Euglenozoa: taxonomy, diversity and ecology, symbioses and viruses

Euglenozoa is a species-rich group of protists, which have extremely diverse lifestyles and a range of features that distinguish them from other eukaryotes. They are composed of free-living and parasitic kinetoplastids, mostly free-living diplonemids, heterotrophic and photosynthetic euglenids, as well as deep-sea symbiontids. Although they form a well-supported monophyletic group, these morphologically rather distinct groups are almost never treated together in a comparative manner, as attempted here. We present an updated taxonomy, complemented by photos of representative species, with notes on diversity, distribution and biology of euglenozoans. For kinetoplastids, we propose a significantly modified taxonomy that reflects the latest findings. Finally, we summarize what is known about viruses infecting euglenozoans, as well as their relationships with ecto- and endosymbiotic bacteria.


Introduction
It is generally accepted that Euglenozoa belong to the most unusual eukaryotes [1][2][3]. This is based on a substantial body of evidence showing that in a number of cellular processes and structures, these almost invariably mono-or bi-flagellated protists departed from what can be considered the 'eukaryotic consensus'. However, this consensus was defined by the studies of just a handful of model organisms, most of which are multicellular [4]. Hence, since the majority of the extant eukaryotic diversity is hidden in protists [5], we prefer to use a 'protist-centric' view, which postulates that these unicellular forms actually are the eukaryotic standard, while the other lineages represent departures from the norm.
The phylum Euglenozoa splits into three well-defined lineages-euglenids, kinetoplastids and diplonemids-with different life strategies and distinct morphologies, yet still unified by a number of common features [6]. Although the euglenids are sometimes further subdivided into Euglenida and Symbiontida [3], both groups are usually treated together due to their morphological similarity, and we still cannot compare their genomic features in the absence of such data from the latter taxon [7]. A recent multigene phylogenetic reconstruction pointed to the potentially sister relationship between Symbiontida and Glycomonada (Kinetoplastea + Diplonemea) [8], suggesting that Symbiontida may become a separate group when more data become available (tree A).
Apart from summarizing taxonomic works, the euglenozoans are almost never treated together in the literature. The kinetoplastid flagellates are by far the best-studied representatives (almost exclusively from a parasitology-centric perspective), with most attention given to the causative agents of serious diseases, such as sleeping sickness, Chagas disease and leishmaniases [9,10]. The diplonemids, as detailed below, were considered a marginal group with no ecological relevance. That has changed recently [11,12], but still very few molecular data other than 18S rRNA are available for this almost exclusively marine group. Finally, the photosynthetic and heterotrophic euglenids are ecologically significant, primarily in freshwater ecosystems, and have potential in biotechnologies [2,13].
The striking differences in lifestyles and cellular (ultra)structure obscure the significant similarities in basic molecular processes. Firstly, all these groups distinguish themselves from other eukaryotes by transcribing nuclear genes in a polycistronic manner [14]. In neither case are the co-transcribed genes functionally related, which distinguishes them from the prokaryotic operons. The usually very long polycistronic mRNA is subsequently processed into monomeric transcripts, which are subject to another process that is found in eukaryotes rather infrequentlytrans-splicing. At the 5 0 end of each monocistronic mRNA, short spliced leader (SL) RNA, already equipped with a methylated cap, becomes attached. The corresponding SL RNA gene is invariably multicopy, and highly conserved, yet with minor species-specific differences [15].
The similarities do not stop there. In their single or dual flagella, all euglenozoans evolved an extra-axonemal structure termed the paraflagellar rod, which supports their flagella [16]. The paraflagellar rod has a characteristic lattice-like structure, which is composed of dozens of proteins, phylogenetically restricted to euglenozoans. It is reduced only in the endosymbiont-containing trypanosomatids and the amastigotes of Leishmania [17]. Studied so far only in kinetoplastids, the paraflagellar rod not only increases propulsion of the cell [18], but also participates in morphogenetic and metabolic roles, as well as in environmental sensing [19]. While all these synapomorphies were probably present in the euglenozoan common ancestor, euglenids, diplonemids and kinetoplastids have acquired significant differences over the course of evolution. This is particularly striking in the case of cis-splicing, since spliceosomal introns are almost absent in the latter group [20], while they are abundantly present in euglenids and diplonemids, many being seemingly non-canonical [11,21]. Another clear difference rests in the size of both nuclear and mitochondrial genomes. The dearth of high-quality data for nuclear genomes of euglenids and their absence in the case of diplonemids are due to the large size and repetitive character of the latter. The transcriptomes from both groups contain an extremely high number of protein-coding genes, probably reflecting their metabolic versatility [6,13]. The situation is quite different in kinetoplastids, the parasitic lifestyle of which led to gene reduction and streamlining [6]. Moreover, due to their small and compact genomes, they belong to the most sequenced eukaryotes [22].
Unexpected differences among the main euglenozoan lineages recently became apparent for their mitochondrial genomes and transcriptomes. Kinetoplastids harbour in their mitochondrial DNA in the form of relaxed (rarely supercoiled) circular molecules, either catenated or free, of two types-maxicircles and minicircles, with the former carrying all protein-coding genes, while the latter encode guide RNA genes required for the editing of the maxicircle transcripts [23]. The size of maxicircles is rather uniform, while the minicircles come in different variants [24]. In diplonemids, the single type of non-catenated circles uniquely encodes fragments of protein-coding genes, the transcripts of which have to be massively trans-spliced and edited in order to become translatable [25]. However, in both groups, the mitochondrial DNA is inflated, and its transcripts are extensively edited [26]. This contrasts with euglenids that lack any form of editing in their mitochondrion, which also contains a small genome composed of heavily fragmented linear molecules [27]. Probably, the most important difference among these groups is the presence of a secondary green plastid solely in euglenids, which have acquired it after their divergence from other euglenozoans [2,28].
Until recently, our knowledge of different groups within euglenozoans was much influenced by the availability of full-size nuclear genome sequences. While hundreds of high-quality genomes are available for trypanosomatids [22], only one such genome is available for bodonids [29] and euglenids [13], respectively, and none for diplonemids. However, this is bound to change soon, mostly due to the ever-decreasing costs and improving sequencing technologies. Recent comparative analyses of molecular features among kinetoplastids, euglenids and diplonemids were based on transcriptomes available for all of them [30].
Future studies of euglenozoans will be heavily influenced by the accessibility of their representatives to (efficient) genetic manipulations. The amenability of trypanosomatids to a range of genetic tools turned them into arguably the functionally best-studied protists [31], while most other groups significantly lag behind. However, this unfavourable situation has changed recently, as first reports of genetic modifications of bodonids, diplonemids and euglenids have been published [32][33][34][35][36]. Anticipated improvement of the methodologies of forward and reverse genetics, which would allow mediumor high-throughput functional analyses in these taxonomic groups, almost guarantee major discoveries. Euglenozoa is a very peculiar group, encompassing organisms strikingly dissimilar in their ecology, ranging from autotrophy to obligate parasitism. This inevitably influenced their classification in the era of the two-kingdoms-oflife paradigm. Kinetoplastids and diplonemids were historically considered predominantly as protozoa, and thus the International Code of Zoological Nomenclature (ICZN) was used for their nomenclature, while euglenids have been classified by different authors as either protozoa or algae. This ambiguity is reflected in their nomenclature, which has been governed in parallel by the ICZN as well as the International Code of Botanical Nomenclature (ICBN) and the International Code of Nomenclature for algae, fungi and plants (ICN), which replaced the latter in 2011. Apart from formal differences, such as the rules on citing authorship of names and emendations, this led to significant issues that include certain taxa having different names depending on the selected system (zoological or botanical). This concerns names of family-group taxa, which have different suffixes depending on the system and, more importantly, names of genera, which may be valid according to one code, but regarded as junior homonyms, and therefore replaced with different names. In addition, the ICZN jurisdiction does not extend above the family-group level, whereas ICN does not have such a restriction. Here, the nomenclature of kinetoplastids and diplonemids follows the ICZN, while for some euglenid groups, the ICN is used by default with the valid names according to the ICZN indicated.

. Free-living kinetoplastids
The common ancestor of Kinetoplastea apparently was a freeliving benthic bacterivorous organism using the anterior flagellum for motion and transporting food particles to the cytostome, while the posterior one ensured gliding on the substrate (plate A, 1,2,4-10, [12][13][14]. This lifestyle is still preserved by a large proportion of kinetoplastids [37]. They inhabit permanent and temporary water bodies with various levels of salinity (freshwater to hypersaline) and some species royalsocietypublishing.org/journal/rsob Open Biol. 11: 200407 were shown to tolerate the transition from marine water to freshwater, and vice versa [38][39][40][41]. Kinetoplastids are very numerous in benthic communities, where they constitute 5-20% of the total biomass of all heterotrophic flagellates, second only to euglenids, suggesting their important role in controlling bacterial growth [42]. They are abundant in seawater ice and some species can be cultured even from the pelagic zone, demonstrating their presence there at least at the cystic stage [43,44]. Recent studies using molecular methods demonstrated that in many water bodies, most of the kinetoplastid biomass is created by neobodonids, of which Neobodo, Rhynchomonas and Dimastigella are the most frequent ones ( plate A, 7,9,14) [45][46][47]. An extensive analysis of free-living kinetoplastids in hundreds of globally collected oceanic samples revealed their abundance being 0.14%, with highest abundance in the mesopelagic zone. Their community structure and richness are significantly influenced by oxygen concentration, salinity and temperature [48].
Moreover, many kinetoplastids live in the soil and readily settle on various organic substrates, such as faeces, composts, etc. [49,50]. Most of the free-living flagellates graze on bacteria with the help of cytostomal lips or, as in Rhynchomonadidae, a flagellum-attached motile proboscis. In addition to bacterial cells, their digestive vacuoles can contain microalgae or detritus particles. Parabodo caudatus ( plate A, 13), being a relatively large (up to 20 µm long) species, exerts both bacterivory and predation, while Rhynchobodo spp. ( plate A, 4) are obligatory predators devouring other flagellates [37].
Some kinetoplastids are known to form cysts, which help them to survive adverse conditions, for example, pass through the digestive system of an animal and settle in its faeces after their discharge [37]. Moreover, some of these flagellates become very tolerant to harsh environments even at the active (non-encysted) stage. This resulted in a series of records of free-living kinetoplastids (Parabodo caudatus, Dimastigella trypaniformis and Procryptobia tremulans) from stool and urine samples or urine-impregnated animal cage bedding sometimes misinterpreted as evidence of their parasitic nature [51][52][53][54]. Interestingly, such tolerance in parabodonids apparently preadapted them to parasitism, which originated in this group at least twice [55].

