Morphological and molecular description of Rhadinorhynchus laterospinosus Amin, Heckmann & Ha, 2011 (Acanthocephala, Rhadinorhynchidae) from marine fish off the Pacific coast of Vietnam

Rhadinorhynchus laterospinosus Amin, Heckmann & Ha, 2011 (Rhadinorhynchidae) was described from a single female collected from a trigger fish, Balistes sp. (Balistidae) from the northern Pacific coast of Vietnam in Halong Bay, Gulf of Tonkin. More recent collections of fishes in 2016 and 2017 revealed wider host and geographical distributions. We report this Acanthocephala from nine species of fish representing six families (including the original record from Balistes sp.) along the whole Pacific coast of Vietnam. The fish species are Alectis ciliaris (Carangidae), Auxis rochei (Scombridae), Auxis thazard (Scombridae), Leiognathus equulus (Leiognathidae), Lutjanus bitaeniatus (Lutjanidae), Megalaspis cordyla (Carangidae), Nuchequula flavaxilla (Leiognathidae), and Tylosurus sp. (Belonidae). We provide a complete description of males and females of R. laterospinosus, discuss its hook metal microanalysis using EDAX, and its micropores. Specimens of this species characteristically have lateral trunk spines bridging the anterior ring of spines with posterior field of ventral spines and a proboscis with 15–19 longitudinal alternating rows of 21–26 hooks each varying with host species. We demonstrate the effect of host species on the distribution and size of the trunk, proboscis, proboscis hooks, trunk spines, and reproductive structures. The molecular profile of this acanthocephalan, based on 18S rDNA and cox1 genes, groups with other Rhadinorhynchus species and further seems to confirm the paraphyly of the genus, which is discussed.


Introduction
Most of the recent taxonomic work on the Acanthocephala from Vietnam has been reported by the Amin-Heckmann-Ha team since 2000. A number of acanthocephalan species from freshwater and marine fish, amphibians, reptiles, birds, and mammals were previously described in Vietnam [3,[9][10][11][12][13]16]. Additionally, 11 species of acanthocephalans were collected from marine fish off the eastern seaboard of Vietnam in Halong Bay in 2008 and 2009. Of these, six new species of Neoechinorhynchus Stiles & Hassall 1905, one new species of Heterosentis Van Cleave, 1931, and two new species of Rhadinorhynchus Lühe 1911 were described [8,14,15]. Four other species of Echinorhynchid acanthocephalans from marine fishes in Halong Bay were described [4] and five other new species from fishes and amphibians of eight collected host species were also described. Three other species of Rhadinorhynchus and one species of Gorgorhynchus were otherwise previously reported from marine fishes in Vietnam by other observers [19].
Fifteen species of acanthocephalans in five families were more recently collected from fishes on the Pacific coast and amphibians in central Vietnam in 2016 and 2017. In the present report, we describe males and females of R. laterospinosus, which was originally described from a single female specimen, from extensive collections of fishes along the Pacific coast of Vietnam and provide a molecular profile of that species based on small subunit ribosomal DNA (18S rDNA) and partial mitochondrial cytochrome c oxidase 1 (cox1) genes. Furthermore, its phylogenetic relationships with other Rhadinorhynchus and closest-related species are analyzed and discussed.

Collections
Collections of 215 specimens of R. laterospinosus from nine species of fish in six families in 2016 and 2017 along the Pacific coast of Vietnam are detailed in Table 1 along with infection parameters, geographical locations and museum numbers of deposited material at the Harold W. Manter Laboratory, Nebraska State Museum, Lincoln, Nebraska.

