Growth and Chloroplast Replacement of the Benthic Mixotrophic Ciliate Mesodinium coatsi

Abstract While the ecophysiology of planktonic Mesodinium rubrum species complex has been relatively well studied, very little is known about that of benthic Mesodinium species. In this study, we examined the growth response of the benthic ciliate Mesodinium coatsi to different cryptophyte prey using an established culture of this species. M. coatsi was able to ingest all of the offered cryptophyte prey types, but not all cryptophytes supported its positive, sustained growth. While M. coatsi achieved sustained growth on all of the phycocyanin‐containing Chroomonas spp. it was offered, it showed different growth responses to the phycoerythrin‐containing cryptophytes Rhodomonas spp., Storeatula sp., and Teleaulax amphioxeia. M. coatsi was able to easily replace previously ingested prey chloroplasts with newly ingested ones within 4 d, irrespective of prey type, if cryptophyte prey were available. Once retained, the ingested prey chloroplasts seemed to be photosynthetically active. When fed, M. coatsi was capable of heterotrophic growth in darkness, but its growth was enhanced significantly in the light (14:10 h light:dark cycle), suggesting that photosynthesis by ingested prey chloroplast leads to a significant increase in the growth of M. coatsi. Our results expand the knowledge of autecology and ecophysiology of the benthic M. coatsi.

THE ciliate Mesodinium species inhabit a broad range of aquatic environments, from freshwater to marine ones, and from water column (planktonic) to sandy sediments (benthic), throughout the world. Since the ciliate genus Mesodinium was established by Stein in 1863, a total of 10 species have been described to date, including M. acarus, M. fimbriatum, M. chamaeleon, M. cinctum, M. coatsi, M. major, M. pulex, M. pupula, M. velox, and M. rubrum; of these, the first two species have been observed in freshwater and the others in marine or estuarine environments (Borror 1972;Dragesco 1963;Foissner et al. 1999;Garcia-Cuetos et al. 2012;Kahl 1935;Lohmann 1908;Moestrup et al. 2012;Nam et al. 2015;Tamar 1992;Taylor et al. 1971). The taxonomy of Mesodinium species has been studied based mainly on morphological characteristics, such as the shape and size of the cell, unique structures of the tentacle, and number of cirri and kinetids (Borror 1972;Dragesco 1963;Kahl 1935;Lohmann 1908;Tamar 1992;Taylor et al. 1971), that have caused the controversy of the genus alternation to Mesodinium or Myrionecta, as well as ambiguous classification between the species. Phylogenetic studies conducted along with careful ultrastructural observations, however, have clarified the phylogenetic positions of at least six Mesodinium species, including M. pulex, M. pupula, M. rubrum, M. major, M. chamaeleon, and M. coatsi, the latter three of which have only recently been added to the genus Moestrup et al. 2012;Nam et al. 2015).
Among Mesodinium species, the ecophysiology of the planktonic species M. rubrum has been the best studied for a long time. It is well known as a species that causes massive, nontoxic blooms in coastal and estuarine areas worldwide (Crawford 1989;Lindholm 1985;Taylor et al. 1971). Furthermore, early ultrastructural studies of M. rubrum isolated from field samples reported that the cells of this ciliate contained chloroplasts, mitochondria, and nucleomorphs originating from its cryptophyte prey (Gustafson et al. 2000), which were previously thought to represent incomplete endosymbionts (Hibberd 1977;Oakley and Taylor 1978;Taylor et al. 1969Taylor et al. , 1971. Since cultures of M. rubrum have been established (Gustafson et al. 2000;Yih et al. 2004), however, it has been revealed that M. rubrum acquires functional chloroplasts, nuclei, and other cellular organelles from its cryptophyte prey, which mainly belong to the genera Geminigera and Teleaulax, in order to maintain stable photosynthetic performance and growth Johnson and Stoecker 2005;Johnson et al. 2006Johnson et al. , 2007Kim et al. 2016Kim et al. , 2017. For this reason, Mesodinium species has come into the spotlight as one of the model organisms from which clues to the evolutionary history of plastid acquisition may be drawn. However, very little is known about such ecophysiological features as trophic level, feeding behavior and process, and chloroplast retention and function of Mesodinium species other than M. rubrum. Only M. pulex and M. chamaeleon, which are both benthic species, have been studied from the establishment of culture, and growth and grazing responses, and chloroplast dynamics of the latter species has recently been reported (Jakobsen et al. 2006;Johnson et al. 2004;Moeller and Johnson 2017;Moestrup et al. 2012;Tarangkoon and Hansen 2011). However, the amount of ecophysiological information available for other benthic Mesodinium species still remains insufficient.
We have been maintaining a culture of the benthic species Mesodinium coatsi by providing it with benthic cryptophyte Chroomonas species as prey. Using these cultures of M. coatsi and its cryptophyte prey, in this study we investigated: (1) the effects of different cryptophyte prey on the growth of M. coatsi; (2) the replacement of ingested prey chloroplasts within M. coatsi; and (3) the growth responses of fed M. coatsi in both light/ dark conditions and total darkness. Through this work, we provide new insights into the basic autecology and ecophysiology of the benthic species M. coatsi.