Parasitic, mutualistic and commensal non-trypanosomatid kinetoplastids
Multiple transitions to various forms of symbiosis can be observed in all orders of Kinetoplastea except Eubodonida (tree B). The earliest branch within this group, Prokinetoplastida, does not contain any described free-living forms. The ectoparasitic Ichthyobodo, affecting both freshwater and marine fish, is generally similar to free-living kinetoplastids ( plate A, 17). In contrast with other symbiotic forms, it anchors on the host epithelium with its rostrum forming an attachment disc and a cytostome process, which is inserted directly into the cytoplasm of the host cell and feeds by myzocytosis [56]. Accumulation of parasites on the epithelium of gills and fins leads to tissue necrosis, which often entails the death of fish, especially fingerlings. Dissemination of Ichthyobodo spp. occurs using a free-swimming stage lacking the rostrum [57]. Another prokinetoplastid genus, Perkinsela, represents one of the most simplified symbiotic eukaryotes, permanently living in the cytoplasm of amoebae, such as Paramoeba ( plate A, 14) [58]. The nature of their relationships is mutualistic as judged by reciprocal metabolic dependence of the two partners, evident from the study of their genomes [59]. Since both Ichthyobodo and Paramoeba live on fish gills, it was proposed that an ancestral Ichthyobodo-like flagellate had been engulfed, but not digested, by an amoeba, and eventually evolved into the endosymbiont Perkinsela [58]. Perkinsela is tightly associated with cosmopolitan Paramoeba, lacks any traces of flagellum ( plate A, 14), and has an extremely reduced metabolism, as well as the largest known mitochondrial DNA [59].
The currently unclassified flagellate Desmomonas prorhynchi, a parasite of the turbellarian Prorhynchus, shares two features with Ichthyobodo: polykinetoplast DNA and attachment to the host cell by an appendage at the anterior end. However, this structure performs an exclusively mechanical function, while feeding is supposed to occur via osmotrophy [60]. Another flagellate with an uncertain taxonomic position, Cephalothamnium cyclopum, is the only described colonial kinetoplastid, which attaches to freshwater copepods [61]. Like its free-living relatives, this flagellate feeds by intercepting bacterial cells with its anterior flagellum and directing them to cytostomal opening. Given that C. cyclopum uses its host only as a substrate and the only inconvenience from its presence may consist of decreased hydrodynamic characteristics of the crustacean, this kinetoplastid is considered as ectocommensal.
Azumiobodo hoyamushi is a neobodonid parasite of ascidians, of which the most important is the sea pineapple Halocynthia roretzi, a cultivated edible species popular in Korea and Japan ( plate A, 5). By invading the tunic, this flagellate is responsible for the so-called soft tunic syndrome, associated with high mortality rates [62]. Being seasonal, it survives the period of high temperatures in resistant cysts attached to the substrate [63,64]. Another parasitic neobodonid is the recently described Allobodo chlorophagus, invading the utricles and the main filaments of the green siphonal alga Codium fragile and feeding on its chloroplasts and starch granules ( plate A, 11) [65].
Parasitic parabodonids are represented by the genera Trypanoplasma and Cryptobia, which often used to be combined into one genus due to morphological similarity. Trypanoplasma spp. are extracellular parasites of fish bloodstream transmitted by haematophagous leeches during bloodfeeding ( plate A, 15) [66]. However, at least two species, T. salmositica and T. bullocki, can exit to the body surface where they reside in the mucus and can be transmitted to other fish by direct contact [57]. After ingestion with the blood, parasites multiply in the leech crop without significant changes in morphology and migrate to the proboscis sheath, wherefrom they are transmitted to the bloodstream of another fish [67]. The severity of infection-acute to chronic-apparently depends more on the level of mutual adaptation as well as the individual variation of host immunity between host and parasite than on parasitaemia [68][69][70]. There is only a single record of a trypanoplasma in a non-fish host, namely in a salamander [71].
The genus Cryptobia can be subdivided into three distinct ecological groups: (i) ectoparasites of fish, (ii) endoparasites of invertebrates and (iii) endoparasites of fish. Members of the first group, represented by Cryptobia carassii, live on fish gills and are regarded as commensals feeding on dead royalsocietypublishing.org/journal/rsob Open Biol. 11: 200407 epithelium and microorganisms [57]. However, gill infections by C. branchialis are associated with high mortality in adult cultured carps, goldfish and catfish as well as in juvenile grass carp [72]. Lamellasoma bacillaria, described as a monoflagellated kinetoplastid living on the fish gills, may also belong to this group of cryptobiae [73].
In invertebrates, Cryptobia infections are quite diverse in terms of localization within the host and taxonomic groups parasitized. Some of them (Cryptobia helicis, C. innominata and C. carinariae) were found attached to the epithelium of spermatheca of floating sea snails and various pulmonates [74][75][76]. The vagina of haematophagous leeches often serves as a habitat of C. vaginalis ( plate A, 16) [77], while C. udonellae was described from the excretory system of an ectoparasitic marine worm [78]. Other species were described from the intestine of a chaetognath (C. sagittae) and a freshwater planarian (C. dendrocoeli), the latter of which was also detected in the eggs, pointing to a potential transovarial transmission [79,80]. It is presumed that Cryptobia spp. from the reproductive system are transmitted via sexual contacts [75,81], while the ectoparasites of aquatic animals should have free-swimming swarmers, although this has not been confirmed [73]. The cryptobiae found in the intestinal contents of frogs and lizards appear to be accidentally ingested parasites of invertebrates, as judged by the morphology of the flagellates and uniqueness of such records [82,83].
In the third group of cryptobiae, encompassing piscine intestinal flagellates, six out of seven described species are known as specific parasites of marine fish. These species do not display any pathogenic effect and therefore are usually considered as commensals [57]. The only known freshwater representative of this group, C. iubilans, infects various cichlid fishes and causes gastroenteritis often associated with invasion of other organs, leading to high mortality [84][85][86]. Сryptobiae belonging to this group can be transmitted directly by ingestion from water and by feeding on infected corpses [57].
Jarrelia atramenti, a flagellate described from the blowhole mucus of a pygmy sperm whale, appears to be a harmless commensal feeding on detritus and/or bacteria [87]. Its resemblance to parasitic parabodonids, proposed to be evidence of their relatedness, may be in fact a parallelism caused by similar living conditions. Indeed, flexible body and flagellar attachment evolved independently in Cryptobia, Trypanoplasma and Dimastigella.

Trypanosomatids
The family Trypanosomatidae contains exclusively obligate parasites and represents the most diverse kinetoplastid group in terms of the number of species described and/or revealed using molecular typing [88][89][90]. Among parasitic protists, it has the widest host range: animals ( predominantly insects and vertebrates), flowering plants and even ciliates [91]. Based on the type of life cycle, trypanosomatids are usually subdivided into two non-taxonomic groups. Monoxenous species develop in a single host, whereas dixenous switch between two, of which one serves as a vector. Molecular phylogenies suggest that the most recent common ancestor of trypanosomatids was a monoxenous parasite of insects [92,93], with the dixenous lifestyle emerging independently at least three times in distantly related lineages of these flagellates [55].

Monoxenous trypanosomatids
Most trypanosomatid genera are monoxenous and the overwhelming majority of their species parasitize two large groups of insects: Diptera and Heteroptera (i.e. flies and true bugs, respectively) [91,94]. Among other insects, used by them as hosts are Hymenoptera (bees, bumblebees, wasps and sawflies), Siphonaptera (fleas), Blattodea (cockroaches), Lepidoptera (moths) and Trichoptera (caddis flies). The single records of monoxenous trypanosomatids from a louse (Anoplura), a planthopper (Homoptera), a scorpion fly (Mecoptera) and a domestic cricket (Orthoptera) may refer to accidental non-specific infections [91]. The adaptation to insects, which are omnipresent, extremely diverse and abundant animals, probably predetermined the transition of these flagellates to other hosts. Trypanosomatids invaded Acari (ticks and mites) and freshwater ciliates living sideby-side with insects, vertebrates and plants. The two latter host groups are associated with dixenous trypanosomatids, although monoxenous species have also been occassionally reported from them [95,96]. The presence of trypanosomatids in nematodes and molluscs [97] may indicate a more complex evolutionary pathway of these flagellates, but first it requires confirmation with modern methods.
The ancestral and still most common lifestyle of monoxenous trypanosomatids includes stages that inhabit insect gut, usually being attached to its wall, and some either active (i.e. flagellate) or inactive (endomastigote or cyst-like amastigote) cells are discharged with faeces. Other insects become infected by feeding on contaminated substrates or directly on fresh faeces (coprophagy) [98]. In addition, the parasites can be transmitted between insects via cannibalism and predation, although the latter way is probably responsible only for the transmission of non-specific transient infections [99]. Some monoxenous trypanosomatids can migrate within insects to other locations in order to facilitate transmission [89]. Thus, parasitism in Malpighian tubules of female firebugs ensures timing the mass production of infective cystlike amastigotes of Blastocrithidia papi to oviposition [100]. Haemocoel invasion allows the inheritance of Herpetomonas swainei between developmental phases of the host saw fly [101], while in the case of Leptomonas pyrrhocoris, this increases the efficiency of transmission by cannibalism [102]. The role of intracellular stages, which are very rare in life cycles of monoxenous trypanosomatids, is uncertain [103]. However, the potential to live intracellularly probably preconditioned transition of these flagellates to dixeny (see below) and parasitism in ciliates. The latter has been repeatedly described from various ciliate species where it was always associated with the macronucleus and, at least in some cases, effective transmission between host cells was observed pointing to specific relationships [104][105][106][107].
Although most monoxenous trypanosomatids are considered non-pathogenic or even commensals [98], this view is influenced by the fact that their effects on the hosts are poorly known and have been investigated in only a few practically important or model insect species. It was shown that trypanosomatid infections lead to elevated mortality rates in triatomine bugs, honeybees, sawflies, eye gnats, fruit flies, firebugs and water striders [101,[108][109][110][111][112]. Other adverse consequences of trypanosomatid infections on insects include delayed development, decrease in body weight, disturbed digestion and excretion, lower endurance, impaired foraging efficiency and lower fecundity [113][114][115][116][117][118]. The above effects royalsocietypublishing.org/journal/rsob Open Biol. 11: 200407 have a significant impact on host fitness, and thus trypanosomatids play an important role in controlling the population sizes of their hosts.

