Isolation of Novel Trypanosomatid, Zelonia australiensis sp. nov. (Kinetoplastida: Trypanosomatidae) Provides Support for a Gondwanan Origin of Dixenous Parasitism in the Leishmaniinae

The genus Leishmania includes approximately 53 species, 20 of which cause human leishmaniais; a significant albeit neglected tropical disease. Leishmaniasis has afflicted humans for millennia, but how ancient is Leishmania and where did it arise? These questions have been hotly debated for decades and several theories have been proposed. One theory suggests Leishmania originated in the Palearctic, and dispersed to the New World via the Bering land bridge. Others propose that Leishmania evolved in the Neotropics. The Multiple Origins theory suggests that separation of certain Old World and New World species occurred due to the opening of the Atlantic Ocean. Some suggest that the ancestor of the dixenous genera Leishmania, Endotrypanum and Porcisia evolved on Gondwana between 90 and 140 million years ago. In the present study a detailed molecular and morphological characterisation was performed on a novel Australian trypanosomatid following its isolation in Australia’s tropics from the native black fly, Simulium (Morops) dycei Colbo, 1976. Phylogenetic analyses were conducted and confirmed this parasite as a sibling to Zelonia costaricensis, a close relative of Leishmania previously isolated from a reduviid bug in Costa Rica. Consequently, this parasite was assigned the name Zelonia australiensis sp. nov. Assuming Z. costaricensis and Z. australiensis diverged when Australia and South America became completely separated, their divergence occurred between 36 and 41 million years ago at least. Using this vicariance event as a calibration point for a phylogenetic time tree, the common ancestor of the dixenous genera Leishmania, Endotrypanum and Porcisia appeared in Gondwana approximately 91 million years ago. Ultimately, this study contributes to our understanding of trypanosomatid diversity, and of Leishmania origins by providing support for a Gondwanan origin of dixenous parasitism in the Leishmaniinae.


