P. micans PMCJH99, isolated from Jinhae Bay in 1999 and P. minimum PMJH00, isolated from Jinhae Bay in 2000 were grown at 20°C in enriched f/2 seawater media (Guillard and Ryther 1962) without silicate, under a 14 : 10 h light-dark cycle, using cool white fluorescent light (50 μmol photons m⁻² s⁻¹).

Transmission electron microscopy (TEM) analysis was used to confirm predation by P. micans on P. minimum (Jeong et al. 2005b). We incubated predator and prey cells in a 250-mL polycarbonate (PC) bottle for 24 h. A 50 mL aliquot from the PC bottle was transferred to a 50-mL centrifuge tube, and the cells were fixed for 1.5 to 2 h by the addition of glutaraldehyde (final concentration 2.5%) in culture medium. Cells were centrifuged and the pellet was agarized. After several rinses with culture medium the cells were postfixed in 1% (w/v) osmium tetroxide in deionized water, then dehydrated using a graded ethanol series (50, 60, 70, 80, 90, and 100% [all v/v] ethanol, followed by two washes with 100% ethanol). The cells were embedded in Spurr’s low viscosity resin (Spurr 1969), sectioned using a RMC MT-XL ultramicrotome (Boeckeler Instruments Inc., Tucson, AZ, USA), and post-stained with 3% (w/v) aqueous uranyl acetate followed by lead citrate. Stained sections were viewed with a JEOL-1010 electron microscope (Jeol Ltd., Tokyo, Japan).

We measured the ingestion and clearance rates of P. minimum by P. micans as a function of incubation time. Dense photosynthetic cultures of P. micans and P. minimum were grown for approximately 1 month and transferred to 10 L PC bottles. 1 mL aliquots from the PC bottles were removed at various times, and cell counts were made using a compound microscope. The experimental starting concentrations of P. micans and P. minimum were established by addition of appropriate volumes of dense cultures. We established triplicate cultures for each of 1) an experimental treatment comprising a mixture of P. micans and P. minimum, 2) a control comprising the predator (P. micans) alone, and 3) a control comprising the prey (P. minimum) alone. In addition, we established an additional bottle containing a filtrate of the predator and prey culture (seawater control).

We used mixed filtrates of the two organisms as a basal medium to ensure consistent water conditions across treatment and controls. To achieve this a predator culture was filtered through a 0.7 μm GF/F filter, and volumes of this filtrate (equal to the volume of predator culture added into the predator control and experimental bottles for each predator-prey combination) were added into the prey control and seawater control bottles. A prey culture was also filtered through a 0.7 μm GF/F filter, and volumes of this filtrate (equal to the volume of prey culture added into the prey control and experimental bottles) were added into both the predator control and seawater control bottles. Two liters of f/2 medium were added to all bottles, which were then filled to 6 L with freshly filtered seawater. Each bottle was then fitted with a cap through which three silicon tubes were inserted. To determine the cell densities (cells mL⁻¹) of predator and prey at the beginning of the experiment and after 1–5 d of incubation, a 20 mL aliquot was removed from each bottle through one of the tubes at each sampling time; 10 mL was fixed with 5% (v/v) Lugol’s solution and 10 mL was fixed with 4% (v/v) formalin. More than 300 cells in three 1 mL Sedgwick-Rafter counting chambers were enumerated. The treatment and control bottles were placed on a shelf without being refilled after subsampling, and incubated at 20°C as described above. To account for the effect of pH on ingestion rates (see next section), incubation of prey control bottles was continued until pH values attained the same pH found in the experimental bottles on the fifth day of incubation, and the prey cell density was determined daily.

For each sampling interval the predation rates of P. minimum by P. micans (cells predator⁻¹ d⁻¹) in the experimental bottles were calculated as described by Jeong et al. (2005c), with the exception of the calculation of growth rate (k). An increase in pH during incubation is known to affect the growth rates of P. micans and P. minimum (Hansen et al. 2007). Therefore, to obtain unique k values for P. minimum and P. micans in the experimental bottles at a given pH, we used the measured k values in the prey and predator control bottles at the same pH as in the experimental bottles.

A 30 mL aliquot was removed from each bottle at each sampling time, and the concentrations of chlorophyll a (chl-a) were measured as described by Arar and Collins (1997).

Seawater C_T and A_T for all bottles were measured using coulometric and potentiometric titration in a VINDTA system (Marianda, Kiel, Germany). The accuracy and precision of C_T and A_T measurements were checked daily against seawater reference materials with known C_T and A_T values (certified by A. Dickson, Scripps Institution of Oceanography, San Diego, CA, USA). The measurement precisions were approximately ±1.5 μmol kg⁻¹ for both C_T and A_T (a total of 14 measurements for each parameter, one set of measurements per day).