Phytomonas
Some trypanosomatids acquired the ability to live in plants, on which their bug hosts feed and, thus, became dixenous. These flagellates belong to the genus Phytomonas and parasitize phloem, fruits, latex or seeds of various plants [119,120]. The bug hosts serve as vectors and, since the contaminative route of transmission to plants is not very effective, the parasites migrate from the intestine through haemocoel to salivary glands [96]. Here, the infective endomastigotes are formed, which are inoculated into plant juices with the bug's saliva during feeding [103,121,122]. Interestingly, in some species, no development occurs in the host gut, which is then used only for the transit of flagellates [103,123]. At least one phytomonad species, P. nordicus, became secondarily monoxenous, since it inhabits a predatory pentatomid bug [124]. The pathogenicity of Phytomonas for insects remains unknown, while their effect on plants ranges from asymptomatic infections to serious diseases of cultural plants [120]. Phytomonas francai living in lactiferous ducts of manioc is associated with root dystrophy; P. leptovasorum causes phloem necrosis and subsequent lethal wilt of coffee trees; P. staheli obstructing phloem of oil and coconut palms accounts for acute wilt in these plants; and an unnamed phytomonad is responsible for the withering of red ginger [96]. These diseases have a high impact on agriculture in developing countries and result in serious economic losses [119].

Leishmania and related dixenous genera
The genera Leishmania, Porcisia and Endotrypanum represent a monophyletic group, whose parasitism in blood-sucking sandflies (Phlebotominae) allowed them to become dixenous parasites of mammals [125]. Secondarily, some Leishmania spp. changed either the vertebrate host or the vector: the subgenus Sauroleishmania switched from mammals to lizards and snakes, while the subgenus Mundinia started using biting midges (Ceratopogonidae) instead of sandflies [126][127][128]. Leishmania is most species-rich genus and many of its members are human parasites, which drew most attention to this group, while the information about Porcisia and Endotrypanum is scarce. The development of leishmaniae in vectors is confined to the intestine, although there are some differences between subgenera in the localization of the proliferative procyclic promastigotes (midgut, pylorus and/ or hindgut) [127]. However, they eventually migrate to the anterior midgut, where they destroy the chitin lining of the stomodaeal valve and secrete a gel plug obstructing the alimentary canal, thus disturbing the normal sucking process [129]. An infected sandfly regurgitates the plug with metacyclic flagellates into the vertebrate bloodstream and due to the inability to swallow the blood makes more attempts increasing chances of spreading the parasites [130]. In the vertebrate, the metacyclic promastigotes are quickly taken up by phagocytic cells and proliferate in their phagolysosomes as amastigotes [131]. Depending on the behaviour of infected macrophages, leishmaniasis manifests itself as either cutaneous (skin ulcers), mucocutaneous (sores in the mucosa of nose, mouth or throat) or visceral, which affects internal organs such as the liver, spleen and bone marrow and is usually fatal without treatment [127].
About 20 species of Leishmania, belonging to the subgenera Leishmania, Viannia and Mundinia parasitize humans. They are responsible for up to one million new cases of leishmaniasis annually, of which up to 90 000 correspond to the visceral form [132]. The visceral form of the disease can be spread even outside the endemic areas either venereally or congenitally [133][134][135][136]. In addition to humans, Leishmania was reported to infect about 70 species of mammals (rodents, carnivores, xenartrans, hyraxes, marsupials, chyropterans, ungulates lagomorphs and primates), with most cases being asymptomatic. The only notable exceptions are canine visceral leishmaniasis, with severe symptoms in over 50% of cases [137], and rare cases of atypical cutaneous leishmaniasis in cows and horses [138,139]. Porcisia living in porcupines and Endotrypanum parasitizing sloths and squirrels do not appear to produce any symptoms [125]. However, Leishmania colombiensis (now assigned to Endotrypanum) is known to cause both cutaneous and visceral leishmaniasis-like diseases in humans [140,141].

Trypanosoma
Trypanosoma is a very speciose genus enclosing approximately 500 species or over 60% of all described species of the family Trypanosomatidae ( plates B and C, . While the frog-infecting type species (T. sanguinis = T. rotatorium) described by Gruby already in 1843 may be of marginal importance, ever since trypanosomes became the best-known protists. Life cycles of these flagellates vary considerably as they parasitize all classes of vertebrates (from agnathans to mammals) and are transmitted by a wide range of vectors including blood-sucking insects (flies, bugs, fleas and lice), ticks, leeches and even vampire bats [9,91]. In vertebrates, they occur most frequently as trypomastigotes, rarely as epimastigotes or amastigotes, while in invertebrates, they predominantly exhibit most trypomastigote or epimastigote morphology, or infrequently occur as promastigotes and amastigotes [92].
From the practical point of view, the most important mammalian trypanosomes were traditionally subdivided into two sections or intrageneric taxons that follow distinct developmental programmes [142]: Salivaria (derived from saliva), with the best-known member being the Trypanosoma brucei complex causing human African sleeping sickness and nagana in livestock and other animals ( plate C, 32-34), terminate development in the salivary glands mouthparts of the vector and are transmitted to a vertebrate host by bite. Stercoraria (stercus = dung) exemplified by Trypanosoma cruzi complex species causing Chagas disease (plate C, [37][38], terminate development in the rear part of the digestive tract of the vector with the transmission to the vertebrate host being contaminative by excrements. With the advent of molecular phylogenetics, it became obvious that neither mammalian trypanosomes in general nor any of the two proposed sections represent monophyletic groups, and therefore they do not deserve a taxonomic status [89,143,144]. Nevertheless, the words salivarian/stercorarian still can refer to the type of development within the vector. Non-mammalian trypanosomes generally follow the stercorarian developmental programme, but in leeches, parasites migrate to the proboscis sheath to be transmitted during blood-sucking [57]. royalsocietypublishing.org/journal/rsob Open Biol. 11: 200407 Trypanosoma brucei evansi and T. b. equiperdum are two notable exceptions ( plate C, [33][34], as they lost the capacity to survive in the gut of an insect vector [145]. The former subspecies therefore switched to mechanical transmission, which allowed it to use non-specific vectors, while the latter adapted to the direct (venereal) transmission and thus became a monoxenous parasite [146]. In other species, direct transmission can also occur, but is facultative [147][148][149][150]. The most (in)famous species are Trypanosoma brucei and T. cruzi, which cause serious human diseases-sleeping sickness and Chagas disease, respectively [151]. The first one is transmitted by tsetse flies in Africa and invades various tissues, but primarily the blood and adipose tissue [152], as freeswimming trypomastigotes and eventually infects cerebrospinal fluid with fatal consequences [153]. Being a serious public health threat in the past, this disease is now on the way to elimination [154]. Trypanosoma cruzi is transmitted by triatomine bugs among a wide range of mammalian hosts, in which it develops in various organs and tissues as intracellular amastigotes [155]. In most cases, the disease does not manifest clinical signs at the beginning, but during the prolonged chronic phase, it significantly undermines health in the human population of South and Central America leading to increased mortality rates [156]. Trypanosoma rangeli has the same geographical distribution and vectors as T. cruzi and is also able to infect humans but appears to be non-pathogenic [157]. Some tsetse-transmitted African trypanosomes, such as T. vivax, T. congolense and T. brucei brucei, cause serious diseases in livestock, collectively named African animal trypanosomiasis. These diseases are associated with high mortality rates and lead to significant damage in animal husbandry, although some local breeds and wild animals acquired tolerance to them [158]. For the overwhelming majority of trypanosome species, their effects on the host are not known and they are often considered as non-or subpathogenic and can cause observable disease only under stress conditions. This is exemplified by piscine trypanosomes, which seem to be well tolerated in wild fish populations [159]. However, in farmed fish or wild juvenile individuals, infections are associated with high mortality rates due to anaemia, anorexia and tissue damage [160][161][162][163].

