MBE Advance Access originally published online on December 16, 2006
Molecular Biology and Evolution 2007 24(3):723-731; doi:10.1093/molbev/msl200
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Articles |
Phylogenetic Analyses of Nuclear, Mitochondrial, and Plastid Multigene Data Sets Support the Placement of Mesostigma in the Streptophyta
* Départment de Biochimie, Centre Robert Cedergren, Canadian Institute of Advanced Research, Université de Montréal, Montréal, Québec, Canada
Botanisches Institut, Lehrstuhl I, Universität zu Köln, Köln, Germany
E-mail: herve.philippe{at}umontreal.ca; michael.melkonian{at}uni-koeln.de.
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
All extant green plants belong to 1 of 2 major lineages, commonly known as the Chlorophyta (most of the green algae) and the Streptophyta (land plants and their closest green algal relatives). The scaly green flagellate Mesostigma viride has an important place in the debate on the origin of green plants. However, there have been conflicting results from molecular systematics as to whether Mesostigma diverges before the Chlorophyta/Streptophyta split or is an early diverging flagellate member of the Streptophyta. Previous studies employed either a limited taxon sampling (plastid and mitochondrial genomes) or a small number of phylogenetically informative sites (single nuclear genes). Here, we use large data sets from the nuclear (125 proteins; 29,319 positions), mitochondrial (33 proteins; 6,622 positions), and plastid (50 proteins; 10,137 positions) genomes with an expanded taxon sampling (21, 13, and 28 species, respectively) to reevaluate the phylogenetic position of Mesostigma. Our study supports the placement of Mesostigma in the Streptophyta (as an early diverging lineage) and provides evidence that systematic biases have played a role in generating some of the previous conflicting results. Importantly, we demonstrate that using an increased taxon sampling as well as more realistic models of evolution allows increasing congruence among the nuclear, mitochondrial, and plastid data sets.
Key Words: Mesostigma • Chlorokybus • phylogenomics • taxon sampling • systematic error • heterotachy
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Since its discovery in 1894 (Lauterborn 1894
), the ubiquitous, but nonabundant, freshwater green flagellate Mesostigma has had an inconspicuous research history until Manton and Ettl (1965) described the ultrastructure of the spectacular scale covering of the cell and its 2 flagella, which allied the organism with the then newly recognized green algal class Prasinophyceae (Christensen 1962). Ultrastructural studies in the early 1970s renewed interest in the Prasinophyceae, when morphologically similar scales were found on flagellate reproductive cells of algae such as the stonewort Chara corallina (Moestrup 1970), that had been implicated in the evolution of embryophyte land plants based on ultrastructural features of mitosis/cytokinesis and the flagellar apparatus (Pickett-Heaps and Marchant 1972). The search for a flagellate member of the land plant lineage among the prasinophyte flagellates remained unsuccessful until a cruciate flagellar root system with 2 multilayered structures (MLSs) was eventually described in Mesostigma viride (Rogers et al. 1981; Melkonian 1983, 1989), which linked this organism to both major lineages of green plants, the Chlorophyta (with cruciate flagellar root systems) and the Streptophyta (with unilateral flagellar root systems and a MLS), the latter including the embryophyte land plants (Bremer et al. 1987). Because many structural features of Mesostigma, however, resembled those of other prasinophyte flagellates such as Pyramimonas, Mesostigma was often regarded as closely related to such prasinophytes (Moestrup and Throndsen 1988; Melkonian 1990; Moestrup 1991), and it was not until molecular phylogenetic analyses of nuclear-encoded genes (small subunit [SSU] rDNA and actin) revealed the affiliation of Mesostigma with the Streptophyta (Melkonian et al. 1995; Bhattacharya et al. 1998; Marin and Melkonian 1999) that this view changed (but for a traditional taxonomic treatment of Mesostigma, see Moestrup 2002). The initial molecular phylogenetic analyses renewed interest in this organism and sparked a flurry of further studies aiming to determine the "true" position of Mesostigma in the tree of life. Unfortunately, depending on the data set used, conflicting conclusions were reached (for reviews, see Lewis and McCourt 2004; McCourt et al. 2004): whereas phylogenetic analyses based on chloroplast and mitochondrial genes and genomes led to the conclusion that Mesostigma is the earliest green plant divergence branching before the split into the Chlorophyta and Streptophyta (Lemieux et al. 2000; Turmel, Ehara, et al. 2002; Turmel, Otis, and Lemieux 2002b), other studies using a single chloroplast gene (rbcL; Delwiche et al. 2002) or a combination of 4 genes from the chloroplast, mitochondrial, and nuclear genomes (Karol et al. 2001) concluded that Mesostigma represents the earliest divergence within the Streptophyta. In the meantime, molecular data accumulated that revealed a number of embryophyte traits in Mesostigma, thus refuting its close affiliation with other prasinophyte flagellates (Nedelcu et al. 2006; Petersen et al. 2006; Simon et al. 2006). The availability of EST data from Mesostigma (Simon et al. 2006) afforded the possibility to perform a large-scale multigene phylogenetic analysis to reevaluate its phylogenetic position. Here, we report multigene phylogenetic analyses of large nuclear, mitochondrial, and chloroplast data sets. The phylogenetic analyses based on the nuclear, mitochondrial, and plastid data sets are consistent with the conclusion that Mesostigma is a member of the Streptophyta. We discuss possible reasons for the incongruence between previous studies. |
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Construction of the Nuclear, Plastid, and Mitochondrial Data Sets
The nuclear data set is based on an available alignment (TREEBASE, accession number SN2312) to which EST and trace sequences downloaded from National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/) from Mesostigma viride, Selaginella moellendorffii, Volvox carteri, Scenedesmus obliquus, Prototheca wickerhamii, Helicosporidium sp., Scherffelia dubia, Ostreococcus tauri, Galdieria sulphuraria, and Chondrus crispus were added as described (Philippe et al. 2004
). Among the Viridiplantae, the 15 species with the largest number of genes sequenced, that represent the major groups were selected. Four red algae and 2 glaucophytes were used to root the tree. Unambiguously aligned sequence blocks were extracted with Gblocks (Castresana 2000) and manually verified.
