Molecular Phylogeny of the Myxobolus and Henneguya Genera with Several New South American Species

The present study consists of a detailed phylogenetic analysis of myxosporeans of the Myxobolus and Henneguya genera, including sequences from 12 Myxobolus/Henneguya species, parasites of South American pimelodids, bryconids and characids. Maximum likelihood and maximum parsimony analyses, based on 18 S rDNA gene sequences, showed that the strongest evolutionary signal is the phylogenetic affinity of the fish hosts, with clustering mainly occurring according to the order and/or family of the host. Of the 12 South American species studied here, six are newly described infecting fish from the Brazilian Pantanal wetland. Henneguya maculosus n. sp. and Myxobolus flavus n. sp. were found infecting both Pseudoplatystoma corruscans and Pseudoplatystoma reticulatum; Myxobolus aureus n. sp. and Myxobolus pantanalis n. sp. were observed parasitizing Salminus brasiliensis and Myxobolus umidus n. sp. and Myxobolus piraputangae n. sp. were detected infecting Brycon hilarii.


Introduction
The phylum Myxozoa harbors a diverse group of metazoan parasites characterized by multicellular spores, with polar capsules containing an extrusible polar filament [1]. Parasites of this phylum have become increasingly important as new species are continually emerging as significant threats to the development of both farmed and natural environment fish [2].
Within myxozoans, 60 genera of the class Myxosporea and two of the class Malacosporea have been established. Among the myxosporeans, the genera Myxobolus Bütschli, 1882, and Henneguya Thélohan, 1892 are the most specious, and harbor species that have an important impact on their fish hosts [1,2,14,15]. These two genera are considered separated groups mainly due the presence of caudal appendages in Henneguya spp. [16]. However, phylogenetic studies using 18 S rDNA sequence data do not support a phylogenetic separation between Henneguya and Myxobolus [17][18][19].
The traditional approach for taxonomic classification of Myxozoa is based on morphological characteristics, primarily of spores, but also of the plasmodia, as well as host and organ and/or tissue specificity. Nevertheless, spore morphology is the main basis of classification and identification, as the spores are relatively rigid and exhibit low intraspecific morphological variability [20][21][22]. However, molecular analyses have been an important tool in the study of these parasites since the late 1990s [8,20,23]. Molecular biological methods have now been employed in myxozoan research, using a number of different approaches, such as the morphological differentiation of similar species [24][25][26]; study of host and tissue specificity [27]; elucidation of life cycle [11,23]; study of the phylogenetic relationships of the group [20,[28][29][30][31][32] and determination of the phylogenetic position of the myxosporeans inside the metazoans [33].
Some studies have attempted to establish the phylogenetic position of Myxozoa within Metazoa, but the classification of these parasites still remains uncertain, with evidence of the proximity of the myxozoans to both cnidarians and bilaterians, depending on the phylogenetic analysis used [33,34]. However, in a recent study, Nesnidal et al. [35] constructed a phylogenetic tree based on a dataset including 128 genes, pointing out that these new genomic data unambiguously support the evolution of the parasitic Myxozoa from Cnidaria.
Other studies, such as those of Eszterbauer [36], Milanin et al. [30], Gleeson and Adlard [37], Hartigan et al. [5] and Adriano et al. [28], have investigated the phylogenetic relationships of the species within a particular genus. These studies have demonstrated that many factors influence species clustering, such as the phylogenetic proximity of the host, tissue tropism, geographic distribution and morphology characteristics.
The present study describes six new myxosporean species based on morphological and 18 S rDNA data and provides a detailed molecular phylogenetic analysis of the Henneguya and Myxobolus genera. This analysis includes the addition of sequences from 12 species of parasites of South American fishes and reveals the relationship of these species with parasites of fishes from other continents.

Materials and Methods
All the fish handling was approved by the ethics committee for animal welfare of the Faculdade de Zootecnia e Engenharia de Alimentos of the Universidade de São Paulo (FZEA/USP), in accordance to the Brazilian legislation (Federal Law nu 11.794 from October 8th, 2008). The field samplings were carried out with permission of SISBIO (System Authorization and Information on Biodiversity) of the Ministry of Environment of Brazil (Authorization No. 15507-1).