Introduction
The success of Leishmania species, the complexity of their dixenous life cycle, and the intricacy of their host-parasite interactions implies a relationship between host, parasite and vector that has evolved over millions of years, certainly predating the appearance of humankind. Evidence for this ancient origin was first identified in the form of Paleoleishmania proterus; a trypanosomatid discovered in a fossilised Palaeomyia burmitis sand fly that became trapped in Burmese amber approximately 100 million years ago (MYA) [1]. A second fossilised specimen of the extinct sand fly Lutzomyia adiketis contained a trypanosomatid parasite assigned the name Paleoleishmania neotropicum [2]. This specimen was preserved in amber from the Dominican Republic and was dated at 20 to 30 million years old [2]. While these findings provide insights into the ancient origins of Leishmania, the evolutionary and biogeographical history of this genus remains a hotly debated topic, and multiple theories have been proposed [3][4][5][6].
The Palaearctic origins theory suggests that Leishmania originated in the Old World and dispersed to the New World via the Bering land bridge which was open during the Eocene epoch [6][7][8]. Amastigotes of the~100 million year old Paleoleishmania proterus were observed in Cretaceous reptilian blood [9], supporting that the reptile-infecting Sauroleishmania subgenus evolved first in the Palearctic. However, this requires that the Sauroleishmania form a sister clade to all other Leishmania species [3,7], and implies that adaptation to mammals, possibly murid rodents, occurred later when reptiles declined during the global cooling episode that denotes the Eocene to Oligocene transition [6,8,10]. Alternatively, the Neotropical origins hypothesis suggests Leishmania appeared in the Neotropics between 34 and 46 MYA and was dispersed to the Nearctic by rodents (i.e. porcupines) via the Panamanian land bridge [11]. The parasites were then dispersed further, from the Nearctic to the Palaearctic via the Bering land bridge [3,6].
The Multiple Origins hypothesis, also known as the Neotropical/African Origins hypothesis [6], considers the origins of the Euleishmania, comprising the Leishmania, Viannia, and Sauroleishmania subgenera, and the Paraleishmania [7] which presently includes Endotrypanum and the newly established genus, Porcisia Shaw, Camargo and Teixeira, 2016 [12]. This hypothesis supposes that the Euleishmania and Paraleishmania existed as separate lineages prior to the breakup of Gondwana. Upon the opening of the Atlantic Ocean, the Euleishmania evolved into the Sauroleishmania and Leishmania subgenera in the Old World, and the Viannia subgenus evolved from the Euleishmania that remained in the New World [7]. This theory also supposes that an ancestor of the few known Neotropical Leishmania (Leishmania) species was later dispersed from the Old World to the New World via the Bering land bridge [3,6]. The Supercontinents hypothesis represents a variation of the Multiple Origins theory, and proposes that the Euleishmania and Paraleishmania diverged approximately 90 to 100 MYA, and that an ancestor to Leishmania, Endotrypanum and Porcisia evolved from a monoxenous trypanosomatid on Gondwana between 90 and 140 MYA [3]. This hypothesis was discussed several years ago by Yurchenko et al. [4], though more recently explored by Harkins et al. [3], who also provided phylogenetic support. Inclusion of an Australian Leishmania species in phylogenies from that study also allowed calibration of time trees at a speciation event (a node) that likely arose when Australia became completely separated from South America, via Antarctica, approximately 40 MYA [3]. However, the separation of these continents was a highly protracted event, beginning during the early Cretaceous period and resulting in a large rift valley between Australia and Antarctica as early as 125 to 105 MYA [13]. Consequently, calibration of this node at 40 MYA represents a minimum time point for the vicariant event that separated the Australian Leishmania parasite from its ancestors in the Neotropics.
There has been an intense effort amongst trypanosomatid taxonomists in recent years to increase our knowledge of trypanosomatid diversity and better understand the evolutionary relationships between members of this important group of parasites [12,14,15]. These endeavours have required detailed molecular and morphological characterisation of newly isolated species to avoid misclassification and subsequent confusion for later investigators [15]. This work has led to several new developments, including establishment of new genera and the reassignment of "old" parasites to different genera [12,[14][15][16][17][18][19]. Despite these recent advances, knowledge of Australia's indigenous Leishmaniinae remains incredibly scarce. Extended periods of geographical isolation have resulted in Australia's unique and often peculiar fauna. Indeed, this uniqueness is reflected in Australia's native Leishmania parasite which, curiously, is thought to be transmitted in the bite of a day feeding midge (Diptera: Ceratopogonidae), rather than a phlebotamine sand fly [20]. Given Australia's unique fauna, surveying its insects for endogenous trypanosomatids could contribute markedly to our understanding of trypanosomatid diversity and uncover evolutionary relationships that were previously elusive.
As a contribution to these efforts, we describe the detailed molecular and morphological characterisation of a novel trypanosomatid isolated from the Australian native black fly, Simulium (Morops) dycei Colbo, 1976. Phylogenetic analyses confirmed this parasite as a sibling species to Leptomonas costaricensis; a trypanosomatid previously isolated from a reduviid bug in Costa Rica [4]. In a recent appraisal of trypanosomatid taxonomy, Espinosa et al. [12] argued that L. costaricensis was phylogenetically distant from other Leptomonas spp. and should be placed in a separate genus. Consequently, the genus Zelonia n. gen Shaw, Camargo and Teixeira (2016) [12] was established to accommodate this organism (henceforth Zelonia costaricensis) and its nearest relatives. Accordingly, the Australian parasite isolated in this study was assigned the name Zelonia australiensis sp. nov. Assuming that the separation of Z. costaricensis and Z. australiensis occurred as a result of vicariance, when Australia and South America separated, we suggest their divergence took place between 36 and 41 MYA, at least [21]. Using this event as the calibration point for a phylogenetic time tree, the clade containing the dixenous parasites Leishmania, Endotrypanum and Porcisia i.e. the Euleishmania and Paraleishmania, was estimated to have diverged from a monoxenous ancestor in Gondwana during the mid-Cretaceous, approximately 91 MYA. Ultimately, this study contributes to our understanding of trypanosomatid diversity, and of Leishmania origins, by providing support for a Gondwanan origin of dixenous parasitism in the Leishmaniinae.

Study location and insect trapping
Insect collection was performed following approval by the University Technology Sydney Animal Care and Ethics Committee. Insect trapping was performed near the location selected by Dougall et al. [20] (Table 1, S1 File) as it was considered suitable for the isolation of other tropical trypanosomatids and would provide an opportunity to re-isolate the Australian Leishmania parasite [22], thereby confirming its persistence in the region. Note that at the time of writing, the name Leishmania 'australiensis' had been used to describe this Australian Leishmania parasite in the scientific literature [6], and in an Australian government document [23], in the absence of any formal description. Consequently, the name Leishmania 'australiensis' is a nomen nudum and is no longer available as a species name. To prevent continued use of this nomen nudum, the present study includes a formal description of this Australian Leishmania species, referred to henceforth as Leishmania macropodum sp. nov., Barratt

Insect identification
Trapped midges and flies were identified with the aid of keys and descriptions [20,[24][25][26][27]. Fly specimens were dissected and mounted using the method described by Craig et al. [28]. In some cases, DNA was extracted from flies for barcoding purposes prior to identification by morphology. A DNA extraction method described by Lawrence et al.
[29] (S1 File) was employed that conserved the exoskeleton for downstream morphological identification.

