A new order of fishes as hosts of blood flukes (Aporocotylidae); description of a new genus and three new species infecting squirrelfishes (Holocentriformes, Holocentridae) on the Great Barrier Reef

A new genus and three new species of blood flukes (Aporocotylidae) are described from squirrelfishes (Holocentridae) from the Great Barrier Reef. Holocentricola rufus n. gen., n. sp. is described from Sargocentron rubrum (Forsskål), from off Heron Island, southern Great Barrier Reef, and Lizard Island, northern Great Barrier Reef, Australia. Holocentricola exilis n. sp. and Holocentricola coronatus n. sp. are described from off Lizard Island, H. exilis from Neoniphon sammara (Forsskål) and H. coronatus from Sargocentron diadema (Lacepède). Species of the new genus are distinct from those of all other aporocotylid genera in having a retort-shaped cirrus-sac with a distinct thickening at a marginal male genital pore. The new genus is further distinct in the combination of a lanceolate body, X-shaped caeca, posterior caeca that are longer than anterior caeca, a single, post-caecal testis that is not deeply lobed, a post-caecal, post-testis ovary that is not distinctly bi-lobed, and a post-ovarian uterus. The three new species can be morphologically delineated based on the size and row structure of the marginal spines, as well by total length, oesophagus and caecal lengths, and the position of the male genital pore, testes and ovary relative to the posterior extremity. The three species of Holocentricola are genetically distinct from each other based on cox1 mtDNA and ITS2 rDNA data, and in phylogenetic analyses of 28S rDNA form a well-supported clade sister to species of Neoparacardicola Yamaguti, 1970. This is the first report of aporocotylids from fishes of the family Holocentridae and the order Holocentriformes.


Introduction
Fishes of Heron and Lizard Islands, on the southern and northern Great Barrier Reef, respectively, have been the focus of extensive blood fluke research over the last two decades; thirty aporocotylid species have been reported from 14 teleost families from these locations. As part of a PhD study, Nolan and Cribb [34,37,38] [39] later described two Cardicola species from Lizard Island, one from each a lutjanid and a scombrid, and Nolan et al. [40] described a new species of Phthinomita from an apogonid. As part of another PhD study, Yong et al. [70][71][72] described three species of Cardicola, one from each of an apogonid, balistid and chanid, and two species of Psettarium Goto & Ozaki, 1930 from tetraodontiforms. Yong and Cribb [67] described a new genus and species from a tetraodontid, and Yong et al. [68] surveyed butterflyfishes (Chaetodontidae) from the Great Barrier Reef, reporting Elaphrobates chaetodontis (Yamaguti, 1970) Yong, Cribb & Cutmore, 2021 from 19 chaetodontid species. Recent blood fluke surveys at these locations have led to the re-collection of many of these known species [15], but examination of fish families not previously surveyed is revealing further aporocotylid richness in the region.
During recent helminthological examinations of fishes from off Heron and Lizard Islands, blood flukes were collected from three holocentrid species. These specimens represent a genus and three species, new to science, which are formally described and characterised phylogenetically below.

Materials and methods Ethics
Fishes were handled and euthanised following all applicable institutional, national and international guidelines for the care and use of animals. Fishes were collected under Great Barrier Reef Marine Park Authority Permits G16/38038. 1

