| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Division of Life Sciences, Rutgers University, Nelson Hall, 604 Allison Road, Piscataway, NJ 08854, USA
2 Department of Plant Biology, Michigan State University, East Lansing, MI 48824, USA
3 Department of Entomology, Rutgers University, Cook College, Blake Hall, 93 Lipman Drive, New Brunswick, NJ 08901, USA
Correspondence
Stacy Brosnan
szimmer{at}rci.rutgers.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Published online ahead of print on 8 November 2002 as DOI 10.1099/ijs.0.02518-0.
The GenBank/EMBL/DDBJ accession numbers for the LSU rDNA sequences determined in this study are AY130223–AY130242 and AY130814–AY130826, as listed in Table 1
.
Nexus files for three datasets generated for phylogenetic analysis, the SSU rDNA, the LSU rDNA and a combined dataset, a complete alignment of all 34 taxa and a proposed secondary structure of the portion of the LSU rDNA sequenced in this analysis from Euglena gracilis are available as supplementary material in IJSEM Online.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
-1,3-glucan storage product known as paramylon, an intranuclear mitotic spindle with a persistent nucleolus, condensed chromosomes throughout the cell cycle and paraxial rods associated with the flagella (Leedale, 1967
). The evolutionary relationships among these taxa are unclear from the available data, although various classification schemes have been proposed (Stein, 1878; Klebs, 1883; Senn, 1900; Hollande, 1952; Huber-Pestalozzi, 1955; Leedale, 1967; Bourrelly, 1970).
The taxonomic system in common use today has its roots in that proposed by Hollande (1952), who created a scheme that was based largely on body shape (radial symmetry versus asymmetry), nutritional mode (phototrophic, osmotrophic or phagotrophic), flagellar structure and degree of metaboly (rigid versus metabolic). His classification consisted of three major groups, each of which included both pigmented and colourless genera. The presence or absence of chloroplasts was considered to be of minor importance. Later, Leedale (1967) incorporated new physiological and electron microscopic information into the Hollande scheme and created a modified scheme that recognized six separate orders based on similar morphological features. Like Hollande, he grouped photosynthetic and colourless forms together.
Recently, a re-evaluation of the phylogeny of the euglenophytes has been conducted using molecular data (Thompson et al., 1995; Montegut-Felkner & Triemer, 1997; Linton et al., 1999, 2000; Preisfeld et al., 2000, 2001; Moreira et al., 2001; Müllner et al., 2001; Milanowski et al., 2001; Nudelman et al., 2003) in combination with morphological characteristics (Leander & Farmer, 2000, 2001a, b; Leander et al., 2001).
The earliest molecular study (Thompson et al., 1995) was conducted using the chloroplast gene encoding the ribulose-1,5-bisphosphate carboxylase/oxygenase large subunit (rbcL). Their primary objective, however, was to analyse Euglena chloroplast intron evolution, and they therefore focused entirely on the genus Euglena. A high degree of homoplasy in rbcL and misidentification of some taxa prevented Thompson et al. (1995) from strongly supporting their phylogeny. Milanowski et al. (2001) later sequenced the plastid SSU (16S) rDNA for 17 species belonging to the Euglenales. Their analysis also led to the conclusion that the genus Euglena is not monophyletic; however, this gene provided only weak resolution of the internal nodes of the phylogeny.
More recent studies focusing on the nuclear SSU rDNA have confirmed previous speculations, derived from morphological data, that phagotrophy arose early in euglenoid evolution, with the subsequent acquisition of phototrophy, osmotrophy and parasitism (Montegut-Felkner & Triemer, 1997; Linton et al., 1999; Preisfeld et al., 2000). The osmotrophic taxa were polyphyletic, and osmotrophy was shown to have evolved many times. In some instances, osmotrophy arose secondarily due to the loss of photosynthesis (Müllner et al., 2001). In the Euglena clade, for example, the presence of the osmotrophic taxa Astasia longa and Khawkinea quartana indicated that they had lost their chloroplasts (Linton et al., 1999, 2000; Müllner et al., 2001), which correlated with previous data demonstrating the presence of a partial chloroplast genome in A. longa (Seimeister & Hachtel, 1990; Gockel et al., 1994).
The phototrophic taxa formed a monophyletic assemblage at the apex of the tree consisting of two clades, one with taxa having two emergent flagella (Eutreptiales sensu Leedale) and the other with taxa having one emergent flagellum (Euglenales sensu Leedale) (Preisfeld et al., 2001; Moreira et al., 2001). The Eutreptiales diverged prior to the Euglenales and were a paraphyletic group (Preisfeld et al., 2000, 2001; Müllner et al., 2001; Moreira et al., 2001). Within the Euglenales, the phototrophic genera Euglena, Phacus and Lepocinclis were also paraphyletic (Linton et al., 1999, 2000; Müllner et al., 2001; Leander & Farmer, 2001b; Nudelman et al., 2003).
