Nutritional intake of Aplanochytrium (Labyrinthulea, Stramenopiles) from living diatoms revealed by culture experiments suggesting the new prey–predator interactions in the grazing food web of the marine ecosystem

Labyrinthuleans (Labyrinthulea, Stramenopiles) are recognized as decomposers in marine ecosystems but their nutrient sources are not fully understood. We conducted two-membered culture experiments with labyrinthuleans and diatoms to discover where labyrinthuleans obtain their nutrients from. The results showed that Aplanochytrium strains obtained nutrients by consuming living diatoms. Aplanochytrium cells did not release digestive enzymes into the medium, but adhered to diatom cells via the tip of their characteristic ectoplasmic net system to obtain nutrients from them. The chloroplast and cell contents of the diatoms shrank and were absorbed, and then the number of Aplanochytrium cells rapidly increased as multiple aplanospores were released. To estimate the effect of labyrinthulean organisms including Aplanochytrium on marine ecosystem, we explored the dataset generated by the Tara Oceans Project from a wide range of oceanic regions. The average proportion of all labyrinthulean sequences to diatom sequences at each station was about 10%, and labyrinthulids, oblongichytrids, and aplanochytrids were the major constituent genera, accounting for more than 80% of labyrinthuleans. Therefore, these groups are suggested to greatly affect the marine ecosystem. There were positive correlations between aplanochytrids and phototrophs, green algae, and diatoms. At many stations, relatively large proportions of aplanochytrid sequences were detected in the size fraction larger than their cell size. This implied that Aplanochytrium cells increased their particle size by adhering to each other and forming aggregates with diatoms that are captured by larger zooplankton in the environment, thereby bypassing the food web pathway via aplanochytrids to higher predators. The intake of nutrients from diatoms by aplanochytrids represents a newly recognized pathway in the grazing food chain in the marine ecosystem.

As mentioned above, aplanochytrids are cosmopolitan and abundant heterotrophs that interact with zooplankton, so they may have important ecological roles that have been overlooked so far. However, the aplanochytrids, like other labyrinthuleans, were cultured in the artificial media containing organic matter in the most previous studies, because they have been convinced as decomposers targeting organic matter such as marine detritus (e.g., [14,31]). So, the nutrient sources of aplanochytrids have never been determined. Therefore, the role and impacts of aplanochytrids in the marine ecosystem are unknown. In this study, we discovered that the aplanochytrids actively consume nutrients from living diatoms and then increase their biomass before reproducing in two-membered culture experiments. Diatoms are the major primary producers in coastal areas, so this discovery suggests the new prey-predator interactions in the grazing food web of the marine ecosystem.

Two-membered culture and cell counting
After 2 weeks of incubating diatom strains in f/2 medium, the diatom cells were harvested, washed, and resuspended in f/2 medium. Zoospores of labyrinthulean strains were inoculated into the diatom cultures. The method to induce zoospores was as follows: vegetative cells were spread on d-GPY agar plates and incubated at 20˚C or 25˚C for 1 to 2 days. After colonies became visible, seawater was added into the agar plate and incubated for several hours to allow zoospores to be released into the seawater. The Aplanochytrium spp. strain 'unidentified thraustochytrid 1' (NBRC 110846) did not produce zoospores, so vegetative cells of these strains were inoculated instead. After culturing the vegetative cells in d-GPY liquid medium for 1 week, the cells were harvested, washed, and resuspended in f/2 medium. Fuchs-Rosenthal counting chambers (Minato Medical, Tokyo, Japan) were used for cell counting and observed under BX60 differential interference contrast (DIC) microscope (Olympus, Tokyo, Japan). We defined the cells with intact chloroplasts as living and also defined the cells with shrunk chloroplasts (or empty cells) as dead. The cells of Aplanochytrium and Skeletonema were continuously counted in single-and two-membered cultures in order to compare the growth process in each culture condition. The cells of Skeletonema strain NIES 324 and Aplanochytrium strain KMPB N-BA-107 in a single-membered culture in f/2 medium were counted. Also, the cells of Aplanochytrium and Skeletonema in a two-membered culture in f/2 medium were counted. Cell counts were performed every two days (0, 2, 4, 6, 8, 10 days after starting the culture experiments).

