Levisunguis subaequalisCurran, Overstreet, Collins & Benz, 2014, was recently described from the lungs of the definitive hosts, softshell turtles, Apalone ferox (Schneider, 1783), and Apalone spinifera aspera (Agassiz, 1857) as well as the viscera of an intermediate host, the western mosquitofish, Gambusia affinis (Baird and Girard, 1853). However, the original account lacked molecular data. Furthermore, histological examination of infected host tissues in the original account of L. subaequalis did not reveal any pathological changes in the intermediate host. The present work provides a robust morphological description of the nymph and novel molecular data from the 18S and 28S ribosomal gene regions and the cytochrome c oxidase subunit 1 (COI) mitochondrial gene. Phylogenetic analyses using Bayesian inference and maximum likelihood analysis with concatenated sequence data from these 3 regions, as well as each region individually, placed the turtle pentastomid L. subaequalis as a sister clade to the crocodilian pentastomids of the genus Sebekia Sambon, 1922. While only concatenated phylogenetic analyses agreed with the currently accepted classification of the Eupentastomida and phylogenetic signal assessment indicated that the concatenated data set yielded the most phylogenetic signal, data from more taxa are still needed for robust phylogenetic inferences to be made. The intensity of infection ranged from 2 to 171 nymphs per fish, compared with the highest previously reported intensity of 6. These high-intensity infections with L. subaequalis were characterized by the nymphs occupying 5–50% of the coelomic cavity of G. affinis. However, despite this heavy parasite infection, fish exhibited minimal pathology. Observed pathology was characterized by compression or effacement of organs adjacent to the nymphs, particularly liver, swim bladder, and intestines, as well as the formation of granulomas around shed pentastomid cuticles. Nonetheless, the morphological and molecular data provided in the present work will bolster future efforts to identify this pentastomid in other hosts where pathology may be present in addition to aiding in the advancement of the field of molecular pentastomid systematics.
Eupentastomida Waloszek, Repetski & Maas, 2006, is an enigmatic class of the extant crustacean endoparasites of the phylum Pentastomida Huxley, 1869, with adults infecting the respiratory tracts or lungs of mammalian, reptilian, or avian definitive hosts and nymphs generally encysted in vertebrate or invertebrate intermediate hosts (Christoffersen and de Assis, 2015). The nomenclature of pentastomids is characterized by a number of taxonomic anomalies including misspelled taxa, incorrect attributions of taxonomic authority names and dates, and incorrectly nominated type species, as recently cataloged and corrected by Poore (2012). The systematics of this class have culminated in the present classification system of Christoffersen and de Assis (2013), which recognizes 141 nominal species in 26 genera. This system was derived by combining previous morphology-based phylogenies of pentastomids by Almeida and Christoffersen (1999) and Junker (2002). While the aforementioned authors have invested considerable effort into resolving the nomenclature of pentastomids and relationships between pentastomid taxa using morphological characters, application of molecular techniques may further clarify these relationships. Both the adult and the nymph of the pentastomid Levisunguis subaequalis Curran, Overstreet, Collins & Benz, 2014, have been recently described (Curran et al., 2014). Adults were described from the Gulf Coast spiny softshell turtle, Apalone spinifera aspera (Agassiz, 1857), and the Florida softshell turtle, Apalone ferox (Schneider, 1783), while the nymph was described from the western mosquitofish, Gambusia affinis (Baird and Girard, 1853). Additionally, nymphs of Levisunguis sp., identified only at the genus level, have been reported from G. affinis by Carpenter (2018). Although morphological data are available for the adult of both sexes and to a lesser extent the nymph of L. subaequalis, there are at present no molecular data for this species. Also, no description of pathological effects associated with infection in the intermediate host is presently available. Furthermore, while this species was described immediately after the publication of the currently accepted classification of pentastomids, the phylogenetic placement of Levisunguis in this system is not clear because Christoffersen and de Assis could not reference the as-yetundescribed L. subaequalis. The present work provides the first molecular data for this species and tentative phylogenetic analyses used to infer the placement of L. subaequalis within the context of current pentastomid systematics. Phylogenetic signal assessment was also carried out to assess the relative taxonomic utility of 18S and 28S ribosomal genes and the cytochrome c oxidase subunit 1 (COI) mitochondrial gene as well as concatenated alignments of all 3 loci. Also provided is a detailed morphological description of the nymph and a description of pathologic changes associated with high-intensity infections in the only identified intermediate host, G. affinis. These data will bolster future identification of this pentastomid and aid in inferring the evolutionary history of pentastomids by allowing for more robust phylogenetic analyses.