Methods
Freshly collected acanthocephalans were extended in water until proboscides were everted and fixed in 70% ethanol for transport to our Institute of Parasitic Diseases (IPD) in Arizona, USA for processing and further studies. Worms were punctured with a fine needle and subsequently stained in Mayer's acid carmine, destained in 4% hydrochloric acid in 70% ethanol, dehydrated in ascending concentrations of ethanol reaching 100% (24 h each), and cleared in 100% xylene then in 50% Canada balsam and 50% xylene (24 h each). Whole worms were then mounted in Canada balsam. Measurements are in micrometers, unless otherwise noted; the range is followed by the mean values between parentheses. Width measurements represent maximum width. Trunk length does not include proboscis, neck, or bursa.
Line drawings were created by using a Ken-A-Vision microprojector (Ward's Biological Supply Co., Rochester, New York), which uses cool quartz iodine 150 W illumination with 10Â, 20Â, and 43Â objective lenses. Images of stained whole mounted specimens were projected vertically on 300 series Bristol draft paper (Starthmore, Westfield, Massachusetts), then traced and inked with India ink. Projected images were identical to the actual specimens being projected.
Specimens were deposited in the University of Nebraska's State Museum's Harold W. Manter Laboratory (HWML) collection in Lincoln, Nebraska, USA. Accession numbers are noted in Table 1.

Scanning electron microscopy (SEM)
About 15 specimens from four host species that had been fixed and stored in 70% ethanol were processed for SEM following standard methods [36]. These included critical point drying (CPD) in sample baskets and mounting on SEM sample mounts (stubs) using conductive double sided carbon tape. Samples were coated with gold and palladium for 3 min using a Polaron #3500 sputter coater (Quorum (Q150 TES) www. quorumtech.com) establishing an approximate thickness of 20 nm. Samples were placed and observed in an FEI Helios Dual Beam Nanolab 600 (FEI, Hillsboro, Oregon) Scanning Electron Microscope, with digital images obtained in the Nanolab software system (FEI, Hillsboro, Oregon) and then stored on a USB for future reference. Samples were received under low vacuum conditions using 10 kV, spot size 2, 0.7 Torr using a GSE detector.

EDXA (energy dispersive X-ray analysis)
Standard methods were used for preparation, similar to the SEM procedure. Eight specimens were examined and positioned with the above SEM instrument which was equipped with a Phoenix energy-dispersive X-ray analyzer (FEI, Hillsboro, Oregon). X-ray spot analysis and live scan analysis were performed at 16 kV with a spot size of five and results were recorded on charts and stored with digital imaging software attached to a computer. The TEAM *(Texture and Elemental Analytical Microscopy) software system (FEI, Hillsboro, Oregon) was used. Data were stored on a USB. The data included weight percent and atom percent of the detected elements, following correction factors, and were stored on a USB. All figures on the USB can be viewed by contacting the second author. The hooks were cut and scanned at two positions (tip and middle) with a gallium beam (LIMS) using a dual beam scanning electron microscope. The alignment of the hook previous to cutting generated a cross section of the area.

Ion sectioning of hooks
A dual-beam SEM with a gallium (Ga) ion source (GIS) was used for the LIMS (Liquid Ion Metal Source) part of the process. The gallium beam (LIMS) is a gas injection magnetron sputtering technique whereby the rate of cutting can be regulated. The hooks of six acanthocephalans were centered on the SEM stage and cross-sectioned using a probe current between 0.2 nA and 2.1 nA according to the rate at which the area is cut. The time of cutting is based on the nature and sensitivity of the tissue. Following the initial cut, the sample also goes through a milling process to obtain a smooth surface. The cut was then analyzed with X-ray at the tip, middle, and base of hooks for chemical ions with an electron beam (Tungsten) to obtain an X-ray spectrum. Results were stored with the attached imaging software then transferred to a USB for future use. The intensity of the GIS was variable according to the nature of the material being cut.

