Biodiversity of symbiotic microalgae associated with meiofaunal marine acoels in Southern Japan

Sample collection, light microscopy, and establishing microalgal cultures

Macroalgal and sediment samples were collected from intertidal and subtidal zones at seven sites in Okinawa, Ishigaki, and Kochi (Fig. 1 and Table S1) from April–October 2022. Both macroalgal and sediment samples were collected from all sites. Areas exposed to sunlight which had seagrass and corals in the vicinity were targeted. Samples were collected in plastic containers and transported back to the laboratory for further processing. Acoels were detached from macroalgae and sediment samples by replacing seawater with an isotonic magnesium chloride solution (Schockaert, 1996). The solution was stirred and then filtered through a 60–100 μm mesh. The mesh was then placed in a petri dish with filtered seawater and removed after 10–20 minutes. The contents of both the petri dish, and the material retained on the mesh, were then examined under an Olympus CKX53 (Olympus, Tokyo, Japan) inverted microscope and Olympus SZ61 (Olympus, Tokyo, Japan) stereomicroscope. Acoels were identified based on their planula body-shape, and other characteristics including the statocyst, eyespots, rhabdoid gland cells, and microalgal symbionts (Achatz et al., 2013; Jondelius & Jondelius, 2020). Using hand-drawn glass capillary tubes, acoels were washed multiple times with filtered seawater to eliminate any contaminants, and isolated into 24-well tissue culture plates containing filtered and autoclaved seawater. Acoels were imaged with either an Olympus CKX53 (Olympus, Tokyo, Japan), or Zeiss Axioskop 2 Plus (Zeiss, Oberkochen, Germany) connected to a Canon EOS Kiss X8i digital camera (Canon, Tokyo, Japan). Individual acoels were either preserved in 99.5% ethanol in 1.5 ml microcentrifuge tubes or broken apart under a dissection microscope and the internal microalgae were transferred to 24-well tissue culture plates. A small portion of acoels (isolates 22–29) were broken and washed in multiple wells to check for multiple symbionts. When cultures were visually pigmented to the unaided eye, they were transferred to small petri dishes (Tetraselmis) or full-sized petri dishes (dinoflagellates). Microalgal cultures were maintained in half-concentration Daigo’s IMK Medium (Wako Pure Chemical Industries, Tokyo, Japan) and were transferred to new dishes with fresh culture media every month. Where possible, after host acoels were broken apart, host material was transferred into a 1.5 ml microcentrifuge tube for DNA extraction. Cultures were maintained at 25 °C (12:12, light:dark hours) in Biotron Incubator LH-350 (NK Systems, Tokyo, Japan). Motile stages of Tetraselmis and Symbiodiniaceae were mounted on glass slides and imaged under a Zeiss Axioskop 2 Plus microscope (Zeiss, Oberkochen, Germany) connected to a Canon EOS Kiss X8i digital camera (Canon, Tokyo, Japan).

DNA extraction, PCR amplification and sequencing

Genomic DNA from cultured microalgae and whole-acoel isolates was extracted according to manufacturer protocols with one of the following kits: DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany), QuickExtract FFPE DNA Extraction Kit (Lucigen, Middleton, WI, USA), or MasterPure Complete DNA and RNA Purification Kit (Epicentre, Madison, WI, USA) (Table S1).

For acoel isolates, nested PCR was performed on extracted DNA using general eukaryotic primers (Tables S1 and S2), and the products were cloned with an NEB PCR Cloning Kit (New England Biolabs, Tokyo, Japan) and then sequenced. For cultured microalgal material, nested PCRs were performed and then directly sequenced, or cloned and then sequenced. An additional Tetraselmis culture (K19), established from acoels collected in 2019 from Ikenoura, was also sequenced. All PCR reactions were first amplified using the SR1–28-1483R primer pair and then nested. From both the whole-acoel isolates and microalgal cultures, the 18S rRNA gene (henceforth, 18S) was amplified for Tetraselmis using SR1–18SRF and SR4–SR12 primer pairs, and a partial 28S rRNA gene (henceforth, 28S) was amplified for Symbiodiniaceae using the D1RF1–28-1483R primer pair. Apart from general eukaryotic primers, extracted host DNA was also amplified using a specific primer pair, Tear F–Tear R, designed from the sequences obtained from cloning. Polymerases used in the PCRs and thermocycler conditions are listed in Table S1; primers are listed in Table S2.

