Mesodinium rubrum Lohmann 1908 is a cosmopolitan species that recurrently forms ciliate red tides in diverse marine environments (Taylor et al. 1971, Lindholm 1985, Crawford 1989, Yih et al. 2013). M. rubrum is able to carry out photosynthesis as well as phagotrophic feeding on prey organisms such as cryptophytes (Yih et al. 2004a, Johnson and Stoecker 2005, Park et al. 2007, Hansen et al. 2012, 2013) and heterotrophic bacteria (Myung et al. 2006). In turn, M. rubrum is known to be an important prey item for many protistan and metazoan grazers at higher trophic level (Sullivan and Gifford 2004, Yih et al. 2004b, Liu et al. 2005, Park et al. 2006, Reguera et al. 2012, Lee et al. 2014). This species usually co-occurs with bacterioplankton (Powell et al. 2005, Jeong et al. 2013) and M. rubrum blooms sometimes succeed those of bacterioplankton (Jeong et al. 2013). Therefore, mixotrophy of M. rubrum is most likely a very important phenomenon for the balanced maintenance of healthy marine environments. To understand the ecology of M. rubrum in marine food web systems, further exploration on the unique aspects of phagotrophy in M. rubrum on several kinds of prey is still desired.

The phototrophic prokaryote Synechococcus is a ubiquitous cyanobacterium in marine ecosystem (Johnson and Sieburth 1979, Waterbury et al. 1979, Marañón et al. 2003, Huang et al. 2012) with its cosmopolitan distribution from tropical to polar waters (Walker and Marchant 1989, Burkill et al. 1993, Landry et al. 1996, Powell et al. 2005). Synechococcus spp. numerically dominate the abundance of phytoplankton in marine environments (Glibert et al. 2004, Murrell and Lores 2004). In addition, Synechococcus sometimes contributes significantly to phytoplankton biomass and primary production in marine ecosystem (Glover et al. 1986, Li 1994, Jeong et al. 2013). It is known to be one of the major contributors to CO₂ and nutrient uptake from the ocean waters (Marañón et al. 2003). Therefore, the growth and mortality of Synechococcus are important factors in understanding the cycling of biomaterials in marine microbial food webs.

Several protistan grazers are known to ingest Synechococcus (Christaki et al. 1999, 2002, Jeong et al. 2005, 2010, 2012, Apple et al. 2011, Strom et al. 2012). The marine ciliate M. rubrum has been found to co-occur with Synechococcus spp. in the coastal waters (Lignell et al. 2003, Jeong et al. 2013, Liu et al. 2013). Therefore to better understand Mesodinium bacterivory in microbial food webs, we investigated the predator-prey relationships between M. rubrum and Synechococcus.

We explored whether M. rubrum is able to feed on Synechococcus. We also measured the ingestion rates of M. rubrum on Synechococcus as a function of prey concentration. In addition, we estimated grazing coefficients attributable to M. rubrum on co-occurring Synechococcus using our data for ingestion rates obtained from the laboratory experiments and data on the abundance of Mesodinium and Synechococcus in the field. The results of the present study provide a basis for improved estimation and understanding the population dynamics of M. rubrum in marine ecosystems.

For isolation and cultivation of Mesodinium rubrum strain MR-MAL01 (Table 1) plankton samples were collected from Gomso Bay, Korea, during May 2001 when the water temperature and salinity were 18.0℃ and 31.5, respectively. A culture of M. rubrum was established by serial single-cell isolations (Yih et al. 2004a). The cryptophyte Teleaulax amphioxeia strain CR-MAL01 (Yih et al. 2004a) was offered as prey of M. rubrum. Both M. rubrum and T. amphioxeia were maintained at 20℃ in f/2 medium (Guillard and Ryther 1962) without silicate under continuous illumination of 20 μmol photons m^-2 s^-1 of cool white fluorescent light in the walk-in incubator system of the Marine Biology Research and Education Center, Kunsan National University. The phototrophic prokaryote Synechococcus strain CC9311 (clade I) (Table 1) was also grown at 20℃ in f/2 medium (Guillard and Ryther 1962) without silicate under continuous illumination of 20 μmol photons m^-2 s^-1. This strain has two phycoerythrin proteins (PE I and PE II) (Ong and Glazer 1991).

The equivalent spherical diameter and cell volume of M. rubrum (Table 1) was measured using an electron particle counter (Coulter Multisizer II; Coulter Corporation, Miami, FL, USA). The carbon contents for M. rubrum was estimated from cell volume according to Menden-Deuer and Lessard (2000). The cell volume and carbon content for Synechococcus sp. was adopted from Apple et al. (2011).

