Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

I. Introduction

Land plants have a chloroplast (plastid) genome (plastome) with a basic canonical organization that is similar to that of their algal ancestors (see Chap. 3). This represents one of the most evolutionary conserved genomic structures in nature. However, from this basic organization, several structural changes have occurred on various evolutionary branches. Here we review aspects of plastomes of extant land plants, except for seed plants (see next chapter). The main lineages include the nonvascular bryophyte lineages (hornworts, liverworts and mosses), the lycophytes and monilophytes. The latter, which include leptosporangiate ferns and horsetails, are also referred to elsewhere as ‘ferns’ (e.g., Pryer et al. 2004; Schneider et al. 2009). Seed plants appear to be the sister to monilophytes (Pryer et al. 2001). Our current understanding of relationships among these lineages is depicted in Fig. 4.1. We begin with an overview of the taxa and structural aspects of plastomes. We then summarize the major events of gene and intron loss in plastomes of non-seed land plants. Next we discuss the phenomenon of RNA editing, a process that occurs at much higher rates in non-seed land plants than in seed plants.

Fig. 4.1.
figure 1

Our current understanding of relationships among major land plant lineages. The extant bryophytes represent a grade of three lineages with liverworts shown sister to all other extant land plants and hornworts shown sister to extant vascular plants. Vascular plants include the lycophytes, monilophytes and seed plants. Four major monilophyte lineages are shown as an unresolved polytomy sister to seed plants.

Full size image

II. Techniques and Overall Plastome Organization

Until about the mid-1990s, restriction site mapping was the main approach to inferring plastome organization. The technique involves digesting DNA with restriction endonucleases, separating the DNA fragments on an agarose gel and then transferring them to a membrane. The fragments on this membrane are then probed with labeled plastid DNA from a well-characterized species, or fragments of the same species cut with a different restriction enzyme. After careful analysis, a coarse-scale map of the plastome can be constructed. The first such physical map was that of the Zea mays plastome (Bedbrook and Kolodner 1979). Mapping studies also indicated that, within plant cells, the plastome exists in two orientations (Palmer 1983), a pattern that is maintained by a form of homologous recombination (so-called flip-flop recombination; Stein et al. 1986). Subsequently, plastomes of many species were mapped (reviewed by Palmer 1985), verifying that in most (but not all) lineages, plastomes map to a circle with a large single copy region (LSC) and a small single copy region (SSC) ­separated by two copies of an inverted repeat (IR), which include the ribosomal RNA genes (Palmer 1985). Fine-scale mapping requires nucleotide sequencing, which is easier and cheaper with today’s techniques. The first two plastomes to be completely sequenced were those of the flowering plant tobacco (Nicotiana tabacum; Shinozaki et al. 1986), and the liverwort Marchantia polymorpha (Ohyama et al. 1986). These data confirmed the earlier inferences on overall plastome organization that had been deduced from mapping studies.

Most green plant plastomes map to a circle of about 150 kb. However, the largest reported plastome, that of the green alga Floydiella terrestris, is more than 500 kb (Brouard et al. 2010). Most plant cells contain many copies of the plastome; even plants with a single plastid (e.g., the unicellular green alga Chlamydomonas reinhardtii) can contain many copies of the plastome. At the other extreme, wheat cells have more than 50 plastids per cell and more than 300 plastome copies per plastid (Boffey and Leech 1982). Thus, although the plastome is a small genome compared to its nuclear counterpart, plastid DNA makes up a significant proportion of total cellular DNA, as much as 20% in some species (Boffey and Leech 1982).

Plastid DNA is not assembled into chromosomes and it does not reside in the plastid as a population of free circular molecules. Rather, several plastomes are organized, with proteins and RNA, into structures known as nucleoids (Sato et al. 2003). Most nucleoids are attached to the envelope membrane, but mature chloroplasts can also have nucleoids associated with the thylakoid membrane (Sato et al. 2003). It is likely that nucleoid structure plays an important role in plastome replication, transcription and post-transcriptional modification. However, the general relationships between plastome packaging and these processes remain poorly understood (Bock 2007).

