Three new species of Creptotrema (Trematoda, Allocreadiidae) with an amended diagnosis of the genus and reassignment of Auriculostoma (Allocreadiidae), based on morphological and molecular evidence

Diversity of Creptotrema was investigated using morphological and molecular tools, including data for Creptotrema creptotrema (type-species). Three new species, parasites of Brazilian fishes, are described: Creptotrema conconae n. sp. (type-host, Imparfinis mirini Haseman), Creptotrema schubarti n. sp. (type-host, Characidium schubarti Travassos) and Creptotrema megacetabularis n. sp. (type-host, Auchenipterus osteomystax (Miranda Ribeiro)). The diagnosis of the genus was amended to include new features. The new species differ from each other mainly in terms of body shape, relative sucker size, and testes position. DNA sequences were obtained from Creptotrema spp. from Brazil, including 28S, ITS and COI. Genetic divergences among the new species and C. creptotrema varied from 2.1 to 5.2% (21–49 bp) for 28S, and 6.6 to 16.4% (21–45 bp) for COI. Phylogenetic analysis (28S) placed the newly generated DNA sequences of Creptotrema in a clade (Creptotrema clade sensu stricto) composed of C. creptotrema, the new species described herein, and all species previously described as Auriculostoma, revealing that Auriculostoma is best interpreted as a synonym of Creptotrema based on the principle of priority of zoological nomenclature. Creptotrema funduli, the single sequence of the genus previously available, was not grouped within the Creptotrema clade sensu stricto, suggesting the need for reevaluation of the taxonomic status of this species. Our results showed that Creptotrema represents a monophyletic genus of trematodes widely distributed across the Americas, which currently comprises 19 valid species of parasites of teleosts and anurans.


Introduction
Allocreadiidae Looss, 1902 includes relatively small digeneans that appear to be restricted to freshwater systems distributed across the Americas, Asia, and Europe [11,24]. The cercariae develop in sphaeriid clams (Bivalvia) and encyst in aquatic arthropods (primarily insects and crustaceans). The adult forms inhabit the digestive system of teleosts and, occasionally, of reptiles (snakes), and amphibians (salamanders and frogs) [11].
The delimitation of genera in Allocreadiidae is usually defined by the possession of a combination of four non-unique morphological characters: the number of muscular "papillae" in the oral sucker; posterior extent of the intestinal ceca; posterior extent of the uterus, and anterior extent of the vitelline follicles. This approach has led to several controversies, resulting in polyphyletic groups and controversial genera and subgenera [11].
Another "papillose genus" of allocreadiids of freshwater fishes from the Neotropical region is Auriculostoma Scholz, Aguirre-Macedo & Choudhury, 2004. This genus includes ten valid species of parasites of teleosts, most of which parasitize small-bodied characiforms from South America [25,26]. According to Scholz et al. [58], the synapomorphy of Auriculostoma, relative to all other allocreadiid genera, is the presence of one pair of ventrolateral muscular papillae and one pair of prominent auricular dorsolateral papillae. However, recent studies [24,53] using scanning electron microscopy (SEM) of specimens belonging to this genus refute the morphological description proposed by Scholz et al. [58], demonstrating the presence of a single pair of muscular lobes on either side of the oral sucker, thus raising questions about the classification of this group. Although molecular analyses, including the sequencing of Auriculostoma spp., have suggested the monophyly of the group, the lack of genome sequences of other "papillose genera" of allocreadiids from the Neotropical region has raised questions about the systematics of this trematode family.
As part of our long-term studies on the biodiversity of fish parasites from the tributaries of the Upper Paraná River basin in Brazil and streams from the Paranapanema River basin, three new species of Creptotrema are described, isolated from teleosts belonging to Siluriformes and Characiformes, supported by morphological and molecular data. For the first time, DNA sequences of Creptotrema spp. from Brazil were obtained, including the large subunit gene (28S rDNA), the internal transcribed spacer (ITS) region of ribosomal DNA, and the cytochrome c oxidase I gene (COI mtDNA) of mitochondrial DNA, including data for C. creptotrema from the type-locality and type-host.
The generic diagnosis of the genus was amended to accommodate the proposed new species and to reassign the species previously described as Auriculostoma (new combinations). The phylogenetic relationships among Creptotrema spp. and other digenean parasites of fishes were also evaluated, including the sequence of Creptotrema funduli Mueller, 1934 from North America, the single previously sequenced species also considered a member of Creptotrema.

Host sampling and parasitological procedures
All fishes were individually euthanized, stored in plastic bags, and placed in a styrofoam box with ice for transportation to the laboratory for necropsy. The intestines of the fishes were removed and analyzed using a stereomicroscope. Some of the collected specimens of digeneans were preserved directly in 96% ethanol for molecular analyses; others were placed in hot water (~60°C) and subsequently stored in 70% ethanol for morphological procedures or fixed directly in hot 70% ethanol for SEM preparations.