Culture of the benthic ciliate Mesodinium coatsi
A culture of M. coatsi (MC01-LOHABE) was established by isolating single cells from samples collected from the sandy sediments of Mohang beach (35°34 0 N, 126°30 0 E), Korea, on October 17, 2011 (Nam et al. 2015). Briefly, surface sand samples were collected with a spoon during low tide, and then were transported directly to the laboratory. Individual ciliate cells were isolated using a drawn glass pipette, washed eight times in sterile seawater, and transferred to a 24-well plate (SPL Lifesciences, Gyeonggido, Korea) containing 1 ml of 30 psu f/2-Si medium (+ 5% v/v soil extract) at 20°C under a 14:10 h light:dark cycle with cool-white fluorescent light at 80 lmol photons/m 2 /s. Light intensity was measured with a photometer (Biospherical Instruments, Model QSL-2101, San Diego, CA). M. coatsi culture has since been maintained in three stock cultures by separately adding the marine benthic cryptophytes Chroomonas sp. 07 (gCR07-LOHABE), Chroomonas sp. 09 (gCR09-LOHABE), and Chroomonas sp. 12 (gCR12-LOHABE) as prey over the past 7 years. Prey were supplied and replenished whenever they were depleted in the cultures. Once M. coatsi cell density increased, the cells were transferred to a 125 ml KIMAX â glass culture flask (KIMBLE, Vineland, NJ) and were kept well-fed by adding sufficient amounts of prey. All of the cultures used in this study were non-axenic.