Potential paralogs were identified and removed as described (Philippe et al. 2004). When including all orthologous proteins that are available from at least 11 out of the 21 used species, the data set contains 125 proteins, totaling 29,319 amino acid positions. On average, 37% of the amino acids are missing.
The mitochondrial data set consists of 33 proteins, a total of 6,622 amino acid positions, from 13 species, including the jakobid Reclinomonas and 4 red algae that were used as an outgroup. The plastid data set consists of 50 plastid-encoded proteins, a total of 10,137 amino acid positions, from 28 species, including a glaucophyte and 8 red plastid-containing eukaryotes used as an outgroup. In all, 13% and 5% of the amino acid positions are missing in the mitochondrial and plastid data sets, respectively. Both data sets were aligned with ClustalW (Thompson et al. 1994), refined manually using MUST (Philippe 1993), and filtered from ambiguously aligned positions with Gblocks (Castresana 2000).
Phylogenetic Analyses
The concatenated data sets of nuclear, mitochondrial, and plastid sequences were analyzed by maximum parsimony (MP) and maximum likelihood (ML). MP analyses were performed using PAUP* 4.0 b10 (Swofford 2002) with the Tree Bisection-Reconnection search and 10 random addition of species. ML analyses were performed using PhyML 2.4 (Guindon and Gascuel 2003) and TREEFINDER (Jobb et al. 2004) with an evolutionary model consisting of the WAG matrix of amino acid substitution (Whelan and Goldman 2001), estimated amino acid frequencies, and a gamma distribution split into 4 categories to model the rate heterogeneity among sites (WAG + F + 
). The reliability of each internal branch was evaluated based on 1,000 (MP) or 100 (ML) bootstrap replicates.
Because the probability of getting trapped on a local maximum using heuristic searches is high when large data sets are used (Salter 2001), we also performed exhaustive ML analyses on a set of topologies obtained by constraining several groups. In general, groups supported by 100% bootstrap values (BVs) in previous analyses were constrained. In some cases, because constraining all the groups supported at 100% BV resulted in a too small set of topologies, we let some of them free (see information about constrained groups in the figures). In all cases, the 3 alternative positions for Mesostigma were exhaustively tested. The constraints allowed us to reduce the number of topologies to be analyzed to 81, 405, and 675 for the nuclear, plastid, and mitochondrial data sets, respectively. For each data set, the likelihood of each topology was calculated using Tree-Puzzle (Schmidt et al. 2002) with the WAG + F + model on the concatenated data set. Sitewise likelihood values were calculated by PAML (Yang 1997) and used to perform RELL bootstrap analyses (Kishino et al. 1990; Bapteste et al. 2002) with 10,000 replicates to assess the statistical support for each unconstrained branch.
Amino Acid Coding to Reduce Compositional Bias
To reduce the possible impact of compositional bias, we recoded the amino acids into 4 functional groups. To do that, we used the same coding as Hrdy et al. (2004) but combined the aromatic Phenylalanine, Tyrosine, and Tryptophane (FYW) and the hydrophobic Methionine, Valine, Isoleucine, and Leucine (MVIL) amino acids in a single category and coded the rare cysteine as missing data. This allowed the use of a general time reversible (GTR) matrix with 4 character states implemented in most programs. The parameters of the GTR matrix were estimated by PAUP* using a Neighbor-Joining tree. The 3 sets of topologies (from nuclear, plastid, and mitochondrial data sets) were exhaustively analyzed by Tree-Puzzle with a GTR + F + model. RELL bootstrap analyses (10,000 replicates) were performed as described above.
Removal of the Fast-Evolving Sites to Reduce the Impact of Some Systematic Errors
To reduce the impact of systematic errors, we eliminated the fast-evolving sites (Philippe, Delsuc, et al. 2005). To do that, we calculated the sitewise rates by PAML on an alignment that does not contain Mesostigma using the ML topology from which Mesostigma was pruned. This strategy was used because the phylogenetic position of Mesostigma may influence the estimation of the sitewise evolutionary rates potentially biasing the site removal approach (Rodriguez-Ezpeleta et al. 2007). The sites were then sorted according to their evolutionary rates and progressive removals of the fastest sites (1,000 each time) were performed. RELL bootstrap analyses (1,000 replicates) were performed after each removal and plotted against the alignment size.