Experiment was designed to measure the ingestion and clearance rates of M. rubrum as a function of the PC when fed on Synechococcus sp.

We prepared dense cultures of M. rubrum (12,000 cells mL^-1) and Synechococcus sp. that were separately grown phototrophically in f/2 medium (Guillard and Ryther 1962) without silicate under continuous illumination of 20 μmol photons m^-2 s^-1. Three 1 mL aliquots were subsampled from each M. rubrum culture for the cell counting under a light microscope (Olympus BH2; Olympus Co., Tokyo, Japan). For the Synechococcus cell counting, 5 mL aliquots from each Synechococcus culture were removed and then fixed with formalin (final conc. = 4%). The fixed sample was stained using DAPI (final conc. = 1 μM) and then filtered onto 25-mm polycarbonate black membrane filters of 0.2 μm-pore-size. The Synechococcus cells on the membranes were observed under an epifluorescence microscope (Olympus BH2; Olympus Co.) with UV light excitation at a magnification of ×1,000.

The initial concentrations of M. rubrum and Synechococcus were established using a pipette to deliver predetermined volumes of known cell concentration to the bottles. Triplicate 80 mL experimental bottles (containing mixtures of M. rubrum and Synechococcus), triplicate prey control bottles (containing Synechococcus only) and triplicate predator control bottles (containing M. rubrum only) were also established. All the bottles were placed on a shelf and incubated at 20℃ under illumination of 20 μmol photons m^-2 s^-1 of cool white fluorescent light.

After 1-, 10-, 20-, and 30-min incubation periods, 5 mL aliquots were removed from each bottle, and then fixed with formalin. The fixed samples were stained using DAPI and then filtered onto 3 μm-pore-sized polycarbonate white membrane filters. Then, the cells of M. rubrum with Synechococcus as well as Synechococcus inside a M. rubrum were enumerated under an epifluorescence microscope with UV, blue, and green-light excitation at a magnification of ×1,000 by scanning the M. rubrum body at consecutive intervals of 1 to 2 μm focal depth along the z-axis. We tried to minimize the concentration of heterotrophic bacteria in the M. rubrum culture. For the experiments we subsampled M. rubrum from the upper thin layer with high density of M. rubrum using a siphon, and then diluted to the target concentrations by adding auto claved seawater to the subsamples. Thus, the initial concentrations of heterotrophic bacteria in the experimental bottles were <16% of Synechococcus concentrations.

The ingestion rate (IR; cells predator^-1 h^-1) was calculated by linear regression of the number of Synechococcus per M. rubrum cell as a function of incubation time as in Sherr et al. (1987).

The clearance rate (CR; mL predator^-1 h^-1) was calculated as:

, where IR (cells predator^-1 h^-1) is the ingestion rate of the M. rubrum predator on the Synechococcus prey and PC (cells mL^-1) is the prey concentration.

Ingestion and clearance rates were calculated using the equations of Frost (1972) and Heinbokel (1978). Data for IRs (cells predator^-1 h^-1) were fitted to a Michaelis-Menten equation:

, where I_max = the maximum ingestion rate (MIR; cells predator^-1 h^-1), x = PC (cells mL^-1), and K_IR = the PC sustaining 1/2 I_max.

We estimated the grazing coefficients attributable to Mesodinium rubrum on Synechococcus spp. by combining field data on the abundance of M. rubrum and Synechococcus with the ingestion rates of the M. rubrum on Synechococcus sp. obtained in the present study. Field data on the abundance of M. rubrum and co-occurring Synechococcus used in this estimation were originally obtained using the water samples from Masan Bay (2004-2005), Korea (Jeong et al. 2013).

The grazing coefficient (g day^-1) was calculated as:

, where CR (mL predator^-1 h^-1) is the clearance rate of a M. rubrum predator on Synechococcus prey at a given PC and GC is a predator concentration (cells mL^-1). CR values were corrected using Q₁₀ = 2.8 (Hansen et al. 1997) because in situ water temperatures and the temperature used in the laboratory for this experiment (20℃) were sometimes different.

We found that M. rubrum ingested Synechococcus cells (Fig. 1). This should be the first report of grazing by red tide ciliate M. rubrum on Synechococcus. M. rubrum commonly co-occurred with Synechococcus in many marine ecosystems (Lignell et al. 2003, Jeong et al. 2013, Liu et al. 2013). Thus, M. rubrum should be considered to be a potentially grazer on Synechococcus in marine planktonic food webs.