Cultivation of parasites from insects
Insects were pooled and crushed with a spatula in~200 μL of PBS. The resulting suspension was used to inoculate a Leishmania culture medium based on the medium previously described by Dougall et al. [20]. The parasite cultures obtained were initially contaminated with a Fusarium sp. fungus. As the parasite cells outnumbered the fungi, the cultures were axenised by serial dilution such that the fungi were diluted out resulting in a pure promastigote culture. To facilitate downstream promastigote counting experiments, a liquid medium was developed and optimised to establish the ideal haemoglobin content (S1 File).

Light microscopy and transmission electron microscopy
To examine the morphology of cultured promastigotes, a Leishman stain was performed (Sigma-Aldrich) on cell-dense promastigote cultures, in accordance with the manufacturer's instructions. Cell morphology was examined by oil emersion light microscopy (1000X magnification) using a Leica DM1000 microscope (Leica Microsystems). To examine their ultrastructural features, cultured promastigotes were embedded in low melting point agarose and prepared for transmission electron microscopy using standard procedures (S1 File). Following this, ultrathin sections were cut from the agarose and examined using a Hitachi H-7650 Transmission Electron Microscope (USA).

DNA extraction and Polymerase Chain Reaction (PCR)
For extraction of total DNA from parasites, approximately 1 mL of dense promastigote culture was placed in a 1.5 mL tube and the cells were pelleted by centrifugation at 300 g for 15 minutes. The supernatant was discarded and DNA was extracted from the pellet using an EZ1 DNA tissue extraction kit (QIAGEN) and a BioRobot EZ1 DNA extracting robot (QIAGEN) according to the manufacturer's instructions. The DNA was eluted in a volume of 50 μL for downstream PCR analysis. PCR primers were designed to amplify the 18S rRNA gene and three protein coding genes; the glycosomal glyceraldehyde 3-phosphate dehydrogenase (gGAPDH), RNA polymerase II largest subunit (RPOIILS), and heat shock protein 70 (HSP70) genes ( Table 2). To generate PCR products from insects for barcoding purposes, a set of previously published primers were used to amplify fragments of the cytochrome C oxidase subunit I (COI) and II (COII) genes, the 18S rRNA gene, and the 28S rRNA gene ( Table 2). Each PCR was prepared using reagents provided in the BIOTAQ PCR Kit (Bioline) (S1 File). The PCR products were subjected to electrophoresis on 2% agarose gels stained with GelRed, and visualised under UV light.

Sequencing of PCR products
The PCR products were excised from agarose gels using a sterile scalpel blade. Amplicons were extracted from gel slices using a QIAquick Gel Extraction Kit (QIAGEN) according to the manufacturer's instructions. Sequencing was performed by the service provider Macrogen (South Korea) on an ABI 3730XL capillary sequencer. Ambiguous, low quality bases were manually trimmed from the ends of sequences which were then assembled using CAP3 [30]. Sequences generated from PCR amplicons of gGAPDH and RPOIIL displayed several 'dualpeaks', where two bases were superimposed at the same base position along the sequence. Furthermore, the multi-copy ITS1 DNA sequences of trypanosomatids can differ between copies, making direct sequencing of ITS1 amplicons difficult [31]. Cloning of these amplicons was performed to overcome this issue, so that individual clones could be sequenced. These amplicons were cloned using a TOPO TA cloning kit for sequencing (Thermo Fisher Scientific). Cloning reactions were prepared according to the manufacturer's instructions (S1 File), and sequencing of cloned PCR fragments was carried out directly from the purified plasmid, twice in the forward and reverse directions, by the service provider Macrogen. Sequencing was performed using the universal T3 and T7 primers (Table 2), which possess priming sites flanking the amplicon insertion site.

Phylogenetic analysis
Phylogenetic trees were constructed to infer the evolutionary relationship between this newly isolated trypanosomatid and other related parasites. S1 Table lists all GenBank accession numbers for sequences generated in this study and those published by others that were used to construct phylogenetic trees. Multiple sequence alignments were performed using the MEGA software package, version 7.0.14 [34]. Alignments were manually curated to improve accuracy, and phylogenetic analysis was performed using MEGA. Trees were inferred using three methods: the Maximum Likelihood (ML) method based on the Tamura-Nei model [35], the Minimum Evolution (ME) method [36], and the Neighbour-Joining (NJ) method [37]. For ML trees, initial trees for the heuristic search were obtained automatically by applying the

Effect of haemoglobin on growth
Promastigote growth was investigated in four liquid media differing in haemoglobin content (M0 to M3) (S1 File). Growth was observed in all media including M0 which contained no haemoglobin although the highest cell densities were observed in M3, which contained the highest haemoglobin concentration (Fig 2). In all media, promastigote growth peaked at day 3 and numbers plateaued by day 4. Promastigote numbers steadily decreased until the experiment was terminated on day 6.

