The evolution of paralogous enzymes MAT and MATX within the Euglenida and beyond
© Szabová et al.; licensee BioMed Central Ltd. 2014
Received: 20 August 2013
Accepted: 30 December 2013
Published: 11 February 2014
Methionine adenosyltransferase (MAT) is a ubiquitous essential enzyme that, in eukaryotes, occurs in two relatively divergent paralogues: MAT and MATX. MATX has a punctate distribution across the tree of eukaryotes and, except for a few cases, is mutually exclusive with MAT. This phylogenetic pattern could have arisen by either differential loss of old paralogues or the spread of one of these paralogues by horizontal gene transfer. Our aim was to map the distribution of MAT/MATX genes within the Euglenida in order to more comprehensively characterize the evolutionary history of MATX.
We generated 26 new sequences from 23 different lineages of euglenids and one prasinophyte alga Pyramimonas parkeae. MATX was present only in photoautotrophic euglenids. The mixotroph Rapaza viridis and the prasinophyte alga Pyramimonas parkeae, which harbors chloroplasts that are most closely related to the chloroplasts in photoautotrophic euglenids, both possessed only the MAT paralogue. We found both the MAT and MATX paralogues in two photoautotrophic species (Phacus orbicularis and Monomorphina pyrum). The significant conflict between eukaryotic phylogenies inferred from MATX and SSU rDNA data represents strong evidence that MATX paralogues have undergone horizontal gene transfer across the tree of eukaryotes.
Our results suggest that MATX entered the euglenid lineage in a single horizontal gene transfer event that took place after the secondary endosymbiotic origin of the euglenid chloroplast. The origin of the MATX paralogue is unclear, and it cannot be excluded that it arose by a gene duplication event before the most recent common ancestor of eukaryotes.
KeywordsMethionine adenosyltransferase Horizontal gene transfer Deep paralogy Gene evolution Euglenozoa
Methionine adenosyltransferase (MAT) is a cytosolic ubiquitous enzyme that synthesizes S-adenosyl-L-methionine (SAM), a molecule that is one of the most important metabolites in living cells. SAM serves as the major methyl donor to phospholipids, DNA, RNA and other small molecules and is the second most widely used enzyme substrate after ATP [1, 2]. MAT is a well-conserved enzyme that is encoded in the genomes of most eukaryotes, eubacteria, and archaebacteria (which have a highly divergent version of the gene) and has been well studied at the primary, secondary, and tertiary structural levels [3–5]. Except for the mammalian MAT II, which is a hetero-oligomer , members of the MAT family are homo-oligomers that usually form tetramers consisting of four identical subunits; the two active sites are located between the subunits in each dimer . Mammalian MAT III and archaeal MATs form dimers .
Multiple sequence alignments of MAT genes from a wide diversity of eukaryotes demonstrated a paralogue of MAT, named MATX, with distinctive features that are absent in all other eukaryotic MATs. These features include four specific insertions and a large number of unique substitutions . The recombinant MATX from Euglena gracilis has been found to function as a homo-dimer with activities comparable to MATs from other eukaryotes . Molecular phylogenetic analyses clearly showed that MATX is related to other eukaryotic MATs, but it forms a long branch in the eukaryotic subtree . The majority of MATX paralogues occur in four distantly related groups of photosynthetic eukaryotes: haptophytes, photosynthetic euglenids, diatoms, and dinoflagellates. MATX was also detected in a pelagophyte alga Aureococcus anophagefferens. All organisms possess either the MAT or the MATX form of the gene, with the exception of five diatom species that have both paralogues and A. anophagefferens that harbors two different homologues of MAT in addition to MATX [8, 10].
A similar punctate distribution of two paralogues with the same function was reported for “elongation factor 1-alpha” (EF-1α) and its paralogue “elongation factor like” (EFL), which are highly conserved members of a GTPase superfamily involved in translation. Like MAT/MATX, the EF-1α/EFL paralogues have a patchy distribution across the tree of eukaryotes and rarely occur together in the same organism. EFL has been localized so far in eight groups of unrelated organisms: dinoflagellates, haptophytes, cercozoans, green algae, choanoflagellates, fungi, diatoms, and radiolarians [11–17].
The punctate distributions of MAT/MATX and EF-1α/EFL across the tree of eukaryotes can be explained by two scenarios: (1) a deep paralogy, whereby both paralogues were present in an ancient common ancestor followed by differential loss of one or the other paralogue in descendant lineages; and (2) a horizontal (syn., lateral) gene transfer (HGT), whereby a more recent origin of one paralogue (most likely the less frequent one, such as MATX) in one lineage of eukaryotes is followed by the spread of this paralogue to other distantly related lineages via horizontal transfer.
These scenarios differ in their assumptions. The first scenario hypothesizes coexistence and probably co-expression of both paralogues in one cell for a long time without negative effects on the organism. This scenario explains the distribution purely by vertical transmission. In this case, MATX must have originated by gene duplication from the MAT already present in the common ancestor of all MATX containing taxa. This organism was very ancient and not very distantly related, maybe identical, to the most recent common ancestor of eukaryotes. Since that time, MAT and MATX must have been propagated side by side in the genomes of the descendants to much more recent nodes of eukaryotic evolution and in some cases (diatoms) even to extant organisms.