The need for a revision of euglenoid taxonomy is apparent. To accomplish this, however, a clear understanding of the relationships among species is necessary. The chloroplast genes examined to date are unable to resolve internal nodes of the photosynthetic taxa and are not applicable to the colourless genera. Because mitochondrial genes evolve more rapidly (Woese, 1996), they are inappropriate for examination of more distant evolutionary relationships. Therefore, a portion of the nuclear LSU rDNA was chosen to resolve further the phylogeny of the photosynthetic euglenoids. Organisms were included from each of the external clades of the SSU rDNA-based phylogeny to create a combined dataset and compare the LSU and SSU rDNA phylogenies.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Currently, two taxa included for analysis are incorrectly identified in the UTEX culture collection (Culture Center of Algae, Austin, TX, USA): Euglena agilis UTEX 1605 is misidentified as Euglena pisciformis var. obtusa, its synonym (Linton et al., 1999), and Phacus oscillans UTEX 1285 is misidentified as Phacus caudata (see Linton et al., 2000). Twenty-nine of the 34 taxa sequenced for the SSU rDNA came from the same cultures as those from which the LSU rDNA was sequenced. Colacium vesiculosum, Euglena oxyuris, Hyalophacus ocellatus, Lepocinclis bütschlii, Phacus acuminatus and Phacus pleuronectes were sequenced from different cultures for the LSU and SSU rDNA. Four of these six organisms were sequenced from Korean isolates for the SSU rDNA and were unavailable for further study. H. ocellatus and Phacus acuminatus were sequenced from the cultures SAG 228.80 and UTEX 1288, respectively, for the SSU rDNA. Peranema trichophorum, a colourless phagotrophic species, and Eutreptiella sp., a metabolic photosynthetic species with two emergent flagella, were sequenced as outgroups.
|
) with Chelex 100 resin (Bio-Rad) or the DNeasy Tissue kit (Qiagen) using the animal-tissues protocol. A 1·3–1·6 kb region of the LSU rDNA starting at stem 11 (Larsen, 1992) and ending at stem 35a was amplified using 15–20 ng total genomic DNA or 10 µl DNA-containing solution from the Chelex extraction. A 100 µl reaction contained 10 µl DNA (20 ng), 59·5 µl distilled, deionized water, 10 µl 10x reaction buffer with 6 µl 25 mM MgCl2, 4 µl 10 mM dNTPs (Applied Biosystems), 5 µl each primer (10 mM) and 0·5 µl DNA polymerase (5 U µl-1; Sigma). Amplification was carried out using the following program: a hot start at 94 °C for 3 min, followed by 30 cycles of denaturation at 94 °C for 2 min, annealing at 37–45 °C (depending on strain) for 2 min and extension at 72 °C for 4 min, finishing with an extension at 72 °C for 11 min and an indefinite hold at 4 °C. Primers specific to the LSU rDNA in euglenoids were created by comparing the LSU rDNA of Euglena gracilis, which had been sequenced previously by M. N. Schnare and others (GenBank accession no. X53361), with both published LSU rDNA primers (Freshwater & Bailey, 1998; Palumbi, 1996) and the LSU rDNA of other distant taxa, such as Mus. Three published LSU rDNA primers were modified slightly and an additional 17 primers were created (Table 2). The entire LSU rDNA fragment used in this analysis was amplified for each organism using primers 1F and C1R. The PCR product was then sized, cut from an agarose gel and purified using either a QIAEX II gel extraction kit (Qiagen) or the MinElute gel extraction kit according to manufacturer's protocol. The purified DNA was sequenced using an ABI 377 dye-terminator cycle sequencer (Perkin Elmer Applied Biosystems). The entire 1·3–1·6 kb fragment was sequenced for most organisms using three forward primers and three reverse primers: 1F, 2F, 1XF, 2R, 1XR and C1R. The additional primers were created for regions that were not adequately conserved for all taxa sequenced. Forward and reverse primers were used to generate the final LSU rDNA sequence for each organism. Any discrepancies among sequences from different primers were resolved by reading the ABI chromatogram for those primers.
|
, using the genetic data environment (GDE 2.2) program (Smith et al., 1994). The secondary structure of Euglena gracilis was used as a guide in the alignment of the SSU rDNA, as described by Linton et al. (2000). LSU rDNA sequences were aligned with the published Euglena gracilis secondary-structure model of Wuyts et al. (2001) and Schnare et al. (1996), as recommended by Kjer (1995), using Microsoft Word on a PC. Stem–loop structures were numbered according to Larsen (1992). Both of the published secondary-structure models were based on a single species of Euglena and were used as a starting point to infer LSU rDNA structure. With our additional sequences, we were able to design a euglenoid secondary-structure model for the D2 region, using comparative analysis (Gutell et al., 1994; Kjer, 1995), as shown in Fig. 1. The portion of the LSU rDNA sequenced contained three intron regions. These regions were determined for each sequence by comparing them with the published Euglena gracilis sequence from which the introns had been spliced out. These same introns were also located within regions of conserved secondary structure and could be determined by flanking stem regions as well. Only conserved regions that could be unambiguously aligned were used in the analyses. Because of length heterogeneity and insertion of introns in the portion of the LSU rDNA sequenced, a large number of nucleotides could not be used in the analyses. Of the 1300–1600 bp sequenced, only 683 nucleotides were used in the parsimony, distance, maximum-likelihood and Bayesian analyses. Of the 1800–2200 bp sequenced for the SSU rDNA, 1518 were used in the phylogenetic analyses.
|
). Taxa from highly supported clades were sequentially removed and 100 000 random trees were generated to ensure that signal was not present for only one or two strongly supported clades. Signal was also evaluated for relationships among clades by randomly choosing nine taxa (two outgroups and one taxon from each of the seven clades) for each dataset. Ten sets of nine taxa were generated randomly to ensure that signal was not based on relationships of specific taxa in the analysis (Kjer et al., 2001). The left skew of tree distributions from an exhaustive search for each set of nine taxa was examined and the g1 statistic was compared with the significance levels presented by Hillis & Huelsenbeck (1992). Taxa were then removed from the set of nine until signal became random. One of the outgroup taxa, Peranema trichophorum, was removed first, followed by the taxa representing either clade G or clade F (see Fig. 2).
|
) and Bayesian analyses were run using MrBayes 2.01 (Huelsenbeck & Ronquist, 2001).