MATERIALS AND METHODS
Parasite collection
In July 2018, 44 specimens of G. affinis were collected via hand netting from an ornamental pond in Birmingham, Alabama. Fish were morphologically identified as G. affinis and not Gambusia holbrooki Girard, 1859, according to Boschung and Mayden (2004) based on the presence of 6 dorsal fin rays, 9 anal fin rays, 12 pectoral fin rays, and 13 caudal fin rays (Suppl. Fig. S1). Several of these fish were visibly infected with endoparasites (Fig. 1). These fish were transported to the Aquatic Parasitology Lab at Mississippi State University's College of Veterinary Medicine for parasitological examination as part of a routine diagnostic investigation. Fish were euthanized with an overdose of MS-222 (Western Chemical Inc., Ferndale, Washington). Eight fish were individually dissected in 0.9% physiological saline in glass dishes under a dissecting microscope to collect parasites. Afterward, the dissected fish remnants were individually digested by incubation at 40 C for 3 hr in 500-ml Erlenmeyer flasks with 200 ml of 0.5% pepsin solution acidified with citric acid (Kim et al., 2013). The contents of each flask were then screened in brass sieves with a 25-µm aperture. Screened contents were examined in glass dishes under a dissecting microscope to collect any parasites not collected during the initial necropsy. All collected parasites, from both necropsies and pepsin digestion, were identified as nymphal pentastomids and fixed directly in 70% ethanol for morphological and molecular characterization. The abdominal cavities of 20 additional G. affinis were injected with 10% neutral-buffered formalin (NBF), and the whole fish were fixed directly in 10% NBF for histopathology.
Figure 1.
Ventral surface of a Gambusia affinis (Baird and Girard, 1853) heavily infected with nymphal Levisunguis subaequalis Curran, Overstreet, Collins & Benz, 2014.
Abundance, intensity, and prevalence
Abundance, intensity, and prevalence, as defined by Bush et al. (1997), were calculated for L. subaequalis collected from G. affinis using Quantitative Parasitology 3.0 (Rózsa et al., 2000) with 95% confidence limits calculated using the Clopper-Pearson Method (Clopper and Pearson, 1934) with 2,000 bootstrap replications.
Morphological characterization
Three hologenophores of nymphal L. subaequalis were made using the methods described by Woodyard et al. (2019a). Briefly, the cephalothorax was cut from each individual nymph using a sterile scalpel blade. The removed cephalothorax was then dehydrated in a series of 4 ethanol washes (70–100%), cleared in methyl salicylate, and mounted on a glass slide in Canada balsam. Each corresponding hindbody was then placed in 1.5-ml tubes with 180 µl of Buffer ATL (Qiagen, Hilden, Germany) and 20 µl of proteinase K and incubated overnight at 55 C. The hindbody was recovered, dehydrated in a series of 4 ethanol washes from 70–100%, cleared in methyl salicylate, and mounted on the same slide as the corresponding cephalothorax. Ten nymphal paragenophores were also made using the above methods without DNA extraction. Three additional nymphs were directly mounted in Hoyer's medium (Hempstead Halide, Galveston, Texas) to better visualize chloride pores, which were not as easily seen in nymphs mounted in Canada balsam. Additionally, the oral cadre was excised from each of the 2 nymphs with scalpel blades and mounted directly in Hoyer's medium to compare with the excised oral cadre according to Curran et al. (2014). Hook and oral cadre measurement schemes were adapted from Fain (1961) (Fig. 2).