For analyses of the internal organs, the digenean specimens were stained with chloride carmine, Mayer's carmalumen, or Gomori's trichrome, cleared in eugenol, and mounted on permanent slides using Canada balsam. Some specimens were mounted in Hoyer's or Gray and Wess' medium to highlight the visualization of the eggs.
Morphometrical and morphological analyses of these specimens were performed using a computerized image analysis system with differential interference contrast (DIC) (Leica Application Suite, V3; Leica Microsystems, Wetzlar, Germany). The morphological descriptions followed the recommendations of Travassos et al. [68], Kohn [29], and observations by Scholz et al. [58] and Razo-Mendivil et al. [53]. All measurements are presented in micrometers (lm) and are expressed as means, followed by the range in parentheses. Illustrations of the structures were produced with the aid of a camera lucida mounted on a Leica DMLS microscope with phase-contrast optics. The prevalence and mean abundance of infections were calculated according to Bush et al. [10].
The SEM analyses were conducted on the specimens of C. creptotrema at the Laboratório de Microscopia Eletrônica e de Microanálise (LMEM) of the State University of Londrina (UEL), in the municipality of Londrina, Paraná State, Brazil. The specimens were processed through an increasing series of alcoholic dehydration (70%-100%) (adapted from [2] Genomic DNA was extracted from the species described herein (three specimens of C. creptotrema, and specimens of the new species described herein: one specimen of Creptotrema conconae n. sp., one specimen of Creptotrema schubarti n. sp., and four specimens of Creptotrema megacetabularis n. sp.), using the DNeasy Blood & Tissue Kit (Qiagen, Valencia, CA, USA), following the manufacturer's protocol. Fragments of 28S rDNA, ITS rDNA, and COI mtDNA were amplified using the primers and cycling conditions presented in Table 1. Conventional polymerase chain reaction (PCR) amplifications were performed on a final volume of 25 lL using ready-togo PCR beads (GE Healthcare) with extracted DNA (3.0 lL) and 1.0 lL of each PCR primer. PCR products (2.0 lL) were run on an agarose gel (1%) using GelRed™ fluorescent nucleic acid dye and loading buffer to confirm amplicon size and yield. PCR amplicons were purified using the QIAquick PCR Purification Kit (Qiagen), following the manufacturer's instructions. Automated sequencing was performed directly on purified PCR products using a BigDye v.3.1 Terminator Cycle Sequencing Ready Reaction kit on an ABI 3500 DNA genetic sequencer (Applied Biosystems). The new sequences were assembled and edited using Sequencher v. 5.2.4 (Gene Codes, Ann Arbor, MI, USA).
The obtained ITS rDNA sequences were deposited in the GenBank database. However, no phylogenetic analysis was conducted because of the shortage of sequences of this marker from allocreadiids deposited in the databank, culminating in an alignment with a reduced number of base pairs. The phylogenetic analyses presented herein were performed using sequences of 28S rDNA and COI mtDNA. Two independent datasets were created: the first contained the newly generated 28S rDNA sequences of C. creptotrema and three new Creptotrema spp., along with 79 published sequences of Allocreadiidae retrieved from GenBank and species from the genera Prosthenhystera Travassos, 1922 (Callodistomidae), Dicrocoelium Dujardin, 1845 (Dicrocoeliidae), Degeneria Campbell, 1977, and Phyllodistomum Braun, 1899 (both Gorgoderidae), used as the outgroup. The second contained the newly generated COI mtDNA sequences of C. creptotrema and the three new Creptotrema spp., along with 16 published sequences of the Allocreadiidae retrieved from GenBank and Phyllodistomum parasiluri Yamaguti, 1934, and Dicrocoelium spp. as the outgroup ( Table 2).
The alignments were performed using the MUSCLE algorithm implemented on Geneious 7.1.3 [27] with default settings. The presence of stop codons and indels for the COI mtDNA alignment was verified by amino acid translation using the trematode mitochondrial code table on Geneious 7.1.3 [27]. To evaluate the occurrence of substitution saturation, the I ss indices were estimated using DAMBE 5 software [72].

Phylogenetic analyses
The best-fitting models for nucleotide substitution selected by jModelTest software [51] using the Akaike information criterion were GTR + I + G for the 28S rDNA and HKY + G for the COI mtDNA datasets. Phylogenies were reconstructed for each alignment under Bayesian inference (BI) using MrBayes v. 3.2 and maximum-likelihood (ML) on RAxML v. 8 [61], both implemented on the CIPRES web portal [40]. The BI was run using four Markov chain Monte Carlo searches with 10,000,000 generations and sampling tree topologies every 100 generations. The burning was set to the first 25% of generations; the consensus trees were estimated using the remaining topologies. The nodes with posterior probabilities greater than 0.90 were considered well supported. The ML analyses were estimated using random starting trees and 1000 bootstrap replicates; nodes with bootstrap values greater than 70 were considered to be well supported. The final trees were visualized and edited using FigTree v. 1.3.1 [52] and CorelDRAW X6.