Specimens of six fish species were caught in two Brazilian states. In the Brazilian Pantanal wetland (Pantanal National Park -17u5094899S, 57u2491499W [distance variation of 3 km], municipality of Poconé, Mato Grosso State), fishes of three species from the family Pimelodidae were examined, including 21  Immediately after capture, the fish were transported alive to the field laboratory where they were euthanized by benzocaine overdose and then measured and necropsied.
Organs or tissues infected with Myxozoa plasmodia were fixed in absolute ethanol for molecular analyses and in 10% buffered formalin for morphological analyses. Each sample was composed of several plasmodia grouped together according to phenotypic characteristics such as morphologic appearance, host species and tissue tropism.
In the laboratory, the plasmodia fixed in formalin were disrupted, and a small sample of the spores was transferred to a glass slide. A total of 30 spores of each sample were measured and photographed under a Leica DM 1000 light microscope equipped with Leica Application Suite version 1.6.0 image capture software. All measurements were performed according to the patterns established by Lom and Arthur [22]. These slides were then fixed with methanol and stained with Giemsa to be deposited in the collection of the Museum of Natural History, Institute of Biology, State University of Campinas (UNICAMP), state of São Paulo, Brazil.
DNA was extracted from the samples fixed in ethanol using the DNeasyH Blood & Tissue Kit (Qiagen) following the manufacturer's instructions. The product was then quantified in a NanoDrop 2000 (Thermo Scientific, Wilmington, USA) spectrophotometer at 260 nm.
Polymerase chain reaction (PCR) was carried out, according to Adriano et al. [28] at a final volume of 25 ml using the primers MX5-MX3 (Table 1). These primers failed to amplify the samples of Henneguya corruscans Eiras et al., 2009 taken from P. corruscans and P. reticulatum, Henneguya maculosus n. sp. taken from P. corruscans and Myxobolus pantanalis n. sp. These samples were amplified with the primer pairs ERIB1-ACT1R and MYXGEN4f-ERIB10 (Table 1), which amplified two overlapping fragments of approximately 1,000 bp and 1,200 bp respectively of the 18 S rDNA gene. The other primers listed in Table 1 were used only in sequencing reactions.
The amplification reactions were conducted with 10-50 ng of genomic DNA, 2.5 ml of 16 Taq DNA Polymerase buffer (Invitrogen by Life Technologies, MD, USA), 0.75 ml of MgCl 2 (1.5 mM), 0.5 ml of dNTPs (0.2 mM), 0.5 ml of each primer (0.2 mM), 0.25 ml of Taq DNA polymerase (2.5 U) (Invitrogen By Life Technologies, MD, USA) and MilliQ purified water. The PCR amplifications were performed in an AG 22331 Hamburg Thermocycler (Eppendorf, Hamburg, Germany). The PCR program consisted of 35 cycles of denaturation at 95uC for 60 s, annealing at 56uC for 60 s and extension at 72uC for 90 s, preceded by an initial denaturation at 95uC for 5 min and followed by a terminal extension at 72uC for 5 min.
The amplified products were analyzed via agarose gel electrophoresis, and their sizes were estimated by comparison with the 1 kb Plus DNA Ladder (Invitrogen by Life Technologies, CA, USA). The purified products were sequenced using the BigDyeH Terminator v3.1 Cycle Sequencing kit (Applied Biosystems TM ) in an ABI 3730 DNA Analyzer (Applied Biosystems TM Inc., CA, USA) using the primers listed in Table 1.
The sequences were visualized, assembled and edited using BioEdit 7.1.3.0 [38] and, for each sequence, a standard nucleotide-nucleotide BLAST (blastn) search was conducted [39] to verify their similarity to other myxosporean sequences in GenBank.
Representative sequences of the 18 S rDNA gene of myxosporeans of the Henneguya and Myxobolus genera infecting almost every order of fish host available in GenBank were included in the analysis. Since the regions of the 18 S rDNA gene are highly conserved, sequences shorter than 1,000 nucleotides found in the GenBank were not included in the analysis in order to maintain as high a resolution of the resulting trees as possible and to avoid loss of information due to shortening of the aligned sequences and the appearance of too many alignment gaps, as indicated by Rosenberg and Kumar [40].
