- Split View
-
Views
-
Cite
Cite
Jörn Petersen, René Teich, Burkhard Becker, Rüdiger Cerff, Henner Brinkmann, The GapA/B Gene Duplication Marks the Origin of Streptophyta (Charophytes and Land Plants), Molecular Biology and Evolution, Volume 23, Issue 6, June 2006, Pages 1109–1118, https://doi.org/10.1093/molbev/msj123
- Share Icon Share
Abstract
Independent evidence from morphological, ultrastructural, biochemical, and molecular data have shown that land plants originated from charophycean green algae. However, the branching order within charophytes is still unresolved, and contradictory phylogenies about, for example,the position of the unicellular green alga Mesostigma viride are difficult to reconcile. A comparison of nuclear-encoded Calvin cycle glyceraldehyde-3-phosphate dehydrogenases (GAPDH) indicates that a crucial duplication of the GapA gene occurred early in land plant evolution. The duplicate called GapB acquired a characteristic carboxy-terminal extension (CTE) from the general regulator of the Calvin cycle CP12. This CTE is responsible for thioredoxin-dependent light/dark regulation. In this work, we established GapA, GapB, and CP12 sequences from bryophytes, all orders of charophyte as well as chlorophyte green algae, and the glaucophyte Cyanophora paradoxa. Comprehensive phylogenetic analyses of all available plastid GAPDH sequences suggest that glaucophytes and green plants are sister lineages and support a positioning of Mesostigma basal to all charophycean algae. The exclusive presence of GapB in terrestrial plants, charophytes, and Mesostigma dates the GapA/B gene duplication to the common ancestor of Streptophyta. The conspicuously high degree of GapB sequence conservation suggests an important metabolic role of the newly gained regulatory function. Because the GapB-mediated protein aggregation most likely ensures the complete blockage of the Calvin cycle at night, we propose that this mechanism is also crucial for efficient starch mobilization. This innovation may be one prerequisite for the development of storage tissues in land plants.
Introduction
Glyceraldehyde-3-phosphate dehydrogenases (GAPDH) are prominent examples of homologous isoenzymes that are essential for glycolysis and Calvin cycle (Cerff 1982). In land plants they are nuclear encoded, but their evolutionary ancestry is quite different. The glycolytic GapC gene has probably a proteobacterial (mitochondrial) origin, whereas the photosynthetic GapA was acquired from the cyanobacterial ancestor of the plastid (Martin and Cerff 1986; Brinkmann et al. 1987; Martin and Schnarrenberger 1997). In contrast to glycolytic GapC, which is not influenced by the redox state of the cell, Calvin cycle GapA is susceptible to light/dark regulation via the thioredoxin system (Buchanan 1984). The essential regulator is CP12, a small protein that mediates the aggregation of GAPDH and phosphoribulokinase (PRK) to catalytically inactive protein complexes (Pohlmeyer et al. 1996; Wedel, Soll, and Paap 1997). This sophisticated mechanism allows a metabolic switch from photosynthetic Calvin cycle to oxidative pentose phosphate pathway (OPPP) at night (Schnarrenberger, Oeser, and Tolbert 1973; Schnarrenberger, Flechner, and Martin 1995). It first appeared in cyanobacteria and was subsequently inherited by land plants in the course of primary endosymbiosis (Wedel and Soll 1998; Tamoi et al. 2005). Two important GAPDH gene duplications correlate with the evolutionary transition from green algae to land plants. First, a new photosynthetic GAPDH named GapB originated from the GapA/B gene duplication in green plants (Brinkmann et al. 1989). This gene recruited the regulatory GAPDH-binding module from CP12, and the corresponding protein exhibits a characteristic carboxy-terminal extension (CTE) (Pohlmeyer et al. 1996). Protein extracts of land plant chloroplasts contain homotetrameric A4 as well as heterotetrameric A2B2 isoenzymes (Cerff 1979; Cerff and Chambers 1979), and due to their CTE, the latter are autonomously capable of forming high molecular GAPDH complexes depending on the dark/light redox status of the chloroplast (Baalmann et al. 1996; Scheibe et al. 2002). Second, a duplicate of the cytosolic GapC designated GapCp acquired a transit peptide that directs this enzyme into plastids (Meyer-Gauen et al. 1994, 1998). GapCp is a key enzyme of plastid glycolysis, a novel pathway of land plants that is restricted to heterotrophic tissues (Petersen, Brinkmann, and Cerff 2003).
