Annotated checklist of fish cestodes from South America

Abstract An exhaustive literature search supplemented by a critical examination of records made it possible to present an annotated checklist of tapeworms (Cestoda) that, as adults or larvae (metacestodes), parasitize freshwater, brackish water and marine fishes, i.e. cartilaginous and bony fishes, in South America. The current knowledge of their species diversity, host associations and geographical distribution is reviewed. Taxonomic problems are discussed based on a critical evaluation of the literature and information on DNA sequences of individual taxa is provided to facilitate future taxonomic and phylogenetic studies. As expected, the current knowledge is quite uneven regarding the number of taxa and host-associations reported from the principal river basins and marine ecoregions. These differences may not only reflect the actual cestode richness but may also be due to the research effort that has been devoted to unravelling the diversity of these endoparasitic helminths in individual countries. A total of 297 valid species, 61 taxa identified to the generic level, in addition to unidentified cestodes, were recorded from 401 species of fish hosts. Among the recognized cestode orders, 13 have been recorded in South America, with the Onchoproteocephalidea displaying the highest species richness, representing c. 50% of all species diversity. The majority of records include teleost fish hosts (79%) that harbour larval and adult stages of cestodes, whereas stingrays (Myliobatiformes) exhibit the highest proportion of records (39%) among the elasmobranch hosts. Fish cestodes are ubiquitous in South America, being mostly recorded from the Warm Temperate Southeastern Pacific (WTSP; 31%) for marine hosts and the Amazon River basin (45%) for freshwater ones. The following problems were detected during the compilation of literary data: (i) unreliability of many records; (ii) poor taxonomic resolution, i.e. identification made only to the genus or even family level; (iii) doubtful host identification; and (iv) the absence of voucher specimens that would enable us to verify identification. It is thus strongly recommended to always deposit representative specimens in any type of studies, including faunal surveys and ecological studies. An analysis of the proportion of three basic types of studies, i.e. surveys, taxonomic and ecological papers, has shown a considerable increase of ecological studies over the last decade.


Introduction
Tapeworms (Cestoda) are a monophyletic assemblage of flatworms (Phylum Platyhelminthes) and they are obligate internal parasites of vertebrates. Their complex lifecycles include one or more intermediate hosts in a wide array of animal phyla (mostly arthropods) and they are exclusively transmitted perorally, i.e. via the food chain (Caira and Littlewood 2013;Littlewood et al. 2015). The cestodes are the second speciesrichest group of platyhelminths, with more than 5000 species in 751 recognized genera that have radiated through marine, freshwater and terrestrial environments (Waeschenbach et al. 2012;Caira and Littlewood 2013).
Cestodes parasitizing elasmobranchs and teleost fishes in at least one stage of development comprise one of the most diverse lineages of tapeworms , only comparable in species richness with cyclophyllidean cestodes, parasites of tetrapods (Caira and Littlewood 2013). Since these parasites usually exhibit a strict host specificity, they are considered suitable models for studies of host-parasite co-evolution  or even helping in circumscribing species boundaries of cestode hosts (Caira and Jensen 2015).
South America is a megadiverse continent, including at least five of the world's biodiversity 'hotspots', more than 30.000 km of coastline and two of the 10 largest freshwater drainage systems of the world, i.e. the Amazon and Paraná River basins, which is reflected in its species-rich ichthyofauna (Myers et al. 2000;Miloslavich et al. 2011;Reis 2013). Bearing that in mind, one might expect a high diversity of fish cestodes as well, even though there are no comprehensive checklists or other faunistic studies encompassing the whole continent that could provide an overview of the cestode diversity, except those with regional focus (e.g. Thatcher 2006 for Amazonia; Tantaléan and Huiza 1994 for Muñoz and Olmos 2008 for Chile).
Studies on fish cestodes from South America date back to the early 19th Century, when C. A. Rudolphi described Anthocephalus macrourus Rudolphi, 1819 from an unidentified sparid fish and Anthocephalus interruptum Rudolphi, 1819 (both cestodes of the order Trypanorhyncha) from Trichiurus lepturus Linnaeus off the Brazilian coast, even though these species are no longer valid (Campbell and Beveridge 1996). Subsequently, K. M.  described several species that are now included in three different orders. Both workers studied cestodes collected by renowned naturalists, such as F. Sellow, I.F.W.M. Von Olfers and J. Natterer. With few exceptions, the number of descriptions and/or reports have considerable increased in the 20th Century and a large amount of information has been generated, yet many studies are faunal surveys dispersed in regional journals that are not readily accessible.
Detailed taxonomic studies combining morphological and molecular approaches have recently expanded our knowledge at lower and higher taxonomic levels, mostly under the framework of the National Science Foundation (Planetary Biodiversity Inventory program) funded project called "A survey of the tapeworms (Cestoda: Platyhelminthes) from vertebrate bowels of the earth" (see http://tapewormdb.uconn.edu). This project funded, amongst others, intensive research on fish cestodes in South America, which were mainly undertaken by A. de Chambrier, T. Scholz and A. A. Gil de Pertierra for teleost hosts, and V. A. Ivanov, F. P. L. Marques and F. Reyda for elasmobranch hosts. The present paper aims at addressing the following objectives: (1) to provide for the first time an annotated checklist that summarizes records of cestodes in marine and freshwater fishes from South America, including detailed information on their hosts, site of infection, geographical distribution, stage of development and molecular data; (2) to critically assess some doubtful reports; and (3) to depict the problems that impede a better understanding of the diversity and host associations of cestodes in South America.

