsD-region located behind the mitochondrial nad4L gene (Supplementary Data 3) and x1920, originating from the bacterial MfdD-RimM-TrmD-RpL19 region inserted upstream of rps4 (Fig. 7c and Supplementary Data 3). In the latter case, the gene for MfdD is severely 5’-truncated, the gene for RimM appears intact, the TrmD reading frame is degenerated by a stop codon and the RpL19 sequence is 3’ truncated (Fig. 7c).
In some cases, top sequence similarities clearly identify an origin of HGT sequences from Rickettsiales, but not necessarily from within the Holosporaceae. Most notable is an assembly of sequence inserts of altogether more than 7 Kbp downstream of the atp1 gene (Supplementary Data 3) that carries the suite of genes encoding murA and two CoA-carboxylase subunits in opposite direction, which are nearly unaffected by sequence degeneration except for in-frame stop codons (Supplementary Data 4). At nucleotide level, a top similarity of 74% identical nucleotides is observed with a not further identified Rickettsiaceae bacterium isolate PMG_002 (Supplementary Data 5). Intriguingly, the VirB8/B9 coding sequences of two P-type conjugative transfer proteins are located ca. 1.2 Kbp downstream of the murA homology. Top nucleotide sequence similarities are again observed for Rickettsiaceae bacterium isolate PMG_002. However, the virB8/B9 and the MurA coding sequences are not connected in that genome, indicating either a related and yet unidentified Rickettsiales donor, separate gene transfers or subsequent rearrangements after copy-transfer into the Haplopteris mtDNA.
Finally, two large stretches of protein coding regions are located upstream of nad9, running in opposite direction (labeled Bact-ORF 1 and 2, respectively), but cannot be assigned taxonomically owing to much lower similarities of only 40-50% at protein sequence level: A hypothetical ORF, possibly RNA polymerase, highest similarity with a Magnetovibrio sp. database entry (MBM08139) and a DNA-polymerase, highest similarity with a Zoogloea sp. sequence entry (KAB2964018).
A complete leptosporangiate fern mitogenome assembly
The large Pteridaceae family contains about 10% of extant ferns species36,37,38. Among them, the Adiantoid and Vittarioid sub-families (see Fig. 6) show particularly high levels of substitution rate heterogeneity39,40. We became interested in the taxon given the apparently dynamic evolution of mitochondrial introns and RNA editing22. Likewise, a striking diversity of chloroplast RNA editing has been reported in the genus Adiantum41.
The genomes of the two endosymbiotic organelles in H. ensiformis here reported are prime examples highlighting the discrepancy between the conservative evolution of chloroplast DNA and the highly dynamic evolution of vascular plant mitochondrial genomes. Among the recently reported flowering plant mitogenomes, the one of the holoparasite Ombrophytum subterraneum42 is a case in point documenting not only a multi-chromosomal structure but multiple evidence for HGT from its host plants. Like in other cases of multi-chromosomal plant mitogenomes, e.g., the hundreds of different mitochondrial chromosomes in some Silene species, no efforts are made any more to come up with the display of a hypothetical, and likely misleading, master circle43,44.
As a representative for the large clade of leptosporangiate ferns with ca. 10,000 species, the H. ensiformis mitogenome again adds to the list of astonishing molecular peculiarities that are hallmarks of vascular plant mtDNAs45,46. Given its extraordinary complexity, it comes as no surprise that no other leptosporangiate mitogenome has previously been assembled despite multiple NGS efforts including the water fern genera Azolla and Salvinia47,48 or the flying spider-monkey tree fern Alsophila spinulosa49 or, most recently, the model ferns Adiantum capillus-veneris50 and Ceratopteris richardii51. Likewise no mitogenome was reported in a genome assembly effort for the lycophyte Isoetes taiwanensis52, possibly owing to even higher complexity than the ones reported previously for Isoetes engelmannii and Selaginella moellendorffii7,8.
We hence speculate that highly complex mitogenomes like the one reported here may be a general feature of leptosporangiate ferns also outside of the Polypodiales. Intriguingly though, and despite its highly dynamic structural evolution, the H. ensiformis mtDNA contains a surprisingly rich set of classic mitochondrial genes (Table 1) when compared to the gene complement of other taxa including the eusporangiate ferns53.
Mitochondrial intron dynamics in ferns
More notable than the gene complements are the diverging mitochondrial intron complements now identified in H. ensiformis in comparison to the previously analyzed eusporangiate ferns, extending earlier conclusions that much more intron dynamics is present in monilophytes in comparison to their seed plant sister clade21,22,33,53. The Haplopteris mitogenome reveals retention of evidently ancient introns that have been lost in the eusporangiate ferns like cox1i395g1, cox2i373g2, rps14i114g2 and rrnLi833g2 (Table 1).
