First Report of Blood Fluke Pathogens with Potential Risk for Emerging Yellowtail Kingfish (Seriola lalandi) Aquaculture on the Chilean Coast, with Descriptions of Two New Species of Paradeontacylix (Aporocotylidae)

Blood flukes are digeneans that infect wild and farmed fish that can cause a severe and potentially lethal disease in farmed fish. These parasites are undetectable in the larval stage based on macroscopic observations in the definitive host with the infection becoming evident when eggs accumulate in the branchial vessels. There are nine known species of the genus Paradeontacylix and seven exclusively parasitize Seriola spp. from several geographical areas. Seriola lalandi aquaculture farms are emerging at various localities in northern Chile. Here, we report, for the first time, two blood fluke species parasitizing S. lalandi in the Southeastern Pacific (Chile). In the laboratory, the gills and heart of fish were removed. The retained blood flukes were separated according to the infection site, fixed in 70% or 95% ethanol for taxonomic and molecular analysis, respectively. Morphometrical differences among the fluke species were evaluated with a principal component analysis (PCA) using proportional body measurements. Phylogenetic trees were constructed based on 28S rDNA, cox1 mDNA using Bayesian inference (BI), and maximum likelihood (ML). Based on morphology, morphometry, and molecular analyses, two new species are proposed: P. humboldti n. sp. from the gills and P. olivai n. sp. from the heart of S. lalandi. Both were clearly distinguished from other species of Paradeontacylix by a combination of morphologic features (posterior tegumental spines, testes arrangement, body size). The genetic distance (based on cox1) among species was >10%. P. humboldti n. sp. and P. olivai n. sp. are sister species (with a common ancestor) independent of P. godfreyi from S. lalandi in Australia. The newly identified parasites may pose a risk to farmed S. lalandi as aporocotylids have been the cause of diseases in farmed fish from other geographical areas. In addition, some cages of S. lalandi are currently maintained in an open circulating system, which could favor the transmission of these parasites (if involved hosts are present in the environment).


Introduction
Blood flukes are digeneans that infect the circulatory systems of wild and farmed fish [1-3] and can cause a severe disease in farmed Seriola spp. [4][5][6][7]. The disease occurs because the eggs accumulate in the gill filaments, leading to gill hyperplasia, egg encapsulation in the gills and ventricle, and papillae formation due to endothelial proliferation in the afferent branchial arteries [8]. In addition, hatching miracidia may cause multiple lesions and microhemorrhages, which in turn can trigger an inflammatory response and result in anemia [5]. This is especially the case when a massive hatch occurs, as has been reported for Sanguinicola inermis in Cyprinus carpio [9] and Cardicola sp. in Sparus aurata [10]. This pathology could be harmful in the aquaculture industry due this infection becoming Paradeontacylix humboldti n. sp. The specimens P. olivai n. sp. were obtained from the heart and they were bigger than specimens P. humboldti n. sp. obtained from the blood vessels ( Figure 1). P. olivai n. sp. was present in wild and farmed S. lalandi with an intensity of infection varying between 1 and 5 worms per fish heart (mean intensity = 2.3; prevalence = 20%). All of them were adults with eggs. About 39 eggs of P. olivai were recovered from gill filaments of farmed fish ( Figure S1). P. humboldti was present only in farmed fish with an intensity of infection of 20 worms recovered from an afferent branchial artery.

Results
All specimens collected in this study were morphologically identified as Paradeontacylix spp. From here on, those will be referred to as Paradeontacylix olivai n. sp and Paradeontacylix humboldti n. sp. The specimens P. olivai n. sp. were obtained from the heart and they were bigger than specimens P. humboldti n. sp. obtained from the blood vessels ( Figure 1). P. olivai n. sp. was present in wild and farmed S. lalandi with an intensity of infection varying between 1 and 5 worms per fish heart (mean intensity = 2.3; prevalence = 20%). All of them were adults with eggs. About 39 eggs of P. olivai were recovered from gill filaments of farmed fish ( Figure S1). P. humboldti was present only in farmed fish with an intensity of infection of 20 worms recovered from an afferent branchial artery.