Molecular methods
Total genomic DNA was extracted from four specimens of R. laterospinosus from Auxis rochei preserved in 70% ethanol using a Qiagen™ (Valencia, California, USA) DNeasy Ò Tissue Kit, and following the manufacturer's instructions. Partial nuclear small subunit ribosomal DNA (18S rDNA) and partial fragments of mitochondrial cytochrome c oxidase 1 (cox1) gene were amplified (50 lL total volume) using ExcelTaq TM SMOBIO Ò PCR Master Mix (Taiwan) containing: 5Â concentrated master mix, that is, a mixture of recombinant Taq DNA polymerase, reaction buffer, MgCl 2 (2 mM), dNTPs (0.2 mM), and enzyme stabilizer; 0.25 lM of each PCR primer and 2 lL of extracted gDNA. Primer pairs and amplification conditions used were as follows.
Partial fragments of the cox1 gene were amplified using the primers LCO1490 (forward, 5 0 -GGTCAACAAATCATAAA-GATATTGG-3 0 ) and HCO2198 (reverse, 5 0 -TAAACTT-CAGGGTGACCAAAAAATCA-3 0 ) [23] under the following thermocycling conditions: initial denaturation at 95°C for 15 min followed by 40 cycles (denaturation for 5 min at 80°C, followed by 1 min 30 s at 92°C, annealing for 1 min at 42°C, and extension for 2 min at 72°C), and a final extension step at 72°C for 10 min. In every PCR run, a negative and a positive control were used to detect any potential contamination and to have a reliable sample to compare with, respectively. PCR amplicons were sequenced directly for both strands using the same PCR primers.
Sequences were assembled and edited using Mega v6 [47] and submitted to GenBank under accession numbers: MK457183 -MK457185 (18S) and MK572741-MK572744 (cox1). Sequences were aligned using Muscle as implemented in MEGA v6 together with published sequences of Rhadinorhynchus and most closely-related published sequences to members of this genus. Rotaria rotatoria (Pallas, 1776) was used as the outgroup in both the 18S (DQ089736) and cox1 (EU499879) datasets. Both alignments (18S: 760 nt positions of which eight were excluded prior to analysis; cox1: 537 nt positions of which 26 were excluded prior to analysis) were used for comparative sequence analysis.
The SeaView v4 interface [27] was used to select blocks of evolutionarily conserved sites. Maximum likelihood (ML) and Bayesian inference (BI) algorithms were used for phylogenetic tree reconstruction after determination of the best-fit model of nucleotide substitution with jModelTest v2.1.4 [22] using the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC), respectively. For the ML algorithm, the best-fitting model selected was the GTR + G model (nst = 6, rates = gamma, ngammacat = 4) both for the 18S and cox1 datasets. In the case of BI, the best-fitting model was TVMef + G (nst = mixed, rates = gamma, ngammacat = 4) for the 18S dataset and TrN + G (nst = 6, rates = gamma, ngammacat = 4) for the cox1 dataset. ML analyses were performed in PhyML v3.0 [30] with a non-parametric bootstrap of 100 replicates. BI analyses were carried out with MrBayes v3.2.6 [42] on the CIPRES Science Gateway v3.3 [39]. Log likelihoods were estimated over 10,000,000 generations using Markov Chain Monte Carlo (MCMC) searches on two simultaneous runs of four chains, sampling trees every 1000 generations. The first 25% of the sampled trees were discarded as "burnin" and a consensus topology and nodal support estimated as posterior probability values [35] were calculated from the remaining trees. Pairwise genetic distance matrices were calculated using the "uncorrected p-distance" model implemented in MEGA v6.

Results
Rhadinorhynchus laterospinosus was originally described from one female specimen collected from an individual triggerfish, Balistes sp. (Linn.) (Balistidae), from the Pacific coast at Halong Bay in May of 2009. It has since been found in eight other species of fish in five other families along the Pacific coast of Vietnam from the north at Hai Phong and Quang Binh to the south at Nha Trang and Binh Thuan (Table 1). We have studied specimens from all host species but provide measurements of specimens from the more extensive collections from two hosts, Auxis rochei (Lacépède) and Auxis thazard (Lacépède). The description is inclusive of morphometric differences noted between specimens collected from these two-host species ( Table 2).   (Table 2). Dorsal hooks slightly shorter and more slender that stouter and more sharply curved ventral hooks (Figs. 6 and 7). Hooks slightly arched ( Fig. 12) with thin grooved cortical layer and thick core (Figs. 13 and 14), smallest anteriorly, largest at middle, gradually smaller posteriorly except at basal circle of abruptly larger hooks (Figs. 10-12 and 29). Hook roots simple, markedly shorter than blades, directed posteriorly (Figs. 6 and 7). Neck prominent, slightly longer than wide posteriorly, with paired sensory pores (Figs. 15, 16 and 29). Proboscis receptacle double-walled, about twice as long as proboscis with cephalic ganglion near its middle. Lemnisci digitiform, equal, uniformly broad throughout, slightly shorter than receptacle (Fig. 1). Gonopore terminal in males but subterminal in females at level of posterior abrupt narrowing of trunk.

