sign in sign up
<<
Home , Bodonida, Bodo (excavate), Bodo saltans, Cryptobia, Euglenid, Euglenozoa

Genes Genet. Syst. (2017) 92, p. 35–42 Phylogenomic analysis of Kinetoplastea 35 Global Kinetoplastea phylogeny inferred from a large-scale multigene alignment including parasitic species for better understanding transitions from a free-living to a parasitic lifestyle

Euki Yazaki1, Sohta A. Ishikawa2*, Keitaro Kume1,3, Akira Kumagai4, Takashi Kamaishi5, Goro Tanifuji6, Tetsuo Hashimoto1,7 and Yuji Inagaki1,7† 1Graduate School of and Environmental Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8572, Japan 2Department of Biological Sciences, Graduate School of Sciences, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan 3Graduate School of Systems and Information Engineering, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8573, Japan 4Miyagi Prefecture Fisheries Technology Institute, 97-6 Sodenohama, Watanoha, Ishinomaki, Miyagi 986-2135, Japan 5National Research Institute of Aquaculture, Fisheries Research Agency, 422-1 Nakatsuhamaura, Minami-Ise, Mie 516-0913, Japan 6Department of Zoology, National Museum of Nature and Science, 4-1-1 Amakubo, Tsukuba, Ibaraki 305-0005, Japan 7Center for Computational Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8577, Japan

(Received 12 October 2016, accepted 23 November 2016; J-STAGE Advance published date: 20 February 2017)

All members of the order known to date are parasites that are most likely descendants of a free-living ancestor. Trypanosomatids are an excel- lent model to assess the transition from a free-living to a parasitic lifestyle, because a large amount of experimental data has been accumulated for well-studied mem- bers that are harmful to and livestock ( spp. and spp.). However, recent advances in our understanding of the diversity of trypano- somatids and their close relatives (i.e., members of the class Kinetoplastea) have suggested that the change in lifestyle took place multiple times independently from that which gave rise to the extant trypanosomatid parasites. In the cur- rent study, transcriptomic data of two parasitic kinetoplastids belonging to orders other than Trypanosomatida, namely Azumiobodo hoyamushi (Neobodonida) and Trypanoplasma borreli (Parabodonida), were generated. We re-examined the transition from a free-living to a parasitic lifestyle in the of kinetoplas- tids by combining (i) the relationship among the five orders in Kinetoplastea and (ii) that among free-living and parasitic species within the individual orders. The former relationship was inferred from a large-scale multigene alignment including the newly generated data from Azumiobodo and Trypanoplasma, as well as the data from another parasitic kinetoplastid, sp., deposited in GenBank; and the latter was inferred from a taxon-rich small subunit ribosomal DNA align- ment. Finally, we discuss the potential value of parasitic kinetoplastids identi- fied in Parabodonida and Neobodonida for studying the evolutionary process that turned a free-living species into a parasite.

Key words: Azumiobodo hoyamushi, Kinetoplastea, multigene phylogeny, para- sites, Trypanoplasma borreli

Edited by Kyoichi Sawamura INTRODUCTION * Corresponding author. E-mail: [email protected] † Corresponding author. E-mail: [email protected] Trypanosomatid flagellates have been studied exten- DOI: http://doi.org/10.1266/ggs.16-00056 sively, as some of them are causative agents of 36 E. YAZAKI et al.