Two cultures separated by a filter in 'Beppu Flasks'
To determine the effect of digestive enzymes from Aplanochytrium cells in the medium, we established separated cultures in 'Beppu Flasks' (Nihon Pall Corporation, Tokyo, Japan) [53], which have two partitions separated by a filter. Using this apparatus, substances such as enzymes that dissolve in the medium can pass through this filter but the organisms cannot. Each partition of the flask contained f/2 medium. Aplanochytrium was inoculated into one part and Skeletonema was inoculated into the other part. The method of the two-membered culture was as described above. The combinations of Aplanochytrium and Skeletonema cultured in this system are shown in Result section. The experiments using 'Beppu Flasks' confirms whether Aplanochytrium cells release components that kill algae, such as digestive enzymes, into sea water or predators that ingest nutrition directly from targeted prey. Statistical analyses were performed using Student's t-test on the combination of Aplanochytrium and Skeletonema cell number shown in Fig 8. Significance was accepted at p < 0.05.

Light microscopy
We conducted light microscopy observations using a BX60 differential interference contrast (DIC) microscope (Olympus, Tokyo, Japan) and captured images with an AxioCam HRc camera (Carl Zeiss, Hallbergmoos, Germany). For continuous observations, we used an Axiovert 200 DIC microscope (Carl Zeiss) and captured digital images with an MC170 camera (Leica, Wetzlar, Germany).

Scanning electron microscopy
The cells were fixed with 2.5% v/v glutaraldehyde and 0.1 M sucrose in 0.1 M cacodylate buffer and were post-fixed with 1.5% OsO 4 . The samples were dehydrated in a graded ethanol series followed by substitution with isoamyl acetate. The samples were successively dried by the CO 2 critical point method. After applying an OsO 4 coating, the samples were observed under a JSM-7200F scanning electron microscope (SEM) (JEOL, Tokyo, Japan).

Analyses of metabarcoding data from Tara Oceans Project
The Tara Oceans project sampled oceanic waters from 2009 to 2013 ( [54][55] https://www.embl. de/tara-oceans/). In this study, we used 766 million raw rDNA sequence reads from 334 size-fractionated plankton samples collected from 47 stations. We focused on two depth layers (surface and deep chlorophyll maximum (DCM)). Sequence information was evaluated after dividing the plankton samples into three size fractions (<20, 20-180 and 180-2000 μm). Sequence identification has been described elsewhere [54]. We confirmed the identification of all 351 operational taxonomic units (OTUs) classified in the Labyrinthulea in the Tara Oceans database. All OTUs were added to the aligned sequence data set (Ueda et al. 2015 [31] and Honda et al. 1999 [3]), aligned with a profile alignment process using the ClustalX software program [56]. The phylogenetic tree was generated by the neighbor-joining (NJ) method using the distances of the Tamura-Nei model with the pairwise deletion option as the gap treatment in MEGA 7 [57]. The correlations were calculated between aplanochytrids and phototrophs (excluding dinoflagellates), chlorophytes, bacillariophytes, and copepods at the sampling sites where aplanochytrids ranged from most abundant to 30th most abundant. The best fit trendlines were estimated by the method of a least-square linear regression and R-squared values were calculated using with Microsoft Excel software (Microsoft Corporation, USA). Statistical analyses were performed using Student's t-test between ratio of thraustochytrid in surface and DCM layers.