Specimen collection
Holocentrid fishes were collected from off Heron Island, southern Great Barrier Reef, and Lizard Island, northern Great Barrier Reef (Queensland, Australia), via spearfishing and hand netting. Some gill filaments were removed and examined for the presence of eggs following Yong et al. [68]. Gill arches were removed and placed in saline solution (0.85% NaCl solution). The hemibranchs of each arch were separated, the branchial arteries removed and squeezed or ripped apart. The hemibranchs were then cut into small pieces and washed using the gut-wash approach of Cribb and Bray [9]. The heart was removed, placed in saline solution and each section opened separately. Some of the ventricle tissue was then squashed and examined for the presence of eggs following Yong et al. [68]. The liver was removed, placed in saline and the vessels in the liver mass cut open. The liver was then roughly ripped apart and washed using the gut-wash approach. The head was then cut in half down the midline and washed using the gut-wash approach. The remaining body was then split along the vertebral column and washed using the gut-wash approach. Aporocotylids were washed in vertebrate saline, fixed by pipetting into near-boiling saline, and preserved in 70% ethanol for parallel morphological and molecular characterisation. Some individual worms were processed for both morphological and molecular analysis (hologenophores, sensu Pleijel et al. [48]). Species were delineated using an integrative interpretation of morphological, ecological, and genetic data, following the criteria of trematode species recognition proposed by Bray et al. [6] (i.e. reciprocal monophyly in the most discriminating available molecular marker + distinction in morphology or host distribution). Prevalence figures combine any evidence of current infection, i.e. adult worms or fresh eggs lodged in gill tissue.

Morphological analysis
Specimens for morphological analysis were washed in fresh water, stained in Mayer's haematoxylin, destained in a solution of 1.0% HCl and neutralised in 1.0% ammonium hydroxide solution. Specimens were then dehydrated through a graded ethanol series, cleared in methyl salicylate and mounted in Canada balsam. Measurements were made using an Olympus SC50 digital camera mounted on an Olympus BX-53 compound microscope using cellSens Standard imaging software. Measurements are in micrometres (lm) and given as a range followed by the mean in parentheses. Where length is followed by breadth, the two measurements are separated by "Â". Drawings were made using an Olympus BX-53 compound microscope and drawing tube.

Molecular sequencing and phylogenetic analysis
Specimens for molecular analysis were processed according to the protocols used by Cribb et al. [12] and Wee et al. [63]. Sequence data were generated from adult worms (whole or 2 S.C. Cutmore and T.H. Cribb: Parasite 2021, 28, 76 hologenophore specimens) and from eggs lodged in gill filaments. Eggs were not removed from the gill, rather the egg mass and gill filament tip were digested together. Following Blasco-Costa et al. [3], three genetic markers were sequenced, the second internal transcribed spacer region (ITS2 rDNA), the large (28S) ribosomal subunit RNA coding region and the cox1 mitochondrial region (cox1 mtDNA). The complete ITS2 rDNA region (with flanking 5.8S and 28S regions) was amplified and sequenced using the primers 3S [30] or GA1 [2] and ITS2.2 [10], the partial D1-D3 28S rDNA region using LSU5 [24], 300F [26], ECD2 [25] and 1500R [54] and the partial cox1 region using Dig_cox1Fa [63] and Dig_cox1R [63]. Geneious Ò version 10.2.6 [21] was used to assemble and edit contiguous sequences. ITS2 and cox1 sequence data generated during this study were aligned in MEGA X [23], with UPGMA clustering for iterations 1 and 2. The cox1 alignment was transferred to Mesquite v.3.31 [28], translated (echinoderm/flatworm mitochondrial code) and inspected for internal stop codons. After the correct reading frame was determined, the first column was removed so that the reading frame began on position one, simplifying position-coding in downstream analyses. The final cox1 dataset was 474 bp. All codon positions in the cox1 dataset were evaluated for substitution saturation, as well as non-stationarity caused by base composition bias. Substitution saturation was assessed using the "Test of substitution saturation by Xia et al." function [65,66] as implemented in DAMBE v. 7.2 [64]; no significant substitution saturation was detected. Nonstationarity was assessed using the v 2 function in PAUP v. 4.0 [57]; significant non-stationarity was not detected. Thus, all codons in the cox1 dataset were used in downstream analyses. An unrooted Neighbor-joining analysis was conducted using MEGA X for the cox1 dataset to explore species boundaries, with the following parameters "Model/Method = No. of differences", "Substitutions to Include = d: Transitions + Transversions", "Rates among Sites = Gamma Distributed" and "Gaps/Missing Data Treatment = Pairwise deletion". Nodal support was estimated by performing 1000 bootstrap replicates. Pairwise differences were estimated for both the ITS2 cox1 datasets using the following conditions: "Variance Estimation Method = None", "Model/Method = No. of differences" and "Substitutions to Include = d: Transitions + Transversions" and "Gaps/Missing Data Treatment = Pairwise deletion".
The partial 28S rDNA sequences generated during this study were aligned with representative sequences of all aporocotylid genera available on GenBank (Table 1). Sequences were aligned using MUSCLE version 3.7 [17] run on the CIPRES portal [29], with ClustalW sequence weighting and UPGMA clustering for iterations 1 and 2. The resultant alignment was refined by eye using Mesquite v.3.31; the ends of the alignment were trimmed, and indels constituting more than three base positions and present in greater than 5% of the sequences in the dataset were removed (leaving a final trimmed dataset of 1254 base positions).
Bayesian inference and maximum likelihood analyses of the 28S dataset were conducted to explore relationships among these taxa. Bayesian inference analysis was performed using MrBayes version 3.2.7 [50] and maximum likelihood analysis using RAxML version 8.2.12 [55], both run on the CIPRES portal. The best nucleotide substitution model was estimated using jModelTest version 2.1.10 [16]. Both the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) predicted the TPM3uf + I + C model as the best estimator; Bayesian inference and maximum likelihood analyses were conducted using the closest approximation to this model. Nodal support in the maximum likelihood analysis was estimated by performing 1000 bootstrap pseudoreplicates. Bayesian inference analysis was run over 10,000,000 generations (ngen = 10,000,000) with two runs each containing four simultaneous Markov Chain Monte Carlo (MCMC) chains (nchains = 4) and every 1000th tree saved. Bayesian inference analysis used the following parameters: nst = 6, rates = invgamma, ngammacat = 4, and the priors parameters of the combined dataset were set to ratepr = variable. Samples of substitution model parameters, and tree and branch lengths were summarised using the parameters sump burnin = 3000 and sumt burnin = 3000. Aporocotylids of chondrichthyans were designated as functional outgroup taxa, following Warren et al. [61].

