Quatrefoil light traps for free-swimming stages of cymothoid parasitic isopods and seasonal variation in their species compositions in the Seto Inland Sea, Japan

Cymothoid parasitic isopods infest a wide range of fish of different taxa living in marine, brackish, and freshwater environments. Most research on the reproductive season of Cymothoidae has been done by collecting or monitoring host fish afflicted with cymothoid parasites. However, collecting ecological data on cymothoid species that infest non-commercial or endangered fishes is complex and challenging. We used a quatrefoil light trap to investigate the seasonal change in species composition of cymothoid free-swimming stages in the Seto Inland Sea, Japan. We also collected preliminary data for efficient light-trap sampling and showed its effectiveness in cymothoid-related research. From October 2020 to December 2021, 613 cymothoid free-swimming stages were sampled monthly. All obtained individuals were identified as Mothocya parvostis (596), Ceratothoa verrucosa (12), and Ceratothoa carinata (5) by DNA barcoding using cytochrome c oxidase subunit I and 16S rRNA gene sequences. Based on the number of M. parvostis mancae collected each month, M. parvostis was anticipated to reproduce from June to December, with two reproduction peaks each year, and C. verrucosa and C. carinata were expected to reproduce in June, July, and September, and September and October, respectively. In addition, free-swimming juveniles were captured, presumably after they had left their optional intermediate hosts. Furthermore, the most effective time to harvest cymothoids with light traps may be during high tide on the night of the new moon. This study serves as a methodological framework for future research on cymothoids using light traps.


Introduction
Cymothoidae Leach, 1818 is a cosmopolitan parasitic isopod family that includes 363 species and 42 genera (Boyko et al., 2008). The hosts of these isopods include a wide range of fish taxa found in marine, brackish, and freshwater settings (Yamauchi, 2016). These parasites adhere to their hosts at four different locations: the opercular cavity, the buccal cavity, the abdominal cavity, and the body surface (Smit et al., 2014). During their life cycle, free-swimming mancae (larvae) infest the host fish and develop into adult cymothoids via juveniles, after which their sex switches from male to female (Smit et al., 2014). Cymothoid parasites are known to negatively impact fish in response to environmental changes (Östlund et al., 2005;Kawanishi et al., 2016), therefore, understanding the diversity, life cycle, and host-parasite relationships of these parasites are essential for understanding and conserving the ecosystem. Reproduction biology is important for understanding the life cycle of cymothoids, especially the dynamics of the free-swimming stages.
Due to the limited collecting methods for cymothoids, very few cymothoid species have been investigated for their reproduction. Most investigations focus on adult or juvenile parasites on collected host fishes (Bello et al., 1997). However, this procedure can only be applied to easily collectable host fishes. Alternative approaches for examining the reproduction of cymothoid species could be derived from other fish research using their planktonic eggs in the sea. For example, egg density of the black sea bream Acanthopagrus schlegelii (Bleeker, 1854) in the ocean is used to estimate their spawning season (Kawai et al., 2017, spawning grounds (Kawai et al., 2021), and spawning time . Since cymothoids breed by releasing free-swimming mancae, it may be possible to determine the reproduction season by collecting mancae from the ocean. The free-swimming cymothoid stages can be caught using light traps (Jones et al., 2008;Saito et al., 2014).
In addition, species identification procedures for cymothoids can interfere with reproduction research. As cymothoids can only be identified based on morphological characteristics of adult females, their mancae, juveniles, and males need to be identified by molecular analysis . The reproductive season can be approximated by employing DNA barcoding to identify free-swimming stages captured by light traps and observing seasonal variation in appearance.
In this study, we investigated seasonal variation in species composition, abundance, and life stages of free-swimming cymothoids collected in the Seto Inland Sea where seven cymothoid species are distributed (Yamauchi et al., 2004;Yamauchi and Nagasawa, 2012;Nagasawa et al., 2014;Nagasawa and Tensha, 2016), and discussed the reproduction season and life cycle of each species. By integrating the light trap collecting method and DNA barcoding identification, this study serves as a methodological framework for investigating the reproduction and life cycles of cymothoids in their free-swimming stages.
The number of cymothoids collected on December 9th, 12th, and 15th of 2020 was compared at different tide levels to assess the effect of tidal range on cymothoid numbers and the best tidal level for their sampling. Each low tide, 1/3 tide, 2/3 tide, and high tide had one sampling operation.
We analyzed the number of cymothoids collected particularly from November 15th to December 15th in 2020 within the sampling period to assess the variation of collected cymothoids and the best days for sampling over the course of a month. Three sampling operations were carried out during the 90 min preceding and following the high tide every 3