Morphometric Analysis
PCA analysis showed that PC1, involving body width, male genital pore-posterior end distance, ovary length, ovary width, oviducal seminal receptacle length, oviducal seminal receptacle width, oötype length, and oötype width, explained 40% of the total variance. PC2, involving number of testes/BL and female genital pore-posterior end distance/BL, explained 18% of the total variance. Together, PC1 and PC2 explained 58% of the variance. P. kampachi and P. ibericus were in one group, and there was some overlap between P. sanguinicoloides and P. godfreyi. However, P. grandispinus, P. balearicus, and the two new species identified in this study were clearly different from each other ( Figure 2).

Morphometric Analysis
PCA analysis showed that PC1, involving body width, male genital pore-posterior end distance, ovary length, ovary width, oviducal seminal receptacle length, oviducal seminal receptacle width, oötype length, and oötype width, explained 40% of the total variance. PC2, involving number of testes/BL and female genital pore-posterior end distance/BL, explained 18% of the total variance. Together, PC1 and PC2 explained 58% of the variance. P. kampachi and P. ibericus were in one group, and there was some overlap between P. sanguinicoloides and P. godfreyi. However, P. grandispinus, P. balearicus, and the two new species identified in this study were clearly different from each other ( Figure 2).

Molecular Analysis
The unique sequences obtained in this study were coded with the following access numbers: MW599287-MW599288 (28S LSU rDNA) and MW598468-MW598470 (cox1 mDNA) (Table 1). Three sequences of P. humboldti n. sp. and five sequences of P. olivai n. sp. were obtained for each gene: 28S LSU rDNA (902 bp in length) and the cox1 mDNA (743 bp in length), respectively (Table 1). Regarding the 28S LSU rDNA, there was no intraspecific polymorphic sites for either species but there were 11 polymorphic sites between the new candidate species (P. humboldti n. sp. and P. olivai n. sp.). Regarding the cox1 gene, there were no intraspecific polymorphic sites among cox1 sequences of P. humboldti n. sp. while only one intraspecific polymorphic site was detected for P. olivai n. sp. and 67 polymorphic sites were detected between the new candidate species.