Figure 2.
Line drawings of taxonomically informative structures of nymphs of Levisunguis subaequalis Curran, Overstreet, Collins & Benz, 2014, mounted in Hoyer's medium. In situ oral cadre and measurement scheme for length and width. Oral cadre measurements are as follows: TOCL, total excised oral cadre length, OCL, oral cadre length and OCW, oral cadre width. Hook and fulcrum measurement schemes as follows: AC, blade length; AD, hook length; BC, base length; AB, hook gape; CD, plateau length; FL, fulcrum length; DAP, nymphal accessory piece length, ventral view.
Histopathological evaluation
Twenty formalin-fixed fish were decalcified in Kristensen's solution for 3 hr, serially trimmed in transverse sections or longitudinally bisected dorsoventrally along the midline, placed into individual tissue cassettes, and processed routinely. Paraffin-embedded tissues were sectioned at 4 µm, adhered to positively charged slides, and stained with hematoxylin and eosin. An additional section of 1 fish was stained with Verhoeff van Gieson stain, an alternative to Movat's pentachrome for highlighting the sclerotized openings of pentastomids, according to previously published methods (Sheehan and Hrapchak, 1980).
Molecular characterization and phylogenetic analysis
DNA was extracted from 3 hologenohpores and 1 whole nymph (DNeasy Blood and Tissue Kit, Qiagen, Hilden, Germany). The 18S and 28S ribosomal RNA gene regions were amplified and sequenced using previously published primers (Table I) and thermal cycling parameters (Table II). New primers were designed to facilitate the amplification of a partial COI gene sequence from pentastomids in the present study. Additionally, partial COI from G. affinis was amplified using the LCO1490/ HCO2198 primer set. The fin from an infected G. affinis (Fig. 1) was removed, and DNA was extracted as above. A section of the trunk of an additional G. affinis (Suppl. Fig. S1) was also subjected to DNA extraction. PCR was carried out using extracted DNA from G. affinis to molecularly confirm host species identity and aid in the design of pentastomid-specific COI primers. PCR reactions consisted of 10 µl of Phusion Hot Start II High-Fidelity PCR Master Mix (Thermofisher Scientific, Waltham, Massachusetts), 1 µl of each primer (10 nmol), 7 µl of nuclease-free water, and 1 µl of DNA template (∼10 ng). PCR products were visualized by electrophoretic migration in 0.8% agarose gels buffered with sodium borate and stained with ethidium bromide. Amplicons were excised, purified (Qiagen Gel Extraction Kit, Qiagen), and bidirectionally sequenced using the same primers used for amplification (Eurofins Genomics, Louisville, Kentucky). The resultant electropherograms were assembled into contiguous sequences in Geneious 11.1.5 (Kearse et al., 2012) and checked for ambiguities. Pentastomid sequences for the 18S, 28S, and COI regions generated in the present study were aligned with all available sequences of pentastomids downloaded from GenBank. Sequences that contained ambiguous bases and those that would not allow the alignments to be trimmed owing to insufficient overlap with other sequence data were excluded from the analyses. Sequences used in the present analyses are indicated in Suppl. Table S1. Separate alignments were generated for each of these loci using GUIDANCE2's implementation of the MAFFT algorithm (Katoh and Standley, 2013) on the GUIDANCE2 Web Server (Landan and Graur, 2008; Sela et al., 2015). Isolates for which data from all 3 loci were available were also used for concatenated phylogenetic analyses. Once constructed, alignments were trimmed manually, and all sites containing gaps were removed in Geneious. Single-locus alignment lengths were as follows: 18S (238 bp), 28S (941 bp), COI (157 bp). Alignment lengths for each region