Myxozoan infection in thinlip mullet Chelon ramada (Mugiliformes: Mugilidae) in the Sea of Galilee

Mullets (Mugilidae) are economically important fish in Israel. Two species of mugilids (i.e., the thinlip mullet Chelon ramada and the flathead grey mullet Mugil cephalus) have been stocked in the Sea of Galilee (Lake Kinneret) in order to increase fishermen’s income and lake water quality. These catadromous species do not reproduce in the lake, consequently, fingerlings have been introduced every year since 1958. Few additional mugilid species have been introduced unintentionally together with these two species, including C. labrosus. Following a survey of myxozoan infections in the Sea of Galilee, we described Myxobolus pupkoi n. sp. infecting the gill arches of C. labrosus, and reported Myxobolus exiguus from visceral peritoneum and gall bladder of C. ramada. Our study indicates that the parasites infecting C. ramada and C. labrosus belong to a lineage of myxozoans infecting mugilids. This result suggests that the infection took place in the Mediterranean Sea, where the fingerlings were caught, before their introduction into the Sea of Galilee. Since 2018 only farm-raised fingerlings have been introduced. We thus recommend to closely monitor the presence of these parasites in the future to determine if the presence of parasites disappear with the introduction of farm-raised fingerlings.


Results
Myxozoan parasites were searched for in ten M. cephalus and twenty-three C. ramada specimens sampled in the Sea of Galilee on November 2020. While no myxozoan parasite was observed in M. cephalus, plasmodia were found in the gall bladder, in the visceral peritoneum lining the intestine and the gill arches of four C. ramada specimen. Each of the four fish was infected by a single parasite. No infection was found in other organs such as kidney, liver, eyes, brain, fins, muscles, and scales. The 18S rRNA sequences from the plasmodia from gall bladder and visceral peritoneum were identical to M. exiguus reported from Portugal (sequence accession number: MH236070) 14 , except for three insertions in the last 7 bp of the sequence, which most probably correspond to sequencing errors since most are located in the primer region. The parasite obtained from gill arches did not match any other myxosporean sequence available in the NCBI database.

Myxobolus exiguus Thélohan, 1895 from the Sea of Galilee.
Plasmodia: Small plasmodia (about 0.5 mm), visible with naked eye, rounded, creamish white, freely floating in the bile or present in the visceral peritoneum lining the intestine, containing 300-500 myxospores per plasmodium.

Myxospore:
The morphological data are based on spores isolated from two plasmodia, one originating from the gall bladder and another from the visceral peritoneum. Few morphological differences were found between the spores of M. exiguus infecting the visceral peritoneum and the gall bladder of C. ramada.
Polar capsules pyriform and bottle-necked, equal in size and arranged side by side with two prominent pores at the anterior side of the myxospore measuring 3.79 ± 0.12 (3.     The new species is very similar in size to M. parenzani, which has been described to have round myxospores (~ 5.4 × 5.4 µm). However, it differs from the present species in having larger polar capsules. The polar capsules of M. pupkoi n. sp. are 2.4 ± 0.20 µm long and occupy about half of the myxospores. Conversely, the myxozpores of M. parenzani have been described to be ~ 2 µm long, to be positioned along the membrane (rather than side by side) and to only occupy the upper third of the spore. Although M. pupkoi n. sp. and M. parenzani infect the same host the morphological differences mentioned above justify the definition of a novel species.
Phylogenetic analysis. The phylogenetic analyses included 79 sequences of the 18S rRNA gene from the mugiliform-infecting lineage of myxobolid, as well as four outgroup sequences. The trees reconstructed based on maximum likelihood and Bayesian criteria only differ in the position of a few low-supported branches (Fig. 3). These trees are in agreement with previous studies on myxozoans infecting mugilids 10,11,14 .
The phylogenetic tree divides the mugiliform-infecting lineage into two clades. The first harbors only Myxobolus ramadus and Myxobolus cochinensis (ML bootstrap support BP = 99; Bayesian posterior probability PP = 0.83). The second, which includes most species (ML = 75; PP = 0.98), is composed of three subclades. One subclade includes myxobolid lineages, infecting mainly fish hosts from the genus Chelon, and is where M. exiguus and the newly described species Myxobolus pupkoi n. sp. branch. The second subclade includes myxobolid lineages infecting mainly fish hosts from the genus Mugil. Both subclades include myxozoans from annelid hosts, for which the fish host is unknown (Fig. 3). The third subclade includes M. suppamattayai and Tetraspora discoidea (ML = 80; PP = 0.99).