Surprisingly though, the heavily recombining mtDNA of H. ensiformis has not resulted in disrupted group II introns like in angiosperms or in gymnosperms where ever more transitions to trans-splicing have been observed recently54. In contrast, most group II introns (11 of 15) found to be trans-splicing in at least one seed plant lineage (cox2i373g2, nad1i394g2, nad1i728g2, nad2i542g2, nad2i1282g2, nad4i461g2, nad4i976g2, nad4i1399g2, nad5i1455g2, nad7i209g2 and rpl2i846g2) are present in conventional cis-arrangements in the Haplopteris mtDNA fully in line with the early evolutionary conclusion that trans-splicing introns in seed plants originate from cis-arranged ancestors in early-branching plant lineages55,56. We speculate that transitions from cis- to trans-splicing group II introns may rely on co-evolving protein splicing factors that are present in the seed plant but not in the monilophyte lineage. The remaining four group II introns known to exist in a trans-splicing state in at least some spermatophytes (cox2i691g2, nad1i669g2, nad5i1477g2 and nad7i917g2) have been lost altogether from the H. ensiformis mitogenome (Table 1).
In the light of the above, it is all the more surprising to find rrnLi825g1 as a group I intron of yet unclear ancestry in a trans-splicing arrangement in the idiosyncratic rrnL gene makeup in the H. ensiformis mitogenome. In contrast to the numerous examples of trans-splicing group II introns in plant organelles alone, trans-splicing group I introns appear to exist much more rarely in nature. First reports of trans-splicing group I introns in the Isoetes engelmannii mitogenome7 and in the mtDNA of Trichoplax57 have been followed by recognition of trans-splicing group I intron cox1i744g1 in Helicosporidium mtDNA58 and of two trans-spliced group I introns in Gigaspora margarita mtDNA59. Remarkably, the trans-splicing group I intron cox1i395g1 in the Isoetes engelmannii mtDNA7 exists in a conventional, cis-arranged version in the H. ensiformis mitogenome.
The mitochondrial rrnL gene is entirely devoid of introns not only in seed plants and lycopyhtes but also in hornworts28 and even in the eusporangiate ferns20, whereas it features four introns in the H. ensiformis mtDNA: rrnLi825g1, rrnLi833g2, rrnLi1897g1 and rrnLi1928g1 (Fig. 5). The latter two group I introns have already been documented serendipitously in sequence samplings covering parts of the mitochondrial rrnL gene60 and are conserved also outside of Pteridaceae in taxa as least as distant as tree ferns (Cyatheales, e.g. Plagiogyria stenoptera, accession DQ647877). The origins of these introns remain unclear as they neither share sequence similarities anywhere else in the plant lineage nor among fungi as is the case for the rampant invader group I intron cox1i726g1 sporadically occurring in angiosperms61,62.
Verifying native and foreign sequences in the organelle genomes
The examples of the H. ensiformis organelle genomes reported here document that parallel analysis of transcriptome along with genome NGS sequencing data is essential to characterize genes as functional or dysfunctional owing to the complex maturation processes including C-to-U and U-to-C RNA editing and cis-splicing or trans-splicing of split genes and introns. Moreover, the highly complex mitogenome of H. ensiformis is a prime case showing that very careful investigations of alien sequences in NGS data to tell them apart from native chloroplast DNA or bacterial contaminations may be needed in such complex cases of organelle genome structures. Aside from our experimental verifications, we note that parts of rickettsial DNA inserts in H. ensiformis also seem to be present in database entries reporting partial mtDNA sequences of the ferns Asplenium nidus (FR669448) and Dryopteris crassirhizoma (MW732172). Moreover, we suggest the careful re-evaluation of some fern cpDNA entries (like Dipteris conjugata KP136829, Polypodium vulgare MT984517, Cystopteris protrusa KP136830 or Selliguea yakushimensis MN623352) that seem to contain mtDNA stretches as possible artifacts. Certainly, however, once verified and when the likely similarly complex mitogenomes of those taxa would be assembled in the future they may document interesting inter-organellar DNA transfer from mitochondria to chloroplasts.
Lateral sequence transfers: cpDNA insertions in the Haplopteris mitogenome
The first evidence for lateral transfer of promiscuous chloroplast DNA into a plant mitogenome has already been documented 40 years ago63 and thereafter found to be merely a standard feature in many seed plant mtDNAs. Most interestingly, among the numerous cpDNA insertions in the mtDNA of H. ensiformis we now find examples evidently documenting ancient cpDNA features that even are not present any more in the recent plastomes, as here seen for the mysterious morffo elements. The origin and dynamics of these only recently described mobile ORFs in fern organelles25 is presently still enigmatic. Here, we find that apparently intact morffos in the cpDNA of H. elongata have evidently disintegrated in the plastome of H. ensiformis (Fig. 1b) but that their counterparts and evidence for yet other morffos is present in its mitogenome as evolution’s mis-placed witnesses.