African (sleeping sickness), Chagas dis- agent of ichthyobodosis, Ichthyobodo necator: Callahan ease, and . Besides their clinical impor- et al., 2002) and intracellular parasites in the amoe- tance, these flagellates possess intriguing properties that bozoan Paramoeba pemaquidensis (i.e., Perkinsela sp. are shared by only a few or no other . Try- or Ichthyobodo-related organism: Dyková et al., 2003; panosomatids are known to possess a unique - Caraguel et al., 2007; Dyková et al., 2008; Feehan et al., derived , the , that encloses glycolytic 2013; Lukeš et al., 2014). Parabodonida includes fish enzymes (Opperdoes and Borst, 1977; Gualdrón-López parasites that cause cryptobiosis in salmonid and cypri- et al., 2012). Mitochondria of trypanosomatids contain nid fishes (e.g., salmositica and Trypanoplasma a complex network of two types of circular DNA mol- borreli: Woo and Poynton, 1995), as well as the snail ecules, maxicircles and minicircles, and their mitochon- parasite C. helicis (Leidy, 1846). Among the known neo- drial mRNAs undergo intricate and distinctive editing bodonids, there is a single parasitic member, Azumiobodo prior to translation (Lukeš et al., 2002, 2005). The 5′ hoyamushi, which infects ascidians and causes soft tunic termini of mRNAs from trypanosomatid nuclear genomes syndrome (Hirose et al., 2012). As the transition from a also undergo post-transcriptional modification (Campbell free-living to a parasitic lifestyle occurred after the diver- et al., 2003; Michaeli, 2011). Although the properties gence of the extant parabodonids/neobodonids, the para- described above are also observed in phylogenetic rela- sites in Parabodonida and Neobodonida are potentially tives of trypanosomatids (i.e., members of other orders in useful to retrace the evolutionary path from a free-living the class Kinetoplastea; see below), trypanosomatids, for to a parasitic lifestyle. which various experimental techniques in molecular and For a deeper understanding of the evolution of parasit- cell biology (e.g., genetic modification) are available, have ism in Kinetoplastea, a well-resolved, taxon-rich phylog- been the center of research on Kinetoplastea. eny is indispensable. Deschamps et al. (2011) analyzed Trypanosomatida, together with Eubodonida, Parabo- an alignment of 64 proteins and elucidated the relation- doida, Neobodonida and Prokinetoplastida, comprise the ship among Trypanosomatida, Eubodonida, Neobodonida class Kinetoplastea (Moreira et al., 2004; Simpson et al., and Parabodonida. However, their analyses contained 2006). All known members of Trypanosomatida and of two potential limitations. First, the alignment analyzed Prokinetoplastida are parasites (Simpson et al., 2006; in Deschamps et al. (2011) contained no prokinetoplastid Lukeš et al., 2014). However, the remaining three orders species. Second, each of Parabodonida and Neobodonida are dominated by free-living members, and only a few or was represented by only a single free-living species but none of the members are known to be parasitic. Pre- no parasitic member. In this study, we overcame these viously published phylogenies of small subunit ribo- limitations by analyzing a new alignment of 43 pro- somal DNA (SSU rDNA) sequences have constantly and teins (43-gene alignment), which covered all five orders robustly united Neobodonida, Parabodonida, Eubodon- in Kinetoplastea, and Parabodonida and Neobodonida ida and Trypanosomatida, excluding Prokinetoplas- were represented by both free-living and parasitic mem- tida (Simpson et al., 2002; Moreira et al., 2004; von der bers. Combining the global Kinetoplastea phylogeny, Heyden et al., 2004). This tree topology suggested that updated by analyzing the 43-gene alignment, with a Trypanosomatida and Prokinetoplastida acquired para- taxon-rich SSU rDNA phylogeny, we discuss the transi- sitic lifestyles separately (Moreira et al., 2004; Simpson tion from a free-living to a parasitic lifestyle in the evolu- et al., 2006; Lukeš et al., 2014). Owing to their impor- tion of Kinetoplastea. tance in public health, the origin of in the extant trypanosomatids is one of the major questions in MATERIALS AND METHODS the evolution of Kinetoplastea. To address this question, the precise relationship among Neobodonida, Parabodon- Cultures, RNA extraction and sequencing The lab- ida, Eubodonida and Trypanosomatida has been explored oratory culture of A. hoyamushi established by Hirose et mainly by analyzing SSU rRNA genes or genes encoding al. (2012) was grown and maintained in sea water con- highly conserved proteins, but has not been resolved with taining 2% heat-inactivated fetal bovine serum (Gibco, high statistical support (Dolezel et al., 2000; Simpson et Thermo Fisher Scientific, Waltham, Massachusetts, USA) al., 2002; Moreira et al., 2004; Simpson et al., 2004; von at 17 °C. Trypanoplasma borreli ATCC50836 was pur- der Heyden et al., 2004; Deschamps et al., 2011). A phy- chased from the American Type Culture Collection, and logenetic analysis of 64 genes encoding highly conserved grown in live-infusion-tryptose medium (Fernandes and proteins successfully designated Eubodonida as the clos- Castellani, 1966) at 17 °C. Total RNA was extracted est relative of Trypanosomatida (Deschamps et al., 2011). from the harvested cells using Trizol (Thermo Fisher Compared to pathogenic trypanosomatids, other Scientific, Waltham, Massachusetts, USA), following the parasitic members in Kinetoplastea have received less manufacturer’s protocol. Construction of a cDNA library research attention. Two types of parasites belong to and subsequent sequencing by the Illumina HiSeq2500 Prokinetoplastida: fish parasites (e.g., the causative system were performed at Hokkaido System Science Phylogenomic analysis of Kinetoplastea 37