Promastigote morphology
Leishman stained smears and wet preparations of cultured parasites revealed several cell morphotypes. Images of these forms are provided in Fig 3. Transmission electron microscopy performed on cultured promastigotes confirmed the presence of ultrastructural features consistent with the Leishmaniinae and similar to the descriptions of Zelonia costaricensis (Fig 4) [4].

Molecular characterisation of parasites
BLAST searches carried out on the parasite sequences generated in this study (GenBank Accessions KY273490 to KY273515) suggested the parasite was of the subfamily Leishmaniinae. The PCR-RFLP assay generated a restriction pattern for the isolate that differed when compared to that produced for the other species tested (Fig 5). Seventeen unique ITS1 DNA clones (GenBank Accessions KY273499 to KY273515), four unique gGAPDH clones (GenBank Accessions KY273493 to KY273496) and three unique RPOIILS clones (GenBank Accessions KY273490 to KY273492), were generated. The L. seymouri sequences generated in this study for gGAPDH, HSP70 and the 18S rRNA genes (GenBank Accessions KY273516, KY273519 and KY273517, respectively) were identical to Leptomonas spp. sequences already available in Gen-Bank (Accessions: AF047495, FJ226475 and KP717895, respectively), supporting the accuracy of sequences generated using this workflow. However, the RPOIILS sequence generated in this study (GenBank Accession: KY273518) differed by six bases to a previously published L. seymouri sequence which may indicate the sequence was derived from a different strain (Gen-Bank Accession: AF338253).

Phylogenetic analysis
Phylogenetic trees were constructed from concatenated alignments of 18S rDNA and gGAPDH sequences (Fig 6), and 18S rDNA, gGAPDH, RPOIILS and HSP70 sequences (Fig 7) to infer the phylogenetic relationship between this novel trypanosomatid and other related parasites. Concatenated sequence alignments were employed as these are generally considered more robust for inferring phylogenetic relationships [15]. For each alignment, phylogenies inferred using the ML, NJ and ME methods showed the same structure. Both phylogenies positioned this parasite in the subfamily Leishmaniinae, basal to the clade occupied by Leishmania, Endotrypanum and Porcisia. The phylogeny generated from the 18S rDNA and gGAPDH concatenated sequence inferred Z. costaricensis as the sibling species to this new parasite, with a bootstrap percentage of at least 99, across 1000 replicates for each phylogenetic method used (ML, NJ and ME). Based on this result and the morphological characteristics previously described, this parasite was assigned to the genus Zelonia and will hereafter be referred to as Zelonia australiensis sp. nov. Once this classification was established, a phylogenetic time tree was constructed using concatenated sequences of the 18S rDNA and RPOIILS genes, given that these phylogenetically informative sequences were available for many Leishmaniinae. The node representing the divergence of Z. australiensis and Z. costaricensis was selected as a calibration point. This node was set at 36 to 41 MYA which is the estimated time period that Australia and South America became completely separated [21], representing a minimum time for the separation of these taxa. Using this calibration point, an ancestor to Leishmania, Endotrypanum and Porcisia was predicted to have appeared approximately 91 MYA (Fig 8), inferring a Gondwanan origin for dixenous parasitism in the Leishmanaiinae subfamily [3]. Fig 8 also infers that the divergence of Z. australiensis from Z. costaricensis, and Leishmania macropodum from other Mundinia parasites occurred around the same time, just prior to the Eocene to Oligocene transition, which occurred between 33 and 34 MYA.

