当前位置: 首页 > 期刊 > 《分子生物学进展》 > 2005年第1期 > 正文
编号:11175501
Ancestors of Trans-Splicing Mitochondrial Introns Support Serial Sister Group Relationships of Hornworts and Mosses with Vascular Plants
http://www.100md.com 《分子生物学进展》
     * Institut für Zellul?re und Molekulare Botanik, Abt. Molekulare Evolution, Universit?t Bonn, Bonn, Germany; Abt. Molekulare Botanik, Universit?t Ulm, Ulm, Germany

    Correspondence: E-mail: volker.knoop@uni-bonn.de.

    Abstract

    Some group II introns in the organelle genomes of plants and algae are disrupted and require trans-splicing of the affected exons from independent transcripts. A peculiar mitochondrial nad5 gene structure is universally conserved in flowering plants where two trans-splicing introns frame a tiny exon of only 22 nucleotides, and two additional conventional group II introns interrupt the nad5 reading frame at other sites. These four introns are absent in the liverwort Marchantia polymorpha, which carries a group I intron at an unrelated site in nad5. To determine how intron gains and losses have sculptured mitochondrial gene structures in early land-plant evolution, we have investigated the full nad5 gene structures in the three bryophyte classes and the fern Asplenium nidus. We find the single Marchantia group I intron nad5i753 present as the only intervening sequence in both closely (Corsinia and Monoclea) and distantly related (Noteroclada, Bazzania, and Haplomitrium) liverwort genera. In a taxonomically wide spectrum of mosses (Sphagnum, Encalypta, Timmia, Ulota, and Rhacocarpus); however, we additionally identify the angiosperm-type group II introns nad5i230 and nad5i1455. The latter is a cis-arranged homolog to one of the two angiosperm trans-splicing introns, notably the first of its kind in mosses. In the hornwort Anthoceros, the "moss and liverwort–type" group I intron nad5i753 is absent, and, besides nad5i230 and nad5i1455, intron nad5i1477 is present as the second ancestral group II intron which has evolved into a trans-splicing arrangement in angiosperms. The influence of highly frequent RNA editing, most notably in the genera Haplomitrium, Anthoceros, and Asplenium, on phylogenetic tree construction is investigated and discussed. Taken together, the data (1) support a sister group relationship of liverworts as a whole to all other embryophytes, (2) indicate loss of a group I and serial entries of group II introns in the nad5 gene during early evolution of the nonliverwort lineage, and (3) propose a placement of hornworts as sister group to tracheophytes.

    Key Words: Bryophytes ? evolution ? phylogeny ? group I and group II introns ? RNA editing

    Introduction

    Complexity and large sizes are among the prime reasons why only five complete land-plant (embryophyte) mitochondrial DNA sequences are available: the liverwort Marchantia polymorpha (Oda et al. 1992) and the angiosperms Arabidopsis thaliana (Unseld et al. 1997), sugar beet (Kubo et al. 2000), rice (Notsu et al. 2002), and, most recently, rapeseed (Handa 2003). The most significant factor contributing to the 10-fold to 100-fold size increase of plant versus animal chondriomes are intergenic regions, but additional genes and introns also contribute their share. Some two-dozen organellar group II and group I introns are components of a typical land-plant (embryophyte) chondriome. The mitochondrial introns have gained additional interest because their presence serves to identify or to confirm clades at higher taxonomic ranks in phylogenetic analyses. A clade of nonliverwort embryophytes (NLE) has been identified by the occurrence of three mitochondrial introns (Qiu et al. 1998), and the distinction of the three bryophyte classes mosses, liverworts, and hornworts has been supported by differential occurrence of mitochondrial introns (Pruchner et al. 2002). The complete chondriome sequences of algae closely related to the embryophyte lineage, such as those of the genera Chara (Turmel et al. 2003) and Chaetosphaeridium (Turmel, Otis, and Lemieux 2002a), are extremely important for phylogenetic rooting purposes. The general lack of orthologous introns in the mtDNA of the algae suggests a major invasion of novel introns into mitochondrial coding sequences coinciding with the establishment of the earliest land-plant lineages.

    Besides the increased proportion of noncoding DNA, recombinational activity in angiosperm chondriomes is a complicating factor that results in genomic complexity, sometimes as coexisting alternative genomic arrangements of the mtDNA. A peculiar outcome of recombination events on evolutionary timescales is the disruption of the three mitochondrial genes nad1, nad2, and nad5 in a total of five group II intron sequences, which have led to trans-splicing arrangements in angiosperms. The nad genes encode subunits of complex I of the respiratory chain, the NADH-ubiquinone-oxidoreductase. The origins of those trans-splicing introns in land-plant evolution have been traced and have led to the identification of cis-arranged ancestors in early branching plant lineages: moniliformopses, lycophytes, and, in one case, a hornwort (Malek, Brennicke, and Knoop 1997; Malek and Knoop 1998). Additionally, the terminal intron in nad1 has been subject to independent cis-to-trans transition among angiosperms later in evolution (Qiu and Palmer 2004).