Green algae together with land plants (green plants; Viridiplantae), red algae (Rhodophyta), and glaucophytes represent the three primary lineages of photosynthetic eukaryotes (Delwiche 1999). It is very likely that they have a common origin and obtained their plastids through a single primary endosymbiosis with a cyanobacterium (Bhattacharya and Medlin 1995; Douglas 1998; Rodriguez-Ezpeleta et al. 2005) even if alternative scenarios were also proposed (Nozaki et al. 2003; Stiller, Reel, and Johnson 2003). For more than two decades, the most investigated glaucophyte Cyanophora paradoxa was regarded virtually as a synonym for the most ancient eukaryotic alga, mainly due to presumably ancestral plastids (cyanelles) harboring a cyanobacterial peptidoglycan wall and carboxysomes (Pfanzagl et al. 1996; Löffelhardt, Bohnert, and Brynat 1997). However, the discussion about the first-diverging lineage has recently been stimulated by contradictory phylogenies of plastid genomes and concatenated nuclear genes placing either glaucophytes or rhodophytes in a basal position (Rodriguez-Ezpeleta et al. 2005). In contrast to the unresolved relationships between the primary lineages, there is convincing evidence that land plants evolved from charophycean green algae (Graham 1993; Qiu and Palmer 1999). Both lineages are now classified as Streptophyta, whereas the remaining prasino-, trebouxio-, ulvo- and chlorophycean green algae are designated as Chlorophyta (Bremer 1985; Bremer et al. 1987). The charophytes comprise five orders, Charales, Coleochaetales, Zygnematales, Klebsormidiales, and the monotypic order Chlorokybales, whose representative Chlorokybus atmophyticus is generally regarded as the most ancient charophycean species (Melkonian, Marin, and Surek 1995; Graham, Cook, and Busse 2000; Karol et al. 2001; Turmel et al. 2002). The identity of the closest charophycean relative to terrestrial plants remains uncertain (Petersen, Brinkmann, and Cerff 2003), whereas liverworts are probably the earliest land plant lineage, diverging more than 475 MYA (Qiu et al. 1998; Wellman and Gray 2000; Wellman, Osterloff, and Mohiuddin 2003). However, a unicellular scaly green alga successfully resisted all attempts of incontestable classification, subsumed in a recent review as “the enigma of Mesostigma” (McCourt, Delwiche, and Karol 2004). Mesostigma viride was initially regarded as a primitive prasinophyte (Mattox and Stewart 1984) until molecular data gave contradictory evidence either for a streptophycean affiliation (Melkonian 1989; Bhattacharya et al. 1998; Marin and Melkonian 1999; Martin et al. 2002) or a basal placement prior to Chlorophyta and Streptophyta (Lemieux, Otis, and Turmel 2000; Turmel et al. 2002).
Here we present phylogenetic analyses of 23 new GAPDH and CP12 sequences from mosses, a liverwort, representatives of all green algal orders, and the glaucophyte C. paradoxa. The distribution of GapA and GapB sequences and the phylogenetic analyses reveal that the GapA/B gene duplication occurred in a common ancestor of land plants, charophytes, and M. viride, thus allowing a reliable positioning of this unicellular green alga.
Materials and Methods
Algal Material
The green algae C. atmophyticus (SAG 48.80), Coleochaete scutata (SAG 3.90), Spirogyra sp. (SAG 170.80), Klebsormidium flaccidum (SAG 121.80), Scherffelia dubia (SAG 17.86), Spermatozopsis similis (SAG 1.85), and the glaucophyte C. paradoxa (Pringsheim strain; SAG 29.80) were obtained from the “Sammlung von Algenkulturen” at the University of Göttingen. The algal material of the ulvophycean green alga Cladophora rupestris was collected on the North Sea island Helgoland, the charophycean green alga Chara vulgaris was identified in a pond near Braunschweig, and the moss Sphagnum cuspidatum was sampled from a sphagnum bog (Uchter Moor) (Petersen, Brinkmann, and Cerff 2003). Material of Chlorella sp. was donated by Ralf Kämmerer (FU Berlin).
Isolation of RNA and Construction of cDNA Libraries
Total RNA from C. atmophyticus, S. dubia, S. similis, Chlorella sp., and C. paradoxa was isolated according to Meyer-Gauen et al. (1998), and Poly (A)+ mRNA was prepared as described in Henze et al. (1995). The λNM1149 cDNA libraries were constructed using the TimeSaver cDNA Synthesis Kit (Biotech Europe GmbH, Amersham Pharmacia, Freiborg, Germany) following the manufacturer's instruction. The construction of the two λZAPII cDNA libraries from the liverwort Marchantia polymorpha and the charophyte C. vulgaris has been previously described (Petersen, Brinkmann, and Cerff 2003), and the λNM1149 library of the moss Physcomitrella patens was provided by William Martin (University of Düsseldorf). The isolation of mRNA and preparation of cDNA libraries from M. viride have been described by Simon et al. (2006).