Although plastomes are typically depicted as circles, most plastid DNA is not in this form in a living plant cell (Bendich 2004; Bock 2007). Researchers have found linear plastomes, concatenated pieces representing multiple plastomes (sometimes circular) (Bendich 2004), and even branched forms (Oldenburg and Bendich 2004a). This variety of possible conformations is likely a function of both phylogenetic divergence and stage of plastome replication. The plastome replication process itself is also poorly understood (Bock 2007), and several mechanisms have been proposed. Early models involved bidirectional replication similar to that in bacteria, resulting in displacement (D) loops (Kolodner and Tewari 1975b). Rolling circle amplification (RCA) could also be used to achieve additional replication (Kolodner and Tewari 1975a). A double D-loop mechanism has also been proposed (Kunnimalaiyaan and Nielsen 1997). However, these models have been challenged, based on the degree of linear DNA observed (Bendich 2004), and a recombination-dependent mechanism was instead proposed (Oldenburg and Bendich 2004b). The challenge of studying replication is making observations during the actual process. Alternatively, researchers can examine the signature of replication, which can be deduced from variation in base composition. Studies of mitochondrial genomes found that regions accumulate adenine-to-guanine transitions due to deamination during the single-stranded phase of replication. This is because A  →  G transitions accumulate evenly over time whereas the accumulation of C  →  T substitutions is complex and asymptotic (Krishnan et al. 2004). Thus, gradients in A/G composition, especially for non-coding DNA, is a function of total amount of time spent in the single-stranded phase, and therefore can reveal origins and directions of replication. This approach was used recently to show that A/G composition gradients are most consistent with the earlier models (bidirectional and RCA) across a wide range of published green plant plastomes (Krishnan and Rao 2009). Direct testing of these models is now needed. Meanwhile, evidence continues to accumulate for a role of recombination-dependent replication in Arabidopsis, especially as a repair process for maintaining plastome integrity (Rowan et al. 2010). Clearly, the evidence suggests that more than one replication process appears to be operating, and the result is a complex population of molecules representing the plastome. Regardless, most land plant plastomes map to a circle and have a fairly conserved set of protein and RNA encoding genes. The map of the plastome of the whisk fern Psilotum nudum is depicted in linear fashion in Fig. 4.2 as a guide to this overall structure.

Fig. 4.2.
figure 2

Psilotum nudum plastid genome structure. Genes (colored boxes) on the right side of the map are transcribed in the top down direction, whereas those on the left side are transcribed bottom up. The tRNA genes are indicated by the three-letter amino acid code followed by the anticodon. Intron-containing genes are show with an asterisk (*); the trans-spliced gene rps12 is shown with two asterisks (**). The two horizontal red lines along the genome indicate the insertion/deletion events unique to all monilophytes and the two grey boxes along the genome indicate the inverted repeats. Note a fragment of ycf2 is found in the inverted repeat.

Full size image

A. Bryophytes

The bryophytes represent a grade of three extant lineages (Fig. 4.1, Mishler and Churchill 1984; Nickrent et al. 2000; Renzaglia et al. 2007; Shaw and Renzaglia 2004). Several phylogenetic analyses lead to the hypothesis that liverworts are sister to all other extant land plants (Qiu et al. 1998) and hornworts are sister to extant vascular plants (e.g., Groth-Malonek et al. 2005; Qiu et al. 2006, 2007; Qiu 2008). Recent findings of cryptospores from the early Middle Ordovician (c. 473–471 mya Rubinstein et al. 2010) may represent ­liverworts or at least their ancestors. A broad-scale phylogenetic analysis of liverworts reveals several key lineages. The earliest branching lineage, Haplomitriopsida, is the sister to all remaining extant liverworts. There is then a major split between the complex thalloid liverworts (Marchantiopsida) and a heterogenous clade (Jungermanniopsida) which includes two clades (Metzgeriidae and Pellidae) of simple thalloid taxa (which is therefore a paraphyletic group) and a monophyletic “leafy” clade (Jungermanniidae) which excludes a few taxa previously considered as leafy (Forrest et al. 2006). The complex-thalloid liverwort Marchantia polymorpha was the first plant for which the chloroplast genome was sequenced (Ohyama et al. 1986). Later, the complete mitochondrial genome of M. polymorpha was also sequenced (Oda et al. 1992), providing yet another important genomic resource for nonvascular plants. A second liverwort plastome was recently sequenced (Wickett et al. 2008b), that of the only known parasitic bryophyte, the simple-thalloid liverwort Aneura mirabilis. Nonphotosynthetic plants often lose plastid genes that are associated with photosynthetic functions (Wickett et al. 2008b; Wolfe et al. 1992). Indeed, A. mirabilis has lost some of the same genes as has the parasitic angiosperm Epifagus virginiana (Wolfe et al. 1992). However, the loss of only a subset of these genes in A. mirabilis suggests that this liverwort is in an earlier stage of acquiring a parasitic life history stage (Wickett et al. 2008a, b).