Growth responses to different cryptophytes
These experiments were performed to investigate the growth responses of M. coatsi to provided different cryptophyte prey. M. coatsi cells originally grown on Chroomonas sp. 09 and then starved for 2 d were used for the following two experiments (1 and 2). Prior to the experiment, the absence of prey in the culture was confirmed under an inverted microscope at 100X magnification (Olympus IX51).
Experiment 1: Chroomonas spp. and Teleaulax amphioxeia as prey Three benthic cryptophytes, Chroomonas spp. 07, 09, and 12, as well as the planktonic cryptophyte T. amphioxeia were used as prey for M. coatsi. Experiment 1 was performed on well plates to allow for the precise enumeration of Chroomonas spp. cells due to the strong adhesion of these prey types to the substrate. The initial predator:prey ratio was adjusted to achieve a ratio of 1:10, with concentrations of M. coatsi and each prey cell being about 20 and 200 cells/ml, respectively. Mixtures of prey and predator (1 ml) were distributed among three wells, and either prey-only or predator-only controls were also respectively established in three additional wells (in triplicate) of a 48-well plate. Each plate in all treatments and controls was fixed by adding 20 lL of Lugol's solution (final concentration 2%) into each well at 2-to 4-d intervals for 14 d. Whole M. coatsi cells in the well were counted directly under an inverted microscope (Olympus IX51), whereas abundances of different cryptophyte prey were enumerated using a Palmer-Maloney chamber. Unfortunately, unlike other cryptophyte prey, the enumeration of Chroomonas sp.
12 cells could not be done at each sampling time because of their aggregative characteristics.
Experiment 2: Rhodomonas spp. and Storeatula sp. as prey Five benthic cryptophytes, Rhodomonas spp. 01, 02, 03, and 04, and Storeatula sp. (CCMP1868), were used as prey for M. coatsi. The initial predator:prey ratio was adjusted to achieve ratios from 1:2 to 1:5, depending on the type of cryptophyte prey used. Mixtures of prey and predator were distributed among three culture flasks, and either prey-only or predator-only controls were also respectively established in three additional culture flasks. Aliquots (2 ml) withdrawn from each flask at 1-to 2-d intervals over 15 d were fixed with acid Lugol's solution (final concentration 2%). All experimental cultures (from experiments 1 and 2) were placed on a shelf under the same culture condition described above. The abundances of M. coatsi and different cryptophytes were enumerated using a Sedgewick-Rafter chamber under the inverted microscope (Olympus IX51) at 100X and 200X. Growth rates of M. coatsi (l) provided with different cryptophyte prey were calculated using the exponential growth equation, l = ln(N 2 /N 1 )/(t 2 À t 1 ), where N 2 and N 1 are cell concentrations at time t 2 and time t 1 , respectively, and t 2 À t 1 is the time interval between samplings. The time interval (t 2 À t 1 ) was selected only for the exponential phase, mostly 2 d after the start of experiment 1 and the day after abundance of non-fed M. coatsi (control) began to decrease in experiment 2. The ingestion rates were calculated according to the method of Jeong and Latz (1994).

Cross-feeding experiment: chloroplasts turnover
The aim of this experiment was to examine how fast the retained prey chloroplasts are replaced with new ones from recently ingested prey. To track chloroplast replacement easily and clearly, two cryptophytes with differently colored chloroplasts (the green chloroplasts of Chroomonas sp. 09 and reddish-brown chloroplasts of Rhodomonas sp. 03 on which M. coatsi grew best in the prior experiments 1 and 2) were used as chloroplast donors. M. coatsi cells initially grown on Chroomonas sp. 09 and then starved for 2 d were allowed to feed on Rhodomonas sp. 03 for 9 d in a culture flask (Corning Inc., Corning, NY, USA) at a predator:prey ratio of approximately 1:3.

Light microscopy
Light micrographs of live M. coatsi and cryptophyte prey were taken at 1,000X magnification using a photomicrographic system (AxioCam HRC, Carl Zeiss Inc.) coupled to a bright-field microscope equipped with differential interference contrast optics (Axio imager A2; Carl Zeiss Inc.).

DNA extraction, PCR amplification, and sequencing
Aliquots (3 ml) withdrawn from each cryptophyte culture were subjected to centrifugation for 5 min at 7168 g. The DNA was then extracted and purified using the Genomic DNA Extraction Kit (Bioneer, Daejeon, Korea). Gene fragments of cryptophyte nuclear SSU rRNA were amplified by polymerase chain reaction (PCR) using the common primer set (EukA/EukB) or a newly designed primer set (GCRSSU-F/GCRSSU-R) ( Table 2). PCR conditions were as follows: 94°C for 5 min followed by 40 cycles of 94°C for 45 s, 55°C for 30 s, and 72°C for 80 s followed by 72°C for 10 min. A semi-nested PCR was also conducted for the detection of gene fragments of Rhodomonas and Storeatula species using a second pair of primers, 18SNF2 and EukB. For this semi-nested PCR, the first PCR product was initially amplified using the universal primer set for 20 cycles. Subsequently, 3 ll of the first PCR product was then used as a template and run through the amplification again, as described above. The PCR products were visualized through EcoDye TM (SolGent, Daejeon, Korea) stained 1% agarose gel electrophoresis and purified using a PCR purification kit (Bioneer). The amplified DNA was sequenced using an ABI3730xl DNA Analyzer at the