Equally weighted parsimony analyses were conducted on the three datasets. The equally weighted parsimony analysis (heuristic search with random stepwise addition and 10 000 repetitions) carried out to find a minimum-length tree(s) was performed using tree bisection/reconstruction branch swapping, ACCTRAN character state optimization, MULTREES on, gaps equivalent to missing data and multistate taxa coded as uncertainty. All nucleotides were treated as unordered. Tree support was examined using both the decay index for each node (Bremer, 1994; Donoghue et al., 1992) and non-parametric bootstraps (Felsenstein, 1985). Decay indices were calculated using AutoDecay 4.0 (Eriksson, 1998). For bootstrapping, 1000 pseudoreplicates were performed, each with 10 random addition searches.
To examine the influence of possible nucleotide biases, a log-det minimum evolution analysis (Lockhart et al., 1994) was performed on each dataset. Only parsimony-informative characters were included in the analysis, as suggested by Lockhart et al. (1994). The outgroup, Peranema trichophorum, was removed from the LSU rDNA and combined analyses because some ambiguously aligned regions of the Peranema LSU rDNA sequence were coded as missing, which creates undefined distances in log-det. A heuristic search with starting trees obtained via neighbour-joining using a tree bisection/reconstruction branch-swapping algorithm was used. One hundred bootstraps were generated to examine support for the analyses using 100 pseudoreplicates, with one random addition search.
Evolutionary models were examined for each dataset using both maximum-likelihood and Bayesian analyses. The evolutionary model and parameters were selected after running MODELTEST 3.06 (Posada & Crandall, 1998). For both datasets, MODELTEST indicated the need for a general time-reversible model with a gamma correction for among-site rate variation and invariant sites according to both the likelihood ratio test and the Akaike information criterion (AIC). The only difference between the model selected by the likelihood ratio test and the model selected by the AIC was that the former model assumed equal nucleotide frequencies and only provided values for two rate-matrix substitutions, leaving the others as a value of 1·0. Initial Bayesian analyses were run with a general time-reversible model (nst=6) with rates set to invgamma and nucleotide frequencies set to equal. Each analysis was performed using four Markov chains, with 500 000 cycles for each chain. Trees were saved to a file every 100 cycles, and the first 1000 trees were discarded. Therefore, a majority-rule consensus tree was created from the remaining 4000 trees to examine the posterior probabilities of each clade. The resulting Bayesian evolutionary model parameters were examined for each dataset after completion of the analysis using the ‘sump’ command. The mean, variance and 95 % confidence interval for the percentage of invariable sites, gamma, the nucleotide frequencies and the general time-reversible R-matrices are given (Table 3). Bayesian analyses were also run using the specified likelihood ratio test and AIC parameters from MODELTEST (Table 4). Since the likelihood ratio test and AIC parameters generated nearly identical tree topologies, only trees based on the AIC parameters are presented. Maximum-likelihood analyses (heuristic search with random stepwise addition and 10 repetitions, tree bisection/reconstruction branch swapping, starting branch length using the Rogers–Swofford approximation method and MULTREES on) were performed using both the parameters specified in the aic of MODELTEST and the mean value of the parameters produced from the initial Bayesian analyses. One hundred bootstraps were run for each maximum-likelihood analysis using the fast stepwise-addition option. For the Bayesian and maximum-likelihood analysis of the LSU, the outgroup was constrained as monophyletic.
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
), and shares 95 % sequence similarity. It was originally described as Phacus pyrum var. ovata Playfair (1921), and was later moved to the genus Lepocinclis by Conrad (1934). Because of its similarity in morphology and rDNA sequences, we suggest moving L. ovata back to its original placement in the genus Phacus as Phacus pyrum var. ovata. However, it should be noted that a 5 % sequence divergence exists between Phacus pyrum and Phacus pyrum var. ovata (syn. L. ovata). Whether this divergence is sufficient to warrant the establishment of a novel species is unknown at present.
The culture Trachelomonas conspersa SAG 1280-1 is currently incorrectly named and has a tangled taxonomic history. Originally, T. conspersa was renamed Strombomonas verrucosa var. conspersa by Deflandre (1930) in his monograph establishing the genus Strombomonas. However, Popova (1955) questioned the validity of the species S. verrucosa Defl. (1930) because of its similarity in morphology to another species of Strombomonas described by Deflandre (1930), Strombomonas acuminata (Schmarda) Defl. (1930). On the basis of the original material of S. acuminata and S. verrucosa, she showed that the cell morphology is very variable for both species, and the shape, size and ornamentation of the lorica depend on the age of the cells and their culture conditions. The name S. acuminata takes priority over S. verrucosa and, therefore, in the opinion of Popova (1966), only one taxon exists, S. acuminata (Schmarda) Defl. (1930), and all other names should be considered as synonyms. Therefore, the morphological forms S. verrucosa and S. verrucosa var. conspersa are currently synonymized with S. acuminata. It should be noted that these two taxa have respective sequence divergence of 1·4 and 1·8 % for the LSU rDNA and SSU rDNA, and S. verrucosa var. conspersa is morphologically distinguishable from S. verrucosa. However, until additional Strombomonas species are sequenced, the amount of divergence necessary to classify a taxon as either a variety or a novel species is unknown. S. verrucosa and S. verrucosa var. conspersa will thus be considered as S. acuminata until additional sequence information becomes available.