(Sapporo, Hokkaido, Japan). We generated 401,725,240 anticipated that such sequences would probably group and 433,374,224 paired-end, 100-base reads from the with the and Dictyostelium sequences, with the Azumiobodo and Trypanoplasma libraries, respectively specific affinity to the latter (i.e., amoebozoan) sequences, (deposited in NCBI Sequence Read Archive under acces- in the phylogenetic analyses described below. sion numbers SRX2210809 and SRX22109115, respec- Sixty-four single-gene alignments (each containing the tively). The two sets of initial reads were separately 14 species described above) were automatically generated assembled into 28,134 contigs (Azumiobodo) and 18,591 by MAFFT 7.205 (Katoh et al., 2002; Katoh and Standley, contigs (Trypanoplasma) by Trinity (Grabherr et al., 2013). Ambiguously aligned positions were manually 2011; Haas et al., 2013). Transcriptomic data generated excluded prior to the phylogenetic analyses described from Perkinsela by the Illumina HiSeq2000 system were below. The alignments were subjected separately to retrieved from NCBI Sequence Read Archive (accession maximum-likelihood (ML) phylogenetic analyses with number ARX255943), and assembled into 18,600 contigs the LG model (Le and Gascuel, 2008), incorporating by Trinity. The contig data of Azumiobodo and Trypano- empirical amino acid frequencies and among-site rate plasma are freely available at https://sites.google.com/ variation approximated by a discrete gamma distribu- site/eukiyazaki/home/data-archive/. tion with four categories (LG + Γ + F model). The tree search was started from ten maximum-parsimony (MP) Phylogenetic alignments We prepared a multi- trees, each of which was generated by random stepwise gene alignment including sequences from Azumiobodo, addition of sequences by RAXML 8.0.3 (Stamatakis, Trypanoplasma and Perkinsela by following Deschamps 2014). Although not shown here, the ML trees inferred et al. (2011), which assessed the phylogenetic relationship from 21 of the 64 single-gene alignments failed to recon- among seven kinetoplastids based on 64 protein-coding struct the monophyly of kinetoplastids, and were omit- genes. First, we prepared four sets of the 64 genes— ted from the multigene alignment described below. The those of , T. cruzi, remaining 43 single-gene alignments listed in Supple- and L. infantum—by surveying in NCBI (http://www.ncbi. mentary Table S1 were then concatenated into a single nlm.nih.gov/), TriTrypDB (http://tritrypdb.org/tritrypdb), alignment containing 6,842 amino acid positions (43-gene and Sanger Institute databases (http://www.sanger. alignment). Prior to phylogenetic analyses, we excluded ac.uk/resources/databases/). The putative Trypanosoma/ both Dictyostelium and Naegleria sequences, which were Leishmania proteins identified in the first survey were used as “probes” to detect Paramoeba sequences poten- then used as queries for TBLASTX against the contig tially contaminating the Perkinsela transcriptomic data. data of Azumiobodo, Trypanoplasma and Perkinsela. We We sampled and aligned SSU rDNA sequences from recovered Azumiobodo, Trypanoplasma and Perkinsela 20 trypanosomatids, 10 parabodonids, 10 eubodonids, 15 transcripts that matched query sequences with E-values neobodonids, five prokinetoplastids, three diplomenids below 10 −10 as candidates encoding the proteins of and a single (64 taxa in total). The 64 SSU interest. The candidate transcripts were subjected to rDNA sequences were aligned by MAFFT, and ambigu- BLASTX against the NCBI nr protein database to con- ously aligned positions were manually excluded. The firm that these transcripts encode proteins that were final SSU rDNA alignment contained 1,179 nucleotide homologous to the queries in the initial BLAST analysis positions. (TBLASTX). Finally, the selected transcripts of Azumio- Both 43-gene and SSU rDNA alignments are avail- , Trypanoplasma and Perkinsela were conceptually able at https://sites.google.com/site/eukiyazaki/home/data- translated into amino acid sequences by EMBOSS Tran- archive/kinetoplastida. seq (http://www.ebi.ac.uk/Tools/st/emboss_transeq/). We repeated the same procedure above for three kinetoplas- Phylogenetic analyses The 43-gene and SSU rDNA tids (Rhynchomonas nasuta, Procryptobia sorokini and alignments were subjected to ML phylogenetic analyses : GenBank accession numbers HO651677– by RAXML 8.0.3. The detailed settings of tree searching HO651928), the amoebozoan Dictyostelium discoideum were the same as those used in the single-gene analy- (NCBI BioProject: PRJNA201), the heterolobosean ses described above, but we assigned the LG + Γ + F (NCBI BioProject: PRJNA43691), the model for ML analyses of the 43-gene alignment, and the diplonemid papillatum (TBestDB organism GTR + Γ model (Rodríguez et al., 1990) for those of the ID: DP) and the euglenid gracilis (NCBI SRA: SSU rDNA alignment. One hundred bootstrap repli- ERX324117). Dictyostelium and Naegleria were included cates were generated from each alignment, and subjected in the alignments to identify sequences originated from to tree searching as described above. The resultant boot- the amoebozoan Paramoeba pemaquidensis (i.e., the host strap trees were used to calculate ML bootstrap support organism of Perkinsela), which potentially contaminate values (MLBPs). the Perkinsela transcriptomic data. If the Paramoeba The two alignments were also analyzed with Bayesian sequences were misassigned as Perkinsela sequences, we method. The 43-gene alignment was subjected to Phy- 38 E. YAZAKI et al. loBayes 3.3 (Lartillot and Philippe, 2004; Lartillot et al., the Parabodonida clade comprising Procryptobia soro- 2009) with the CAT + Poisson model. Two Markov chain kini and Trypanoplasma borreli, (iii) the Neobodonida Monte Carlo (MCMC) runs were conducted for 10,000 clade comprising Rhynchomonas nasuta and Azumiobodo cycles with burn-in of 2,500 (the maxdiff value were hoyamushi, and (iv) Perkinsela sp. representing Proki- 0.00013). Subsequently, the consensus tree with branch netoplastida. All the internal nodes in the ingroup were lengths and Bayesian posterior probabilities (BPPs) was supported by MLBPs of 100% and BPPs of 1.0, except the calculated from the remaining trees. The SSU rDNA clade of Rhynchomonas and Azumiobodo, which received alignment was subjected to MrBayes 3.2.3 (Huelsen- an MLBP of 97% and a BPP of 0.99 (Fig. 1). Prior to beck and Ronquist, 2001; Huelsenbeck and Ronquist, this study, the phylogenetic position of Prokinetoplastida 2004). We assigned the same substitution model as in has been assessed only by single-gene alignments includ- the ML method described above. The MCMC run was ing kinetoplastid species sampled from all five orders performed with one cold and three heated chains with (Callahan et al., 2002; Moreira et al., 2004; Simpson et default chain temperatures. We ran 1,000,000 genera- al., 2004; Breglia et al., 2007; Hirose et al., 2012) or by tions, and sampled log-likelihood scores and trees with four-gene and 11-gene alignments with restricted taxon branch lengths every 1,000 generations. The first 25% samplings (Tanifuji et al., 2011; Cenci et al., 2016). Thus, generations were discarded as burn-in. The consensus the current study provides the first phylogenomic support tree with branch lengths and BPPs were calculated from for the earliest-branching status of Prokinetoplastida in the remaining trees. Kinetoplastea (Fig. 1).