Taxonomy
This section contains nomenclatural changes and according to the ICZN requirements for publications in online-only journals, this work has been registered in Zoobank: urn:lsid:zoobank.org:pub:81EA01C5-8989-4BBD-9C64-04D81132307D.
Possess kinetoplast, represented by one or several large masses of mitochondrial DNA termed kinetoplast DNA (kDNA). Four kinetoplast types are distinguished: eukinetoplast-dense network of interlocked DNA circles, prokinetoplast-single compact mass not organized into a network, polykinetoplast-several clusters scattered over the mitochondrial lumen, and pankinetoplast-a diffuse mass occupying a large portion of a mitochondrial lumen [164]. Ancestrally, kinetoplastids bear two heterodynamic flagella, of which one or both were lost in some lineages; mitochondrial RNA undergoes editing represented by deletions and insertions of uridine residues. Some features considered for a long time to be defining (e.g. polycistronic transcription of nuclear genes, trans-splicing via spliced leader RNA, compartmentalized glycolysis, base J, etc.) were recently shown to be present also in other euglenozoan lineages [6].
Note: Until relatively recently, all kinetoplastids have been classified into two large groups-bodonids (freeliving, ectocommensals, ecto-or endoparasitic biflagellate species) and trypanosomatids (exclusively endoparasitic uniflagellate species). However, the 18S rRNA gene-based molecular phylogenetic analysis showed paraphyly of bodonids, which were subsequently separated into four orders royalsocietypublishing.org/journal/rsob Open Biol. 11: 200407 [165]. Thus, now the term 'bodonids' for the designation of non-trypanosomatid kinetoplastids is deprecated. As judged by available environmental sequences, the diversity of kinetoplastids is much broader than that described to date, and there are some undiscovered lineages potentially representing new high-level taxa (up to the subclass) [166].
• Subclass Prokinetoplastia . The phylogenetic group enclosing the genera Ichthyobodo and Perkinsela as judged by 18S rRNA gene-based trees [165].  [168,169].  Etymology: The new name and the two previous ones share the Greek roots κρυπτός (hidden) and αὖλαξ (furrow), referring to the distinctive feature of the genus, and the feminine gender.
Note: All described species were once assigned to Rhynchobodo and the diplonemid Hemistasia based on light microscopy [179]. Molecular phylogenetic inference showed that flagellates identified as Rhynchobodo and Cryptaulax/Cryptaulaxoides are unrelated [180]. Note: Authorship of this name is often attributed to Lackey who mistakenly used it instead of Rhynchomonas [183]. However, only Vørs made the name available by providing the genus description and specifying the type species [182]. ▪ Family Rhynchomonadinae Cavalier-Smith, 2016. Solitary, free-living; biflagellate, the anterior flagellum adheres to the flexible proboscis and they move together, the posterior flagellum is used for gliding and attached to the body at least in its proximal part; cytopharynx not supported by a microtubular rod [170]. ▪ Genus Dimastigella Sandon, 1928. Free-living, in soil or freshwater; anterior flagellum significantly longer than the proboscis, posterior flagellum attached to the cell body across the whole length of the latter; cytostome on or under rostrum; phagotrophic; polykinetoplastic [184,185] [3,165]. Previously, parasitic representatives of this group were considered a single lineage and often lumped into one genus (Cryptobia), but molecular phylogenetic analyses showed their polyphyly [187]. ▪ Family Cryptobiidae Poche, 1911 emend. Kostygov. Clade uniting the genera Cryptobia and Parabodo based on 18S rRNA gene phylogenies [68,188]. ▪ Genus Cryptobia Leidy, 1846. Parasites/ commensals of fish (on gills or in the gut) or various invertebrates (in the lumen of reproductive, digestive or excretory organs) [73,78,189]; recurrent flagellum attached to cell body without the formation of undulating membrane, its posterior part is used for attachment to host epithelium; conspicuous ventral furrow; phagotrophic with well-developed but miniaturized cytopharynx in most species; subpellicular microtubules extend to whole-cell length; pro-or pankinetoplastic [81,164] ( plate A, 16). Type species: Cryptobia helicis Leidy, 1846. Note: The genus may be paraphyletic with respect to Parabodo as judged by sequence data on two species from invertebrates and fish, although the relationships are poorly resolved [190]. ▪ Genus Parabodo Skuja, 1939 emend. . Free-living, solitary; posterior flagellum free; the cytostome is placed at the anterior end of the cell making the latter bifurcate, well-developed cytopharynx; subpellicular microtubules extend to whole-cell length; prokinetoplastic [191,192] [194]; phagotrophic or osmotrophic; cytostomecytopharyngeal complex fully developed only in a few representatives, while the majority has no cytopharynx, and cytostome is present as a shallow pit or completely absent [195,196]. For a long time, the classification was based on the presence of the following morphotypes in the cell cycle: promastigote (elongated with apical flagellum and prenuclear kinetoplast), choanomastigote (shortened, with apical flagellum and prenuclear kinetoplast), opisthomastigote (elongated, with apical flagellum and postnuclear kinetoplast), opisthomorph (shortened, with apical flagellum and postnuclear kinetoplast), epimastigote (with lateral flagellum attached to the cell body and prenuclear kinetoplast), trypomastigote (with lateral flagellum attached to the cell body and postnuclear kinetoplast), amastigote/endomastigote (flagellum not emerging from the pocket, the first variant predominantly used when flagellum is very short) and cyst-like amastigote (compact cells with dense cytoplasm, completely lacking flagellum). Single family. The taxonomic changes introduced here follow the guidelines specified for this group previously [197]. ▪ Family Trypanosomatidae Doflein, 1901. With the same definition as the order (tree C). Note: Historically, all monoxenous genera were often termed 'lower trypanosomatids', but after the switch to the phylogeny-oriented paradigm, this concept has been abandoned [92]. As an alternative, a colocation 'insect trypanosomatids', which has no evolutionary connotations, is often used.
Note: Historically, mammalian trypanosomes were subdivided into the sections Stercoraria (Herpetosoma, Megatrypanum and Schizotrypanum) and Salivaria (Duttonella, Nannomonas, Pycnomonas and Trypanozoon). Being longestablished and of practical importance for the community of medical and veterinary doctors, this mammalian-centred classification is incorrect from the phylogenetic point of view (see Biology of Kinetoplastea), as it not only does not royalsocietypublishing.org/journal/rsob Open Biol. 11: 200407 correspond with the known diversity of trypanosomes but also cannot accommodate species transmitted by other modes. Moreover, mammalian trypanosomes are not monophyletic and the names stercoraria and salivaria, which were proposed to reflect the differences between the modes of transmission, are non-taxonomical and have only historical value.
In parallel, the genus Trypanosoma was divided into several subgenera based on rather subtle morphological differences. Interestingly, this historical taxonomical classification in general corresponds well with the current phylogenetic analysis. However, despite the indisputable usefulness of this classification, in recent decades, the usage of subgenera was largely omitted ( probably because of anticipated-but not materialized-conflicts between the morphology-and phylogenybased systems). As a consequence, newly emerging clades in the expanding phylogenetic trees were named after the bestknown representatives, causing confusion. Here, we revive the original subgeneric concept, and by erecting several new subgenera achieve mutual harmonization of the old morphological and the modern phylogenetic approaches. We believe that this taxonomical system reflects best the true diversity of these important parasites.
Species for which only morphological but no sequence information is available are marked with an asterisk (*). We have mapped the hosts (classes of vertebrates) onto the phylogenetic trees, highlighting associations among clades and their vertebrate host.
• 'the aquatic clade' (monophyletic) (plate B, [18][19][20][21][22][23]. Note: Molecular phylogenies confirmed the monophyly of the genus Trypanosoma [198] and its subdivision into the aquatic and terrestrial clades [199][200][201]. In most reconstructions, the aquatic clade was further split into the 'fish/turtle' and 'amphibian' subgroups. While the Tree C. Trypanosomatidae. A tree summarizing multiple phylogenetic reconstructions, mostly 18S rRNA gene-based. Dashed line denotes an unstable clade, which is disrupted when certain taxa are included into the analysis. For the genus Leishmania, relationships between its four subgenera are also shown. Highlighting denotes types of the life cycles (see graphical legend).
Votýpka and Kostygov. Diagnosis: Leech-transmitted parasites of aquatic vertebrates. Morphologically variable medium to large conspicuously elongated trypomastigotes with a notably bent body, undulating membrane including free flagellum, and with the kinetoplast situated close to the posterior end of the body. Defined by 18S rRNA-based phylogenetic analyses. Type species: Haematomonas cobitis Mitrophanow, 1883 (= Trypanosoma cobitis).
Note: Mayer in 1843 found in the blood of a frog (Rana esculenta) captured in Germany two organisms that he named Amoeba rotatoria and Paramaecium loricatum, which were later recognized as first-ever described trypanosomes [209]. Gruby published later the same year a description of a haemoflagellate from the blood of a frog in France and named it Trypanosoma sanguinis (from Greek trypanon, an auger; soma, body) [210]. In 1901, Doflein created the family Trypanosomidae, in which the genus Trypanosoma was subdivided into three subgenera including the nominotypical one-Trypanosoma with T. sanguinis Gruby 1843 as a type [205]. In 1926, International Commission on Zoological Nomenclature accepted T. rotatorium Mayer 1843 as the senior synonym of T. sanguinis Gruby 1843 [211].
• 'the terrestrial clade' (monophyletic) ( plate C, [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40]. Note: The internal classification of trypanosomes from terrestrial hosts is rather confusing. Mammalian trypanosomes, subdivided into the sections stercoraria and salivaria, were mostly singled out, followed by the avian trypanosome branch(es). Additionally, new clades appeared gradually in phylogenetic studies, the inclusion of which is a significant problem. Moreover, these abovementioned groups are not monophyletic, making the internal system of the terrestrial clade unstable. We have attempted to rectify the situation by building a system that accommodates taxonomic units, for which sequence information is available. ○ avian subgenera ( paraphyletic) Based on comparative morphology and developmental cycles, avian trypanosomes were hypothesized to be closely related to the subgenus Megatrypanum that infects ruminants [212,213]. This assumption is now supported by phylogenetic and phylogenomic studies that have also shown paraphyly of avian trypanosomes, represented by at least three distinct lineages (T. avium, T. corvi and T. bennetti clades) [144,214,215]. These groups are distinguishable by the morphology of bloodstream stages and the kinetoplast thickness [216,217]. ▪ Subgenus Trypanomorpha Woodcock, 1906 emend. Votýpka Diagnosis: Medium to large size trypomastigotes (40-100 μm) with longitudinal striations (myonemes), central oval nucleus, prominent undulating membrane and small kinetoplast to which the free flagellum is anterior. Cells in culture have kinetoplast exceptionally thick (greater than 500 nm). Defined by 18S rRNA-based phylogenetic analyses; cosmopolitan distribution.
Note: For the type species described from a European little owl Athene noctua [218], neither culture nor sequence data are available. Both exist for T. thomasbancrofti Šlapeta, 2016 and also for less clearly defined T. avium Danilewsky, 1885 and T. gallinarum Bruce et al., 1911 [214-217,219]. The vaguely defined species Trypanosoma avium was described in 1885 by Danilewsky from birds, but the type material was not preserved [220]. In 1903, Laveran proposed to restrict this species name to parasites of owls [221]; however, the name was often used to designate any bird trypanosome.
Diagnosis: Medium to large trypomastigotes (about 40-80 μm) in the bloodstream of bird hosts with longitudinal striations (myonemes), nucleus positioned centrally and posterior kinetoplast; the thickness of kinetoplast in cultured cells is less than 500 nm. Defined by 18S rRNA-based phylogenies; cosmopolitan distribution.
Etymology: The generic name refers to the fact that trypanosomes come from bird hosts, the order Aves (the Latin name for a bird is avis).
Note: Baker [222] emended T. corvi and restricted the use of the name to large trypanosomes from non-American corvids and also from other bird families. The species was re-described [223] and phylogenetically characterized [214] and forms the T. corvi clade along with T. culicavium Votýpka et al., 2012. Although morphologically indistinguishable from the subgenus Trypanomorpha (T. avium clade), Aviotrypanum (T. corvi clade) is not directly related to it [144,216,217], justifying separate treatment. ▪ Subgenus Ornithotrypanum Votýpka, subgen. nov.
Diagnosis: Small to medium-size non-striated avian trypanosomes (less than 40 µm in length) with kinetoplast situated close to the posterior end of the body. Kinetoplast thickness of cells in culture below 500 nm. Defined by 18S rRNA-based phylogenetic analyses; cosmopolitan distribution. Type species: Trypanosoma bennetti Kirkpatrick, Terway- Thompson and Iyengar, 1986, here designated.
Etymology: The generic name refers to the fact that trypanosomes come from bird hosts, the order Aves (the ancient Greek name for a bird is órnῑs; ὄρνῑς).
Note: Out of a number of morphologically described species parasitizing wild birds, for five of them  [144,224,225]. Their divergence justifies the establishment of a new subgenus-Ornithotrypanum.
Phylogenomic approach revealed that the T. bennetti (= Ornithotrypanum) and T. avium (= Trypanomorpha) clades are nested within the mammalian clade and are paraphyletic with respect to the ruminants-infecting Trypanosoma theileri Laveran, 1902 [144], thus breaking the monophyly of mammalian Tree D. Trypanosoma. A tree summarizing multiple phylogenetic reconstructions, mostly 18S rRNA gene-based. All species or a selection of the most important ones (marked with three dots at the end), for which affiliation with a given subgenus and/or clade was confirmed using molecular phylogenetic methods, are listed. Highlighting denotes orders of vertebrate hosts (see graphical legend). Human parasites are underlined. Type species are shown in bold and the names accepted as senior subjective synonyms are marked with an asterisk.
Note: This globally distributed species found in more than 100 rodent species ( predominantly Rattus spp.) and rarely also in humans [226], is transmitted by the ingestion of fleas Xenopsylla cheopis and Nosopsyllus fasciatus or in their faeces. Considered largely non-pathogenic for rodents, it can cause serious disease in unnatural hosts. After its introduction in synanthropic rats to Christmas Island, it drove the endemic rat Rattus macleari to extinction, being the only known cases of a trypanosomatid responsible for the extinction of its host species.
The questions whether three phylogenetically related species with Megatrypanum-like morphology and basal phylogenetic position, T. talpae (European mole, Talpa europaea), T. sapaensis (white-toothed shrew, Crocidura dracula) and T. anourosoricus (mole shrew, Anourosorex yamashinai) should be included into the subgenus Herpetosoma remains unresolved [230]. ▪ Subgenus Megatrypanum Hoare, 1964 Diagnosis: Large trypomastigotes (40-100 µm) with long pointed posterior extremity; medium nonterminal kinetoplast located near the nucleus and far from the posterior end of the body; long free flagellum; reproduction as epimastigotes in the mammalian host.
Note: Other members are ovine T. melophagium Flu, 1908 using a sheep ked (Melophagus ovinus) as a vector, while caprine *T. theodori Hoare, 1931 is transmitted by a goat ked (Lipoptena capreoli). Two species of cervid trypanosomes, European and North American T. cervi Kingston and Morton, 1975 and newly described Pan-American T. trinaperronei Teixeira, Camargo and García, 2020, are transmitted by deer keds (Lipoptena cervi and L. mazamae) [231]. Tabanids probably transmit *T. stefanskii Kingston et al., 1992 from a roe deer (Capreolus capreolus) [232] and *T. ingens Bruce et al., 1909 from African antelopes [233]. Vectors remain unknown for the closely related simian T. cyclops Weinman, 1972 and *T. lucknowi Weinman, White and Antipa, 1984 from Macaca monkeys. Morphological [234,235] and phylogenetic studies [199,236] justify inclusion of these two species into the subgenus Megatrypanum. Note: Probably least settled within the former stercoraria is the taxonomy within the so-called T. cruzi superclade. Its members have been hypothesized to primarily be parasites of bats (Chiroptera), from which they have expanded into other mammals. Two species, T. cruzi Chagas, 1909 and T. rangeli, both restricted to the New World, are capable of infecting humans. The majority of known invertebrate vectors of these trypanosomes belong to true bugs (Heteroptera). The T. cruzi superclade incorporates two subgenera: Schizotrypanum and Aneza, as well as several (un)named clades and species complexes [237][238][239]. Diagnosis: Relatively small trypomastigotes (15-25 μm), typically C-or S-shaped in blood smears with short pointed posterior extremity, a large subterminal kinetoplast and long free flagellum. In mammals, reproduction takes place in form of the intracellular amastigotes. Type species: Trypanosoma cruzi Chagas, 1909; a causative agent of Chagas disease in humans, transmitted by triatomine bugs (e.g. Triatoma, Rhodnius) [240]. Note: Trypanosoma cruzi (also known as T. cruzi cruzi or T. cruzi sensu stricto) has a very high molecular and phenotypic heterogeneity, reflected by the existence of seven genetically distinct lineages (or discrete typing units, DTUs) termed TcI-TcVI and Tcbat [241]. An impartial comparison of this conspicuous genetic diversity, which corresponds very well with life cycles, clinical manifestations and host specificity, with the situation of the T. brucei complex, reveals a striking and untenable discrepancy between these two key species complexes. While in the T. brucei complex, five DTUs have the status of five different royalsocietypublishing.org/journal/rsob Open Biol. 11: 200407 (sub)species (see below), the T. cruzi complex has so far not been split into subspecies and sticks to the DTU code. Therefore, we propose that the same system, based on subspecies, should be applied for both species complexes. We urge our colleagues working with T. cruzi to implement such a system.
The current T. cruzi complex is accompanied by three (sub)species from bats with typical Schizotrypanum morphology: T. Diagnosis: Small to medium-size trypomastigotes (25-35 µm) with long pointed posterior extremity, medium subterminal kinetoplast and long free flagellum, are all similar to the subgenus Herpetosoma. At least, some species (e.g. T. rangeli) produce metacyclic stages in the salivary glands of its triatomine bug vectors, and therefore are not strictly speaking stercorarians (although transmission via faeces also occurs).
Type species: Trypanosoma rangeli Tejera, 1920. Note: T. rangeli is restricted to South America and has a wide mammalian host range including humans, for which it is non-pathogenic; the only known vectors are triatomine bugs of the genus Rhodnius. While another member of the subgenus, T. conorhini Donovan, 1909, is found worldwide in rats and Indonesian primates and is transmitted by triatomine bug T. vespertilionis Edmond and Etienne Sergent, 1905 infecting bats is widely distributed in Africa and Europe, where it is transmitted by Cimex spp. Sequence data are also available for T. teixeirae Barbosa et al., 2016 found in the blood of Australian flying foxes. ○ 'mammalian subgenera (salivaria)' (monophyletic) ▪ Subgenus Duttonella Chalmers, 1908 Diagnosis: Small to medium-size trypomastigotes (21-26 µm) with large and usually terminal kinetoplast, small to rounded posterior extremity and long free flagellum. Development in the insect vector is confined to the proboscis and the adjacent cibarial pump.
Type species: Trypanosoma vivax Ziemann, 1905; a causative agent of souma in cattle and other ungulates in Africa and South America (following its import from western Africa).
Note: Although only one described species has been formally assigned to this subgenus, phylogenetic analyses revealed a complex of species related to T. vivax [242]. Previously described *T. Note: T. congolense is further split into three types/ 'subspecies' (Savannah, Forest and Kilifi) that are, arguably, sufficiently different to warrant species status due to different virulency for domestic animals. Other species infect ungulates and monkeys (T. simiae Bruce et al., 1909 which is represented by two 'good' species) or were detected in tsetse flies like T. godfreyi McNamara, Mohammed and Gibson, 1994 and several unnamed species [243,244]. ▪ Subgenus Pycnomonas Hoare, 1964 Diagnosis: Small trypomastigotes (8-20 µm) with very short pointed posterior extremity, small subterminal kinetoplast and short free flagellum. Vector development takes place in the midgut and salivary glands of tsetse flies (Glossina spp.).
Note: Although this subgenus contains only one described species, two unnamed species were found in tsetse flies [243] and in a wide variety of domestic and free-living ungulates [244]. ▪ Subgenus Trypanozoon Lühe, 1906 Diagnosis: Pleomorphic trypomastigotes represented by long slender (mean length 30 μm) with long free flagellum and short stumpy forms (mean length 18 µm) with no free flagellum; both have a small subterminal kinetoplast.
Note: According to the life cycle, transmission mode, vectors, vertebrate hosts and clinical manifestations, five (sub)species are recognized and widely accepted: Trypanosoma brucei brucei (in ungulates, the causative agent of nagana in cattle; transmitted by tsetse that restrict its distribution to sub-Saharan Africa), T. brucei rhodesiense (causative agent of acute sleeping sickness in humans; game animals and livestock are primary reservoir; vectored by tsetse, sub-Saharan Africa), T. brucei gambiense (chronic sleeping sickness in humans; some domestic animals are reservoir; vectored by tsetse, sub-Saharan Africa), T. brucei evansi (causative agent of trypanosomiasis in camels, horses, cattle, buffalo, dogs and pigs, called surra in Africa and Asia and murrina in South America; transmitted mechanically by blood-sucking insects and vampires) and T. brucei equiperdum (causes dourine in horses in Asia, Africa, South America and Europe; transmitted sexually). The latter two subspecies are closely related and are unique in being so-called petite mutants of T. brucei [145]. ○ 'other terrestrial subgenera' ( paraphyletic) ▪ Subgenus Australotrypanum Votýpka and Kostygov, subgen. nov. Diagnosis: Morphologically highly variable trypomastigotes in the blood of marsupials and bats in Australia. Defined by 18S rRNA-based phylogenetic analyses.
Etymology: The generic name refers to the fact that trypanosomes come from hosts of the order Crocodilia.
Note: T. grayi transmitted by tsetse [251][252][253] clusters together with three recently described species, T. terena Teixeira and Camargo, 2013, T. ralphi Teixeira and Camargo, 2013 and T. kaiowa Teixeira and Camargo, 2019 transmitted by insect vectors [254,255], into a strongly supported monophyletic clade [255]. Based on morphology, *T. cecili Lainson, 1977 could also belong to this subgenus. All described crocodilian trypanosomes form the monophyletic crocodilian clade (subgenus Crocotrypanum) of the terrestrial lineage and are transmitted by insect vectors. Trypanosoma clandestinus Teixeira and Camargo, 2016, transmitted among caimans by leeches, is not related to this group and is nested within the aquatic lineage (subgenus Haematomonas) [255]. ▪ Subgenus Squamatrypanum Votýpka and Kostygov, subgen. nov. Diagnosis: Morphologically variable medium to large-sized trypomastigotes with multi-folded undulating membrane including free flagellum and kinetoplast located near the nucleus. Defined by 18S rRNA-based phylogenetic analyses.
Etymology: The unusual combination of hosts (see below) was used for the subgenus name combining the Latin name of reptiles (Squamata) and mammals (Mammalia).
Note: This clade brings together trypanosomes from diverse hosts, namely lizards, snakes, rodents and marsupials. Based on their morphology, T. lainsoni Naiff and Barrett, 2013 from rodents and T. freitasi Rêgo, Magalhães and Siqueira, 1957 from marsupials used to belong to the subgenus Megatrypanum; however, this taxonomic classification does not reflect their phylogenetic position. While T. varani Wenyon, 1908 described from a Nile monitor lizard (Varanus niloticus) in Sudan [256] and later found in a Ghanaian ball python (Python reginus) [257] represents the only Afrotropical species, other three species within this subgenus were described in American reptiles: T. serpentis Viola et al., 2009 from Brazilian snake Pseudoboa nigra [258], T. scelopori Ayala, 1970 from North American western fence lizard (Sceloporus occidentalis) [259] and T. cascavelli Pessôa and Da Biasi, 1971 from a South American rattlesnake (Crotalus durissus) [260]. The latter species also survives in the blood of Neotropical marsupials [261]. T. freitasi and T. gennarii Marcili, 2017, were described from Didelphis and Monodelphis opossums, respectively [262,263], yet their host spectrum is even broader [261]. Finally, T. lainsoni, originally described from Amazonian rodents [264], can infect South American marsupials and bats [261]. ○ 'incertae sedis species Note: Only the type species is currently assigned to this genus. The referred tree topology is strongly gGAPDH-dependent, prone to artefacts due to compositional bias in nucleotide sequences of this gene [103,[273][274][275]  • Subgenus Viannia Lainson and Shaw, 1987. Cause cutaneous and mucocutaneous forms of leishmaniasis in mammals; restricted to South America; transmitted by sandflies, in which they develop in the midgut and hindgut [127]. Type species: Leishmania braziliensis Vianna, 1911.
• Subgenus Sauroleishmania Ranque, 1973. Live in the blood cells of lizards and snakes, predominantly in the mononuclear phagocyte system, but reported also from erythrocytes and thrombocytes [276]. Transmitted by sandflies, in which they develop in the hindgut [127]. May represent a mixture of two distinct dixenous taxa, one of which is defined morphologically (intraerythrocytic trypomastigotes and/or epimastigotes in sloths) and another phylogenetically (Leishmania-like flagellates related to former L. herreri and parasitizing skin and visceral organs of various mammals as amastigotes [125,271,281]. The latter is transmitted by phlebotomine sand flies as promastigotes [125]  rootlet, short free anterior flagellum, long posterior flagellum attached to the cell body and situated in shallow longitudinal groove; subpellicular microtubules present only under groove; membranous sacs under plasmalemma; micropores; elongated mitochondrion parallel to the flagellar groove, with tubular cristae, possesses anterior dilation ('kinetoplast') with multiple osmiophilic bodies; no traces of oral apparatus [299].
Note: Assignment of this genus to kinetoplastids is not justified, since its 'kinetoplast', as judged by Feulgen staining, does not contain DNA. The presence of cortical membranous sacs, micropores and a striated rootlet [299] strongly suggest that this flagellate is a member of Alveolata Cavalier-Smith, 1991. Note: although many features of the genus are reminiscent of kinetoplastids, it was also considered to be a stramenopile [182,301]