Mosses are a diverse clade of more than 12,000 ­species, representing about eight main extant ­lineages (Cox et al. 2004; Goffinet and Buck 2004; Newton et al. 2000; Wahrmund et al. 2009, 2010). Two complete moss plastomes have been sequenced: that of the model species for molecular genetic studies, Physcomitrella patens (Sugiura et al. 2003) and the desiccation-tolerant species Syntrichia (= Tortula) ruralis (Oliver et al. 2010). These plastomes differed by a large (71 kb) inversion in the large single copy (LSC) region, with S. ruralis possessing the apparently ancestral organization. Further analysis revealed that the inversion is unique to the Funariidae (Goffinet et al. 2007). This inversion is the largest plastome reorganization reported to date for land plant plastomes, and appears to represent a single evolutionary event (Goffinet et al. 2007).

Hornworts represent the third main clade of nonvascular land plants, with about 400 extant species (Bateman et al. 1998). Hornworts are probably sister group to the vascular plants (Groth-Malonek et al. 2005; Qiu et al. 2007). Phylogeny within the hornworts has been examined by Duff and coworkers (Duff et al. 2004; 2007). Currently, there is only a single published complete plastome sequence of a hornwort, Anthoceros formosae (Kugita et al. 2003b). This plastome has a very high level of RNA editing (Kugita et al. 2003a), as do several mitochondrial and plastid genes in most hornworts studied (Duff and Moore 2005; Duff 2006). More details on RNA editing are provided later in this chapter.