Separate Analysis to Reduce the Impact of Heterotachy
To reduce systematic errors due to rate heterogeneity among sites through time (heterotachy), we performed separate analyses on previously defined partitions of the data set. To do that, we proceeded as described above for the exhaustive analysis but allowed branch lengths, alpha parameter of the gamma distribution, and stationary amino acid frequencies to be estimated independently for each partition. In order to evaluate the performance of the separate versus the concatenated model, we used the second-order Akaike Information Criterion (Akaike 1973), AICc (Hurvich and Tsai 1989): AICc = 2logL + 2K + 2K(K + 1)/n – K – 1, K being the number of free parameters of the model and n the number of positions in the data set. The number of free parameters was calculated as 1 alpha parameter + 2s – 3 branch lengths (s being the number of species) and 19 amino acid frequencies.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Phylogenetic Analysis Based on Nuclear Genes
We assembled a data set of 125 evolutionarily conserved orthologous proteins (29,319 amino acid positions) from 15 different taxa of Viridiplantae and 4 red algae and 2 glaucophytes as outgroups to gain more insight into the phylogenetic position of M. viride. The ML phylogeny inferred from the concatenation of the 125 proteins clearly placed Mesostigma in a sister-group position to the 6 other streptophyte taxa included in the analysis (fig. 1). This position of Mesostigma is supported by 100% BVs in all analyses (fig. 1). Almost all nodes in the tree are well resolved (except for the 2 long-branch thermophilic red algae), and the overall tree topology agrees well with phylogenies derived from SSU rDNA sequence comparisons (e.g., Marin and Melkonian 1999
). The amino acid coding into 4 functional groups (to reduce the impact of compositional bias), the removal of fast-evolving sites, or the separate analysis of 2 partitions (ribosomal and nonribosomal proteins) did not change the results (data not shown).
|
Phylogenetic Analysis Based on Mitochondrial Genes
The data set consists of 33 mitochondrion-encoded orthologous proteins (6,622 amino acid positions) from 8 different taxa of Viridiplantae and 4 red algae and the jakobid flagellate Reclinomonas as outgroups. The ML phylogeny inferred from the concatenation of the 33 proteins again placed Mesostigma in a sister-group position to the 5 other streptophyte taxa included in this analysis (fig. 2). This position is, however, only weakly supported by BVs (no support in MP, 65–85% BV support in the ML [TREEFINDER and PhyML] and exhaustive analyses) (fig. 2). Again we note that the tree topology is not only largely congruent with that of the nuclear-encoded data set (considering the different taxon sampling) but overall is also not in conflict with several previous single-gene phylogenies. Our phylogeny, however, is in conflict with the phylogenetic analysis based on mitochondrial proteins presented by Turmel, Otis, and Lemieux (2002b)
, who showed that Mesostigma branched at the base of the Viridiplantae. To understand this discrepancy, we did the following additional experiments (summarized in the Supplementary Material online, supplementary figs. S1 and S2): 1) we reduced our data set to the same proteins used by Turmel, Otis, and Lemieux (2002b), that is, 4,842 amino acids; 2) we reduced the taxon sampling to the same species used by these authors (8 species); 3) we reduced both the number of proteins and the taxon sampling; and 4) finally, we analyzed all data sets with (supplementary fig. S1) and without (supplementary fig. S2) taking rate heterogeneity among sites into account. The results of Turmel, Otis, and Lemieux (2002b) could only be reproduced using their taxon sampling and without using a gamma distribution to model the rate heterogeneity among lineages. Applied to our original data set, the amino acid coding into 4 functional groups or the separate analysis of 2 partitions (ribosomal and nonribosomal proteins) does not change the results; however, the removal of fast-evolving sites improves the bootstrap support value for the placement of Mesostigma as sister group of streptophytes (97% when the 1,500 fastest evolving sites are eliminated; data not shown).
|
Phylogenetic Analysis Based on Plastid Genes
The data set consists of 50 plastid-encoded orthologous proteins (10,137 amino acid positions) from 19 taxa of Viridiplantae and 8 eukaryote taxa with red-algal type plastids and the glaucophyte Cyanophora as outgroups. The ML tree inferred from the concatenation of the 50 proteins places Mesostigma together with the streptophyte genus Chlorokybus (fig. 3A) with strong support (100% BV in all analyses). Mesostigma + Chlorokybus (MC) emerge at the base of the Viridiplantae with low bootstrap support in the ML analysis (64–68% BV) and no support in the MP analysis. If Chlorokybus was excluded from the analysis, the support for the basal position of Mesostigma increased significantly in the ML analyses (85–93% BV; summarized in the Supplementary Material online, supplementary fig. S3). If Mesostigma, however, was excluded from the analysis (fig. 3B), Chlorokybus grouped with the other Streptophyta in a basal position with BVs of 75–92%. In these 3 cases, the amino acid coding into 4 functional groups or the removal of fast-evolving sites did not change the results (not shown).
|
However, these 2 approaches do not overcome tree-reconstruction artifacts due to heterotachy, that is, rate variation across sites through time (Lopez et al. 2002
), which are most likely present in plastid data sets (Lockhart et al. 2006). In an attempt to detect heterotachy, we divided the data set into 3 functional classes: translation (ribosomal proteins), RNA polymerase (A, B, and B' subunits of the RNA polymerase), and photosynthesis (the remaining proteins, all directly implicated in photosynthesis except 2—acetylCoA carboxylase and cytochrome biogenesis protein). As shown in figure 4, the differences on the branch lengths inferred from the 3 data sets are extreme. In particular, streptophytes and chlorophytes evolve about 3 times faster with respect to the remaining species (including Mesostigma and Chlorokybus) in the RNA polymerase data set and not in the other 2. As expected, important differences in the BV for the position of MC are observed when analyzing these tree data sets independently (table 1). The translation data set (2,199 amino acid positions) supports the placement of MC in the Streptophyta with 96% BV, whereas the support for the same relationship is nonexistent with the RNA polymerase data set (1,498 amino acid positions). Albeit being the largest (6,449 amino acid positions), the photosynthesis data set does not discriminate among the 3 alternatives.