    The nad5 gene structure of angiosperms is particularly unique: two trans-splicing introns frame a small exon of only 22 nt (Knoop et al. 1991). The large upstream exon has been introduced as a phylogenetically informative locus, and new introns in nad5 have shown up in the analysis of pteridophytes (Vangerow, Teerkorn, and Knoop 1999) and in a hornwort (Beckert et al. 1999). Subsequent studies have integrated this internal region of nad5 into multigene analyses that have identified Charales instead of Coleochaetales as the immediate extant sister group to embryophytes (Karol et al. 2001).

    Because the phylogenetic screenings so far excluded the four angiosperm introns in nad5, we have reasoned that a more complete analysis of the large nad5 gene may reveal further valuable cladistic information. We here report the results in the bryophyte classes and find further support for a deep liverwort versus nonliverwort dichotomy in land plants. Two of the angiosperm-type group II introns are shared with mosses, one of which (nad5i1455) is the first cis-arranged moss homolog to a trans-splicing intron of angiosperms. A tracheophyte-hornwort clade finds support in the loss of a group I intron invariably present in mosses and liverworts (nad5i753) and gain of another group II intron (nad5i1477) that later evolves into a trans-splicing status in spermatophytes.

    Materials and Methods

    Molecular Work

    Total nucleic acids were extracted from plant material with the Plant DNeasy kit (Qiagen) according to the protocol supplied by the manufacturer. Primers bordering the extended nad5 region UD (fig. 1) are n5up1 (5'-GCAGGNTTTT TYGGNCGTTT TCT-3') and n5do (5'-AACATNRCAA AGGCATAATG ATA-3'). Two primers pairing in the KL region analyzed earlier were used alternatively to amplify the upstream region in combination with n5up1: nad5-1 (5'-CTCCAGTYAC CAACATTAGC ATAAAAAAAG T-3') or nad5-2 (5'-GCRAGWCCWA CTCCCTCCCA TCC-3'), respectively. Primer Hans (5'-TGTCATGATG CGCCCATTCT TATGG-3') pairing in the KL region was used in combination with n5do to amplify the downstream region encompassing nad5i1455. PCR amplification assays contained 1 μl template DNA (approximately 10 ng–1 μg), 10 μl 10xPCR buffer (100 mM Tris/HCl pH 8.85, 250 mM KCl, 50 mM (NH4)2SO4, and 20 mM MgSO4), 250 mM of each dNTP, 0.25 μg of each primer, 2.5 U DNA polymerase, and double distilled water to 100 μl. Different commercially available thermostable DNA polymerases were used; for example, a mixture (90:1) of Taq DNA Pol (Gibco BRL) and Pwo DNA Pol (Boehringer Mannheim). A typical amplification assay included an initial denaturation (5 min at 94°C) followed by 35 cycles with 1 min denaturation at 94°C, 1 min annealing at 50°C to 55°C, and 2 min 30 s synthesis at 72°C and a final step of synthesis for 6 min at 72°C. Genomic DNA of Anthoceros agrestis was cut with BamHI, cloned into pBluescript SK II + (Stratagene), and a positive clone was retrieved by hybridization as an alternative strategy when PCR amplification attempts had failed repeatedly (see Results). PCR fragments were blunt-end ligated into pBlueskript II SK+ (Stratagene) or sequenced directly. Positive clones were sequenced with a Thermosequenase kit (Amersham) using Cy5-fluorescence labeled oligonucleotides and run on an Alf Express sequencer (Pharmacia).

    FIG. 1.— The nad5 gene structure with group II introns depicted as ellipses and group I intron i753 as a square. Group II introns i230, i1872, i1455, and i1477 are conserved in angiosperms (gray shading), the latter two in a trans-disrupted arrangement. A central region of nad5 (K to L) excluding the angiosperm-type introns has been used previously in phylogenetic analyses and has revealed nad5i753 as universally conserved in mosses and liverworts, nad5i1242 conserved in lycophytes and moniliformopses, nad5i391 present only in the lycopod Huperzia selago, and nad5i881 only in the hornwort Anthoceros punctatus. Subject of the present study is the extended nad5 region (U to D).

    Phylogenetic Analyses

    In a complete alignment, intron sequences of nad5i230 (1,113 nt), i753 (911 nt), and i1455 (2,831 nt) were added separately to the coding region and were fused into a single NEXUS formatted file (6,886 nt). Sequences of the angiosperm trans-splicing introns were included up to the ends of conserved homology. Nonhomologous and large-indel regions in the introns were excluded in phylogenetic analyses. RNA editing sites were annotated for optional exclusion in the phylogenetic analyses. Phylogenetic tree construction was done with the PAUP* version 4.0b10 software package (Swofford 2003) and the MrBayes version 3.0b4 program (Huelsenbeck and Ronquist 2001). Node confidences were determined by sampling 10,000 bootstrap replicates of maximum-parsimony searches (five random sequence additions each) or 4,900 trees (four chains; 500,000 generations; burn-in set to 100 trees) in Bayesian searches, respectively. Trees were drawn with TreeView version 1.6.1 (author R.D.M. Page, http://taxonomy.zoology.gla.ac.uk/rod/treeview.html).