Pairwise genetic distances within and among sequences were calculated using the Kimura-2-parameter (K2P) model and a bootstrap procedure with 1000 replicates in the MEGA7 program [28,32].  (Continued on next page)  prepharyngeal level (neck region) present or absent. Tegument smooth. Remnants of eyespots present (most evident in young specimens, adults in the early development phase) or absent. Oral sucker subterminal, rounded or funnel-shaped with variable morphology (a ventral anterior horseshoe-shaped structure and presence of a discrete single pair of muscular lobes ["auricles"] on either side of oral sucker, stretching from the ventral side to lateral area, but not extending to dorsal region, or a single pair of muscular lobes on either side of oral sucker, with a broad base, stretching from ventrolateral to dorsolateral side forming an auricular dorsolateral papilla, with a free end). Small dome-like papillae over the margin of oral sucker (around [19][20][21] and ventral sucker (around 6), with a variable arrangement. Ventral sucker round, pre-equatorial to equatorial, with similar size or larger than oral sucker. Prepharynx short or absent. Pharynx subspherical. Esophagus short to moderately long. Intestinal ceca blind, bifurcating at anterior margin of ventral sucker; ceca end immediately or well posteriorly beyond testes. Genital pore median, at level of intestinal bifurcation or slightly anterior or posterior, equidistant between ventral sucker and cecal bifurcation. Seminal vesicle bipartite; pars prostatica and cirrus present. Cirrus unarmed, eversible. Cirrus-sac well-developed, long, usually forming several loops, sometimes overlapping ventral sucker or reaching the ovary posteriorly. Testes in posterior half of body, oval or elongate, entire or slightly lobed, tandem, symmetrical or oblique, intra or extracecal. Ovary posterolateral to ventral sucker, slightly pre-equatorial to equatorial, pretesticular, entire to lobed, submedian. Seminal receptacle posterior to ovary. Laurer's canal with dorsolateral opening. Vitelline follicles lateral, exclusively extracecal or extra-and intracecal, with large or small follicles extending from level of oral sucker to posterior end of body, completely separated into two fields or confluent (scarcely or not) in post-testicular region. Uterus pretesticular, between anterior testis and genital pore, but loops may extend to the post-testicular region; intra-or extracecal. Eggs operculate, unembryonated, without spines or filaments; small knob opposite to operculum present or absent. Excretory pore terminal at posterior region of body; excretory vesicle I-shaped to claviform, reaching to anterior testis. Intestinal parasites of freshwater fishes and anurans in the Americas. Remarks: Travassos et al. [68] originally described the defining characters of Creptotrema as the presence of one pair of muscular ventrolateral lobes (= "papillae" in original description) associated with the oral sucker, and a uterus that does not reach the region posterior to the testes. In the present study, the generic diagnosis was amended, adding new features to accommodate the three new species proposed in the present study, and also amending the reassignment of the species previously described as belonging to Auriculostoma (new combinations).
According to Scholz et al. [58], the synapomorphy of Auriculostoma to all other allocreadiid genera is the presence of two pairs of muscular lobes (originally described as "papillae"), one ventrolateral and one pair prominent, auricular dorsolateral. However, Razo-Mendivil et al. [53], using SEM photomicrographs of specimens of Auriculostoma totonacapanense Razo-Mendivil, Mendoza-Garfias, Pérez-Ponce de Léon & Rubio-Godoy, 2014, demonstrated the presence of a single narrow muscular lobe on either side of the oral sucker, with a broad base, stretching from the ventrolateral to the dorsolateral side, and free ends (see Fig. 2 of [53]). The same morphology was confirmed using SEM analysis for Auriculostoma lobata [24] and Auriculostoma tica [25]. SEM analysis also revealed the presence of small dome-like papillae over the margins of oral and ventral suckers that were poorly visualized by light microscopy. In the present study, we adopted the term "lobes" to refer to prominent and muscular structures ("auricles") present in the superior margin of the oral sucker, while the term "papillae" was used to refer to these small dome-like structures over the oral and ventral suckers.
Following the analyses of the species originally described as Creptotrema and Auriculostoma deposited in museums, it was possible to observe the same morphological pattern of the oral sucker described by Razo-Mendivil et al. [53] for the majority of species that varied in the presence or absence of dorsal free ends. The same morphological pattern was observed in the SEM analyses of the newly collected specimens of C. creptotrema in the present study (Figs. 1A-1C), in vivo specimens (of the new species described and C. creptotrema analyzed herein), and the illustrations provided in the descriptions of new species over the years.
Therefore, the main amendments to the original diagnosis of the genus Creptotrema are the presence of remnants of eyespots in some species (numerous and evident in young and adults in the early development phase, versus not present in the original description by Travassos et al. [68]); the presence of ovary and testes with entire or slightly lobed margins, versus simply entire in the original description of Creptotrema; the presence of small dome-like papillae over the margin of the oral sucker and ventral sucker, with variable arrangements (more visible by SEM analysis); and the presence of an oral sucker subterminal, rounded or funnel-shaped, with variable morphology of the lobes. The oral sucker may form a discrete lobe on either side of the oral sucker, without prominent dorsolateral lobes ("auricles") and free ends on the dorsal view (as originally described for the type-species C. creptotrema and other congeners such as C. lynchi and C. megacetabularis n. sp.); or assuming the same morphology described by Razo-Mendivil et al. [53] and observed for most species previously attributed to Auriculostoma and some species of Creptotrema (C. paraense, and the new species described herein as C. conconae n. sp.).