in the concatenated analyses were as follows: 18S (1,191 bp), 28S (1,014 bp), and COI (650 bp), for a total alignment of 2,855 base pairs. The disparity in alignment lengths for single loci between the concatenated and single-locus phylogenies is due to the inconsistency of sequence quality and length of sequences available in GenBank for pentastomids. Optimum DNA substitution models were determined for the 18S and 28S alignments as well as each codon position in the COI alignments using the Bayesian Information Criterion in MEGA7 (Kumar et al., 2016). Additionally, substitution saturation indices were calculated for each codon position in both COI alignments using DAMBE v. 7.0.13 (Xia, 2018) to ensure that the phylogenetic utility of these data was not adversely affected by substitution saturation. Substitution models for each locus in the single-locus alignments were as follows: 18S (K2+G), 28S (T92+G), COI codon position 1 (TN93), COI codon position 2 (T92+G), and COI codon position 3 (JC). Substitution models for each locus for the concatenated alignment were as follows: 18S (K2+G), 28S (K2+G), COI codon position 1 (HKY+G), COI codon position 2 (HKY), and COI codon position 3 (T92+I). These alignments were then concatenated in MEGA7, and a partitioned, concatenated Bayesian phylogenetic analysis was performed in MrBayes 3.2.6 (Ronquest and Huelsenbeck, 2003; Altekar et al., 2004) using Markov chain Monte Carlo searches of 2 simultaneous runs of 4 chains with sampling of 100th tree for 1 × 106 generations, ensuring the value of the standard deviation of the split frequencies reached <0.1. In addition to Bayesian phylogenetic inference, maximum likelihood analysis was conducted on the same alignments with the aforementioned partitioning scheme and models using IQ-Tree (Nguyen et al., 2015) on the IQ-Tree Web Server (Trifinopoulos et al., 2016). Single-locus analyses were carried about using the above methods with the exception of not being concatenated in MEGA7 prior to analyses by MrBayes and IQ-Tree. In addition to the aforementioned molecular methods of tree construction, a cladogram of the currently accepted morphology-based classification of pentastomids, with the addition of the genus Levisunguis, was constructed using a text editor. Trees constructed in MrBayes and IQ-Tree, as well as the cladogram constructed via a text editor, were formatted in FigTree 1.4.3 (Rambaut, 2016) and Adobe Illustrator 23.0.2.
Table I
Primers used for DNA amplification.
Table II
Thermal cycling parameters used for DNA amplification.
To assess the similarity of molecular data from L. subaequalis and other molecularly characterized pentastomid species, 18S, 28S, and COI sequences for L. subaequalis in the present study were searched in the GenBank, and up to the 10 closest nominal species matches, where data availability allowed, were downloaded and aligned as above. Geneious was then used to assess distances between these sequences.
To assess the relative phylogenetic utility of each of the aforementioned gene targets, phylogenetic informativeness profiles were constructed using PhyDesign (López-Giráldez and Townsend, 2011). The phylogenetic tree constructed using the concatenated alignment was formatted with FigTree, arbitrarily ultrametricized with Mesquite (Maddison and Maddison, 2018), and uploaded to the PhyDesign web server along with the concatenated, partitioned alignment file used in the concatenated phylogenetic analyses. The resultant phylogenetic informativeness profile images were formatted with Adobe Illustrator 23.0.2.
RESULTS
Intensity, abundance, and prevalence
Levisunguis subaequalis was detected in 100% (8/8) of dissected G. affinis. Intensity ranged from 2 to 171 nymphs with a mean intensity of 36.8 (95% bootstrap confidence limit 12.8–112).
