Background

Organellar genomes are widely used as a source of genetic information in evolutionary studies, mainly due to the haploid character and the presence in hundreds to thousands of copies in each cell [1, 2]. In the majority of known organisms the mitogenomes and plastomes are maternally inherited, resulting in presence only single haplotypes of theses genomes in the organism. Several studies described heteroplasmy of plastid genomes [3], however, the most of the studies did not reveal intraindividual polymorphism [4, 5] supporting organellar genomes as a resources for evolutionary studies.

The sequences of complete mitochondrial genomes are mainly used in phylogenetics, phylogeography and population genetics of animals and fungi [6,7,8], while in plants sciences plastid genomes are mainly used for these purposes.

Compared to the seed plants, the organellar and especially plastid genomes are poorly explored in bryophytes. Up to the date only 15 plastid and 48 mitochondrial complete genome sequences are known for bryophytes genera. Moreover, most of sequenced genomes belongs to just four moss families Funariaceae [9], Grimmiaceae [10, 11], Orthotrichaceae [12,13,14,15] and Sphagnaceae [16].

The mitochondrial genomes of early land plants are known from their stable structure in comparison to the seed plants [12, 17]. The liverworts are the oldest evolutionary lineage of sporophytic plants and the most genetically diverse. However, despite high nucleotide variation at inter- and intrageneric level the gene content and order remain almost unchanged since the deepest nodes of liverworts diversification [18,19,20]. The only observed changes were the intron losses of atp1 and cox1 genes in the leafy liverworts group [20] and pseudogenization of the nad7 gene in the majority of the liverworts except of Treubia lacunosa [18].

This stability seems to be associated with the lack of repetitive sequence in the mitogenomes of early land plants, which are common in seed plants [12]. However, the mitogenomes of the liverworts are poorly explored, even in comparison to the mosses, where up-to-date complete mitochondrial genomes sequences of 6 genera are known.

The available data is even more scarce in case of chloroplast genomes, limited to the genera Marchantia, Pellia, Aneura and Ptilidium.

The liverwort Gymnomitrion concinnatum (Lightf.) Corda belongs to the family Gymnomitriaceae H. Klinggr. This group includes ten genera (Acrolophozia, Apomarsupella, Gymnomitrion, Herzogobryum, Marsupella, Nanomarsupella, Nothogymnomitrion, Paramomitrion, Poeltia, and Prasanthus), the most numerous of which are Gymnomitrion (27 species) and Marsupella (26 species) [21]. Historically, only two widespread genera (Gymnomitrion and Marsupella) were considered part of Gymnomitriaceae. Based on the circumscription of the genus Gymnomitrion presented by Váňa et al. [21], there are seven species recorded in Poland and Slovakia. Most of them grow in the Tatra Mountains (Western Carpathians). Gymnomitrion concinnatum is acidophilus, epilithic and epigeic liverwort that grows on magma (granite) and metamorphic (mostly gneissic) rocks and crystalline slates. Most often it occurs on shelves and crevices of rock walls, less often in loose alpine (and subnival) grasslands and snow-beds with predominance of bryophytes [22]. From Central Europe, phytocoenoses with high occurrence of G. concinnatum were described as Gymnomitrietum concinnati Herzog 1943 ex Philippi 1956 (class: Grimmietea alpestris Hadač et Vondráček in Ježek et Vondráček 1962) (compare [22,23,24,25]). The known genetic resources of the Gymnomitriaceae are limited to the sequences of ITS and three chloroplast loci [26, 27] of the genera Gymnomitrion, Herzogobryum, Marsupella and Prasanthus.

In the present study we sequenced, assembled, annotated and analysed organellar genomes of Gymnomitrion concinnatum, which provide new insights into evolution of mitogenomes and plastomes in liverworts.

Results & discussion

The characteristics of chloroplast genome of Gymnomitrion concinnatum

The plastome of Gymnomitrion concinnatum is 120,994 bp long with a structure typical for most plants, including a pair of IR regions (each of 8704 bp) separated by LSC (81,701 bp long) and SSC (20,179 bp) regions (Fig. 1). The plastome is almost 2000 bp longer than the second longest known leafy liverwort plastid genome of Ptilidium pulcherrimum, however length seems to be variable at the genus level. Comparative analysis of the chloroplast genomes of six Aneura pinguis cryptic species revealed that the length of the chloroplast genomes ranged from 120,698 to 121,140 bp [19]. The first sequenced liverworts plastome of Marchantia paleacea [28] and later sequenced of the same species differ in length by 390 bp. The changes in length of the plastome could have evolutionary significance, however with limited availability of liverworts plastome sequences, it is too early to conclude about possible variation of plastome size in the leafy liverworts lineage.