Although the uterus was originally described as pretesticular, it may extend to the inter-testicular (as observed in C. guacurarii comb. n.) or post-testicular regions, depending on the number of eggs present in the uterus and the maturity of the specimens (as amended for C. creptotrema in this study, and described for C. schubarti n. sp. and C. megacetabularis n. sp.); it may occupy the extracecal region (as described for C. lynchi and C. megacetabularis n. sp.) or the intra-cecal region (as in other congeners).
The morphological and molecular evidence presented in this study supports the recognition of Auriculostoma as a synonym of Creptotrema.  [1,7].
Site of infection: intestine.
Prevalence and mean abundance (present study): one host analyzed and infected with 17 specimens.
Redescription: (Based on specimens collected for the present study: 5 completely developed adults, stained with chloride carmine). Body small, round to elongated, 1248 (1120-1405) long, 564 (513-663) wide. Tegument smooth. Oral sucker subterminal, rounded to funnel-shaped, 163 (141-206) long, 188 (175-194) wide, with ventral anterior horseshoe-shaped flaps, and a discrete single ventrolateral muscular lobe ("auricles") on either side of oral sucker, stretching from ventral region to lateral area, but not extending to dorsal region. Ventral sucker Remarks: The original description of C. creptotrema [68] was revised by Kohn [29] to resolve inconsistencies about morphological features such as the absence of spines in the tegument (originally described as spinous), and to provide a consistent description of the cirrus sac (absent in the original description). The redescription provided by Kohn [29] was made based on the same specimens analyzed by Travassos et al. [68], deposited at the Helminthological Collection of the Oswaldo Cruz Institute (CHIOC).
In the present study, we collected new specimens of C. creptotrema obtained from the type-host M. obtusidens (Anostomidae) and type-locality (Emas Waterfalls) in Brazil and compared them with all the materials previously deposited in the CHIOC by Travassos et al. [68]. The specimens from the present study were obtained from a single host and presented morphological variations in body size and shape, the ratio between oral and ventral sucker, and the number of eggs in the uterus.
Some specimens were similar to those described by Travassos et al. [68] (Figs. 2C-2D; Figs. 3A-3B), presenting a small body, round to elongated, ventral sucker noticeably larger than oral sucker, and few large eggs in the uterus (less than 10), commonly distributed in the pretesticular or testicular region. Others were similar to specimens of C. lynchi reported in fishes belonging to Anostomidae from Emas Waterfalls (Cachoeira de Emas) by Kohn et al. [30] (Figs. 2A-2B; Figs. 3C-3D). These are probably all fishes belonging to the Megaleporinus genus, considering that the natural distribution of Hypomasticus copelandii Steindachner, 1875 (= Leporinus copelandii) ranges from the Paraíba do Sul River to Mucuri River, including Doce River and Itapemerim River, Brazil [6]). These digenean specimens exhibited an elongated body, and although the ventral sucker was larger than the oral sucker, both appeared proportionally smaller when compared to the total body length. The uterus presented numerous eggs (more than 10), exceeding the testis and/or extending to the posterior region of the body (for detailed morphological and morphometric data, see Table 3).
Although morphological variations have been observed among the new specimens of C. creptotrema, there is no intraspecific genetic divergence between the sequences of these different morphotypes (see "Molecular data and phylogenetic inference" in the "Results" section). This may indicate that the morphological differences observed are probably due to the different stages of the ontogenetic development of the specimens. Therefore, it is suggested that the specimens of C. creptotrema studied by Travassos et al. [68] in the original description might represent adults in the early development phase. On the other hand, the specimens reported by Kohn et al. [30] as C. lynchi in anostomid fishes from Emas Waterfalls were interpreted as completely developed adult specimens belonging to C. creptotrema, since C. lynchi was originally described in toads (Table 4) from Colombia [9]. Differences in the morphology of adults in the early development phase and completely developed adults were also observed for C. conconae n. sp. and C. megacetabularis n. sp. described in this study, reinforcing this hypothesis. The differences are mainly related to the number of eggs in the uterus (fewer eggs in adults in the early development phase and eggs larger proportionally to the total body size), differences in oral sucker morphology (predominantly subspherical in adults in the early development phase) and differences in oral and ventral sucker ratio (see their respective descriptions).
Previous studies on Creptotrema spp. based on light microscopy may have misinterpreted the morphology of the oral sucker. SEM photomicrographs of the specimens of C. creptotrema collected during the present study (Figs. 1A-1C) confirm the morphology of the oral sucker. This structure presents as ventral anterior discrete horseshoe-shaped flaps and a discrete single ventrolateral muscular lobe on either side of the oral sucker, stretching from the ventral side to the lateral area, but not extending to the dorsal region, without free dorsal ends. The presence of a tegument, papillous or rugose, on the suckers described by Kohn [29] was not observed in our light microscopy and SEM analyses. Material deposited: holotype CHIOC 39595, paratypes CHIOC 39596a-g; 39597a-b, and vouchers CHIBB 645 L-651 L.