Materials and methods
Parasite-host and host-parasite checklists for fish cestodes from South America were compiled on the basis of an exhaustive search of literature published until August 2016; abstracts of meetings, theses and reports without primary data were not considered. The bibliographic search was complemented by the information gathered from Helminthological Abstracts, Host-Parasite Database of the Natural History Museum, London (Gibson et al. 2005), Global Cestode Database (GDC) , Google Scholar, ScienceDirect, Web of Knowledge, as well as some previously published books (e.g. Palm 2004;Thatcher 2006). The classification of cestodes proposed by Khalil et al. (1994) is basically followed, but it is updated based on revisional papers on individual cestode orders or molecular phylogenetic studies at the ordinal level, such as Kuchta et al. (2008) for bothriocephalideans and diphyllobothriideans, Olson et al. (2010) for trypanorhynchs,  for onchoproteocephalideans, phyllobothriideans and 'tetraphyllideans', Healy et al. (2009) and Ruhnke et al. (2015) for rhinebothriideans, and Jensen et al. (2016) for lecanicephalideans.
The species are arranged according to taxonomic categories and are presented in alphabetical order followed by data on their hosts (species name, class and family), habitat, site of infection, stage of development, marine ecoregion according to Spalding et al. (2007), river basins or lakes, country and references (between parentheses). All cestodes presented herein follow the most recent taxonomic literature and the validity of individual taxa or the reliability of their records were critically assessed by the present authors, who consulted with experts for some tapeworm groups.
Host species are arranged in taxonomic and then alphabetical order. The scientific names of hosts have been updated based on Froese and Pauly (2016) and supplement-ed by the most recent taxonomic papers for certain problematic taxa (e.g. Cichla Bloch & Schneider, Pseudoplatystoma Bleeker and Zungaro Bleeker).
The following abbreviations are used for collections: The following abbreviations are used for molecular markers:

18S
small subunit of the nuclear ribosomal RNA gene; ITS1 first nuclear ribosomal internal transcribed spacer; 5.8S 5.8S ribosomal RNA gene; ITS2 second nuclear ribosomal internal transcribed spacer; 28S large subunit of the nuclear ribosomal RNA gene; 16S large subunit of the mitochondrial ribosomal RNA gene; cox1 cytochrome c oxidase I The following abbreviation is used for records of metacestodes in the host-parasite list: L larvae. * Asterisks in the parasite-host list indicate the type species of the genus.

Family Proteocephalidae La Rue, 1911
[Even though recent molecular data suggest that most of the traditionally recognized subfamilies are artificial, i.e. non-monophyletic, we are following Woodland's subfamilial classification for practical reasons]
Notes: type host; it was originally reported as Silurus sp., but this genus only

Taxa incertae sedis
Anindobothrium anacolum (Brooks, 1977 (Brooks 1977). Notes: type host. The genus Anindobothrium is likely a member of the Rhinebothriidea (see Ruhnke 2011) and it is already treated as such in the GCD within Anthocephalidae .

Paragrillotia sp.
Dipturus trachyderma (Elasmobranchii: Rajidae); marine; site of infection and stage of development not given; WTSP; Chile (Leible et al. 1990). Notes: host reported as Raja trachyderma. Neither were the three known species of Paragrillotia Dollfus, 1969 described from rays nor reported from Southeastern Pacific (see Beveridge and Justine 2007a). Since the vouchers were apparently not deposited, we considered this record doubtful.

Nybelinia africana Dollfus, 1960
Pseudupeneus maculatus (Actinopterygii: Mullidae); marine; body cavity; metacestode; TSA; Brazil (Palm 2004 (Pereira and Boeger 2005). Note: the taxonomic status of this species is problematic since the type material is a mixture of different species under the same name (Palm 2004).