Three pairs of taxa were found to be identical in both their SSU and LSU rDNA sequences. They were Trachelomonas oblonga var. punctata Lemm. and Trachelomonas hispida (Perty) Stein em. Defl., Phacus alatus Klebs and Phacus pleuronectes (O.F.M.) Dujardin and Phacus acuminatus Stokes and Phacus brachykentron Pochm. The strain of T. oblonga var. punctata obtained from the culture collection was misidentified and should be renamed as T. hispida.
In the genus Phacus, different taxonomic issues arise. Phacus alatus and Phacus pleuronectes (the type species for the genus) are morphologically distinct on the basis of their overall body shape and pellicle structure, yet they have identical SSU and LSU rDNA sequences. Phacus alatus was originally called Phacus pleuronectes var. triquetra and we recommend the use of the varietal name. The varietal status recognizes the differences in morphology with respect to the type species, and retention of the species designation reflects the genetic identity of the two gene sequences we have examined.
Similarly, Phacus acuminatus and Phacus brachykentron are both members of the section Proterophacus Pochmann and share many similarities. Both taxa are described as having oval or egg-shaped bodies that terminate in a pointed tail. One to three rounded paramylon grains are present in both. The size ranges for the two species overlap substantially and the two taxa have identical LSU and SSU rDNA sequences. In general, the cell body of Phacus acuminatus is wider in the posterior half than in the anterior half. However, the degree of widening varies among the varieties of the species, from very wide in Phacus acuminatus var. javana (Pochm.) Huber-Pestalozzi (or subsp. javana Pochm.) to no widening in Phacus acuminatus var. iowensis Allorge et Jahn. In summary, the morphological variability between the two species overlaps sufficiently that, when viewed along with the identity of LSU and SSU rDNA sequences, these taxa can be considered to be the same. Since Phacus acuminatus Stokes (1885) takes precedence over Phacus brachykentron Pochmann (1942), Phacus brachykentron should be considered as Phacus acuminatus.
Secondary structure in D2 and intron
Structure had not been proposed for the D2 region of the LSU rDNA by either of the published secondary-structure models (Wuyts et al., 2001; Schnare et al., 1996). By comparing all 34 euglenoid LSU sequences, we found covarying positions throughout this region that conformed to the models of Gutell et al. (1994) for other eukaryotes. The secondary-structure alignment for Euglena gracilis, based on comparisons with the other euglenoid LSU sequences, is available as supplementary material in IJSEM Online (Supplementary Fig. A), together with a complete alignment of all 34 taxa (Supplementary Fig. B). The proposed secondary-structural model for the D2 region is shown in Fig. 1. Secondary structures for stems 14 and 16 could not be determined using comparative analysis (Gutell et al., 1994; Kjer, 1995) for all 34 euglenoid sequences. Stems 18A, B and C varied in length among taxa and could not be aligned unambiguously. The terminal loop of stem 31A included a large insertion for a number of taxa, causing it to vary in length from 25 to 348 bp. The three introns present in the portion of the LSU rDNA sequenced were located between stems 13 and 18A (intron 2), stems D2_1 and D2_2 (intron 3) and stems D2_3' and D2_6 (intron 4). One taxon, K. quartana, lacked intron 2, and two taxa, Euglena stellata and Phacus acuminatus, lacked intron 3. Within intron 4, we found a stem–loop structure for all but three taxa that was supported by multiple compensatory base changes, shown for E. gracilis in Supplementary Figs A and B.
Signal analysis
Both the LSU and SSU rDNA possessed non-random phylogenetic signal for relationships within clades, but did not contain signal for relationships among clades. The g1 statistic for the SSU rDNA remained significant at the 1 % level (Hillis & Huelsenbeck, 1992) as taxa were removed from strongly supported clades until only eight taxa remained, at which point signal became insignificantly different from random. At this point, only one taxon from each clade and one outgroup remained. The g1 statistic for the LSU rDNA remained significant at the 1 % level until only seven taxa remained (one outgroup and one taxon from six of the seven clades). To examine the lack of signal for relationships among clades further, 10 sets of nine taxa (two outgroups and one taxon from each of the seven clades) were used. Taxa were removed sequentially until signal became insignificantly different from random. For the SSU rDNA, all 10 sets of nine taxa were significant at the 1 % level, with g1 statistics ranging from -1·9924 to -1·1713. However, with the removal of the outgroup Peranema, all 10 datasets became insignificantly different from random, with g1 statistics ranging from 0·1230 to -0·2304. Since Peranema was the last taxon to be removed when we were testing for total signal of the SSU rDNA dataset, it was necessary to check whether all of the signal for the SSU rDNA was dependent on Peranema. Therefore, Peranema was removed from the dataset every five taxa to examine whether signal was still significant. If signal was still significant, it was reinserted into the dataset and another five taxa were removed. The result was that the removal of Peranema did not affect the total signal until only nine taxa remained. The LSU rDNA dataset also was significant at the 1 % level for all 10 sets of nine taxa, with g1 statistics ranging from -0·5994 to -0·3080. The removal of Peranema from the dataset made the signal for four of the 10 datasets not significantly different from random. Four of the remaining datasets were significant at the 1 % level, and the other two were significant at the 5 % level. Subsequent removal of the taxon representing either clade G or clade F (Fig. 2) left only one dataset with signal significantly non-random at the 5 % level. The removal of the taxon representing clade A (Fig. 2) eliminated the significant signal.