General results
Six species of Holocentridae were examined for aporocotylids on the Great Barrier Reef: 30 Neoniphon sammara (Forsskål), 10 Myripristis murdjan (Forsskål), nine S. diadema (Lacepède), nine S. spiniferum (Forsskål), three Sargocentron caudimaculatum (Rüppell), and two S. rubrum from off Lizard Island; and 17 S. rubrum and one N. sammara from off Heron Island. Adult aporocotylids were collected from N. sammara and S. diadema off Lizard Island and from S. rubrum off Heron Island. Adult worms were found in the branchial arteries, heart, vessels of the liver, head split wash and body split wash, but only in one or two of these sites in any individual fish. Eggs lodged in the gill tissue were found in all hosts infected by adult worms, as well as in four S. rubrum, three N. sammara and one S. diadema from which no adults were found; eggs were always concentrated in clusters at the tips of small numbers of gill filaments. Eggs lodged in heart tissue were found in just a single S. rubrum which was also infected by adult worms. Sequence data were generated for all host species/infection location combinations, from adults and from eggs. Four cox1 genotypes are present (Fig. 1), with one from only N. sammara at Lizard Island, one from only S. diadema at Lizard Island, one from S. rubrum at Heron and Lizard Islands, and one from S. rubrum at only Heron Island. Based on genetic, morphometric, and ecological data, we recognize the four genotypes as representing four distinct species belonging to a new genus; three are formally described, with the fourth lacking suitable morphological material.
Family Aporocotylidae Odhner, 1912 Genus Holocentricola n. gen.  Body lanceolate, ventrally concave, broadest at level of testis or caeca, with distinct terminal notch at posterior end, and distinct bulge sometimes present at level of uterus. Tegumental spines arranged in ventro-marginal transverse rows for entire body length, straight for most of body length, those in final 5-10 rows slightly curved with small hook on tip. Rosethornshaped or fused spines absent. Oral sucker poorly delineated, weakly muscularised, bearing concentric rows of fine spines. Mouth ventrally subterminal. Oesophagus almost straight to gently sinuous, thick-walled. Caeca form X-shape; intestinal bifurcation in middle third of body. Anterior caeca equal to subequal in length, much shorter than posterior caeca. Posterior caeca equal to subequal in length. Testis single, roughly rectangular, with margins irregularly lobed, immediately posterior to posterior margin of posterior caeca, usually extends laterally beyond lateral nerve cords. External seminal vesicle absent. Vas deferens sometimes widening posteriorly. Cirrus-sac retort-shaped, rounded anteriorly, dramatically narrowed posteriorly; anterior rounded portion contains seminal vesicle and pars prostatica; posterior narrow portion notably thickened at marginal genital pore, contains ejaculatory duct. Seminal vesicle round to ovoid, restricted to anterior, rounded portion of cirrus-sac, joining coiled pars prostatica. Ejaculatory duct long. Male genital pore on sinistral margin at distinct to indistinct marginal notch. Ovary oblong, roughly rectangular or wedge-shaped, medial, with margins irregularly lobed, immediately posterior to testis, usually extending laterally beyond lateral nerve cords. Oviducal seminal receptacle present. Oötype posterior to rest of genitalia, medial to submedial. Uterus weakly convoluted, passing anteriorly between oviduct and dextral side of cirrus-sac, ventrally overlapping posterior portion of ovary, then passing posteriorly, sinistral to cirrus-sac, to female genital pore; distal portion of uterus often forming prominent egg reservoir, creating distinct marginal bulge. Female genital pore dorsal, sinistro-submedial, separate from and anterior to male pore. Eggs in utero ovoid to subspherical, very thin-shelled, anoperculate. Vitellarium follicular, distributed from just posterior to dorsal nerve commissure to posterior half of testis or level of ovary, laterally exceeding nerve cords, largely confluent anterior to testis. Excretory vesicle small, saccular. Excretory pore at apex of terminal notch. In circulatory system of holocentrid fishes.
Site in host: Ventricle, branchial arteries, vessels of liver, wash of head split, wash of body split.
Etymology: This species is named from the Latin rufus (red) for the type and only host, the Red squirrelfish.