    Results

    The nad5 gene structure with all intron insertion sites known so far in the plant lineage is shown in figure 1. A nomenclature that numbers organelle introns according to the preceding homologous nucleotide in the continuous reading frame of Marchantia polymorpha (Dombrovska and Qiu 2004) is adapted here. Four group II introns in nad5 are universally conserved in angiosperms. Two of these are conventional cis-arranged introns (nad5i230 and nad5i1872), two others (nad5i1455 and nad5i1477) are disrupted into trans-arrangements (Knoop et al. 1991; Pereira et al. 1991). Cis-arranged homologs of the latter two have been identified in the fern Asplenium (nad5i1455) and in the lycophyte Isoetes and the hornwort Anthoceros (nad5i1477), respectively (Malek and Knoop 1998). Independently, a large central part of the nad5 gene, the KL region between i230 and i1455 (fig. 1), has been introduced for phylogenetic analyses in bryophytes and pteridophytes. A group I intron, nad5i753, has been found universally conserved in all mosses and liverworts (Beckert et al. 1999), and group II intron nad5i1242 has been discovered in pteridophytes (Vangerow, Teerkorn, and Knoop 1999). Two further introns have been discovered in the course of those studies but are taxonomically of much more restricted occurrence. Group II intron i881 has been identified in the hornwort Anthoceros punctatus (Beckert et al. 1999) and nad5i391, a paralog of nad5i1242, has been identified in the lycophyte Huperzia selago (Vangerow, Teerkorn, and Knoop 1999).

    With the present study, we have tried to maximize the phylogenetic analysis of nad5 by designing PCR primers matching terminally conserved sequences. Intron sizes made the design of further internal primers to amplify the gene region necessary (see Materials and Methods). In Charales algae, none of the eight introns shown in figure 1 is present, as is evident from the full Chara vulgaris chondriome sequence now available (Turmel, Otis, and Lemieux. 2003), and as we can confirm with the sequences of the closely related Charales genera Lamprothamnium and Nitella.

    Among mosses, we have chosen to include both the basal branching genus Sphagnum and two (basal) members each from the derived clusters Hypnanae (Ulota and Rhacocarpus) and Dicrananae (Encalypta and Timmia) in our taxon sampling (table 1). Among liverworts, we have analyzed marchantiid liverworts (complex thalloids), considered to be closely and less closely related to Marchantia polymorpha (Corsinia and Monoclea, respectively), the simple thalloid and leafy liverworts from the classically defined orders Metzgeriales (Noteroclada) and Jungermanniales (Bazzania), and the isolated order of as yet unclear affiliation, Calobryales (Haplomitrium). PCR amplification from DNAs of hornwort and fern genera proved to be notoriously difficult, possibly because of the high degree of RNA editing in nad5 (see below), which makes primer design a guesswork. Even multiple rounds of primer redesigning did not allow for successful amplification of the nad5 region covering intron i1455 in hornworts. However, we have been able to identify a 4.7-kb genomic clone from a random BamHI library of Anthoceros agrestis DNA by hybridization that overlaps flanking PCR-derived Anthoceros sequences.

    Table 1 Taxon Sampling

    As in Marchantia polymorpha (Oda et al. 1992), angiosperm-type group II intron nad5i230 is likewise absent in all five liverworts we have investigated now (fig. 1). In contrast, we have found intron nad5i230 universally conserved in all five moss genera. Among mosses, intron size ranges from 831 nt in Encalypta streptocarpa to 875 nt in Sphagnum fallax. Intron nad5i230 is of very similar size in angiosperms (829 nt in Arabidopsis thaliana). The intron is conserved in gymnosperms where sizes can exceed 1,000 nt (Wang, Tank, and Sang 2000). To bridge the phylogenetic gap to seed plants, we have included the fern Asplenium nidus in our studies, for which some sequence information of downstream regions was already available (see below). Expectedly, nad5i230 is present in Asplenium and is 1,022 nt long. Most notably, intron nad5i230 is also present in the hornwort Anthoceros agrestis (855 nt). Hence, nad5i230 is clearly a new example of a nonliverwort embryophyte intron. Strikingly, all nad5i230 intron sequences consistently feature the same mysterious deviations from the group II consensus structures that had also been observed in the angiosperms: The 5' terminus TGGCG replaces the canonical GTGCG motif, and a GY is present instead of AY at the 3' end (fig. 2A). A high degree of sequence conservation in the 5' intron region and around intron domains V and VI further indicates a stable vertical transmission.

    FIG. 2.— Group II introns conserved in the nad5 genes of nonliverwort land plants: nad5i230 (A), nad5i1455 (B), nad5i1477 (C), and nad5i1872 (D). Shown are the terminal regions of conservation in the sequence alignment of the taxon set under consideration up to the first indel in domain I and after the last indel in domain IV. Sequence variations are indicated by the IUPAC ambiguity code. The branch-point A in domain VI for lariat formation during splicing is shown in italics. Size variability in domain VI and size ranges for the internal region of domains I through IV in nad5i230 and nad5i1872 are indicated. No sizes can be attributed to the trans-splicing introns nad5i1455 and nad5i1477 in angiosperms, but the respective numbers of nucleotides in Anthoceros are given.