Phylogenetic analyses
All analyses for the SSU rDNA, LSU rDNA and the combined dataset support the same seven external clades (Fig. 2). Although there is strong support for the external clades in most analyses, there is both weak and variable support for the relationship among those clades (Fig. 2; and see Figs 3–5 below).
|
|
|
a), log-det (Fig. 3b), Bayesian (Fig. 3c) and maximum-likelihood analyses (Fig. 3d) all generated the same seven external clades, but the relationships among those clades varied. Clades A, B, D and E are strongly supported in all analyses. Phacus agilis falls within clade C, though with little support. Bootstrap numbers for clade C represent the support for the clade excluding Phacus agilis. Trachelomonas volvocinopsis var. spiralis falls within clade A in all SSU rDNA analyses, though with little support. Clades F and G are moderately supported, with some taxa (H. ocellatus and Phacus acuminatus) forming long branches within the tree.
The topologies of the SSU rDNA are not significantly different from the combined Bayesian analysis (Fig. 2). Clades F and G are basal to a clade containing A, B, C, D and E. The relationships among clades A, B, C, D and E vary depending on the type of analysis used. This is not unexpected, considering that the signal for relationships among clades, according to the g1 statistic, is insignificantly different from random. Unlike the combined analysis, the SSU rDNA dataset forms a monophyletic grouping of clade E (the metabolic Euglena) and clade B (Trachelomonas) in all analyses except log-det, which again forms a monophyletic loricate clade (A and B). However, with the exception of the Bayesian analysis, there is little support for this relationship.
LSU rDNA.
Parsimony (Fig. 4a), log-det (Fig. 4b), Bayesian (Fig. 4c) and maximum-likelihood analyses (Fig. 4d) all generated the same seven external clades; however, as in the previous analyses, the relationships among those clades varied. Clades A, B and D were strongly supported in all analyses. Clades C and E were moderately supported in all analyses. There was strong support for some of the relationships within clades G and F for the parsimony, log-det and maximum-likelihood analyses, but there was little support for the clades themselves. Both Bayesian and maximum-likelihood analyses (Fig. 4c, d) generated the same tree topology. Unlike the parsimony, log-det and maximum-likelihood analyses, the Bayesian analyses strongly support six of the seven external clades and relationships among clades A and B and clades D, E, F and G.
The differences in tree topology for the LSU rDNA are largely due to the variability of the clade that is represented as most basal in each tree. Unlike the SSU rDNA and the combined analysis, which place clades F and G at the base of the phylogeny, clades C, D and E are all most basal in one of the trees represented in the analyses of the LSU rDNA. Clades G and F form a clade in all four analyses, but never fall at the base of the tree. The LSU rDNA dataset also differs from the SSU rDNA and combined analysis in its placement of Euglena tripteris, which falls as a sister taxon to the larger F and G clade. However, like the combined analysis, clades A and B are monophyletic in all analyses and include T. volvocinopsis var. spiralis as a sister taxon in the two Bayesian and maximum-likelihood analyses.
Combined dataset.
Parsimony (Fig. 5a), log-det (Fig. 5b), Bayesian (Figs 2 and 5c) and maximum-likelihood analyses (Fig. 5d–e) all generated the same seven external clades. Although relationships among clades vary, some general trends in the combined dataset can be observed.
The loricates (clades A and B and T. volvocinopsis var. spiralis) are monophyletic in all analyses, with the exception of the parsimony analyses, in which T. volvocinopsis var. spiralis is sister to clade C. However, there is no support for the relationships of clade A/B to clades C and D in the parsimony tree. There is strong support for T. volvocinopsis var. spiralis being sister to the two strongly supported loricate clades Strombomonas (A) and Trachelomonas (B) in the log-det and Bayesian analyses. The loricates are also sister to clade C in all analyses, with support of 87 % (Fig. 2) and 84 % (Fig. 5c) in the Bayesian trees, but with only weak bootstrap support (Fig. 5a, b, d, e) in the other trees. Clade D is sister to clade C in all analyses except the maximum-likelihood analysis using the same evolutionary model as in the Bayesian analysis (Fig. 5e). However, there is no support for that relationship. Clade E falls either as sister to the A, B, C, D clade (Figs 2 and 5a, b, d) or as sister to a separate F and G clade (Fig. 5c, e). Neither relationship has strong support. Clades F and G remain sister to each other in all analyses and sometimes form a separate clade, though never with strong support.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
). However, the relationship among the clades varies. This is not surprising, since, according to the g1 statistic, signal for relationships among clades for both the SSU and LSU rDNA is not significantly different from random.