Remarks
Holocentricola rufus was found in all body sites examined, with adult worms in the heart (specifically the ventricle), branchial arteries of the gills, the major vessels of the liver, as well as in the wash of head split (gills already removed), and wash of entire body split (head and gills removed); however, specimens of this species were most commonly found infecting the branchial arteries. cox1 sequence data were generated for samples from all five infections sites and from eggs lodged in the tips of gill filaments; all sequences form a strongly supported clade in the neighbor-joining analysis, with no division by infection location. No adults were recovered from the single infection from Lizard Island but an ITS2 sequence was generated from eggs lodged in gill tissue; this sequence is identical to those from adult samples from Heron Island.
Site in host: Heart, branchial arteries, vessels of liver, wash of head split.
Etymology: This species is named from the Latin exilis (slender or thin) for the type and only host, the Slender squirrelfish.
Excretory vesicle small, saccular; paired collecting ducts not traceable. Excretory pore at apex of terminal notch.

Remarks
Holocentricola exilis was found in four of the five body sites examined (heart, branchial arteries, vessels of liver, wash of head split), but was most commonly found in the wash of the head split and the branchial arteries. The intensity of infection of H. exilis was notably lower than that found for H. rufus Type host: Sargocentron diadema (Lacepède), Crown squirrelfish (Holocentriformes: Holocentridae).
Site in host: Ventricle, branchial arteries, vessels of liver, wash of head split.
Prevalence: 2 of 9 (adults in one). Intensity: 4 worms in single fish from which adult worms were recovered.
Etymology: This species is named from the Latin coronatus (crowned) for the type and only host, the Crown squirrelfish.
Excretory vesicle small, pyriform; paired collecting ducts not traceable. Excretory pore at apex of terminal notch.

Remarks
Infections of H. coronatus were found in two of nine S. diadema; eggs were lodged in the gills of both infected hosts, and four adult worms (one worm in each of the four infected body sites) in one of the two. Sequence data for this species were derived from eggs in gill tissue.
Site in host: Branchial arteries, vessels of liver, wash of head split.
Intensity: 1 worm per fish, when adult worms were detected.