Morphological characterization
Measurements from 18 nymphs: Reported measurements are rounded to the nearest micron unless otherwise stated and in the following format: range (average) Body 6.00–7.88 (6.77) mm long with 65–76 (71) annuli (Fig. 3). Oral cadre open anteriorly, 125–223 × 92–111, in situ (Fig. 4A). Excised oral cadre total length 300–320 (322) (n = 2) (Fig. 4B). Each hook bearing a sclerotized accessory piece (Fig. 4C). Fine, cuticular spines continuous along posterior of each annulus. Chloride pores 6–8 (7) in diameter, arranged in 2–3 irregular, horizontal rows along the middle each annulus (Fig. 4D). Hook and fulcrum measurements, compared with those of nymphs of Sebekia mississippiensis Overstreet, Self & Vliet, 1985, are indicated in Table III. Measurements from individual nymphs provided in Suppl. Table S2. Vouchers are deposited in the Harold W. Manter Laboratory of Parasitology, University of Nebraska, Lincoln, Nebraska under the following accession numbers: hologenophores (HWML210023–HWML210025) and paragenophores (HWML210027–HWML210036).
Figure 3.
Nymph of Levisunguis subaequalis Curran, Overstreet, Collins & Benz, 2014, from Gambusia affinis (Baird and Girard, 1853) mounted in Hoyer's medium. Color version available online.
Figure 4.
Photomicrographs of qualitative characters of ventrally mounted nymphs of Levisunguis subaequalis Curran, Overstreet, Collins & Benz, 2014 mounted in Hoyer's medium. (A) In situ oral cadre. (B) Excised oral cadre. (C) Hook and dorsal accessory piece. (D) Annuli showing fine cuticular spines continuous along posterior of each annulus on posterior margin and chloride pores arranged in irregular rows of 2–3 on each annulus.
Table III
Comparative morphological data for nymphs of Levisunguis subaequalis from the present study and Sebekia mississippiensis (Woodyard et al., 2019a).*
Histopathological description of Levisunguis subaequalis in Gambusia affinis
Nymphs of L. subaequalis were detected in 55% (11/20) of fish examined by light microscopy. All infected fish were gravid females. The nymphs occupied approximately 5–50% of the coelomic cavity by visual estimation, either free or encapsulated in pseudocysts formed by a few layers of epithelioid macrophages and fibroblasts (Fig. 5). The pentastomids typically compressed, dissected between, or displaced adjacent organs, particularly liver lobules, the swim bladder, and intestinal loops in all infected fish (Fig. 5). Aside from hepatocellular atrophy, coelomic distension, and moderate to severe thinning and atrophy of the ventral body wall musculature, no significant pathology was seen in the host viscera. The cephalothorax region of the pentastomids contained 2 pairs of retracted, curved, refractile hooks surrounded by skeletal myofibers (Fig. 6A). Bilaterally symmetric, tortuous head glands and hemolymph flanked the cephalic ganglion and esophagus (Fig. 6B). The glands produced a homogeneous, eosinophilic material (presumptive chitinous cuticle). The esophageal lumen of each specimen contained deep brown pigment consistent with hematin from digested host erythrocytes (Fig. 6B). Large acidophilic glands persisted down the length of the nymph along with thin peripheral bands of skeletal muscle, abundant vacuolated space, multifocal basophilic cells, and an intestine (Fig. 6C, D). The body of each nymph was pseudosegmented and lined by a chitinous cuticle with randomly distributed sclerotized openings containing annular rings, which stained black with Verhoeff van Gieson stain (Fig. 6C–F). The cuticle was occasionally shed, which incited infiltrates of epithelioid macrophages and granuloma formation in the mesentery. In fewer numbers within the mesentery of some fish, smaller granulomas contained acanthocephalans and larval nematodes, whereas the intestinal lumen occasionally contained adult nematodes. A slide used for histological assessment in the present study is deposited in Harold W. Manter Laboratory of Parasitology, University of Nebraska, Lincoln, Nebraska (HMWL-210026).
Figure 5.
Subgross magnification of a longitudinal, midline section through the body of a Gambusia affinis (Baird and Girard, 1853). Nymphs of Levisunguis subaequalis Curran, Overstreet, Collins & Benz, 2014, fill approximately 50% of the coelom and compress the liver (arrows), intestine (arrowheads), and swim bladder (asterisk). H&E stain. Color version available online.