Representative DNA sequence: 1107 bp long sequence of the 28S rDNA gene -GenBank OK044374; 1073 bp long sequence of complete ITS -GenBank OK047367; 447 bp long sequence of COI mtDNA -GenBank OK075288, all obtained from a specimen parasitizing I. mirini.
Etymology: The epithet refers to the Dr. Maria Conceição Zocoller-Seno, retired professor at São Paulo State University, Campus of Ilha Solteira, Brazil. During her career, Dr. Zocoller-Seno worked as a researcher in the field of Animal Parasitology. Affectionately nicknamed by students and co-workers as "Concon" (in mention of the initial of her first name "Conceição"), she was a great motivator in the academic and teaching career of several students. Thus, in recognition of her performance, we name the new species as a representation of our gratitude and appreciation.
Description: (Based on 2 specimens stained with Mayer's carmalumen and 8 specimens stained with Gömöri's trichrome). Body elongated, 1463 (994-1857) long, 298 (126-421) wide, slightly constricted at pre-pharyngeal level (neck region). Remnants of eyespots present (most numerous and evident in adults in the early development phase), sparse, restricted to anterior first-third of body. Oral sucker subterminal, subspherical to funnel-shaped, with 156 (126-188) long, 161 (127-209) wide, with a single muscular lobe on either side of the oral sucker ("auricles"), with a broad base, stretching from ventrolateral to dorsolateral side and presenting free dorsal ends.  Less than 10 eggs More than 10 eggs More than 10 eggs More than 10 eggs * Measurements of the diameter. The adult in the early development phase and completely developed adult forms of Creptotrema creptotrema from this study were recovered from a single host, and specimens showing both morphologies were sequenced and used in the 28S rDNA and COI mtDNA analyses; there was no intraspecific genetic divergence between these morphotypes. Remarks: Creptotrema conconae n. sp. was placed in this genus in the context of the current proposal that considers Auriculostoma synonymous to Creptotrema. This new species is consistent with the original generic diagnosis of Auriculostoma [58] and later reviewed by Razo-Mendivil et al. [53], compared to that originally described as the type-species C. creptotrema. The main morphological characteristic of this new species, compatible with the new diagnosis of Creptotrema proposed herein, is the presence of a subterminal oral sucker with a single muscular lobe on either side of the oral sucker, with a broad base, stretching from the ventrolateral to the dorsolateral side, and with free dorsal ends.
Creptotrema conconae n. sp. is morphologically similar to Creptotrema astyanace comb. n. (= Auriculostoma astyanace) (type-species of Auriculostoma), sharing characteristics such as testes in tandem, oval, and with entire margin (not lobed); uterus completely pretesticular, extending to the genital pore; cirrus-sac long and sinuous, and distribution of the vitelline follicles (both intra-and extracecal and extensively confluent in the post-testicular region) (Table 5). However, the new species and C. astyanace comb. n. differs by the smaller total body length and width of C. conconae n. sp. (on average 1463 Â 298 versus 2900 Â 488 in C. astyanace comb. n.); in the shape of the oral sucker, varying from subspherical to funnel-shaped in the new species versus funnel-shaped in C. astyanace comb. n.; and in the position of the ovary (postequatorial and sinistral in the new species versus pre-equatorial and median in C. astyanace comb. n.). Although the eggs are operculated and similar in shape in both species, those observed in C. conconae n. sp. are larger (62 Â 44) and less numerous (no more than 40 eggs in completely developed adults) versus smaller (55-59 long Â 34-41 wide) and numerous (40-60 eggs) in C. astyanace comb. n. (see [58]).
Creptotrema conconae n. sp. is the first species of the genus described in heptapterid fishes and is morphologically similar to other species described in fishes belonging to Pimelodidae (Siluriformes) from South America, such as Creptotrema platense comb. n. (= Auriculostoma platense) and C. paraense from Argentina and Brazil, respectively ( Table 4). The new species differs from these other congeners in its smaller body, narrower and more elongated than that described for other species of parasites of siluriform fishes. The new species presents a long and sinuous cirrus-sac, exceeding the ventral sucker posteriorly and reaching the ovary, but not longer than that described for C. paraense (around 400 versus 319 lm in the new species) versus a short cirrus-sac that does not reach beyond the posterior margin of the ventral sucker observed in C. platense comb. n.