Biology
The body of knowledge on the biology of diplonemids comprises a large number of environmental 18S rRNA sequences, few cultured and sequenced species, as well as several formally described species lacking sequence data and not available in culture. Until recently, diplonemids have been perceived as a small and unimportant group of euglenozoans. However, deep-sea sampling and extensive metabarcoding surveys in the past two decades uncovered extraordinary diversity of marine planktonic diplonemids [12,302,303]. Their discovery began with the recovery of environmental 18S rRNA sequences from deep-sea planktonic and hydrothermal samples, which together formed a novel well-supported clade, sister to Diplonemidae [304,305]. Subsequently, a diplonemid-focused study of several oceanic regions uncovered considerable diversity within the new clade by amplifying 18S rRNA, designated as deep-sea pelagic diplonemids (DSPD) I clade [302]. The same study identified another small lineage called DSPD II.
A breakthrough came with a V9 region metabarcoding survey by the Tara Oceans expedition, which uncovered remarkable abundance and diversity of DSPD in the tropical and subtropical sunlit ocean, expanding the number of potential diplonemid species to over 12 300 [303]. Further analysis of combined datasets from photic and mesopelagic zones identified as many as approximately 45 000 diplonemid species, thus qualifying them among the most species-rich planktonic eukaryotes in the ocean [12]. In the most recent study, extended with smaller datasets from the Arctic, Adriatic Sea and anoxic Cariaco Basin, the number of species increased to approximately 67 000, designating diplonemids as the most diverse and fifth most abundant eukaryotic clade [48]. However, fluorescence in situ hybridization studies or those based on the V4 region of 18S rRNA reported significantly lower abundance [306][307][308].
Analysis of extended Tara Oceans datasets identified that 97% of diplonemid diversity is confined to the DSPD I clade, or eupelagonemids, whereas classic diplonemids (or Diplonemidae), hemistasiids and DSPD II accounted for 1% each [12,48]. The distribution of eupelagonemids showed clear depth stratification: although their sequences were recovered from the surface water down to the abyssopelagic zone [309], royalsocietypublishing.org/journal/rsob Open Biol. 11: 200407 they are most diverse and abundant in the mesopelagic zone. However, multiple lineages of eupelagonemids show cosmopolitan distribution without a clear biogeographic pattern and a rather weak relation to abiotic factors [310]. Eupelagonemids generally show preference for dark and moderately oxygenated environments, but were occasionally detected under anoxic conditions [48]. It is still not known what drives the high diversity of eupelagonemids given the relative homogeneity of the physico-chemical conditions in the deep ocean, especially in dysphotic and aphotic zones. It has been suggested that different species might use different nutrient resources [302]. Indeed, the co-occurrence analyses showed very few obvious patterns of interaction with other components of marine plankton, among which are positive correlations with parasitic dinoflagellates and stramenopiles as well as with bacteria and bacterivorous stramenopiles, indicating possible bacteriovory and parasitic lifestyle of some eupelagonemids [12] ( plate E, [67][68][69][70][71][72][73][74][75][76].
Diplonemids have been occasionally reported from freshwater ecosystems, such as the case of Rhynchopus amitus [318] and Diplonema ambulator from a freshwater aquarium [319]. Subsequently, metabarcoding approach identified a low number of diplonemids in geographically isolated deep lakes, such as Baikal [320], and lakes in Japan [321], Switzerland and the Czech Republic [322]. Further systematic metabarcoding screening of lakes using diplonemid-specific primers might uncover so far overlooked diversity of freshwater diplonemids.
The fourth diplonemid clade, Hemistasiidae, is represented by several hundred species, so far found in the photic zone [12]. Hemistasia-like flagellates have been frequently detected in coastal waters of North and Baltic seas, Mediterranean, the Sea of Japan and around Australia and Antarctica [313,315,[323][324][325], pointing to their cosmopolitan distribution from cold to tropical regions. All hemistasiids were described from planktonic samples, except for Artemidia motanka isolated from an aquarium [311] and a hemistasiid associated with shallow-water sediments [313].
Classic diplonemids and hemistasiids are exclusively heterotrophic organisms, mostly known as eukaryovores, and displaying a wide array of lifestyles, such as ectocommensalism, predation, scavenging and opportunistic endoparasitism. Several species were found to parasitize lobsters, clams [180,326] and plants [319], while others were referred to as epibionts of crabs [327], lobsters [180], plants [300,315,316] and algal biofilms (Diplonema sp. 4 ATCC 50232). Another common trophic mode for both groups is predation and/or scavenging on planktonic algae and small invertebrates, including diatoms, dinoflagellates, green algae, prymnesiophytes and copepods [314,315,318,323]. Bacteriovory was described from only two species and seems rather uncommon [300,328].  Type species: Eupelagonema oceanica Okamoto and Keeling, 2019. Note: DSPD II (deep sea pelagic diplonemids II)-small planktonic clade, well-supported phylogenetically, known exclusively from sequences of the V9 region of 18S rRNA [12], without cultured or formally described representatives; morphology and ultrastructure not known.