Molecular Analysis
The unique sequences obtained in this study were coded with the following access numbers: MW599287-MW599288 (28S LSU rDNA) and MW598468-MW598470 (cox1 mDNA) (Table 1). Three sequences of P. humboldti n. sp. and five sequences of P. olivai n. sp. were obtained for each gene: 28S LSU rDNA (902 bp in length) and the cox1 mDNA (743 bp in length), respectively (Table 1). Regarding the 28S LSU rDNA, there was no intraspecific polymorphic sites for either species but there were 11 polymorphic sites between the new candidate species (P. humboldti n. sp. and P. olivai n. sp.). Regarding the cox1 gene, there were no intraspecific polymorphic sites among cox1 sequences of P. humboldti n. sp. while only one intraspecific polymorphic site was detected for P. olivai n. sp. and 67 polymorphic sites were detected between the new candidate species.  The final alignment of all datasets (including the Paradeontacylix spp. sequences from GenBank) resulted in 910 bp for 28S LSU rDNA and 416 bp for cox1 mDNA. Phylogenetic reconstructions based on the total molecular evidence (910 bp + 416 bp = 1326 bp) resulted in the same general topology for both inference methods (ML and BI) ( Figure 3). The species of Paradeontacylix were clustered into a single monophyletic clade, which was strongly supported by a high posterior probability in the BI analysis (pp = 1) and by the bootstrap support value in the ML analysis (bML = 99) ( Figure 3). In each phylogenetic tree there were three subclades with moderate-to-strong statistical support: one involved P. humboldti n. sp. and P. olivai n. sp., the second clade (which is a sister clade of the new candidate species) involved P. balearicus and P. grandispinus, and the third clade involved P. ibericus and P. kampachi ( Figure 3). P. godfreyi appeared as a basal species within the genus Paradeontacylix.  Etymology: The species is named in honor of German naturalist Alexander von Humboldt.
Specimens: Holotype and one paratype have been deposited in the MNHNCL, encoded as: PLAT-15021 and PLAT-15022, respectively.
Based on two whole-mounted gravid adult specimens. Body measurements are given in Table 3. Body smooth, elongated, dorsoventrally flattened, lancet-shaped ( Figure 4A). Longer than wide by 12-13 (12.5, n = 2) times. Marginal tegumental spines ventrolateral, 286-436 (361, n = 4) rows on either side of body; same size throughout body, averaging 4 long by 1 wide (maximum width at spine base), 4-9 per row, decreasing in number toward both extremities to about 4 at anterior and posterior ends. Posteriorly, 9 large tegumental spines, conspicuous, claw-like distally, arranged in 4 longitudinal rows each comprising Based on cox1 mDNA, P. humboldti n. sp. and P. olivai n. sp. had genetic distances >10% between them and between them and the other species of Paradeontacylix. The pairwise sequence divergences for each of the two molecular markers are shown in Table 2. Table 2. Genetic distance between Paradeontacylix spp based on 28S LSU rDNA and cox1 mDNA, respectively. Lower half shows the percentage differences between the paired comparisons using 28S LSU rDNA (based on 910 bp) and upper half shows the percentage differences between the paired comparisons using cox1 (based on 416 bp). Etymology: The species is named in honor of German naturalist Alexander von Humboldt. Specimens: Holotype and one paratype have been deposited in the MNHNCL, encoded as: PLAT-15021 and PLAT-15022, respectively.

Remarks
The P. humboldti n. sp. displays all the diagnostic characteristics of the genus Paradeontacylix McIntosh, 1934. Here, we have initially classified it as a species of the genus Paradeontacylix according to testes distribution as irregular (random, without a pattern) and regular (in two well-defined rows). Based on this criterion, the first group includes P. grandispinus, P. balearicus, P. sanguinicoloides, and P. godfreyi with irregular testes distribution, while a second group includes: P. kampachi, P. buri, P. ibericus, and P. megalaspium, which present regular distribution. The distribution of testes in the new species P. humboldti n. sp. is irregular. Additionally, it differs significantly in the number of testes, intestine shape, and ovary shape from P. grandispinus (63-69 vs. 19-32 testes; H-shaped vs. X-shaped intestine; oval-shaped vs. heart-shaped ovary, respectively) and P. balearicus (63-69 vs. 20-26 testes; H-shaped vs. X-shaped intestine; oval-shaped vs. shield-shaped ovary, respectively). It also differs from P. sanguinicoloides in the shape of the large posterior tegumental spines (claw vs. rose thorn, respectively), ovary shape (oval-shaped vs. heart-shaped, respectively) and position of female genital pore (antero-sinistral vs. antero-mesal of male genital pore, respectively). P. humboldti n. sp. differs from P. godfreyi in the ovary shape (oval-shaped vs. heart-shaped) and testes position in the body (middle and posterior vs. posterior third). It is also approximately half of the maximum body widths of P. sanguinicoloides (155-179 vs. 330, respectively) and P. godfreyi (155-179 vs. 357-566, respectively). P. humboldti n. sp. is half the body length of P. godfreyi (1858-2352 vs. 3739-4215, respectively). Additionally, the position of the female genital pore, which opens without crossing vas deferens, is only shared with S. godfreyi.      Etymology: The species is named in honor of Dr. Marcelo Oliva, who has studied marine parasites for over 40 years in Chile.
Specimens: Holotype and one paratype have been deposited in the MNHNCL, encoded as: PLAT-15019 and PLAT-15020, respectively.