The sequences of Kudoa thyrsites and Kudoa alliaria were used as an outgroup. To avoid long-branch attraction effect due to distantly related outgroup sequences, two separated analyses were performed using both maximum likelihood and maximum parsimony methods. The first analysis was conducted with a set of aligned sequences containing outgroup sequences, and the second was performed with the same dataset, but excluding the outgroup. Since there were no substantial differences for topology and bootstrap values between the trees (with and without the outgroup), the tree containing the outgroup sequences was used as the base topology of the final tree.
A total of 114 sequences were aligned using ClustalW implemented in the program BioEdit 7.1.3.0 [38] with default settings. Manual adjustments were performed by eye to correct the alignment and remove ambiguous positions from the dataset, resulting in an alignment with a total of 2,270 characters, including alignment gaps. For Myxobolus spp. parasites of fish from the Eurasia Palearctic region, only sequences considered valid by Molnár [41] were used.
The best evolutionary model of nucleotide substitution using the Akaike Information Criterion (AIC) was determined with the program jModeltest 0.1 [42], which identified the general time reversible model (GTR+G) as the best fit. From the data, nucleotide frequencies (A = 0.2347, C = 0.1809, G = 0.2742, T = 0.3101) and the rates of the six different types of nucleotide substitution (AC = 1.2478, AG = 3.5282, AT = 1.5099, CG = 0.7026, CT = 4.1695, GT = 1.0000) were estimated. The alpha value of the gamma distribution parameter was 0.4200. These parameters were employed in maximum likelihood (ML) analysis, which was performed using PhyML 3.0 [43] software, with bootstrap confidence values calculated using 500 replicates.
A maximum parsimony (MP) analysis was performed using PAUP* 4.0b10 [44], with a starting tree obtained via stepwise addition and a heuristic search with 320,172 replicates, with a tree bisection-reconnection (TBR) branch swapping algorithm and random sequence addition. Clade supports were assessed via a bootstrap analysis of 1,000 replicates.
The resulting trees were visualized with TreeView 1.6.6 [45] and edited and annotated in Adobe Photoshop (Adobe Systems Inc., San Jose, CA, USA).
The sequences of the South American species were aligned to produce a pairwise similarity matrix using MEGA 5.0 [46], with possible gaps and/or missing data treated via complete deletion, to determine the relationships among them and to verifying the possible occurrence of the same myxosporean infecting different tissues or organs, geographic locations or host species.

Nomenclatural Acts
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix ''http://zoobank.org/''. The LSID for this publication is: urn:lsid:zoobank.org:pub:20E2B90B-801D-43E4-B6D0-CE0561FFF7EB. The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS.

Results
Of the fishes examined in this study, the specimens of P. and H. pellucida, these were the first 18 S rDNA sequences obtained. For H. corruscans, the sequence produced herein was obtained from specimens found to infect P. reticulatum, which is a different host species than that from which it was originally described, and for H. eirasi, H. multiplasmodialis and M. cordeiroi, the new sequences were longer than the ones available in GenBank. The undescribed species, morphometric data, spores photomicrographs and drawings of the undescribed species are presented in the description provided below (Figs. 1 and 2; Table 2).
Description. White and elongated plasmodia, measuring 0.5 to 1.5 mm, found in the gill filaments of P. corruscans and P. reticulatum. The plasmodia exhibited spores in different stages of development. The mature spores were ellipsoidal from the frontal view, with the body measuring 13.760.6 mm in length, 4.160.2 mm in width, 3.060.2 mm in thickness and 17.561.0 mm in the caudal process. The polar capsules were elongated, filled the entire anterior half of the spore and were equal in size, measuring 5.660.5 mm in length and 1.660.2 in width. The polar filaments exhibited 6-7 turns. The measurements of the spores and polar capsules of the parasites found infecting P. reticulatum are presented in Table 2.
Type host. Pseudoplatystoma corruscans Spix & Agassiz, 1829; Etymology. The specific name refers to the spots in the skin of the host, which is known in Portuguese as pintado (in English ''spotted surubim''). In Latin maculosus = spotted (in English) = pintado (in Portuguese).