Reverse Transcriptase–Polymerase Chain Reaction Amplification
Expressed GAPDH genes were amplified via reverse transcriptase–polymerase chain reaction (RT-PCR). Degenerate primers were designed based on the universally conserved N- and C-terminal GAPDH sequence motifs GINGF (5′-GSNATHAAYGGNTTYGG-3′) and WYDNE (5′-CCAYTCRTTRTCRTACCA-3′), respectively. Reverse transcription, PCR amplification, cloning, and sequencing were performed as previously described (Petersen, Brinkmann, and Cerff 2003; Petersen et al. 2006).
Isolation and Sequencing of cDNA Clones
The clones from M. polymorpha and C. vulgaris libraries were isolated with homologous RT-PCR probes for GAPDH genes, whereas all other libraries were screened with previously identified GapA and GapB probes under heterologous conditions (Petersen, Brinkmann, and Cerff 2003). Screening for CP12 of M. polymorpha was performed with a homologous probe, amplified via RT-PCR with primers specific for expressed sequence tag (EST) clone AU082009. All clones were sequenced on both strands. Due to the isolation of partial cDNA clones, GapA sequences of C. atmophyticus and M. viride were then assembled by merging cDNA and RT-PCR sequence information.
Identification of EST Clones by Database Analyses
Both GapA sequences of the ulvophycean green alga Acetabularia acetabulum were obtained from the “Protist EST Program” database (University of Montreal), whereas the CP12 sequence from the red alga Galdieria sulphuraria was identified from the Galdieria database (GDB; Michigan State University). Deduced GAPDH and CP12 amino acid sequences of land plants were used as query sequences for Blast searches in the National Center for Biotechnology Information (NCBI) database (TBlastN; EST database), and the respective clones were assembled into contigs (Supplementary Table S1, Supplementary Material online).
Sequence Handling and Phylogenetic Analyses
The newly established GAPDH clones, assembled EST Contigs (see above), and already annotated reference sequences (GenBank) were aligned with ClustalX and manually refined using the ED program of the MUST software package (for software citations, see Petersen et al. 2006). The data set contained 74 GAPDH sequences and 325 unambigously aligned amino acid positions. It was analyzed by all four standard methods of phylogenetic reconstruction (Bayesian inference, maximum likelihood [ML], distance [NJ], and maximum parsimony), including bootstrap analyses (or equivalent methods), to estimate the support for internal nodes of the phylogenies. The analyses were performed as previously described (Petersen et al. 2006). A MrBayes consensus tree is shown, where the posterior probabilities are displayed as percent values. Alternative topologies were analyzed with the approximately unbiased (AU) test using the CONSEL package (Shimodaira and Hasegawa 2001).
Results
Isolation of GAPDH and CP12 Sequences
In this study, we established six GapB and 15 GapA cDNA sequences (Supplementary Table S1, Supplementary Material online). Both, GapA and GapB clones, were identified from the liverwort M. polymorpha, the charophycean green algae C. scutata, C. vulgaris, Spirogyra sp., and C. atmophyticus, as well as from the unicellular green alga M. viride. GapA sequences were established from all other species analyzed in this study (P. patens, S. cuspidatum, K. flaccidum, S. similis, Cladophora ruspestris, S. dubia, Chlorella sp., C. paradoxa). However, exhaustive efforts to identify GapB from mosses (Physcomitrella, Sphagnum) were unsuccessful. In agreement with our experiments, database analyses of Physcomitrella show that GapB is absent from more than 120,000 EST clones as well as from genomic TRACE files (NCBI) representing at least a fourfold coverage of the whole genome. In accordance with our phylogenetic analyses (see below), the most probable explanation for this absence is a secondary loss of GapB genes in mosses; the same argument might also apply for the charophyte Klebsormidium. In addition to GAPDH sequences, we characterized two full-length CP12 cDNA clones from the green alga M. viride and the liverwort M. polymorpha. All reference sequences were retrieved from the databases (Supplementary Table S1, Supplementary Material online), and the deduced amino acid sequences were used for alignments and phylogenetic analyses.
Phylogenetic Analyses of GapA and GapB Sequences
Figure 1 shows a Bayesian consensus tree based on the phylogenetic analysis of 57 eukaryotic GapA and GapB sequences rooted with 16 cyanobacterial Gap2 outgroup sequences (see also Supplementary Fig. S1, Supplementary Material online). The data set comprises streptophyte sequences of various land plants including mosses (Physcomitrella, Sphagnum) and a liverwort (Marchantia). Moreover, it contains representatives of all five charophycean orders Coleochaetales (Coleochaete), Charales (Chara), Zygnematales (Spirogyra), Klebsormidiales (Klebsormidium), and Chlorokybales (Chlorokybus) as well as the unicellular green flagellate Mesostigma (Graham and Wilcox 2000). Several representatives of the Chlorophyta from the classes Chloro- (Chlamydomonas, Spermatozopsis, Scenedesmus), Ulvo- (Cladophora, Acetabularia), Trebouxio- (Chlorella), and Prasinophyceae (Scherffelia) are also included. The two remaining primary photosynthetic lineages are represented by sequences of Cyanophora (Glaucophyta) and five red algae (Rhodophyta).