B. Lycophytes

The lycophytes include a large assemblage of both extant and extinct lineages. Extant groups include the heterosporous Isoetopsida with about 150 species of Isoetes (quillworts) and about 700 species of Selaginella (spikemosses). The remaining extant lineage is the homosporous Lycopodiopsida (clubmosses) of which about 300 species are known, including Lycopodium, Huperzia and related genera. Extinct lineages include many fossil species, especially from the late Silurian (about 420 mya) through the Carboniferous (about 300 mya, Kenrick and Crane 1997). Ancient representative of this group of plants formed many of the fossil coal beds. Photosynthesis in these plants harnessed the sun’s energy, which is now used as one major source of fossil fuels. These extinct lycophytes were large plants; some reached 30 m, whereas today’s species are less than 1 m. As a group the lycophytes appear to be a sister group to Euphyllophytes (monilophytes plus seed plants, see below). This early split is supported both by analysis of morphology in fossil taxa (Kenrick and Crane 1997) and extant taxa (Kranz and Huss 1996). However, an additional convincing piece of evidence comes from analysis of plastome organization. Monilophytes and seed plants possess a 30 kb inversion in the LSC relative to lycophytes and bryophytes (Raubeson and Jansen 1992). Further details of the organization of lycophyte plastomes came from restriction site mapping of an Isoetes plastome (Duff and Schilling 2000), which confirmed the overall similarity of the lycophyte and bryophyte plastomes. The first complete plastome sequence of a lycophyte was that of Huperzia lucidula (Wolf et al. 2005). Since then, additional plastomes have been sequenced from the ­heterosporous genera, Selaginella moellendorffii, S. uncinata and Isoetes flaccida (Karol et al. 2010; Tsuji et al. 2007). Although lycophytes share structural similarities with bryophytes, the former do have some unique features. For example, ycf2 normally resides in the LSC in most plastomes, but has been translocated to the SSC in I. flaccida, with the 5′ end now incorporated into the IR. In addition, the chlL/chlN gene cluster has been inverted in I. flaccida so that it is now adjacent to ycf2 rather than ycf1 as in H. lucidula. The ycf2 translocation and the chlL/chlN inversion occur in neither of the Selaginella plastomes. Both Selaginella plastomes differ considerably in gene order from other plastomes (Karol et al. 2010). An approximately 14-kb region has been translocated from the LSC to the IR/SSC in both Selaginella plastomes. The genes included in this translocation differ slightly between S. uncinata and S. moellendorffii. In addition, rps4 is in the IR in Selaginella and marks one endpoint of the translocated segment. The other endpoint resides in the SSC and is marked by psbD in S. moellendorffii. In S. uncinata, the same endpoint includes three additional genes (trnE-UUC, trnY-GUA and trnD-GUC), which remain in the LSC adjacent to ycf2 in S. moellendorffii. Selaginella uncinata also has a ∼20-kb LSC inversion (psbI to rpoB-trnC-GAC), a duplication of the psbK/trnQ-UUG region, and translocation of petN from the LSC to the SSC. These features appear to be unique to S. uncinata (Karol et al. 2010). Because complete plastome sequences are available from only four species of lycophytes, it is not yet possible to infer the phylogenetic extent of all plastome changes. Additional taxon sampling will be needed to understand more fully how recent and extensive these changes are.

C. Monilophytes (Ferns)

Monilophytes represent another group of vascular plants with an extensive fossil history. Here we consider four main extant lineages: (1) leptosporangiate ferns (about 11,000 species), (2) a clade that includes whisk ferns (Psilotum and Tmesipteris) and the Ophioglossales, (3) Marattioid ferns and (4) Horsetails (Equisetum). Data from plastid and nuclear gene sequences (Pryer et al. 2001) and morphology (Kenrick and Crane 1997; Schneider et al. 2009) find support for monophyly of a clade that includes these four ­lineages. Together the clade is called moniliformopses (Kenrick and Crane 1997), monilophytes (Pryer et al. 2004), or ferns sensu lato (Schneider et al. 2009). Further resolution of relationships among these four groups has not yet been achieved. Although monophyly of monilophytes has support from analyses of extant taxa, analyses that include fossil taxa has questioned this idea (Rothwell and Nixon 2006).

The first monilophyte to have a plastome sequenced was Psilotum nudum (GenBank accession #AP004638 from 2002, see Fig. 4.2). Several phylogenetic studies support inclusion of the ophioglossoid ferns with the whisk ferns (Pryer et al. 2001, 2004; Qiu et al. 2007), but so far no complete plastome from the ophioglossoid ferns has been published. Complete plastome sequences are available from one horsetail (Karol et al. 2010), one marattioid fern (Roper et al. 2007) and four leptosporangiate ferns (Der 2010; Gao et al. 2009; Wolf et al. 2003, 2011).

An inversion in the LSC involving trnG-GCC to trnT-GGU is found in all fern plastomes and no other land plant plastomes (Karol et al. 2010), thus providing further evidence for monilophyte monophyly. Within the leptosporangiate ferns, a series of additional inversions has occurred, two of which (18 kb and 21 kb respectively) result in a reverse gene order within the IR (Wolf et al. 2010). An additional pair of inversions occurred more recently in the LSC of a large clade of ferns (the “polypods”, Wolf et al. 2010).