Following the nested reaction, PCR products were purified using Polyethylene Glycol (PEG) and directly sequenced or cloned. Acoel DNA inserts were screened by colony PCR using the S1512A–S1513A primer pair provided in the NEB cloning kit. Colony-PCR products were cut using a HAEIII restriction enzyme (Takara, Shiga, Japan) following the manufacturer’s protocol. Digested products were checked on a 1% agarose gel. A minimum of four replicates of each banding pattern were selected. Select colony-PCR products were purified again with PEG, and then sequenced with BrilliantDye Terminator v3.1 (NimaGen, Nijmegen, Netherlands), according to manufacturer protocol on a 3130 genetic analyzer (Applied Biosystems, Waltham, MA, USA). Occasionally, additional sequencing primers were used to join sequence fragments (refer to Table S2).

Phylogenic analyses of the 18S and 28S rRNA genes

Geneious Prime 2023.1.2 (https://www.geneious.com) was used to process Sanger sequencing reads. Cloning primers and low-quality regions (HQ% ≤ 50) were manually trimmed and assembled with the “De Novo Assemble” tool. Tetraselmis and dinoflagellate sequences were identified with NCBI’s BLAST search tool, sorted into their respective data sets, and compiled with existing data from Genbank. Multiple Alignment using Fast Fourier Transform (MAFFT) (Katoh et al., 2002) within Geneious was then used to align both data sets. Alignments were manually trimmed, resulting in a final alignment of 1,807 bps and 1,391 bps for Tetraselmis 18S and Symbiodiniaceae 28S, respectively.

Iqtree v1.6.12 (Nguyen et al., 2015) was used to select the best-fit model for both Maximum-Likelihood and Bayesian analyses. Under the Akaike Information Criterion with correction (AICc), the models TN+F+R2 and TIM+3+F+R3 were selected for Tetraselmis and Symbiodiniaceae, respectively. Bootstrap analysis was performed on 1,000 pseudo-replicates. Under the Bayesian Information Criterion (BIC), the general time reversible (GTR) substitution model was selected. MrBayes 3.2.7a (Huelsenbeck & Ronquist, 2001) was used to calculate Bayesian Posterior Probabilities. Four Markov Chain Monte Carlo (MCMC) chains were run with the general time reversible (GTR) substitution model for 10,000,000 generations and sampled every 100^th generation. The rate variation across sites were drawn from a gamma distribution with a proportion being invariable (lset rates = invgamma). Other parameters were set to the default settings. Chain convergence was checked with Tracer v1.7 and the posterior probability values were obtained from the consensus tree after the first 25% was discarded as burn-in. Molecular Evolutionary Genetics Analysis version 11 (MEGA11) (Tamura, Stecher & Kumar, 2021) was used to compute the uncorrected pairwise distance for Symbiodiniaceae sequences. Default settings were used: no variance estimation method, uniform rates among sites, and complete deletion of gaps and missing data.

Sequences generated from this study were deposited to NCBI. Accession numbers are listed in Table S1.

Sampling and identification of host acoels

A total of 66 acoels were sampled from seven sites (Fig. 1; Table S1). Acoels consisted of four main morphotypes that were differentiated based on symbionts, as well as features including ocelli, rhabdoid type, and body shape. The first (Fig. 2A) and second (Fig. 2B) morphotype harbored exclusively Tetraselmis. These were distinguishable, based on the numerous blunt-ended colored rhabdoids seen only in the second morphotype. A third type of acoel harbored exclusively Symbiodiniaceae and did not have any visible ocelli or colored rhabdoids (Fig. 2C). The fourth type were acoels collected from Ikenoura that contained Tetraselmis and Amphidinium. These also possessed three posterior caudal lobes (Fig. 3).