Remarks. When compared with the other Henneguya spp. that have been described infecting Pseudoplatystoma fishes, H. maculosus n.
sp. has spores with a morphology similar to that of H. corruscans. However, H. maculosus n. sp. has a longer caudal process, and the site of infection within the gills is different, as H. maculosus n. sp. forms plasmodia in the filaments, whereas H. corruscans forms round plasmodia on the lamellae. Apart from the length of the caudal process and the site of infection within the gills, the morphological features of the two species are very similar to each other, leading to inconclusive definitions regarding the separation of these species. However, combined with the differences mentioned above, molecular data (Table 3, Fig. 3) support the conclusion that these are distinct species.
Sequences consisting of 1,930 and 1,496 nucleotides were produced from the samples of H. maculosus n. sp. infecting P. corruscans (GenBank accession number KF296344) and P. reticulatum (GenBank accession number KF296345) respectively. Between the sequences from the samples collected in the two hosts, there were variations in 29 nucleotides, which represent 1.94% of the comparable total. The South American species that presented the most similar sequence to H. maculosus n. sp. found in P. corruscans was H. eirasi, which exhibited 111 different nucleotides (9.25% of the total). The comparisons with the sequence from the sample found in P. reticulatum are displayed in Table 3.
Description. Yellow and spherical plasmodia measuring 1 to 5 mm and containing spores in different stages of development were found in the gill arch of. P. corruscans and P. reticulatum. The mature spores were oval shaped from the frontal view. The spores obtained from plasmodia collected from P. corruscans measured 9.260.2 mm in length and 6.560.3 mm in width. From the lateral view, the spores measured 4.260.2 mm, and the valves were symmetrical. The polar capsules were similar in size and measured 4.560.2 mm in length and 1.660.1 mm in width ( Table 2). The polar filaments presented 4-5 turns. The measurements of the spores and polar capsules of the parasites found infecting P. reticulatum are displayed in Table 2 Etymology. The specific name refers to the yellow plasmodia formed by the parasite in the host, as in Latin, the word flavus = yellow.
Remarks. This is the first report of a species of the genus Myxobolus infecting fishes of the genus Pseudoplatystoma. The morphology of this parasite closely resembles that of Myxobolus pantanalis n. sp. (also described in this article). However, they are found in different host species and at different infection sites within the gills, with M. pantanalis n. sp. being found infecting the gill filaments of S. brasiliensis and M. flavus n. sp. obtained from the gill arch of P. corruscans and P. reticulatum. The molecular data show that these two myxosporeans do not share a close phylogenetic history. Compared to two other pimelodid-infecting Myxobolus species, M. cordeiroi, which infects Z. jahu, and Myxobolus absonus Cellere et al., 2002 infecting Pimelodus maculatus Lecépède, 1803 Sequences with lengths of 1,520 and 1,530 bases of the 18 S rDNA gene were obtained from the hosts P. corruscans (GenBank accession number KF296346) and P. reticulatum (GenBank accession number KF296347) respectively. The difference between the sequences was 28 nucleotides, which represents 1.88% of the comparable total (Table 3). When compared to the other South American species, the smallest amount of interspecific differences was observed between M. flavus n. sp. infecting P. corruscans and H. corruscans infecting P. reticulatum, corresponding to 224 different nucleotides (15.17%). Comparisons with the sequence from the sample found in P. reticulatum are displayed in Table 3.
Description. Small white plasmodia, measuring approximately 0.4 mm in length found in the liver of S. brasiliensis. The mature spores were oval shaped from the frontal view and measured 12.660.5 mm in length and 8.360.3 mm in width. From the lateral view, the spores measured 5.560.3 mm, and the valves were smooth and symmetrical. The meeting of the two valves produced a tenuous sutural line. The two polar capsules showed the same size and measured 5.760.3 mm in length and 2.960.2 mm in width ( Table 2). The polar filaments exhibited 7-8 turns.
Site of infection. Liver. A total of 1,757 bases of the 18 S rDNA gene of M. aureus n. sp. were sequenced (GenBank accession number KF296348). According to the similarity matrix (Table 3), the South American species that carries the most similar sequence to M. aureus n. sp. is Myxobolus piraputangae n. sp. (also described here), with 199 different nucleotides detected between them (15.40%) ( Table 3).