A common origin of GapA and GapB genes from green plants (Streptophyta and Chlorophyta), Rhodophyta, and Glaucophyta, which all originated via primary endosymbiosis with a cyanobacterium, is solidly supported by bootstrap values and Bayesian posterior probabilities between 99% and 100%. Surprisingly, Cyanophora groups together with the green lineage (80%–98%), which is divided into three distinct subtrees, whereas the red algae represent the most basal branch. All chlorophyte GAPDH sequences form a monophyletic group branching basal to the GapA and GapB subtrees of streptophytes. Thus, the GapA/B gene duplication occurred in a common ancestor of land plants, charophytes, and Mesostigma (fig. 1). Mesostigma represents the first branching lineage in both subtrees followed by the charophytes. The two GapA subtrees are only moderately supported by statistical analyses, reflecting the high degree of sequence identity within the green lineage (at least 75%; data not shown). The GapB subtree is, in contrast, solidly supported (89%–100%; fig. 1) because the branch length subsequent to the gene duplication but before the divergence of Mesostigma is two times longer for the GapB subtree than for GapA (fig. 1). This observation is corroborated by the GapB-specific sequence pattern including two insertions of one amino acid (fig. 2 [i]) and at least three specific sequence positions (fig. 2 [ii]), which are not found in GapA genes. However, a comparison of the average branch length of the respective subtrees after this short phase of accelerated evolution reveals that GapB genes exhibit a remarkable degree of sequence conservation that is clearly higher than that of the respective GapA counterparts (fig. 1).
GapA and GapB sequences of the unicellular green alga Mesostigma congruently branch prior to charophytes, while Chlorokybus is the most basal charophyte in the slowly evolving GapB subtree. The position of Klebsormidium among bryophytes and seed plants within the GapA subtree obtains no significant statistical support; actually, phylogenetic analyses with reduced data sets indicate that this placement is the result of a long-branch attraction (LBA) artifact caused by fast-evolving seed plant sequences (data not shown). Low support from both GapA and GapB subtrees allows no certain prediction about the closest charophycean relative of land plants. Therefore, a complementary approach is applied to resolve these evolutionary relationships by identification of specific sequence signatures. The most parsimonious explanation for shared insertions, deletions, and conserved amino acid positions between different genes is a common ancestry. We detected five GapB-specific positions in the alignment that may correlate with the evolutionary relationships (fig. 2 [iii]). Their pattern is completely different between land plants and the green algae Mesostigma and Chlorokybus. In contrast, Spirogyra and Chara share three and four of these positions, respectively, whereas the pattern of Arabidopsis and Coleochaete are even identical.
Examination of the Phylogenetic Position of Cyanophora with AU Tests
The reliability of the unexpected affiliation of C. paradoxa together with green plants was analyzed with the AU test. Based on the phylogenetic tree shown in figure 1, we placed Cyanophora at all alternative positions within the tree and tested a total of 143 topologies. A summary of the relevant topologies is presented in Supplementary Table S2, Supplementary Material online (series A). Our AU test significantly rejects all trees that place Cyanophora basal to or within red algae (A3 + A4), whereas a total of 28 topologies were not significantly rejected. A total of 26 of them can be excluded in a biological context because they place this glaucophyte either within green plants (A2) or within cyanobacteria (A6). Among the two remaining meaningful topologies, the best (ML) tree proposes a green affiliation (A1) and the alternative tree shows the familiar basal placement as first-diverging photosynthetic lineage (A5). The RELL bootstrap support, which allows a relative weighting of all tested topologies and adds up to 1.0, clearly favors the best tree (0.5750, A1 vs. 0.0130, A5).
Our phylogenetic analyses demonstrate that the GapA sequences from Cyanidioschyzon and Porphyra have by far the longest branches within the rhodophycean clade (Supplementary Fig. S1, Supplementary Material online), and further analyses revealed that they are probably subject to LBA (ingroup LBA; data not shown). Therefore, we removed these two sequences from the tree and performed a second series of AU tests with 139 topologies (Supplementary Table S1, Supplementary Material online; series B). The results are comparable with those of the first analysis, but the resampling estimated log-likelihood (RELL) bootstrap support for the best tree increases from 0.5750 to 0.7920 (fig. 1). If those topologies where Cyanophora groups within green plants (B2) are also considered (B1 + B2), the suggested green affiliation reaches almost maximum support (0.9482), whereas a position basal to rhodophytes and green plants decreases to a negligible value (0.0050).