|
|
A way to reduce artifacts due to heterotachous behaviors is to use a separate model (Yang 1996
; Kolaczkowski and Thornton 2004). However, a fully separate analysis considering independent parameters for each one of the 50 proteins only slightly decreases the support for the probably incorrect basal position of MC. In fact, there are too many free parameters (3,650) in this model, and the AIC (table 1) indicates that it is the one that worst fits the data. In contrast, when only 3 partitions (translation, RNA polymerase, and photosynthesis) are considered, only 146 additional parameters with respect to the concatenate model are needed, and the fit is better in this case (table 1). Interestingly, support for the sister group of MC and streptophytes increases from 32% to 57% with the partially separate model, indicating that heterotachy accounts at least in part for the misplacement of MC in the plastid data set. |
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
The incongruence between 2 single-gene phylogenies can be the result of 1) paralogy (generated by gene duplication), lateral gene transfer, or lineage sorting, 2) stochastic error, derived from the use of too few phylogenetically informative sites, and 3) systematic error, arising from the presence of nonphylogenetic signal in the data that is not accounted for in the tree-reconstruction models employed (Phillips et al. 2004
). Nonphylogenetic signals mainly derive from variable evolutionary rates across lineages leading to the well-known long-branch attraction (LBA) artifact, heterogeneous nucleotide/amino acid compositions leading to artificial attraction of taxa with the same bias, and heterogeneity of the evolutionary rate of a given position through time, that is, heterotachy (Philippe, Delsuc, et al. 2005).
The extent of taxon sampling may also account for incongruence between different phylogenies. Empirical evidence (e.g., Brinkmann et al. 2005; Philippe, Lartillot, et al. 2005) argues for a rich taxon sampling in phylogenetic analyses because this enables better detection of multiple substitutions and thus recognition of LBA artifacts (Felsenstein 1978; Hendy and Penny 1989). Sometimes, however, the number of extant lineages is extremely sparse and it may never be possible to attain a rich or balanced taxon sampling.
Previous single-gene/few-gene phylogenetic analyses trying to assess the phylogenetic position of Mesostigma may have suffered from some or all the above problems and yielded conflicting results. Given these restrictions, the conclusions from these analyses are often contradictory and somewhat limited. Phylogenies based on the nuclear-encoded SSU rDNA (Melkonian et al. 1995; Marin and Melkonian 1999) and actin-coding genes (Bhattacharya et al. 1998) suggested the divergence of Mesostigma at the base of the Streptophyta, although the basal divergences within the Streptophyta remained unresolved. Phylogenies derived from plastid-encoded genes are incongruent. The concatenation of the 2 plastid-encoded rRNA genes reported phylogenies that consistently placed Mesostigma at the base of the Chlorophyta and Streptophyta (Turmel, Ehara, et al. 2002), whereas phylogenies based on the rbcL gene or on 4 genes from the 3 genomes (atpB and rbcL from the plastid genome, nad5 from the mitochondrial genome, and SSU rDNA from the nuclear genome) placed Mesostigma at the base of the Streptophyta (Karol et al. 2001; Delwiche et al. 2002).
Phylogenomics, the genome-scale approach to phylogenetic inference, is thought to overcome the limitations of single-gene phylogenies by combining many genes and ultimately complete genomes (Philippe, Delsuc, et al. 2005). The use of large data sets theoretically overcomes incongruence because such data sets reduce the impact of stochastic error when more and more genes are considered. Several empirical studies have confirmed these predictions (e.g., Qiu et al. 1999; Madsen et al. 2001; Bapteste et al. 2002); however, conflicting results have also emerged (such as the question of the monophyly of the Ecdysozoa, Lophotrochozoa, and Protostomia; see Philippe, Lartillot, et al. 2005). Systematic error, caused by the presence of nonphylogenetic signals in the data, is not expected to disappear with the addition of data because, unlike stochastic error, it does not average out over sites. If nonphylogenetic signal is strong enough, it will cause the tree-reconstruction method to be inconsistent and lead to an incorrect, but statistically strongly supported tree (Felsenstein 1978; Phillips et al. 2004).
Two multigene analyses have been carried out to date that specifically address the phylogenetic position of Mesostigma in the Viridiplantae, and both refer to organelle phylogenomics. Lemieux et al. (2000) sequenced the entire chloroplast DNA of Mesostigma (118,360 bp) and analyzed a subset (53) of the 135 proteins encoded on the plastome (10,629 amino acid positions) with a taxon sampling that contained 3 Embryophyta and 3 Chlorophyta (and Cyanophora paradoxa as the outgroup). The tree topology in which Mesostigma diverged before the Streptophyta and Chlorophyta was strongly favored over alternative topologies that placed Mesostigma at the base of either the Streptophyta or the Chlorophyta. It must be noted, however, that their taxon sampling was very limited and lacked, for example, other streptophyte algae, and only one outgroup taxon was used. Furthermore, the probabilistic methods used (ML analysis under the JTT-F model) for tree reconstruction assumed a uniform rate of substitution. Additional studies from this group increased taxon sampling by adding complete plastid genomes of both streptophyte algae and Chlorophyta (Turmel, Otis, and Lemieux 2002a; Pombert et al. 2005, 2006; Turmel et al. 2005, 2006) but did not address the phylogenetic position of Mesostigma. Other multigene analyses using chloroplast proteins gave mixed results concerning the placement of Mesostigma (Martin et al. 2002, 2005). In a second approach, Turmel, Otis, and Lemieux (2002b) sequenced the mitochondrial genome of M. viride (42,424 bp) and analyzed a subset (19) of the 65 proteins encoded on the chondriome (4,139 amino acid positions) with a taxon sampling that contained 2 embryophytes and 2 Chlorophyta (plus 3 red algae as outgroups). The tree topology in which Mesostigma diverged before the Streptophyta and Chlorophyta was supported by BV of 100% in PROTML, distance, and MP analyses assuming a uniform substitution rate across sites. However, when rate variation across sites (8 gamma categories) was taken into consideration, in ML analyses (JTT model), the BV dropped significantly and support for this topology was low (63% or 67%, excluding or including invariant sites, respectively).