Origins of bacterial sequences in the Haplopteris mtDNA
The role of horizontal gene transfers (HGT) is increasingly appreciated not only for bacterial evolution but also in the evolution of eukaryotic genomes including ferns64,65. In several cases, host-parasite interactions are key to the events of HGT42,66 and such interactions in nature may also be responsible for gene transfer into fern mtDNA67. In particular, after first reports on horizontal plant-to-plant transfer of mtDNAs68,69, it is meantime well understood that HGT has contributed to many seed plant mitogenomes70. The most outstanding example is the case of the early-branching flowering plant A. trichopoda having integrated into its mitogenome not only numerous stretches of mtDNA from other angiosperms but also the near-complete mitogenomes of two mosses19,71. A new dimension of HGT into plant mitogenomes opened up with the discovery of tRNA genes from chlamydial origins into very early tracheophyte lineages29. Yet more recently, it was found that sequences of fungal origin have been horizontally transferred early into mitogenomes of the Orchid family72.
Here, we now report on a multitude of Rickettsia-like genome insertions as one of the most astonishing findings emerging from the assembly and analysis of the H. ensiformis mitogenome. Intriguingly, Rickettsiales are known to be obligate intracellular parasites. Their phylogenetic relatedness to the progenitor of the eukaryotic mitochondrion is a matter of ongoing debate73,74. The bacterial DNA inserts in the H. ensiformis mtDNA are most likely derived from species closely related to (Cand.) Caedibacter acanthamoebae35 of the Holosporales (or, alternatively, Holosporaceae among Rickettsiales). Caedibacter (or Caedimonas) endosymbionts transfer the killer trait to their Paramecium hosts75,76. Other Rickettsia are known to be associated with arthropods, leeches and protists and Rickettsia-like organisms (RLOs) and are associated not only with human or animal diseases but also with numerous plant diseases, for example the Rickettsia endosymbiont of the tobacco whitefly Bemisia tabaci. Moreover, Rickettsia have also been associated with a papaya disease77 and have been identified eustigmatophyte algae78 and in the green alga Mesostigma viride79. It will be very interesting to see whether Rickettsial or related bacterial DNA insertions will also be identified in further leptosporangiate mitogenomes and to ultimately identify the exact donor species and the biological mechanisms of the HGT processes. Notably, the xenologous bacterial DNA regions identified in the H. ensiformis mitogenome include similarities to the IS481 family transposase and virB8 and virB9 homologs (Supplementary Data 3). The latter genes are commonly found among mobile IS elements and are associated with conjugative gene transfer amongst Rickettsia species80.
Conclusions
The leptosporangiate fern mitogenome reported here shows that routine pipelines for organelle genome assemblies are prone to fail when mitogenomes become as complex as in the case of the H. ensiformis mitochondrial DNA. Notably, mitogenomes were not produced during genome assemblies from NGS data obtained recently for several ferns and a lycophyte, as mentioned above. It will be highly interesting to see whether our findings will help to carefully re-consider and re-analyze the NGS data in those cases and whether lateral and horizontal gene transfers will have affected mitogenomes and complicated their assemblies to similar or possibly even higher degrees. The presence of foreign sequences arising from lateral transfer of cpDNA or horizontal transfer of bacterial DNA will demand careful re-investigation and additional experimental confirmation similar to the approaches that we have reported here. Further discoveries of rickettsial, chlamydial or possibly sequences of yet other bacterial origins will motivate to explore the physical interactions in nature of the donor bacteria and fern prothallia as the likely entry points for the horizontally transferred sequences. Similarly intriguing is the surprising retention of cis-arranged group II introns in the highly recombinant mitogenome of H. ensiformis, possibly indicating that only intron-binding proteins present in seed plants but not in fern nuclear genomes have allowed for the cis-to-trans transition of their ribozymic structures.
Methods
Plant material
The Bonn University Botanic Garden kindly provided plant material for H. ensiformis (xx-0-BONN-24687) and V. lineata (xx-0-BONN-17295). Species identities were independently verified by PCR amplification and sequencing of the rbcL and atpA locus revealing complete sequence identities with independent sequence accessions (H. ensiformis KX164999 and MH359250 and V. lineata EF473712 and KU744782).
Organelle genome sequencing and assembly
The Qiagen DNeasy Plant Mini Kit was used for DNA isolation and the Sigma-Aldrich plant RNA isolation Kit for RNA preparation followed by ribosomal RNA depletion using the RiboMinus Plant Kit for RNA‐Seq (Thermo Fisher Scientific). RNA quality was checked with an Agilent 2100 Bioanalyzer system for RIN values of at least 8.5. Genome and transcriptome sequencing (paired-end whole genome sequencing and RNAseq) was done commercially at the BGI on an Illumina platform. After further quality control of DNA and RNA samples at the BGI, paired-end reads of 150 and 100 nt length were produced on the Illumina HiSeq X Ten and Hiseq 2500 platform, respectively. Raw read data were evaluated with FastQC v0.11.9 (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). No adapter artifacts could be detected for 31,091,834 clean RNA reads with a mean sequence quality of 36 (>99.9% base call accuracy) for forward and reverse reads (Supplementary Data 6). The MEGAHIT v1.2.9 software81,82 was used for de novo whole genome assembly with independent runs using three different settings (“strict”, “default” and “relaxed”). The “strict” assembly was run with the minimum starting k-mer size and small number of increments of k-mer size iterations (--k-min 21 --k-step 16). The third assembly was run with parameters set for a high ultra-complex metagenomics dataset (--k-min 27 --k-step 10). Complete assembly of cpDNA was performed with NOVOPlasty 2.3.283 using contigs from the MEGAHIT assembly with conserved chloroplast genes as seeds. The sequencing and assembly statistics for our study are available in Supplementary Data 6. This file contains information on the quality control measures taken after sequencing, as well as details on the assembly of the organellar genomes and the identification of RNA editing sites.