Phylogenetic analyses
The alignment of each cryptophyte species' nuclear gene sequences was constructed using the Clustal X algorithm and refined by eye using the Genetic Data Environment (GDE 2.4) program (Smith et al. 1994). A maximum likelihood (ML) tree with 2,000 bootstrap replicates was inferred for each alignment using RAxML version 8 (Stamatakis 2014). Prior to Bayesian analysis, we performed a likelihood ratio test using Modeltest, version 3.7 (Posada and Crandall, 1998) to determine the best model for the combined dataset. Bayesian analysis was run using MrBayes 3.2.5 (Ronquist et al. 2012), with four Metropolis-coupled Markov chain Monte Carlo (MCMC) processes run for 20,000,000 generations, keeping one tree every 1,000 generations. The data were analyzed with a single GTR + I + G model and the following parameters were specified: Prset revmatpr = dirichlet (1.3820, 3.4431, 1.1997, 1.9270, 6.5590, 1.0); statefreqpr = dirichlet (0.2779, 0.1934, 0.2571, 0.2717); shapepr = exponential (0.6718); and pinvarpr = fixed (0.5991) for the nuclear SSU rRNA genes. The first 8000 trees were discarded as burn-in. Trees were visualized using the Figtree v.1.4.2 program.

Phylogenetic positions of the cryptophytes used in this study
Cryptophytes, including the nine experimental strains used in this study, were divided into seven distinct clades in the phylogenetic tree derived from the analysis of their nuclear SSU rRNA genes (Fig. 2), including two clades comprising the monospecific genera Proteomonas (clade 1) and Falcomonas (clade 3). Rhodomonas spp. 01, 02, 03, and 04, and Storeatula sp., were included in clade 2, and Chroomonas spp. 07, 09, and 12 were placed in clade 4. The planktonic T. amphioxeia was included in clade 5.
In the nuclear SSU rRNA-based phylogeny, all cryptophyte prey used in this study were placed in different phylogenetic positions, except for Rhodomonas spp. 02 and 03, which were in the same position.

Chloroplast replacement
When sufficient numbers of Rhodomonas sp. 03 were supplied as prey (Fig. 6B), the chloroplasts of most M. coatsi cells (98%) completely switched from the old, green chloroplasts of Chroomonas sp. 09 to the new, reddish-brown chloroplasts of Rhodomonas sp. 03 within 4 d (Fig. 6A). Some M. coatsi cells (11%) were even able to entirely exchange all of their old chloroplasts with new ones from recently ingested prey within a single day. After 6 d, the chloroplasts from the previous prey (Chroomonas sp. 09) were completely replaced with those of new prey (Rhodomonas sp. 03). The chloroplast turnover time was similar in the reverse situation. When M. coatsi cells starved for 2 d while retaining chloroplasts of Rhodomonas sp. 03 were exposed to Chroomonas sp. 09 again, the chloroplasts of most M. coatsi cells (96%) were again replaced with those of Chroomonas sp. 09 within 4 d (Fig. 6A).

Figure 2
The phylogenetic tree of cryptophytes inferred from Bayesian analysis and RAxML based on nuclear SSU rDNA sequences in this study. The sequences of the nine cryptophytes determined in this study are indicated by black boxes. At internodes, tree support values are represented in terms of both posterior probabilities and bootstrap values, respectively. Values of <0.5 or 50% are shown as "À". Mismatches of the branch between the Bayesian and RAxML tree are indicated by a "*".