Observations of nutrient intake from diatoms by Aplanochytrium cells
Aplanochytrium spp. strains have a simple life history, in which the vegetative cells release aplanospores (see [14,36,58]). The vegetative cells were spherical with a diameter of 5-10 μm and developed ectoplasmic nets of occasionally more than 50 μm in length (

Comparison of assimilation from diatoms among labyrinthuleans
The numbers of labyrinthulean cells and Skeletonema living/dead cells in the two-membered culture were counted (Fig 6). Previously, it was confirmed that there was almost no increase in the number of labyrinthulean cells in single culture in algal inorganic medium (i.e., f/2 medium). Fig 6 shows the labyrinthulean strains arranged in order of increasing cell density. The cell densities of Aplanochytrium sp. strains SEK 602 and 717 were significantly higher than those of other strains. Correspondingly, the abundance of the Skeletonema living/dead cells in two-membered cultures were significantly lower with these Aplanochytrium sp. strains than with the other strains. The abundance of living Skeletonema cells was low, indicating that Aplanochytrium sp. strains fed on Skeletonema cells faster than the Skeletonema cells grew. Three other Aplanochytrium spp. strains (KMPB N-BA-107, SEK 754, SEK 758) also showed comparatively high cell densities, and resulted in low abundance of Skeletonema cells in twomembered cultures. In these experiments, all Aplanochytrium strains obtained nutrition from Skeletonema cells via the ectoplasmic nets, but no other tested organisms were able to obtain nutrition from Skeletonema cells.

Continuous cell counting of Aplanochytrium and Skeletonema in singleand two-membered cultures
To explore the predator/prey relationship between Aplanochytrium and Skeletonema, cells were continuously counted in single-and two-membered cultures. First, the cells of Nutritional intake of Aplanochytrium from living diatoms Skeletonema strain NIES 324 were counted in a single axenic culture in algal inorganic medium. The cell density increased gradually until the 6 th day after the start of culture, and thereafter the number of cells became constant (stationary phase) ( Fig 7A). Second, the cells of Aplanochytrium strain KMPB N-BA-107 were counted a single axenic culture in algal inorganic medium (Fig 7C). There was no increase in cell density during the 10-day incubation period. In a two-membered culture of Aplanochytrium and Skeletonema cells, the cell density of Skeletonema did not increase, but that of Aplanochytrium cells continued to increase  logarithmically for 10 days (Fig 7B). These results indicated that Aplanochytrium cells grew by absorbing nutrients from Skeletonema cells.