RNA reads were assembled with the Trinity v2.8.2 software84,85. The BLAST 2.9.0+86 suite was used to initially identify contigs with chloroplast or mitochondrial gene content using available lycophyte and fern organelle genomes. Raw read data and assembly data have been deposited under BioProject accession no. PRJNA862965. Whole genome assembly contigs could be clustered into two sets based on k-mer coverage (mt/cp 1:10). After connecting, contigs were verified by PCRs (see below), contigs of the mitochondrial genome were connected by hand.
Verification of mtDNA arrangements
PCRs were used to independently verify the highly complex arrangements of the H. ensiformis mitochondrial DNA resulting from multiple repetitive sequences and insertions of chloroplast and xenologous DNA. PCR amplicons were designed with primers anchoring in neighboring sequence regions of evident mitochondrial identity, preferably coding regions. Special care was taken to investigate repeated sequences for potential recombination creating alternative arrangements of flanking sequences. To best avoid false positives suggesting active recombination resulting from artificial template switches we used a strategy of template-switch avoiding tsa-PCRs. To that end, a mix of gel-eluted PCR fragments containing a repeat sequence in different sequence environments (AB and CD) was used to obtain products reflecting a reciprocal exchange of flanking regions (AD and CB). A series of template dilutions (1:10, 1:20, 1:30, 1:40, 1:50 and 1:60) and numbers of PCR cycles (15, 20, 25 and 30) were tested and adjusted to determine the threshold for artificial production of template-switch products.
Identification of mitochondrial genes avoided routine pipelines but started from homologs in the mtDNA of diverse taxa, including the liverwort Marchantia polymorpha, the lycophyte P. squarrosus and the gymnosperm Ginkgo biloba. Gene identities were verified by the parallel transcriptome studies to confirm intron splicing and C-to-U and U-to-C RNA editing. Identification of tRNA genes combined the use of tRNAscan-SE87 and sensitive BLASTN searches using a tRNA query set including the recently identified chlamydial tRNA xenologues in early tracheophytes20,29. To identify DNA similarities including repeats, chloroplast DNA or xenologous bacterial DNA insertions, we used sensitive BLASTN or XBLAST similarity searches (word sizes = 7 or 3, respectively) and strict random expectancy threshold cutoffs of 1e−10. On nucleotide level, this translates approximately into identification of identical repeated sequences of ca. 40 bp (i.e. slightly larger than the conserved domain V of group II introns) or respective larger, but less similar, regions.
The primers used in our PCR experiments on the verification of mtDNA arrangements, transcriptome studies, and the determination of Rickettsial DNA insertions within the mtDNA of H. ensiformis can be found in Supplementary Data 7.
Transcriptome studies and determination of RNA editing sites
Transcriptome studies were used to determine all intron splicing sites. The identification of RNA editing sites was done as previously described28. Briefly, DNA and RNA reads were mapped against the organelle contig sequences using GSNAP88 and JACUSA89 was used to determine RNA-DNA differences. Thresholds were set to minimally 30 reads and RNA editing efficiencies of at least 1% for chloroplast and at least 5% for mitochondrial transcripts for a strict determination of C-to-U and U-to-C RNA editing sites, respectively. Selected loci were analyzed independently by RT-PCR-based cDNA analyses and sequencing for various reasons like unexpectedly inefficient RNA editing at certain sites or because of coexisting pseudogene copies as discussed under results (chloroplast rpoC1 and mitochondrial genes atp1, atp8, nad5, nad7, rrnL and the rpl6-rps13-rps11 co-transcript).
Statistics and reproducibility
H. ensiformis DNA and RNA for Illumina sequencing were prepared from the same isolate from the Botanic Garden of the University Bonn, accession 24687. PCR experiments were conducted on the same samples sent in for Illumina sequencing and in parallel for at least one other specimen of H. ensiformis accession 24687 taken from a different location in the Botanic Garden.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
Haplopteris ensiformis primary nucleotide sequence reads are submitted to the sequence read archive (SRA) under BioProject accession number PRJNA862965. The assembled chloroplast genome is deposited under accession number OM867544 and the assembled mitogenome chromosomes are available under accession numbers OM867545–OM867553. Oligonucleotide sequences for PCR experiments are provided in Supplementary Data 7. Uncropped versions of the electrophoretic gels shown in Figs. 4c and 7a have been deposited as Supplementary Fig. 5.