Results and discussion
The database compiled from the available literature on fish cestodes in South America comprises records of 297 species recognized as valid as well as unidentified ones included in 120 genera and 32 families, associated with 401 cartilaginous and bony fish hosts (Tables 1, 2). Among the recognized 19 orders of tapeworms, 13 have been found in marine and freshwater systems in South America (excluding the doubtful reports of the Caryophyllidea). The recently erected order Onchoproteocephalidea, which accommodates several taxa previously placed in the tetraphyllidean family Onchobothriidae and the entire former order Proteocephalidea, is the most diverse group, being represented by 148 species in 43 genera. The tapeworm with the widest spectrum of definitive hosts is Rhinebothroides freitasi (Rhinebothriidea) that parasitizes nine species of stingrays of the genus Potamotrygon, even though it exhibits only a mesostenoxenous specificity, i.e. occurrence limited to a single host genus. Conversely, members of five orders, namely Amphilinidea, Cathetocephalidea, Diphyllidea, Lecanicephalidea, 'Tetraphyllidea' and most likely Gyrocotylidea (see the checklist records for details), showed only a single fish host (oioxenous specificity). It is also worth noting the usually broad spectrum of intermediate teleost hosts for metacestodes, mainly diphyllobothriideans, 'tetraphyllideans' and trypanorhynchs, which is reflected in the higher number of actinopterygian (315) than chondrichthyean (86) hosts. However, the stingray Potamotrygon motoro harbours the highest number of cestodes (17) belonging to the species-rich genera Acanthobothrium Blanchard, 1848, Potamotrygonocestus Brooks & Thorson, 1976, Rhinebothrium Linton, 1890and Rhinebothroides Mayes, Brooks & Thorson, 1981, in addition to Paroncomegas araya.
A total of 208 species of tapeworms are found across seven major ecoregions of South American coast (one additional species is found in Galapagos), being the highest species richness reported from WTSP (66) and WTSA (60), whereas 209 species are found throughout six major river basins of South America (Fig. 1). The major number of species comes from the Amazon and Paraná River basins, with 95 and 80 species, respectively. At least four species were reported from particular lakes, mostly parasitizing osmeriforms and salmoniforms in Argentina and Chile.
The number of taxonomic studies has been steadily growing since 1940, but only 16 papers were based on an integrated taxonomy approach, using molecular data as an important tool. The number of general parasitological surveys has also increased since the beginning of the last century, whereas ecological studies have launched the first publications only in the mid-sixties, with a peak in the last sixteen years, noticeably higher than the previous period (Fig. 2). *Only sequences of cestodes collected in South America were considered.

Taxonomic resolution
Among the genera of fish cestodes reported from South America, one half was either identified only at generic level or they were specifically identified in some reports and at generic level in others. The numerous papers published in the last 30 years, mostly those ecological ones (see Fig. 2) focused on marine teleost hosts as models, include a high number of records of unidentified larvae. Most of them corresponded to the Table 2. Survey of fish hosts that harbour cestodes in South America. Figure 1. The geographical distribution of tapeworms in South America associated with their fish hosts from the major marine ecoregions of Spalding et al. (2007) and river basins in the continent. Each species may occur in more than one basin or ecoregion.
achieved and those parasites in fish hosts exhibit lower taxonomic resolution than endohelminths parasitizing birds and mammals. The accurate identification of larval stages of cestodes is usually challenging, because they lack key morphological traits that are present in their adult forms, and studies dealing with their genetic characterization are rare in South America (Rozas et al. 2012). An even more important concern is the high number of records of uni- dentified diphyllobothrid plerocercoids in teleosts (see the Parasite-Host list), because these metacestodes can infect humans who consume raw or undercooked fish and may cause a disease known as diphyllobothriosis (Scholz et al. 2009;Kuchta et al. 2015). Larval trypanorhynchs are the only exception, because they may be precisely identified based on their tentacular armature (Palm 2004;. For instance, all three valid species of Pterobothrium Diesing, 1850 originally described from South America have teleost fishes as type hosts. One of the main obstacles that hampers our understanding of the diversity of fish cestodes in South America is the deficient knowledge of their life cycles and failure to match the morphologically amorphous or divergent larval forms to their adult stages; to date, no life cycle studies have been undertaken in this continent. Jensen and Bullard (2010) performed the most comprehensive study combining molecular and morphological approaches to elucidate life cycles of marine cestodes from four metazoan phyla in the Gulf of Mexico. They found as many as eight larval types which could be associated with their adult forms and provided a useful morphological key for the 15 recognized types, including larvae of the currently recognized Onchoproteocephalidea, Phyllobothriidea, Rhinebothriidea and 'Tetraphyllidea'.
Unlike the poor taxonomic resolution of marine larvae from teleosts, adult forms, typically those infecting freshwater catfishes (Siluriformes) and potamotrygonid sting-rays (Potamotrygonidae), have been fairly well-documented (Reyda and Marques 2011;de Chambrier et al. 2015b). Their characterisation using modern descriptive tools, e.g. scanning electron micrographs and molecular data, associated with the traditional morphological approach, deeply contributed to the improvement of their taxonomic resolution and to elucidating the high cestode diversity associated with these groups of hosts. Miloslavich et al. (2011) estimated the fish diversity in five subregions along the South American coast and suggested the occurrence of more than 5000 species in these marine systems. Reis (2013) estimated a value slightly higher for fishes from freshwater drainage systems in South America, c. 5400 species. Considering that predictions for estimating the global species richness of parasites suggest that they exceed twice the number of their hosts (Dobson 2008) and that only 4% of the potential fish hosts have been scrutinized for cestodes in South America, it is straightforward to conclude that our knowledge of the diversity of these parasites is far from adequate. Similar results were also found for trematodes infecting freshwater fishes in the same continent (Choudhury et al. 2016) and it may be valid also for others groups of helminths.