How do the results obtained in the present study compare with those previously published? We have assembled a large nuclear data set (125 proteins, 29,319 positions) with a reasonable taxon sampling (15 Viridiplantae, among them 6 streptophytes and 8 chlorophytes + Mesostigma). Phylogenetic analyses, involving different methods of tree reconstruction and addressing likely causes of systematic error such as compositional bias and fast-evolving sites, lead us to conclude that Mesostigma is an early branching member of the Streptophyta. Because only one other streptophyte alga (Closterium) was included in the analysis, we cannot yet address the relationship between Mesostigma and other streptophytes such as Chlorokybus, Klebsormidium, or Chaetosphaeridium, which must await the generation of EST data from these organisms. Similarly, we note that early branching chlorophytes such as Pyramimonas (Nakayama et al. 1998) are still lacking. In general, the results obtained from multigene analyses of the nuclear data set corroborate earlier analyses of nuclear-encoded single genes (SSU rDNA, actin) that placed Mesostigma in the Streptophyta.
For the mitochondrial data set, our results are in accordance with the nuclear data set but are in conflict with the mitochondrial protein phylogeny of Turmel, Otis, and Lemieux (2002b). Additional analyses adjusting the data set (from 6,622 to 4,842 amino acid positions) and taxon sampling (from 13 to 8 taxa) to those used by Turmel, Otis, and Lemieux (2002b) revealed the likely reasons for the discrepancy: poor taxon sampling in the ingroup (lack of other streptophyte algae) as well as in the outgroup (only the long-branch red algae were chosen) appeared to be responsible and the number of positions used was less important as long as the rate heterogeneity among lineages was modeled (see Results). However, when no gamma distribution was used, bootstrap support for the placement of Mesostigma with the streptophytes was abolished. This is in accordance with the data of Turmel, Otis, and Lemieux (2002b) who showed that the bootstrap support for the position of Mesostigma at the base of the Streptophyta and Chlorophyta was lowered when the rate heterogeneity among lineages was modeled. We conclude that the phylogeny of mitochondrial proteins places Mesostigma in the Streptophyta when taxon sampling is improved and the rate heterogeneity among lineages is modeled.
The phylogeny derived from the plastid data set reveals that the inclusion of Chlorokybus, which strongly groups with Mesostigma, decreases the bootstrap support for the basal position of Mesostigma in the Viridiplantae (fig. 3) and that different subsets of the complete data set provide conflicting results (fig. 4 and table 1). When the translational proteins are used, significant support for the placement of Mesostigma and Chlorokybus (MC) in the Streptophyta is obtained (this is also true for the data sets with Mesostigma or Chlorokybus only), whereas the RNA polymerase data set significantly rejects this relationship (table 1; supplementary tables S1 and S2, Supplementary Material online). This can be explained by the disproportionately fast evolutionary rate of the streptophyte RNA polymerase, which attracts this group to either the Chlorophyta or the outgroup (fig. 4C). Albeit its large size (6,449 positions) and no apparent evolutionary rate heterogeneities (fig. 4B), the photosynthesis data set does not support or reject any of the 3 alternatives (table 1; supplementary tables S1 and S2, Supplementary Material online). This lack of resolution is likely due to the slow evolutionary rate of these proteins.
The plastid data set illustrates tree-reconstruction artifacts due to heterotachy (Kolaczkowski and Thornton 2004). There exist highly heterotachous branch lengths for the different functional classes that cannot be acknowledged when a single set of branch lengths is used to analyze the concatenation of the 50 plastid proteins. As a result, an erroneous topology (MC at the base of green plants) is recovered. Although a separate model a priori defined by 3 function-based gene partitions may correct for heterotachy, the improvement with respect to concatenation is only marginal (from 32% to 57% BV for the sister group of MC and streptophytes). This suggests that even if the branch length differences are extreme between the 3 partitions, among-gene heterotachy is not the only cause of systematic error in this data set (Philippe, Zhou, et al. 2005). For instance, changes in the proportion of variable sites across the phylogeny within genes, which is the case for the RNA polymerase subunits (Lockhart et al. 2006), are not corrected by the type of separate model we used.
The separate model of the 50 proteins has a lower fit to the data because it implies much more free parameters, but only slightly improving the likelihood—note that previous studies (Bapteste et al. 2002; Philippe et al. 2004; Rodriguez-Ezpeleta et al. 2005) suggested, using an inadequate AIC approximation (see Posada and Buckley 2004), that the separate model was better. In any case, the difficultly of correctly locating Mesostigma with plastid proteins constitutes an interesting case study for testing new, more realistic, models of sequence evolution.
Apparently, Mesostigma attracts Chlorokybus to a position in the tree that does not conform to its typical streptophyte traits such as a subapical flagellar insertion and unilateral flagellar root system in the zoospores, which are morphological synapomorphies of the Streptophyta (Rogers et al. 1980; Lewis and McCourt 2004). Given that the evolutionary rate of Mesostigma is higher than that of Chlorokybus, this suggests that the position of Mesostigma in phylogenetic trees based on plastid data sets is affected by systematic errors. According to Jeffroy et al. (2006), phylogenies based on multiple genes can be biased by systematic errors and should be carefully scrutinized for possible tree-reconstruction errors. Supporting their conclusions, here, we have shown the importance of the use of 1) probabilistic methods that aim to capture real substitution patterns (Steel 2005)—in the mitochondrial data set and 2) an increased taxon sampling to corroborate the results—in the plastid and mitochondrial data sets.