The second scenario assumes that one (MATX) can be horizontally transferred and is capable of functional replacement of the MAT form soon after the transfer. Our previous work on the model systems of Euglena gracilis and Trypanosoma brucei indicates that MATX fulfills the assumptions for both of these scenarios, because this paralogue can be co-expressed with MAT and can immediately take over its function . By contrast, EFL was capable of long-term co-expression, but was not able to functionally replace EF1-α. Based on these results, neither of the two evolutionary scenarios can be refuted for MAT/MATX. However, in the case of EF1-α/EFL, HGT is apparently more difficult and likely played a less important role in the evolutionary history of this paralogue couple .
There are several questions associated with the putative HGT explanation for the origin and distribution of the MATX paralogue that remain unanswered. For instance, under what circumstances would the highly divergent MATX evolve within one recent group of eukaryotes and in which lineage could it happen? One hypothesis posits that MATX evolved during a secondary endosymbiotic origin of plastids from the endosymbiont copy of the MAT gene, which was released from purifying selection and underwent accelerated sequence evolution . Therefore, an analysis of the distribution of MAT/MATX in euglenids provides an opportunity to evaluate this possibility.
The Euglenida is a large group of marine and freshwater eukaryotic flagellates with diverse modes of nutrition, including phagotrophy, osmotrophy, photoautotrophy, and a recently discovered example of mixotrophy (a euglenid capable of both phagotrophy and photosynthesis) [19, 20]. Photosynthetic and secondarily osmotrophic euglenids (i.e., colorless euglenids that have lost photosynthesis) form a monophyletic group that is the sister lineage to the mixotrophic Rapaza viridis and is nested within a paraphyletic assemblage of phagotrophic euglenids. It is inferred that the secondary chloroplast was gained through secondary endosymbiosis in the most recent common ancestor of all photosynthetic euglenids, including R. viridis[19–22]. The marine flagellate Pyramimonas (Pyramimonadales, Prasinophyta) is inferred to be the closest known relative of the euglenid chloroplasts (Turmel et al. 2009). In this study, we investigated the distribution of MAT and MATX in euglenids and Pyramimonas in order to evaluate whether the origin of MATX occurred simultaneously with the secondary endosymbiotic origin of the euglenid chloroplast. These data were also expected to provide insights into whether euglenids were the first group of eukaryotes to evolve the MATX paralogue.
MAT and MATX phylogeny and distribution of MATX in euglenids
Sources of sequences applied in this study
▲ KF383286 MAT
▲ KF383290 MATX
In the phylogenetic tree (Figure 1), MATX paralogues formed a well-supported clade that was separated from the MAT paralogues by a long stem. The tree was rooted by five bacterial outgroups within the MAT paralogues, with Trichomonas vaginalis MAT being the most basal branch. However, the backbone topology of the MAT tree was weakly supported, and the MATX branch was situated only one node apart from prokaryotes. We used Kishino Hasegawa (KH), weighted KH (WKH), Shimodaria Hasegawa (SH) and weighted SH (WSH) tests to evaluate whether the root position between MAT and MATX paralogues is significantly worse than the suggested root on the T. vaginalis branch. The tests showed that this root position cannot be excluded (p = 0.076 for KH and WKH, p = 1.00 for SH and p = 0.945 for WSH).
The MATX sequences from photoautotrophic euglenids formed a well-supported subclade (bootstrap 77%) within the more inclusive MATX clade and branched as the sister group to a clade consisting of Lactuca, dinoflagellates and Dendroctonus. The MAT sequences from the heterotrophic euglenids clustered together with kinetoplastids; the MAT sequence from P. parkeae branched together with other green algae; and the MAT sequences from M. pyrum and P. orbicularis clustered with ciliates and Aureococcus, respectively.
Comparison of MATX and SSU rRNA gene phylogeny
Results of topology tests
7*10 -6 /0
MATX excl. APLD (2)
1*10 -4 /2*10 -4
5*10 -5 /4*10 -6
MATX rooted (3)
8*10 -6 /0
1*10 -5 /0
1*10 -4 /0
MATX rooted exl. APDL (4)
2*10 -4 /2*10 -4
MATX euglenids (5)
Similarly we compared the MATX topology (Additional file 4: Figure S2) with the SSU rRNA gene tree (Additional file 3: Figure S3) and manual species topologies (Additional file 1) of the subclade of photosynthetic euglenids. In this case, the tests showed that the euglenid “species topologies” cannot be rejected (p > = 0.003).
Distribution of MAT and MATX paralogues in euglenids
We detected MATX only in photoautotrophic euglenids. Rapaza viridis, which contains secondary chloroplasts and represents the earliest diverging lineage within the photoautotrophic clade, apparently possesses only the MAT form of the gene; the same holds for the heterotrophic euglenids (Petalomonas, Distigma and Peranema) and Pyramimonas parkeae, which contains the closest known relative of the euglenid chloroplast. Therefore, our results suggest that MATX is specific for the clade of photoautotrophic euglenids after the split of Rapaza. We also found two exceptions within the clade of photoautotrophic euglenids; P. orbicularis and M. pyrum both possess the MAT and MATX paralogues in their cDNAs, so both genes are transcribed in these species. The MATX form in these two species is located within the MATX clade with other photoautotrophic euglenids, while the MAT form is unrelated to euglenid MATs; the MAT of P. orbicularis branches together with the MAT sequences from Aureococcus, and the MAT in M. pyrum branches together with the MAT sequence from ciliates. These facts are most likely explained by two independent horizontal gene transfers of MATs from two different sources into two different lineages of euglenids.