Seawater pH values for all bottles were calculated from C_T and A_T measurements using the carbonic acid dissociation constants of Mehrbach et al. (1973) as refitted in different functional forms by Dickson and Millero (1987). This set of thermodynamic constants has proved to be the most consistent for laboratory (Lee et al. 1996, Lueker et al. 2000, Millero et al. 2006) and field (Wanninkhof et al. 1999, Lee et al. 2000, Millero et al. 2002) measurements of carbon parameters. Given the uncertainty (±1.5 μmol kg⁻¹) in C_T and A_T measurements, the predicted pH values based on C_T and A_T measurements were accurate to ±0.005 units. To monitor changes in seawater pH in experiments such as these it is recommended that measurements of C_T and A_T be performed daily, because conventional pH measurement using glass electrodes does not provide stable signals in high pH solutions (pH > 8.5).

The measurements of C_T and pH were performed in parallel with measurements of the growth, ingestion, and clearance rates of P. minimum by P. micans, as described in the preceding section.

As mixotrophic dinoflagellates can perform multiple activities (including photosynthesis, feeding, digestion, and/or respiration), we choose the C_T assay over the ¹⁴C-based short-term incubation method, as the former provides a measure of the net result of multiple activities. We calculated the C_T uptake rate by P. micans cells (nmol C_T dinoflagellate⁻¹ d⁻¹) in predator control bottles at each sampling interval (1 day) by dividing the reduction in C_T (nmol C_T g⁻¹ d⁻¹ or μmol C_T kg⁻¹ d⁻¹) by the mean predator concentration (cells mL⁻¹). The C_T uptake rate by P. minimum cells in prey control bottles was also calculated by dividing the reduction in C_T by the mean prey concentration at each time interval. The mean predator (and prey) concentration at each interval was calculated following the method described by Jeong and Latz (1994).

At each sampling time the expected total C_T uptake rate by the populations of P. micans and P. minimum (μmol C_T kg⁻¹ d⁻¹) in the experimental bottles was determined by summing the calculated uptake rates of their equivalent individual populations in the predator and prey control bottles, respectively. In this calculation the C_T uptake rate by the population of P. micans (nmol C_T g⁻¹ d⁻¹, or μmol C_T kg⁻¹ d⁻¹) in the experimental bottles at each time was obtained by multiplying the mean C_T uptake rate for P. micans cells (nmol C_T dinoflagellate⁻¹ d⁻¹) measured in the predator control bottles by the mean P. micans concentration (cells mL⁻¹) in the experimental bottles. Similarly, the C_T uptake rate by P. minimum was calculated by multiplying the mean C_T uptake rate of P. minimum cells measured in the prey bottles by the mean P. minimum concentration in the experimental bottles. To minimize the effect of pH on the C_T uptake rate, the mean C_T uptake rates (nmol C_T dinoflagellate⁻¹ d⁻¹) obtained from the prey and predator control bottles at the same pH as in the experimental bottles were used.

Experimental data were treated with linear regression analysis of variance. If the p-value is smaller than 0.05, the ANOVA test implies that a relation does exist between variables. Using this result along with the scatter plot of the figure consisting of a pair of variables (one for x-axis, and one for y-axis), it can be concluded that the relationship between two variables is linear.

A TEM image (Fig. 1) shows a P. minimum cell within a cell of the P. micans predator. Using an inverted microscope we previously captured an image of P. micans engulfing a P. minimum cell through the sutures (Jeong et al. 2005b).

With increasing incubation time, the abundance of P. micans in the experimental treatment bottles increased from 550 cells mL⁻¹ (day 0) to 3,596 cells mL⁻¹ (day 5), whereas abundance in the predator control bottles increased from 562 to 4,290 cells mL⁻¹ (Fig. 2A). Over the same period the abundance of P. minimum in the experimental treatment bottles increased from 3,820 to 16,639 cells mL⁻¹, whereas their abundance in the prey control bottles increased from 3,878 to 30,333 cells mL⁻¹ (Fig. 2B).