The specific ingestion rates (SIRs) of M. rubrum on Synechococcus linearly increased with increasing PCs up to 1.9 × 10⁶ cells mL^-1, but exhibit sigmoidal saturation at higher concentrations (Fig. 2). When the data were fitted to Eq. (2), the MIR of M. rubrum was 2.1 cells predator^-1 h^-1 (0.5 pg C predator^-1 h^-1). In addition, M. rubrum was able to acquire up to 12.6 pg C from Synechococcus daily. The maximum clearance rate of M. rubrum on Synechococcus was 4.2 nL predator^-1 h^-1 (Table 2).

In comparison with MIRs of other red tide organisms when fed on Synechococcus, the MIR of M. rubrum on Synechococcus is quite much lower (Table 3). The maximum of volume SIR (VSIR) of M. rubrum was also lower than that of the other predators. M. rubrum is also able to feed on heterotrophic bacteria with higher SIR than that for Synechococcus (Myung et al. 2006). In addition, M. rubrum fed on exclusively cryptophyte prey species when offered a variety of algal prey species (Park et al. 2007, Myung et al. 2013). Thus, such kind of multiple prey species for the phagotrophy of M. rubrum might have evoked a type of partitioned ingestion with differential prey preferences.

The ingestion rates of the red tide organisms on Synechococcus were affected by the PC. The K_IR (the mean PC sustaining 1/2 I_max of MIR) of 1.2 × 10⁶ cells mL^-1 for M. rubrum feeding on Synechococcus was relatively higher than that for other predators (Table 3). Therefore, the ingestion of M. rubrum on Synechococcus was less sensitive than that of other red tide organisms to the concentration of prey cells under prey-limited conditions.

The MIRs of heterotrophic nanoflagellates, mixotrophic dinoflagellates, heterotrophic dinoflagellates, and ciliates feeding on Synechococcus sp. are in general positively correlated with the predator’s equivalent spherical diameter (p < 0.01, ANOVA) (Table 4, Fig. 3). This relationship suggests that the sizes of the protistan grazers may be an important factor affecting their ingestion rates on Synechococcus. However, the MIR of M. rubrum on Synechococcus is relatively quite lower than that of the other protistan grazers with the exceptions in Picophagus flagellates, Bodo saltans, and Gonimonas pacifica. Furthermore, M. rubrum exhibited jumping behavior (Fenchel and Hansen 2006). This jumping behavior of M. rubrum may be partially responsible for the minimal ingestion. The MIR of Ochromonas sp. feeding on Synechococcus was higher than that of the other protistan grazers with the exception Eutinnuis sp. (Table 4). In addition, the SIR of Ochromonas was higher than that of the other protistan grazers. Thus, Synechococcus was the optimal prey for Ochromonas sp. not for M. rubrum among bacteriovorus protistan grazers.

The MIR of M. rubrum on Synechococcus provided in the present study was higher than that of the small heterotrophic nanoflagellate Picophagus flagellates (0.18 pg C predator^-1 h^-1), Bodo saltans (0.12 pg C predator^-1 h^-1) on Synechococcus (Guillou et al. 2001). However, the MIR of M. rubrum on Synechococcus was lower than that of Pseudobodo sp. (0.68 cells predator^-1 h^-1) (Dolan and Šimek 1998, Christaki et al. 2002). Therefore, M. rubrum may sometimes compete with the heterotrophic nanoflagellates for the common prey Synechococcus if they co-occur.

M. rubrum was found to be able to feed on the cyanobacterium Synechococcus sp. as well as the cryptophyte Teleaulax amphioxeia and heterotrophic bacteria (Yih et al. 2004a, Myung et al. 2006). M. rubrum grew well when supplied with T. amphioxeia (Yih et al. 2004a). However, M. rubrum did not sustain growth when only heterotrophic bacteria or Synechococcus were offered as prey (personal observation, data not shown here). Therefore, Synechococcus may not make a critical contribution to the population growth of M. rubrum in natural environments, but become supplementary prey.

M. rubrum is known to require both food uptake and photosynthesis for sustainable growth. The prey cells are used as sources for energy, carbon, and nutrients. Accordingly, M. rubrum seems to be able to perform photosynthesis using kleptoplastids from a cryptomonad T. amphioxeia while taking up heterotrophic bacteria and Synechococcus as phosphorus and nitrogen source.