Remarks
The distribution of testes in P. olivai n. sp. is regular, as they are distributed in two rows as in P. kampachi, P. buri, P. ibericus, and P. megalaspium. However, P. olivai n. sp.

Discussion
The detection and correct identification of potential pathogenic parasites are an important as a first step to support emergent aquaculture [3,22]. The combination of morphologic and molecular characteristics is a strong and reliable method to identify species [32]. The conservative 28S LSU gene is an efficient marker for analyzing the phylogeny of digeneans at taxonomic levels, such as the genus and family [33,34], while the use of a mitochondrial marker as a DNA barcode has been useful for species discrimination [15,35]. Furthermore, using two or more independent loci (as in this study) provides advantages when estimating species-level relationships and testing hypotheses regarding species delimitation [36]. By analyzing new DNA sequences from P. humboldti n. sp. and P. olivai n. sp., our analysis complemented the previous analyses on the phylogenetic relationships between species of Paradeontacylix reported [15,16]. We found that P. godfreyi from S. lalandi in the Indian Pacific Ocean, south Australia, is located at the basal position within the Paradeontacylix spp. clade with strong nodal support. However, this result could be related to the genetic marker because previous authors used ITS-2 fragment, which is more variable than 28S gene (domains C1-D2). Regardless of this difference in tree topology, previous authors [15,16] demonstrated absence of influence of the host-phylogeography on the phylogenetic relationship of Paradeontacylix spp. in S. dumerili. More precisely, P. grandispinus (found in Japan) and P. balearicus (found in the Mediterranean) are genetically related as well as P. kampachi (found in Japan) and P. ibericus (found in the Mediterranean). Additionally, we found that the two new species, P. humboldti n. sp. and P. olivai n. sp. from S. lalandi in the southeastern Pacific, belong to the same (monophyletic) clade and represent a sister clade of P. balearicus and P. grandispinus. This suggests that they underwent a relatively recent divergence within the phylogeny of the genus Paradeontacylix. These results, however, must be confirmed using other more resolutive molecular markers.
The final host fish of Paradeontacylix spp., such as S. dumerili and S. lalandi, are known to be long-distance migratory fish species with genetically structured populations across their extensive geographical distributions [37][38][39][40]. Two populations of S. dumerili occur in the northeast Atlantic [39] and two other populations occur in the northwest Pacific [37]. This suggests that populations of this host species have undergone genetic divergence in the past as a consequence of historical processes [41,42] and that the ancestral parasite species had existed before the separation of the S. dumerili populations between Japan and the Mediterranean [15]. This would explain how a host fish species can come to harbor a pair of genetically related Paradeontacylix species (e.g., S. dumerili harbors P. grandispinus (in Japan) and P. balearicus (in the Mediterranean), and S. dumerili also harbors P. grandispinus (in Japan) and P. balearicus (in the Mediterranean)), which are reported to be restricted by the current geographical distribution of their host species populations. Similarly, there are at least four genetically distinct S. lalandi populations worldwide [18,40] with a single population of S. lalandi distributed in the south Pacific [40]. However, the spatio-temporal genetic structure for S. lalandi from the southeastern Pacific coast (the same area as in the present study) [43] and two populations of S. lalandi (on the Australian and New Zealand coasts) have been reported [44]. This could explain the existence of the two new species of Paradeontacylix in S. lalandi in the southeastern Pacific (P. humboldti n. sp. and P. olivai n. sp.), which differ from the species P. godfreyi recorded in S. lalandi in the Indian Pacific [17]. Therefore, further studies are required to clarify whether one or more Paradeontacylix species parasitize S. lalandi on the Australian coast and to know whether this parasite presents high or low host specificity as suggested by Hutson and Whittinton [17].