Mixotrophic growth of Mesodinium coatsi in both light/dark and dark conditions
Mesodinium coatsi abundance greatly increased after feeding on prey (Chroomonas sp. 09 or Rhodomonas sp. 03), but its growth responses differed markedly as a function of light (Fig. 7). In the light/dark condition, the growth of M. coatsi in the presence of either Chroomonas sp. 09 or Rhodomonas sp. 03 was considerably enhanced relative to that in total darkness (one-way ANOVA, Tukey's HSD test, P < 0.005), with growth rates in light/dark conditions being 0.54 (AE0.03) d À1 and 0.60 (AE0.02) d À1 on each respective prey type, compared to 0.15 (AE0.02) d À1 and 0.22 (AE0.01) d À1 in equivalent conditions under darkness. In darkness, M. coatsi in the presence of prey (whether Chroomonas sp.09 or Rhodomonas sp. 03) showed positive growth, whereas in the absence of prey it showed no or negative growth.

Prey specificity of Mesodinium coatsi
The ciliate Mesodinium species are phagotrophs that rely on prey ingestion for growth (Gustafson et al. 2000;Tarangkoon and Hansen 2011;Yih et al. 2004). The prey  specificity of some Mesodinium species has been explored in previous studies. The mixotrophic species M. rubrum is known to feed on cryptophytes belonging to the Teleaulax/Plagioselmis/Geminigera (TPG) clade (Gustafson et al. 2000;Hansen and Fenchel 2006;Johnson and Stoecker 2005;Yih et al. 2004), but it prefers to feed on a certain cryptophyte species so that it can undergo sustained growth. The temperate M. rubrum showed an extraordinarily high growth rate when fed T. amphioxeialike species compared to when it was fed other TPG cryptophytes Johnson et al. 2016;Myung et al. 2011;Park et al. 2007;Raho et al. 2014), which is consistent with field observations of this species. Most natural M. rubrum cells have been reported to predominately retain T. amphioxeia-like plastids (Herfort et al. 2011;Johnson et al. 2016;Nishitani et al. 2010). While M. rubrum has a relatively narrow prey range, the benthic species M. pulex feeds on a wide range of prey organisms, including the cryptophytes Guilardia theta, Rhodomonas sp., Teleaulax sp., and T. amphioxeia, as well as the dinoflagellate Heterocapsa rotundata, but its growth differs depending on what prey organisms it ingests (Johnson et al. 2004;Tarangkoon and Hansen 2011). The mixotroph M. chamaeleon has also been reported to ingest prey belonging to at least five different genera of cryptophytes (i.e. Chroomonas mesostigmatica, Guillardia theta, Hemiselmis cryptochromatica, Storeatula major, and Teleaulax amphioxeia), but it showed a distinct prey preference for Storeatula major (Moeller and Johnson 2017;Moestrup et al. 2012), which is smaller than the Storeatula sp. CCMP1868 strain used in our study.
Our study showed that M. coatsi can, although temporarily, feed on a relatively broad array of cryptophyte prey, similar to M. chamaeleon. M. coatsi was also able to ingest phylogenetically distinct prey belonging to four different cryptophyte genera (Chroomonas, Rhodomonas, Storeatula, and Teleaulax), but not all of these ingestions have supported its sustained growth and thus its growth responses varied as a function of the prey species it Figure 5 Growth rates of the benthic ciliate Mesodinium coatsi when fed different cryptophytes (A) Chroomonas spp. and Teleaulax amphioxeia, (B) Rhodomonas spp. and Storeatula sp. Growth rates were calculated from the experiments in Fig. 3 and 4. Bars are means AE SE for triplicate cultures. Note that while all offered Chroomonas spp. supported considerable growth of M. coatsi, only Rhodomonas sp. 03 supported the significant positive growth of M. coatsi. The other Rhodomonas spp. and Storeatula sp. allowed for