Effect of digestive enzymes from Aplanochytrium cells in medium
As mentioned above, the diatom cells were digested by Aplanochytrium cells, suggesting that Aplanochytrium cells may produce and secrete digestive enzymes. Therefore, we investigated Nutritional intake of Aplanochytrium from living diatoms whether digestive enzymes were released into the medium using the 'Beppu Flask' system. The results of control tests are shown in Fig 8A-8C. First, Aplanochytrium and Skeletonema were cultured in each of the two compartments, As the results in this experiment, Skeletonema cells grew while Aplanochytrium cells hardly proliferated (Fig 8D). This indicated that Aplanochytrium cells did not make soluble secretions of diatoms as nutrients, and did not release digestive enzymes into the medium. Next, Aplanochytrium and Skeletonema cells were incubated in one compartment as a two-member culture, while Aplanochytrium cells were cultured alone in the other compartment. The Aplanochytrium cells grew well in the two-membered culture with Skeletonema cells, but hardly proliferated in the single culture ( Fig 8E). These results suggested that Aplanochytrium cells digested the diatoms with degrading enzymes, but the degradants did not disperse in the medium. When Aplanochytrium and Skeletonema cells were cocultured in one compartment and Skeletonema cells were cultured alone in the other, the Skeletonema cells cultured alone grew well, but those co-cultured with Aplanochytrium cells did not ( Fig 8F). This result indicated that some substances that kill diatoms, such as the digestive enzymes, were not dispersed in the medium from Aplanochytrium cells as an influential concentration. Together, these findings indicated that Aplanochytrium cells produce decomposing enzymes by contact with diatom cells and can efficiently absorb all the degradation products.
We estimated the contributions of the major abundant OTUs in surface and DCM layers (Fig 9A). Three phylogenetic groups (aplanochytrids, oblongichytrids and labyrinthulids) accounted for 80%-90% of total labyrinthulean sequences on average. In the surface layer, aplanochytrids accounted for 25.3% (4.5%-72.5%), oblongichytrids for 30.4% (2.3%-65.4%), and labyrinthulids for 33.7% (3.8%-84.0%) of the total number of sequences identified as Labyrinthulea. In the DCM layer, aplanochytrids accounted for 24.1% (4.1%-56.3%), oblongichytrids accounted for 26.8% (5.2%-51.1%), and labyrinthulids accounted for 30.8% (0.0- 60.9%) of the total number of sequences identified as the Labyrinthulea. The average ratios of labyrinthulean sequences to diatoms and to all organisms in surface and DCM layers across all sampling sites are shown in Fig 9B and 9C. The average ratio of labyrinthulean sequences to diatom sequences at each station was about 10% at both depths, while the ratio of labyrinthuleans to all organisms was higher in the DCM than in the surface water. Fig 10 shows the relative contributions of aplanochytrids in the different size fractions in the Tara Oceans V9 data set. The seawater samples from each sampling site were filtered through differently sized filters, and then aplanochytrid sequences were detected in the subsamples with different particle sizes. The average proportions of aplanochytrid sequences out of total sequences in the filtered samples with particle sizes less than 20 μm, 20 to 180 μm, and 180 to 2000 μm were 84.2%, 14.8%, and 2.7%, respectively. A single vegetative cell of Aplanochytrium is ca. 5-10 μm, so these results suggested that aplanochytrids formed aggregates with each other and/or with floating matter including digested diatoms, or were eaten by large zooplankton. At 18 sites (43% of the 42 sampling sites), the proportion of aplanochytrids sequences in the subsample with particles of >20 μm exceeded 25%. At three sites, aplanochytrids accounted for more than half of the sequences in the subsample with particles of >20 μm.
We investigated the correlations between aplanochytrids and phototrophs (excluding dinoflagellates), chlorophytes, bacillariophytes, and copepods at the sampling sites where aplanochytrids ranged from most abundant to 30 th most abundant (Fig 11). There were positive correlations between aplanochytrids and phototrophs (excluding dinoflagellates), chlorophytes, and bacillariophytes (R 2 = 0.15, 0.29, and 0.10, respectively). Although aplanochytrids can ingest nutrients from diatoms, it is noteworthy that the highest correlation was between aplanochytrids and chlorophytes. This suggested that aplanochytrids may also ingest nutrients from green algae or their secreted substances. There was almost no correlation between aplanochytrids and copepods (R 2 = 0.01).