Remarks
The present report represents an expansion of our understanding of R. laterospinosus since its description from only one female in 2011 [8] from a trigger fish, Balistes sp. from the northern Pacific coast of Vietnam at Cat Ba Island, Halong Bay, Gulf of Tonkin. The single female had a proboscis with 18 longitudinal rows of 24 hooks each, and eight ventral and 18 lateral spines in the posterior field of trunk spines connecting anteriorly with the anterior field of trunk spines. The collection of over 200 specimens from eight additional hosts along the Pacific coast of Vietnam provided an opportunity to describe males, lemnisci, the female reproductive system, and eggs for the first time, and to clarify the dorso-ventral differentiation of proboscis hooks that were inaccurately declared as "similar in shape and size, and in their posteriorly directed angle of projection from proboscis" [8] with the availability of more specimens for study. The new description made it possible to examine the relationship between host species and the expression of certain morphometric parameters. Specimens from A. thazard had larger size of trunk, some proboscis hooks, proboscis receptacle, testes, anterior and posterior cement glands, and Saefftigen's pouch, but relatively fewer and smaller trunk spines than specimens from A. rochei (Table 2).

Energy dispersive X-ray analysis (EDXA)
We report the X-ray scans and metal composition of large and small proboscis hooks (Figs. 34, 35 and Tables 3, 4) and trunk spines (Table 5 and Fig. 36) of R. laterospinosus that were cut with a gallium beam (LMIS) and viewed with a dual beam scanning electron microscope with X-ray capabilities (EDXA). There are variable levels of calcium, phosphorus, and sulfur depending on the type of hook and whether readings are made at the base, core, tip or edge of hooks. Other common elements of living organisms (carbon and oxygen) and elements used for specimen preparation (gallium, palladium, gold) are not included in the analysis. In large hooks, the calcium and phosphorus levels were highest at the center of the hook base (Table 3 and Fig. 34). In small hooks, calcium and phosphorus were highest at hook tips (Table 4 and Fig. 35). Sulfur was high in both spine tip and base compared to calcium and phosphorus ( Table 5, Fig. 36).

Molecular results
Three partial 18S rDNA (741-767 nt) and four cox1 (606-622 nt) sequences were generated from four adult specimens (two males and two females) of R. laterospinosus. While 18S rDNA sequences were identical (only the longest one was thus included in the corresponding phylogenetic trees), intraspecific sequence divergence for cox1 ranged between 0.008 and 0.018% (5-11 nt difference). Table 6 provides data for the sequences retrieved from GenBank and used in the phylogenetic analyses based on the two alignments. While both ML and BI algorithms produced trees with identical topology for the 18S gene (Fig. 37), a slightly different topology was observed for the cox1 gene (Figs. 38 and 39).
According to phylogenetic analyses based on the cox1 gene, the four newly generated sequences for the R. laterospinosus grouped, with low support, with a clade formed by representatives of Bolbosoma and the species Neorhadinorhynchus nudus   difference), and remained apart from a third clade which included the only available published sequence on this gene for Rhadinorhynchus (0.238-0.242%, 122-124 nt difference from newly generated sequences).