Evolution of parasitism in Kinetoplastea We ana- RESULTS AND DISCUSSION lyzed an SSU rDNA alignment, in which taxon sampling Global Kinetoplastea phylogeny revisited We was much richer than that of the 43-gene alignment, to updated the large-scale multigene (phylogenomic) illustrate the sporadic distribution of parasitic and free- alignment used in Deschamps et al. (2011) by adding living species in Kinetoplastea (Fig. 2). Prokinetoplas- Trypanoplasma, Azumiobodo and Perkinsela, and re- tida was excluded from the clade of Trypanosomatida, examined the global phylogeny of Kinetoplastea. The Eubodonida, Parabodonida and Neobodonida with an ML tree and MLBPs inferred from a 43-gene alignment MPBP of 100% and a BPP of 1.0. The monophylies of are presented in Fig. 1. As the topology inferred from Eubodonida, Parabodonida and Neobodonida were recov- Bayesian method was identical to that from the ML ered with MLBPs of 63–91% and BPPs of 0.71–1.0, while method, we only mapped BPPs on the ML tree shown that of Trypanosomatida was not positively supported. in Fig. 1. Although the 43-gene phylogeny includes Although Paratrypanosoma confusum was excluded from three parasitic kinetoplastids that were absent in the the clade of other trypanosomatids in the SSU rDNA previous phylogenomic analysis, the overall tree topol- phylogeny (Fig. 2), we regard Paratrypanosoma as an ogy (Fig. 1) was similar to that presented by Deschamps ancestral branch in Trypanosomatida based on a series of et al. (2011). Trypanosomatida was found to be tied multigene phylogenetic analyses presented in Flegontov with the other four taxons/clades in the following et al. (2013). We assume that the SSU rDNA phylogeny order: (i) Bodo saltans representing Eubodonida, (ii) presented here failed to place Paratrypanosoma in the correct position due to lack of phylogenetic signal. Thus, the combination of the 43-gene phylogeny, which resolved

Trypanosoma cruzi (0%) the backbone of the tree of Kinetoplastea (Fig. 1), and Trypanosoma brucei (3%) Trypanosomatida the taxon-rich SSU rDNA phylogeny (Fig. 2) enables us (0%) Leishmania major (0%) to depict how and when parasitic species emerged during Bodo saltans (12%) Eubodonida the evolution of Kinetoplastea. We schematically illus- Procryptobia sorokini (35%) Parabodonida trate the evolution of lifestyles in Kinetoplastea in Fig. Trypanoplasma borreli (0%) 97 Rhynchomonas nasuta (1%) 3 (see below for details). As the life cycles of most “free- Neobodonida 0.99 Azumiobodo hoyamushi (37%) living” kinetoplastids are not well understood, we cannot Perkinsela sp. (18%) Prokinetoplastida exclude the possibility that some of them have parasitic Diplonema papillatum (63%) stages. (27%) 0.05 substitutions/site The 43-gene phylogeny united the obligatory para- Fig. 1. Global Kinetoplastea phylogeny inferred from an sitic order Trypanosomatida with Eubodonida (Fig. 1), alignment comprising 43 genes encoding highly conserved pro- which comprises free-living members (see Fig. 2 for the teins. Tree topology was inferred from the maximum-likelihood details). Thus, as discussed in Simpson et al. (2006), (ML) method. Nodes marked by dots were supported by ML bootstrap values of 100% and Bayesian posterior probabilities of Deschamps et al. (2011) and Flegontov et al. (2013), a par- 1.0. For each taxon, the percentage of missing data is presented asitic lifestyle was most likely established after the sepa- in parentheses. Parasites are marked by diamonds. ration of Trypanosomatida and Eubodonida, but before Phylogenomic analysis of Kinetoplastea 39

Neobodo saliens (AF174379) designis (AY425016) Neobodo designis DH (AF464896) Kinetoplastid LFS2 (AF174380) 79 Azumiobodo hoyamushi (AB636162) 0.98 Cruzella marina (AF208878) Acruariola framvarensis (AY963571) Cryptaulaxoides-like sp. TCS-2003 (AY425021) Neobodonida 63 Dimastigella trypaniformis (X76495) 0.71 Dimastigella mimosa (AF208882) Dimastigella trypaniformis (X76494) Rhynchomonas nasuta (AY425025) Cryptaulax sp. ATCC50746 (AY425022) Rhynchobodo sp. Sourhope (AY490212) Rhynchobodo sp. ATCC50359 (U67183) 37