Figure 6.
Structural features of nymphs of Levisunguis subaequalis Curran, Overstreet, Collins & Benz, 2014, infecting the coelom of Gambusia affinis (Baird and Girard, 1853) in longitudinal, transverse, and oblique sections. (A) Two pairs of refractile hooks (arrows) are retracted within the head of the nymph. H&E stain. (B) Within the head, a ganglion (arrowhead) and esophagus containing digested blood (white asterisk) are enclosed on either side by symmetrical head glands (arrows). H&E stain. (C) The body is pseudosegmented (arrows) with a thin layer of subcuticular skeletal muscle (arrowheads), central intestine (asterisk), and large acidophilic glands. H&E stain. (D) Multifocal clusters of basophilic cells are scattered throughout the body (arrows). Note the recently shed cuticle (arrowheads). H&E stain. (E) The cuticle contains randomly disseminated sclerotized openings with hyaline annular rings (arrow). H&E stain. (F) The annular rings stain black with an elastin stain (Verhoeff van Gieson stain). Color version available online.
Molecular characterization and phylogenetic analysis
While Bayesian inference and maximum likelihood phylogenetic analyses of ribosomal 18S and 28S regions and the mitochondrial COI gene all suggested a close evolutionary relationship between Levisunguis, as a distinct clade, and species of Sebekia, these trees exhibited low posterior probabilities and bootstrap support values as well as polytomies, polyphyletic clades, and general disagreement of topology with the current classification of pentastomids (Suppl. Figs. 2–4). Concatenated analyses using all 3 regions also resulted in trees with low support values and placed sequences of Levisunguis as a sister clade to the sebekid pentastomids, which use crocodilian definitive hosts. Together, these 2 groups form a clade for the subfamily Sebekinae. The placement of genera in these concatenated phylogenetic analyses is in accord with the currently accepted, morphology-based organization of the Eupentastomida (Fig. 7), while the single-locus analyses did not agree with each other or the morphology-based classification. Levisunguis subaequalis sequences from the present study are deposited in GenBank under the following accession numbers: 18S (MN065566–MN065569), 28S (MN065506–MN065509), and COI (MN062093–MN062096). The G. affinis COI sequences from the present study are deposited under the following GenBank accession numbers: MN265387–MN265388.
Figure 7.
Cladogram depicting the currently accepted taxonomy of class Eupentastomida adapted from Christoffersen and de Assis (2013) to include Levisunguis based on keys provided therein and the description of L. subaequalis Curran, Overstreet, Collins & Benz, 2014, by Curran et al. (2014). The cladogram is juxtaposed with concatenated phylogenetic analyses from 18S, 28S, and cytochrome c oxidase subunit 1 (COI) from L. subaequalis in the present study and available pentastomid sequences from GenBank. GenBank accession numbers are given for each sequence in the format 18S accession number/28S accession number/COI accession number. Numbers above branches indicate Bayesian posterior probabilities/maximum likelihood bootstrap values. Scale bar represents the number of nucleotide substitutions per site. Color version available online.
Percentage similarity between 18S, 28S, and COI sequences for L. subaequalis from the present study and up to the 10 closest nominal species matches from GenBank are presented in Table IV. Sequences for L. subaequalis were 100% similar to each other for each gene. 18S showed the least variability between species, followed by 28S and COI. Given that no fewer than 4 species were ≥99% similar at 18S, this gene is likely not suited for identification of pentastomids to species.
Phylogenetic signal analysis carried out on the concatenated 18S, 28S, and COI alignment suggests that the concatenated dataset yields more net phylogenetic signal than 18S, 28S, or COI alone. Regarding the individual loci, 28S offers greater signal at the basal end of the tree while COI offers slightly greater support at more distal clades. 18S offered comparatively little phylogenetic signal throughout (Fig. 8).