Of the 66 acoels collected, 36 contained exclusively Tetraselmis symbionts (Figs. 4–6); 26 contained exclusively Symbiodiniaceae (dinoflagellate) symbionts (Figs. 7 and 8); one contained both Symbiodiniaceae and Amphidinium symbionts (Fig. 8B); and three contained a mix of Tetraselmis and Amphidinium symbionts (Fig. 3).

Maximum-Likelihood analysis of 23 host acoel 18S sequences (Fig. 9) formed three genetic groups that corresponded with three different morphotypes: the first and second morphotype within Convolutidae and the third morphotype comprising a clade with Mecynostomidae. Genetic data were not obtained from 43 out of 62 acoel hosts, likely due to host DNA being lost while breaking apart host tissue to free algal symbionts.

No sequences for the fourth morphotype were obtained. Isolates within Convolutidae with corresponding genetic data harbored either Tetraselmis or Symbiodiniaceae symbionts, while isolates sister to Mecynostomidae harbored exclusively Symbiodiniaceae symbionts. Information on the isolate number, symbiont type, locality, and dates can be found in Table S1.

Cultivation of Tetraselmis and dinoflagellate microalgal strains

In the majority of cases, a single microalgal culture was established from one host acoel. Exceptions to these included isolates 19–21 collected from Ikenoura, isolate 61 (Fig. 8B) from Shiraho, and isolates 27 and 28 (Figs. 7H and 7J) from Kabira. Isolates 19–21 contained two types of symbionts, Tetraselmis sp. and Amphidinium gibbosum (Fig. 3). Cultures were established for both microalgae. Isolate 61 contained Symbiodinium sp. and A. gibbosum; but only Symbiodinium sp. was successfully cultured. Isolate 27 contained Cladocopium and Symbiodinium (A_IV), and isolate 28 contained Symbiodinium, subclades A_III and A_IV; all symbionts were cultured. In total, 32 Tetraselmis cultures, 26 Symbiodiniaceae cultures, and two Amphidinium cultures were established (and sequenced) from acoels (Fig. S2).

Molecular diversity of Tetraselmis

Maximum-likelihood analysis of Tetraselmis 18S sequences (Fig. 10) formed four main groups that clustered around T. convolutae, T. astigmatica, T. marina, and T. gracilis. The most often encountered Tetraselmis was clustered with T. astigmatica (26 cultures, and eight sequences from total-host DNA), followed by T. convolutae (four cultures, and three sequences from total-host DNA), T. marina (two cultures), and T. gracilis (one culture). Deeper nodes in the Tetraselmis phylogenetic analyses were largely unresolved.

Molecular diversity of Symbiodiniaceae

Maximum-Likelihood analysis of Symbiodiniaceae 28S sequences (Fig. 11) formed six clusters across three different clades: Miliolidium (one sequence from total-host DNA), Cladocopium (three cultures), and Symbiodinium (23 cultures, and eight sequences from total-host DNA). The Miliolidium clade included isolate 16 (Fig. 7E) that grouped with five other sequences isolated from Foraminifera and Porifera. The (uncorrected pairwise) distance between the type species, Miliolidium leei Pochon & LaJeunesse, 2021, and the Miliolidium sequence obtained in the present study was 0.5%. The Cladocopium clade included strains 27.1–27.3 from isolate 27 which were sister to C. infistulum, C. thermophilum, and C. goreaui. The intra-clade distances between our sequences and these species ranged from 1–4%. These strains had a 1.6% difference to the type species, Cladocopium goreaui. Many sequences in Symbiodinium grouped together with S. natans (13 cultures, and two from total-host DNA); the genetic distance did not exceed 1%. The remaining sequences formed a sister group with either S. tridacnidorum (nine cultures), S. microdriaticum (one culture and four sequences from total-host DNA extractions), or S. pilosum (two sequences from total host DNA extractions); genetic distances between our sequences and sequences they grouped with ranged between 0.5–2%, 2–3%, and 4–6%, respectively.