Sequence Signatures of GapA, GapB, and CP12
The streptophyte GapB gene recruited its regulatory module (CTE) from CP12, and both domains are still alignable as shown in figure 3. The mature CP12 subunit of land plants, chlorophytes, and some cyanobacteria is composed of two domains, both containing two highly conserved cysteine residues with a regulatory function (Supplementary Fig. S2, Supplementary Material online) (Wedel and Soll 1998). The cysteine residues of the N-terminal domain are surprisingly absent from CP12 proteins of rhodophytes, Cyanophora, and Synechococcus, but however, analyses in Synechococcus document that this lack does not influence the formation of the PRK/CP12/GAPDH complex (Tamoi et al. 2005). Nevertheless, all available CP12 sequences exhibit a highly conserved C-terminal domain including the characteristic cysteines (fig. 3). This also applies for the composite alignment of CP12 and GapB (indicated by a gray background in fig. 3).
Discussion
Metabolic Role of GapA, GapB, and CP12 in Cyanobacteria and Plants
In the present study, we analyzed the distribution of plastid Calvin cycle GAPDH (GapA and GapB) in green plants. The streptophyte GapB sequences originated from the GapA/B gene duplication at least 700 MYA (fig. 1; timing see below), and due to the gain of an important additional function, GapB is a prime example of a duplicated gene (Brinkmann et al. 1989; Lynch et al. 2001). The newly recruited CTE (fig. 3) mediates the aggregation of plastid GAPDH to 550-kDa complexes of high molecular mass (HMM) (Baalmann et al. 1996; Martin, Scheibe, and Schnarrenberger 2000; Sparla, Pupillo, and Trost 2002; Buchanan and Luan 2005; Sparla et al. 2005). However, the discovery of a specific regulator of the Calvin cycle, CP12, shows that GapB only tells part of the story (Wedel, Soll, and Paap 1997; Wedel and Soll 1998). This small linker protein is responsible for the formation of a second (600 kDa) aggregate containing PRK and GAPDH. The mechanism was established in a common ancestor of present-day cyanobacteria (Pohlmeyer et al. 1996; Tamoi et al. 2005) probably about a billion years prior to the GapA/B gene duplication (Yoon et al. 2004). The striking similarity between the CTE of GapB and the C-terminal domain of CP12 (fig. 3) documents that GapB acquired its characteristic extension from CP12 (Baalmann et al. 1996; Pohlmeyer et al. 1996). We characterized CP12 cDNAs from the liverwort M. polymorpha as well as the unicellular green alga M. viride, and our comprehensive analyses of the EST databases revealed that CP12 is universally present in green plants, rhodophytes, and the glaucophyte C. paradoxa (fig. 3). The inactivation of the Calvin cycle at night is in land plant chloroplasts probably regulated by the formation of two independent HMM protein aggregates, a 600-kDa PRK/GAPDH/CP12 complex and a 550-kDa A8B8 GAPDH complex (Scheibe et al. 2002). This extensive regulation in plants assures the metabolic flexibility of the plastid, which switches from the reductive Calvin cycle for CO2 fixation during illumination to the OPPP for reduced nicotinamide adenine dinucleotide phosphate generation at night (Schnarrenberger, Oeser, and Tolbert 1973; Schnarrenberger, Flechner, and Martin 1995). An incomplete or missing inhibition of the Calvin cycle would result in futile cycling and a waste of adenosine triphosphate, as recently demonstrated by a cyanobacterial knockout mutant of CP12 that shows a significantly reduced growth rate (Tamoi et al. 2005). It is interesting in this context that diatom plastids, which were recruited via secondary endosymbiosis from a red alga, lost their OPPP (Michels, Wedel, and Kroth 2005), probably due to the replacement of the plastid GapA by a nuclear-encoded GapCI that is not susceptible to thioredoxin-mediated regulation (Liaud et al. 1997, 2000). These examples document the importance of blocking the Calvin cycle at night in the presence of OPPP, but they give no convincing answer for the necessity of two regulatory complexes in land plants. The simplest explanation is that the newly emerged HMM complex tightens this inactivation in the dark, but a more auspicious prediction would be a crucial function of the GapB-dependent complex for streptophycean metabolism. Because PRK inactivation via CP12 reliably prevents futile cycling at night, formation of an independent A8B8 complex likely inactivates the residual GAPDH activity. This may be necessary for efficient starch mobilization because it prevents further glycolytic degradation of glyceraldehyde-3-phosphate in the plastid and allows an export via the triose translocator into the cytosol (Flügge et al. 1989). This promising explanation is compatible with the observation that plastid starch storage is restricted to green plants, whereas rhodophytes and glaucophytes accumulate their polysaccharides (e.g., floridean starch) outside the plastid (van den Hoek, Mann, and Jahns 1996). Therefore, the functional gain of GapB at the base of streptophycean evolution could constitute a metabolic prerequisite for the development of heterotrophic storage tissues in terrestrial plants including those with agricultural relevance as potatoes or cereal grains.