Biology
Since most of the 18S rRNA phylogenies placed symbiontids either within euglenids, or as a sister clade to them (e.g. [170,329,330]), they are often regarded as derived euglenids (e.g. [329,331]). For those reasons, here we discuss the biology of euglenids and symbiontids together. However, a recent phylogenomic reconstruction [8] placed the latter group as a sister to diplonemid-kinetoplastid clade (Glycomonada), suggesting that they should be treated as a separate group within Euglenozoa.
It has been demonstrated that both phagotrophic (e.g. Distigma) and photosynthetic euglenids (e.g. Euglena gracilis; plate F, 85) are remarkably tolerant to various kinds of pollution with heavy metals such as cadmium, chromium or lead, as well as capable to remove these ions from the environment, making these protists potentially suitable for use in bioremediation of heavy metal-rich industrial wastewater [344,345]. The genetic background of heavy metal resistance in euglenids has been examined in detail only in the genus Peranema. Interestingly, although the investigated Peranema sp. strain exhibited the capability for efficient removal of cadmium from wastewater samples, the study revealed that it possesses genes responsible for resistance to a variety of other heavy metals, but not cadmium [346]. Euglenophytes have also been found in waters polluted with diesel oil [347], phenol [348], herbicides and insecticides [349,350], and can survive in highly radioactive water [351]. Some Euglenophytes are also extremophiles (e.g. Euglena mutabilis), as they tolerate very high salinity [352] in extremely acidic environments [353] or in hot mud pools [354].
Euglenids and symbiontids are predominantly free-living, exhibiting a remarkably wide range of nutrition modes, including phagotrophy, osmotrophy and photoautotrophy. Phagotrophic euglenids consume bacteria or microbial eukaryotes, and their prey size correlates with the euglenids' cell size and flexibility. Some phototrophs are capable of pinocytosis, or even phagotrophy of algae in the case of the deepbranching phototroph Rapaza [355] ( plate F, 87). Symbiontids are marine heterotrophs, presumably phagotrophs, as suggested by their ultrastructure. Additionally, the bacteria on their surface probably exchange metabolites with the hosts' mitochondria-related organelles, and it is also possible that they provide a food source for the symbiontids [7].
In contrast with the dominant free-living euglenids, there is an assemblage of eight heterotrophic genera incertae sedis (Michajlowastasia, Parastasiella, Dinemula, Paradinemula, Mononema, Ovicola, Naupliicola and Embryocola) that exhibit obligate parasitic lifestyles. Although their host range is rather narrow, encompassing exclusively freshwater, free-living copepods (specifically the eggs, larvae and digestive tracts of adults), their geographical range spans across the eutrophic freshwater bodies of all continents and nearly all climate zones, covering the range of their host group. Unfortunately, the phylogeny of these eight genera remains unresolved, as virtually all studies of the parasitic euglenids, however extensive, were carried out in the pre-sequencing era [356]. Occasionally, other photosynthetic (Euglenamorpha) and heterotrophic (Heteronema) euglenids have been identified in the gastrointestinal tracts of a very wide range of vertebrate and invertebrate animal hosts; however, it remains disputed whether they are symbionts or parasites [357,358]. Several species of euglenophytes (mainly genus Colacium) have been recognized as parasites or, more likely, epibionts of zooplankton [359][360][361][362][363], while others (e.g. Euglena mutabilis, Trachelomonas hispida) have also been identified in the traps of carnivorous plants, such as Genlisea [364] or Utricularia [365]. Whether euglenophytes are prey, accidental inhabitants, or a part of the specialized community of the carnivorous plants' traps, remains unresolved.
As most aquatic microbial eukaryotes, euglenids and symbiontids are considered cosmopolitan. Our understanding of their distribution is hampered by the limited number of environmental sequencing projects focused on those groups. Despite clear microscopical evidence of phagotrophic euglenids in sediments and phototrophic euglenids in ponds, they are suspiciously rare in environmental sequencing datasets [166,[366][367][368]. It was suggested that euglenids' 18S rRNA is divergent and often longer than the typical eukaryotic one [369,370], and the universal primers are not working efficiently for euglenids. Environmental sequences might be, however, obtained with specific primers designed for a certain group [371]. The same as for diplonemids, the V9 region of 18S rRNA seems to be a more suitable metabarcoding marker, and phototrophic euglenids have been surveyed in the environmental sequences from the TARA Oceans dataset and OSD dataset [339]. Although euglenophytes are overall quite rare in the marine plankton, their distribution is quite broad in the global ocean, with preference for the coastal upwelling zones, where the nutrient availability is higher than in other parts of the ocean [372]. Symbiontids have been investigated in several environmental 18S rRNA-based surveys in different geographical regions, suggesting their cosmopolitan distribution [166,[373][374][375][376]. Euglenids' synapomorphy is a pellicle build of proteinaceous strips beneath the plasma membrane. The strips can be fused together in certain genera; otherwise, the organisms are capable of characteristic 'euglenoid motion' also known as 'metaboly'. Another characteristic feature, shared with Kinetoplastida, is the presence of flagella inserted at the base of the flagellar pocket, and the flagella are conspicuously thickened due to the presence of paraxonemal ( paraflagellar) rods. Additionally, the main storage polymer of most euglenids is paramylon, a distinctive β-1,3-glucan.
Note: Historically, species belonging to class Euglenida were described according to the rules of the ICZN or the ICN due to the presence of photosynthetic organisms within this ancestrally non-photosynthetic group. This has created some confusion, as several taxa bear two different, equally valid names. Moreover, many clades are unstable and assigning a rank and a name to them would be unproductive, since these names are likely to become obsolete very soon. The taxonomy proposed here is a consensus between a strictly organized taxonomy for stable clades and informal nomenclature. Additionally, Euglenophyceae and their subordinate taxa are treated as algae and classified according to ICN (with alternative names consistent with ICZN provided in notes), while all other subordinate taxa of Euglenida are classified according to ICZN (with alternative names consistent with ICN provided in notes).
• Clade Olkaspira Lax and Simpson, 2020. This robustly supported monophyletic clade includes organisms with pellicle composed of S-shaped proteinaceous strips with overhangs, and chisel-shaped feeding apparatus (if the Tree F. Euglenids and symbiontids. A tree summarizing multiple phylogenetic reconstructions, primarily based on 18S rRNA gene sequences, genes encoded by plastid genomes (Euglenophyceae), and a set of nuclear protein-coding genes retrieved from transcriptomes. Polyphyletic genera are marked with an asterisk (*). Lineages with highly unstable position are marked with a dotted line. Possible paraphyly of clades has been further described in the section on Euglenids taxonomy.  Marin and Melkonian, 2003. Cells with one emergent flagellum and one vestigial within the cell; feeding by phototrophy or secondarily by osmotrophy, mostly freshwater [3]; 18S rRNA gene has C in the first position of the Helix 7/8 spacer [381]; introns highly abundant in the plastid genome (always more than 28 and usually more than 51); intron maturase mat1/ycf13 always present in the plastid genome, with mat2, mat5 or both of them usually present [384].
• Family Euglenaceae Dujardin, 1841 emend. Kim et al., 2010. Solitary or colonial, mostly free-living, but some inhabit the digestive tracts of animals; usually possess one emergent flagellum and one non-emergent; some may possess mineralized external shells (lorica); size, number and presence of pyrenoids in chloroplasts varies with the species [3]; ribosomal operon may be present in the plastid genome in one copy or more, but never as two identical inverted repeats [384]. Note: valid name under ICZN is Euglenidae Dujardin, 1841. ▪ Genus Colacium Ehrenberg, 1834. Solitary or colonial cells with envelopes that also form stalks for surface attachment [385]; often sessile (epizoic; attached to copepods) [386]; ribosomal operon present in the plastid genome in one full and one incomplete copy with the same orientation [387].  [381]; ribosomal operon present in the plastid genome in one copy [390]. Type species: Euglena pyrum Ehrenberg, 1832 (= Monomorphina pyrum). Type species and other species available in culture collections; 18S rRNA of multiple species and two full plastid genome sequences (M. aenigmatica and M. parapyrum; non-type species) are available. ▪ Genus Strombomonas Deflandre, 1930. Cells of variable shape and size with discoid or flat chloroplasts with pyrenoids and smooth lorica without collar [392]; ribosomal operon present in the plastid genome in one full and one incomplete copy with opposite orientation [387].  [395] or four flagella composed of longer and shorter pairs [396] (plate F, 86); may possess epi-or endobiotic bacteria [397]; mostly psychrotolerant or psychrophilic [342,396]; 18S rRNA gene possesses a CA insertion after the second position in the loop of Helix 18 [381]; ribosomal operon present in two copies with opposite orientation in the plastid genome, but one copy may be split [170,381]. collection; 18S rRNA sequence of the same species is available. Note: this genus probably encompasses the organism described as Helikotropis okteres [406], as its existence as a separate entity is not supported by any distinguishing morphological feature or molecular data [392]. ▪ Genus Menoidium Perty, 1852. Rigid, flattened and elongated cells with fused pellicle and one emergent flagellum without hairs [407]; microtubule scroll present [404] [378]. Only a recent multigene phylogeny resolves them as a monophyletic sister clade to Euglenophyceae, but with very weak support [8]. Regardless of the phylogenetic uncertainties, these organisms have been informally referred to as 'peranemids' due to their common morphological traits. ▪ Genus Peranema Dujardin, 1841. Biflagellate, highly metabolic cells with longer and thicker anterior flagellum and protruding feeding apparatus, capable of cutting into other cells and sucking in its contents ( plate F, 89); can and will attempt to feed on anything, including bacteria, yeast, microalgae, ink, raw starch and other euglenids [408]; very rapid response to light by rhodopsin-mediated phototaxis [409]; in recent phylogeny resolved as monophyletic [378].  [300,378]; polyphyletic genus, with two strains (ABLN1 and WBF1) branching separately, with weak support, from the otherwise monophyletic sister clade to Teloprocta [378]; resolved as monophyletic in a recent multigene phylogeny; however, the two divergent strains mentioned above were not included in the analysis [8].