Fig. 1
figure 1

Gene map of the chloroplast of Gymnomitrion concinnatum. Genes inside and outside the outer circle are transcribed in counterclockwise and clockwise directions, respectively. The genes are color-coded based on their function. The inner circle visualizes the G/C content

Full size image

The GC content of the G. concinnatum plastome is 34.5% and falls within the range of other known liverworts, were GC content ranges from 26.6% in Marchantia polymorpha to 40.6% in Aneura mirabilis [29], but was higher than in Ptilidium pulcherrimum (33.2%), the first leafy liverwort for which the complete plastid genome was sequenced [30].

As in the closest known relative with a plastome sequence, P. pulcherrimum, the plastome of Gymnomitrion consist of 121 unique genes, including 81 protein-coding genes, 6 genes of unknown function (ycf genes), 4 ribosomal RNAs and 30 transfer RNAs (Additional file 1: Table S2). The gene order and content seems to be stable in leafy (Gymnomitrion concinnatum, Ptilidium pulcherrimum) and simple thalloid liverworts (Aneura pinguis, Pellia endiviifolia). Complex thalloid liverworts (Marchantia paleacea, M. polymorpha) have two more genes: cysA and cysT. Heterotrophic Aneura mirabilis has a reduced plastome due to lost or pseudogenization of genes involved in the process of photosynthesis [29].

Besides slight differences the comparative analysis of two known leafy liverworts plastomes revealed similarity in size and functionality of genes. The plastome of Gymnomitrion concinnatum is 1987 bp longer than Ptilidium pulcherrimum mainly due to over 500 bp insertion between trnH-GUC and ycf2 genes. The second insertion is located within the mentioned ycf2 gene, which is 598 bp longer in Gymnomitrion than in Ptilidium (5828 bp vs 5242 bp).

Features of Gymnomitrion concinnatum mitochondrial genome

The complete mitochondrial genome of G. concinnatum is 162,574 bp in length (Fig. 2), which is in accordance with other reported Marchantiophyta mitogenomes, i.e. Aneura pinguis [19], four Calypogeia species [20], Marchantia paleacea [31], Pleurozia purpurea [17], Treubia lacunosa [18], Tritomaria quinquedentata [20]. The length of aforementioned mitochondrial genomes ranges from 142,510 (T. quinquedentata) to 186,609 bp (M. paleacea). Overall GC content of the mtDNA is 44.7%, which is similar to other known liverworts mitochondrial genomes (42.2–47.4%). The mitogenome of G. concinnatum contains 70 unique genes, including 42 protein-coding genes, 3 ribosomal RNAs and 25 transfer RNAs (Additional file 1: Table S2), which is a typical set of mitochondrial protein-coding genes involved in respiration and protein synthesis. The phylogenetic tree constructed on the basis of 38 protein-coding sequences of mitogenomic sequences of seven liverworts is in accordance with Marchantiophyta clade phylogeny [32] (Fig. 3).

Fig. 2
figure 2

Gene map of the mitochondrion of Gymnomitrion concinnatum. Genes inside and outside the outer circle are transcribed in counterclockwise and clockwise directions, respectively. The genes are color-coded based on their function. The inner circle visualizes the G/C content

Full size image
Fig. 3
figure 3

Phylogenetic relationships among Gymnomitrion concinnatum and other liverworts based on whole set of protein-coding sequences of mitochondrial genomes. The posterior probability values are given at the nodes

Full size image

The nad7 protein-coding sequence was identified in G. concinnatum mitogenome as a pseudogene. The absence of nad7 CDS is not surprising, in view of the fact that previously conducted analyses on liverworts have shown that nad7 occurs as pseudogene in Marchantiopsida and Jungermanniopsida [17, 19, 20, 31, 33]. The only liverworts that preserved functional nad7 belong to Haplomitriopsida clade: Treubia lacunosa and Haplomitrium mnioides [18, 33]. Interestingly, the ORF of the first exon in G. concinnatum nad7 was preserved and showed 84.8 and 82.8% sequence identity to T. lacunosa and H. mnioides, respectively. None of other known liverworts mitochondrial genomes preserved ORF of the first exon of nad7 gene [17, 19, 20, 31, 33].