    Group I intron nad5i753 is a universally conserved group I intron in mosses and liverworts (fig. 1), present in more than 50 genera investigated without exception (Beckert et al. 1999). Expectedly, it is also present in the nad5 gene of the new moss and liverwort genera analyzed now. As observed before, the group I intron is significantly smaller in liverworts (670 nt in Monoclea to 685 nt in Noteroclada) than in mosses (826 nt in Rhacocarpus to 856 nt in Sphagnum). To answer whether absence of nad5i753 may be unique in the hornwort genus Anthoceros, we have analyzed this nad5 gene region for other hornwort genera and have been able to obtain sequences of partial PCR products from Megaceros, Phaeoceros, and Notothylas DNA clearly showing a continuous reading frame and absence of nad5i753 (at the same time also demonstrating absence of intron i881 [fig.1], apparently a unique intron gain in Anthoceros punctatus). Thus, nad5i753 is invariably present in mosses and liverworts but absent in all other land-plant groups.

    Group II intron nad5i1455 is trans-disrupted in angiosperms (fig. 1). A cis-arranged ortholog of this intron (1,825 nt long), the presumable ancestral state, has already been identified in the fern Asplenium nidus in a screening for cis-arranged counterparts (Malek and Knoop 1998). At this position, the nad5 reading frame is continuous in Marchantia polymorpha and this is confirmed for all other liverworts we have investigated now. However, we find intron nad5i1455 to be present in cis-arranged form and to be of significant size in all mosses. Sizes of intron nad5i1455 in mosses range from 2,623 nt in Encalypta to 2,676 nt in Sphagnum. Absence of an intron in mosses at this position was erroneously deduced from approximately cDNA-sized products obtained in PCR amplifications, in hindsight probably because of mispriming in the long intron sequence when the short downstream exon of 22 nt precluded design of alternative primers in that screening strategy (Malek and Knoop 1998). In fact, PCR amplification of this gene region still failed in hornworts, but we have been able to identify a genomic clone from a random library of Anthoceros agrestis DNA, encompassing nad5i1455, which is 2,274 nt long in this species. Again, a highly atypical sequence deviating from the group II consensus with the NAY 3' end being replaced by ACC is conserved between mosses, the hornwort, the fern, and angiosperms (fig. 2B).

    Group II intron nad5i1477 is likewise trans-disrupted in angiosperms (fig. 1). An ortholog of this intron in cis-arranged form (2,391 nt in size) has been identified in Anthoceros crispulus (Malek and Knoop 1998). Again, the short exon between nad5i1455 and nad5i1477 had precluded alternative upstream primer design. Nevertheless, correct priming in the upstream exon combined with downstream mispriming in the intron clearly determines its presence also in the hornwort genera Phaeoceros and Notothylas, as we found in partial PCR products. Hence, nad5i1477 is a common denominator of this bryophyte class and shared with angiosperms (fig. 2C). As in Marchantia, intron nad5i1477 is absent in all liverworts and all mosses we have investigated now.

    Group II intron nad5i1872 is present in all angiosperm nad5 genes investigated and is in the size range of 914 (Vicia faba) to 1,108 nt (Oenothera berteriana). Some partial sequences deposited in the database clearly demonstrate that intron nad5i1872 is also present in gymnosperms (Liepelt, Bialozyt, and Ziegenhagen 2002). We found it absent in all bryophytes we have investigated. However, we have now identified this intron in Asplenium nidus. Therefore, nad5i1872 is obviously a gain in the tracheophyte lineage. Strikingly, the Asplenium intron is only 556 nt long. As in the other group II introns, there is significant sequence conservation of 5' region and domains V and VI (fig. 2D), again making a vertical mode of transmission very likely.

    The Coding Region and RNA Editing

    The now analyzed nad5 region extends over 601 codons (600 in Charales algae). As previously found, the mitochondrial sequences are highly conserved in the marchantiid liverworts and without any requirement for the C to U type of RNA editing typically observed in other land-plant groups (Steinhauser et al. 1999). In fact, the encoded polypeptides are identical in Marchantia and Corsinia. Four codon changes cannot be corrected by RNA editing in Monoclea. Typical plant organellar RNA editing by C-U pyrimidine exchanges is required in all other plants to reconstitute conserved codon identities (fig. 3). A total of 173 such positions are identified in the alignment, 95 of which are unique to a single species; the remaining 78 are shared between at least two taxa, mostly between Anthoceros and Asplenium. Only moderate levels of RNA editing are expectedly observed for the mosses and the jungermanniid liverworts. However, in stark contrast, we find that the now analyzed mitochondrial sequences of the isolated liverwort genus Haplomitrium present a surprise: overall 65 sites in the nad5 sequence require C to U RNA editing to re-establish conserved codon identities (fig. 3). Both the fern Asplenium and the hornwort Anthoceros show the usual requirement of highly frequent RNA editing in both directions of pyrmidine exchange (fig. 3). Most importantly, many of these sites are shared between the hornwort and the fern and bear the obvious danger of homoplasies in phylogenetic analyses.