Discussion
The present study describes M. pupkoi n. sp. and presents a novel tissue of infection, the gall bladder, for M. exiguus. It also reveals their presence in the Sea of Galilee. Morphologic and morphometric differences were observed between the myxospores of M. exiguus found in the visceral peritoneum and in the gall bladder, and their taxonomic identity could only be confirmed based on the 18S rRNA analysis. These data illustrate, once more, the importance of 18S rRNA sequences to complement the morphological identification of myxozoan species 9,10,19 , especially in the case of mugilids, which are infected with closely related myxobolid parasites 10 .
Tissue specificity is a characteristic of many histozoic myxozoans and an important taxonomical character [24][25][26][27] . Unfortunately, because of the small number of infected fish observed, the infected tissues were used for molecular identifications rather than histological preparation. Future work should determine, which tissue of the gill arch is   24,27 . Myxobolus exiguus is a histozoic parasite described from the visceral peritoneum 14 . Its presence in the gall bladder bile, which is a coelozoic localization, is thus surprising. However, it is possible that the plasmodia found in the bile originated from the gall bladder wall or the hepatic bile ducts and thus is of a histozoic origin 25,28 . Both M. exiguus and M. pupkoi n. sp. belong to a well-defined clade of mugilid infecting myxozoans 14 . Members of this clade have also been found in marine, brackish and freshwater annelids present in estuaries 1,10,29 . Since mugilids are not native species from the Sea of Galilee and Jordan Valley system, it is most likely that these myxozoans were introduced to the Sea of Galilee with their mugilid hosts during stocking 5,30 , being the infection originated in the Mediterranean or the coastal plain estuaries of Israel from where the fingerlings originated before their introduction to the lake.
The fact that mugilid fingerlings are hosting numerous myxozoan parasites, sometimes even within a single individual, has been noted by Sharon et al. 11 . It was further suggested that the introduction of wild caught mugilid to new growing area may lead to the spread of their myxozoan parasites 11 . We here show that this prediction likely reflects the course of events that occurred in Israel where the transfer of infected fingerlings from the coastal environment led the presence of alien parasites in the Sea of Galilee.
Myxozoans need two hosts to complete their life cycle, it may thus seem that myxozoans are not a serious threat to the Sea of Galilee since the annelid hosts were not transferred with the fish hosts. However, the establishment of a transmission cycle in that environment using local annelids species cannot be ruled out.  30 . These freshwater annelids are known to harbor myxozoans closely related to the ones described in this work 10 (see Fig. 3). This suggests that the repeated introduction of infected mugillid fingerlings to the Sea of Galilee could have led to the establishment of myxozoan parasites to this new environment. While we did not find myxozoan parasites on the M. cephalus specimen studied during this work, their past introductions from wildcaught fingerlings pose the same threats than C. ramada [2][3][4][5] . Indeed, M. cephalus fingerlings from the north of Israel were found to be infected by another myxobolid species (sequence MF118765, Fig. 3) 11 , which has been considered to be M. adiposus 10 . This observation suggests that more than two myxobolid species may have been introduced to the Sea of Galilee with their mugilid hosts. Since 2018, only farm raised fingerlings of M. cephalus www.nature.com/scientificreports/ have been introduced to the Sea of Galilee. These fingerlings are expected to be parasite-free. If the introduced myxozoans cannot reproduce due to the lack of annelid hosts, we expect to stop observing myxobolid infected mugilids in the future. We thus recommend to closely monitor the presence of these parasites in the next 5 years to determine if these myxozoan parasites disappear.