Other gene structure alterations observed in G. concinnatum mitogenome were intron losses in cox1 and atp1 genes. G. concinnatum cox1 contained 7 introns while other liverworts, reported earlier [17,18,19, 31], contain 9 introns (Additional file 2: Figure S1). cox1 intron loss has been recently reported in another Jungermanniales - Tritomaria quinquedentata and Calypogeia species (loss of 4 and 3 introns, respectively) [20]. All of these observations suggest that intron loss in cox1 is a feature specific to Jungermanniales species. Additionally, the atp1 gene lost both introns in G. concinnatum mitochondrial genome and is intronless. This phenomenon was also reported in previous studies on Calypogeia species, as well as Tritomaria quinquedentata and Treubia lacunosa species [18, 20].

Prediction of RNA editing sites of chloroplast genes

RNA editing is the process that alters the identity of nucleotides in RNA sequence or that add or delete nucleotides so that the mature RNA sequence differs from that defined in the genome [34,35,36]. In order to analyse possible RNA post-transcriptional modifications in protein-coding sequences, 87 plastid CDSs were investigated using PREPACT 2.0 [37]. A total number of 739 RNA editing sites were predicted in chloroplast genes of G. concinnatum. The C to U substitutions accounted for 60.1% (444 substitutions), while U to C substitutions accounted for 39.9% (295) of total RNA editing sites. Three substitutions affected ORF of two CDSs: ccsA and petD, where glutamine codon were altered to stop codon, as well as atpF, where threonine codon was altered to methionine codon which resulted in restoration of ORF of the gene. The substitution in atpF seems to be crucial for providing protein product. Considering aforementioned substitutions, the CDSs of the two genes are consistent with other liverworts. The highest RNA editing site content was observed in petL gene (4.2% of the CDS nucleotides were altered), however the coding sequence of the gene is only 96 bp length. It is worth mentioning that in 8 of 15 CDSs of subunits of photosystem II no RNA editing sites were found (Additional file 3: Table S3).

Prediction of RNA editing sites of mitochondrial genes

A total number of 222 RNA editing sites were predicted in 42 CDSs of G. concinnatum mitogenome. The C to U substitutions accounted for 76.6% (170 substitutions), while U to C substitutions accounted for 24.4% (52) of total RNA editing sites. The plastome CDSs (739 substitutions while 71,379 bp total length) contained 1.6 times as many RNA editing sites as mitogenome CDSs (222 substitutions while 33,534 bp total length). Two substitutions identified within mitogenome altered threonine codon to methionine codon which result in occurrence of start codons (nad2 and nad6 genes) and one substitution altered glutamine codon to stop codon (atp9 gene). Despite the aforementioned three substitutions the predicted translation products of the CDSs are consistent with corresponding CDSs of other liverworts. In fact due to RNA editing modifications the proper ORF of nad2 and nad6 genes are restored. The highest RNA editing site content was observed in ccmFC gene sequence (2.3% of the CDS nucleotides were altered) as well as the average RNA editing site content was the highest among cytochrome c maturation coding genes (1.5%). On the other hand, no substitutions were found in six CDSs: cox2, nad1, rpl6, rpl16, rps4, rps13 and rps19 (Additional file 4: Table S4).