The rate matrices of the combined dataset, the nucleotide composition and the gamma distribution all fall within the 95 % confidence intervals for these parameters for both the SSU and LSU rDNA datasets (Table 3), indicating that they are not evolving at excessively different rates. The combined analysis stabilized the phylogeny, increasing support for some of the internal nodes of the tree; however, because of the lack of signal for these relationships, only tentative conclusions can be drawn about the relationships among the external clades.
Taxonomic revisions
The seven clades found in all trees do not form groupings consistent with the current taxonomy of the Euglenophyta. The genera Euglena and Phacus are not monophyletic. Both genera are split into two separate clades, i.e. clades E and F for Euglena and clades C and G for Phacus. Clade E consists of metabolic Euglena species, whereas the Euglena species in clade F are more rigid. The Phacus species in clade G are the typical flat, leaf-like organisms for which the genus was described. Clade C consists of Phacus species that are more rounded in cross-section. Most of these taxa were originally in the genus Euglena and were later moved to the genus Phacus. A taxonomic revision of Euglena and Phacus is necessary, but additional taxa must still be sequenced before final amendments can be made. However, on the basis of the data available, we recommend the following: since the type species for the genus Euglena, Euglena viridis, resides in clade E, this clade will retain the name Euglena, and a new name should be proposed for clade F; similarly, Phacus pleuronectes (the type species) resides in clade G, and taxa in this clade should retain the name Phacus.
Not enough species of the genus Lepocinclis have been sequenced to determine whether or not it is monophyletic. Most species of Lepocinclis were previously in the genus Euglena and were subsequently moved to Lepocinclis after the genus was established. The Lepocinclis species embedded within the rigid Euglena species in clade F will probably acquire the genus name assigned to clade F.
The loricate genera are monophyletic in most analyses, with a separate clade for the genus Strombomonas (clade A) and another for the genus Trachelomonas (clade B). There is some question as to the placement of T. volvocinopsis var. spiralis. In the SSU rDNA analyses, it groups with the Strombomonas clade; in the LSU rDNA and combined analyses, it sits as a sister taxon to the Strombomonas and Trachelomonas clade. T. volvocinopsis var. spiralis differs in size, shape and ornamentation from the organisms placed within the strongly supported Trachelomonas clade (B). T. volvocinopsis var. spiralis is a small, rounded species with a helical ridge pattern ornamenting its lorica. The other Trachelomonas species in clade B are oblong with ornamentation of pores and spines. The addition of other Trachelomonas species similar to T. volvocinopsis var. spiralis will help to elucidate the placement of these species in the phylogeny and better establish the subgroups within the genus Trachelomonas. Strombomonas will either remain as a separate genus or become a subgroup within Trachelomonas.
The genus Colacium (clade D) is monophyletic in all analyses. The placement of this genus in relation to the other clades remains uncertain, however.
The colourless, osmotrophic species A. longa and K. quartana group with the metabolic Euglena species (clade E), as shown previously by Linton et al. (1999, 2000), indicating that they lost photosynthesis secondarily. Similarly, the colourless H. ocellatus groups with Phacus acuminatus in the flat Phacus clade (G). However, Hyalophacus has a large pairwise divergence from the ingroup taxa and may become a long branch (Figs 3–5).
While the taxa presented in this analysis represent a broad range of photosynthetic euglenoids, it is clear that additional taxa, especially from some of the more speciose genera including Phacus and Trachelomonas, will need to be sampled to resolve the phylogeny in more detail. The presence of the seven strongly supported clades within the euglenoid phylogeny provides a framework for future taxon selection to strengthen our understanding of euglenoid evolution.
Conclusions
Seven strongly supported clades of photosynthetic euglenoids have been identified. Although the relationships among these clades are often weak and may vary, some tentative conclusions can be drawn.
(i) Clades A and B are monophyletic. All taxa that are housed within a lorica are contained in clade A/B, indicating a single origin of the lorica. Taxa within the genus Strombomonas were originally assigned to Trachelomonas. If the Strombomonas clade remains embedded within Trachelomonas, the genus Strombomonas will need to be dissolved.
(ii) Secondary osmotrophy, due to the loss of photosynthesis, has arisen many times, as evidenced by A. longa and K. quartana in clade E and H. ocellatus in clade G.
(iii) The genus Euglena should be limited to those taxa in clade E, which contains the type species, Euglena viridis.
(iv) The genus Phacus should be limited to those taxa in clade G, which contains the type species, Phacus pleuronectes.
(v) The Phacus species in clade C will need to be renamed. Phacus pyrum Stein was originally described as Euglena pyrum by Ehrenberg and later renamed as Lepocinclis pyrum Perty before it became Phacus pyrum Stein. Similarly, Phacus pyrum var. ovata is also known as L. ovata Conrad. The multiple transfers of these taxa among the various genera demonstrate the uncertainty of the current taxonomy. We are now considering retaining the name Lepocinclis and applying it to all taxa in clade C.