    FIG. 3.— Alignments of amino acid translations of the nad5 UD region investigated for six selected species, produced with the GeneDoc software (authored by K. Nicholas, http://www.cris.com/ketchup/genedoc.shtml). Intron insertion sites and intron phases are indicated above the affected codons. C-to-U–type and U-to-C–type of exchanges by RNA editing needed to reconstitute codon identities conserved in the alga Chara are highlighted by white letters on gray or black backgrounds, respectively. Frequent RNA editing is evident for the liverwort Haplomitrium, the hornwort Anthoceros, the fern Asplenium, and, to a lesser extent, the moss Sphagnum and the angiosperm Arabidopsis included for comparison.

    Phylogenetic Information in Exons and Introns

    To address whether the presence of nad5 introns in the basal embryophyte clades could be mapped conclusively and parsimoniously onto a phylogenetic tree topology, we have used the new nad5 sequences for phylogenetic tree construction. Sequences of the two model angiosperms Arabidopsis thaliana and Oryza sativa and the algae Chara vulgaris (Charales) and Chaetosphaeridium globosum (Coleochaetales) were taken from the database and added to our data set.

    We have performed maximum-parsimony (MP) analyses and Bayesian-likelihood analyses in parallel. Because of extreme sequence conservation in the marchantiid liverworts, the phylogenetic analyses of the exons alone completely lacks resolution in this subclade. When intron sequences were included (4,687 sites, 1,304 parsimony informative), we obtained a single most-parsimonious tree (3,777 steps) with good bootstrap support for most nodes (fig. 4). However, the lack of introns conserved across all three bryophyte classes and their general absence in the algae restricts a reasonable sequence-based analyzes for out-group rooting to the coding region (1,726 sites, 570 parsimony informative). Topological differences between the eight MP trees obtained for the coding-only region (1,721 steps) are restricted to arrangements within the liverworts and mosses, whereas bootstrap analysis supports the nodes of the phylogenetic backbone. Somewhat expectedly, exclusion of the intron sequences from the phylogenetic analyses decreases bootstrap support for internal nodes within the mosses and liverworts. The posterior probabilities for nodes obtained in a Bayesian analysis conducted in parallel largely coincide with MP bootstrap support (fig. 4). Exceptions are Haplomitrium, which is placed basal to the Bazzania-Noteroclada clade, and Timmia, which is placed basal to the Rhacocarpus-Ulota clade, in the Bayesian analysis.

    FIG. 4.— Single most-parsimonious tree (3,777 steps) found for phylogenetic analysis of the extended nad5 UD region excluding shared RNA editing positions and ambiguous intron regions (4,687 characters, 1,304 parsimony informative sites). Node significances are given as bootstrap support (10,000 replicates, five random taxon additions in each replicate) for the intron-included data set (first value), the exon-only data set (second value), and as posterior probabilities (third value) in a Bayesian analysis (GTR++I model of sequence evolution; four chains, 500,000 generations, 5,000 trees sampled, 100 discarded as burn-in). Intron presence within respective clades is indicated by gray-shaded boxes.

    The reinclusion of the 78 shared RNA editing sites leads to the artifact of Anthoceros and Asplenium being joined in a clade as sister group to the angiosperms (obviously a type of long-branch attraction through this unique type of homoplasies) but leaving the remaining nodes unaffected. The exclusion of the 95 RNA editing sites unique in individual taxa, on the other hand, only results in shorter terminal branches but leaves the topology unchanged.

    Most importantly, the phylogenetic analysis results in a topology that places hornworts as a sister group to the vascular plants and mosses as the sister clade to the hornwort-tracheophyte group. This topology indeed explains nad5 intron gains and losses parsimoniously (fig. 4). Gain of intron nad5i1477 and loss of group I intron nad5i753 can be interpreted as a synapomorphy of the joint hornwort-tracheophyte clade. The presence of introns nad5i230 and nad5i1455 exclusively shared between mosses, vascular plants, and hornworts to the exclusion of liverworts strongly adds characters to the suggested dichotomy of liverworts and nonliverwort embryophytes and makes other topologies less likely. For example, a hornwort-basal topology with a moss-liverwort clade sister to tracheophytes (Nickrent et al. 2000) would require eight events of nad5 intron gains and losses instead of six, as shown in figure 4.

    Discussion

    Trans-splicing group II introns have first been described in the chloroplast rps12 gene of land plants. Their discoveries have been made nearly simultaneously through the emerging plastome sequencing projects of the liverwort Marchantia polymorpha (Kohchi et al. 1988) and the angiosperm Nicotiana tabacum (Koller et al. 1987), hence immediately indicating an evolutionary ancient development, as later confirmed by further plastome sequences showing general conservation of this trans-splicing intron among land plants.

    Trans-splicing group II introns are not exclusively restricted to embryophytes. Recently two trans-splicing group II introns have been reported in the nad3 gene of Mesostigma viride (Turmel, Otis, and Lemieux 2002b). However, no orthologs of these introns are known in other species, and, consequently, their occurrence is, as yet, not of cladistic value for a phylogenetic placement of this alga. Earlier, trans-splicing had already been reported for the chloroplast psaA gene of the alga Chlamydomonas reinhardtii (Choquet et al. 1988), an organism that is clearly yet much more distantly related to land plants. However, as in the Mesostigma example, the lack of orthologs in related clades precludes its use in cladistic analyses. Mechanistically, however, one of the Chlamydomonas psaA introns is particularly interesting, as it is broken twice, thus, requiring three independent RNAs for splicing (Goldschmidt-Clermont et al. 1991). A similar case of tripartite intron reassembly has later been reported for nad5i1477 in the evening primrose (Knoop, Altwasser, and Brennicke 1997).