<0.50 Bodo edax (AY028451) Bodo uncinatus (AF208884) Bodo saltans (AY490226) Bodo saltans Petersburg (AF208887) Bodo saltans (AY490224) Bodo saltans (AY490231) Eubodonida 91 Bodo saltans (AY490232) 43 1.0 Bodo saltans (AY490230) <0.50 Bodo saltans (AY490233) Bodo saltans (AY490234) Procryptobia sorokini (DQ207592) Procryptobia sorokini ATCC50641 (AY425018) Procryptobia sorokini HFCC98 (DQ207593) 41 Parabodo nitrophilus (AY425019) Parabodonida 96 <0.50 Parabodo caudatus (X53910) 88 1.0 Cryptobia helicis (AF208880) 0.98 Cryptobia bullocki (AF080224) Cryptobia catostomi (AF080226) Trypanoplasma borreli Pg-JH (L14840) Cryptobia salmositica (AF080225) Paratrypanosoma confusum CUL13 (KF963538) Leishmania amazonensis (JX030088) Leishmania infantum JPCM5 (XR001203206) Leishmania major (GQ332361) sp. (X53914) esmeraldas (KT944309) nordicus PhN (KT223609) Phytomonas sp. (KX257483) Herpetomonas ztiplika (HG425174) Herpetomonas tarakana OSR-27 (KR868691) Trypanosomatida 67 Strigomonas oncopelti (AF038025) Jaenimonas drosophilae Fi-01.02 (KP260534) sp. (KX138601) Blastocrithidia triatomae (KX138599) 0.95 VINCH89 (AJ009149) Trypanosoma varani V54 (AJ005279) Trypanosoma cyclops (AJ131958) CAM22b (AJ009159) Trypanosoma brucei (M12676) Trypanosoma pestanai LEM110 (AJ009159) Ichthyobodo sp. GM0503 (AY255800) Ichthyobodo sp. AAN2003(AY228872) 99 Ichthyobodo necator (AY224691) Prokinetoplastida 1.0 Perkinsiella-like sp. (HQ132932) Perkinsela sp. (DQ167481) sp. ATCC50226 (AY425013) Diplonema papillatum ATCC50162 (KF633466) Hemistasia phaeocysticola (AB948221) Euglena gracilis (AF283309)

0.1 substitutions/site 0.4 substitutions/site Fig. 2. Global Kinetoplastea phylogeny inferred from an alignment of small subunit ribosomal DNA sequences. Tree topology was inferred from the maximum-likelihood (ML) method. Nodes marked by dots were supported by ML bootstrap values (MLBPs) of 100% and Bayesian posterior probabilities (BPPs) of 1.0. MLBPs and BPPs are presented only for the nodes that are critical for discussing the evolution of lifestyle in Kinetoplastea. Para- sites are marked by diamonds. the divergence of the extant trypanosomatids including free-living members (Fig. 2). In the SSU rDNA phy- Paratrypanosoma (Fig. 3). All trypanosomatids known logeny (Fig. 2), T. borreli, C. catostomi, C. bullocki and so far are extracellular parasites, but an intracellular C. salmositica formed a robust clade (MLBP of 100% stage has been also reported for Leishmania spp. and and BPP of 1.0). The four parabodonids are com- Trypanosoma cruzi (Tyler and Engman, 2001; Handman monly found in the blood stream, although ectoparasitic and Bullen, 2002). As Leishmania spp. and T. cruzi are forms have also been reported for C. bullocki and C. distantly related in the SSU rDNA phylogeny (Fig. 2), salmositica (Bower and Margolis, 1983; Woo and Wehnert, the ability to invade host was likely acquired 1983). We can conclude that an extracellular parasitic by the two separate lineages after the divergence of try- lifestyle was established on the branch leading to the panosmatids (Fig. 3). Trypanoplasma-Cryptobia clade, as proposed in Simpson Parabodonida contains parasitic members, namely et al. (2006). However, we currently have no evidence to Trypanoplasma borreli and Cryptobia spp., as well as determine whether the ectoparasitic form is the ances- 40 E. YAZAKI et al.