Discussion
The genus Leishmania includes approximately 20 species of protozoan parasite that are the etiological agents of human leishmaniais [6], an important albeit neglected tropical disease. Relative to other protozoan diseases, leishmaniasis is second in importance to malaria as a Leishmania exists on all continents with the exclusion of Antarctica, though its geographical range is focused in the tropics and subtropics [6]. Despite Australia's geographical isolation, representatives of this genus have also been found on this continent [44]. As a consequence of its wide global dispersion patterns, the biogeographical history of Leishmania has been hotly debated for decades and several hypotheses have been proposed. The Palaearctic origins theory suggests that Leishmania originated in the Old World during the early Cenozoic period [8], and was later dispersed to the Nearctic and then the Neotropics via the Bering land bridge, which was open during the Eocene epoch but eventually closed approximately 33 to 35 MYA [6,45]. The discovery of P. proterus fossilised in Burmese amber supports an Old World origin for Leishmania, though the age of the amber (100 to 110 million years old) supports an earlier Cretaceous origin [1,9], consistent with recent phylogenies [3] (Fig 8). Paleoleismania proterus were visible in the proboscis of Palaeomyia burmitis, and amastigotes were noted in reptilian red blood cells within the fly [1,9]. This led to the interpretation that a dixenous life cycle had evolved in the Leishmaniinae roughly 100 MYA in the Old World, and supported that Cretaceous reptiles were the first vertebrate hosts of the earliest dixenous Leishmaniinae [6,8,9]. However, this interpretation is not supported by current phylogenies that do not place the Sauroleishmania in a basal position or sister clade to all other Leishmania species [3,[46][47][48][49] (Fig 8).
While the fossilised forms identified within P. burmitis are compelling and undoubtedly represent an early trypanosomatid [1], inferring evolutionary relationships for protozoa based purely on morphology is precarious. Some of the forms described by Poinar and Poinar could easily represent epimastigotes of Trypanosoma spp. based on the location of the kinetoplast relative to the nucleus [1]. Trypanosoma spp. are basal to all Leishmaniinae and so a dixenous life cycle probably evolved in this genus much earlier [15]. Furthermore, Trypanosoma spp. are known to infect reptiles and some reptile-infecting trypanosomes are transmitted by sand flies [50][51][52]. Mixed trypanosomatid infections are also common in insects [53,54], which further complicates interpretation of such evidence. Additionally, it is well established that trypanosomatids have undergone substantial molecular evolution despite minimal morphological change [55]. This phenomenon has led to erroneous taxonomic assignments, even for taxa that are presently alive today [15]. Consequently, assignment of these organisms to the Trypanosomatidae based on the fossil evidence at hand is warranted [1], though classifying these organisms at any greater resolution is probably tenuous.
The Neotropical origins hypothesis proposes that Leishmania evolved in South America between 34 and 46 MYA [3,6]. Indeed, a Neotropical origin is supported by the evidence available, though the appearance of Leishmania probably occurred much earlier than the Neotropical hypothesis initially proposed (Fig 8) [3,56]. The Neotropical origins theory is also supported by the limited range of the Paraleishmania which are restricted to the New World, Inferred evolutionary relationship between Zelonia australiensis and other trypanosomatids using concatenated 18S rDNA and gGAPDH sequences. This tree was constructed using sequences from 23 trypanosomatids, aligned to a total of 1302 positions with all gaps and missing data eliminated. The structure of this tree was inferred using three statistical methods; the ML method based on the Tamura-Nei model, the ME method [36] and the NJ method [37]. The same tree structure was predicted using each method. The first value at each node is the percentage of trees in which the associated taxa clustered together using the ML method (1000 replicates). The second and third number at each node is the percentage of replicate trees obtained for the ME and NJ methods respectively, in which the associated taxa clustered together in the bootstrap test (1000 replicates) [102]. A solid diamond indicates a node that obtained a value of 100% for all three methods. An open diamond indicates a node that obtained a value of at least 99% for each method. The star highlights the phylogenetic position of Z. australiensis. The bar represents the number of substitutions per site. and are basal to all Euleishmania [3,6,16] (Figs 7 and 8). The Multiple Origins hypothesis suggests the Euleishmania and Paraleishmania evolved on Gondwana prior to the opening of the Atlantic Ocean. When Africa and South America separated, the Euleishmania in the Old World evolved into the Sauroleishmania and Leishmania subgenera, while Euleishmania in the New World evolved into Viannia [3,7]. A very ancient, African origin has been proposed for most Old World Leishmania species given their intimate relationships with certain rodent species and hyraxes; vertebrates that have a highly restricted range [7]. However, the results of this study and others suggest that the Old World Leishmania (Leishmania) parasites originated approximately 30 MYA [3] (Fig 8).
The present study supports a Gondwanan origin for dixenous parasitism in the Leishmaniinae subfamily, inferring the appearance of a common ancestor to the Euleishmania and Paraleishmania at approximately 91 MYA (Fig 8) [3]. This places the origin of the dixenous Leishmaniinae during the breakup of Gondwana when the radiation of mammals first began [57], and is within the lower limit of 90 to 140 MYA proposed recently by Harkins et al. [3]. By 90 MYA, Africa and South America had already separated. Multiple phylogenies suggest that Viannia emerged millions of years later (Fig 8) [3,11,57], implying that their divergence from other Euleishmania was not triggered by the separation of Africa and South America. The   Fig 7. Inferred evolutionary relationship between Zelonia australiensis and other trypanosomatids using concatenated 18S rDNA, gGAPDH, RPOIILS and HSP70 sequences. This phylogenetic tree was constructed using sequences from 15 trypanosomatids, aligned to a total of 2344 positions with all gaps and missing data eliminated. The structure of this tree was inferred using three statistical methods; the ML method based on the Tamura-Nei model, the ME method [36], and the NJ method [37]. The same tree structure was predicted using each method. The first value at each node is the percentage of trees in which the associated taxa clustered together using the ML method (1000 replicates). The second and third number at each node is the percentage of replicate trees obtained for the ME and NJ methods respectively, in which the associated taxa clustered together in the bootstrap test (1000 replicates) [102]. A solid diamond indicates a node that obtained a value of 100% for all three methods. The star highlights the phylogenetic position of Z. australiensis. The bar represents the number of substitutions per site. doi:10.1371/journal.pntd.0005215.g007