Discussion
The two-membered culture experiments showed that only Aplanochytrium cells consumed nutrients from living Skeletonema and Chaetoceros diatoms via their ectoplasmic nets. The Aplanochytrium cell numbers rapidly increased after releasing multiple aplanospores. The intake the nutrients from diatoms by labyrinthuleans was first reported by Gaertner (1979) [59]. In that study, cells of Schizochytrium sp. grew on the surface of the following diatoms: Thalassiosira nordenskioeldii, Chaetoceros sp. and Thalassionema nitzschioides. As the cells of Schizochytrium grew, the protoplasm of the diatoms shrank inside the silica shells. Labyrinthuleans adhered to up to 35% of diatom cells in field samples collected from Rosfjord, Norway. Raghukumar (1986) [60] reported that the cells of Ulkenia visurgensis adhered to 'moribund' diatoms such as Coscinodiscus, Navicula, Nitzschia, Grammatophora, and Melosira. In our two-membered culture experiments, however, there was little growth of Schizochytrium Nutritional intake of Aplanochytrium from living diatoms aggregatum and Ulkenia sp. with Skeletonema diatoms. Taxonomic identification of labyrinthulean cells in environmental samples by microscopic observations is difficult, so species should be confirmed by molecular identification methods. In this study, comparison of cell growth in the two-membered culture experiments suggested that only Aplanochytrium strains were able to absorb nutrients from living diatoms.
It was reported that the distribution of thraustochytrids was related to deep chlorophyll maximum layer in the Hawaiian, Arabian and Pacific seas [30,40,[61][62]. In both of those reports, the cell numbers of thraustochytrids were directly counted after acriflavine dye staining of sulfurylated polysaccharides in the cell wall, so the classification and phylogeny of these thraustochytrids were not clearly shown. Li et al. (2013) [40] showed that 80% of 18S rDNA phylogenetic lineages obtained by sequencing labyrinthulid-specific amplicons aligned with aplanochytrids in Hawaiian sea water. Our experiments showed that Aplanochytrium strains are able to absorb nutrients from living diatoms. Therefore, the correlation between cell density of thraustochytrids and chlorophyll value is reasonable. However, it should be noted that recent quantitative PCR and metagenomic analyses of labyrinthuleans in Chinese coastal habitats showed the negative correlation between aplanochytrids and chlorophyll a [43].
We explored the dataset generated by the Tara Oceans Project from a wide range of oceanic regions to characterize the diversity patterns of labyrinthuleans on a global scale. At almost all sampling stations, labyrinthulids, oblongichytrids, and aplanochytrids were the major constituent genera and together accounted for about 80% of labyrinthulean organisms. The average proportion of aplanochytrids out of total labyrinthulean organisms at each station was 25.3% and 24.1% in the surface and DCM layers, respectively. In our analyses of the Tara Oceans database, we detected positive correlations between aplanochytrids and phototrophs (excluding dinoflagellates), chlorophytes and bacillariophytes. The average ratio of labyrinthuleans to total diatoms was about 10% (Fig 9). Diatoms contribute around 20% of global primary productivity [51,[63][64]. Thus, labyrinthuleans including aplanochytrids may affect the marine ecosystem.
Aplanochytrids have been detected in the gut and fecal pellets of copepods and salps in the tropical Indian and North Pacific Oceans [34,37,39,65]. Hirai et al. (2018) [39] reported that aplanochytrids, which were recognized by 18S rDNA metagenomic analysis, were possibly the major food source of Calanus sinicus, a large copepod common in the coastal waters of the subtropical western North Pacific [66]. The proportion of Aplanochytrium was higher in gut contents than in environmental samples, suggesting that C. sinicus positively selected aplanochytrids as prey [39]. This copepod was shown to actively capture particles with a diameter of 20-50 μm, but few particles smaller than 10 μm [67]. The size of a single vegetative cell of Aplanochytrium is 5-10 μm in diameter; however, we observed that Aplanochytrium cells adhered to each other and formed aggregates with diatoms, resulting in larger particle sizes. In the environment, labyrinthulean organisms are frequently observed to attach to aggregates [40,68]. These aggregates may be consumed by C. sinicus and larger zooplankton. Nutritional intake of Aplanochytrium from living diatoms Labyrinthuleans ingest nutrition via their ectoplasmic nets (e.g., [2,9,[11][12]69]). In the present study, we observed that the tip of the ectoplasmic nets of Aplanochytrium cells adhered to and penetrated into diatom cells, and then the chloroplasts and cell contents of the diatoms shrank and were absorbed. Because the ectoplasmic nets spread out over a large area forming a fine network to search for prey, they can capture bait at least 50 μm in diameter, even though the single cells are 5-10 μm in diameter.