Fig. 2). The affinity between Azumiobodo and Cruzella prompts us to propose serial lifestyle changes in Neo- Trypanosomatida , as follows: the common ancestor of Azumiobodo and Cruzella established a symbiotic relationship with ascidians, and, after separation of the two species, the for- Eubodonida mer became a pathogenic extracellular parasite causing soft tunic syndrome. Although the above scenario needs to be examined in future studies, we are confident that the transition from a free-living to a parasitic lifestyle Parabodonida occurred once after the divergence of the extant neobodo- nids (Fig. 3). Independent from the lifestyle changes discussed above, the transition from a free-living to a parasitic Neobodonida lifestyle was proposed for Prokinetoplastida (Simpson et al., 2006). As all the members belonging to Proki- netoplastida known to date are parasitic, it is straight- forward to assume that they emerged from a parasitic Prokinetoplastida ancestor (Fig. 3). In the SSU rDNA phylogeny (Fig. 2), prokinetoplastids were split into two subclades; one is Fig. 3. Transition from a free-living to a parasitic lifestyle in of parasites that infect the skin, fins and gills of fishes Kinetoplastea. The branching order among Trypanosomatida, (e.g., Ichthyobodo necator) and the other is of intracellu- Eubodonida, Parabodonida, Neobodonida and Prokinetoplastida lar parasites of amoebozoans (e.g., Perkinsela sp.). The is based on Fig. 1. Blue branches/clades indicate free-living conspicuous difference in parasitic mode between the species, while red clades indicate parasitic species. Putative two subclades in Prokinetoplastida—ectoparasitism changes in lifestyle are marked by stars. Due to the lack of a precise phylogenetic position, the parasitic parabodonid Jarrellia and intracellular parasitism—demands a more complex atramenti is omitted from this figure. Orange triangles indicate scenario than that assuming a single transition from the lineages that acquired the ability to invade host cytoplasm, a free-living to a parasitic lifestyle prior to the diver- namely Leishmania spp. and Trypanosoma cruzi in Trypanoso- gence of prokinetoplastids (Fig. 3). Notably, the hosts of matida and the intracellular parasites of marine amoebozoans Perkinsela sp. and its relatives are amoebozoans, which in Prokinetoplastida. parasitize marine (Munday et al., 2001; Young et al., 2007). Thus, it is attractive to hypothesize that tral trait of this clade. Cryptobia helicis (Leidy, 1846), the common ancestor of Perkinsela sp. and its relatives which was found in the seminal receptacle of snails, was an ectoparasite of a marine , then switched its most likely acquired an extracellular parasitic lifestyle host to an amoebozoan parasitizing marine animals, and independent from the Trypanoplasma-Cryptobia clade, finally invaded and settled in the cytoplasm of the amoe- as the SSU rDNA phylogeny united C. helicis with free- bozoan host. The above scenario will be favored if future living Parabodo nitrophilus and P. caudaus (MLBP of surveys find a novel ectoparasitic species that branches 96% and BPP of 1.0; Fig. 2). Altogether, we propose that at the base of the clade of Perkinsela sp. and its rela- parasitism emerged at least twice in Parabodonida (Fig. tives. It is also important to pursue the possibility that 3). The potential third parasitic lineage in Parabodonida some free-living prokinetoplastids have been overlooked is Jarrellia atramenti, found in the mucus of the respi- in natural environments, as environmental sequence data ratory tract of the pygmy sperm whale (Poynton et al., hinted that the full diversity of Prokinetoplastida has yet 2001). As only morphological and no molecular data are to be unveiled (Moreira et al., 2004; von der Heyden et al., available for this species, it will be necessary to assess 2004). We will certainly need to reevaluate how parasit- the phylogenetic relationship between Jarrellia and other ism was established in Prokinetoplastida after its organ- parabodonids to better understand the evolution of para- ismal diversity is sufficiently depicted in the future. sitism in this order. Among the diversity of neobodonids, Azumiobodo is Future perspectives The phylogenetic relationship the sole parasitic member known to date (Hirose et al., among the five orders in Kinetoplastea was found to be 2012). This flagellate was found in the tunics of ascid- largely unchanged before and after incorporating three ians with soft tunic syndrome by histopathology (Kumagai parasitic kinetoplastids, Azumiobodo, Trypanoplasma and et al., 2010). Intriguingly, Azumiobodo appeared to bear Perkinsela, into a phylogenomic alignment (Deschamps et a phylogenetic affinity to a commensal in the ascidian al., 2011; this study). However, we still need to improve intestine, Cruzella marina (Frolov and Malysheva, 2002), taxon sampling in phylogenomic alignments to re-examine in the SSU rDNA analysis (MLBP of 79% and BPP of 0.98; the global Kinetoplastea phylogeny in the future. For Phylogenomic analysis of Kinetoplastea 41 instance, the monophyly of Neobodonida received only Callahan, H. A., Litaker, R. W., and Noga, E. J. (2002) Molecular weak statistical support in the SSU rDNA analysis (Fig. of the suborder Bodonina (Order Kinetoplastida), including the important fish parasite, Ichthyobodo necator. 2). Thus, future phylogenomic analyses including addi- J. Eukaryot. Microbiol. 49, 119–128. tional neobodonids are required to strengthen their mono- Campbell, D. A., Thomas, S., and Sturm, N. R. (2003) Transcrip- phyly inferred from the SSU rDNA alignment. Another tion in kinetoplastid : why be normal? Microbes potential concern is the diversity of Eubodonida. It is Infect. 