Figure 8.
Net and per site phylogenetic informativeness profiles for 18S, 28S, COI, each COI codon position, and concatenated sequence data from 18S, 28S, and COI generated by PhyDesign. GenBank accession numbers are given for each sequence in the format 18S accession number/28S accession number/COI accession number. Values above branches indicate posterior probabilities/bootstrap support values. Scale bar represents the number of nucleotide substitutions per site in the phylogenetic tree, while branch lengths are arbitrary in the morphology-based cladogram. Color version available online.
DISCUSSION
Nymphs in the present study were found to be morphologically consistent with the description of L. subaequalis by Curran et al. (2014), though these authors' description of taxonomically informative hook measurements as “similar in morphology to adult hooks but smaller” complicates efforts to make comparisons. Given that no nymphal vouchers have been deposited for this species, direct comparison with nymphs from the present study is not possible. Nonetheless, the nymphs from the present study agree in annulus counts, chloride pore diameter and distribution, fulcrum lengths, and oral cadre dimensions. Furthermore, the ranges of taxonomically informative hook measurements for L. subaequalis and Sebekia mississippiensis, the only other nominal sebekid reported from G. affinis in the southeastern United States (Curran et al. 2014), do not overlap to any degree (Table III). Given that nymphs from the present study are morphologically consistent with L. subaequalis and not with S. mississippiensis, the pentastomids in the present study are identified as L. subaequalis. Although Curran et al. (2014) placed L. subaequalis in the family Sebekidae, the species was not placed within a subfamily. However, based on the distribution of chloride pores and the absence of an esophageal peg of the oral cadre, L. subaequalis is a member of the subfamily Sebekinae within the family Sebekidae. This is in accordance with the keys to genera provided by Christoffersen and de Assis (2013).
Also of note, the measurements for nymphs of L. subaequalis from the present study were more consistent with 1 of the nymphal specimens in the type series of Sebekia mississippiensis Overstreet, Self & Vliet 1985 (AMNH 538) than all other described nymphs of S. mississippiensis. Hook and fulcra measurements for this nymph were previously reported as follows: AB (42), AC (164), AD (171), BC (124), DAP (122), and fulcrum length (410) (see suppl. data table S1 in Woodyard et al., 2019a). These data suggest that this specimen may be L. subaequalis rather than S. mississippiensis. Boyko (1996) indicates that this specimen, deposited in association with the original description of S. mississippiensis by Overstreet et al. (1985), is from G. affinis collected 4 miles east of Grand Chenier, Louisiana. This specimen likely represents a previously unacknowledged host record for L. subaequalis.
Table IV
Percentage similarity between 18S, 28S, and COI sequences for Levisunguis subaequalis from the present study compared with closest nominal species matches in GenBank. Percentages are based on 402, 378, and 215 base pair alignments for 18S, 28S, and COI, respectively. Few 28S sequences were available for comparison.
Given that the highest previously reported intensity for L. subaequalis in G. affinis is 6 (Curran et al., 2014), intensities in the present study of up to 171 nymphs per G. affinis were expected to induce significant pathologic changes in the host. However, limited adverse changes were observed in the present study, suggesting an old host-parasite relationship. Anecdotally, host motility appeared to have been negatively impacted given that fish were easily netted at the surface. The compression of the swim bladder by nymphs of L. subaequalis, as observed histologically in the present study, may have played some role here. However, such claims remain speculative in the absence of behavior studies comparing infected and uninfected G. affinis. Although the type intermediate host does not appear to be greatly affected by infection with L. subaequalis, potential infectivity and pathology of this pentastomid to other fish species remains unknown.