In conclusion, the inclusion of Mesostigma in the Streptophyta, likely as an early branching lineage, is weakly supported by the mitochondrial and plastid data sets but significantly by the nuclear data set. This corroborates exciting recent findings about land plant–specific molecular/biochemical traits in Mesostigma such as the GapA/B gene duplication (Petersen et al. 2006; Simon et al. 2006), plant-type peroxisomal glycolate oxidase (Stabenau and Winkler 2005; Simon et al. 2006), the bud-induced (BIP) multigene family (Nedelcu et al. 2006), and F-box family proteins (Simon et al. 2006), suggesting that many typical embryophyte traits may have evolved at the level of the unicellular ancestor of the streptophytes before the transition to land, presumably when such a flagellate adapted to a freshwater/brackish habitat (Simon et al. 2006). This underpins the pivotal role that Mesostigma is likely to play in the coming years as a model to unravel the intricacies of the early steps in the evolution of streptophytes.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Figures S1–S3, tables S1 and S2, and the alignments used are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
H.P. acknowledges Génome Québec, the Canadian Research Chair and the Université de Montréal for financial support and the Réseau Québecois de Calcul de Haute Performance for computational resources. N.R.E. has been supported by ‘Programa de Formación de Investigadores del Departamento de Educación, Universidades e Investigación’ (Government of Basque Country). Part of this work was supported by grants from the Deutsche Forschungsgemeinschaft (Be 1779/7-1 and Be 1779/7-2).
|
Footnotes |
|---|
1 These authors contributed equally to this work.
Peter Lockhart, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialAcknowledgementsReferences |
|---|
Akaike H. (1973) Information theory and an extension of the maximum likelihood principle. In Csaki H (Ed.). Proceedings 2nd International Symposium on Information Theory(Akademia Kiado, Budapest (Hungary)) pp. 267–281.
Bapteste E, Brinkmann H, Lee JA, et al. (11 co-authors). (2002) The analysis of 100 genes supports the grouping of three highly divergent amoebae: Dictyostelium, Entamoeba, and Mastigamoeba. Proc Natl Acad Sci USA 99:1414–1419.
Bhattacharya D, Weber K, An SS, Berning-Koch W. (1998) Actin phylogeny identifies Mesostigma viride as a flagellate ancestor of the land plants. J Mol Evol 47:544–550.[CrossRef][Web of Science][Medline]
Bremer K, Humphries CJ, Mishler BD, Churchill SP. (1987) On cladistic relationships in green plants. Taxon 36:339–349.
Brinkmann H, van der Giezen M, Zhou Y, de Raucourt Poncelin G, Philippe H. (2005) An empirical assessment of long-branch attraction artefacts in deep eukaryotic phylogenomics. Syst Biol 54:743–757.
Castresana J. (2000) Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol Biol Evol 17:540–552.
Christensen T. (1962) Alger. In Böcher TW, Lange M, Sørensen T (Eds.). Botanik, Systematisk Botanik vol. 2(Munksgaard, Copenhagen) pp. 1–178.
Delwiche CF, Karol KG, Cimino MT, Sytsma KJ. (2002) Phylogeny of the genus Coleochaete (Coleochaetales, Charophyta) and related taxa inferred by analysis of the chloroplast gene rbcL. J Phycol 38:394–403.[CrossRef][Web of Science]
Felsenstein J. (1978) Cases in which parsimony or compatibility methods will be positively misleading. Syst Zool 27:401–410.[CrossRef][Web of Science]
Guindon S and Gascuel O. (2003) A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol 52:696–704.
Hendy MD and Penny D. (1989) A framework for the quantitative study of evolutionary trees. Syst Zool 38:297–309.[CrossRef][Web of Science]
Hrdy I, Hirt RP, Dolezal P, Bardonova L, Foster PG, Tachezy J, Embley TM. (2004) Trichomonas hydrogenosomes contain the NADH dehydrogenase module of mitochondrial complex I. Nature 432:618–622.[CrossRef][Medline]
Hurvich CM and Tsai C-L. (1989) Regression and time series model selection in small samples. Biometrika 76:297–307.
Jeffroy O, Brinkmann H, Delsuc F, Philippe H. (2006) Phylogenomics: the beginning of incongruence? Trends Genet 22:225–231.[CrossRef][Web of Science][Medline]
Jobb G, von Haeseler A, Strimmer K. (2004) TREEFINDER: a powerful graphical analysis environment for molecular phylogenetics. BMC Evol Biol 4:18.[CrossRef][Medline]
Karol KG, McCourt RM, Cimino MT, Delwiche CF. (2001) The closest living relatives of land plants. Science 294:2351–2353.[CrossRef][Web of Science][Medline]
Kishino H, Miyata T, Hasegawa M. (1990) Maximum likelihood inference of protein phylogeny, and the origin of chloroplasts. J Mol Evol 31:151–160.[CrossRef]
Kolaczkowski B and Thornton JW. (2004) Performance of maximum parsimony and likelihood phylogenetics when evolution is heterogeneous. Nature 431:980–984.[CrossRef][Medline]
Lauterborn R. (1894) Über die Winterfauna einiger Gewässer der Oberrheinebene. Biol Zbl 14:390–398.
Lemieux C, Otis C, Turmel M. (2000) Ancestral chloroplast genome in Mesostigma viride reveals an early branch of green plant evolution. Nature 403:649–652.[CrossRef][Medline]
Lewis LA and McCourt RM. (2004) Green algae and the origin of land plants. Am J Bot 91:1535–1556.
Lockhart P, Novis P, Milligan BG, Riden M, Rambaut A, Larkum T. (2006) Heterotachy and tree building: a case study with plastids and eubacteria. Mol Biol Evol 23:40–45.