Elasmobranch and teleost fish hosts
Contrasting the generally poorly-known diversity of fish cestodes in South America, some groups of hosts have been extensively studied compared to others. Among the elasmobranch hosts, the stingrays (Myliobatiformes) have been steadily examined for tapeworms, exhibiting the highest proportion of records (39%), which were mainly reported from marine and freshwater systems (e.g. Brooks et al. 1981a, b;Reyda and Marques 2011). Regarding teleosts, members of the order Perciformes are the most representative hosts, representing c. 40% of all records among this group. The majority of these studies have been conducted by ecological research teams interested in unravelling the structure of fish parasite communities and, more recently, their use as biological tags for stock discrimination (e.g. Luque et al. 2010;Timi et al. 2010a).
According to Luque and Poulin (2007), the study effort and local priorities of research teams play an important role on the uneven knowledge of parasite species richness in Neotropical fishes. Since cestodes are ubiquitously distributed in fishes from South America, it is likely that the higher the number of elasmobranchs and teleosts examined in parasitological surveys, the higher the number of parasite-host associations that will be identified.

Accurate identification of fish hosts
During the development of this checklist, we have faced several examples of problematic identification and controversial taxonomy of hosts, which may compromise the reliability of any parasitological survey and limit our understanding of host specificity, the rela-tionship between parasite and host phylogenies, as well as the establishment of trophic links elucidated by life-cycle studies (Naylor et al. 2012). Some genera, such as Cichla, Pimelodus, Potamotrygon, Pseudoplatystoma and Zungaro, have a convoluted taxonomic history and their species boundaries can diverge depending upon the approach used. Kullander and Ferreira (2006) for instance, recognized 15 species of Cichla distributed across South American rivers, based on morphological characters. However, Willis et al. (2012) recognized only eight species using multi-locus genetic data, suggesting that the number of Cichla species in South America may have been overestimated.
Therefore, we recommend that parasitologists keep a piece of host tissue in a molecular-grade ethanol for sequencing and to work in synergy with fish taxonomists to be as accurate as possible in fish identification, as already advocated by Naylor et al. (2012) for elasmobranch hosts.  provided a field-sampling protocol that may be useful not only for parasite taxonomists, but also for those who are interested in general host-parasite associations. Poulin et al. (2016) tested the completeness of 25 checklists of metazoan parasites in vertebrate hosts from several geographic regions based on three approaches. None of the studies analyzed performed well and only three of them passed two of the tests. Several obstacles contribute to a lack of completeness of checklists, including: (1) the reliability of information depends on the accuracy of the description or report; (2) geographically biased studies may not reflect the real distribution of diversity; (3) cryptic species, i.e. genetically distinct species that look similar morphologically, may contribute to an underestimate of the true number of species; and (4) only a small fraction of the potential fish hosts in South America have been examined for parasites. To mitigate these issues, we have attempted to critically gather as much information as possible and have obtained expert opinions. Therefore, we hope that we provide here the most robust database up to date that may help in a reliable estimation of the true diversity of fish cestodes in South America.

Conclusions
the National Science Foundation PBI awards Nos. 0818696 and 0818823, Institute of Parasitology (institutional support RVO 60077344) and Czech Science Foundation (P505/12/G112). PVA was supported by a postgraduate fellowship from CNPq.