Inducible Mixotrophy in the Dinoflagellate Prorocentrum Minimum

The Journal of Published by the International Society of Eukaryotic Microbiology Protistologists

Journal of Eukaryotic Microbiology ISSN 1066-5234

ORIGINAL ARTICLE Inducible Mixotrophy in the Dinoﬂagellate Prorocentrum minimum

Matthew D. Johnson

Woods Hole Oceanographic Institution, 266 Woods Hole Road, Woods Hole, Massachusetts, 02543

Keywords ABSTRACT Alveolate; cryptophyte; grazing; nutrient starvation; phagotrophic phototroph; stress; Prorocentrum minimum is a neritic dinoflagellate that forms seasonal blooms Teleaulax amphioxeia. and red tides in estuarine ecosystems. While known to be mixotrophic, previous attempts to document feeding on algal prey have yielded low grazing Correspondence rates. In this study, growth and ingestion rates of P. minimum were measured M.D. Johnson, Biology Department, Woods as a function of nitrogen (-N) and phosphorous (-P) starvation. A P. minimum Hole Oceanographic Institution, 266 Woods isolate from Chesapeake Bay was found to ingest cryptophyte prey when in Hole Road, Woods Hole, MA 02543, USA stationary phase and when starved of N or P. Prorocentrum minimum ingested Telephone number: +1(508) 289-2584; two strains of Teleaulax amphioxeia at higher rates than six other cryptophyte FAX number: +1(508) 289-457-2076; species. In all cases -P treatments resulted in the highest grazing. Ingestion e-mail: [email protected] rates of -P cells on T. amphioxeia saturated at ~5 prey per predator per day, while ingestion by -N cells saturated at 1 prey per predator per day. In the Received: 22 August 2014; revised 12 presence of prey, -P treated cells reached a maximum mixotrophic growth rate November 2014; accepted November 12, 1 1 (lmax) of 0.5 d , while -N cells had a lmax of 0.18 d . Calculations of 2014. ingested C, N, and P due to feeding on T. amphioxeia revealed that phagotrophy can be an important source of all three elements. While P. minimum is a doi:10.1111/jeu.12198 proficient phototroph, inducible phagotrophy is an important nutritional source for this dinoflagellate.

MIXOTROPHY is a widely practiced nutritional mode for 1998; Stoecker et al. 2009). Phagotrophy is an ancestral marine protists that combines phagotrophic heterotrophy trait that persists in all eukaryotic supergroups, and con- and photoautotrophy (Jones 1994; Sanders 1991; Stoecker stitutive mixotrophy has been documented within the Stra- 1998). The important ecological role of mixotrophy in pro- menopiles (e.g. chrysophytes), Alveolates (dinoflagellates), viding access to limiting nutrients has long been recog- Haptophytes, Archaeplastids (green algae), and Hacrobians nized (Bird and Kalff 1986; Caron et al. 1990; Keller et al. (cryptophytes). The retention of this trait among plastid- 1994; Nygaard and Tobiesen 1993). In oligotrophic ocean containing protists in marine foodwebs is particularly con- regions, where micro and/or macronutrients are chronically spicuous within the dinoflagellates (Jeong et al. 2010). limiting, bacterivory by mixotrophic flagellates has been Among constitutive mixotrophs, phagotrophy can be shown to be ubiquitous, shaping both the diversity and broadly categorized as being an inducible or persistent productivity of eukaryotic phytoplankton (Hartmann et al. trait, and this phenomenology of feeding reflects its physi- 2012). In coastal ecosystems, mixotrophy can be impor- ological role as either a source for limiting nutrients or as tant for both maximizing growth potential of certain spe- a major nutritional source for maximizing growth. Photo- cies (Li et al. 1999), or as a nutrient source when synthetic and heterotrophic metabolism in mixotrophs is dissolved nutrients are limiting (Smalley et al. 2003). rarely balanced (Flynn and Mitra 2009; Stoecker 1998), Mixotrophy may be divided into two major classes: (1) and the relative importance of each process varies by spe- the phagotrophic phototrophs or constitutive mixotrophs, cies due to factors such as their photosynthetic growth which possess intrinsic plastids and feed to supplement efficiency (Adolf et al. 2006a; Skovgaard et al. 2000) and their nutrition, and (2) photosynthetic protozoa, acquired need for “growth factors” (Skovgaard 2000). phototrophs, or nonconstitutive mixotrophs, which lack The dinoflagellate Prorocentrum minimum has been docu- their own plastids but sequester them from algal prey or mented to form annual blooms in numerous coastal and host endosymbionts (Flynn and Mitra 2009; Stoecker estuarine ecosystems (Glibert et al. 2012; Heil et al. 2005).