Phylogeny of other dinoflagellates sequenced from acoels

In addition to Symbiodiniaceae, the dinoflagellate Amphidinium was also observed (isolates 19–21) and cultured (isolates 61 and K22) from acoels. Amphidinium cultures were identified to the genus level based on the presence of an operculum. The 28S sequences from these Amphidinium isolates, 61 and K22, clustered with A. gibbosum (Fig. S1), including an A. gibbosum from the acoel Waminoa sp. (AB626894). Additionally, a Heterocapsa sequence (isolate 7) was sequenced, and was positioned sister to H. pseudotriquetra (Fig. S3). It should be noted, however, that Heterocapsa was not cultured or directly observed within acoels during this study.

Meiofaunal marine acoels harbor diverse lineages of Tetraselmis

In this study, we found at least four distinct lineages of Tetraselmis associated with marine acoels. Several Tetraselmis from Kochi and Ishigaki (isolates 19–21, and cultures K19, K22, 58, and 59) were closely related to T. convolutae from Symsagittifera roscoffensis (Parke & Manton, 1967). Tetraselmis convolutae strains from S. roscoffensis from Roscoff, France (RCC1563 and RCC1564) are the only Tetraselmis to be formally described from acoels (Parke & Manton, 1967). The novel Japan isolates from this study formed a sister clade to T. convolutae, but statistical support at shared nodes was weak. Whether Tetraselmis strains from this study are the same species as T. convolutae (from Roscoff) is difficult to conclude based on 18S data alone. Additional morphological data are needed to make a distinction between these closely related strains.

Apart from T. convolutae, we report members of the T. marina, T. gracilis, and T. astigmatica clades as symbionts of acoels for the first time. Specifically, isolate 5 and culture 57 from Kabira were closely related to T. verrucosa (KX904702), a free-living strain collected from Messolonghi, Greece (Chantzistrountsiou et al., 2016). Both Japan isolates were well-nested within the T. marina clade (Fig. 10), with a genetic distance ranging from 0.3–1.1%. The red-orange coloration indicated that culture 57 is T. verrucosa f. rubens, however, measurements of both the position and distribution of the chloroplast is needed to identify these isolates (Chantzistrountsiou et al., 2016; Hori, Norris & Chihara, 1983). Free-living T. marina and T. verrucosa f. rubens have been described (with morphological data) from Japan (Hori, Norris & Chihara, 1983). The present study provides the first molecular data of these Japanese lineages. Isolate 24 from Odo was positioned sister to T. gracilis (KU352751), a free-living strain collected from the Arabian Sea (unpublished NCBI data). However, another unpublished NCBI sequence, T. gracilis (KJ756816), was positioned in a separate clade and represents a strain from the North Sea (Butcher, 1959). Thus, the identity of isolate 24 from Odo, Japan remains uncertain though it is likely distinct from the T. gracilis described by Butcher (1959).

The majority of Tetraselmis sequences generated in this study clustered within the T. astigmatica clade. Isolates from multiple sites around Okinawa and Ishigaki tended to form multiple subclades (Fig. 10, I—IV). Currently, only two sequences that represent T. astigmatica exist (KSN-2002 and CCMP880). The strain CCMP880 (JN376804) is presumably the same strain used in the description of T. astigmatica (Hori, Norris & Chihara, 1982, 1983, 1986; Norris, Hori & Chihara, 1980). Both sequences are from unpublished sources. The topology of the tree shows multiple subclades closely related to T. astigmatica, but molecular phylogenetic analysis based on 18S alone lacks the resolution to clearly resolve these (Piganeau et al., 2011). Of note are inconsistencies found between light micrographs taken of Tetraselmis strains (culture 31, Fig. S2D) and the original T. astigmatica description (Hori, Norris & Chihara, 1982). The most intriguing of these is the presence of a stigma, a feature absent in the description of T. astigmatica by Hori, Norris & Chihara (1982). A more detailed investigation utilizing additional genetic markers (ITS, psbA. and rbcL), as well as ultrastructural data, is warranted to resolve what might be a species complex. Nonetheless, this study is the first to report lineages in T. marina, T. gracilis, and T. astigmatica as symbionts of acoels.