Cyanophora Possesses a Cyanobacterial GapA Gene That Is Closely Related to Green Plants
We present the first study of nuclear-encoded plastid GAPDH as well as CP12 sequences of all three primary photosynthetic lineages (green plants, red algae, and glaucophytes) (figs. 1 and 3). Our phylogenetic analyses clearly indicate the common cyanobacterial origin of eukaryotic GapA (99%–100% support; fig. 1) and suggest the same for CP12 (data not shown). This implies two endosymbiotic gene transfers from the plastid to the host cell nucleus in a common ancestor of the present-day lineages. Thus, GAPDH and CP12 data independently support the commonly accepted prediction that plastids of green plants (Chlorophyta and Streptophyta), rhodophytes, and glaucophytes can be traced back to a single primary endosymbiosis with a cyanobacterium (Bhattacharya and Medlin 1995; Douglas 1998; Delwiche 1999; Rodriguez-Ezpeleta et al. 2005). The branching order of these three primary lineages is still unclear, but in particular the glaucophyte-specific peptidoglycan wall and carboxysomes are generally regarded as ancient cyanobacterial traits (Pfanzagl et al. 1996; Löffelhardt, Bohnert, and Brynat 1997). Accordingly, several molecular analyses of concatenated plastid genomes show a basal position of glaucophytes prior to green plants and red algae, albeit only with weak to moderate statistic support (Martin et al. 1998, 2002; Rodriguez-Ezpeleta et al. 2005). Our analyses of plastid GapA sequences surprisingly favor an alternative placement of Cyanophora with green plants and thus a basal branching of the rhodophytes (fig. 1). In order to assess the confidence level of this green affiliation, we tested 143 alternative topologies of the glaucophyte GapA sequence with the least biased and most rigorous test available to date, the “AU” test (Supplementary Table S2, Supplementary Material online). A close affinity to red algae is significantly rejected by our analyses, but up to a fifth of the tested topologies including a basal position of Cyanophora cannot be rejected, probably due to the high degree of GapA sequence conservation (>75% amino acid identity) in general and of the Cyanophora sequence in particular (fig. 1). However, the RELL bootstrap proportion clearly supports the sistergroup relationship of glaucophytes and green plants presented in figure 1, especially after elimination of the two fast-evolving red algal sequences (Supplementary Table S2, Supplementary Material online; RELL support of 0.95). Even though hidden paralogy or cryptic lateral gene transfer may confound single-gene phylogenies, the same grouping has been obtained in analyses of concatenated nuclear-encoded genes, albeit without significant support (Nozaki et al. 2003; Rodriguez-Ezpeleta et al. 2005). The pattern of presence or absence of chlorophyll b, phycobilins, or flagella leads to contradictory deductions for the relationships of the three basal lingeages (van den Hoek, Mann, and Jahns 1996; Tomitani et al. 1999). Interestingly, Raven (2003) argues that peptidoglycan and carboxysomes are functionally linked in glaucophytes and associated with an adaption to freshwater habitats. Furthermore, considering that this functional trait is shared with the cyanobacterial ancestor, its maintenance in glaucophytes does not justify a basal positioning of this group (Raven 2003). We conclude that the proposed sistergroup relationship with green plants (fig. 1) constitutes a sound working hypothesis against which future studies should be tested. Nevertheless, additional molecular data and especially a better species sampling of red algae and glaucophytes will be required to unequivocally determine the branching order of the three primary lineages.