In Chesapeake Bay P. minimum is transported annually from extracting overnight in 90% acetone at 20 °C, and reading the mouth of the Bay during winter, to the upper bay the sample on a TD-700 (Turner Designs, Sunnyvale, CA) (~240 km) during spring (Tyler and Seliger 1978). Blooms of fluorometer. Photophysiology was evaluated using a fluo- P. minimum in Chesapeake Bay are most frequent between rescence induction and relaxation (FIRe) fluorometer sys- April and May, between temperatures of 12 and 28 °Cand tem (Satlantic, Halifax, NS, Canada). Dark-adapted (15 min) salinities of 4.5–12.8 PSU, and can reach concentrations in samples were put into a cylindrical 1 cm path length cuv- excess of 105 cells/ml (Tango et al. 2005). Nuisance blooms ette and placed in the FIRe fluorometer, where a blue of P. minimum have become more persistent in many tem- (450 nm) LED source was used to analyze samples. This perate estuaries in response to coastal eutrophication (Gli- consisted of a single turnover flash excitation of 80 lsto bert et al. 2012; Heil et al. 2005; Heisler et al. 2008; saturate photosystem II (PSII) and to push fluorescence to a Hodgkiss and Ho 1997; Paerl 1997). The persistence of these maximum, followed by a relaxation sequence that consisted blooms, under high N:P ratios, relative to Redfield propor- of 40 pulses of weak modulated light separated by 60 ls tions, points to flexible nutrient acquisition pathways (Glibert intervals. All curves were fitted using the software FIREP- et al. 2012). In P. minimum these pathways include mixotro- RO (v.1.20, Satlantic, Inc). A summary of how this method- phy (Stoecker et al. 1997), expression of alkaline phospha- ology works and how various physiological parameters are tase during P-starvation (Dyhrman and Palenik 1999), and determined may be found elsewhere (Kolber et al. 1998). ectocellular aminopeptidases that cleave amino acids from Cellular carbon (C), nitrogen (N), and phosphorous (P) dissolved organic matter (Salerno and Stoecker 2009; Stoec- concentrations were determined by filtering 10–30 ml of ker and Gustafson 2003). While previous studies have docu- cells onto a pre-combusted Whatman GF/F filter, for CHN mented mixotrophy in P. minimum in field (Stoecker 1998) and particulate P (PP) analysis. Samples were analyzed and laboratory (Jeong et al. 2005; Salerno 2005) conditions, using a CE-440 Elemental Autoanalyzer (Exeter Analytical details of their grazing kinetics and how feeding relates to Inc., Chelmsford, MA) and a Technicon AutoAnalyzer II nutrient starvation is lacking. The aims of this study were to (SEAL Analytical, Mequon, WI) for CHN and PP analyses, determine: (1) the optimal cryptophyte prey for P. minimum, respectively, at the Analytical Services Department of Uni- (2) the role of inorganic N and P starvation in inducing mixo- versity of Maryland, Center for Environmental Science, trophic feeding, and (3) to determine grazing rates on crypto- Horn Point Laboratory. Cell dimensions for acid Lugol’s phyte prey and its impact on growth of P. minimum. preserved cryptophyte prey were determined using a Ze- iss Axioscope and correcting for the average cell volume shrinkage (Montagnes et al. 1994) MATERIALS AND METHODS