days.
Based on the results of the above two surveys, monthly sampling was carried out to determine the seasonal fluctuation in the species composition of free-swimming cymothoids. Each month, three sampling operations were carried out during the 90 min preceding and following the high tide of the new moon.
After each operation, the trap was pulled up after turning the lights off. Plankton trapped in the net at the bottom of the trap was then packed in bottles and transported on ice to our laboratory. Following Bruce (1985), cymothoids were separated from the collected plankton. Each specimen was measured for its total length and then fixed in 99.5% ethanol. The life stages of cymothoids were identified following Aneesh et al. (2016); in particular, mancae were distinguished from juveniles by the absence of pereopods 7 (presence in juveniles). Water temperature was measured using a HOBO pendant logger UA-002-64 (Onset, MA, USA) placed at the sampling sites at a depth of 1 m. All applicable international, national, and institutional guidelines for the care and use of animals were strictly followed. All animal sample collection protocols complied with the current laws of Japan.

Molecular identification
Total DNA was extracted from the pereopods of the cymothoid specimen using the alkaline lysis method according to the recommended protocol for KOD FX Neo DNA polymerase (Toyobo, Osaka, Japan). The pereopod was mixed with 18 μL of NaOH (50 mM) and incubated at 95 • C for 10 min. The tubes containing 2 μL of Tris-HCl (1 M, pH 8.0) were extensively vortexed and centrifuged at 12,000 rpm for 5 min. The supernatant was separated and frozen at − 30 • C until use in the polymerase chain reaction (PCR). The primers LCO1490-Fujita (5 ′ -ACAAAHCATAAAGATATT-3 ′ ) and HCO2198-Fujita (5 ′ -ACTTCDGGGTGRCCRAAAAATC-3 ′ ) manually modified based on Folmer et al. (1994) were used to amplify partial cytochrome c oxidase subunit I (COI) sequences. Partial 16S rRNA sequences were amplified using primers 16Sar (5 ′ -CGCCTGTTTAACAAAAACAT-3 ′ ), and 16Sbr (5 ′ -CCGGTCTGAACTCAGATCATGT-3 ′ ) (Simon et al., 1994). The total volume for each PCR cocktail was 8.1 μL, which included 1 μL of DNA, 0.78 μL of ultrapure water, 4.06 μL of 2 × PCR buffer, 1.62 μL of dNTP mix, 0.24 μL of each primer (10 μM solutions), and 0.16 μL of KOD FX Neo DNA polymerase. The thermocycler profile for COI consisted of an initial denaturation at 94 • C for 2 min; followed by 35 cycles of denaturation at 98 • C for 10 s, annealing at 50 • C for 30 s, and extension at 68 • C for 45 s; and a final extension at 68 • C for 7 min. The thermocycler profile for 16S rRNA consisted of initial denaturation at 94 • C for 2 min; followed by 35 cycles of denaturation at 98 • C for 10 s, annealing at 50.5 • C for 30 s, and extension at 68 • C for 30 s; and a final extension at 68 • C for 7 min. Dye-terminator methods were used to sequence PCR products with an ABI 3130xl Genetic Analyzer (Applied Biosystems, CA, USA). The determined sequences were deposited in GenBank (LC741450-LC741549, LC741603-LC742126). Basic local alignment search tool (BLAST) was run on each sequence in the National Center for Biotechnology Information database. We established confidence values for identification with BLAST (≥99% similarity and an E-value = 0.0).
Prior to the analysis of the free-swimming cymothoids, the COI and 16S rRNA sequences of an adult male-female pair of Ceratothoa carinata (Bianconi, 1869) collected from the buccal cavity of a Japanese scad Decapterus maruadsi (Temminck and Schlegel,1844) caught in Sagami Bay, Kanagawa, Japan on April 4, 2022, were deposited in GenBank as reference sequences (COI: LC724050, LC724049; 16S rRNA: LC724051, LC724052). This cymothoids were identified as C. carinata because of the features: the contiguous and swollen antennular bases, the dorsal pereon surface with medial longitudinal ridge present, and the concave posterior margin on the pleotelson (Hadfield et al., 2016).