Lopez P, Casane D, Philippe H. (2002) Heterotachy, an important process of protein evolution. Mol Biol Evol 19:1–7.
Madsen O, Scally M, Douady CJ, Kao DJ, DeBry RW, Adkins R, Amrine HM, Stanhope MJ, de Jong WW, Springer MS. (2001) Parallel adaptive radiations in two major clades of placental mammals. Nature 409:610–614.[CrossRef][Medline]
Manton I and Ettl H. (1965) Observations on the fine structure of Mesostigma viride Lauterborn. J Linn Soc Lond Bot 59:175–184.
Marin B and Melkonian M. (1999) Mesostigmatophyceae, a new class of streptophyte green algae revealed by SSU rRNA sequence comparisons. Protist 150:399–417.[Medline]
Martin W, Deusch O, Stawski N, Grünheit N, Goremykin V. (2005) Chloroplast genome phylogenetics: why we need independent approaches to plant molecular evolution. Trends Plant Sci 10:203–209.[CrossRef][Web of Science][Medline]
Martin W, Rujan T, Richly E, Hansen A, Cornelsen S, Lins T, Leister D, Stoebe B, Hasegawa M, Penny D. (2002) Evolutionary analysis of Arabidopsis, cyanobacterial, and chloroplast genomes reveals plastid phylogeny and thousands of cyanobacterial genes in the nucleus. Proc Natl Acad Sci USA 99:12246–12251.
McCourt RM, Delwiche CF, Karol KG. (2004) Charophyte algae and land plant origins. Trends Ecol Evol 19:661–666.[CrossRef][Medline]
Melkonian M. (1983) Mesostigma, a key organism in the evolution of two major classes of green algae and related to the ancestry of land plants. Br Phycol J 18:206.
Melkonian M. (1989) Flagellar apparatus ultrastructure in Mesostigma viride (Prasinophyceae). Plant Syst Evol 164:93–122.[CrossRef]
Melkonian M. (1990) Phylum Chlorophyta: class Prasinophyceae. In Margulis L, Corliss JO, Melkonian M, Chapman DJ (Eds.). Handbook of protoctista(Jones and Bartlett Publishers, Boston (MA)) pp. 600–607.
Melkonian M, Marin B, Surek B. (1995) Phylogeny and evolution of the algae. In Arai K, Kato M, Doi Y (Eds.). Biodiversity and evolution(The National Science Mueseum Foundation, Tokyo (Japan)) pp. 153–176.
Moestrup Ø. (1970) The fine structure of mature spermatozoids of Chara corralina, with special reference to microtubules and scales. Planta 93:295–308.[CrossRef]
Moestrup Ø. (1991) Further studies of presumably primitive green algae, including the description of Pedinophyceae class. Nov. and Resultor gen. nov. J Phycol 27:119–133.[CrossRef][Web of Science]
Moestrup Ø. (2002) Phylum Prasinophyta. In John DM, Whitton BA, Brook AJ (Eds.). The freshwater algal flora of the British Isles(Cambridge University Press, Cambridge (UK)) pp. 281–286.
Moestrup Ø and Throndsen J. (1988) Light and electron microscopical studies on Pseudoscourfieldia marina, a primitive scaly green flagellate (Prasinophyceae) with posterior flagella. Can J Bot 66:1415–1434.[CrossRef]
Nakayama T, Marin B, Kranz HD, Surek B, Huss VAR, Inouye I, Melkonian M. (1998) The basal position of scaly green flagellates among the green algae (Chlorophyta) is revealed by analyses of nuclear-encoded SSU rRNA sequences. Protist 149:367–380.[CrossRef][Web of Science]
Nedelcu AM, Borza T, Lee RW. (2006) A land plant-specific multigene family in the unicellular Mesostigma argues for its close relationship to Streptophyta. Mol Biol Evol 23:1011–1015.
Petersen J, Teich R, Becker B, Cerff R, Brinkmann H. (2006) The GapA/B gene duplication marks the origin of Streptophyta (charophytes and land plants). Mol Biol Evol 23:1109–1118.
Philippe H. (1993) MUST, a computer package of management utilities for sequences and trees. Nucleic Acids Res 21:5264–5272.
Philippe H, Delsuc F, Brinkmann H, Lartillot N. (2005) Phylogenomics. Annu Rev Ecol Evol Syst 36:541–562.[CrossRef]
Philippe H, Lartillot N, Brinkmann H. (2005) Multigene analyses of bilaterian animals corroborate the monophyly of Ecdysozoa, Lophotrochozoa, and Protostomia. Mol Biol Evol 22:1246–1253.
Philippe H, Snell EA, Bapteste E, Lopez P, Holland PW, Casane D. (2004) Phylogenomics of eukaryotes: impact of missing data on large alignments. Mol Biol Evol 21:1740–1752.
Philippe H, Zhou Y, Brinkmann H, Rodrigue N, Delsuc F. (2005) Heterotachy and long-branch attraction in phylogenetics. BMC Evol Biol 5:50.[CrossRef][Medline]
Phillips MJ, Delsuc F, Penny D. (2004) Genome-scale phylogeny and the detection of systematic biases. Mol Biol Evol 21:1455–1458.
Pickett-Heaps JD and Marchant HJ. (1972) The phylogeny of the green algae: a new proposal. Cytobios 6:255–264.[Web of Science]
Pombert J-F, Lemieux C, Turmel M. (2006) The complete chloroplast DNA sequence of the green alga Oltmannsiellopsis viridis reveals a distinctive quadripartite architecture in the chloroplast genome of early diverging ulvophytes. BMC Biol 4:3.[CrossRef][Medline]
Pombert J-F, Otis C, Lemieux C, Turmel M. (2005) The chloroplast genome sequence of the green alga Pseudendoclonium akinetum (Ulvophyceae) reveals unusual structural features and new insights into the branching order of chlorophyte lineages. Mol Biol Evol 22:1903–1918.