    Trans-splicing in the mitochondrial lineage of land plants affects five group II introns (nad1i394, nad1i669, nad2i542, nad5i1455, and nad5i1477) that are disrupted in all angiosperms. Interest in the evolution of trans-splicing has led us to identify cis-arranged precursor introns in early branches of land-plant evolution (Malek, Brennicke, and Knoop 1997; Malek and Knoop 1998). In summary, the picture has emerged that these introns have entered new genomic loci early in the evolution of tracheophytes or, as is now clearly demonstrated, even in the bryophyte lineages. None of these introns is present in Marchantia or other liverworts.

    In contrast to the nad5 introns investigated here, no homologs of the trans-splicing introns in nad1 (Dombrovska and Qiu 2004) and nad2 (Pruchner et al. 2002) mentioned above are present in mosses or hornworts. However, one additional group II intron in nad1, nad1i728, which is subject to independent transitions into trans arrangements now shown to occur frequently and independently among angiosperms (Qiu and Palmer 2004), is conserved as a cis-spliced intron in mosses and hornworts.

    Lacking sufficient characters for cladistic analyses, nearly all possible phylogenetic tree topologies for the three bryophyte classes liverworts, hornworts, and mosses have been suggested in the literature. A topology in which hornworts are the basal-most embryophytes and in which a liverwort-moss clade is sister to tracheophytes was obtained through analyses of the nuclear small rRNA (Hedderson et al. 1996) and of a four-gene data set assembled from small rRNA sequences of all three genomes and the chloroplast rbcL gene (Nickrent et al. 2000).

    Our study, however, adds arguments to the alternative view that liverworts are the sister group to all other embryophytes (Qiu et al. 1998). Any phylogenetic topology in which liverworts would not be placed as the basal-most embryophytes now need to postulate (1) numerous losses of mitochondrial introns conserved in tracheophytes and shared by mosses and/or hornworts (nad5i230, nad5i1455, and nad5i1477 [this contribution], nad2i156 and nad2i1282 [Pruchner et al. 2002], nad4i461, nad7i140, and nad7i209 [Pruchner et al. 2001], cox2i373, cox2i691, and nad1i728 [Qiu et al. 1998]) and (2) gain of liverwort-type introns that are not observed in the other clades. The positional stability of the mitochondrial introns argues against this option. On the other hand, some intron losses along the phylogenetic backbone affecting entire land-plant clades do obviously exist. For example, the topology now suggested by analysis of nad5 would, on the one hand, nicely explain intron nad2i1282 as a further synapomorphy of the hornwort-tracheophyte clade. However, on the other hand, secondary nad2 intron losses need to be postulated in any tree topology. For the tree shown in figure 4, this would be the loss of nad2i156 in hornworts (present in mosses but absent in liverworts) and the loss of nad2i709 in mosses (present in all other land-plant clades). An intron serendipitously identified in the nad1 gene of a moss (Malek and Knoop 1998), nad1i27, has now been characterized as universally conserved between mosses and hornworts (Dombrovska and Qiu 2004). Presence of this intron can be interpreted as a gain in the nonliverwort lineage and subsequent loss in the common ancestor of tracheophytes in the phylogeny presented here.

    Similar larger-scale genomic features of cladistic value are rare in the chloroplast genome, as revealed by the recently completed sequence of the Anthoceros formosae plastome (Kugita et al. 2003). However, loss of the ycf66 gene adds to synapomorphies of the joint hornwort-tracheophyte clade.

    The phylogeny presented here is compatible with the "close relationship" of hornworts and tracheophytes that has been observed in the analysis of chloroplast rDNA and ITS sequences (Samigullin et al. 2002). The discrepancies to the phylogenetic studies based on rRNA sequences mentioned above, however, remain to be explained. Nuclear 18S rRNA studies have resulted in phylogenies in which jungermanniid liverworts have been linked to mosses (Capesius and Bopp 1997)—a phylogeny that has remained unsupported by all subsequent phylogenetic studies. The small rRNA gene even has been labeled as "positively misleading" for an unrelated issue of plant phylogeny (Duvall and Bricker 2004). A recent study has found bryophytes monophyletic with strong support when concatenated amino acid sequences of 51 chloroplast-encoded genes were used for phylogenetic analysis (Nishiyama et al. 2004). It should be kept in mind that in that study, (1) taxon sampling was somewhat biased towards angiosperms, whereas only one species of each bryophyte class was included, (2) that different weakly supported topologies were obtained when nucleotide instead of amino acid sequences were used, and (3) that node confidence was affected by selective inclusion of outgroup algae.