Algal cultures Experimental conditions Prorocentrum minimum (CBJR10) was isolated by MDJ Nutrient starvation was induced by transferring log-phase from the mouth of the James River, Chesapeake Bay, dur- P. minimum cells into f/2 (-Si) media without either phos- ing a cruise in October of 2011. Teleaulax amphioxeia phorous (-P treatment) or nitrate (-N treatment), and allow- strains GCEP01 and GCEP02 were isolated from Eel Pond, ing cells to starve under batch culture conditions. Falmouth, MA, in May 2008 by Dr. Mengmeng Tong, Assessment of P. minimum grazing on various crypto- Hanusia phi (CCMP 325), Hemiselis andersenii (CCMP phyte species was determined by adding nutrient replete 439), Guillardia theta (CCMP 2712), Proteomonas sulcata (NR), -P (starved 2 wk), or -N (starved 2 wk) P. minimum (CCMP 704), and Rhodomonas salina (CCMP 757) were to 24-well polystyrene plates in filtered seawater at obtained from the National Center for Marine Algae and 1 9 103 cells/ml, and adding each cryptophyte prey at Microbiota (formally CCMP), and Storeatula major (SM or 1 9 104 cells/ml. After 24 h, 1 ml was removed and fixed g) was isolated from Chesapeake Bay by Dr. Allen Lewitus with 1% gluteraldehyde (Sigma Chemicals, St. Louis, MO) (Table 1). All cultures were maintained in batch at 18 °C, for at least 1 h at 4 °C, filtered onto a black 2.0 lm poly- using f/2 medium (-Si) (Guillard and Ryther 1962) added to carbonate filter under low pressure, mounted on a micro- seawater (35 PSU), 0.2 lm-filtered and autoclaved (FSW), scope slide with fluorescence-grade immersion oil, and collected from Martha’s Vineyard Sound, MA. Irradiance enumerated using a Zeiss AxioScope A1 under blue (Ex: was measured using a QMSW-SS light meter (Spectrum BP 450-490, BS FT 510; Em: LP 515 nm) and green (Ex: Technologies, Inc. Aurora, IL), and maintained at 14:10 h BP 546/12, BS FT 560; Em: BP 575–640 nm) fluorescent light-dark cycle, at 50 lmol photons/m2/s. light. Frequency and number of phycoerythrin-containing food vacuoles in feeding cells, which appear as circular orange bodies, were enumerated. For all prey treatments, Cellular attributes and enumeration at least 300 P. minimum cells were enumerated. A Cell concentrations were obtained using an Accuri C6 (BD representative micrograph of a P. minimum cell with T. Biosciences, San Jose, CA) flow cytometer, equipped with amphioxeia-containing food vacuoles was obtained using a laser excitation at 488 and 640 nm and emission at 533/30, Zeiss laser-scanning microscope (LSM) 700 fitted with a 585/40, and > 670 nm, or a Zeiss Axioscope (Zeiss, C-apochromat 40x/1.20-objective (Fig. 1). For this image Thornwood, NJ). Chlorophyll a (chl a) was determined by cells were fixed (as above) and stained with 1 lM of the filtering 10–30 ml of culture into a Whatman GF/C filter, nucleic acid dye 40,6-diamidino-2-phenylindole (DAPI;

Table 1. Cryptophyte cultures used in feeding preference experiments and their cell dimensions

Cryptophyte Clone n Length (lm) Width (lm) Volume (lm3)

Rhodomonas salina CCMP 757 27 12.10 (0.97) 6.94 (0.58) 385.29 (78.91) Storeatula major SM, g 34 11.37 (1.15) 6.64 (0.62) 332.79 (75.69) Teleaulax amphioxeia GCEP01 44 12.26 (2.26) 6.28 (0.45) 321.41 (80.81) Guillardia theta CCMP 2712 19 8.62 (2.48) 5.83 (1.30) 172.99 (60.57) Hanusia phi CCMP 325 26 8.24 (1.61) 5.44 (1.07) 172.17 (38.57) Proteomonas sulcata CCMP 704 15 9.63 (1.92) 5.38 (1.23) 157.17 (39.81) Teleaulax amphioxeia GCEP02 27 7.16 (0.57) 4.41 (0.28) 91.75 (16.04) Hemiselmis andersenii CCMP 439 18 6.70 (1.49) 4.44 (0.97) 70.83 (13.64)