Creptotrema conconae n. sp. and C. paraense share more morphological characteristics than C. platense comb. n.. The distribution of vitelline follicles is confluent in the posterior region of the testes in both species (and not confluent in C. platense comb. n.); the arrangement of the testes is in tandem in the new species and C. paraense (versus oblique in C. platense comb. n.) (Table 5). However, C. paraense has larger testes than the new species (210 lm versus 130 lm in diameter in the new species); the testes are arranged   Pre-and intertesticular, intercecal/Eggs between genital pore and mid-level of testes (Continued on next page) in tandem in both species, but in C. paraense, they do not occupy a regular position (one immediately juxtaposed to the other), as observed for the new species. In addition, the ovary of C. paraense has a more rounded and regular shape than that of the new species (obliquely oval to irregular in shape), and the eggs are slightly smaller (56 Â 42 lm versus 62 Â 44 lm in the new species) and more numerous than in the new species. Additionally, the size of the oral sucker is larger than that of the ventral sucker (in proportion) in C. paraense, whereas they are similar in size in the new species. Material deposited: holotype CHIOC 39598 and paratypes CHIOC 39599; 39600 and 39701 Representative DNA sequence: 1278 bp long sequence of the 28S rDNA gene -GenBank OK044373; 442 bp long sequence of COI mtDNA -GenBank OK075293.
Etymology: The epithet refers to the type-host Characidium schubarti.
In addition, Creptotrema schubarti n. sp. can be distinguished from other congeners that have symmetrical to oblique testes (C. creptotrema, C. pati, C. lynchi, and C. megacetabularis n. sp.) by the distribution of eggs in the uterus. The eggs may occupy the region between the genital pore and may exceed the testes region, extending into the extracecal area in the new species, whereas, in other congeners, the eggs are distributed preferentially in the pretesticular and inter-testicular regions (as in C. lynchi), but may extend posteriorly to the testes, but typically not extracecal (as in C. creptotrema and C. megacetabularis n. sp.). In C. pati, the eggs are also distributed in the uterus in the inter-testicular region and may extend to the post-testicular region, but the uterus is intercecal (not extending to the extracecal region as in the new species), and the testes are extracecal (versus intra-cecal in the new species) ( Table 5). Material deposited: holotype CHIOC 39703; paratypes CHIOC 39704 a-b; 39705 and 39706 a-h.
Remarks: This species was erroneously identified as C. creptotrema by Yamada et al. [73]; however, it is herein recognized and described as a new species. As previously reported for C. creptotrema, C. megacetabularis n. sp. also shows variation in the morphology of specimens according to their ontogenetic development (adults in the early development phase are relatively smaller, have fewer eggs in the uterus, and the difference in the oral/ventral sucker is proportionally more evident than in completely developed adults). However, there was no intraspecific genetic divergence within the sequences of the morphotypes identified as adults in the early development phase and completely developed adults (see "Molecular data and phylogenetic inference" section).
The morphology of this new species resembles that of other congeners that also have symmetrical to oblique testes: C. creptotrema, C. pati, C. lynchi, C. guacurarii comb. n., and C. schubarti n. sp. (Table 5). However, the new species can be distinguished by the following features: the presence of a large ventral sucker, with almost twice the size of the oral sucker (oral/ventral sucker ratio nearly 1:2) (this difference may be even more evident in adults in the early development phase), well evidenced by the proportion of the size and width of the body of the parasite, and eggs with a knob opposite to the operculum.
Although there are some similarities in the morphology of this new species and C. lynchi reported from toads from Colombia [9], differences in the average measurements and proportions of these structures can be observed. The testes are elongated in both species, but they are smaller (169 Â 112 lm on average) and symmetrical in C. megacetabularis n. sp. versus larger (300-480 Â 192-252 lm) and symmetrical to oblique in C. lynchi; the cirrus-sac of the new species is smaller, approximately 428 lm versus 444-660 lm in C. lynchi.
The testes are intra-cecal in this new species, C. creptotrema, C. lynchi, C. guacurarii comb. n., and C. schubarti n. sp. (sometimes cecal), but they are extracecal in C. pati. The total length of the body (on average) of the new species (1188 lm) was similar to that observed in C. guacurarii comb. n. (1254 lm) and in the completely developed C. creptotrema adults analyzed in the present study (1248 lm), but was larger than that in C. creptotrema specimens analyzed in the original description (562 lm) and C. schubarti n. sp. (802 lm).

Molecular data and phylogenetic inference
We successfully obtained eight partial sequences of the 28S rDNA: two of C. creptotrema  The 28S rDNA final alignment was 1038 bp long. The I ss indicated no saturation in either the transitions or transversions; the critical index of substitution saturation (I ss.c ) values were above the I ss values.
The 28S rDNA ML and BI phylogenetic analyses presented identical topologies, with most clades being well supported. Both analyses also recovered the Allocreadiidae as monophyletic. In the final 28S rDNA phylogenetic reconstruction (Fig. 7), two main clades were observed within the Allocreadiidae clade (pp = 1; bootstrap = 82). One comprised Acrolichanus auriculatus (Wedl, 1858) Ward, 1917, parasites of sturgeons (Acipenseriformes) from the Holarctic region (pp = 1; bootstrap = 100), and the other included all the other sequences of allocreadiids from Neotropical and Holarctic regions (pp = 0.91; bootstrap = 74). The clade composed of allocreadiids from the Neotropical region comprised sequences of the genera Creptotrematina, Wallinia, Creptotrema (sensu stricto clade, including "Auriculostoma" spp.), Paracreptotrema, Paracreptotrematoides, and Pseudoparacreptotrema; the North American Megalogonia ictaluri Surber, 1928 appears as a sister taxon of this clade. Creptotrema funduli from Mississippi (USA) is basal to the clade composed of allocreadiids from the Neotropical region and M. ictaluri.