Remarks
This putative, undescribed species was found co-infecting, with H. rufus, three individuals of S. rubrum at Heron Island; only three specimens were collected, and just two hologenophores were available for morphological analysis. This species is genetically distinct from but sister to H. rufus in all phylogenetic analyses; despite the shared host and close phylogenetic affinity, the two species are clearly distinct morphologically. 81-138). Specimens of Holocentricola sp. A also have fewer spines per row than H. rufus (7 vs. 8-9), smaller spines (6 long vs. 7-8 long) and shorter spine rows (8-11 wide vs. 11-22). The two hologenophore slides have been lodged as voucher specimens in the QM in the hope that future collecting in the region will enable the description of this species.

Molecular results
cox1 and ITS2 data were generated for all four putative Holocentricola species, and 28S data for three species; genomic DNA of H. coronatus was derived from eggs lodged in gill filaments and thus the amplified 28S sequence was contaminated by host DNA. The four putative species of Holocentricola are clearly genetically distinct based on cox1 mtDNA and ITS2 rDNA data, differing at 34-61 base positions in the cox1 analysis (Fig. 1, Table 2) and 5-13 base positions in the ITS2 analysis. cox1 sequence data for H. rufus (the species for which the most replicate sequences were generated) demonstrated intraspecific variation at 0-5 base positions; no variation was found for three sequences of each of H. exilis and Holocentricola sp. A. Bayesian inference and maximum likelihood analyses of the 28S dataset resulted in identical phylograms (Fig. 4) Table 2. Total pairwise cox1 differences between species of Holocentricola, with number of differences below the diagonal and p-distances above.  This is strikingly different from the seemingly random distribution of eggs across the gills reported for species of other aporocotylid genera (e.g. [8,41,49,71]). We infer that gravid Holocentricola worms are highly mobile in the circulatory system and amass eggs in the reservoir, and, when laden, insert their posterior end into or enter a filament to lay. This interpretation is supported by the presence of the adults in sites throughout the circulatory system (the heart, branchial arteries, vessels of liver, head and body) and the rarity of eggs in the heart tissue of infected holocentrids; of the 81 holocentrids examined during this study just a single S. rubrum had a few eggs lodged in tissues of the ventricle. Eggs are routinely found lodged in the heart tissues (primarily the those of the ventricle) in fishes with a current or recent blood fluke infection (e.g. [52,59,68]); their absence in holocentrids suggests that Holocentricola eggs do not traverse the circulatory system passively. As far as we can determine, this pattern of confinement of the eggs to just a single or a few filament tips has not been reported for any other aporocotylids. We suspect that the paucity of blood fluke reports for holocentrids prior to this study relates to a lack of examination of this family of fishes, rather than an absence of species infecting them. The semi-cryptic nature of holocentrids means they are not commonly seen and seldom collected. As part of the extensive blood fluke sampling during the PhD of Nolan [32][33][34][35][36][37][38] over 1200 individual fishes were examined from the Great Barrier Reef; these 1200 fishes included just a single holocentrid, a N. sammara from off Heron Island [31]. During our own long-term collection program, we have examined over 19,000 individuals of 960 marine teleost species, but just 14 of these species are holocentrids; of these 14 species we have examined just six for blood flukes (those in this study), and adequate numbers (at least 30 individuals, following Cribb et al. [11]) for just one (N. sammara). Given our findings of four blood fluke species in a small number of holocentrid species examined from just two locations, we predict that a more thorough survey of holocentrids in Australia (of which there are 34 species; Bray [4]), and elsewhere globally (a further 54 species; Froese and Pauly [18]), will reveal a rich aporocotylid fauna.

Conclusions
This is the first report of aporocotylids from fishes of the family Holocentridae and the order Holocentriformes. The three new species described here (Holocentricola coronatus, Holocentricola exilis and Holocentricola rufus) are morphologically and genetically distinct from each other and from all other known aporocotylids. We predict that further examination of holocentrids will result in the collection of additional species of Holocentricola.