The grazing coefficients attributable to M. rubrum on co-occurring Synechococcus spp. in Masan Bay, Korea were affected mostly by the abundance of M. rubrum predator. The abundance of M. rubrum and Synechococcus spp. (n = 40) were 11-933 cells mL^-1 and 51-39,509 cells mL^-1, respectively. Grazing coefficients attributable to M. rubrum on co-occurring Synechococcus spp. were 0.001 to 0.036 day^-1 (Fig. 4).

To our knowledge, prior to this study, there had been no reports on the estimation of grazing impact by Mesodinium on co-occurring populations of Synechococcus. Grazing coefficients derived from studies in Masan Bay in 2004-2005 show that up to 3.6% of Synechococcus populations can be removed by Mesodinium in approximately 1 day. High mean abundance of Synechococcus (3,568 cell mL^-1) and relatively low mean abundance of Mesodinium (99 cell mL^-1) in coastal waters are responsible for the resulted relatively lower grazing coefficients. The results of the present study suggested that M. rubrum was not able to control the whole Synechococcus populations. However, the ingestion rates of protists are known to be affect by light and nutrition conditions (Jeong et al. 1999, Myung et al. 2006, Berge et al. 2008). Therefore, the lower grazing impact of Mesodinium on co-occurring Synechococcus may also be affected by light and nutrients conditions as well as the prey availability.

Complex mixotrophy in the marine ciliate Mesodinium rubrum. Unique status of kleptoplastidy in M. rubrum was shown by the highly organized chloroplast-mitochondria complexes from its cryptophyte prey (Johnson et al. 2007, Nam et al. 2012), long-time functioning of the kleptoplastids (Myung et al. 2013), and even the karyoklepty from the ingested cryptophyte prey cells (Johnson et al. 2007). Bacterivory in M. rubrum was noteworthy as an alternate source of essential microelements and cell carbon (Myung et al. 2006). Present study confirms that M. rubrum also feed on Synechococcus cells, one of the most abundant single cell phototrophs in the sea. In the euphotic zone of the oligotrophic open ocean Synechococcus spp. predominates the upper euphotic layers, and sometimes perform nitrogen fixation, to be fed into open ocean food web (Phlips et al. 1989, Walker and Marchant 1989). Equipped with quite unique and complex mixotrophic arrays of metabolism such as phagotrphy on hetrotrophic bacteria and Synechococcus as well as digestion, kleptoplastidy, and karyoklepty after the ingestion of cryptophyte prey, M. rubrum might be able to form recurrent and massive blooms in diverse marine environments (Crawford 1989, Herfort et al. 2011, Yih et al. 2013).

Metabolic importance of Synechococcus for Mesodinium rubrum. C requirements for the growth of M. rubrum can be met by photosynthesis using kleptoplastids from prey cryptophytes as well as C from ingested cryptophytes, heterotrophic bacteria, and autotrophic bacteria like Synechococcus spp. (Yih et al. 2004a, Myung et al. 2006, 2011, 2013, Park et al. 2007). Maximum contribution to the growth of M. rubrum by ingested cryptophytes, heterotrophic and phototrophic bacteria were estimated to reach 5.5, 6.2, and 1.2% body carbon ciliate^-1 day^-1, respectively (Yih et al. 2004a, Myung et al. 2006).

New trophic pathways from marine phototrophic prokaryotes. Present study showed that M. rubrum is able to feed on phototrophic prokaryotes the most abundant microorganisms in the ocean (Johnson and Sieburth 1979, Waterbury et al. 1979, Ferris and Palenik 1998). M. rubrum has thus long been involved in the newly recognized trophic pathways between diverse marine organisms and Synechococcus, which further emphasizes the ecological importance of M. rubrum as a new model organism with multiple layers of mixotrophy. Currently, more information about the population dynamics of M. rubrum is needed to understand the relative importance of its Synechococcus feeding for the frequent success of the mixotrophic ciliate in the sea.

Phagotrophy of the phototrophic ciliate Mesodinium rubrum on the cyanobacterium Synechococcus, one of the most abundant single cell phototrophs in the sea, was firstly confirmed by the feeding experiment in the present study. By the unique and complex mixotrophic arrays including phagotrophy on heterotrophic bacteria and Synechococcus as well as digestion, kleptoplastidy and karyoklepty after the ingestion of cryptophyte prey, thus, M. rubrum can form recurrent and massive blooms in diverse marine environments. The new trophic pathway from Synechococcus to diverse predators linked by M. rubrum might further emphasize the ecological importance of M. rubrum as a marine model organism with multiple layers of mixotrophy.