References
Gualberto, J. M. & Newton, K. J. Plant mitochondrial genomes: dynamics and mechanisms of mutation. Annu. Rev. Plant Biol. 68, 225–252 (2017).
Arrieta-Montiel, M. P. & Mackenzie, S. A. In Plant Mitochondria. Advances in Plant Biology, Vol. 1 (ed. Kempken, F.) 65–82 (Springer, 2011).
Cole, L. W., Guo, W., Mower, J. P. & Palmer, J. D. High and variable rates of repeat-mediated mitochondrial genome rearrangement in a genus of plants. Mol. Biol. Evol. 35, 2773–2785 (2018).
Chevigny, N., Schatz-Daas, D., Lotfi, F. & Gualberto, J. M. DNA repair and the stability of the plant mitochondrial genome. Int. J. Mol. Sci. 21, 328 (2020).
Christensen, A. C. Plant mitochondria are a riddle wrapped in a mystery inside an enigma. J. Mol. Evol. 89, 151–156 (2021).
Liu, Y. et al. The mitochondrial genome of the lycophyte Huperzia squarrosa: the most archaic form in vascular plants. PLoS ONE 7, e35168 (2012).
Grewe, F., Viehoever, P., Weisshaar, B. & Knoop, V. A trans-splicing group I intron and tRNA-hyperediting in the mitochondrial genome of the lycophyte Isoetes engelmannii. Nucleic Acids Res. 37, 5093–5104 (2009).
Hecht, J., Grewe, F. & Knoop, V. Extreme RNA editing in coding islands and abundant microsatellites in repeat sequences of Selaginella moellendorffii mitochondria: the root of frequent plant mtDNA recombination in early tracheophytes. Genome Biol. Evol. 3, 344–358 (2011).
Chaw, S.-M. S. M. et al. The mitochondrial genome of the gymnosperm Cycas taitungensis contains a novel family of short interspersed elements, Bpu sequences, and abundant RNA editing sites. Mol. Biol. Evol. 25, 603–615 (2008).
Guo, W. et al. Ginkgo and Welwitschia mitogenomes reveal extreme contrasts in gymnosperm mitochondrial evolution. Mol. Biol. Evol. 33, 1448–1460 (2016).
Kan, S.-L., Shen, T.-T., Gong, P., Ran, J.-H. & Wang, X.-Q. The complete mitochondrial genome of Taxus cuspidata (Taxaceae): eight protein-coding genes have transferred to the nuclear genome. BMC Evol. Biol. 20, 10 (2020).
Sullivan, A. R. et al. The mitogenome of Norway spruce and a reappraisal of mitochondrial recombination in plants. Genome Biol. Evol. 12, 3586–3598 (2020).
Jackman, S. D. et al. Organellar genomes of white spruce (Picea glauca): assembly and annotation. Genome Biol. Evol. 8, 29–41 (2015).
Putintseva, Y. A. et al. Siberian larch (Larix sibirica Ledeb.) mitochondrial genome assembled using both short and long nucleotide sequence reads is currently the largest known mitogenome. BMC Genomics 21, 654 (2020).
Richardson, A. O., Rice, D. W., Young, G. J., Alverson, A. J. & Palmer, J. D. The ‘fossilized’ mitochondrial genome of Liriodendron tulipifera: ancestral gene content and order, ancestral editing sites, and extraordinarily low mutation rate. BMC Biol. 11, 29 (2013).
Alverson, A. J. et al. Insights into the evolution of mitochondrial genome size from complete sequences of Citrullus lanatus and Cucurbita pepo (Cucurbitaceae). Mol. Biol. Evol. 27, 1436–1448 (2010).
Rodríguez-Moreno, L. et al. Determination of the melon chloroplast and mitochondrial genome sequences reveals that the largest reported mitochondrial genome in plants contains a significant amount of DNA having a nuclear origin. BMC Genomics 12, 424 (2011).
Sloan, D. B. et al. Rapid evolution of enormous, multichromosomal genomes in flowering plant mitochondria with exceptionally high mutation rates. PLoS Biol. 10, e1001241 (2012).
Rice, D. W. et al. Horizontal transfer of entire genomes via mitochondrial fusion in the angiosperm Amborella. Science 342, 1468–1473 (2013).
Guo, W., Zhu, A., Fan, W. & Mower, J. P. Complete mitochondrial genomes from the ferns Ophioglossum californicum and Psilotum nudum are highly repetitive with the largest organellar introns. New Phytol. 213, 391–403 (2017).
Knie, N., Grewe, F. & Knoop, V. Monilophyte mitochondrial rps1 genes carry a unique group II intron that likely originated from an ancient paralog in rpl2. RNA 22, 1338–1348 (2016).