Evolution of the MAT and MATX paralogues
We will focus on how well the observed data fit within the context of the two alternative hypotheses for the evolution of MAT and MATX in euglenids in particular and eukaryotes in general: (A) the deep paralogy scenario and (B) the horizontal gene transfer scenario (Figure 3). Let us first suppose that the deep paralogy scenario (Figure 3A) is correct. This scenario requires at least four independent losses of the MATX gene to explain its distribution in euglenids and many more losses of MATX to explain its distribution within the tree of eukaryotes. Gene losses are frequent events and many losses are not in themselves unlikely. Slightly suspicious, however, is the discrepancy in the number of MAT losses versus the number of MATX losses in this scenario. MAT was lost in euglenids (and within the Euglenozoa) only once, while MATX was lost at least four times only within euglenids. A similar disproportion of losses is present in the tree of eukaryotes. If we compare the MAT/MATX history to the case of EF-1α/EFL, the discrepancy is not as significant in the EF-1α/EFL case; the occurrence of EFL is more fragmented not only in euglenids but also in other eukaryotic groups [15–17, 23]. To our knowledge, it is impossible to evaluate the significance of the observed disproportion between the number of losses of one paralogue compared to the other, so we must conclude that in this respect our observations do not contradict the deep paralogy scenario.
Moreover, if the deep paralogy scenario is correct (Figure 3A), then we would expect both paralogues MAT and MATX to be present in the most recent common ancestor of all MATX-containing taxa, which is likely identical to the most recent common ancestor of eukaryotes. If so, then we would expect that the root of the tree in Figure 1 will be positioned between the MAT and MATX lineages. This is true for EF-1α/EFL tree . In the case of MAT/MATX, the bacterial outgroups form the sister branch to MAT of Trichomonas vaginalis, and the MATX clade is positioned within the MAT lineages. However, the bootstrap values supporting the backbone of the MAT/MATX tree are very low (Figure 1), and the root position on the MATX branch was not rejected by the statistical tests. In this respect our data do not contradict the deep paralogy scenario.
The deep paralogy scenario also assumes that the two paralogues can be co-expressed together in one organism. The observation that the two paralogues are simultaneously present in the transcriptomes of two different euglenids (P. orbicularis and M. pyrum), five diatoms, and Aureococcus demonstrates that this is indeed possible. Moreover, we have confirmed this fact experimentally on the model system of Euglena gracilis and Trypanosoma brucei. In this respect the data do not contradict the deep paralogy scenario.
Finally, the deep paralogy scenario expects that the relationships between the eukaryotic groups in the MATX part of the tree will correspond to the accepted eukaryotic phylogeny, because the gene, despite being lost in many lineages, has evolved vertically. This is apparently not true, because MATX sequences in dinoflagellates form a relatively robust sister branch to MATX sequences in euglenids (bootstrap = 81%), even though dinoflagellates are in fact more closely related to apicomplexans, ciliates, stramenopiles (including diatoms) and haptophytes. More importantly, the conflict between the global MATX phylogeny and the species phylogeny of the MATX containing taxa was significant in statistical tests. Within the clade of photoautotrophic euglenids, the MATX phylogeny also differed from species tree, but this difference was not significant. In this last respect, therefore, our data do contradict the scenario of deep paralogy followed by differential losses in its purest form. In order to explain this observation, we must invoke either horizontal gene transfers within the MATX clade or at least two more gene duplications and subsequent differential losses of putative paralogues within the MATX clade. The latter case would assume that some ancestral organisms would harbor at least four paralogues of this enzyme, which is inconsistent with the observation that most extant species contain only one paralogue (see Additional file 1); therefore, we conclude that MATX has not evolved vertically.
Let us now suppose that the horizontal gene transfer scenario is correct. The first assumption of this scenario is that the MATX paralogue is capable of horizontal transfer. The ability of the MATX paralogue to substitute the function of MAT has been proven experimentally in E. gracilis and T. brucei. In this study, we have also revealed two relatively clear cases of MAT horizontal transfers from different sources into P. orbicularis and M. pyrum. In order to explain the distribution of MATX in euglenids through HGT, we only require a single horizontal gene transfer shortly after Rapaza viridis split from the other photoautotrophic euglenids (Figure 3B); only a few more horizontal gene transfers would be necessary to explain the distribution of MATX in all eukaryotes. Taken together, the data suggests that MATX is capable of HGT and the number of required events is low. In this respect, the data do not contradict the horizontal gene transfer scenario.
The second assumption of the HGT scenario is that there was a eukaryotic group in which the MATX first evolved and then subsequently spread into other lineages of eukaryotes. Such a group would ideally appear as a paraphyletic assemblage near the very base of MATX clade. At the same time, the root of the MAT/MATX tree would be situated inside the MAT paralogues. The data collected so far do not suggest any source group, because the taxa with MATX either form monophyletic groups (e.g., euglenids and dinoflagellates) or have unclear phylogenetic positions (e.g., diatoms, haptophytes and Aureococcus). Our working hypothesis that the MATX originated during the secondary endosymbiotic origin of the euglenid chloroplast () is not supported by the fact that the MATX paralogue is absent in both Rapaza viridis and the closest relative of the euglenid chloroplast, Pyramimonas. Moreover, the MATX paralogues in euglenids do not form a paraphyletic group, but instead form a robust clade within the more inclusive MATX clade. The position of the root between MAT and MATX lineages cannot be rejected, and both paralogues might have been present in the common ancestor of all eukaryotes. The current data are in this respect not in direct conflict but, at the same time, they are also not supportive of the horizontal gene transfer scenario.