Reconciling novel molecular data into the contemporary taxonomic framework of Tetraselmis

Assigning species names to molecular clades and reconciling novel (molecular) data within this genus remains difficult. Classification of Tetraselmis species is based on the morphological characteristics of cells including symmetry, presence/position of the stigma, and the ultrastructure of the pyrenoid and flagella (Butcher, 1959; Hori, Norris & Chihara, 1982, 1983, 1986; Marin, Matzke & Melkonian, 1993; Norris, Hori & Chihara, 1980). Currently, molecular data (18S) is not available for over half the accepted (~30) Tetraselmis species. Only three species (T. jejuensis, T. indica, and T. verrucosa) have type strains for which both molecular and morphological data is available (Arora et al., 2013; Chantzistrountsiou et al., 2016; Hyung et al., 2021). Moreover, different sequences with the same name (T. convolutae, T. gracilis, and T. chuii) appear in different clades, further clouding Tetraselmis taxonomy (Arora et al., 2013; Hyung et al., 2021). This study established 32 cultures of Tetraselmis from acoels providing a vital resource to address fundamental issues surrounding Tetraselmis taxonomy and systematics.

Marine acoels also contain a diversity of dinoflagellate microalgae

The majority of dinoflagellate sequences generated from acoels in this study belonged to the family Symbiodiniaceae. Identification of these symbionts in acoels (except for those in Waminoa) has been completely based on observational data and lacks corresponding molecular information (Achatz, 2008; Achatz & Hooge, 2006; Hooge & Smith, 2004; Yamasu, 1982). Kunihiro & Reimer (2018) found that Waminoa in Okinawa contained only a single lineage of Cladocopium but speculated that acoels harbor a much more diverse range of Symbiodiniaceae.

Novel (28S) sequences generated in this study formed six clusters across Symbiodinium, Cladocopium, and Miliolidium: four in Symbiodinium, and one each in Cladocopium and Miliolidium (Fig. 11). This study is the first report of four lineages of Symbiodinium and Miliolidium, as symbionts of acoels. Within Symbiodinium, one cluster was positioned in subclade AIII; the sequences were grouped with S. natans. The presence of S. natans in cultures established from cnidarian or other metazoan hosts has been attributed to their abundance in the environment (e.g., contamination) or possible background symbiosis. They have also not been established as a symbiont under experimental conditions (LaJeunesse et al., 2018). This is intriguing, as acoels may be a possible reservoir, serving as a primary host for S. natans.

The remaining three clusters, one in subclade AII and two in subclade AIV, could represent novel lineages. These latter three formed separate branches with moderate support. The presence of S. natans in cultures established from cnidarian or other metazoan hosts has been attributed to their abundance in the environment (e.g., contamination) or possible background symbiosis. They have also not been established as a symbiont under experimental conditions (LaJeunesse et al., 2018). This is intriguing, as acoels may be a possible reservoir, serving as a primary host for S. natans. Other Symbiodinium sequences from this study did not group with known species in the phylogenetic analyses. The identity (or novelty) of these Symbiodinium strains will have to be confirmed through sequencing of a more variable genetic region (e.g., ITS), and by morphological comparisons.