The newly generated sequences of Creptotrema spp. formed a clade referred to as Creptotrema sensu stricto (pp = 1; bootstrap = 99), with Wallinia spp. recovered as its sister group. The Creptotrema sensu stricto clade included the sequences of the type-species C. creptotrema, the sequences of the new species described herein, and all sequences previously deposited in the GenBank as Auriculostoma spp. Creptotrema funduli, the only sequence of the genus previously deposited in GenBank, was not positioned within the Creptotrema sensu stricto clade, suggesting that this species might not belong to Creptotrema and should be reevaluated (Fig. 7).
Among the Creptotrema sensu stricto clade, C. schubarti n. sp. and C. guacurarii comb. n. clustered together, both recovered from hosts belonging to Characidium from Brazil and Argentina, respectively. These species were closely related to C. creptotrema, forming a sister group of Creptotrema tica comb. n. (= A. tica) (all species preferentially parasitize characiform fishes). Creptotrema megacetabularis n. sp. (from siluriforms belonging to Auchenipteridae) appeared as a sister taxon in this group. Creptotrema conconae n. sp. (the only species that parasitizes heptapterids) appeared as a sister clade of this group, clustering together with C. astyanace comb. n. and C. lobata comb. n.
For the partial 28S rDNA, the interspecific genetic divergences found among the sequences of the new species and C. creptotrema (type-species) varied from 2.1 to 5.2% (21-49 bp); among Creptotrema spp. (except C. funduli) and species previously described as Auriculostoma, it varied from 0.4% to 5.0% (6-49 bp); from Creptotrema spp. (including the new sequences and Auriculostoma spp. reassigned herein) and Wallinia spp. (its sister clade), it varied between 2.9% and 6.9% (39-63 bp). There was no intraspecific genetic divergence within the sequences of C. creptotrema and C. megacetabularis n. sp. that presented differences between their morphology. For the other species of this genus, only one sequence per species was available; therefore, intraspecific comparisons could not be inferred. See Supplementary Table S1 for information on the genetic divergence values using the 28S rDNA gene.
The final COI mtDNA alignment was 352 bp long. The I ss indicated no saturation in either the transitions or transversions, and the critical index of substitution saturation (I ss.c ) values were above the I ss values.
For the mtCOI gene, the genetic divergences between all Creptotrema spp. varied from 6.6 to 16.4% (21-45 bp) (see Supplementary Table S2). No intraspecific genetic divergence was observed between the three C. creptotrema sequences; for the other species of this genus, only one specimen was sequenced.

Discussion
Information regarding the species diversity of allocreadiids from Neotropical freshwater fishes has increased recently as a result of integrative taxonomic approaches involving morphological and phylogenetic analyses [25]. However, there is still a shortage of information about these digeneans that may generate inconsistencies in species delimitations (e.g., Crepidostomum, Paracreptotrema, Megalogonia, and herein demonstrated for species of Auriculostoma and Creptotrema), their relationships, and host-parasite associations [17,24,25,49].
The accurate morphological analyses of type-specimens of Creptotrema deposited in museums and the phylogenetic positioning among isolates provided in this study allowed us to formally describe three new species and to amend the diagnosis of Creptotrema. We observed intraspecific morphological variation among specimens of C. creptotrema collected in this study, C. conconae n. sp., and C. megacetabularis n. sp. However, the absence of intraspecific genetic divergence between the sequences of the different morphotypes of each species (reported as C. creptotrema and C. megacetabularis n. sp., herein referred to as adults in the early development phase and completely developed adults), reinforces the hypothesis that  Table 2). these morphological differences represent distinct stages of ontogenetic development of these species.
Although we have found low interspecific genetic divergences between the partial 28S rDNA sequences of Creptotrema schubarti n. sp. and C. guacurarii comb. n. (0.4%, 6 bp), the set of morphological differences presented herein supports the erection of the new species. These are the only sequences of the Creptotrema clade sensu stricto generated from parasites of fishes belonging to the genus Characidium from Brazil and Argentina, respectively, which may have contributed to the low divergence observed. The 28S rDNA represents a highly conserved fragment and is not an ideal molecular marker for the delimitation of species or for making inferences about interspecific divergences. Unfortunately, to date, there are no other deposited sequences of C. guacurarii comb. n. from other molecular markers useful for species delimitation, such as the mtCOI gene, which would enable a more precise analysis of molecular divergences between them. However, a comprehensive and meticulous analysis of the morphological features of both species provided important evidence for their delimitation.