Gene order and repeat sequences

Liverworts are considered to be conservative in terms of mitochondrial gene order evolution [12, 17, 18]. All complete liverwort mitochondrial genome sequences published have shown that the gene order is preserved in Marchantiophyta. These analyses included six species of liverworts from the orders Treubiales, Marchantiales, Pleuroziales, Metzgeriales and Jungermanniales, spanning the Marchantiophyta [17,18,19,20, 31]. Therefore Gymnomitrion concinnatum was expected to preserve the same mitochondrial gene order. In order to verify structural homology of G. concinnatum mitogenome with known species of liverworts, the mitogenome sequences of G. concinnatum and C. integristipula were aligned using MAUVE software [38] and then visualized as genome maps comparison using Circos [39]. The analysis revealed that G. concinnatum mitogenome structure is rearranged in comparison with other Marchantiophyta mitogenomes. In comparison with C. integrisipula, five locally collinear blocks (LCB) were identified among these two mitochondrial genomes (Fig. 4). First and last LCB (A and E) were located within the same regions on both mitogenomes. However middle three LCBs of G. concinnatum (B, C and D) were arranged in a different order. Two of these three LCBs were inverted in relation to C. integristipula mitogenome. All LCBs started and ended within intergenic spacers. The mitogenome structure was in accordance with mapping to reference and de novo assembly approaches with high coverage values observed at LCB junctions. Additionally, the aforementioned structure of mitogenome was independently confirmed with the use of PCR method (Additional file 5: Figure S2). The above observations cast doubt on the idea that mitogenome structure of liverworts or even whole Bryophyte is highly conserved as suggested in previous studies [12, 17, 18].

Fig. 4
figure 4

Comparison of G. concinnatum (left) and C. integristipula (right) mitochondrial genome structure and gene order. The outer track visualizes the genes, tRNA and rRNA order of both mitogenomes. The inner track visualizes rearrangements of G. concinnatum mitogenome and its regions (as LCBs) relatively to C. integristipula mitogenome as representative of other liverworts. The coloured links between LCBs represent location of the same regions. The strandedness of each region is also preserved: outer blocks represent the forward strand, while inner blocks the reverse strand

Full size image

Rearrangements of mitochondrial genome structure of seed plant are usually connected with occurrence of repeated sequences [40,41,42]. Repeats of different sizes have been observed in higher plants mitogenomes: large (> 500 bp), intermediate (50–500 bp) and small (< 50 bp). Large and intermediate repeats are considered to be involved in recombination of mitochondrial genomes structure [41, 43]. Such repeated sequences can pair up, recombine and form different configuration of mitogenome as a result [17]. Considering the above and the fact that mitogenome structure rearrangements among Marchantiophyta and Bryophyta have not been reported previously, further investigation on sequence repeats and recombination was undertaken in this study.

Consequently, repeated sequences, exceeding 100 bp in size, were identified in G. concinnatum mitogenome sequence. This analysis distinguished ten pairs of sequence repeats, varying from 107 bp to 566 bp length, which overall account for 3.1% of mitochondrial genome sequence. It turned out that the two longest repeats, 566 bp (97.7% sequence identity) and 435 bp length (95.9% sequence identity) were located on junctions between aforementioned LCBs. Therefore, the two repeated sequences of first pair (R1) were located on the edges of B-C and D-E LCBs, while the two repeated sequences of second pair (R2) were located on the edges of A-B and D-C LCBs (Fig. 4).

Considering unusual configuration of G. concinnatum mitogenome and location of repeated sequences, the model of recombination was proposed (Fig. 5). It is likely that two rearrangements within mitogenome have occurred one after the other. First, starting with LCBs arranged in order common for liverworts i.e. A, B, C, D, E, the rearrangement within R1 repeats has taken place resulting with following order of LCBs: A, B, D (inverted), C (inverted) and E. Next, the second rearrangement within R2 repeats has occurred resulting with following order of LCBs: A, D, B (inverted), C (inverted) and E, which is mitochondrial genome configuration identified in G. concinnatum (Fig. 4).

Fig. 5
figure 5

Probable recombination model that occurred in mitochondrial genome of Gymnomitrion concinnatum. First stage (a) represents the mitochondrial structure typical for the rest of known mitogenomes of liverworts while the last stage (e) represents mitochondrial structure that was identified in G. concinnatum. Rearrangement within R1 repeats (b) results with intermediate configuration of mitogenome (c). Second rearrangement (d), within R2 repeats, finally leads to  configuration of G. concinnatum mitogenome. The colored lines represent different LCBs. The two pairs of repeated sequences are depicted (pair of white squares and pair of white circles)