Our data are not entirely consistent with either of the two scenarios for MAT/MATX evolution in their purest forms. The hypothesis of deep paralogy followed by differential losses is rejected by the fact that MATX did not evolve purely by vertical transmission. The hypothesis of a more recent origin of MATX followed by spread via horizontal gene transfers is complicated by the absence of a source of the first MATX paralogue and the fact that both paralogues could be present in the most recent common ancestor of all eukaryotes. Therefore, we infer that the MATX paralogue spread among eukaryotes via HGT; however, the original source of MATX is not yet known and it could originate by gene duplication from MAT in the last eukaryotic common ancestor.
We also infer that euglenids were not the group in which the MATX paralogue evolved. Instead, a foreign MATX paralogue substituted the ancestral euglenid MAT paralogue in a single horizontal gene transfer event that occurred after the secondary endosymbiotic origin of the euglenid chloroplast (Figure 3B). Although the donor of the euglenid MATX paralogue is not known, the MATX paralogue, once established, may have evolved vertically within the clade of photoautotrophic euglenids. Two photoautotrophic euglenids (P. orbicularis and M. pyrum) regained a new version of the MAT paralogue by recent horizontal gene transfers from two different eukaryotic lineages and now contain both paralogues. Overall, the case study of MAT/MATX illustrates the complex evolutionary histories of some eukaryotic genes and highlights the prevalence of gene duplications, differential losses of paralogues, and horizontal gene transfer events during the course of eukaryotic evolution.
Euglenid strains and culture conditions
All cultures used in this study are listed in Table 1. Strains of Eutreptiella gymnastica (SCCAP K-0333), Trachelomonas sp. (SCCAP K-1380) and Pyramimonas parkeae (SCCAP K-0007) were obtained from the Scandinavian Culture Collection of Algae and Protozoa (SCCAP). Strains of Monomorphina pyrum (CCAP 1261/4B) and Monomorphina aenigmatica (CCAP 1261/9) were obtained from the Culture Collection of Algae and Protozoa (CCAP). Distigma sp. was isolated from samples collected from freshwater sediment from Czech Republic (50°27’N, 13°20’E). This culture was not monoeukaryotic and contained various other protists, therefore, we used a method of single cell cloning by serial dilution to obtain a monoclonal Distigma sp. culture. Rapaza viridis was isolated and cultured from marine sediment samples from Canada (48° 47.551’ N, 125° 06.974’ W) . Euglena clara (SAG 25.98), Euglena gracilis (SAG 1224-5/25), Euglena proxima (SAG 1224-11a), Eutreptia viridis (SAG 1226-1c), were obtained from the Culture Collection of Algae at Goettingen, Germany. Euglena stellata (UTEX 372), Trachelomonas volvocina (UTEX 1327), Monomorphina parapyrum (UTEX 2354) and Euglenaria anabaena (UTEX 373) were obtained from the Culture Collection of Algae at the University of Texas, Austin Texas, USA. Euglena viridis (ATCC PRA110) was from the American Type Culture Collection, Manassas, Virginia, USA and Eutreptiella braarudii (CCMP 1594) was obtained from the National Center for Marine Algae and Protozoa, East Boothbay, Maine, USA. Phacus inflexus (ACOI 1336) and Phacus orbicularis (ACOI 996) were obtained from the Coimbra Collection of Algae, Coimbra, Portugal. Culture of Petalomonas cantuscygni (CCAP 1259/1) was provided by Dr. Mark Farmer at the University of Georgia, Athens, Georgia, USA and it was originally obtained from the Culture Collection of Algae and Protozoa. Strombomonas accuminata NJ, S 716 and Trachelomonas ellipsoidalis NJ, ST1 are cultures maintained in the Triemer lab which were originally isolated from pond samples from New Jersey, USA; Lepocinclis tripteris MI 101 and Lepocinclis playfairiana MI 102 are cultures isolated from ponds near Michigan State University, East Lansing, MI, USA.
DNA, RNA isolation and preparation of cDNA
Genomic DNA from Eutreptiella gymnastica, Trachelomonas sp., Pyramimonas parkeae, Monomorphina pyrum, Monomorphina aenigmatica, and Distigma sp. was extracted from strains using the Qiagen Blood and Tissue kit and total RNA was isolated from 150 ml of well-grown cultures (approx. 25*106 cells) using TRIzol Reagent (Invitrogen). Total RNA from Rapaza viridis was isolated using Ambion® RNAqueous-Micro Kit (Life technologies). mRNA was purified from total RNA with the use of Dynabeads mRNA Purification Kit (Invitrogen). cDNA was then prepared using Smarter PCR cDNA Synthesis Kit (Clontech) according to the manufacturer’s protocol with 15 to 27 cycles of cDNA amplification (depending on the amount of mRNA used in the first-strand synthesis).