The molecular evidence demonstrated herein identified Auriculostoma as a synonym of Creptotrema, and supports the monophyly of the clade Creptotrema sensu stricto. Based on our morphological analyses, it is possible to affirm that the morphology of the oral sucker and muscular lobes ("auricles") is highly variable among Creptotrema spp. (considering the species previously allocated in Auriculostoma), and represents a shared feature among other "papillose" allocreadiids as a result of convergent evolution. Therefore, this character should not be considered sufficient to delimitate and distinguish species of Creptotrema or genera of "papillose" allocreadiids, since this character arose or was secondarily lost several times during the evolutionary history of the group [46].  Table 2). We recommend the use of different tools for consistent morphological and molecular analyses to diagnose the species.
According to our 28S rDNA phylogenetic inference (see Results section), the sequence of C. funduli was not grouped within the Creptotrema sensu stricto clade. Considering that this sequence was not sourced from the type-host or type-locality [15,46], we hypothesize that the specimen might not belong to Creptotrema and may represent a new genus yet to be properly described, corroborating the paraphyly of the genus recently proposed by Pérez-Ponce de León et al. [46]. After analyzing the types deposited in museums and based on our molecular data, C. funduli is proposed herein as species inquirenda; future studies must be conducted to resolve its taxonomic status.
Although the morphology of C. funduli resembles that of other species assigned to Creptotrema, this species shows conspicuous differences in its diagnostic characteristics. Creptotrema funduli was considered incorrectly assigned to this genus by Manter [38]; however, the species continued to be allocated erroneously to this genus [26]. Manter [38] also analyzed 10 co-types and observed conspicuous finger-like lobe projects inward from the dorsal wall of the oral cavity, which is not compatible with the original description of Creptotrema, corroborating the molecular analysis here.
Creptotrema mulleri Coil & Kuntz, 1960, the only species supposedly belonging to this genus and described outside the American continent, was considered a synonym of Crepidostomum farionis [38]. Manter [38] examined the type and paratype specimens of C. mulleri and observed four partially retracted papillae (= lobes) on the dorsal surface of the oral sucker. Considering this morphological feature and the fact that C. mulleri was described from specimens collected in salmonids (trout) from Turkey, its supposed relationship with Creptotrema spp. was not accepted in this study.
Creptotrema is now one of the most widely distributed genera of trematodes across the Americas, represented by 19 valid species of parasites of freshwater teleosts and anurans (15 from South America, two from Central America, and two from North America), and South America seems to be an important center of diversification for these allocreadiids. Excluding the nominal species C. funduli, which requires a taxonomic and systematic review, it is possible that the Neotropical species of Creptotrema (Creptotrema sensu stricto clade) comprises a separate monophyletic group as suggested by Choudhury et al. [13]. Although the definitive hosts of Creptotrema spp. are preferentially freshwater fishes belonging to multiple orders (Characiformes, Gymnotiformes, Perciformes, and Siluriformes), excluding C. lynchi as originally described in toads, there seems to be a close relationship between this lineage of digeneans and characiforms, in which 10 of the 19 valid species have been reported (Table 4).
Crepidostomum was originally described to include parasites of salmoniforms, but some species have also been reported in cyprinodontiforms, hiodontiforms, perciforms, gadiforms, and scorpaeniforms ( [21] and references cited in Table 2). The validity of many species of Crepidostomum has been questioned because few characters can be used for species delimitation, some of which are shared among other "papillose" allocreadiids, and some of the available molecular data are not supported by voucher specimens [21,49,66]. In our analyses using the 28S rDNA gene, the clade composed of Margotrema spp., C. cornutum, and C. auritum may indicate a closer relationship among these parasites (all sequences from the USA and Mexico), indicating the need for supplementary studies regarding their morphology and new molecular data to investigate whether they may belong to the same genus or if they may be erected in a new genus.
Considering that South America is exceedingly rich in freshwater fish fauna (approximately 5160 species, according to Reis et al. [57]), less than 5% of these potential hosts have been examined for parasites, making it difficult to predict the total trematode diversity [12]. This percentage is much higher in North America (45%) and Central America (32%); however, these subcontinents comprise a lower species richness of freshwater fishes (1213 and 299 species, respectively) than South America. This highlights the need for further parasitological surveys throughout the Americas to improve knowledge about the parasite species richness and host-parasite relationships, especially for South America, along with the use of different tools for morphological analyses (e.g., light microscopy and SEM) and molecular studies (employment of different markers for phylogenetic analyses). Our results describe new species, molecular data, and important information, increasing our understanding of the relationships among allocreadiids, especially Creptotrema, contributing substantially to clarifying the phylogeny of this genus.

Conflict of interest
The authors declare that they do not have any conflict of interest.

Supplementary material
The supplementary material of this article is available at https://www.parasite-journal.org/10.1051/parasite/2021065/olm Table S1. Nucleotide divergence (p-distance expressed in %) estimated for the 28S rDNA among Creptotrema spp. and selected digeneans Table S2. Nucleotide divergence (p-distance expressed in %) estimated for the mitochondrial cytochrome c oxidase I (COI mtDNA) among Creptotrema spp. and selected digeneans