Description. White, round-to-elongated plasmodia measuring approximately 0.2 to 1.2 mm in length and found in the posterior end (distal region) of the gill filaments of S. brasiliensis. The mature spores were oval shaped from the frontal view and measured 9.360.4 mm in length and 6.560.4 mm in width. The two polar capsules were elongated, pear shaped and equal in size and measured 4.260.5 mm in length and 2.060.1 mm in width ( Etymology. The specific name refers to the location in which the infected fishes were caught, the Brazilian Pantanal. In Latin, pantanalis = wetland. Remarks. In S. brasiliensis, M. salminus was previously described as infecting the gills. However, the tissues infected by these two species are different, as M. pantanalis n. sp. is found in the surface of the filaments, and M. salminus is reported to infect the blood vessels of the gill filaments. These two species are also very different morphologically, as the spores of M. pantanalis n. sp. are   (Table 3), the smallest amount of interspecific differences was found with H. maculosus n. sp. parasitizing P. corruscans, corresponding to 290 different nucleotides (15.45%) ( Table 3).
Description. White and spherical plasmodia, measuring 0.4 to 0.8 mm, found in the spleen of B. hilarii. The mature spores were ellipsoid in shape from the frontal view and measured 13.560.7 mm in length and 7.860.4 mm in width. From the lateral view, the spores measured 7.760.1 mm, and the valves were symmetrical. The two polar capsules were oval shaped and of equal size, measuring 5.160.4 mm in length and 2.760.3 mm in width, and the polar filaments exhibited 4-5 turns, arranged obliquely with respect to the axis of the capsules (  Etymology. The specific name refers to the location in which the host is observed, the Brazilian Pantanal wetland, as in Latin, umidus = wet. Remarks. When compared to the other Myxobolus spp. found infecting B. hilarii, M. umidus n. sp. differs from M. oliveirai, Myxobolus brycon Azevedo et al., 2011, and Myxobolus piraputangae n. sp. (also described in this article) in morphological and morphometric characteristics and other phenotypic features, such as the preferential site of infection, as this is the only species already described infecting the spleen of B. hilarii. Regarding species parasites of other characiform hosts, Myxobolus serrasalmi Walliker, 1969 exhibits spores with a similar size and is also found in the spleen of Serrasalmus rhombeus Linnaeus, 1766, but it has longer and wider polar capsules than M. umidus n. sp., as well as distinct host species.
A 1,307 base sequence of the 18 S rDNA gene was generated from M. umidus n. sp. (GenBank accession number KF296350). The South American species whose sequence presented the smallest amount of genetic differences compared to M. umidus n. sp. was M. piraputangae n. sp., with 135 different nucleotides (10.44%) ( Table 3).
Description. White and spherical plasmodia, measuring 0.5 to 1.0 mm, found in the kidney of B. hilarii. Mature spores were circular to slightly ellipsoidal in frontal view, measuring 10.160.5 mm in length and 8.760.5 mm in width. From the lateral view, the spores measured 6.760.3 mm, and the valves were Etymology. The specific name refers to the common name of the host, which is the piraputanga in Brazil.
Remarks. Compared to the other Myxobolus species infecting B. hilarii, M. piraputangae n. sp. exhibited different spore morphology, presenting circular spores, rather than pear shaped spores, as in M. oliveirai and M. brycon, or ellipsoidal spores, as in M. umidus n. sp. The dimensions of these species are also very different, as M. oliveirai and M. umidus n. sp. exhibit larger spores, while those of M. brycon are much smaller than the spores of M. piraputangae n. sp. Regarding the other Myxobolus spp. was found infecting other characiform fishes, M. macroplasmodialis was described infecting S. brasiliensis and Myxobolus cunhai Penido, 1927 was described infecting Pygocentrus piraya Cuvier, 1819, which have spores similar in size to those of M. piraputangae n. sp., however, they differ in other phenotypic aspects, such as spore morphology, the site of infection and the host species.
A 1,296 nucleotide sequence of the 18 S rDNA gene from M. piraputangae n. sp. was generated (GenBank accession number KF296351). The South American species presenting the sequence with the smallest genetic variation from that of M. piraputangae n. sp. was M. umidus n. sp., corresponding to 135 different nucleotides (10.44%) ( Table 3).

Phylogenetic Analyses
The partial 18 S rDNA gene sequences obtained from 15 samples in this study, plus three others from parasites of South American fishes, were compared with each other and with other myxozoan sequences from other geographic regions available in GenBank. BLAST searches demonstrated that these sequences exhibit a close relationship with those of other myxosporeans that have been previously sequenced, but they were not identical to any of them.