(vi) The rigid Euglena and Lepocinclis species in clade F should be combined under a single name. What that name will be remains uncertain. If clades F and G ultimately form a monophyletic assemblage, clade F may need to be combined with clade G under Phacus. Alternatively, a new name will need to be generated for the taxa in clade F. Interestingly, the taxon originally described as Phacus tripteris was moved to Euglena tripteris by Klebs (1883), moved back to Phacus tripteris by Pochmann (1942) and again moved back to Euglena tripteris by Huber-Pestalozzi (1955). Perhaps it is time to move it back to its original placement under Phacus.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bremer, K. (1994). Branch support and tree stability. Cladistics 10, 295–304.[CrossRef]
Conrad, W. (1934). Matériaux pour une monograpie due genre Lepocinclis Perty. Arch Protistenkd 82, 203–249 (in French).
Deflandre, G. (1930). Strombomonas nouveau genre de l'Euglenacees. Arch Protistenkd 69, 551–614 (in French).
Donoghue, M. J., Olmstead, R. G., Smith, J. F. & Palmer, J. D. (1992). Phylogenetic relationships of Dipsacales based on rbcL sequences. Ann MO Bot Gard 79, 333–345.
Eriksson, T. (1998). AutoDecay version 4.0. Program distributed by the author. Department of Botany, Stockholm University, Sweden.
Felsenstein, J. (1985). Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783–791.[CrossRef]
Freshwater, D. W. & Bailey, J. C. (1998). A multigene phylogeny of the Gelidiales including nuclear large-subunit rRNA sequence data. J Appl Phycol 10, 229–236.[CrossRef]
Gockel, G., Hachtel, W., Baier, S., Fliss, C. & Henke, M. (1994). Genes for components of the chloroplast translational apparatus are conserved in the reduced 73-kb plastid DNA of the nonphotosynthetic euglenoid flagellate Astasia longa. Curr Genet 26, 256–262.[CrossRef][Medline]
Goff, L. & Moon, D. (1993). PCR amplification of nuclear and plastid genes from algal herbarium specimens and algal spores. J Phycol 29, 381–384.[CrossRef]
Gutell, R. R., Larsen, N. & Woese, C. R. (1994). Lessons from an evolving rRNA: 16S and 23S rRNA structures from a comparative perspective. Microbiol Rev 58, 10–26.
Hillis, D. M. & Huelsenbeck, J. P. (1992). Signal, noise, and reliability in molecular phylogenetic analyses. J Hered 83, 189–195.
Hollande, A. (1952). Classe des Eugléniens (Euglenoidina Butschli, 1884). In Traité de Zoologie, vol. 1, pp. 238–284. Edited by P. P. Grassé. Paris: Mason (in French).
Huber-Pestalozzi, G. (1955). Das Phytoplankton des Süsswassers. IV. Euglenophyceen. Edited by A. Thienemann. Stuttgart: E. Schweizerbart'sche (in German).
Huelsenbeck, J. P. & Ronquist, F. (2001). MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17, 754–755.
Kjer, K. M. (1995). Use of rRNA secondary structure in phylogenetic studies to identify homologous positions: an example of alignment and data presentation from the frogs. Mol Phylogenet Evol 4, 314–330.[CrossRef][Medline]
Kjer, K. M., Blahnik, R. J. & Holzenthal, R. W. (2001). Phylogeny of Trichoptera (caddisflies): characterization of signal and noise within multiple datasets. Syst Biol 50, 781–816.[CrossRef][Medline]
Klebs, G. (1883). Über die Organization einiger Flagellatengruppen und ihre Beziehungen zu Algen and Infusorien. Untersuchungen Bot Inst Tübingen 1, 233–362 (in German).
Larsen, N. (1992). Higher order interactions of 23S rRNA. Proc Natl Acad Sci U S A 89, 5044–5048.
Leander, B. S. & Farmer, M. A. (2000). Comparative morphology of the euglenid pellicle. I. Patterns of strips and pores. J Eukaryot Microbiol 47, 469–479.[CrossRef][Medline]
Leander, B. S. & Farmer, M. A. (2001a). Comparative morphology of the euglenid pellicle. II. Diversity of strip substructure. J Eukaryot Microbiol 48, 202–217.[CrossRef][Medline]
Leander, B. S. & Farmer, M. A. (2001b). Evolution of Phacus (Euglenophyceae) as inferred from pellicle morphology and SSU rDNA. J Phycol 37, 143–159.[CrossRef]
Leander, B. S., Triemer, R. E. & Farmer, M. A. (2001). Character evolution in heterotrophic euglenids. Eur J Protistol 37, 337–356.[CrossRef]
Leedale, G. F. (1967). Euglenoid Flagellates. Englewood Cliffs, NJ: Prentice Hall.
Linton, E. W., Hittner, D., Lewandowski, C., Auld, T. & Triemer, R. E. (1999). A molecular study of euglenoid phylogeny using small subunit rDNA. J Eukaryot Microbiol 46, 217–223.[Medline]
Linton, E. W., Nudelman, M. A., Conforti, V. & Triemer, R. E. (2000). A molecular analysis of the euglenophytes using SSU rDNA. J Phycol 36, 740–746.[CrossRef]
Lockhart, P. J., Steel, M. A., Hendy, M. D. & Penny, D. (1994). Recovering evolutionary trees under a more realistic model of sequence evolution. Mol Biol Evol 11, 605–612.
Milanowski, R., Zakry
, B. & Kwiatowski, J. (2001). Phylogenetic analysis of chloroplast small-subunit rRNA genes of the genus Euglena Ehrenberg. Int J Syst Evol Microbiol 51, 773–781.[Abstract]
Miyamoto, M. M. & Walter, M. F. (1995). Testing species phylogenies and phylogenetic methods with congruence. Syst Biol 44, 64–76.