Cladocopium sequences (isolate 27) grouped with other undescribed species, separate from C. goreaui and other described coral symbionts within that same clade. Our Cladocopium sequences (cultures 27.1–27.3) could not be compared with symbionts of Waminoa acoels—because different genes (ITS-2 and psbAncr) were utilized (Kunihiro & Reimer, 2018). It is important to note that a distinct strain of Symbiodinium (culture 27.4) was also established, along with Cladocopium, from isolate 27. This potentially shows some adaptability in acoels, like corals that also host a variety of Symbiodiniaceae at the same time.

A Miliolidium sequence was obtained from isolate 16. The genus Miliolidium has been reported as a symbiont of Foraminifera and Porifera, or as free-living (Pochon & LaJeunesse, 2021). It is unclear if our isolate is novel compared to Miliolidium leei, the type species. While sequences in this clade are genetically divergent from each other, our data did not have an associated culture that could aid in differentiating these species (morphologically). Still, the diversity of hosts that acquire Miliolidium make this genus an interesting target for understanding symbiosis between microalgae, metazoans, and protists.

Amphidinium and Heterocapsa were also sequenced in this study. Amphidinium from isolate 61 (Fig. 8B) and those from Ikenoura (from 2019 and 2022) (Fig. 3) were identified as A. gibbosum based on morphology and molecular phylogenics (Fig. S1). Amphidinium gibbosum has been reported from Waminoa litus and Heterochaerus langerhansi (Hikosaka-Katayama et al., 2012; Taylor, 1971). Both Waminoa and Heterochaerus are relatively large acoels (measuring approximately 2 mm). As Amphidinium are much larger, in comparison to other acoel symbionts, they may occur more frequently in larger acoels. Future collections, targeting larger acoels, may uncover a higher diversity of Amphidinium. Heterocapsa species have never been reported as a symbiont of acoels. The single Heterocapsa sequence in our data likely represents a food, or other environmental contaminant, not a symbiont of acoels. Heterocapsa was neither abundant nor cultured in our experiments.

Identity of acoels belonging to Convolutidae collected in this study

Fourty-four Tetraselmis-containing acoels were collected in this study. These are all believed to be Convolutidae, based on the presence of microalgae (Atherton & Jondelius, 2022; Jondelius & Jondelius, 2020). We observed two morphotypes, distinguished by the presence of colored rhabdoids (compare Figs. 2A to 2B). Of the 16 Convolutidae isolates sequenced (using 18S), individuals with colored rhabdoids formed a clade consisting of 14 identical sequences. This included isolate 33 which did not possess colored rhabdoids and contained Symbiodinium instead of Tetraselmis. This clade resembles the genus Convoluta—based on the distribution of colored rhabdoid gland cells, and body size (Hooge & Tyler, 2008). Only two individuals without colored rhabdoids (isolates 1 and 17) were sequenced. Both differed from any known Symsagittifera or Convolutriloba sequences; but they are believed to belong to one of these families as sagittocysts were visualized in a representative photo taken under oil-immersion (Fig. 2A). Isolates 1 and 17 (Figs. 1B and 1C) were only imaged with an inverted microscope, and sagittocysts were not apparent. Three of the isolates from Ikenoura (isolates 19–21), although lacking corresponding sequences, were identified as putative Amphiscolops, based on the three posterior caudal lobes and the presence of two symbionts (Tetraselmis and Amphidinium). This is similar to other Amphiscolops species.

Even though a large portion of the acoels collected in this study were not sequenced, it is likely that these represent a variety of lineages in Convolutidae. For example, light micrographs of isolates 5 and 57 (Figs. 6K and 6L) have a sparse distribution of rhabdoids, a blunt posterior end, and a centrally-located statocyst (not anterior) (Achatz, 2008; Achatz et al., 2010). Similarly, isolates 58 and 59 (Figs. 4J and 4K) have notably rounded anterior and posterior ends. Molecular identification of acoels was challenging because host individuals were broken apart to establish microalgal cultures. This likely lowered the success rate (~70% of hosts were not sequenced). Breaking apart the hosts also made histological observations impossible.