In case of E. gracilis, M. parapyrum, S. accuminata and L. playfairiana the total RNA was extracted by grinding wet biomass in liquid nitrogen followed by purification using RNA/DNA Maxi Kit (Qiagen); mRNA, whenever used for cDNA synthesis, was purified from total RNA using Qiagen Oligotex mRNA Maxi Kit. cDNA was prepared using Smart (later Smarter) cDNA synthesis Kit (Clontech) or by similar technology provided by MINT cDNA synthesis Kit (Evrogen). cDNA libraries were normalized using Trimmer cDNA normalization Kit (Evrogen). The resulting normalized cDNA was adapted for Roche 454 sequencing by performing a multiple last amplification step, pooling the PCR products in order to achieve the overall amount of cDNA acceptable for sequencing.
For the remaining euglenid strains, total RNA was isolated using RNAzol RT RNA Isolation Reagent (Molecular Research Center, Inc.). High level purification of total RNA was achieved using MEGAclear Kit (Ambion). Next, mRNA was isolated using MIcroPoly(A)Purist Kit (Ambion). Preparation of cDNA suitable for the next generation sequencing was according to cDNA Rapid Library Preparation Manual (Roche, GS FLX Titanium Series, later GS FLX + Series - XL+).
Amplification, sequencing and assembly
In case of Pyramimonas parkeae, Eutreptiella gymnastica, Trachelomonas sp., Distigma sp., Monomorphina aenigmatica and Monomorphina pyrum we have amplified the MAT or MATX genes from cDNA template using slightly modified primers of Kamikawa et al. : Forward primer MATA3-F (5’-GAGYMMGTSAVYGARGGYCAYCCXGACAA-3‘) directed at the consensus amino acid (aa) sequence GHPDK and the reverse primer MATB3-R (5’-CCRTGNGCNCCCCADCCDCCRTAXGT-3’) directed at the eukaryotic consensus aa sequence TYGGWGAH inside a conserved block. Amplification was carried out in 25-μl reactions with 1.5 μl of the diluted cDNA as a template using EmeraldAmp MAX PCR Master Mix (TaKaRa Bio Inc.) and the following program: a hot start at 95°C for 4 min, followed by 35 cycles of denaturation at 95°C for 30 s, annealing at 55°C for 60 s and extension at 72°C for 90 s, finishing with an extension at 72°C for 15 min. The PCR products were excised from the gel, cloned into pGEM-T Easy Vector System (Promega) and sequenced. The new sequences were deposited in GenBank under the accession numbers listed in Table 1.
Small subunit (SSU) ribosomal RNA gene from E. gymnastica was amplified from genomic DNA with “universal” eukaryote SSU primer pairs Medlin A (5’-CTGGTTGATCCTGCCAG-3‘), Medlin B (5’-TGATCCTTCTGCAGGTTCACCTAC-3’) described by Medlin et al. . Amplification was carried out using the following program: a hot start at 95°C for 4 min, followed by 35 cycles of denaturation at 95°C for 30 s, annealing at 55°C for 60 s and extension at 72°C for 90 s, finishing with an extension at 72°C for 15 min. Medlin A, Medlin B, EPA-23 (5’- GTCATATGCTTYKTTCAAGGRCTAAGCC -3’), EPA-2286 (5’- TCACCTACARCWACCTTGTTACGAC -3’) according to Müllner et al.  and our primers SSU 633-F (5’- GGCAGCAGGCRCGCAAATTGC -3’) and SSU 2031-R (5’- TCAACCAGACAAATCACTYCACCAA -3’) were used for sequencing of PCR products.
Small subunit (SSU) ribosomal RNA gene from L. playfairiana and M. parapyrum was amplified from genomic DNA with nuclear SSU primers 18S_1A (AAYCTGGTTGATCCTGCCAGT) and 18S_1520B (TGATCCTTCTGCAGGTTCACCTAC). Amplifications were carried out using 5 min of denaturation at 94°C and 30 cycles of the following: 94°C for 30 s, 45°C – 50°C for 1 min, 72°C for 2 min, a final extension at 72°C for 11 min. For sequencing of PCR products were used primers 18S_1A, 18S_1520B, 18S_300F (WGGGTTYGATTCCGGAG), 18S_528F (CGGTAATTCCAGCTCC), 18S_516R (ACCAGACTTGCYCTCC), 18S_960F (TTTGACTCAACRCGGG) and 18S_1055R (CGGCCATGCACCACC).
For the 454 sequences obtained from cDNAs, the raw reads (SFF File format) from 454 were filtered to remove reads shorter than 50 bp and all reads which had more than 30% of the bases with a Phred quality score less than 30 using NGS QC TK  were excluded. The resulting high quality reads were assembled using Roche's proprietary "Newbler" software version 2.6 with "cDNA" option. Assembled contigs shorter than 200 bp were excluded.
The full length of euglenid MATX genes were 1290 bp. Some of the sequences were incomplete: P. orbicularis (length 1257 bp), M. pyrum (length 906 bp), M. aenigmatica (length 843 bp), Trachelomonas sp. (length 909 bp) and E. anabaena (length 1266 bp). The length of the MAT genes were 1167 bp for P. cantuscygni, 1137 for P. orbicularis, 795 bp for R. viridis, 774 bp for M. pyrum, 765 bp for Distigma sp. and 720 bp for P. parkeae.