Viruses in Euglenozoa
Viruses are the most abundant and widespread life form on our planet. During several billion years of coevolution, viruses have developed specific mechanisms allowing them to infect virtually any cellular organism [426]. It was estimated that viruses lyse about 20% of oceanic protists daily and, thus, play a major role in regulating the Earth's biogeochemical cycle [427]. Euglenozoa are no exception to this rule, although viral diversity was thoroughly investigated only in kinetoplastids [428]. The reason for such discrepancy is obvious-this is by far the best-studied group, which includes several parasites of medical or economic importance [10]. There is no doubt that representatives of Euglenida, Diplonemea and Symbiontida can be infected by viruses, but this has not been verified experimentally. Kinetoplastids possess DNA and RNA viruses. The only documented case of a DNA virus is the one infecting freeliving Bodo saltans [429]. This Bodo saltans virus belongs to the family Mimiviridae and its genome of about 1.39 Mb is among the largest described genomes of giant viruses. The functional role this virus may play in the biology of bodonids remains to be elucidated, but the plethora of acquired adaptation traits (such as the mechanism to facilitate membrane fusion, interference competition, contracted translation machinery and inflated genome with numerous genome rearrangements) makes this virus an interesting model for future studies. Also of note is that the abundance of such nucleo-cytoplasmic large DNA viruses was estimated at 10 4 -10 5 genomes ml −1 in the photic zone and 10 2 -10 3 genomes ml −1 of water in the oxygen minimum zone of the World Ocean [430].
The situation with RNA viruses is more complex. Here, we will only discuss viruses with known genetic structure and will not cover older reports of the mere presence of virus-like particles (reviewed by Grybchuk et al. [428]). The best-studied cases are of Leishmania RNA viruses (Leishmaniavirus spp., LRVs of the family Totiviridae). Discovered in the late 1980s in representatives of the Leishmania subgenus Viannia [431], the very first Leishmaniavirus LRV1 was sequenced in the early 1990s [432], and its biological role was uncovered about 20 years later [433]. Its presence is linked to the increased metastatic potential, parasite burden, immune response in mouse models of leishmaniasis and frequent treatment failures [434,435]. Notably, Old World leishmanias L. (Leishmania) possess a phylogenetically related Leishmaniavirus, LRV2, which is widespread in isolates of L. major [436,437], but its role in the disease progression is unknown. The phylogenies of viruses and their respective hosts are mainly congruent, suggesting long-term coevolution [438]. In addition to Leishmania, representatives of Leishmaniavirus LRV3 and LRV4 have been documented in another group of trypanosomatids, Blechomonas spp. [439]. These viruses have probably been acquired from Leishmania during co-infections.
The most successful group of viruses infecting trypanosomatids are bunyaviruses (LBVs, proposed family Leishbunyaviridae). They infect multiple Crithidia and Leptomonas spp. [440,441], Leishmania (Mundinia) martiniquensis [442], and at least one isolate of Phytomonas sp. [441]. The wide distribution of these viruses can be explained by their encapsulated structure, which promotes easy dispersion in co-infections.
The viral load in L. seymouri is extremely high, indicating that this virus may enhance Leishmania virulence in the case of Leishmania donovani-Leptomonas seymouri co-infections [444]. Other viruses are less widespread and, in many instances, appear to be restricted to a particular trypanosomatid host, as can be exemplified by Tombus-like viruses and Ostravirus in Leptomonas pyrrhocoris [441].
6. Endo-and ectosymbioses in Euglenozoa 6.1. Kinetoplastid endosymbionts Reports on endosymbionts in kinetoplastids are rare, but a few cases have been described in detail. Endosymbiosis between a bacterium and a trypanosomatid host occurred independently at least twice in the evolutionary history of trypanosomatids, in which neither hosts nor bacteria are closely related. The members of subfamily Strigomonadinae engaged in endosymbiotic relationship with Ca. Kinetoplastibacterium spp., representatives of Alcaligenaceae family (Burkholderiales; β-proteobacteria). Three genera of Strigomonadinae-Angomonas, Strigomonas and Kentomonas-are considered to share an endosymbiont-bearing ancestor, in which the reductive evolution of the endosymbiont genome occurred prior to the radiation of the host genera [274,445]. Each member of Strigomonadinae carries a different Kinetoplastibacterium species, which co-evolved together with its host [285]. Another trypanosomatid, Novymonas esmeraldas (most closely related to Leishmania), established endosymbiosis with Ca. Pandoraea novymonadis, representing another family (Burkholderiaceae) of Burkholderiales [275].
The presence of endosymbionts in all these cases likely compensates for the inability of their hosts to synthesize certain metabolites, such as haem, nucleotides, and several amino acids and vitamins, which are provided by the bacteria [275,285,445]. Both P. novymonadis and Kinetoplastibacterium spp. feature genomes, which are strongly reduced compared to related free-living β-proteobacteria, nevertheless preserving genes necessary for nutritional provisioning of their hosts [195]. Both endosymbiotic associations are permanent with bacteria being transmitted vertically; however, the association between Kinetoplastibacterium and Strigomonadinae is considered more ancient than that of N. esmeraldas and P. novymonadis [89,195]. The latter partnership is characterized by the lack of stringent control over the number of bacteria, less extensive genome reduction, higher GC content, and the presence of TCA and amino acid synthesis pathways [89].
Outside of these two systems, the presence of endosymbiotic bacteria was reported from a free-living freshwater kinetoplastid Bodo curvifilus [446], while the recently studied endosymbionts of Bodo saltans have been assigned to Paracaedibacter, with possible role in defensive endosymbiosis [447]. Finally, the trypanosomatid Phytomonas borealis isolated from the midgut of spiked shieldbugs also harbours endosymbionts [448], although their taxonomic identity and function remain unknown.