Different gene targets have varying efficacy for resolving taxonomic uncertainty at multiple taxonomic rankings (Hwang and Kim, 1999). Given the tumultuous history of pentastomid systematics, the recent, morphology-based reorganizations of the evolutionary relationships of pentastomids (Almeida and Christoffersen, 1999; Junker, 2002; Christoffersen and de Assiss, 2013) warrant confirmation using molecular data. However, the current dearth of sequence data suitable for such analyses confounds efforts at phylogenetic inference. Both single-locus Bayesian inference and maximum likelihood analyses using the 18S (Suppl. Fig. S2), 28S (Suppl. Fig. S3), and COI (Suppl. Fig. S4) regions as well as concatenated Bayesian and maximum likelihood inference using all 3 regions are compelling enough to either support or refute the current, morphology-based classification of the Eupentastomida (Fig. 7). Single-locus analyses using all available sequence data for pentastomids from 18S, 28S, and COI exhibit polytomies, polyphyletic clades, and low support values. Furthermore, these trees agree with neither each other nor the currently established, morphology-based placement of taxa within the class Eupentastomida. Only the concatenated analyses using 18S, 28S, and COI data agree with currently accepted pentastomid taxonomy based on morphological characters. However, the bootstrap and posterior probability values are still low to offer robust support for the currently accepted taxonomy of this class. The wealth of phylogenetic trees recovered in the present study afforded some opportunity to offer a limited assessment of the phylogenetic signal from the 18S, 28S, and COI regions as well as concatenated data from all 3 regions. Phylogenetic informativeness profiles constructed from each region using PhyDesign (Fig. 8) indicate the concatenated dataset offers the greatest net phylogenetic signal, with 28S and COI providing the greatest single-locus contributions to net phylogenetic informativeness. 28S offers greater informativeness at more basal clades, while COI offers slightly higher support at more distal branches. It is unfortunate that 28S, the region that seems to offer the greatest phylogenetic signal of any single loci compared in the present study for distinguishing genera of pentastomids, is also the region with the least available data (Suppl. Table S1). Still, it must be noted that these analyses are based on the limited data presently available for these regions. More sequence data from 18S, 28S, and COI are needed to more thoroughly assess the utility of these markers for phylogenetic inference of the evolutionary relationships between pentastomids. To date, only Barton and Morgan (2016) and Woodyard et al. (2019a) provide 18S, 28S, and COI sequence data from the same specimens for 4 of the 141 currently described pentastomid species for the analyses carried out in the present study. While phylogenetic analyses using concatenated ribosomal and mitochondrial sequence data are increasingly being used for phylogenetic inference in other parasitic groups as evidenced by recent works on species ranging from Clinostomidae (Caffara et al., 2016; Woodyard et al., 2017; Briosio-Aguilar et al., 2019) to Emeriidae (Ogedengbe et al., 2018; Woodyard et al., 2019b), the present molecular data from pentastomids are not suitable for such analyses. While there are over 141 described pentastomid species belonging to 26 genera (Christoffersen and de Assis, 2013), molecular data, including the data from the present study, are publicly available for only 14 species from 11 genera. To this end, the authors urge any researchers undertaking molecular characterization of pentastomids to make publicly available sequence data from the 18S, 28S, and COI genes to accommodate future studies. Although phylogenetic signal assessment in the current study suggests these regions have the most phylogenetic utility when combined, the respective phylogenetic utilities of these regions, either alone or in combination, are not yet well established for pentastomids. While the present molecular analyses cannot be said to definitively support the placement of the genus Levisunguis within the family Sebekidae, as established based on the morphological assessment by Curran et al. (2014) and the present study, there is little reason to doubt the current placement of the genus. Nonetheless, it is clear that data from other pentastomid taxa are needed to further infer the phylogeny of these enigmatic parasites.
ACKNOWLEDGMENTS
This work was supported by the Colleges of Veterinary Medicine at Mississippi State University and the University of Georgia. This work was also supported by the 2018 Student Subunit of the American Fisheries Society at Mississippi State University by an Undergraduate Research Scholarship awarded to E.T.W. and the 2018 Mississippi Chapter of the American Fisheries Society Student Research Scholarship, also awarded to E.T.W.