III. The Inverted Repeat Boundaries

Plastome IRs from most plants typically house a similar gene content, which includes primarily rRNA and tRNA genes (Jansen et al. 2007; Palmer and Stein 1986; Turmel et al. 2007). This is seen also in some leptosporangiate ferns where, except for a few early-diverging clades, the IR itself is inverted (Wolf et al. 2003). Most of the variation in IR gene content occurs at the ends of the IR. This “ebb and flow” of the IR boundaries into and out of the LSC and SSC regions has been attributed to effects of recombination and gene conversion (Goulding et al. 1996). Effects of these positional changes have been seen in related species at the nucleotide level in several species of Nicotiana (Goulding et al. 1996). Furthermore, when comparing distantly related lineages of land plants, several plastomes exhibit unique IR boundaries that differ from the basic theme (Karol et al. 2010). But this is not always the case: other distantly related taxa have very similar IR boundaries. For example, Marchantia polymorpha, two mosses and Equisetum arvense were identical in gene content at both ends of the IR. This suggests that whereas the ends of the IR clearly ebb and flow in some lineages, in other lineages they appear to be rather stable, at least at the scale of gene order (Karol et al. 2010).

IV. Changes in Gene and Intron Content

Most plastomes sequenced to date contain a very similar repertoire of genes. The most significant exceptions are plastomes from parasitic plants in which many photosynthetic genes are lost or pseudogenized (Wickett et al. 2008b; Wolfe et al. 1992). Overlaid on the basic pattern are found a few genes that are absent in some sequenced plastomes. Some of these genes seem to have been lost multiple times based on their phylogenetic distribution (See Fig. 4.3). These include infA and ycf1. Other genes appear to be distinctly present or absent in particular clades. Here we briefly list these latter patterns based on what we know is a very limited sample of plastomes (especially for non-seed land plant clades). We ignore many that are specific to only one plastome, except where that plastome is the sole representative (such as the single published hornwort plastome).

Fig. 4.3.
figure 3

Phylogenetic results using nucleotide data. Phylogenetic analyses were performed using 49 plastome gene sequences from 45 completely sequenced plastomes, including 39 land plants and six charophycean algae. The nucleotide alignment from Karol et al. (2010) was used as a starting point (49 genes from 43 taxa). To this we incorporated into the alignment sequence data from two new leptosporangiate fern plastomes: Cheilanthes lindheimeri (Wolf et al. 2011) and Pteridium aquilinum (Der 2010). Maximum likelihood analyses were performed on the Cyberinfrastructure for Phylogenetic Research (CIPRES) Portal (v.3.1, Miller et al. 2009) using RAxML-HPC (v. 7.2.7, Stamatakis 2006, 2008) with 200 bootstrap replicates. Third codon positions were excluded to avoid problems associated with relatively rapidly evolving sites. The best tree (−ln  =  195205.737395) is shown with bootstrap proportions drawn above branches. The relationships among major fern lineages are weakly supported, though monophyly of the ferns is strongly supported. The branch leading to Selaginella was drawn to one-half scale to accommodate this figure. Note that the sister relationship of the liverwort and mosses is strongly supported and is in contrast to the cladogram shown in Fig. 4.1. This relationship was also recovered by Karol et al. (2010) when divergent taxa (Selaginella spp.) were excluded from phylogenetic analyses. Furthermore, Renzaglia and Garbary (2001) concluded that characters related to sperm cell development were compelling evidence for the monophyly of liverworts plus mosses, a clade they referred to as Setaphytes. Names of lost plastid genes are shown with arrows. An asterisk (*) following a gene name indicates that this gene has been lost in at least two lineages independently. The rare gain of a plastid gene (matK) is also indicated in the green algae.

Full size image

The genes ccsA and rpoA are absent from the plastomes of two mosses (Syntrichia ruralis and Physcomitrella patens, Oliver et al. 2010; Sugiura et al. 2003), petN is lacking in S. ruralis, and cemA is absent from both Selaginella plastomes. Mosses and liverworts lack rps16, but the gene is present in hornworts and some vascular plants. The genes matK and rps15 are pseudogenes in the hornwort (Kugita et al. 2003b). The gene cluster chlB, chlL, and chlN is absent from Psilotum nudum and angiosperms. The gene psaM is lacking from the three polypod ferns (Adiantum capillus-veneris, Cheilanthes lindheimeri and Pteridium aquilinum), as well as from the two Selaginella plastomes and the majority of seed plant plastomes. Seed plant plastomes lack rpl21, as do the two Selaginella plastomes. The parasitic liverwort A. mirabilis has lost several genes (including several ndh genes) and many others exist as possibly recent pseudogenes (Wickett et al. 2008a).