    As in other difficult issues of phylogenetic analysis, an extension of taxon sampling to maximize the inclusion of independent characters from all three plant genomes will provide an ultimate answer in the near future. The present study adds to the view that mitochondrial exons and introns in particular are a valuable resource in that regard. Contrasting modes of intron evolution in the two plant organelle genomes are obvious. Whereas most plastome introns apparently have been gained at an algal level of organization, mitochondrial introns seem to have been gained after establishment of the embryophyte lineage on land. Mycorrhizal and other endophytic fungal symbionts that can interact with bryophytes and that had been present at the time of establishment of land plants (Ligrone, Pocock, and Duckett 1993; Read et al. 2000; Redecker, Kodner, and Graham 2000) allow for speculations on differential gains of mitochondrial introns in the liverwort and nonliverwort lineages, potentially originating form distinct fungal donors.

    Acknowledgements

    We thank Dr. Yin-Long Qiu (Ann Arbor, Michigan) for discussing unpublished data and providing Haplomitrium mnioides DNA. We are grateful to Dr. Jochen Heinrichs (G?ttingen, Germany) for providing plant material of Monoclea gottschei and Noteroclada confluens. Finally, we thank the Deutsche Forschungsgemeinschaft DFG for financial support.

    References

    Beckert, S., S. Steinhauser, H. Muhle, and V. Knoop. 1999. A molecular phylogeny of bryophytes based on nucleotide sequences of the mitochondrial nad5 gene. Plant Syst. Evol. 218:179–192.

    Capesius, I., and M. Bopp. 1997. New classification of liverworts based on molecular and morphological data. Plant Syst. Evol. 207:87–97.

    Choquet, Y., M. Goldschmidt-Clermont, J. Girard-Bascou, U. Kück, P. Bennoun, and J. D. Rochaix. 1988. Mutant phenotypes support a trans-splicing mechanism for the expression of the tripartite psaA gene in the C. reinhardtii chloroplast. Cell 52:903–913.

    Dombrovska, O., and Y. L. Qiu. 2004. Distribution of introns in the mitochondrial gene nad1 in land plants: phylogenetic and molecular evolutionary implications. Mol. Phylogenet. Evol. 32:246–263.

    Duvall, M. R., and E. A. Bricker. 2004. 18S gene trees are positively misleading for monocot/dicot phylogenetics. Mol. Phylogenet. Evol. 30:97–106.

    Goldschmidt-Clermont, M., Y. Choquet, J. Girard-Bascou, F. Michel, M. Schirmer-Rahire, and J. D. Rochaix. 1991. A small chloroplast RNA may be required for trans-splicing in Chlamydomonas reinhardtii. Cell 65:135–143.

    Handa, H. 2003. The complete nucleotide sequence and RNA editing content of the mitochondrial genome of rapeseed (Brassica napus L.): comparative analysis of the mitochondrial genomes of rapeseed and Arabidopsis thaliana. Nucleic Acids Res. 31:5907–5916.

    Hedderson, T. A., R. L. Chapman, and W. L. Rootes. 1996. Phylogenetic relationships of bryophytes inferred from nuclear-encoded rRNA gene sequences. Plant Syst. Evol. 200:213–224.

    Huelsenbeck, J. P., and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754–755.

    Karol, K. G., R. M. McCourt, M. T. Cimino, and C. F. Delwiche. 2001. The closest living relatives of land plants. Science 294:2351–2353.

    Knoop, V., M. Altwasser, and A. Brennicke. 1997. A tripartite group II intron in mitochondria of an angiosperm plant. Mol. Gen. Genet. 255:269–276.

    Knoop, V., W. Schuster, B. Wissinger, and A. Brennicke. 1991. Trans splicing integrates an exon of 22 nucleotides into the nad5 mRNA in higher plant mitochondria. EMBO J. 10:3483–3493.

    Kohchi, T., K. Umesono, Y. Ogura, Y. Komine, K. Nakahigashi, T. Komano, Y. Yamada, H. Ozeki, and K. Ohyama. 1988. A nicked group II intron and trans-splicing in liverwort, Marchantia polymorpha, chloroplasts. Nucleic Acids Res. 16:10025–10036.

    Koller, B., H. Fromm, E. Galun, and M. Edelman. 1987. Evidence for in vivo trans splicing of pre-mRNAs in tobacco chloroplasts. Cell 48:111–119.

    Kubo, T., S. Nishizawa, A. Sugawara, N. Itchoda, A. Estiati, and T. Mikami. 2000. The complete nucleotide sequence of the mitochondrial genome of sugar beet (Beta vulgaris L.) reveals a novel gene for tRNA(Cys)(GCA). Nucleic Acids Res. 28:2571–2576.

    Kugita, M., A. Kaneko, Y. Yamamoto, Y. Takeya, T. Matsumoto, and K. Yoshinaga. 2003. The complete nucleotide sequence of the hornwort (Anthoceros formosae) chloroplast genome: insight into the earliest land plants. Nucleic Acids Res. 31:716–721.

    Liepelt, S., R. Bialozyt, and B. Ziegenhagen. 2002. Wind-dispersed pollen mediates postglacial gene flow among refugia. Proc. Natl. Acad. Sci. USA 99:14590–14594.

    Ligrone, R., K. Pocock, and J. G. Duckett. 1993. A comparative ultrastructural-study of endophytic basidiomycetes in the parasitic achlorophyllous hepatic Cryptothallus-Mirabilis and the closely allied photosynthetic species Aneura-Pinguis (Metzgeriales). Can. J. Bot. 71:666–679.