The similarity matrix of the 18 S rDNA gene sequences of the South American species showed that among the myxosporeans that infect more than one host, the intraspecific variation ranged from 0.91%, between the two isolates of H. multiplasmodialis found infecting P. corruscans and P. reticulatum, to 2.32%, between the samples of H. corruscans found infecting P. corruscans and P. reticulatum ( Table 3). The M. macroplasmodialis sequences from S. brasiliensis sampled in the Brazilian Pantanal and the Mogi Guaçu River were identical. The smallest interspecific difference was observed between H. multiplasmodialis from P. reticulatum and H. corruscans from P. corruscans, which was 6.26%.
The phylogenetic analyses resulted in similar ML and MP topologies. Therefore, only the ML tree was used to represent the basic tree topology (Fig. 3), in which the clusters occurred mainly according to the phylogenetic affinities of the fish hosts. The tree was divided initially into two major clades, A and B. Clade A was formed by Myxobolus/Henneguya species associated exclusively with fishes belonging to the order Cypriniformes; clade B was further divided into two other clades (C and D); clade C consisted of one Myxobolus species parasite of anguiliforms and four Henneguya spp., parasites of salmoniforms; clade D was further divided to form two subclades (D1 and D2), both of Myxobolus/Henneguya species, which are parasites of different fish families. Subclade D1 harbored, in one branch, all the Myxobolus species parasites of salmoniforms, except Myxobolus cerebralis Hofer, 1903, which in our topology, was clustered outside the ingroup, much closer to the outgroup. In the other branch of the subclade D1, M. cordeiroi, a parasite of Z. jahu, a South American pimelodid, and Henneguya basifilamentalis Molnár et al., 2006 and Henneguya mystusia Sarkar, 1985, parasites of Hemibagrus nemurus Valenciennes, 1840, an Asian bagrid, were clustered. The branch D1 also harbored four Myxobolus species, parasites of South American characiforms of the bryconid family. The speciose subclade D2 harbored Myxobolus/Henneguya species parasites of marine perciforms and mugiliforms, two freshwater perciforms (Henneguya creplini Gurley, 1894 and Henneguya doori Guilford, 1963), plus Henneguya lobosa Cohn, 1895 and Henneguya psorospermica Thelohan, 1895 parasites of esociforms, and the South American M. oliveirai and H. pellucida, parasites of characiforms of the families Bryconidae and Characidae, respectively. M. pantanalis n. sp., another species that is parasite of a South American bryconid, appeared alone in a branch away from the other two groups of characiform parasites. Still in D2, five Henneguya spp., parasites of fishes of the family Ictaluridae, two Myxobolus spp., parasites of fishes of the pangasiid family, and five species (four Henneguya and one Myxobolus) that are parasites of fishes of the family Pimelodidae formed a large group composed of species that are parasites of siluriforms (Fig. 3).

Discussion
In this report, based on analyses of the 18 S rDNA gene sequences and morphological features, six novel species of myxozoans are described, one of which belongs to the genus Henneguya and five to Myxobolus. Of these six new species, four Myxobolus spp. were found infecting bryconid hosts, with M. aureus n. sp. and M. pantanalis n. sp. being parasites of S. brasiliensis and M. umidus n. sp and M. piraputangae n. sp. parasitizing B. hilarii. In S. brasiliensis and B. hilarii four myxosporean species have been described previously, with M. macroplasmodiaslis and M. salminus parasitizing S. brasiliensis and M. oliveirai and M. brycon parasitizing B. hilarii. Both M. flavus n. sp. and H. maculatus n. sp. are described here as infecting two pimelodid species, P. corruscans and P. reticulatum. The observation of these myxosporeans in these two hosts expands the number of species described as infecting these pimelodids. The parasites previously reported in these fish species include H. corruscans, H. eirasi and H. multiplasmodialis, which were found infecting both hosts in the natural environment in Brazilian rivers, in addition to Henneguya pseudoplatystoma Naldoni et al., 2009, which was recorded infecting hybrid fish, resulting from a cross between P. corruscans and P. reticulatum, on fish farms in the São Paulo and Mato Grosso do Sul states, in Brazil.