Zumkeller, S. M., Knoop, V. & Knie, N. Convergent evolution of fern-specific mitochondrial group II intron atp1i361g2 and its ancient source paralogue rps3i249g2 and independent losses of intron and RNA editing among Pteridaceae. Genome Biol. Evol. 8, 2505–2519 (2016).
Knie, N., Grewe, F., Fischer, S. & Knoop, V. Reverse U-to-C editing exceeds C-to-U RNA editing in some ferns – a monilophyte-wide comparison of chloroplast and mitochondrial RNA editing suggests independent evolution of the two processes in both organelles. BMC Evol. Biol. 16, 134 (2016).
Guo, W., Grewe, F. & Mower, J. P. Variable frequency of plastid RNA editing among ferns and repeated loss of uridine-to-cytidine editing from vascular plants. PLoS ONE 10, e0117075 (2015).
Robison, T. A. et al. Mobile elements shape plastome evolution in ferns. Genome Biol. Evol. 10, 2558–2571 (2018).
Lenz, H. et al. Introducing the plant RNA editing prediction and analysis computer tool PREPACT and an update on RNA editing site nomenclature. Curr. Genet. 56, 189–201 (2010).
Song, M. et al. A novel chloroplast gene reported for flagellate plants. Am. J. Bot. 105, 117–121 (2018).
Gerke, P. et al. Towards a plant model for enigmatic U-to-C RNA editing: the organelle genomes, transcriptomes, editomes and candidate RNA editing factors in the hornwort Anthoceros agrestis. New Phytol. 225, 1974–1992 (2020).
Knie, N., Polsakiewicz, M. & Knoop, V. Horizontal gene transfer of chlamydial-like tRNA genes into early vascular plant mitochondria. Mol. Biol. Evol. 32, 629–634 (2014).
Lenz, H., Hein, A. & Knoop, V. Plant organelle RNA editing and its specificity factors: enhancements of analyses and new database features in PREPACT 3.0. BMC Bioinformatics 19, 255 (2018).
Oldenkott, B., Yamaguchi, K., Tsuji-Tsukinoki, S., Knie, N. & Knoop, V. Chloroplast RNA editing going extreme: more than 3400 events of C-to-U editing in the chloroplast transcriptome of the lycophyte Selaginella uncinata. RNA 20, 1499–1506 (2014).
Bégu, D., Castandet, B. & Araya, A. RNA editing restores critical domains of a group I intron in fern mitochondria. Curr. Genet. 57, 317–325 (2011).
Bégu, D. & Araya, A. The horsetail Equisetum arvense mitochondria share two group I introns with the liverwort Marchantia, acquired a novel group II intron but lost intron-encoded ORFs. Curr. Genet. 55, 69–79 (2009).
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 15, 1888–1903 (2003).
Suzuki, H. et al. Draft genome sequence of Caedibacter varicaedens, a kappa killer endosymbiont bacterium of the ciliate Paramecium biaurelia. Genome Announc. 3, 23–25 (2015).
Schuettpelz, E., Schneider, H., Huiet, L., Windham, M. D. & Pryer, K. M. A molecular phylogeny of the fern family Pteridaceae: assessing overall relationships and the affinities of previously unsampled genera. Mol. Phylogenet. Evol. 44, 1172–1185 (2007).
Schuettpelz, E. & Pryer, K. M. Fern phylogeny inferred from 400 leptosporangiate species and three plastid genes. Taxon 56, 1037–1050 (2007).
Schuettpelz, E. & Pryer, K. M. in Biology and Evolution of Ferns and Lycophytes (Ranker, T. A. & Haufler, C. H.) 395–416 (Cambridge University Press, 2008).
Grusz, A. L., Rothfels, C. J. & Schuettpelz, E. Transcriptome sequencing reveals genome-wide variation in molecular evolutionary rate among ferns. BMC Genomics 17, 692 (2016).
Rothfels, C. J. & Schuettpelz, E. Accelerated rate of molecular evolution for vittarioid ferns is strong and not driven by selection. Syst. Biol. 63, 31–54 (2014).
Fauskee, B. D., Sigel, E. M., Pryer, K. M. & Grusz, A. L. Variation in frequency of plastid RNA editing within Adiantum implies rapid evolution in fern plastomes. Am. J. Bot. 108, 820–827 (2021).
Roulet, M. E. et al. Multichromosomal structure and foreign tracts in the Ombrophytum subterraneum (Balanophoraceae) mitochondrial genome. Plant Mol. Biol. 103, 623–638 (2020).
Kozik, A. et al. The alternative reality of plant mitochondrial DNA: one ring does not rule them all. PLoS Genet. 15, 1–30 (2019).
Sloan, D. B. One ring to rule them all? Genome sequencing provides new insights into the ‘master circle’ model of plant mitochondrial DNA structure. New Phytol. 200, 978–985 (2013).
Knoop, V. Plant mitochondrial genome peculiarities evolving in the earliest vascular plant lineages. J. Syst. Evol. 51, 1–12 (2013).