A group II intron, along with its encoded maturase gene (matK) invaded the trnK-UUU gene in charophycean algae after the divergence of chlorophytes and charophytes. All chloropycean algae and some early diverging charophycean algae (Mesostigmatophyceae and Chlorokybophyceae) do not contain this intron. More derived charophycean algae (Charophyceae, Coleocheatophyceae and Zygnematophyceae) have the intron. There is one lineage (Klebsormidiophyceae, which is sister to Charophyceae, Coleocheatophyceae, Zygnematophyceae and land plants) where we do not yet fully know the condition of trnK. A large clade of leptosporangiate ferns has subsequently lost trnK-UUU and its intron (Wolf et al. 2010, 2011), yet matK remains. The introns of clpP are variable across land plants, with some plastomes having two, and others having one intron in this gene, but there appears to be no distinct phylogenetic pattern (Karol et al. 2010).

Thus, although plastome gene content tends to be well-conserved among land plant lineages, several clade-specific gene losses are apparent.

V. RNA Editing

The central dogma of molecular genetics requires conservation of information from genomic DNA through messenger RNA to the final amino acid sequence of a protein. However, detailed studies of the various products of transcription and translation have found exceptions to this conservation. Considerable post-translational modification occurs to proteins. In addition to the various aspects of RNA processing that occur, an independent post-transcriptional stage is RNA editing. This process alters the nucleotides in the primary transcript so that the messenger RNA differs from the genomic encoding sequence (See Chap. 13). RNA editing is found throughout eukaryotes, and is especially common in organellar genomes (reviewed by Tillich et al. 2006). In plastome genes from seed plants, the process occurs at fewer than 40 sites and about ten times that number have been reported in ferns and hornworts. In most cases, cytosines are edited to uracils, but in hornworts and ferns, additional uracil-to-cytosine edits have been reported (Kugita et al. 2003a; Wolf et al. 2004).

RNA editing requires both cis- and trans-acting factors. Cis-acting factors include the actual site to be edited. Other cis-acting factors include upstream and downstream recognition sequences (Kobayashi et al. 2008). However, the latter appear to have no obvious pattern across sites. This might be because the trans-acting factors (nuclear-encoded proteins) are likely to be of several types (Hammani et al. 2009). To date, over 20 different nuclear factors have been associated with RNA editing in Arabidopsis (see Stern et al. 2010), most of which are pentatricopeptide repeat (PPR) proteins (Kotera et al. 2005; Okuda and Shikanai 2008). These proteins are characterized by tandem repeats of a degenerate 35 amino acid motif, and several PPR gene subfamilies are found across eukaryotic lineages.

The functions of RNA editing are not obvious. Several authors have argued that RNA editing repairs errors in genomic sequences (Jobson and Qiu 2008; Stern et al. 2010). However, this seems far less efficient than a simple nucleotide substitution at the DNA level of the genome, which would require no further action. An additional role has been implicated in gene regulation, whereby RNA editing varies with developmental stage and could be used to restore correct translation when the gene product is needed (Hirose et al. 1999). This has been observed in a few cases in animals, but seems to play a minor role in plants (Stern et al. 2010). It seems more likely that the enzymes that edit RNA have evolved for other cellular functions and their editing ability then releases selective constraints for the edited sites in genes. In fact, some of these other functions of editing enzymes are known. In primates, the APOBEC family of RNA editing enzymes includes cytosine deaminases that act to restrict infection from retroviruses (Bransteitter et al. 2009). Further research is needed on the RNA editing factors of Arabidopsis and other plants if we are to understand further the function and cellular significance of RNA editing.