The findings presented above reveal the notable fact that P. corruscans and P. reticulatum, which are species that are phylogenetically very close, can host the same myxosporean species. These results are in accordance with the host specificity criteria noted for Henneguya and Myxobolus species by Molnár [47] and Molnár et al. [48].
The similarity matrix showed that among the myxosporean species that are able to infect both P. corruscans and P. reticulatum, the genetic divergence between the samples obtained from the different hosts was only 0.91% for H. multiplasmodialis, 1.88% for M. flavus n. sp., 1.94% for H. maculosus n. sp. and 2.32% for H. corruscans. We consider these levels to reflect intraspecific divergences. On the other hand, the interspecific differences observed among the sequences of the morphologically distinct species varied from 6.26% to 27.37% (Table 3).
Might be expected that, because of the sampling method employed, in which several cysts from the same parasite species collected from each host species were grouped together to produce a satisfactory amount of DNA, we would detect intraspecific variations among samples from the same host. This might be observable on the sequencing electropherograms, but was not in fact detected. These results suggest that, although more specific studies regarding the population structure of this species with more suitable molecular markers are required, there exist different lineages of the same species infecting these two phylogenetically similar hosts, probably on a speciation process.
There is no exact value for determining whether particular differences in 18 S rDNA genes can be used to discriminate intraor interspecific variation in members of the phylum Myxozoa [49]. Thus, this decision must be made for each individual case, always with the aid of other aspects of the biology and/or ecology of the organisms, such as their morphology, tissue and/or organ tropism, type of host and geographic area [49]. In taxonomic analyses of Myxozoa, Molnár et al. [50] [21] even though only 0.5% variation was detected between their 18 S rDNA sequences. Because the samples presented different infection sites and spores with significantly different lengths, the authors concluded that they were separate species. The same conclusion was made by Easy et al. [51], who stated that even though Myxobolus intramusculi Easy et al., 2005 and Myxobolus procerus Kudo, 1934 were both found in the muscles of Percopsis omiscomaycus Walbaum, 1792 and presented differences in only 2% of their 18 S rDNA sequences, they were considered two separate species, because M. intramusculi was found in intramuscular tissue, whereas from M. procerus was found infecting the adjacent connective tissue. This is the first study investigating the phylogenetic relationships among several Henneguya spp. and Myxobolus spp. infecting South American fishes. In the ML and MP trees produced in this study, the characteristic presenting the most apparent phylogenetic signal was the relationship of the fish hosts, with clusters occurring mainly according to the order and family. This pattern can be clearly observed in almost all clades. Clade A was composed exclusively of Myxobolus/Henneguya species, parasites of fishes of the order Cypriniformes, indicating that this group has a distinct evolutionary origin from the other species analyzed. Following the tendency to group according to the order/family of the fish host, Henneguya spp. parasites of salmonids formed a subclade in clade C, while Myxobolus spp. parasites of salmonids formed a subclade in D1. However, M. cerebralis appears separated from the Myxobolus spp. parasites of salmonids, revealing that this group is not monophyletic. The South American species, parasites of characiform fishes also formed polyphyletic clusters, with species occurring in subclades D1 and D2. However, in subclade D1 there is a further subclade composed exclusively of four Myxobolus spp. parasites of the family Bryconidae. In clade D2, the characiform parasites M. oliveirai (from a bryconid host) and H. pellucida (from a characid host) clustered together, and M. pantanalis n. sp. clustered away from the other characiform parasites.
Regarding the Myxobolus/Henneguya species parasites of siluriforms, M. cordeiroi, parasite of a pimelodid and H. basifilamentalis and H. mystusia, parasites of an Asian bagrid, all clustered in a subclade of D1, while in a subclade of D2, all other species parasites of siluriforms (six Henneguya spp. of North American ictalurids, two Myxobolus spp. of an Asian pangasiid and four Henneguya spp. and one Myxobolus of South American pimelodids), formed a grouping composed exclusively of parasites of this host order. Despite the polyphyletic clustering observed among parasites of siluriforms, the large majority of the species clustered in a monophyletic subclade in D2, and even M. cordeiroi, H. basifilamentalis and H. mystusia clustered together in D1.