A potential novel family of acoels containing Symbiodiniaceae outside the Convolutidae

One of the more exciting discoveries from this study was a novel clade of acoels containing symbionts that fall outside the Convolutidae. The acoels from this clade contained a diversity of Symbiodiniaceae symbionts from Symbiodinium (subclades A_II, A_III, and A_IV) and Miliolidium. This is potentially the first report of non-convolutids containing microalgal symbionts. A prominent diagnostic characteristic of the family Convolutidae is the presence of microalgal symbionts—no other acoel family is known to contain microalgal symbionts (Achatz et al., 2013; Jondelius & Jondelius, 2020). This novel clade (seven individuals; isolates 2, 10, 12–16) was grouped sister to the Mecynostomidae, with only limited ML/BPP support (69/0.98). The shape and size of copulatory organs, or the presence of bursal nozzles could not be observed under bright field microscopy because the microalgal symbionts were dense. Most individuals also were not fully mature. Therefore, it was difficult to compare the morphology of these outstanding acoels to a contemporary taxonomic key (Jondelius & Jondelius, 2020). Histological slides and sperm axoneme structure are needed to confirm the identity (family) of this clade (Achatz et al., 2010).

Microalgal symbionts are diverse, but what drives preference?

The acoel-microalgal symbiosis is likely driven by the needs of the host acoel. If juvenile acoels do not take up symbionts from the environment, they fail to fully mature (Douglas, 1983). Vertical transmission of microalgae from parent acoels to offspring is rare and is only known in Waminoa (Hikosaka-Katayama et al., 2012). Acoels provide nitrogen for microalgal symbionts, and in return the microalgae provide necessary photosynthates for the acoel (Meyer, Provasoli & Meyer, 1979). The mechanism behind the recognition and selection of symbionts is still unknown, however, the acoel Symsagittifera roscoffensis is selective towards T. convolutae (Arboleda et al., 2018).

Observations in the present study suggest that meiofaunal acoels prefer either Tetraselmis or Symbiodiniaceae. Acoels appear to be repeatedly associated (loyal) with one of these symbionts. Acoels with and without colored rhabdoids all hosted Tetraselmis symbionts. However, isolate 33, a convolutid hosting Symbiodinium, was also observed. Isolate 33 was genetically identical to 13 other convolutids that all harbored Tetraselmis. This means that a single acoel species can contain either Tetraselmis or Symbiodiniaceae; but this was not observed concurrently within the same individual.

There were subtle patterns that could be observed within our dataset at the species level. Isolates 5 and 57, which were morphologically different from other acoels, harbored Tetraselmis belonging to the T. marina clade. Similar observations were made with isolates 58 and 59, with symbionts that were closely related to T. convolutae. Arboleda et al. (2018) demonstrated that S. roscoffensis could associate with six other species of Tetraselmis, albeit with varying levels of abundance. For example, T. rubens and T. subcordiformis appeared less dense within the host acoel, relative to T. convolutae. A larger sample size, based on molecular data from acoel hosts, would be needed to substantiate further claims regarding host-symbiont preference. In our study, dinoflagellates (Amphidinium) and Tetraselmis appeared together in only one acoel, the putative Amphiscolops from Ikenoura. Tetraselmis from these isolates were closely related to S. roscoffensis (Parke & Manton, 1967) and Amphiscolops oni (Asai et al., 2022). Tetraselmis found in these acoels seems to be limited to a single lineage, but this may have resulted from a sampling bias, as Amphiscolop-like acoels were not collected from Okinawa or Ishigaki. This association of Amphidinium and Tetraselmis might also be attributed to the relative size of the acoel (e.g., Amphiscolops) (Achatz, 2008; Asai et al., 2022; Ogunlana et al., 2005; Trench & Winsor, 1987).

Regarding seasonality, no clear patterns were observed with respect to either Tetraselmis or Symbiodiniaceae. While seasonality cannot be completely disregarded as a driver of symbiotic preference, regular sampling at set locations targeting the same acoel species would have to be performed, to address such questions.