RNA editing can cause problems for comparative analyses of nucleotide sequences. Most ­phylogenetic analyses are based on alignment of orthologous genomic sequences. However, if RNA editing occurs, these DNA sequences represent the unedited versions. Should one use the genomic sequences or the edited versions? The latter can only be inferred accurately by using mature RNA transcripts to generate cDNA. Until this is done, one does not know which sites have been edited. For analyses of seed plants, this dilemma is trivial because RNA editing rates are so low. But in ferns, lycophytes and some bryophytes, the effect on the outcomes of analyses can be significant. In hornworts, RNA editing rates are so high that the same site can be C to U edited in some taxa and U to C edited in other taxa (Duff and Moore 2005). When phylogenetic analyses of hornworts use cDNA sequences, the results are different from those from genomic sequences (Duff and Moore 2005; Duff 2006). Removal of edited sites does not help, because that reduces the amount of potentially useful phylogenetic signal. The solution can only be attained once we know the evolutionary stability of RNA editing itself. If relatively stable, then the fact that a site is edited provides an evolutionary marker. If sites come and go rapidly, then RNA editing sites are homoplastic and the results of phylogenetic analysis of cDNA sequences will be misleading. The answer will depend on the relative levels of homoplasy in genomic sequences versus RNA editing sites, and this is likely to vary across clades of land plants.

VI. Phylogenetic Analyses

Over the last few decades single gene phylogenetic analyses have served as powerful tools for reconstructing the evolutionary history of every major lineage of life on Earth (Donoghue and Cracraft 2004). Reduced costs and improvements in sequencing technologies have allowed several genes to be sequenced across a broad range of taxa for phylogenetic reconstruction (Holton and Pisani 2010; Nickrent et al. 2000; Qiu et al. 2007; Shalchian-Tabrizi et al. 2008). Indeed, with new second-generation sequencing technologies, complete plastome sequences are now being generated at an ever increasing rate (Cronn et al. 2008; Wolf et al. 2011). We reanalyzed the plastome alignment of Karol et al. (2010) and included two new leptosporangiate fern taxa (Cheilanthes lindheimeri and Pteridium aquilinum). This analysis included 49 plastome genes from 45 green plant taxa and the results are shown in Fig. 4.3. The overall topology is consistent with results presented in Karol et al. (2010), with the two new fern taxa found in a monophyletic leptosporangiate clade. Relationships among the major monilophyte lineages remained weakly supported. Most of the currently available land plant plastome sequences are from seed plants, with very few available from the presumed sister clade, monilophytes. With additional data from other fern representatives, including ophioglossoid ferns, it will become possible to gain further insight into early land plant evolution as well as the patterns and processes that shape the evolution of plastomes.

VII. Future Directions

Currently, the distribution of complete plastome sequence data is biased toward angiosperms. In general, clades more distantly related to angiosperms are less well sampled. There are especially critical clades in the algae for which no representative plastome sequence is available (e.g., Klebsormidiophyceae, Coleochaete). Although obtaining the actual DNA sequence is relatively easy, limiting steps in plastome sequencing mostly involve isolating plastome DNA. Although this can be done through various centrifugation and other procedures (Jansen et al. 2005), there are some alternative approaches. If the plastome component of total DNA is high then a total genomic shotgun sequence can provide sufficient data from which the plastome sequence can be assembled (Wolf et al. 2011). A more cost-effective approach involves multiplex sequencing-by-synthesis on the Illumina platform (Cronn et al. 2008). In this protocol, more than a hundred plastomes can be sequenced simultaneously. However, custom probes or PCR-primers will be needed for each major clade, the range of these depending on sequence divergence levels. One problem with the shotgun genome approach is that it may not be possible to distinguish genuine reads of plastome DNA from those that are plastid DNA that has been transferred to the nucleus (Bock and Timmis 2008). To some extent, this is a problem for all approaches to plastome studies, but the problem is exacerbated by short reads and the use of total genomic DNA extractions. Regardless, the prospects seem good for filling many of the critical clade gaps in the next few years. This should ease the trend away from recent exemplar studies (with a few, though critical taxa) toward more taxon-dense studies with broad phylogenetic breadth. Although such a trend may not always uncover much new in terms of phylogenetic hypotheses, it is sure to show us more details of the evolution of plastomes themselves.