The MAT and MATX protein sequences were aligned in ClustalX , the SSU rRNA gene sequences were aligned in MAFFT (http://www.genome.jp/tools/mafft/) using G-INS-I option . The alignments were manually refined in BioEdit 18.104.22.168. . The regions, which could not be unambiguously aligned, were excluded from the analyses.
A phylogeny of eukaryotic MAT and MATX was inferred from 123 sequences using 347 aligned amino acid positions; the phylogenetic relationships within the MATX clade were inferred from 41 sequences and 405 positions; the phylogenetic relationships within the euglenid subgroup of the MATX clade were inferred from 21 sequences and 399 alignment positions. Maximum likelihood trees were estimated by RAxML_HPC version 2.3.3  using the best fitting models as determined by Prottest (http://darwin.uvigo.es/software/prottest2_server.html)  and 10 replicates of starting tree construction. The models were PROTGAMMALG for MAT + MATX and MATX of euglenids and PROTGAMMAWAG for analysis of eukaryotic MATX clade. Bootstrap supports (BS) were calculated from 500 replicates. Bayesian trees were estimated by MrBayes version 3.1.2 (Ronquist and Huelsenbeck 2003) using the WAG + GAMMA + Invariants + covarion model of substitution. In case of MAT + MATX analysis (Figure 1), two MCMC were run for 5 860 000 generations, trees from the first 1000 000 generations were discarded as burn-in. In case of MATX analysis (Figure 2), two MCMC were run for 17 775 000 generations, trees from the first 2 818 500 generations were discarded as burn-in.
For the purposes of topology testing, pruned and rooted data sets of MATX clade were analyzed – 40 sequences (only one Karenia brevis sequence was used), 36 sequences (without Aureococcus, Prymnesium, Dendroctonus and Lactuca) and both previous data sets rooted by Trichomonas and Escherichia (i.e. 42 and 38 sequences). All alignments contained 405 amino acid positions and were analysed as described above. Phylogenetic trees of SSU rDNA were inferred by maximum likelihood method from the corresponding set of taxa – 40 and 36 sequences in unrooted, 42 and 38 sequences in rooted analyses of MATX clade and 21 sequences of MATX containing euglenids. Unrooted and rooted SSU alignments contained 1525 and 1282 positions respectively. A maximum likelihood trees were estimated by RAxML_HPC version 2.3.3  using the GTRGAMMA model of nucleotide substitution, 10 replicates of starting tree construction and BS were calculated from 500 replicates.
All data sets and trees generated in this study have been deposited in TreeBASE (study accession number is 15062).
The Kishino Hasegawa (KH)  and Shimodaria Hasegawa tests  implemented in Consel 0.1j  were used for topology testing. We have decided not to report the results of approximately unbiased test  because we have realized that the test behaves very unstably for our data sets; re-testing of the same data sets produced very different p-values that sometimes differed in significance. Regarding the significance or non-significance at the p = 0.001 level, the results of the AU tests were in agreement with the results of KH and SH tests in most cases; however due to their instability, we have decided to report only the results of KH and SH tests.
A set of 503 topologies was created in order to test whether the relationships between MATX paralogues are in conflict with the relationship of MATX containing taxa as inferred from SSU rDNA sequences. This set of topologies contained the best topology inferred from an analysis of the MATX protein alignment by RAxML, 500 topologies from bootstrap permutations of the MATX alignment generated by RAxML, the best tree inferred by RAxML from the SSU rRNA alignment of the same set of taxa, and the manually constructed topology reflecting the current view of species relationships. The latter two topologies representing species trees are given in Additional file 1 and in Additional file 2: Figure S1 and Additional file 3: Figure S3. Site likelihoods for topologies 1–501 were inferred by Treepuzzle 5.2.  using MATX gene alignment, WAG + I + Γ model of amino acid substitution and parameter values inferred from the topology nr. 1. Site likelihoods for topologies 502 and 503 were inferred by Treepuzzle using MATX gene alignment, WAG + I + Γ model of amino acid substitution and parameter values inferred from these topologies. The sets of site likelihoods were then compared by the KH, weighted KH (WKH), SH and SH (WSH) test in Consel 0.1j . The tests were performed for (1) the full set of MATX paralogues from 40 taxa, (2) a set of MATX paralogues, excluding MATX from Aureococcus, Prymnesium, Dendroctonus and Lactuca, (3) data set 1 rooted by Trichomonas and Escherichia, (4) data set 2 rooted by Trichomonas and Escherichia, and (5) a set of MATX paralogues from euglenids.
The same tests were used to evaluate whether or not the root position between MAT and MATX paralogues can be rejected. For these tests, we used topology shown in Figure 1, 500 bootstrap topologies calculated from the same alignment, and a topology that differed from Figure 1 only in the position of prokaryotic outgroups that were moved on the branch separating MAT and MATX paralogues. The tests were performed as described above.
Availability of supporting data
All the supporting data are included as additional files.