relatively lower growth than the four-species mentioned above. Negative growth was found when the ciliate was exposed to T. amphioxeia and Rhodomonas sp. 02. Lower case letters on A and B indicate significant differences among treatments at the P < 0.05 level (one-way ANOVA, Tukey's HSD test). An erroneous negative growth rate, which was obtained when the ciliate was fed Rhodomonas sp. 02, was excluded in the statistical test shown in B. consumed. These results may be due to several factors. It seems unlikely that prey size would affect the observed differences in growth responses because M. coatsi is capable of ingesting prey of various cell sizes, ranging from 7.4 (Chroomonas sp. 09) to 19.1 lm (Storeatula sp. CCMP1868), using its retractable-extensible tentacles. Rather, it is likely that the differential growth responses were related to differences in habitat between the predator and prey organisms. For example, M. coatsi did not grow when offered T. amphioxeia, and eventually died out. T. amphioxeia is planktonic and prevalent throughout the water column (Herfort et al. 2011;Johnson et al. 2016;Peterson et al. 2013), whereas M. coatsi mainly inhabits benthic sedimentary environments. Such a spatial separation may have caused some of the difficulty for M. coatsi in exploiting the planktonic T. amphioxeia due to low natural encounter rates between them, which may in turn, if any, have resulted in the ingestion of T. amphioxeia by M. coati not at rates that have any effect on the population dynamics of planktonic cryptophyte prey, as indicated from the near-zero level of the ingestion rate. Third, chloroplast type of its cryptophyte prey may also have a significant effect on the growth of M. coatsi. It is known that cryptophytes with green or blue-green chloroplasts (e.g. Chroomonas, Hemiselmis, and Komma) possess phycocyanin pigments, while those with red or reddish-brown chloroplasts (e.g. Geminigera, Guillardia, Hanusia, Plagioselmis, Rhinomonas, Rhodomonas, Storeatula, and Teleaulax) have phycoerythrin pigments (Gantt 1979;Gantt et al. 1971;Hoef-Emden 2008). In our experiment, M. coatsi achieved sustained growth on all phycocyanin-containing Chroomonas spp., whereas the growth responses of M. coatsi fed Rhodomonas spp. and Storeatula sp., which have chloroplasts with phycoerythrin pigments, were distinct and variable. While Rhodomonas sp. 03 supported a growth rate of M. coatsi as high as that with Chroomonas spp., Rhodomonas spp. 01 and 04, and Storeatula sp. (CCMP1868), did not support sustained growth in the long run, although they did temporarily support the growth of M. coatsi. Given that Rhodomonas spp. 02 and 03 occupy the same phylogenetic position and have similar cell sizes, it is noteworthy that M. coatsi showed opposite growth responses to these two species, although the reason for this currently remains unknown. Garcia-Cuetos et al. (2012) and Moeller and Johnson (2017) observed that most benthic Mesodinium species frequently retain green chloroplasts in nature, indicating that such species, including M. chamaeleon and M. coatsi, may prefer to feed on phycocyanin-containing cryptophytes over phycoerythrin-containing cryptophytes in their benthic environments. Nonetheless, we have occasionally observed M. chamaeleon or M. coatsilike species containing both green and reddish-brown chloroplasts in field samples (M. Kim, pers. observ.). Along with our field observations, the remarkable growth of M. coatsi on Rhodomonas sp. 03 in our laboratory experiment suggests that M. coatsi can also exploit some phycoerythrin-containing members of the Rhinomonas/ Rhodomonas/Storeatula clade, depending on the composition of the benthic cryptophyte prey community.