Morphometric comparisons
The observed relationship between host species and size and even shape of acanthocephalans observed in this study ( Table 2) has been previously demonstrated for other acanthocephalans including Echinorhynchus salmonis Müller, 1784 whose variability in the size of taxonomically important structures such as the trunk, proboscis hooks, proboscis, testes, etc. has been attributed to host species. Such relationships have been reported in Lake Michigan where male and female specimens from bloater, Coregonus hoyi (Gill) (Salmonidae) achieved not only larger size but also different body form (broad anteriorly) compared to the slender specimens from rainbow smelt, Osmerus mordax (Mitchell) (Osmeridae) [17]. The larger and heavier worms from bloater invariably showed a higher regression coefficient (adjusted coefficient of determination) compared to those from smelt in all characters including size of trunk, proboscis, longest proboscis hooks, receptacle, testes, lemnisci, and eggs. The taxonomic implications of this variability were discussed (Amin and Redlin, 1980). Earlier, Amin [1] demonstrated a similar relationship for Acanthocephalus dirus (Van Cleave, 1931) Van Cleave and Townsend, 1936 in Wisconsin fishes. Females of the same developmental stage recovered during the same period were found to have attained larger sizes in certain hosts than in others with the largest females being found in Lepomis macrochirus Rafinesque. The size of the trunk in males was also found to follow the same pattern. Similarly, testes also attained a larger size in males recovered from Catostomus commersonii Lacépède (Catostomidae) than in males from Semotilus atromaculatus (Mitchill) (Cyprinidae). Amin [1] stated that these size variations "result from differential growth rates of these worms in the various host intestinal environments (and) are probably mediated by certain host specific factors."

Distribution
Amin [2] and Amin et al. [8] recognized 38 valid species of Rhadinorhynchus and invalidated 30 others. Only five more species of Rhadinorhynchus were described since, four from marine fishes off Australia [43] and Rhadinorhynchus  Table 3). Insert: SEM of a lateral and cross gallium cut hook. Figure 35. Energy Dispersive X-ray spectrum of the tip of a small hook of a Rhadinorhynchus laterospinosus specimen showing high levels of calcium and phosphorus but less calcium than large hooks (see Table 4). Insert: SEM of posterior hooks and hook tips in cross gallium cuts.

Morphological comparisons
Morphologically, Rhadinorhynchus stunkardi Gupta et Fatma, 1987 from India is the only other species of Rhadinorhynchus that has lateral trunk spines connecting the anterior and posterior fields of trunk spines like R. laterospinosus. Rhadinorhynchus stunkardi, however, has only 3-4 posterior trunk spines on the ventral side, only 8-10 proboscis hook rows each with 24-26 small hooks that reach a maximum length of only 46, considerably larger eggs, 120-150 Â 25-28, and a terminal gonopore [31].

Energy dispersive X-ray analysis (EDXA)
The results of the X-ray analysis (Tables 3-5 and Figs. 34-36) of gallium cut hooks and spines of R. laterospinosus show that the large hooks in the mid-proboscis and the small posterior hooks had a high level of sulfur at the tip edge, which is consistent with the base of the hooks. This element along with calcium and phosphorus aid in the mineralization and hardening of the outer layer of hooks, similar to the enamel layer of the mammalian tooth (calcium phosphate apatite) [32]. The base center of large hooks shows increased levels of calcium (45.30%) and phosphate ions (14.87) ( Table 3), comparable to the inner core of mammalian teeth [6]. Sulfur levelsshowed a higher differential concentration at the edge than the middle of cut hooks (Tables 3 and 4). This element is part of the prominent outer layer of most acanthocephalan hooks and is a major contributor to the hardening process of this attachment structure. There is a difference in the distribution of calcium ions in the smaller hooks in relation to large hooks, this level being highest in the core and base of large hooks (45/30%) but highest at the tip of small hooks (30.57%) (Tables 3 and 4). A similar EDAX study of the proboscis hooks of Echinorhynchus baeri Kostylew, 1928 showed that large hooks have higher calcium, phosphorus, and sulfur than miniature rootless hooks [6]. Comparable patterns for the numerous trunk spine gallium cuts (Table 5) demonstrate the rigid nature of the spine which is explained by the X-ray scans (Fig. 36). There is a reasonably high level of phosphorus, calcium and especially sulfur at the tip (18.23%) and base (11.64%) of the spine, which have mineralized to form the rigid support. The X-ray scans of the gallium cut hooks and spines help explain the morphological nature of R. laterospinosus and identify its unique "personality" [44]. The uniqueness of the metal analysis as expressed by X-ray scans appears to be species-specific and can be regarded as a fingerprint of key diagnostic value that is just as important as molecular analysis.   This was well demonstrated in the study of Rhadinorhynchus oligospinosus Amin and Heckmann, 2017 from mackerels in the Pacific Ocean off Peru [5], among others.