The chl-a concentration in the predator and prey control bottles increased (day 0–5) from 14.5 to 76.9 ng mL⁻¹, and 8.7 to 28.9 ng mL⁻¹, respectively (Fig. 3A & B), whereas the chl-a concentration in the experimental treatment bottles increased from 20.2 to 82.4 ng mL⁻¹ over the same period (Fig. 3C). Over the same period (day 0–5) the pH increased considerably in the predator control (from 8.36 to 9.47), the prey control (from 8.27 to 8.73), and the experimental treatment (from 8.38 to 9.57) bottles (Fig. 4A–C), whereas the pH in the seawater control bottle remained approximately unchanged (from 8.10 to 8.12).

During the incubation period (day 0–5) the concentration of C_T in the predator control and prey control bottles decreased from 1,800 to 1,100 μmol C_T kg⁻¹ and from 1,850 to 1,550 μmol C_T kg⁻¹, respectively (Fig. 5A & B). The C_T value for the experimental treatment bottles decreased from 1,790 to 1,060 μmol C_T kg⁻¹ (Fig. 5C), whereas The concentration of C_T in the seawater control bottle remained between 2,092 and 2,099 μmol C_T kg⁻¹.

Seawater pH is known to affect the growth rates of P. micans and P. minimum; in general, the higher the pH the lower the growth rate. The pH in the predator control and experimental treatment bottles increased more rapidly than in the prey control bottles. Therefore, our measurements of abundance of P. minimum, pH, and C_T in the prey control bottles were extended up to day 12 (Fig. 6), as opposed to day 5 in the predator control and experimental bottles. The abundance of P. minimum in the prey control bottles increased from 35,708 cells mL⁻¹ at day 6 to 68,583 cells mL⁻¹ at day 12 (Fig. 6A). The pH increased in the prey control bottles from 8.85 to 9.48 (Fig. 6B) between day 6 and day 12, and the C_T concentrations decreased from 1,480 to 1,100 μmol C_T kg⁻¹ (Fig. 6C).

The C_T uptake rates by P. micans in the predator control bottles (0.049–0.118 nmol C_T dinoflagellate⁻¹ d⁻¹) (Fig. 7A) were an order of magnitude higher than those for P. minimum in the prey control bottles (0.004–0.011 nmol C_T dinoflagellate⁻¹ d⁻¹) (Fig. 7B). However, the chl-a specific C_T uptake rates of P. micans in the predator control bottles (112–220 μmol C_T [mg chl-a] ⁻¹ h⁻¹) were similar in magnitude to those of P. minimum in the prey control bottles (98–254 μmol C_T [mg chl-a] ⁻¹ h⁻¹). The C_T uptake rates (per cell) of P. micans and P. minimum were significantly positively correlated with the C_T concentration (p < 0.05, linear regression ANOVA) (Fig. 7).

The total C_T uptake rate of the combined populations of P. micans and P. minimum in the experimental bottles was 123–161 μmol C_T kg⁻¹ d⁻¹, which was lower than the expected total C_T uptake rate (136–212 μmol C_T kg⁻¹ d⁻¹) solely by a phototrophic growth of P. micans and P. minimum (Fig. 8A). With increasing incubation time the difference in the total C_T uptake rate (expected total C_T uptake rate - measured total C_T uptake rate) increased from 8 to 56 μmol C_T kg⁻¹ d⁻¹ (Fig. 8B). Even for C_T uptake corrected for the effect of pH, the measured total C_T uptake rate was lower than the expected rate (132–176 μmol C_T kg⁻¹ d⁻¹) (Fig. 8C), and the reduction in the total C_T uptake rate arising from the mixotrophy of P. micans increased from 6 to 25 μmol C_T kg⁻¹ d⁻¹ (Fig. 8D).

The higher pH (or lower C_T) observed in the prey control bottles between day 6 and day 12 could have lowered the growth rate of P. minimum. Thus, we calculated two ingestion rates of P. minimum by P. micans; one rate corrected for pH effects and the other not so corrected. To account for the effect of pH on ingestion rate, we calculated the ingestion rate of P. minimum by P. micans by replacing the growth rate of P. minimum in the experimental bottles by that obtained for P. minimum in the prey control bottles at similar pH and C_T levels.

Without correcting for pH effects the apparent predation rate of P. minimum by P. micans was 236–2,784 cells d⁻¹ at day 5 (Fig. 9A). In contrast, the pH-corrected predation rate of P. minimum by P. micans was 630–1,946 cells d⁻¹ at day 5 (Fig. 9B).

Our results indicate that the differences between the expected and measured rates of C_T uptake in the experimental treatment bottles were significantly positively correlated with the rate of predation of P. minimum by P. micans (p < 0.05 for both, linear regression ANOVA) (Fig. 10A & B). These relationships suggest that feeding by P. micans on P. minimum may be an important factor in controlling respective rates of carbon uptake.