Mower, J. P., Sloan, D. B. & Alverson, A. J. in Plant Genome Diversity, Vol. 1. (Wendel, J. F. et al.) 123–144 (Springer Vienna, 2012).
Sessa, E. B. et al. Between two fern genomes. Gigascience 3, 15 (2014).
Li, F.-W. et al. Fern genomes elucidate land plant evolution and cyanobacterial symbioses. Nature 4, 460–472 (2018).
Huang, X. et al. The flying spider-monkey tree fern genome provides insights into fern evolution and arborescence. Nat. Plants 8, 500–512 (2022).
Fang, Y. et al. The genome of homosporous maidenhair fern sheds light on the euphyllophyte evolution and defences. Nat. Plants 8, 1024–1037 (2022).
Marchant, D. B. et al. Dynamic genome evolution in a model fern. Nat. Plants 8, 1038–1051 (2022).
Wickell, D. et al. Underwater CAM photosynthesis elucidated by Isoetes genome. Nat. Commun. 12, 1–13 (2021).
Mower, J. P. Variation in protein gene and intron content among land plant mitogenomes. Mitochondrion 53, 203–213 (2020).
Guo, W., Zhu, A., Fan, W., Adams, R. P. & Mower, J. P. Extensive shifts from cis to trans splicing of gymnosperm mitochondrial introns. Mol. Biol. Evol. 37, 1615–1620 (2020).
Malek, O., Brennicke, A. & Knoop, V. Evolution of trans-splicing plant mitochondrial introns in pre-Permian times. Proc. Natl Acad. Sci. USA 94, 553–558 (1997).
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 4, 1599–1609 (1998).
Burger, G., Yan, Y., Javadi, P. & Lang, B. F. Group I intron trans-splicing and mRNA editing in the mitochondria of placozoan animals. Trends Genet. 25, 381–386 (2009).
Pombert, J. F. & Keeling, P. J. The mitochondrial genome of the entomoparasitic green alga Helicosporidium. PLoS ONE 5, e8954 (2010).
Pelin, A. et al. The mitochondrial genome of the arbuscular mycorrhizal fungus Gigaspora margarita reveals two unsuspected trans -splicing events of group I introns. New Phytol. 194, 836–845 (2012).
Qiu, Y.-L. et al. The deepest divergences in land plants inferred from phylogenomic evidence. Proc. Natl Acad. Sci. USA 103, 15511–15516 (2006).
Cho, Y., Qiu, Y. L., Kuhlman, P. & Palmer, J. D. Explosive invasion of plant mitochondria by a group I intron. Proc. Natl Acad. Sci. USA 95, 14244–14249 (1998).
Cusimano, N., Zhang, L.-B. & Renner, S. S. Reevaluation of the cox1 group I intron in Araceae and angiosperms indicates a history dominated by loss rather than horizontal transfer. Mol. Biol. Evol. 25, 265–276 (2008).
Stern, D. B. & Lonsdale, D. M. Mitochondrial and chloroplast genomes of maize have a 12-kilobase DNA sequence in common. Nature 299, 698–702 (1982).
Wickell, D. A. & Li, F. On the evolutionary significance of horizontal gene transfers in plants. New Phytol. 225, 113–117 (2020).
Gao, J.-G. Tracking the evolutionary innovations of plant terrestrialization. Gene 769, 145203 (2020).
Mower, J. P., Stefanović, S., Young, G. J. & Palmer, J. D. Plant genetics: gene transfer from parasitic to host plants. Nature 432, 165–166 (2004).
Davis, C. C., Anderson, W. R. & Wurdack, K. J. Gene transfer from a parasitic flowering plant to a fern. Proc. R. Soc. B Biol. Sci. 272, 2237–2242 (2005).
Bergthorsson, U., Adams, K. L., Thomason, B. & Palmer, J. D. Widespread horizontal transfer of mitochondrial genes in flowering plants. Nature 424, 197–201 (2003).
Won, H. & Renner, S. S. Horizontal gene transfer from flowering plants to Gnetum. Proc. Natl Acad. Sci. USA 100, 10824–10829 (2003).
Mower, J. P., Jain, K. & Hepburn, N. J. The role of horizontal transfer in shaping the plant mitochondrial genome. Adv. Bot. Res. 63, 41–69 (2012).
Taylor, Z. N., Rice, D. W. & Palmer, J. D. The complete moss mitochondrial genome in the angiosperm Amborella is a chimera derived from two moss whole-genome transfers. PLoS ONE 10, e0137532 (2015).
Sinn, B. T. & Barrett, C. F. Ancient mitochondrial gene transfer between fungi and the orchids. Mol. Biol. Evol. 37, 44–57 (2020).
Fan, L. et al. Reply to: Phylogenetic affiliation of mitochondria with Alpha-II and Rickettsiales is an artefact. Nat. Ecol. Evol. 6, 1832–1835 (2022).