Results
We collected 589 mancae and 24 juveniles of Cymothoidae with the light trap. All individuals were successfully identified by DNA barcoding, with 596 individuals identified as Mothocya parvostis Bruce (1986), 12 individuals as Ceratothoa verrucosa (Schioedte and Meinert, 1883), and five individuals as C. carinata (Table 1; Fig. 3; Supplementary  Table S1). The mancae of M. parvostis were collected from June to December with abundance peaking in June and from September to December (Fig. 4). The juveniles of M. parvostis were collected from August to December, with the highest abundance in October (Fig. 5). The mean total length of M. parvostis did not overlap between mancae and juveniles ( Table 1). The mancae of C. verrucosa were collected in June, July, September, and October, with the highest abundance in July (Fig. 4). A single C. verrucosa juvenile was collected in July 2021 (Fig. 5). The total length of C. verrucosa juveniles was much larger than that of mancae (Table 1). No juvenile C. carinata were collected, and mancae were collected in September and October 2021 (Figs. 4 and 5). No cymothoids were collected from January to May. The dominant species for all months, except July 2021, was M. parvostis. In July 2021, Table 1 Results of sampling of cymothoid free-swimming stages collected by the quatrefoil light trap, and molecular identification using cytochrome c oxidase subunit I (COI) and 16S rRNA genes. Data for each individual are presented in Supplementary Table S1 C. verrucosa was the dominant species. The water temperatures at the sampling site ranged from a minimum of 11.24 • C to a maximum of 33.43 • C (Fig. 6). There are no data for water temperature from February 9th to April 13th and from October 31st to November 21st , 2021 due to challenges with the logger. DNA barcoding identified all cymothoids collected from November to December 2020 as M. parvostis. When the number of M. parvostis collected was compared to the daily tide level (low tide, 1/3 tide, 2/3 tide, and high tide), the species appeared to be most abundant on all survey days at high tide (Fig. 7). The number of M. parvostis collected on each sampling day increases during spring tide and decreases during neap tide (Fig. 8). Especially abundant M. parvostis were obtained during a spring tide at the new moon.