Montegut-Felkner, A. E. & Triemer, R. E. (1997). Phylogenetic relationships of selected euglenoid genera based on morphological and molecular data. J Phycol 33, 512–519.[CrossRef]
Moreira, D., López-García, P. & Rodríguez-Valera, F. (2001). New insights into the phylogenetic position of diplonemids: G+C content bias, differences of evolutionary rate and a new environmental sequence. Int J Syst Evol Microbiol 51, 2211–2219.[Abstract]
Müllner, A. N., Angeler, D. G., Samuel, R., Linton, E. W. & Triemer, R. E. (2001). Phylogenetic analysis of phagotrophic, phototrophic and osmotrophic euglenoids by using the nuclear 18S rDNA sequence. Int J Syst Evol Microbiol 51, 783–791.[Abstract]
Nudelman, M. A., Rossi, M. S., Conforti, V. & Triemer, R. E. (2003). Phylogeny of Euglenophyceae based on small subunit rDNA sequences: taxonomic implications. J Phycol 39, 226–235.[CrossRef]
Palumbi, S. R. (1996). Nucleic acids. II. The polymerase chain reaction. In Molecular Systematics, 2nd edn, pp. 205–247. Edited by D. M. Hillis, C. Moritz & B. K. Mable. Sunderland, MA: Sinauer.
Playfair, G. J. (1921). Australian freshwater flagellates. Proc Linnean Soc NSW 46, 99–146.
Pochmann, A. (1942). Synopsis der Gattung Phacus. Arch Protistenkd 95, 81–252 (in German).
Popova, T. G. (1955). Opredelitel Presnovodnych Vodoroslej SSSR, vol. 7, Euglenovyge Vodorosli. Moscow & Leningrad: Sovetskaia Nauka (in Russian).
Popova, T. G. (1966). Flora Sporovyky Rastenij SSSR, vol. 8, Euglenophyta [Gen. Trachelomonas, Strombomonas, Eutreptia, Euglena]. Moscow & Leningrad: Izd. Akademii Nauka (in Russian).
Posada, D. & Crandall, K. A. (1998). MODELTEST: testing the model of DNA substitution. Bioinformatics 14, 817–818.
Preisfeld, A., Berger, S., Busse, I., Liller, S. & Ruppel, H. G. (2000). Phylogenetic analyses of various euglenoid taxa (Euglenozoa) based on 18S rDNA sequence data. J Phycol 36, 220–226.[CrossRef]
Preisfeld, A., Busse, I., Klingberg, M., Talke, S. & Ruppel, H. G. (2001). Phylogenetic position and inter-relationships of the osmotrophic euglenoids based on SSU rDNA data, with emphasis on the Rhabdomonadales (Euglenozoa). Int J Syst Evol Microbiol 51, 751–758.[Abstract]
Schnare, M. N., Damberger, S. H., Gray, M. W. & Gutell, R. R. (1996). Comprehensive comparison of structural characteristics in eukaryotic cytoplasmic large subunit (23S-like) ribosomal RNA. J Mol Biol 256, 701–719.[CrossRef][Medline]
Seimeister, G. & Hachtel, W. (1990). Structure and expression of a gene encoding the large subunit of ribulose-1,5-bisphosphate carboxylase (rbcL) in the colourless euglenoid flagellate Astasia longa. Plant Mol Biol 14, 825–833.[CrossRef][Medline]
Senn, G. (1900). Euglenineae. In Die natürlichen Pflanzenfamilien, pp. 173–185. Edited by A. Engler & K. Prantl (in German).
Smith, S. W., Overbeek, R., Woese, C. R., Gilbert, W. & Gillevet, P. M. (1994). The genetic data environment: an expandable GUI for multiple sequence analysis. Comput Appl Biosci 10, 671–675.
Sogin, M. L. & Gunderson, J. H. (1987). Structural diversity of eukaryotic small subunit ribosomal RNAs: evolutionary implications. Ann N Y Acad Sci 573, 125–139.
Stein, F. R. (1878). Der Organismus der Infusionsthiere. III Abt. Der Organismus der Flagellaten. Leipzig: W. Engelmann (in German).
Stokes, A. C. (1885). Some apparently undescribed Infusoria from fresh water. Am Nat 19, 18–27.[CrossRef]
Swofford, D. L. (2002). PAUP*: phylogenetic analysis using parsimony (*and other methods), version 4.0b10. Sunderland, MA: Sinauer.
Thompson, M. D., Copertino, D. W., Thompson, E., Favreau, M. R. & Hallick, R. B. (1995). Evidence for the late origin of introns in chloroplast genes from an evolutionary analysis of the genus Euglena. Nucleic Acids Res 23, 4745–4752.
Woese, C. R. (1996). The world of ribosomal RNA. In Ribosomal RNA: Structure, Evolution, Processing, and Function in Protein Biosynthesis, pp. 23–47. Edited by A. Zimmermann & A. E. Dahlberg. Boca Raton, FL: CRC Press.
Wuyts, J., De Rijk, P., Van de Peer, Y., Winkelmans, T. & De Wachter, R. (2001). The European large subunit ribosomal RNA database. Nucleic Acids Res 29, 175–177.
| ||||||||||||||||||||||||||||||||||||