The Presence of GapB Places Mesostigma As the Earliest Streptophyte Branching Basal to Charophytes and Land Plants
Unique events in evolution are excellent cornerstones to clarify the relationship of present-day lineages reliably. Thorough analyses of endosymbiotic gene transfers offer the chance to reconstruct very ancient interkingdom linkages (Harper and Keeling 2003; Petersen et al. 2006), whereas genome or gene duplications provide insight into more recent intrakingdom evolution (Lynch and Conery 2000; Petersen, Brinkmann, and Cerff 2003). In this study, we investigated the timing of the GapA/B gene duplication in relation to the origin of land plants. The GapB-specific C-terminal domain allows an unambiguous identification of the respective gene (fig. 3), and a closer look at the databases reveals that GapB is universally present and highly expressed in seed plants (data not shown) but absent from the genome of Chlamydomonas. Here we present a broad survey of plastid GapA and GapB genes from green plants including 20 newly established sequences and document that Chlorophyta exclusively harbor GapA genes (fig. 1). Because GapB was found in the liverwort M. polymorpha, a typical member of the probably most ancient lineage of land plants, the gene duplication must have occurred in a common ancestor of all terrestrial plants more than 500 MYA (Qiu et al. 1998; Wellman and Gray 2000; Wellman, Osterloff, and Mohiuddin 2003). However, the identification of GapB sequences in nearly all major groups of charophytes (fig. 1) shows that the gene duplication obviously occurred much earlier. The question about the closest relative of land plants generated multiple conflicting hypotheses, but the most convincing data suggest Charales and/or Coleochaetales as the most promising candidates (Huss and Kranz 1997; Bhattacharya and Medlin 1998; Graham, Cook, and Busse 2000; Karol et al. 2001; Petersen, Brinkmann, and Cerff 2003; McCourt, Delwiche, and Karol 2004). Although, our GAPDH analyses (fig. 1) allowed us to investigate this topic in both GapA and GapB subtrees, the statistic support either for C. vulgaris or for C. scutata is weak, probably because of the limited phylogenetic signal due to the high degree of sequence conservation. The comparison of several GapB-specific sequence positions reveals the presence of identical sequence signatures in Arabidopsis and Coleochaete (fig. 2), an observation that was previously also made for cytosolic GapC sequences (Petersen, Brinkmann, and Cerff 2003). However, unequivocal molecular data are required to support or reject the assumption that the differentiated morphology of stoneworts (Charales) (van den Hoek, Mann, and Jahns 1996) actually reflects an advanced evolutionary position. The basal position of C. atmophyticus, initially proposed based on its sarcinoid thallus (Rogers, Mattox, and Stewart 1980), is now substantiated by analyses of the complete plastid genome (Turmel et al. 2002) of the only species within the Chlorokybales. The respective branching in our GapB subtree (fig. 1) and the divergent pattern of GapB-specific sequence positions (fig. 2) are congruent with this conclusion.
A crucial finding of this study is the presence of GapB in the enigmatic unicellular green alga M. viride (McCourt, Delwiche, and Karol 2004). This scaly flagellate was previously classified as a prasinophyte (Mattox and Stewart 1984), but ultrastructural features proposed a placement together with charophytes (Melkonian 1989). In addition, molecular data are inconclusive, supporting either a branching prior to green plants (Chloro- and Streptophyta) on the basis of mitochondrial and plastid genomes (Lemieux, Otis, and Turmel 2000; Turmel, Otis, and Lemieux 2002) or a placement together with charophytes on the basis of single-gene and plastid genome phylogenies (Bhattacharya et al. 1998; Marin and Melkonian 1999; Martin et al. 2002). However, the discovery of GapB in Mesostigma provides unequivocal support for the streptophycean affiliation of this green alga. Furthermore, our phylogenetic analyses support a placement at the base of all charophytes and reject a position together with or basal to chlorophytes (fig. 1). Thus, the GapA/B gene duplication occurred in the common ancestor of Mesostigma, charophytes, and land plants and coincides with the emergence of streptophytes. The exact timing of this event, which accordingly marks the origin of land plant evolution, is quite difficult, but recent molecular estimates for the split of streptophytes and chlorophytes range from 700 to 1150 MYA (Douzery et al. 2004; Yoon et al. 2004).
William Martin, Associate Editor
We are grateful toward Naiara Rodriguez-Ezpeleta, Denis Baurain, and Lorenz Bülow for helpful comments on the manuscript and thank Carina Grauvogel for critical discussions. The sequence data from A. acetabulum were produced by the Protist EST Program (http://amoebidia.bcm.umontreal.ca/public/pepdb/agrm.php) and those of G. sulphuraria were provided by the Michigan State University GDB (http://genomics.msu.edu/sequence_data.html). Major financial support, including a Ph.D. stipend for R.T., was received from the Deutsche Forschungsgemeinschaft (CE 1/27-2).
References
Baalmann, E., R. Scheibe, R. Cerff, and W. Martin.
Bhattacharya, D., and L. Medlin.
Bhattacharya, D., K. Weber, S. S. An, and W. Berning-Koch.
Bremer, K., C. J. Humphries, D. Mishler, and S. P. Churchill.
Brinkmann, H., R. Cerff, M. Salomon, and J. Soll.
Brinkmann, H., P. Martinez, F. Quigley, W. Martin, and R. Cerff.
Buchanan, B. B.
Buchanan, B. B., and S. Luan.
Cerff, R.