Discussion
In this study, we successfully performed quantitative collections of three free-swimming cymothoid species using a quatrefoil light trap. Although the quatrefoil light trap was developed for collecting fish larvae (Floyd et al., 1984), the method, together with DNA barcoding identification, is useful for studying cymothoid life cycles. Incidentally, free-swimming mancae die within 1 week-10 days after detaching from the brood pouch of cymothoid females if they could not find any hosts (Menzies et al., 1955;Hatai and Yasumoto, 1980;Sandifer and Kerby, 1983). Their starvation resistances affect estimating accuracy of the reproduction month of cymothoids using this method. Hickford and Schiel (1999) reported that the number of fish larvae collected by the light trap method negatively correlated with lunar illumination, with fish larvae being more abundant at new moons than at full moons. In this study, we documented changes in the number of M. parvostis collected by light trapping over a month. Free-swimming M. parvostis appeared more abundant at spring tides, especially at new moons than at full moons. The number of free-swimming cymothoid individuals collected by light trapping may be affected by both lunar irradiance and tidal effect. Since mancae of Cymothoidae use surface tension to float on the water's surface (Adlard and Lester, 1995), a higher number of mancae floating offshore may reach the traps during spring tides when the difference in tidal range is greater. In addition, the number of M. parvostis collected by with a light trap was lowest at low tide and highest at high tide. This method of collecting cymothoids with a light trap was considered most efficient at high tide on a new moon night. However, because the collected number of zooplankton depends on various physical and chemical factors such as flow speed, turbidity and depth (Hickford and Schiel, 1999;Mcleod and Costello, 2017), the factors affecting the number of cymothoids collected by the light trap method need to be examined in more detail.
In DNA barcoding using a database such as GenBank, sequences labeled with incorrect species names are problematic for species identification (Shen et al., 2013). However, the reference sequences (Table 1) used in this study can be trusted for the following reasons. The COI (LC159573, LC412904, LC549122-LC549151) and 16S rRNA sequences (LC159462, LC416622, LC549152-LC549178) of M. parvostis, were morphologically and genetically proven to be derived from M. parvostis by Fujita et al. (2020Fujita et al. ( , 2022. The COI (LC159556, LC160317) and 16S rRNA sequences (LC159444) of C. verrucosa, were registered by Hata et al. (2017). They did not report the morphology of these individuals. However, these two individuals were collected from the buccal cavities of the Crimson seabream Evynnis tumifrons (Temminck and Schlegel, 1843) and the red seabream Pagrus major (Temminck and Schlegel, 1844); C. verrucosa is the only cymothoid species reported in the buccal cavities of these fish species (Yamauchi, 2016). The COI (LC724050, LC724049) and 16S rRNA sequences (LC724051, LC724052) of C. carinata, were sequenced from the female and male morphologically identified in this study.
Seven species of cymothoids are reported in the Seto Inland Sea: Anilocra clupei Williams and Bunkley-Williams, 1986, C. carinata, C. verrucosa, andElthusa sacciger (Richardson, 1909), M. parvostis and Nerocila japonica Schioedte and Meinert, 1881, and N. phaiopleura Bleeker, 1857(Yamauchi et al., 2004Yamauchi and Nagasawa, 2012;Nagasawa et al., 2014;Nagasawa and Tensha, 2016). In this study, three species of Cymothoidae, namely M. parvostis, C. verrucosa, and C. carinata, were collected with a light trap from June to December and identified by DNA barcoding. In the collected cymothoids, both manca and juvenile M. parvostis were substantially more abundant. In addition, no cymothoids were collected from January to May. It is possible that the lower water temperatures (lower than 15 • C; Fig. 6) during this period either prevented cymothoid reproduction, or reduced the activity of mancae, thus reducing the area over which they could spread. Inoue (1941) proposed the life cycle of M. parvostis (previously identified as Irona melanosticta Schioedte and Meinert, 1884; Bruce, 1986) and suggested that M. parvostis mancae are released from adult females of M. parvostis around June. In addition, M. parvostis mancae infestation on juveniles of A. schlegelii was observed from June to August , on juveniles of the cobaltcap silverside Hypoatherina tsurugae (Jordan and Starks, 1901) from July to October, and on juveniles of the yellowfin seabream Acanthopagrus latus (Houttuyn, 1782) from October to December . In this study, free-swimming mancae of M. parvostis were collected from June to December, and not from January to May. This result suggests that M. parvostis reproduce in the Seto Inland Sea from June to December. However, the number of collected mancae varied, with higher numbers collected in December 2020, June 2021, September 2021, and October 2021. In addition, only a few M. parvostis mancae were collected in July, the only month that the species was not dominant. Therefore, M. parvostis likely has two reproduction peaks per year, one in spring and another in autumn. Sanada (1938) collected a total of 120 C. verrucosa individuals from P. major over a 10-month period in the Seto Inland Sea, and after observing mancae in the brood pouches of several females at a specific time of year he concluded that C. verrucosa mancae leave the brood pouch of females in late August. However, this does not mean that females do not release mancae in other seasons. In this study, C. verrucosa mancae were collected in June, July, and October, suggesting that C. verrucosa reproduction occurs over a wider period than the one proposed by Sanada (1938).
Despite numerous taxonomic studies, the life cycle and reproduction of C. carinata remain unknown. Our specimens collected in September and October 2021 are the first records of free-swimming C. carinata mancae. The collection suggests that C. carinata may reproduce in September and October. However, because C. carinata was not observed in October 2020, and only a small number of individuals was collected overall, further investigation is required to clarify its reproduction season.
In this study, we verified the utility of a new research method for the life cycles of cymothoids combined with quantitative light trap collection and DNA barcoding identification. We observed and recorded the reproductive season of three cymothoid species and provide evidence supporting the hypothesis that M. parvostis juveniles have a freeswimming period after leaving optional intermediate hosts. Thus, combining quantitative collection of free-swimming cymothoids by light trapping and species identification by DNA barcoding is an effective method for clarifying reproduction, distribution, and morphology of the free-swimming stages of cymothoids.

Author contributions
HF designed and executed the experiments, collected samples, and wrote the manuscript. KK and DD performed data analyses, provided useful ideas on research design, and edited the manuscript. TU acted as supervisor, provided useful ideas on research design, and edited the manuscript. All authors have reviewed and approved the final manuscript.

Declaration of competing interest
HF, KK, DD, and TU declare that they have no conflict of interest.