5, 1231–1240. not clear whether the eubodonids identified so far rep- Caraguel, C. G. B., O’kelly, C. J., Legendre, P., Frasca, S. Jr., Gast, R. J., Desrés, B. M., Cawthorn, R. J., and Greenwood, resent the true diversity of this order, as the SSU rDNA S. J. (2007) Microheterogeneity and coevolution: an exami- phylogeny (Fig. 2) implies that the diversity of Eubodon- nation of rDNA sequence characteristics in ida is considerably lower than that of any other order in pemaquidensis and its prokinetoplastid . J. Kinetoplastea. Thus, future studies may identify novel Eukaryot. Microbiol. 54, 418–426. kinetoplastid flagellates that bear phylogenetic affinities Cenci, U., Moog, D., Curtis, B. A., Tanifuji, G., Eme, L., Lukeš, J., and Archibald, J. M. (2016) Heme pathway evolution in to the currently known eubodonids. If such kinetoplas- kinetoplastid . BMC Evol. Biol. 16, 109. tid flagellates exist, it would be worthwhile to incorporate Deschamps, P., Lara, E., Marande, W., López-García, P., Ekelund, them into phylogenomic analyses to reflect the proper F., and Moreira, D. (2011) Phylogenomic analysis of kineto- diversity of Eubodonida. supports that trypanosomatids arose from within Because of the serious threats to public health they bodonids. Mol. Biol. Evol. 28, 53–58. Dolezel, D., Jirk , M., Maslov, D. A., and Lukeš, J. (2000) Phy- pose, members of Trypanosomatida have attracted more ů logeny of the bodonid flagellates (Kinetoplastida) based on research attention than other kinetoplastids. From an small-subunit rRNA gene sequences. Int. J. Syst. Evol. evolutionary biological perspective, trypanosomatids Microbiol. 50, 1943–1951. are regarded as model organisms for studying how and Dyková, I., Fiala, I., Lom, J., and Lukeš, J. (2003) Perkinsiella when kinetoplastid flagellates acquired a parasitic life- amoebae-like of Neoparamoeba spp., rela- tives of the kinetoplastid Ichthyobodo. Eur. J. Protistol. style and pathogenicity (Deschamps et al., 2011; Lukeš 39, 37–52. et al., 2014). Nevertheless, we also anticipate that Dyková, I., Fiala, I., and Pecková, H. (2008) Neoparamoeba spp. Parabodonida and Neobodonida will provide insights and their eukaryotic endosymbionts similar to Perkinsela into the mechanism involved in the transformation of a amoebae (Hollande, 1980): coevolution demonstrated by free-living species into a parasite. In Parabodonida, the SSU rRNA gene phylogenies. Eur. J. Protistol. 44, 269–277. snail parasite Cryptobia helicis showed a specific affin- Feehan, C. J., Johnson-Mackinnon, J., Scheibling, R. E., Lauzon- Guay, J. S., and Simpson, A. G. B. (2013) Validating the iden- ity to the free-living members Parabodo caudatus and P. tity of Paramoeba invadens, the causative agent of recurrent nitrophilus (Fig. 2). This change in lifestyle took place mass mortality of sea urchins in Nova Scotia, Canada. Dis. after the divergence of parabodonids (Fig. 3), and we may Aquat. Organ. 103, 209–227. have a chance to pinpoint the set of genes that played a Fernandes, J. F., and Castellani, O. (1966) Growth characteris- pivotal role in the change by comparing the genomic and tics and chemical composition of Trypanosoma cruzi. Exp. Parasitol. 18, 195–202. transcriptomic data of the parasite and its closest free- Flegontov, P., Votýpka, J., Skalický, T., Logacheva, M. D., Penin, living relatives. Similarly, the comparison between two A. A., Tanifuji, G., Onodera, N. T., Kondrashov, A. S., Volf, P., closely related neobodonids, Azumiobodo and Cruzella, a Archibald, J. M., et al. (2013) Paratrypanosoma is a novel parasite and a commensal of ascidians, respectively, may early-branching trypanosomatid. Curr. Biol. 23, 1787– provide insights into the origin of parasitism and patho- 1793. Frolov, A. O., and Malysheva, M. N. (2002) Ultrastructure of the genicity. flagellate Cruzella marina (Kinetoplastidea). Tsitologiia 44, 477–484. S. A. I. was supported by research fellowships from the Japan Grabherr, M. G., Haas, B. J., Yassour, M., Levin, J. Z., Thompson, Society for the Promotion of Science (JSPS) for Young Scientists D. A., Amit, I., Adiconis, X., Fan, L., Raychowdhury, R., Zeng, (Nos. 2400007 and 2802725). This work was supported in part Q., et al. (2011) Full-length transcriptome assembly from by grants from JSPS awarded to Y. I. (23117006 and 16H04826) RNA-Seq data without a reference genome. Nat. Biotechnol. and T. H. (23117005 and 15H05231). The phylogenetic anal- 29, 644–652. yses conducted in this work have been carried out under the Gualdrón-López, M., Brennand, A., Hannaert, V., Quiñones, W., “Interdisciplinary Computational Science Program” in the Cen- Cáceres, A. J., Bringaud, F., Concepción, J. L., and Michels, ter for Computational Sciences, University of Tsukuba. P. A. M. (2012) When, how and why glycolysis became com- partmentalised in the Kinetoplastea. A new look at an REFERENCES ancient organelle. Int. J. Parasitol. 42, 1–20. Haas, B. J., Papanicolaou, A., Yassour, M., Grabherr, M., Blood, P. Bower, S. M., and Margolis, L. (1983) Direct transmission of D., Bowden, J., Couger, M. B., Eccles, D., Li, B., Lieber, M., et the haemoflagellate Cryptobia salmositica among Pacific al. (2013) De novo transcript sequence reconstruction from salmon (Oncorhynchus spp.). Can. J. Zool. 61, 1242–1250. RNA-seq using the Trinity platform for reference genera- Breglia, S. A., Slamovits, C. H., and Leander, B. S. (2007) Phylog- tion and analysis. Nat. Protoc. 8, 1494–1512. eny of phagotrophic () as inferred from Handman, E., and Bullen, D. V. R. (2002) Interaction of hsp90 gene sequences. J. Eukaryot. Microbiol. 54, 86–92. Leishmania with the host macrophage. Trends Parasitol. 42 E. YAZAKI et al.