    Malek, O., A. Brennicke, and V. Knoop. 1997. Evolution of trans-splicing plant mitochondrial introns in pre-Permian times. Proc. Natl. Acad. Sci. USA 94:553–558.

    Malek, O. and V. Knoop. 1998. 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.

    Nickrent, D. L., C. L. Parkinson, J. D. Palmer, and R. J. Duff. 2000. Multigene phylogeny of land plants with special reference to bryophytes and the earliest land plants. Mol. Biol. Evol. 17:1885–1895.

    Nishiyama, T., P. G. Wolf, M. Kugita et al. (12 co-authors). 2004. Chloroplast phylogeny indicates that bryophytes are monophyletic. Mol. Biol. Evol. 21:1813–1819.

    Notsu, Y., S. Masood, T. Nishikawa, N. Kubo, G. Akiduki, M. Nakazono, A. Hirai, and K. Kadowaki. 2002. The complete sequence of the rice (Oryza sativa L.) mitochondrial genome: frequent DNA sequence acquisition and loss during the evolution of flowering plants. Mol. Genet. Genom. 268:434–445.

    Oda, K., K. Yamato, E. Ohta, Y. Nakamura, M. Takemura, N. Nozato, K. Akashi, T. Kanegae, Y. Ogura, Kohchi, T., and K. Ohyama. 1992. Gene organization deduced from the complete sequence of liverwort Marchantia polymorpha mitochondrial DNA: a primitive form of plant mitochondrial genome. J. Mol. Biol. 223:1–7.

    Pereira de Souta, A., M.-F. Jubier, E. Delcher, D. Lancelin, and B. Lejeune. 1991. A trans-splicing model for the expression of the tripartite nad5 gene in wheat and maize mitochondria. Plant Cell 3:1363–1378.

    Pruchner, D., S. Beckert, H. Muhle, and V. Knoop. 2002. Divergent intron conservation in the mitochondrial nad2 gene: signatures for the three bryophyte classes (mosses, liverworts, and hornworts) and the lycophytes. J. Mol. Evol. 55:265–271.

    Pruchner, D., B. Nassal, M. Schindler, and V. Knoop. 2001. Mosses share mitochondrial group II introns with flowering plants, not with liverworts. Mol. Genet. Genom. 266:608–613.

    Qiu, Y. L., Y. R. Cho, J. C. Cox, and J. D. Palmer. 1998. The gain of three mitochondrial introns identifies liverworts as the earliest land plants. Nature 394:671–674.

    Qiu, Y. L., and J. D. Palmer. 2004. Many independent origins of trans-splicing of a plant mitochondrial group II intron. J. Mol. Evol. 59:80–89.[ISI][Medline]

    Read, D. J., J. G. Duckett, R. Francis, R. Ligrone, and A. Russell. 2000. Symbiotic fungal associations in ‘lower’ land plants. Philos. Trans. R. Soc. Lond. B Biol. Sci. 355:815–830.

    Redecker, D., R. Kodner, and L. E. Graham. 2000. Glomalean fungi from the Ordovician. Science 289:1920–1921.

    Samigullin, T. K., S. P. Yacentyuk, G. V. Degtyaryeva, K. M. Valiejo-Roman, V. K. Bobrova, I. Capesius, W. F. Martin, A. V. Troitsky, V. R. Filin, and A. S. Antonov. 2002. Paraphyly of bryophytes and close relationship of hornworts and vascular plants inferred from analysis of chloroplast rDNA ITS (cpITS) sequences. Arctoa 11:31–43.

    Steinhauser, S., S. Beckert, I. Capesius, O. Malek, and V. Knoop. 1999. Plant mitochondrial RNA editing. J. Mol. Evol. 48:303–312.

    Swofford, D. L. 2002. PAUP*: phylogenetic analysis using parsimony (*and other methods). Version 4. Sinauer Associates, Sunderland, Mass.

    Turmel, M., C. Otis, and C. Lemieux. 2002a. The chloroplast and mitochondrial genome sequences of the charophyte Chaetosphaeridium globosum: insights into the timing of the events that restructured organelle DNAs within the green algal lineage that led to land plants. Proc. Natl. Acad. Sci. USA 99:11275–11280.

    ———. 2002b. The complete mitochondrial DNA sequence of Mesostigma viride identifies this green alga as the earliest green plant divergence and predicts a highly compact mitochondrial genome in the ancestor of all green plants. Mol. Biol. Evol. 19:24–38.

    ———. 2003. 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.

    Unseld, M., J. R. Marienfeld, P. Brandt, and A. Brennicke. 1997. The mitochondrial genome of Arabidopsis thaliana contains 57 genes in 366,924 nucleotides. Nat. Genet. 15:57–61.

    Vangerow, S., T. Teerkorn, and V. Knoop. 1999. Phylogenetic information in the mitochondrial nad5 gene of pteridophytes: RNA editing and intron sequences. Plant Biol. 1:235–243.[ISI]

    Wang, X. Q., D. C. Tank, and T. Sang. 2000. Phylogeny and divergence times in Pinaceae: evidence from three genomes. Mol. Biol. Evol. 17:773–781.(Milena Groth-Malonek*, Da)