Fig 8. Phylogenetic Time Tree inferring the evolutionary relationships between
Zelonia australiensis and other trypanosomatids using concatenated 18S rDNA and RPOIILS sequences. This tree was constructed using sequences from 29 trypanosomatids, aligned to a total of 784 positions with all gaps and missing data eliminated. The structure of this tree was inferred using three statistical methods; the ML method based on the Tamura-Nei model, the ME method [36], and the NJ method [37]. The same tree structure was predicted using each method. The predicted minimum divergence times for each node i.e. the values outside the brackets, were predicted using the RelTime method [40]. Estimated divergence times greater than 1 MYA are rounded to the nearest whole number. The error calculated for the divergence time at each node is shown in S2 Fig. Regarding values within brackets, the first number is the percentage of trees in which the associated taxa clustered together using the ML method (1000 replicates). The second and third number is the percentage of replicate trees obtained for the ME and NJ methods respectively, in which the associated taxa clustered together in the bootstrap test (1000 replicates) [102]. An open diamond indicates a node that obtained a value of 99% or greater for each method. A solid diamond indicates a node that obtained a value of 100% for all methods. A solid circle represents nodes that obtained a value of 60% or less for each method. A solid square represents a collapsed node. The star highlights the phylogenetic position of Z. australiensis. Branches are colour coded to indicate the current dispersion pattern for the different species. Note that Leishmania infantum is also found in European countries flanking the presence of the Leishmania subgenus in the New World is often discussed as a migration from the Old World to the New [58]. However, based on current evidence, an alternate scenario is proposed. Approximately 50 MYA the climate in the northern hemisphere was tropical and a series of land bridges, shallow seas and island chains connected Europe, North America and Asia [11,59]. These land bridges were probably endemic for Leishmania and allowed movement of host and vector between the Old World and the New until approximately 35 to 33 MYA, when these bridges disappeared [45]. Paleontological evidence supports an exchange of primate and rodent species between North and South America during the same period, indicating that the Panamanian land bridge was also open [60]. The disappearance of these northern land bridges coincides with the sharp drop in temperature that signifies the Eocene to Oligocene transition [61]. This also coincides with the inferred emergence of the New World Leishmania (Leishmania) spp. approximately 30 MYA [3] (Fig 8). By 33 MYA, these once tropical northern land bridges were uninhabitable for sand flies, probably forcing the range of Leishmania and other tropical species south towards the Neotropics in the New World, and out of Northern Europe, towards Africa and South East Asia in the Old World. The presence of L. (L.) amazonensis/mexicana complex organisms in China supports this scenario [3,62].
The subgenus Mundinia Shaw, Camargo and Teixeira 2016 was recently established to accommodate members of what was previously referred to as the L. enrietti complex [12]. While Mundinia are widely dispersed, L. (M.) enrietti itself was initially isolated from guinea pigs in Brazil and is probably native to the Neotropics [63]. A related organism, Leishmania (Mundinia) martiniquensis, was later identified on the Caribbean Island of Martinique, detected in immunocompromised patients presenting with cutaneous leishmaniasis (CL) and visceral leishmaniasis (VL) [64][65][66]. Parasites of the Mundinia subgenus have since been identified in Thailand i.e. Leishmania sp. 'siamensis', as a cause of human VL, predominantly in immunosuppressed patients [67][68][69][70]. As discussed by other investigators [46], Leishmania 'siamensis' represents a nomen nudum, and is shown inverted commas here as a consequence. Leishmania 'siamensis' was detected in horses from the USA and central Europe [71,72], and in Swiss cows [73]. The geographical range of L. 'siamensis' and L. (M.) martiniquensis is known to overlap given the recent detection of L. (M.) martiniquensis in Thailand [46], resulting in misidentification in some cases [46,74]. Additionally, a unique Mundinia parasite was only recently identified as a cause of human CL in Ghana [46], though this organism is yet to be named. Leishmania (M.) macropodum is also a member of the Mundinia subgenus, and is recognised as a cause of CL in Australian native macropods [44,75].
The global dispersion pattern of Mundinia is difficult to explain, though the current range of L. (M.) martiniquensis may be related to human activities such as international shipping and trade, facilitating the movement animals i.e. livestock, companion animals and rodents, between regions that would have otherwise been non-traversable. Indeed, rats have been pivotal to the global dispersion of other parasites via this route [76]. Furthermore, Mundinia parasites are not necessarily restricted to sand fly vectors, which could facilitate their adaptation to new regions [20,22]. As a consequence of these dispersion patterns, it is difficult to infer where Mundinia originally appeared.