———.
Cerff, R., and S. E. Chambers.
Delwiche, C. F.
Douzery, E. J., E. A. Snell, E. Bapteste, F. Delsuc, and H. Philippe.
Flügge, U. I., K. Fischer, A. Gross, W. Sebald, F. Lottspeich, and C. Eckerskorn.
Graham, L. E., M. E. Cook, and J. S. Busse.
Harper, J. T., and P. J. Keeling.
Henze, K., A. Badr, M. Wettern, R. Cerff, and W. Martin.
Huss, V. A. R., and H. D. Kranz.
Karol, K. G., R. M. McCourt, M. T. Cimino, and C. F. Delwiche.
Lemieux, C., C. Otis, and M. Turmel.
Liaud, M. F., U. Brandt, M. Scherzinger, and R. Cerff.
Liaud, M. F., C. Lichtle, K. Apt, W. Martin, and R. Cerff.
Löffelhardt, W., H. J. Bohnert, and D. A. Brynat.
Lynch, M., and J. S. Conery.
Lynch, M., M. O'Hely, B. Walsh, and A. Force.
Marin, B., and M. Melkonian.
Martin, W., and R. Cerff.
Martin, W., T. Rujan, E. Richly, A. Hansen, S. Cornelsen, T. Lins, D. Leister, B. Stoebe, M. Hasegawa, and D. Penny.
Martin, W., R. Scheibe, and C. Schnarrenberger.
Martin, W., and C. Schnarrenberger.
Martin, W., B. Stoebe, V. Goremykin, S. Hansmann, M. Hasegawa, and K. Kowallik.
Mattox, K. R., and K. D. Stewart.
McCourt, R. M., C. F. Delwiche, and K. G. Karol.
Melkonian, M.
Melkonian, M., B. Marin, and B. Surek.
Meyer-Gauen, G., H. Herbrand, J. Pahnke, R. Cerff, and W. Martin.
Meyer-Gauen, G., C. Schnarrenberger, R. Cerff, and W. Martin.
Michels, A. K., N. Wedel, and P. G. Kroth.
Nozaki, H., M. Matsuzaki, M. Takahara, O. Misumi, H. Kuroiwa, M. Hasegawa, T. Shin-i, Y. Kohara, N. Ogasawara, and T. Kuroiwa.
Petersen, J., H. Brinkmann, and R. Cerff.
Petersen, J., R. Teich, H. Brinkmann, and R. Cerff.
Pfanzagl, B., A. Zenker, E. Pittenauer, G. Allmaier, J. Martinez-Torrecuadrada, E. R. Schmid, M. A. De Pedro, and W. Löffelhardt.
Pohlmeyer, K., B. K. Paap, J. Soll, and N. Wedel.
Qiu, Y. L., Y. Cho, J. C. Cox, and J. D. Palmer.
Qiu, Y. L., and J. D. Palmer.
Raven, J. A.
Rodriguez-Ezpeleta, N., H. Brinkmann, S. C. Burey, B. Roure, G. Burger, W. Löffelhardt, H. J. Bohnert, H. Philippe, and B. F. Lang.
Rogers, C. E., K. R. Mattox, and K. D. Stewart.
Scheibe, R., N. Wedel, S. Vetter, V. Emmerlich, and S. M. Sauermann.
Schnarrenberger, C., A. Flechner, and W. Martin.
Schnarrenberger, C., A. Oeser, and N. E. Tolbert.
Shimodaira, H., and M. Hasegawa.
Simon, A., G. Glöckner, M. Felder, M. Melkonian, and B. Becker.
Sparla, F., P. Pupillo, and P. Trost.
Sparla, F., M. Zaffagnini, N. Wedel, R. Scheibe, P. Pupillo, and P. Trost.
Stiller, J. W., D. C. Reel, J. C. Johnson.
Tamoi, M., T. Miyazaki, T. Fukamizo, and S. Shigeoka.
Tomitani, A., K. Okada, H. Miyashita, H. C. Matthijs, T. Ohn, A. Tanaka.
Turmel, M., C. Otis, J.-C. De Cambiaire, J.-F. Pombert, and C. Lemieux.
Turmel, M., C. Otis, and C. Lemieux.
van den Hoek, C., D. Mann, and H. M. Jahns.
Wedel, N., and J. Soll.
Wedel, N., J. Soll, and B. K. Paap.
Wellman, C. H., and J. Gray.
Wellman, C. H., P. L. Osterloff, and U. Mohiuddin.
Author notes
*Institut für Genetik, Technische Universität Braunschweig, Braunschweig, Germany; †Botanisches Institut, Universität zu Köln, Köln, Germany; and ‡Département de Biochimie, Université de Montréal, Montréal, Canada