Our experiments with two cultures separated by a membrane in 'Beppu Flasks' showed that Aplanochytrium cells could absorb nutrients by directly adhering their ectoplasmic nets to diatoms, without releasing digestive enzymes into the medium. Iwata & Honda (2018) [69] showed that ectoplasmic nets recognize the food source by adhering to it and become thicker, during which secretion of digestive enzymes and absorption of digested substances actively occur. This feeding of labyrinthulean organisms including Aplanochytrium is superficially similar to 'diffusion feeding', which is known among the heliozoans, foraminiferans, and radiolarians (see [70]). The tentacles of the above-mentioned protists are supported by massive bundles of microtubules and the attached prey is captured in the food vacuole by phagocytosis. In contrast, the ectoplasmic nets consist of actin filaments, and develop a very thin network structure [71][72][73]. The ectoplasmic net penetrates the diatom cell. The ratio of the volume surrounded by the feeding apparatus to the volume of the cell body is larger in Aplanochytrium than in other protozoa. Therefore, feeding of Aplanochytrium via its ectoplasmic net is an Nutritional intake of Aplanochytrium from living diatoms efficient method to ingest nutrition with a smaller investment than that of diffusion feeding in protists.
In the freshwater environment, the zoosporic true fungi, chytrids, also consume nutrients from phytoplankton, including diatoms, and play an important role as prey for zooplankton in the aquatic ecosystem (see [74]). Chytrids and labyrinthuleans are superficially similar in their morphology at the vegetative stage; that is, both cells are spherical and develop pseudopod-like structures. In Aplanochytrium, vegetative cells develop ectoplasmic nets radially around themselves and wait for diatoms to stick to the nets, similar to the strategy of spiders to capture insects in their webs. In chytrids, after zoospores settle on the surface of diatoms, the chytrid cells penetrate and develop rhizoids inside the diatoms to ingest nutrients. In this case, when large phytoplankton that are inedible by zooplankton are infected by chytrids, nutrients within Nutritional intake of Aplanochytrium from living diatoms host cells are transferred to zooplankton by grazing on the small edible zoospores of parasitic chytrids, in a pathway known as the 'Mycoloop' [74]. Thus, the strategy of nutritional intake using the pseudopod-like structure differs markedly between Aplanochytrium and chytrids. However, the nutrients supplied to zooplankton have some similarities. Chytrid zoospores supply poly-unsaturated fatty acids (PUFAs) to zooplankton [75]. Labyrinthuleans produce poly-unsaturated fatty acids, especially docosahexaenoic acid (DHA) [15,[76][77][78][79][80][81]. Marine fishes contain both DHA and eicosapentaenoic acid (EPA). These fatty acids are important for aquatic ecological processes, and fish in particular require these PUFAs as essential nutrients [82]. In aquaculture, labyrinthuleans have been used to enhance the PUFA content of rotifers and Artemia that serve as food for marine fish larvae [83][84][85]. The survival rate of fish larvae was shown to increase when PUFA-enriched rotifers and Artemia were used as food [86][87][88]. As diatoms accumulate more EPA than DHA, it is easy to predict that diatoms are involved in the accumulation of EPA in fish [89][90]. The present study showed that Aplanochytrium spp. are candidates as sources of DHA for zooplankton and fish. Dinoflagellates, haptophytes, cryptophytes, and bacteria are also candidates as sources of DHA [89,91,92].
As mentioned above, the consumption of living diatoms by Aplanochytrium cells in twomembered culture experiments indicates the possibility there is a newly recognized pathway in the grazing food chain in the marine ecosystem. Aplanochytrium spp. are common in the marine environment, from coastal waters to outer oceans, from surface water to deep sea water, and even on phytoplankton in marine snow. However, we have investigated only two diatoms as target organisms nutritionally ingested by Aplanochytrium cells, so that it is essential to investigate whether aplanochytrids can ingest nutrients from other algae and protists. Also, it is necessary to clarify whether zooplankton can prey on clusters of Aplanochytrium cells and ingest nutrition from them by direct experiments. Furthermore, the ingestion frequency and production efficiency among algae, Aplanochytrium and zooplankton should be clarified by the future researches such as experiments using the mesocosms and detailed field studies. By accumulating such knowledge, the roles and impacts of Aplanochytrium in the marine ecosystem will be clarified.