Methods
Parasite material and morphological identification. Mugil cephalus specimens (n = 10) with a length of 18-20 cm, and Chelon ramada specimens (n = 23) with a length of 15-20 cm freshly collected from the Sea of Galilee, were obtained from a local fisherman and immediately transported to the lab. The fish sampling was performed under the authorization of the Fisheries and Aquaculture Department of the Israeli Ministry of Agriculture and Rural Development (authorization provided on 11.11.2020). The fish specimens were carefully examined externally for the presence of plasmodia followed by a dissection of each fish sample. Gills, scales, brain, kidney, visceral peritoneum, and other body parts were removed and examined under a stereo-microscope. When plasmodia were identified, fresh myxospores were photographed under a compound microscope with a DS-Ri2 Nikon photographic unit. The myxospores were stained with freshly prepared Ziehl-Neelsen and Giemsa solutions 31 . Spore measurements were done using a calibrated ocular micrometer.
DNA isolation and amplification. The samples for molecular work were fixed in absolute alcohol and stored at −20 °C. The extraction of genomic DNA was performed using the Qiagen DNeasy Blood and Tissue Kit according to the manufacturer's instructions. PCR amplification of the 18S rRNA was performed using both universal and myxosporean specific primers [32][33][34] (Table 3 The obtained PCR products were purified using the ExoSAP method 35 . Sequencing was performed using the external primers ER1B1 and ER1B10 together with six additional primers (Table 3) at the DNA Sequencing Unit at Tel-Aviv University on an ABI 3500xl Genetic Analyzer (Applied Biosystems™). The obtained sequences were visualized, assembled and edited using Geneious 11.1.5.

Phylogenetic reconstructions.
To reconstruct the phylogenetic relationships of the Israeli species, we first blasted the obtained 18S rRNA sequences, against the NCBI nr databased (https:// blast. ncbi. nlm. nih. gov/) on 05.09.2021. This allowed us to determine that both sequences belong to a distinct lineage of myxobolid infecting mugilids in the Histozoic III lineage sensus 14,36 . We downloaded all sequence hits from this clade that were longer than 800 bp. The sequences of Henneguya shaharini (EU643630), H. rhinogobii (AB447992), H. cynoscioni (JN017203), and Myxobolus khaliji (KC711053) were used as outgroups.
Because the start and end of sequences can include sequencing errors, we excluded the first and last 10 base pairs of all sequences, except for sequence MH183025 Myxobolus bragantinus for which we removed the first 20 bp based on a manual examination. The webserver Guidance2 37 (last accessed 15.12.2021) was used to align the edited sequences and to remove ambiguously aligned positions. Specifically, the sequences were aligned with the MAFFT algorithm under the options "Max-Iterate: 1000" and "Pairwise Alignment Method: localpair". Positions with a score below 0.93 were removed as well as positions with more than 50% of missing data. Finally, we also removed positions corresponding to the annealing regions of the primers 18Se (5′-CTG GTT GAT CCT GCC AGT -3′) 38 and 18Sr (5′-CTA CGG AAA CCT TGT TAC G-3′) 39 , the primers used to amplify most sequences in our sequence alignment 10,14,40 . The final dataset included 83 taxa and 1781 positions ( Supplementary Files 1 and 2).
Phylogenetic relationships were reconstructed both under the maximum likelihood and the Bayesian criteria. Maximum likelihood analyses were performed with the program IQ-TREE version 1.6.12 41 . The analyses were run with the options -m MFP -b 1000 (i.e., ModelFinder + tree reconstruction + 1000 non-parametric bootstrap replicates). The model selected based on the BIC criterion was the TVM + F + R3 model. Bayesian reconstructions were performed with PhyloBayes MPI version 1.8 42 under the CAT + GTR + GAMMA 4 model. Four chains ran Table 3. PCR primers used for the amplification and sequencing of the 18S rRNA gene.