Full size image

In order to investigate if the same pattern of repeated sequences occurs within mitochondrial genomes of other species of the same division, the analyses of sequence repeats among 6 complete mitogenome sequences of aforementioned liverworts were conducted. In every mitochondrial genome sequence, except A. pinguis, two pairs of repeated sequences exceeding 400 bp and 90% of identity were found. Considering the location of repeated sequences on mitogenomes, three different pairs were distinguished (Fig. 6). Four of six species contained the same pattern of repeated sequences: R2 and R3 repeats, where R2 repeats were located within nad2-rps12 and nad4L-tatC intergenic regions while R3 repeats were located within nad5-nad4 intergenic spacer and the second intron of cob. Interestingly, M. paleacea mitogenome contained only one of aforementioned repeated sequences pairs - R3, but also R1 - identified in G. concinnatum mitochondrial sequence. It seems that mitochondrial genome of G. concinnatum share one pair (R2) identified in C. integristipula, P. purpurea, T. quinquedentata and T. lacunosa and another pair (R1) identified in M. paleacea. However further analysis revealed that R3 pair, absent in G. concinnatum mitogenome, was not completely missing in that genome. The second mate of R3 repeated sequence pair located within second intron of cob gene was still present while deletion exceeding 600 bp within nad5-nad4 intergenic spacer caused disappearance of the first mate of R3 pair.

Fig. 6
figure 6

Location of repeated sequences within alignment of six mitogenomes of liverworts. Three pairs of repeated sequences were identified: R1 (blue blocks), R2 (pink blocks) and R3 (green blocks). The yellow blocks depict coding-sequences of genes on the consensus sequence. The mitochondrial genome of G. concinnatum have been artificially rearranged in order to obtain the same mitogenome structure as in other species

Full size image

Considering the unique pattern of repeated sequences of G. concinnatum mitogenome, it seems plausible that described patterns of repeated sequences might be crucial for maintenance of gene order and the variations of specific patterns may cause recombination of mitochondrial genome and therefore gene order rearrangement. Up to now variations of relative placement of the genes within mitochondrial genomes caused by homologous recombination were identified only among seed plants. The mitochondrial gene order rearrangements have been found on different levels of phylogenetic classification of these plants [8, 41, 44]. The results obtained here provide the first evidence of rearrangements in the structure and gene order of widely considered as evolutionary stable mitochondrial genomes of liverworts.

Conclusions

The results obtained in this study provide the overview of mitochondrial and chloroplast genome structure and gene order of Gymnomitrion concinnatum against the background of known organellar genomes of liverworts. The complete organellar genome sequences of G. concinnatum were fully sequenced for the first time extending the knowledge of the poorly explored organellar genomes of bryophytes. Almost all aspects of organellar genomes evolution such as diversity in gene content, genome size and sequence similarity, seems to be rather conservative in G. concinnatum and highly similar to other liverworts.

Nonetheless the most relevant finding of the study, in contrast, is the discovery of rearrangements in the structure and gene order of G. concinnatum mitochondrial genome. It is the first case of mitochondrial recombination among liverworts. The recombination activity of plant mitochondria plays a major role in the evolution of mitochondrial genomes [45]. Previous studies have concluded that the mitogenome of liverworts exhibits conservative evolution contrary to highly dynamic evolution in seed plants [17]. The findings provided by this study strongly support the hypothesis that gene order rearrangements of mitochondrial genome structure are not just limited to seed plants which is in opposite to generally accepted statement that the mitogenome gene order of liverworts is constant [12]. Considering the fact that up to now only 7 mitogenomes of liverworts, including G. concinnatum, have been fully sequenced it is likely that other mitochondrial genomes of this plant group may be also rearranged in their structure and gene order. However further research providing knowledge on organellar genomes of another species of bryophytes need to be conducted to test above hypothesis.

Methods

Plant material

The specimen details were as follow: Gymnomitrion concinnatum (Lightf.) Corda, Slovakia, High Tatra Mountains, Výsňé Kôprovické sedlo pass (Liptovské kopy), 49.19040°N, 19.96641°E, alt. 1910 m a.s.l., fine-grained screen with high participation of lichens, leg., det. P. Górski, 4.09.2014. The DNA was extracted using the Zymo Plant/Seed DNA kit (Zymo Research, Irvine, CA, USA). One individual from a one year old herbarium specimen was ground with silica beds in a MiniBead-Beater homogenizer for 50 s and subsequently processed according to the manufacturer protocol. DNA quantity was estimated using Qubit fluorometer and Qubit™ dsDNA BR Assay Kit (Initrogen, Carsbad, NM, USA). DNA quality was checked by electrophoresis in 0.5% agarose gel stained with Euryx Simple Save (Eurx, Gdańsk, Poland).