The work on the project was supported by the project "BIOCEV – Biotechnology and Biomedicine Centre of the Academy of Sciences and Charles University" (CZ.1.05/1.1.00/02.0109), from the European Regional Development Fund and by the Czech Science Foundation (P506/11/1320) awarded to VH and by grants to BSL from the Tula Foundation (Centre for Microbial Diversity and Evolution at the University of British Columbia) and the Canadian Institute for Advanced Research, Program in Integrated Microbial Biodiversity. RET was supported by an Assembling the Tree of Life grant from the National Science Foundation (DEB-0830056). Many of the Euglenozoan MATX sequences were generated as part of this larger project. RET would like to thank his collaborators at Virginia Commonwealth University, Dr. Gregory A. Buck (PI on the grant), Dr. Vishal N. Kopardé and Dr. Andrey V. Matveyev for their roles in generating these sequences.
- Cantoni GL: Biological methylation: selected aspects. Annu Rev Biochem. 1975, 44: 435-451. 10.1146/annurev.bi.44.070175.002251.PubMedView ArticleGoogle Scholar
- Chiang PK, Gordon RK, Tal J, Zeng GC, Doctor BP, Pardhasaradhi K, McCann PP: S-adenosylmethionine and methylation. FASEB J. 1996, 10: 471-480.PubMedGoogle Scholar
- Takusagawa F, Kamitori S, Misaki S, Markham GD: Crystal structure of S-adenosylmethionine synthetase. J Biol Chem. 1996, 271: 136-147. 10.1074/jbc.271.1.136.PubMedView ArticleGoogle Scholar
- Graham DE, Bock CL, Schalk-Hihi C, Lu ZJ, Markham GD: Identification of a highly diverged class of S-adenosylmethionine synthetases in the archaea. J Biol Chem. 2000, 275: 4055-4059. 10.1074/jbc.275.6.4055.PubMedView ArticleGoogle Scholar
- Gonzalez B, Pajares MA, Hermoso JA, Alvarez L, Garrido F, Sufrin JR, Sanz- Aparicio J: The crystal structure of tetrameric methionine adenosyltransferase from rat liver reveals the methionine-binding site. J Mol Biol. 2000, 300: 363-375. 10.1006/jmbi.2000.3858.PubMedView ArticleGoogle Scholar
- Kotb M, Kredich NM: S-Adenosylmethionine synthetase from human lymphocytes purification and characterization. J Biol Chem. 1985, 260: 3923-3930.PubMedGoogle Scholar
- Markham GD, Pajares MA: Structure – function relationships in methionine Adenosyltransferases. Cell Mol Life Sci. 2009, 66: 636-648. 10.1007/s00018-008-8516-1.PubMedPubMed CentralView ArticleGoogle Scholar
- Sanchez-Perez GF, Hampl V, Simpson AGB, Roger AJ: A new divergent type of eukaryotic methionine adenosyltransferase is present in multiple distantly related secondary algal lineages. J Eukaryot Microbiol. 2008, 55: 374-381. 10.1111/j.1550-7408.2008.00349.x.PubMedView ArticleGoogle Scholar
- Garrido F, Estrela S, Alves C, Sánchez-Pérez GF, Sillero A, Pajares MA: Refolding and characterization of methionine adenosyltransferase from Euglena gracilis. Protein Expr Purif. 2011, 79: 128-136. 10.1016/j.pep.2011.05.004.PubMedView ArticleGoogle Scholar
- Kamikawa R, Sanchez-Perez GF, Sako Y, Roger AJ, Inagaki Y: Expanded phylogenies of canonical and non-canonical types of methionine adenosyltransferase reveal a complex history of these gene families in eukaryotes. Mol Phylogenet Evol. 2009, 53: 565-570. 10.1016/j.ympev.2009.06.016.PubMedView ArticleGoogle Scholar
- Keeling PJ, Inagaki Y: A class of eukaryotic GTPase with a punctate distribution suggesting multiple functional replacements of translation elongation factor 1α. Proc Natl Acad Sci U S A. 2004, 101: 15380-15385. 10.1073/pnas.0404505101.PubMedPubMed CentralView ArticleGoogle Scholar
- Noble GP, Rogers MB, Keeling PJ: Complex distribution of EFL and EF-1α proteins in the green algal lineage. BMC Evol Biol. 2007, 7: 82-10.1186/1471-2148-7-82.PubMedPubMed CentralView ArticleGoogle Scholar
- Kamikawa R, Inagaki Y, Sako Y: Direct phylogenetic evidence for lateral transfer of elongation factor-like gene. Proc Natl Acad Sci U S A. 2008, 105: 6965-6969. 10.1073/pnas.0711084105.PubMedPubMed CentralView ArticleGoogle Scholar
- Keeling PJ, Palmer JD: Horizontal gene transfer in eukaryotic evolution. Nat Rev Genet. 2008, 9: 605-618. 10.1038/nrg2386.PubMedView ArticleGoogle Scholar
- Gile GH, Faktorova D, Castlejohn CA, Burger G, Lang BF, Farmer MA, Lukes J, Keeling PJ: Distribution and phylogeny of EFL and EF-1α in Euglenozoa suggest ancestral co-occurrence followed by differential loss. PLoS One. 2009, 4: e5162-10.1371/journal.pone.0005162.PubMedPubMed CentralView ArticleGoogle Scholar