In general, even though there are some inconsistencies, as reported above, the tree produced in the present study shows a strong tendency of the Myxobolus/Henneguya species to form clusters based on the phylogenetic relationships of the fish hosts, even if these hosts are from different biogeographic regions, as has been previously reported by other authors [28,32]. This suggests that the origins and radiations of these parasites are very ancient, perhaps as old as the hosts themselves, which, in the case of Osteichthyes, had an Early Silurian origin [52] and Mesozoic radiation [53].
The clustering of myxosporeans according to host phylogeny was also observed in a study by Gleeson et al. [54] which reported a clear separation among multivalvulid myxozoans of the genus Kudoa, parasites of elasmobranch and teleost hosts. However, this separation occurred at a higher taxonomic level of the hosts than that observed in this study for Henneguya and Myxobolus, since some Kudoa spp. have broader host specificity with infections across multiple families and even orders. Similar findings were reported by Gleeson and Adlard [37] who founded that Chloromyxum spp. parasites of elasmobranchs tend to cluster together despite having different geographic origins.
In clade D the occurrence of some clusters of lineages of myxozoans from a marine environment was observed, as were clusters with distribution according to the host order, with one clade composed of Henneguya spp. parasites of marine perciforms, and another formed by Myxobolus spp. parasites of mugiliforms. Eszterbauer [36] suggested that the preference for a particular site of development plays an important role in determining the phylogenetic relationships among myxosporean species, and she also proposed that genetic separation based on tissue and/or organ tropism is a more ancient evolutionary characteristic than host specificity. However, the results reported by Eszterbauer [36] were based on data obtained from a study including only myxosporean species parasitizing cyprinid fishes. This does not mean that the site of infection is not a relevant evolutionary feature, but it appears that organ tropism has more evolutionary influence within clades that are defined by a stronger evolutionary signal, such as the order/family of the fish host. In accordance with Eszterbauer [36] and Fiala [18], our topology shows that tissue tropism has some phylogenetic influence in some places in the tree, especially when considering myxosporean species parasitizing hosts that are phylogenetically close.  et al., 2008). In all these cases, these myxosporean species clustered primarily by host affinity and in a second step, by organ/tissue tropism.
The classic taxonomic distinction between the Myxobolus and Henneguya genera is based on a single morphological character: the presence of the caudal process in Henneguya spp. and its absence in Myxobolus spp. [16]. Our phylogenetic analyses showed that species of the Henneguya and Myxobolus genera were grouped together, and this corroborates the results reported by several other authors in studies conducted in other regions of the world [17][18][19]21]. This phenomenon can be observed in clade A, where Henneguya doneci  (Fig. 3). Thus, the findings of the present study of Myxobolus/Henneguya parasites of South American species, corroborate the idea that, based on analyses of ribosomal genes, there is no support for a phylogenetic separation of the Henneguya and Myxobolus genera. This absence of phylogenetic separation between these two genera led Kent et al. [17] to state that the caudal appendage of Henneguya spp. is not a valid feature for characterization of the genus. These authors also affirmed that the fact that species belonging the Henneguya and Myxobolus genera cluster together in groups with a monophyletic origin suggests that the genetic ability to develop spores with a caudal appendix is broadly distributed within the group, but due to still unknown reasons, only certain lineages express this character. A piece of evidence that provides support for this theory is the occurrence of some Myxobolus spp. that develop spores both with and without a Henneguya-like caudal prolongation within a single plasmodium, as reported for Myxobolus turpisrotundus Liu et al., 2010 [19], Myxobolus bizerti Bahri and Marques, 1996, Myxobolus mulleri Bahri, 1997 and Myxobolus heterosporus Baker, 1963 [55].
To date, phylogenetic studies of myxosporeans have been performed using species parasites from the Nearctic, Palearctic and Australian bioregions. In this study, for the first time, a consistent number of sequences of Myxobolus/Henneguya spp. parasites of fishes from the Neotropical region were included in a phylogenetic analysis with almost all of the valid sequences of species of these two genera available in GenBank. The results reveal that the strongest evolutionary signal for Myxobolus/ Henneguya species is the phylogenetic affinity of the fish hosts, with clusters occurring mainly based on the order and/or family of the host. The origin of the Osteichthyes is estimated to the Early Silurian, around 427 million years ago [52], with Mesozoic radiation [53]. The observation that the phylogenetic affinity of the fish hosts is the strongest evolutionary signal in Myxobolus/ Henneguya species suggests an ancient origin and radiation of this important and complex group of parasites.