Genome sequencing, assembly and annotation

The genomic library was constructed with TruSeq Nano DNA kit (Illumina, San Diego, CA, USA) and was sequenced using HiSeqX (Illumina) to generate 150 bp paired-end reads at Macrogen Inc. (Seoul, Korea) with 350 bp insert size between paired-ends. Afterwards, 20,166,426 sequencing reads were cleaned by removing the adaptor sequences and low-quality reads with Trimmomatic v0.36 [46]. The filtered reads were assembled using SPAdes 3.12.0 [47]. Reference mitogenome sequence of Calypogeia integristipula (NC035977.1) and plastome sequence of Ptilidium pulcherrimum (NC015402.1) were used to identify organellar genomes of G. concinnatum among the generated contigs. To verify de novo assembly iterative mapping was carried out independently for each genome using Geneious R8 software [48]. The chloroplast and mitochondrial genome sequences of G. concinnatum had 1440x and 264x coverage depth, respectively.

Genes were identified and annotated based on the closest known organellar genomes of related species to G. concinnatum i.e., Calypogeia integristipula, Tritomaria quinquedentata, Pleurozia purpurea, Ptilidium pulcherrimum and Aneura pinguis. Predictions were made using Geneious R8 software [48] and BLAST+ 2.8.0 tool [49]. Annotated sequences of G. concinnatum chloroplast and mitochondrial genome were submitted to GenBank under MH705066 and MH705065 accession number, respectively. Circular genome maps were created using the OGDraw software [50]. To verify gene order of G. concinnatum mitogenome the mitogenome sequences of G. concinnatum and C. integristipula were aligned using Mauve 2.3.1 [38]. The comparative analysis of the two mitogenomes was visualized using Circos plot [39].

PCR analysis

The junction regions between rearranged LCBs of G. concinnatum mitogenome were confirmed using PCR. Four primers pairs were designed based on the nucleotide sequences that overlap edges of four LCBs pairs. The sequences of primers with expected amplicons lengths are given in Additional file 6: Table S1. PCR reactions were performed in 25 μL of a reaction mixture containing 20 ng of DNA, 1x PCR buffer, 1.5 mM MgCl2, 200 μM dNTP (dATP, dGTP, dCTP, dTTP), 1.0 μM of each primer and 1 U RUN polymerase (A&A Biotech, Gdańsk, Poland). Reactions were performed under the following thermal conditions: (1) initial denaturation—4 min at a temperature of 94 °C; (2) denaturation—45 s at 94 °C; (3) annealing—50 s at 53 °C for, (4) elongation–60 s at 72 °C; (5) final elongation—7 min at 72 °C. Stages 2–4 were repeated 30 times. PCR products were separated in the QIAxcel capillary electrophoresis system, using the Qiaxcel High Resolution Kit; with the 15–3000 bp alignment marker (Qiagen) and 100–2500 DNA size marker. Standard OM500 settings were used as the electrophoresis program. The size of the obtained amplicons were determined by using BioCalculator software (Qiagen).

Phylogenomics reconstruction

Mitogenomic sequences of seven liverworts i.e. Aneura pinguis, Calypogeia integristipula, Marchantia paleacea, Pleurozia purpurea, Treubia lacunosa, Tritomaria quinquedentata available in GenBank and Gymnomitrion concinnatum presented in this study, were used for the phylogenetic analysis. First, a set of 38 protein-coding sequences (Additional file 1: Table S2), present in each mitogenomes, were extracted, concatenated and aligned using Geneious R8 [48] and MAFFT [51]. Next, based on the alignment, Bayesian analysis was conducted using MrBayes 3.2.1 [52], including M. paleacea as an outgroup. The MCMC algorithm was run for 2,000,000 generations (sampling every 1000) with four incrementally heated chains. The first 1000 trees were discarded as burn-in. The remaining trees were used to generate the consensus tree.

Prediction of RNA-editing sites

To predict editing sites within protein-coding sequences of 87 chloroplast and 42 mitochondrial genes, PREPACT 2.0 [37] tool was used with 0.001 e-value cutoff.