All marine aporocotylids whose life cycles are known use terebellid polychaetes (Nicolea gracilibranchis, Longicarpus modestus, and Reterebella aloba; Terebella sp.; Neoamphitrite vigintipes) as intermediate hosts [6,45,46]. However, the intermediate hosts for Paradeontacylix spp. are unknown, but it has been suggested that, for aporocotylids, direct penetration by cercariae is the dominant infection route in fishes, i.e., the infection could be independent of the host diet [3]. This is critical information as it implies that the life cycle and infection potential of the parasite may be independent of the trophic web. Currently, in northern Chile the emerging aquaculture capture wild fish to improve the genetic variability of the brood fish, and already at least one dead fish has evidenced infection by Paradeontacylix spp. This means that if the parasite exists in the environment (in intermediate hosts and wild fishes), and if there is an open aquaculture system in the region, there is a high probability that the captive fish will acquire the parasite. In the practice, blood flukes of fish are difficult to detect as they infest the host vascular system [3,47]. Montero et al. [48] reported, for the first time, that although P. ibericus infections in farmed S. dumerili were undetectable based on macroscopic observation, there were encysted cercariae (schistosomula) in fixed histological sections of muscle, which could be useful for early diagnosis of this pathology. After experimentally evaluating the parasite development, they reported that small juvenile P. ibericus worms were still found in the muscles and lymphatic system >100 days after the transfer of the fish to tanks, and P. ibericus adults (with a recognizable reproductive system) were recovered 8 months after the transfer. There was a higher intensity of P. ibericus in the girdle muscles, head kidney, and sinus venosus (involving both juveniles and adult worms) while a low intensity (involving only adults) was detected in the gills. In our study, wild fish were infected mainly with P. olivai adults detected in the heart while farmed fish were infected with both species, although predominantly with P. humboldti, and numerous eggs were observed in the gills. Further studies focused to evaluate infection levels, proportion of each species, presence/absence of eggs in the gill, or heart and their variability in the sampling periods will be necessary to elucidate the early biological traits and life cycles of the two new blood fluke species infesting S. lalandi in the southeastern Pacific as it was done by Montero et al. [48].
As different parasites species can have different infection patterns and in turn respond differentially to control treatments, the identification of the two new species of Paradeontacylix described here is crucial to develop a diagnosis protocol of these potential pathogens in farmed fish, which should be accompanied by records on the occurrence of these parasites, monitoring of the fish condition, and mortality due to any of these parasites. This protocol will allow the implementation of preventive management and control of this potential disease in the emerging fish farming industry in Chile.