Current phylogenies suggest that the Ghanian parasite and L. enrietti diverged within the last 10 million years [3,46]. These species have been observed in only a few restricted regions implying that their native range is limited. Perplexingly, this suggests that these two parasites diverged long after the New World separated from Africa. During the Miocene epoch there Mediterranean basin. This time tree was calibrated by setting the node depicting the divergence of Z. australiensis and Zelonia costaricensis at 41 to 36 (mean~39) million years ago; a minimum time estimate for the vicariance event that separated these taxa.
doi:10.1371/journal.pntd.0005215.g008 was a warm period in central Europe which abruptly ended at~14 MYA, when temperatures dropped markedly to a mean annual temperature of~14.8˚C to15.7˚C [77,78]. Consequently, it is unlikely that movement of Leishmania between the Nearctic and Palearctic occurred via the northern land bridges at this time. Alternatively, dispersion of an ancestral Mundinia parasite between the Old World and the New as recently as 10 MYA may have been facilitated by far-travelling marine mammals (seals), or bats, which are potential hosts of Leishmania [79][80][81][82][83]. Alternatively, recent satellite evidence has revealed a scattering of numerous seamounts across the Atlantic Ocean [84]. At 10 MYA, these seamounts may have existed as a large volcanic island chain that allowed movement of terrestrial organisms across the Atlantic, but eventually eroded into the sea [85]. However, it should be noted that these possibilities are purely speculative and not well supported by the evidence at hand.
Australia was considered free of Leishmania until the discovery of L. (M.) macropodum in 2004 [44]. Prior to the present study L. (M.) macropodum had not been formally described. Therefore, the name it was informally assigned i.e. Leishmania 'australiensis', represents a nomen nudum. However, the formal description provided herein resolves this issue. Based on current evidence, the presence of L. (M.) macropodum in Australia is most likely the result of vicariance; the complete separation of Australia from South America by roughly 40 MYA [3,21]. This study infers that the divergence of Z. australiensis from Z. costaricensis, and L. (M.) macropodum from other Mundinia parasites, occurred within approximately 3 million years of each other, approaching the Eocene to Oligocene transition (Fig 8). Given the margins of error associated with such predictions (S2 Fig) and the concurrence between the inferred divergence times of these taxa, the estimates presented here are plausible. This scenario is also consistent with the biogeography of other taxa, including the distribution of the plant genus Nothofagus and that of marsupials, which are generally restricted to parts of Central and South America, Australia and Oceania [3,86].
Novymonas esmeraldas, Z. costaricensis and Z. australiensis are presumably monoxenous trypanosomatids basal to all dixenous Leishmaniinae (Fig 6) [4,16], and probably represent the nearest ancestors of a parasite that transitioned from a monoxenous to a dixenous life cycle [87]. The rigorous growth of Z. australiensis in high haemoglobin concentrations and on chocolate agar is consistent with a haemoprotozoan (Fig 2, S1 File) [88] and/or adaptation to life as a parasite of hematophagous insects, which probably represents the first step in the transition to a dixenous life cycle. While Z. costaricensis was originally isolated from a non-hematophagous reduviid bug, Ricolla simillima, these insects are predatory and may have recently fed on a hematophagous insect prior to the isolation of Z. costaricensis [89]. This is conceivable as Novymonas which was first isolated and described from Niesthrea vincentii (Hemiptera: Rhopalidae) has also been detected in Zelus sp. (an assassin bug) and Culicoides sp. (a hematophagous midge) [16].
As parasites occupying the Novymonas/Zelonia clade (Fig 6) infect varied and disparate hosts, it is difficult to infer their vicariance based on host distribution. Also, given the origins of the Australian Simuliidae, their role in the dispersion of Zelonia is probably limited. Dumbleton [90] suggested that Simulium entered Australia from the north during what was then referred to as the Tertiary period, between 65 and 1.6 MYA. Similarly Crosskey [25] was of the firm opinion that Simulium entered Australia from the north via a land bridge that once connected Australia and New Guinea, but no time was suggested. As Australia drifted north, the interaction of New Guinea as the leading edge to the Australian Plate with the Pacific Plate and others, was complex and is discussed in some detail by Craig et al. [91] in relation to formation of the Solomon Islands. Given the distribution of various segregates of Simulium, colonization of this genus into New Guinea could have occurred as early as the mid Eocene to early Miocene (20 to 40 MYA). Simulium dycei is a member of subgenus Morops that is centred and diverse in New Guinea, an indication it is an older segregate of Simulium that colonized this land mass originally. A good assumption would be that Simulium has been on the Australian land mass for 40 MYA at most [91].