Micropores
The presence of micropores on various trunk regions of specimens of R. laterospinosus (Figs. 18, 22 and 23) suggests differential nutrient absorption related to the diameter and distribution of micropores as appears to be the case in practically all acanthocephalans. We have documented this phenomenon in 16 species of acanthocephalans [33] and a few more since.  (1916,1917), and Sclerocollum rubrimaris Schmidt and Paperna, 1978 were reviewed earlier [7]. The micropore canals appear to be continuous with canalicular crypts that constitute a huge increase in external surface area implicated in nutrient uptake [7].

Molecular analysis
To date, genetic data have been provided for only three species of Rhadinorhynchus: R. laterospinosus (present results), R. pristis, and R. lintoni (see Table 6 for references). The scarcity of molecular profiles described for this genus poses difficulties for correctly determining relationships among its members and with other genera, and adds importance to the new molecular data presented herein.
The lack of congruence between taxonomy and evolutionary history within Rhadinorhynchus observed in the 18S rDNA-and cox1-derived phylogenies has been noted previously by other authors based on morphological [40] and genetic markers (18S and 28S rDNA and cox1 genes) [21,29]. Indeed, while the sequences provided for R. pristis and R. lintoni [48] form a strongly supported clade with members of the genus Pomphorhynchus in the 18S rDNA-derived phylogram, the rest of the available Rhadinorhynchus sequences (including newly generated ones) form a clearly separate group that also includes sequences from T. annulospinosa and G. decapteri. This pattern was highlighted previously [21,29]. While Gregori et al. [29] questioned the genetic identification of the specimens characterized by Verweyen et al. [48], Braicovich et al. [21] attributed this pattern to incorrect assignment to Rhadinorhynchus by García-Varela et al. [24]. In fact, in their revision of the genus, Amin et al. [8] classified R. pristis and R. lintoni from Atlantic and Mediterranean waters as invalid species, which supports the view by Gregori et al. [29] given that specimens collected by Verweyen et al. [48] were from Pacific waters. This solves the paraphyly "problem" observed in these previous phylogenies and in the ones presented herein based on the 18S rDNA gene. Because previously described 18S rDNA Rhadinorhynchus sequences [20,29] and present results group with those provided by García-Varela et al. [24], the suggestion by Braicovich et al. [21] of a misidentification by the latter author could be ruled out. Another specimen belonging to the same clade has recently been classified into the genus Gymnorhadinorhynchus [45]. The null difference between this sequence and the newly generated one for R. laterospinosus points to a need for reclassification of this Gymnorhadinorhynchus sp. specimen most probably into the genus Rhadinorhynchus.
The outcome of the phylogenetic analysis based on the cox1 gene is less complete than the 18S rDNA-based one due to the near absence of cox1 gene sequences for Rhadinorhynchus in GenBank. Even so, it shows conflictive relationships for members of this genus, with present sequences forming a subclade within a group including Bolbosoma members and N. nudus, apart from the group formed by Rhadinorhynchus sp., T. annulospinosa, and G. decapteri. Although the goal of the present study is not to discuss the higher level classification of Paleacanthocephala, the inclusion of the echinorhynchid N. nudus within the Polymorphidae (i.e. Bolbosoma) further demonstrates the extent of these inconsistencies at the suprafamiliar level. In fact, the paraphyly within the palaeacanthocephalan at the family level is well established [25,34,40,48], which highlights the existing problems with their taxonomic arrangement and points to the need for a reclassification based on better morphological, ecological and molecular characterization of their members.
To summarize, following the 18S rDNA-based analysis, a single clade including all the valid species of the genus Rhadinorhynchus described up until now is recognized. However, Rhadinorhynchus relationships in phylogenetic analysis based on cox1 sequences are not so clear, mostly due to the lack of published sequences of this gene so far. Conflicting relationships with other genera (i.e. Gymnorhadinorhynchus, Transvena, Bolbosoma and Neorhadinorhynchus) are apparent in both phylogenies, underlining the importance of elucidating relationships within the Paleacanthocephala in future studies.