- Kamikawa R, Yabuki A, Nakayama T, Ishida K, Hashimota T, Inagaki Y: Cercozoa comprises both EF-1α-containing and EFL-containing members. Eur J Protistol. 2011, 47: 24-28. 10.1016/j.ejop.2010.08.002.PubMedView ArticleGoogle Scholar
- Ishitani Y, Kamikawa R, Yabuki A, Tsuchiya M, Inagaki Y, Takishita K: Evolution of elongation factor-like (EFL) protein in Rhizaria is revised by radiolarian EFL gene sequences. J Eukaryot Microbiol. 2012, 59: 367-373. 10.1111/j.1550-7408.2012.00626.x.PubMedView ArticleGoogle Scholar
- Szabova J, Ruzicka P, Verner Z, Hampl V, Lukes J: Experimental examination of EFL and MATX eukaryotic horizontal gene transfers: coexistence of mutually exclusive transcripts predates functional rescue. Mol Biol Evol. 2011, 28: 2371-2378. 10.1093/molbev/msr060.PubMedView ArticleGoogle Scholar
- Leander BS, Esson HJ, Breglia SA: Macroevolution of complex cytoskeletal systems in euglenids. Bioessays. 2007, 29: 987-1000. 10.1002/bies.20645.PubMedView ArticleGoogle Scholar
- Yamaguchi A, Yubuki N, Leander BS: Morphostasis in a novel eukaryote illuminates the evolutionary transition from phagotrophy to phototrophy: description of Rapaza viridis n. gen. et sp. (Euglenozoa, Euglenida). BMC Evol Biol. 2012, 12: 29-10.1186/1471-2148-12-29.PubMedPubMed CentralView ArticleGoogle Scholar
- Leander BS: Did trypanosomatid parasites have photosynthetic ancestors?. Trends Microbiol. 2004, 12: 251-258. 10.1016/j.tim.2004.04.001.PubMedView ArticleGoogle Scholar
- Takahashi F, Okabe Y, Nakada T: Origins of the secondary plastids of euglenophyta and chlorarachniophyta as revealed by an analysis of the plastid-targeting, nuclear-encoded gene psbO. J Phycol. 2007, 43: 1302-1309. 10.1111/j.1529-8817.2007.00411.x.View ArticleGoogle Scholar
- Gile GH, Novis PM, Cragg DS, Zuccarello GC, Keeling PJ: The distribution of Elongation Factor-1 Alpha (EF-1alpha), Elongation Factor-Like (EFL), and a non-canonical genetic code in the ulvophyceae: discrete genetic characters support a consistent phylogenetic framework. J Eukaryot Microbiol. 2009, 56: 367-72. 10.1111/j.1550-7408.2009.00411.x.PubMedView ArticleGoogle Scholar
- Medlin L, Elwood HJ, Stickel S, Sogin ML: The characterization of enzymatically amplified eukaryotes 16S like ribosomal RNA coding regions. Gene. 1988, 71: 491-499. 10.1016/0378-1119(88)90066-2.PubMedView ArticleGoogle Scholar
- Müllner AN, Angeler DG, Samuel R, Linton EW, Triemer RE: Phylogenetic analysis of phagotrophic, photomorphic and osmotrophic euglenoids by using the nuclear 18S rDNA sequence. Int J Syst Evol Microbiol. 2001, 51: 783-791. 10.1099/00207713-51-3-783.PubMedView ArticleGoogle Scholar
- Patel RK, Jain M: NGS QC toolkit: a toolkit for quality control of next generation sequencing data. PLoS ONE. 2012, 7: e30619-10.1371/journal.pone.0030619.PubMedPubMed CentralView ArticleGoogle Scholar
- Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG: The clustalx windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucl Acids Res. 1997, 24: 4876-4882.View ArticleGoogle Scholar
- Katoh K, Asimenos G, Toh H: Multiple alignment of DNA sequences with MAFFT. Methods Mol Biol. 2009, 537: 39-64. 10.1007/978-1-59745-251-9_3.PubMedView ArticleGoogle Scholar
- Hall TA: BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl Acids Symp Ser. 1999, 41: 95-98.Google Scholar
- Stamatakis A: RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics. 2006, 22: 2688-2690. 10.1093/bioinformatics/btl446.PubMedView ArticleGoogle Scholar
- Abascal F, Zardoya R, Posada D: ProtTest: selection of best-fit models of protein evolution. Bioinformatics. 2005, 21: 2104-2105. 10.1093/bioinformatics/bti263.PubMedView ArticleGoogle Scholar
- Kishino H, Hasegawa M: Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA sequence data, and the branching order in Hominoidea. J Mol Evol. 1989, 29: 170-179. 10.1007/BF02100115.PubMedView ArticleGoogle Scholar
- Shimodaira H, Hasegawa M: Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Mol Biol Evol. 1999, 16: 1114-1116. 10.1093/oxfordjournals.molbev.a026201.View ArticleGoogle Scholar
- Shimodaira H, Hasegawa M: CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics. 2001, 17: 1246-1247. 10.1093/bioinformatics/17.12.1246.PubMedView ArticleGoogle Scholar
- Shimodaira H: An approximately unbiased test of phylogenetic tree selection. Syst Biol. 2002, 51: 492-508. 10.1080/10635150290069913.PubMedView ArticleGoogle Scholar
- Schmidt HA, Strimmer K, Vingron M, Von Haeseler A: TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics. 2002, 18: 502-504. 10.1093/bioinformatics/18.3.502.PubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.