Posada D and Buckley TR. (2004) Model selection and model averaging in phylogenetics: advantages of akaike information criterion and bayesian approaches over likelihood ratio tests. Syst Biol 53:793–808.
Qiu YL, Lee J, Bernasconi-Quadroni F, Soltis DE, Soltis PS, Zanis M, Zimmer EA, Chen Z, Savolainen V, Chase MW. (1999) The earliest angiosperms: evidence from mitochondrial, plastid and nuclear genomes. Nature 402:404–407.[CrossRef]
Rodriguez-Ezpeleta N, Brinkmann H, Burey SC, Roure B, Burger G, Loffelhardt W, Bohnert H, Philippe H, Lang BF. (2005) Monophyly of primary photosynthetic eukaryotes: green plans, red algae, and glaucophytes. Curr Biol 15:1325–1330.[CrossRef][Web of Science][Medline]
Rodriguez-Ezpeleta N, Brinkmann H, Roure B, Lartillot N, Lang BF, Philippe H. (2007) Detecting and overcoming systematic errors in genome-scale phylogenies. Syst Biol In press.
Rogers CE, Domozych DS, Stewart KD, Mattox KR. (1981) The flagellar apparatus of Mesostigma viride (Prasinophyceae): multilayered structures in a scaly green flagellate. Plant Syst Evol 138:247–258.[CrossRef]
Rogers CE, Mattox KR, Stewart KD. (1980) The zoospore of Chlorokybus atmophyticus, a charophyte with sarcinoid growth habit. Am J Bot 67:774–783.[CrossRef][Web of Science]
Salter LA. (2001) Complexity of the likelihood surface for a large DNA dataset. Syst Biol 50:970–978.
Schmidt HA, Strimmer K, Vingron M, von Haeseler A. (2002) TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics 18:502–504.
Simon A, Glöckner G, Felder M, Melkonian M, Becker B. (2006) EST analysis of the scaly green flagellate Mesostigma viride (Streptophyta): implications for the evolution of green plants (Viridiplantae). BMC Plant Biol 6:2.[CrossRef][Medline]
Stabenau H and Winkler U. (2005) Glycolate metabolism in green algae. Physiol Plant 123:235–245.[CrossRef]
Steel M. (2005) Should phylogenetic models be trying to ‘fit an elephant’? Trends Genet 21:307–309.[CrossRef][Web of Science][Medline]
Swofford DL. (2002) PAUP* phylogenetic analysis using parsimony (*and other methods). Version 4. Sunderland: Sinauer Associates.
Thompson JD, Higgins DG, Gibson TJ. (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22:4673–4680.
Turmel M, Ehara M, Otis C, Lemieux C. (2002) Phylogenetic relationships among streptophytes as inferred from chloroplast small and large subunit rRNA gene sequences. J Phycol 38:364–375.[CrossRef][Web of Science]
Turmel M, Otis C, Lemieux C. (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.
Turmel M, Otis C, Lemieux C. (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.
Turmel M, Otis C, Lemieux C. (2005) The complete chloroplast DNA sequences of the charophycean green algae Staurastrum and Zygnema reveal that the chloroplast genome underwent extensive changes during the evolution of the Zygnematales. BMC Biol 3:22.[CrossRef][Medline]
Turmel M, Otis C, Lemieux C. (2006) The chloroplast genome sequence of Chara vulgaris sheds new light into the closest green algal relatives of land plants. Mol Biol Evol 23:1324–1338.
Whelan S and Goldman N. (2001) A general empirical model of protein evolution derived from multiple protein families using a maximum-likelihood approach. Mol Biol Evol 18:691–699.
Yang Z. (1996) Maximum-likelihood models for combined analyses of multiple sequence data. J Mol Evol 42:587–596.[CrossRef][Web of Science][Medline]
Yang Z. (1997) PAML: a program package for phylogenetic analysis by maximum likelihood. Comput Appl Biosci 13:555–556.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  B. Pils and A. Heyl Unraveling the Evolution of Cytokinin Signaling Plant Physiology, October 1, 2009; 151(2): 782 - 791. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Turmel, M.-C. Gagnon, C. J. O'Kelly, C. Otis, and C. Lemieux The Chloroplast Genomes of the Green Algae Pyramimonas, Monomastix, and Pycnococcus Shed New light on the Evolutionary History of Prasinophytes and the Origin of the Secondary Chloroplasts of Euglenids Mol. Biol. Evol., March 1, 2009; 26(3): 631 - 648. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Gruenheit, P. J. Lockhart, M. Steel, and W. Martin Difficulties in Testing for Covarion-Like Properties of Sequences under the Confounding Influence of Changing Proportions of Variable Sites Mol. Biol. Evol., July 1, 2008; 25(7): 1512 - 1520. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Kolaczkowski and J. W. Thornton A Mixed Branch Length Model of Heterotachy Improves Phylogenetic Accuracy Mol. Biol. Evol., June 1, 2008; 25(6): 1054 - 1066. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Khan, N. Parks, C. Kozera, B. A. Curtis, B. J. Parsons, S. Bowman, and J. M. Archibald Plastid Genome Sequence of the Cryptophyte Alga Rhodomonas salina CCMP1319: Lateral Transfer of Putative DNA Replication Machinery and a Test of Chromist Plastid Phylogeny Mol. Biol. Evol., August 1, 2007; 24(8): 1832 - 1842. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||