Sample Collection
Twenty specimens of S. lalandi ranged between 46 and 84 cm fork length were acquired at a fish market in Antofagasta (24 • S), Chile, captured by an artisanal fishery in the nearshore area (24 • S-26 • S) during the summer season (January-February) in 2018 and 2019. In these months, wild S. lalandi migrate to the Chilean coast when the water temperature (17 • C-21 • C) increases [43]. Additionally, one specimen (about 4 kg) of S. lalandi, captured in the summer of 2017 and maintained captive in the hatchery (with an open circulatory marine water intake and effluent system) of the Univ. Antofagasta was found dead and infected with blood flukes in August 2019. The fish was characterized as having an opened mouth and opercula, showing typical signs of suffocation. In the laboratory, the gills and heart of the fish were removed. Each heart was opened and washed in a Petri dish with freshwater, which was then filtered with a sieve. Each gill arch was dissected longitudinally and washed following the same protocol. The retained blood flukes were carefully separated according to the infection site (heart or gill) and fixed in 70% or 95% ethanol for taxonomic identification and molecular analysis, respectively. Parasitological indexes (mean intensity and prevalence) were calculated [49].

Morphological Description and Morphometrical Analyses
The flukes were stained with acetocarmine or Gomori's trichrome, dehydrated in ethanol (70-100%), cleared with clove oil (Sigma-Aldrich, Taufkirchen, Germany), and mounted in Eukitt ® mounting medium (O. Kindler GmbH, Freiburg, Germany). The flukes were photographed (M125 camera; Leica, Wetzlar, Germany) and measured using the ImageJ software [50]. Measurements were made in micrometers (µm) and are given as the range, followed by the mean and the number of structures measured or counted in parentheses. Specimens were drawn using a compound microscope with a drawing tube. The type material was submitted to Museo Natural de Historia Natural (MNHNCL) in Santiago, Chile.
Morphometric analyses involved comparing the measurements of the specimens in this study with the measurements of P. godfreyi, P. sanguinicoloides, P. grandispinus, P. kampachi, P. balearicus, and P. ibericus obtained from original publications ( Table 2). For this, we used the measurements (minimum and maximum) directly reported in the publications while other measurements were estimated from the drawings. P. buri and P. megalaspium were not considered because these species do not have large posterior tegumental spines. The source and number of specimens per host species and geographical area are given in Table 2.
To evaluate differences in morphometry among the fluke species, a principal component analysis (PCA) was performed using proportional body measurements as all relevant fluke measurements are correlated with body length [51]. For this analysis, the following 16 body measurements divided by total body length (BL) were used: maximum body width, number of marginal tegumental spine rows, posterior spine length, esophagus length, anterior caeca-intestine distance, posterior caeca-intestine distance, number of testes, testicular area length, male genital pore-posterior end distance, ovary length, ovary width, oviducal seminal receptacle length, oviducal seminal receptacle width, oötype length, oötype width, and female genital pore-posterior end distance. These analyses were performed using the Statistica 7.0 software (StatSoft Inc., Tulsa, OK, USA).

DNA Extraction and Amplification
The DNA was isolated following a modified version of a protocol reported in [52]. This involved treatment with sodium dodecyl sulphate, digestion with proteinase K, NaCl protein precipitation, and subsequent ethanol precipitation. The DNA was eluted in nucleasefree water and quantified using a BioSpec-nano spectrophotometer (Shimadzu, Japan).

Phylogenetic Reconstruction
The sequences obtained in this study were aligned with sequences of Paradeontacylix spp. obtained from GenBank using Clustal X [56] (Table 1). Visual inspection was then performed using ProSeq v2.91 [55] in order to edit the length of the final dataset. The jModelTest v0.1.1 tool [57] was used to identify the best evolutionary model for each gene. Gene concatenation (LSU + cox1) was performed using Mesquite v2.75 [58]. Phylogenetic trees were constructed based on 28S LSU rDNA, cox1 mDNA, and the concatenated genes using Bayesian inference (BI) and maximum likelihood (ML) analyses. Five members of the Aporocotylidae family, Cardicola forsteri, C. opisthorchis, Psettarium nolani, P. sinense, and Plethorchis acanthus, were selected as outgroups based on their close phylogenetic relationships with the genus Paradeontacylix. Sequences of the outgroup taxa were obtained from GenBank (Table 1).
The BI analyses were conducted using MrBayes [59] with the following parameters: nst = 6 and rates = invgamma according to the evolutionary model determined by jModeltest v0.1.1 for each gene (GTR + G + I for 28S LSU rDNA and TIM2 + G + I for cox1 mDNA, and replaced by GTR + G + I for MrBayes software). Each analysis was performed for 5,000,000 generations, with one run of four chains, sampling every 100 generations. Support for nodes in the BI tree topology was based on posterior probability. The initial 25% was discarded as burn-in. The results were visualized using TRACER v1.7 [60]. The ML analyses were performed using W-IQ-TREE (http://iqtree.cibiv.univie.ac.at/ [61]  Finally, the pairwise p-distances for LSU rDNA and cox1 mDNA sequences among multiple species of Paradeontacylix were analyzed using the MEGA v6 software [62].

Conclusions
P. humboldti n. sp. and P. olivai n. sp. described in the present study constitute two new species of blood flukes that infest Seriola lalandi. The identity of each of the two new species is supported by morphological, morphometric, and molecular data.