Diplonemid endosymbionts
Diplonemids are known for establishing symbiosis with members of Holosporaceae and Rickettsiaceae families, which are exclusively parasitic/endosymbiotic lineages of α-proteobacteria [449]. At present, there are only two reports on endosymbionts in diplonemids: Holosporaceae bacteria inside two Diplonema species [450], and a hemistasiid Namystinia karyoxenos with a Rickettsiaceae endosymbiont [325]. While D. aggregatum and N. karyoxenos contain a single endosymbiont, D. japonicum harbours two species of bacteria from closely related genera, the genomes of which have been sequenced, assembled and analysed [451]. They are severely reduced with similar gene content retained, and lack all energy metabolism pathways, including glycolysis, pentose-P pathway, the TCA cycle and oxidative phosphorylation. Although complete synthesis pathways for amino acids or vitamins are absent, the nutritional role of the endosymbionts cannot be ruled out due to the large number of proteins without known functions. A large portion of their highly reduced genomes is dedicated to secreted proteins that are possibly involved in manipulation of the host metabolism. Similar to Holosporaceae and Rickettsiaceae in other protist hosts, the role of the diplonemid endosymbionts is not clear. However, it was hypothesized that due to the presence of various secretion/toxin systems, the endosymbionts might take part in Defence against bacterial pathogens [451].

Symbiontid symbionts
As suggested by their very name, the distinctive trait of the Symbiontida is their capability of forming permanent, probably obligatory symbiotic relationships [7,41]. Thus far, three distinctive kinds of epibiotic bacteria have been described to thrive on the surface of symbiontid cells: rodshaped, sulfide-oxidizing ε-proteobacteria associated with symbiontid genera Calkinsia and Bihospites [41], extrusive apparatus-bearing cocci with strong resemblance to hypotrich ciliate-associated Verrucomicrobia, endemic to the genus Bihospites [329], and magnetotactic Deltaproteobacteria, associated with multiple unclassified environmental strains of symbiontids [452].
Although the role of magnetosome-bearing, but nonmotile δ-proteobacterial symbionts is clearly to provide their hosts with magnetotaxis [452], the functions of other microorganisms in their relationships with symbiontids seem to be more complex. It has been suggested that the ε-proteobacterial symbionts detoxify the local surroundings to limit the inhibitory effect of sulfide on the cellular respiration of the symbiontids [41], while the Verrucomicrobialike bacteria provide their hosts with a Defence mechanism against predators [329]. Additionally, the epibiotic bacteria can be used by their hosts as an auxiliary food source [329,452]. In exchange, the eukaryotic hosts' role is to provide their epibionts with various metabolites, such as hydrogen, as all symbiontids possess hydrogenosomes [7], and to serve as efficient means of transport along the oxycline in the deep-sea environment [41]. It remains uncertain, however, whether the metabolic coupling between the symbiontids and ε-proteobacteria is limited to the outward hydrogen flux, and, perhaps more importantly, if any metabolic exchange occurs between the Verrucomicrobia-like bacteria and their symbiontid hosts.

Euglenid symbionts
Curiously, the observable affinity of the symbiontids towards prokaryotic partners is, to some extent, shared by their royalsocietypublishing.org/journal/rsob Open Biol. 11: 200407 postulated close relatives-the euglenids. In their case, the tight relationships with bacteria are not as widespread, as a majority of the described euglenid genera have never been observed to harbour any symbionts, but the diversity of these relationships may be substantially greater [397]. Unfortunately, our knowledge of the euglenidbacteria associations remains superficial due to the fact that they have mostly been reported in the pre-genomics and transcriptomics era [419,453,454]. Nonetheless, it is undeniable that these relationships are widespread, as they involve both phagotrophic (genera Petalomonas and Dylakosoma; [419,455]) and photosynthetic euglenids (genera Euglena, Phacus, Lepocinclis and others; [454]), and of rather diverse nature, as euglenids have been observe to harbour bacteria within their cells [389,454], on their surface [419,456], or even both [397]. What is more, a single heterotrophic euglenid species (Anisonema platysomum) has been observed to harbour magnetosomes. It remains unclear whether this typically prokaryotic trait has been acquired by Anisonema from magnetotactic bacteria (similar to those associated with symbiontids), or evolved independently [457].
As indicated by the so far most elaborate studies of euglenid-bacteria associations, involving a phagotrophic euglenid Petalomonas sphagnophila and a strain of photosynthetic Eutreptiella sp., it is evident that euglenids are capable of harbouring multiple distinct bacterial symbionts simultaneously [397,455]. Petalomonas sphagnophila from Canadian peatlands has been observed to carry six different bacteria within its cells, namely two strains of Rickettsiales, one representative of Firmicutes, one γ-proteobacterium, one δ-proteobacterium and one enigmatic, pigmented prokaryote with unidentified affiliation [455]. Moreover, Eutreptiella from the Long Island Sound possesses epibionts classified as Roseovarius, Oceanicaulis (Alphaproteobacteria) and Marinobacter (Gammaproteobacteria), as well as an endobiotic representative of Rickettsiales, though phylogenetically distant from those associated with P. sphagnophila [397]. Unfortunately, except for the hypothesis that the epibiotic bacteria of Eutreptiella supply their host with vitamin B12, well supported by cultivation experiments and transcriptomic data, very little is known about the nature and purpose of the relationships between bacteria and the two aforementioned euglenid hosts [397,455]. In fact, it is uncertain whether these associations are symbiotic at all, especially considering that Rickettsiales are common intracellular parasites of a vast variety of eukaryotes [458].

Note added in proof
After the acceptance of this paper for publication, a taxonomic description of two new kinetoplastid flagellates, marine Papus ankaliazontas and freshwater Apiculatamorpha spiralis, has been published [459]. These new taxa represent free-living representatives of the order Prokinetoplastida. The available 18S rRNA gene-based phylogeny does not allow estimating reliably the relationships of these flagellates with the previously characterized genera Perkinsela and Ichthyobodo due to low statistical supports [459]. Moreover, the inferred position of these new taxa contradicts with the well-supported topology of the recently published phylogenomic tree, which, however, does not include Ichthyobodo [30]. Thus, it is currently premature to classify the two new genera and they are considered as Prokinetoplastida incertae sedis. Below is a short characterization of these forms. P. ankaliazontas: free-living, solitary, eukaryvorous; two unattached flagella, the anterior one with thin undulate mastigonemes; flagellar pocket connected to oblique groove; pronounced rostrum with apical cytostome, tubular cytopharynx supported by prismatic microtubular rod; trichocysts at the anterior end.
A. spiralis: free-living, solitary, eukaryvorous; two unattached flagella the anterior one with fine fiber layer; cell surface with spherical and lamellate scales; pronounced rostrum with apical cytostome, tubular cytopharynx supported by prismatic microtubular rod; trichocysts at the anterior end.
Data accessibility. This article has no additional data. Authors' contributions. J.L. wrote the abstract and Introduction, conceived the idea for this manuscript and supervised the work. A.Y.K. prepared plate D and trees B-D, and wrote the section on Kinetoplastea except the section on taxonomy of Trypanosoma, which was written by J.V. J.V. also prepared plates B and C. V.Y. wrote the section on Viruses in Euglenozoa. D.T. and J.L. wrote the section on Diplonemea and subsections on diplonemid and kinetoplastid endosymbionts, and prepared plates A and E and tree E. A.K. and K.M. wrote the section on Euglenida and Symbiontida, and subsections on euglenid symbionts and symbiontid symbionts, and prepared plate F and trees A and F. All authors reviewed, corrected and accepted the manuscript.