18, 332–334. of marine fish caused by infection with Neoparamoeba Hirose, E., Nozawa, A., Kumagai, A., and Kitamura, S. (2012) pemaquidensis. J. Dis. 24, 497–507. Azumiobodo hoyamushi gen. nov. et sp. nov. (Euglenozoa, Opperdoes, F. R., and Borst, P. (1977) Localization of nine glyco- Kinetoplastea, Neobodonida): a pathogenic kinetoplas- lytic enzymes in a microbody-like organelle in Trypanosoma tid causing the soft tunic syndrome in ascidian aquacul- brucei: the glycosome. FEBS Lett. 80, 360–364. ture. Dis. Aquat. Organ. 97, 227–235. Poynton, S. L., Whitaker, B. R., and Heinrich, A. B. (2001) A Huelsenbeck, J. P., and Ronquist, F. (2001) MRBAYES: Bayes- novel trypanoplasm-like flagellate Jarrellia atramenti n. ian inference of phylogenetic trees. Bioinformatics 17, g., n. sp. (Kinetoplastida: Bodonidae) and from the 754–755. blowhole of a stranded pygmy sperm whale Kogia breviceps Huelsenbeck, J. P., and Ronquist, F. (2004) MrBays 3: Bayesian (Physeteridae): morphology, life cycle and potential pathoge- phylogenetic inference under mixed models. Bioinformat- nicity. Dis. Aquat. Organ. 44, 191–201. ics 19, 1572–1574. Rodríguez, F., Oliver, J. L., Marin, A., and Medina, J. R. (1990) Katoh, K., Misawa, K., Kuma, K., and Miyata, T. (2002) MAFFT: The general stochastic model of nucleotide substitution. J. a novel method for rapid multiple sequence alignment Theor. Biol, 142, 485–501. based on fast Fourier transform. Nucleic Acids Res. 30, Simpson, A. G. B., Gill, E. E., Callahan, H. A., Litaker, R. W., 3059–3066. and Roger, A. J. (2004) Early evolution within kinetoplas- Katoh, K., and Standley, D. M. (2013) MAFFT multiple sequence tids (Euglenozoa), and the late emergence of trypanosoma- alignment software version 7: improvements in performance tids. 155, 407–422. and usability. Mol. Biol. Evol. 30, 772–780. Simpson, A. G. B., Lukeš, J., and Roger, A. J. (2002) The evolution- Kumagai, A., Suto, A., Ito, H., Tanabe, T., Takahashi, K., ary history of kinetoplastids and their . Mol. Kamaishi, T., and Miwa, S. (2010) Mass mortality of cul- Biol. Evol. 19, 2071–2083. tured ascidians Halocynthia roretzi associated with soften- Simpson, A. G. B., Stevens, J. R., and Lukeš, J. (2006) The evo- ing of the tunic and flagellate-like cells. Dis. Aquat. Organ. lution and diversity of kinetoplastid flagellates. Trends 90, 223–234. Parasitol. 22, 168–174. Lartillot, N., and Philippe, H. (2004) A Bayesian mixture model Stamatakis, A. (2014) RAxML version 8: a tool for phylogenetic for across-site heterogeneities in the amino-acid replace- analysis and post-analysis of large phylogenies. Bioinfor- ment process. Mol. Biol. Evol. 21, 1095–1109. matics 30, 1312–1313. Lartillot, N., Lepage, T., and Blanquart, S. (2009) PhyloBayes 3: Tanifuji, G., Kim, E., Onodera, N. T., Gibeault, R., Dlutek, a Bayesian softwar package for phylogenetic reconstruction M., Cawthorn, R. J., Fiala, I., Lukeš, J., Greenwood, S. J., and molecular dating. Bioinformat. 25, 2286–2288. and Archibald, J. M. (2011) Genomic characterization of Le, S. Q., and Gascuel, O. (2008) An improved general amino Neoparamoeba pemaquidensis () and its kineto- acid replacement matrix. Mol. Biol. Evol. 25, 1307–1320. endosymbiont. Eukaryot. Cell 10, 1143–1146. Leidy, J. (1846) Description of a new genus and species of Ento- Tyler, K. M., and Engman, D. M. (2001) The life cycle of zoa. Proc. Acad. Nat. Sci. Philadelphia 3, 100–101. Trypanosoma cruzi revisited. Int. J. Parasitol. 31, 472–481. Lukeš, J., Guilbride, D. L., Votýpka, J., Zíková, A., Benne, R., and von der Heyden, S., Chao, E. E., Vickerman, K., and Cavalier- Englund, P. T. (2002) DNA network: evolution of Smith, T. (2004) Ribosomal RNA phylogeny of bodonid and an improbable structure. Eukaryot. Cell 1, 495–502. diplonemid flagellates and the evolution of euglenozoa. J. Lukeš, J., Hashimi, H., and Zíková, A. (2005) Unexplained com- Eukaryot. Microbiol. 51, 402–416. plexity of the mitochondrial genome and transcriptome in Young, N. D., Crosbie, P. B. B., Adams, M. B., Nowak, B. F., and kinetoplastid flagellates. Curr. Genet. 48, 277–299. Morrison, R. N. (2007) Neoparamoeba perurans n. sp., an Lukeš, J., Skalický, T., Týč, J., Votýpka, J., and Yurchenko, V. agent of amoebic gill disease of Atlantic salmon (Salmo (2014) Evolution of parasitism in kinetoplastid flagellates. salar). Int. J. Parasitol. 37, 1469–1481. Mol. Biochem. Parasitol. 195, 115–122. Woo, P. T. K., and Poynton, S. L. (1995) Diplomonadida, Kineto- Michaeli, S. (2011) Trans-splicing in trypanosomes: machin- plastida and Amoebida ( ). In ery and its impact on the parasite transcriptome. Future Fish Diseases and Disorders. Volume 1: Protozoan and Microbiol. 6, 459–474. Metazoan Infections (ed.: Woo, P. T. K.), pp. 27–96. CABI, Moreira, D., López-García, P., and Vickerman, K. (2004) An Wallingford, UK. updated view of kinetoplastid phylogeny using environ- Woo, P. T. K., and Wehnert, S. D. (1983) Direct transmission of mental sequences and a closer outgroup: proposal for a new a hemoflagellate, Cryptobia salmositica (Kinetoplastida: classification of the class Kinetoplastea. Int. J. Syst. Evol. Bodonina) between rainbow trout under laboratory condi- Microbiol. 54, 1861–1875. tions. J. Protozool. 30, 334–337. Munday, B. L., Zilberg, D., and Findlay, V. (2001) Gill disease