Despite the concurrence between the inferred arrival dates of Simulium in Australia and the appearance of Z. australiensis, it is unlikely that Zelonia was dispersed from South America to Australia via the Nearctic, the Palearctic and then South East Asia to arrive with Simulium. If dispersion of Leishmaniinae via this route occurred during this period, one might expect to encounter close relatives of L. (M.) macropodum or other dixenous species in Papua New Guinea, the Solomon Islands and/or parts of Indonesia, though no such reports exist. Consequently, the available evidence suggests that the separation of Australia from South America gave rise to Z. australiensis and L. (M.) macropodum. Zelonia probably came to infect Simulium when this genus arrived from New Guinea around 40 MYA. Prior to this, Zelonia was likely already in Australia, parasitizing other insect species. Indeed, investigation of other Australian insects such as native reduviids and Culicoides spp. for infection with Z. australiensis is warranted.
Leptomonas spp. are considered monoxenous parasites that are generally of no clinical importance [92][93][94]. However, L. seymouri, originally isolated from the phytophagous cotton stainer bug, Dysdercus suturellus [95], is capable of infecting humans opportunistically, inducing co-infections with L. (L.) donovani [96,97]. Its ability to cause human infections implies that L. seymouri also possesses an alternate hematophagous host [98]. While they are still considered monoxenous, and are continually grouped in basal clades to Leishmania [16,17,99] (Figs 6, 7 and 8), it is plausible that certain monoxenous Leishmaniinae are ancestors of transitional forms that did not complete the switch to a dixenous life cycle. Indeed, monoxenous trypanosomatids occasionally explore the dixenous niche based on multiple reports of infections involving animals and humans [98]. Genome sequencing and transcriptome profiling identified several adaptations in L. seymouri that allow it to persist in the vertebrate host environment [100]. Furthermore, L. seymouri survived for several days in two species of phlebotamine sand fly [100]. Given their close relationship with Leishmania, Leptomonas spp. represent interesting models for studying the transition from a monoxenous to dixenous life cycle, including the evolutionary innovations that enable parasitism of vertebrate hosts [98,100]. Moreover, the ability of L. seymouri to infect humans under some circumstances raises questions as to whether Novymonas and Zelonia are truly monoxenous, or if they might also be capable of infecting vertebrates under some circumstances, occasionally exploring the dixenous niche.
To conclude, we described the first isolation of Zelonia australiensis sp. nov. from the Australian native black fly S. (M.) dycei in Australia's Northern Territory. A detailed molecular and morphological characterisation was performed to establish this assignment, including light and electron microscopy, sequencing and phylogenetic analyses. As a result, Z. australiensis was identified as a sibling taxon to the monoxenous trypanosomatid, Z. costaricensis. Subsequently, the divergence of these species was used as a unique calibration point for a phylogenetic time tree exploring the relationships between several species of the Leishmaniinae subfamily. These analyses inferred the emergence of dixenous parasitism in the Leishmaniinae at approximately 91 MYA, in Gondwana, during the Cretaceous period. Ultimately, this study contributes to our understanding of trypanosomatid diversity by describing a unique Australian species, and to our understanding of Leishmania evolution by providing support for a Gondwanan origin of dixenous parasitism in the Leishmaniinae.  (Figs 3 and 4, S2 File), which is its sibling taxon. When cultured axenically, individuals of Zelonia australiensis exist in a variety of morphotypes as detailed in Fig 3. The species is also defined by a set of unique sequences of the 18S rDNA, gGAPDH, RPOIILS, HSP70 and ITS1 (GenBank Accessions: KY273490 to KY273515).

Etymology:
The species name is derived from the country Australia, where the organism was first isolated.
Type locality: Vicinity of Darwin, Northern Territory, Australia. The precise coordinates of isolation are provided in Table 1.  [20], although experimental infections have been achieved in Culicoides midges [22].
Type locality: Vicinity of Darwin, Northern Territory, Australia.
Type Material: For information on type material contact the investigators who first identified and later isolated L. (M.) macropodum [20,44,75].