Martijn, J., Vosseberg, J., Guy, L., Offre, P. & Ettema, T. J. G. Phylogenetic affiliation of mitochondria with Alpha-II and Rickettsiales is an artefact. Nat. Ecol. Evol. 6, 1829–1831 (2022).
Schrallhammer, M., Castelli, M. & Petroni, G. Phylogenetic relationships among endosymbiotic R-body producer: bacteria providing their host the killer trait. Syst. Appl. Microbiol. 41, 213–220 (2018).
Grosser, K. et al. More than the ‘killer trait’: infection with the bacterial endosymbiont Caedibacter taeniospiralis causes transcriptomic modulation in paramecium host. Genome Biol. Evol. 10, 646–656 (2018).
Davis, M., Ying, Z., Brunner, B., Pantoja, A. & Ferwerda, F. Rickettsial relative associated with papaya bunchy top disease. Curr. Microbiol. 36, 80–84 (1998).
Yurchenko, T. et al. A gene transfer event suggests a long-term partnership between eustigmatophyte algae and a novel lineage of endosymbiotic bacteria. ISME J. 12, 2163–2175 (2018).
Yang, A., Narechania, A. & Kim, E. Rickettsial endosymbiont in the “early-diverging” streptophyte green alga Mesostigma viride. J. Phycol. 52, 219–229 (2016).
Siguier, P., Gourbeyre, E. & Chandler, M. Bacterial insertion sequences: their genomic impact and diversity. FEMS Microbiol. Rev. 38, 865–891 (2014).
Li, D. et al. MEGAHIT v1.0: A fast and scalable metagenome assembler driven by advanced methodologies and community practices. Methods 102, 3–11 (2016).
Li, D., Liu, C.-M., Luo, R., Sadakane, K. & Lam, T.-W. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics 31, 1674–1676 (2015).
Dierckxsens, N., Mardulyn, P. & Smits, G. NOVOPlasty: de novo assembly of organelle genomes from whole genome data. Nucleic Acids Res. 9, gkw955 (2016).
Grabherr, M. G. et al. Full-length transcriptome assembly from RNA-Seq data without a reference genome. Nat. Biotechnol. 29, 644–652 (2011).
Haas, B. J. et al. De novo transcript sequence reconstruction from RNA-seq using the Trinity platform for reference generation and analysis. Nat. Protoc. 8, 1494–1512 (2013).
Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. Basic local alignment search tool. J. Mol. Biol. 215, 403–410 (1990).
Lowe, T. M. & Chan, P. P. tRNAscan-SE On-line: integrating search and context for analysis of transfer RNA genes. Nucleic Acids Res. 44, W54–W57 (2016).
Wu, T. D., Reeder, J., Lawrence, M., Becker, G. & Brauer, M. J. GMAP and GSNAP for genomic sequence alignment: enhancements to speed, accuracy, and functionality. Methods Mol. Biol. 1418, 283–334 (2016).
Piechotta, M., Wyler, E., Ohler, U., Landthaler, M. & Dieterich, C. JACUSA: site-specific identification of RNA editing events from replicate sequencing data. BMC Bioinformatics 18, 7 (2017).
Greiner, S., Lehwark, P. & Bock, R. OrganellarGenomeDRAW (OGDRAW) version 1.3.1: expanded toolkit for the graphical visualization of organellar genomes. Nucleic Acids Res. 47, W59–W64 (2019).
Darty, K., Denise, A. & Ponty, Y. VARNA: interactive drawing and editing of the RNA secondary structure. Bioinformatics 25, 1974–1975 (2009).
Minh, B. Q. et al. IQ-TREE 2: new models and efficient methods for phylogenetic inference in the genomic era. Mol. Biol. Evol. 37, 1530–1534 (2020).
Acknowledgements
The authors wish to thank the Molecular Evolution group for stimulating discussion and productive interaction, especially Philipp Gerke for his contributions to establish pipelines for the analysis of RNA editing in NGS data. The authors gratefully acknowledge the support for this project by computing time provided by the Paderborn Center for Parallel Computing (PC2). Finally, we thank the staff at the Botanical Garden of the university Bonn for providing materials and their kind support.
Funding
Open Access funding enabled and organized by Projekt DEAL.
Author information
Authors and Affiliations
Contributions
S.Z. did wet lab work and established bioinformatic pipelines. M.P. helped with nucleic acid preparations and molecular cloning. S.Z. and V.K. analyzed data and prepared figures. V.K. wrote the manuscript and all authors edited and approved the final manuscript version.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Communications Biology thanks Huan Liu and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary handling editor: George Inglis. Peer reviewer reports are available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Zumkeller, S., Polsakiewicz, M. & Knoop, V. Rickettsial DNA and a trans-splicing rRNA group I intron in the unorthodox mitogenome of the fern Haplopteris ensiformis. Commun Biol 6, 296 (2023). https://doi.org/10.1038/s42003-023-04659-8
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s42003-023-04659-8
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.