Chloroplast replacement
We confirmed that Mesodinium coatsi can replace the chloroplasts of its previous prey with those from newly ingested prey. In comparison to M. rubrum, in which full chloroplast turnover took 2-5 wk depending on the concentration of prey offered Peltomaa and Johnson 2017), we observed that in most M. coatsi old chloroplasts were replaced by new ones within 4 d when available prey was present in sufficient quantities. Furthermore, no differences in the transition time of chloroplasts were observed when the ciliate switched from Chroomonas sp. 09 to Rhodomonas sp. 03, or vice versa, both of which have chloroplasts containing different types of pigment (i.e. phycocyanin and phycoerythrin, respectively). This result may imply that M. coatsi does not preferentially retain certain chloroplasts for supporting phototrophic growth when both prey types are available. M. chamaeleon was also shown to retain different types of chloroplasts by cross-feeding on different types of cryptophyte prey (Moeller and Johnson 2017); M. chamaeleon replaced~50% of its chloroplasts from C. mesostigmatica with those of Storeatula major within 2 d, and replaced nearly 100% of these within 8 d.

Function of retained chloroplasts
When provided cryptophyte prey, Mesodinium coatsi was able to undergo heterotrophic growth in darkness, but its growth was significantly enhanced in the presence of light (using a 14:10 h light:dark cycle), i.e. its mixotrophic growth could be supported by either only feeding on prey or by feeding coupled with light (photosynthesis). In particular, light had a significant effect on the positive biomass increase of M. coatsi (i.e. this was 2.3 times higher in light/dark conditions than in darkness), suggesting that the chloroplasts of its ingested prey are photosynthetically active until being digested. The observed highest growth rate of well-fed M. coatsi was 0.55 d À1 in light, which was similar to the phototrophic growth of M. rubrum, which is able to maintain stable photosynthesis for over a month (Hansen and Fenchel 2006;Johnson and Stoecker 2005;Kim et al. 2017;Peltomaa and Johnson 2017). Nonetheless, we also cannot rule out alternate explanations for the effects of light on growth in M. coatsi. For example, increasing light irradiance often results in the either stimulation of feeding or acceleration of digestion, both of which produce a positive increase in the biomass of protists (Li et al. 1999;Strom 2001;Tarangkoon and Hansen 2011).
In M. coatsi, the photosynthetic capacity and stability of the retained prey chloroplasts seems to be somewhat lower than that in M. rubrum because we observed that the photosynthetic growth of M. coatsi decreased rapidly as soon as prey numbers were depleted. When prey were absent, starved M. rubrum was observed to be capable of dividing three to four times and surviving for up to about 4 mo (Johnson and Stoecker 2005;Johnson et al. 2007;Kim et al. 2016Kim et al. , 2017Nam et al. 2015), whereas in the present study starved M. coatsi gradually lost prey chloroplasts more and more with increasing starvation time and survived for only about 2 wks. The difference in photosynthetic ability and survival time between the two species may be associated with differences in which prey organelles are sequestered and how they are packaged within the ciliates. M. rubrum has been reported to sequester prey chloroplasts and prey nucleus, which remain transcriptionally active in photosynthetic functions (Johnson et al. 2007;Kim et al. 2016Kim et al. , 2017Lasek-Nesselquist et al. 2015), whereas M. chamaeleon and M. coatsi have been observed to harbor whole cryptophyte cells rather than to sequester certain organelles, and then digest these in food vacuoles within a short time period Nam et al. 2015). Such a difference in the degree of retention and acquisition of prey organelles may lead the lack of ability of M. coatsi to control the prey chloroplasts, unlike M. rubrum. Therefore, it appears that M. coatsi have to uptake new prey chloroplasts constantly for their continued survival, like M. chamaeleon (Moeller and Johnson 2017;Moestrup et al. 2012).

CONCLUSIONS
Our study demonstrated that Mesodinium coatsi can use both phototrophic and heterotrophic nutritional strategies. Such a combination of two nutritional modes may allow this ciliate to survive more efficiently in a changing environment. In terms of the categories of mixotrophy recently redefined by Mitra et al. (2016) based on the physiological function of protists, M. coatsi appears to fall within the "Generalist Non-Constitutive Mixotroph (GNCM)" functional group. M. coatsi was able to feed on various types of cryptophytes prey and exploit retained prey chloroplasts, but not for all prey organisms. However, the ability of M. coatsi to use and control stolen plastids seemed to be poor. Therefore, M. coatsi appears to continuously seek new chloroplast donors to support its sustained growth. In addition, its photosynthetic ability was similar to that of M. chamaeleon, but lower than that of M. rubrum, which implies that M. coatsi is evolutionarily intermediate between the heterotrophic and phototrophic Mesodinium species in terms of its acquired phototrophy, similar to M. chamaeleon (Moeller and Johnson 2017). This study provides new information to support a better understanding of chloroplast function and status in mixotrophic benthic Mesodinium species.