Sigma Chemicals) for 10 min, and then filtered onto a and P3, respectively, prey were washed prior to adding to membrane for microscopy (as above). A layered image P. minimum. T. amphioxeia were centrifuged in 50-ml fal- micrograph was constructed with images using differential con tubes, at 1,500 rpm for 30 min, and the supernatant interference contrast and lasers at 405 nm (DAPI) emitting was gently removed and replaced with supernatant from at 445 nm, 488 nm (chlorophyll) emitting at 515 nm, and either the -N or -P, P. minimum culture. This treatment 543 nm (phycoerythrin) emitting at 600 nm (Fig. 1). was designed to determine if elimination of “carry-over” To determine grazing rates of -N and -P P. minimum, nutrients would alter grazing rates. Each well was sampled cultures were maintained (n = 2) in batch, and sampled in at time 0, 24, and 48 h after addition of prey. Prey-only independent experiments for cellular attributes and grazing controls were sampled for each prey concentration in spent rates, after 2, 3, and 4 wk of nutrient starvation. Using 12- media (GF/C filtered) of P. minimum from the respective well plates, P. minimum from batch cultures were starvation treatment, in order to monitor T. amphioxeia combined with late log phase T. amphioxeia prey at prey/ growth in the absence of P. minimum.AP. minimum-only predator ratios between 0.5 and 20, and T. amphioxeia control was used to measure phototrophic growth P. mini- concentrations between 0 and 110,000 cells/ml. In one mum for each treatment. All cell counts for grazing assays experiment for both -N and -P treatments, experiments N3 were conducted on live cells immediately after sampling, using a flow cytometer (above). Growth (l) and grazing rates, including the grazing coefficient (g), clearance rate (F), and Ingestion rate (IR), were calculated using equations described by Frost (1972) and Heinbokel (1978), and refined by Jeong and Latz (1994) and Kim et al. (2008)

Statistical analysis Differences in P. minimum cellular attributes across different treatments were determined using a one-way ANOVA, and comparisons between treatments were clarified using a Tukey HSD test. Differences in log-transformed growth rate of P. minimum and its grazing rates (g, F, I)onT. amphioxeia between nutrient treatments for pooled prey/ predator ratios (2–10) were determined using a multivariate analysis of variance (MANOVA), and significant differences for each variable were clarified using a univariate ANOVA with Tukey HSD post-hoc tests (Bonferroni adjusted a level of 0.025). One-way ANOVAs and MANOVA were done using R (stats packages: AOV, MANOVA, and TukeyHSD functions) (R Development Core Team 2012).

Figure 1 A confocal laser scanning micrograph image of Prorocen- RESULTS trum minimum with four food vacuoles (FV) containing Teleaulax amphioxeia prey. Image is layered with four channels: a differential Effects of nutrient stress on P. minimum interference contrast (DIC) image showing the P. minimum cell in visi- ble light, a UV light (Ex: 405 nm) image exciting the blue DAPI-stained Both N and P starvation resulted in significant changes in nucleus (N) of P. minimum, a blue-light (Ex: 488 nm) image exciting photosynthetic parameters and intracellular nutrient pools. chlorophyll (red) within the plastids (P) of P. minimum as well as T. Photochemical efficiency (Fv/Fm) was significantly lower amphioxeia in the food vacuoles, and a green light (Ex: 543 nm) in all nutrient stress treatments, dropping to 0.179 in -P image exciting the T. amphioxeia phycoerythrin (yellow) in the plastids cells after about 4 wk of stress (Table 2). Chlorophyll within the food vacuoles. Scale bar = 5 lm. concentrations were also lower in all treatments relative

Table 2. Cellular attributes for Teleaulax amphioxeia (TA) prey and Prorocentrum minimum (PM) under nutrient replete (R), nitrogen (N), or phosphorous (P) limiting conditions for 2, 3, and 4 wk

Sample Fv/Fm Chl per cell C per cell Chl:C (9103) N per cell P per cell C:N C:P N:P

TA-Ra 0.511b (0.001) 1.79 (0.07) 89.6 (5.3) 20.2 (2.3) 13.1 (1.26) 3.13 (0.15) 6.73 (0.23) 29.6 (3.1) 4.42 (0.62) PM-R 0.550 (0.002) 4.36 (0.09) 210 (34.1) 21.0 (3.0) 31.5 (5.78) 4.39 (0.94) 6.66 (0.14) 48.1 (2.6) 7.22 (0.24) PM-P2c 0.413* (0.005) 3.51* (0.26) 308 (23.7) 11.4* (1.72) 26.2 (0.56) 1.52* (0.11) 11.8 (0.65) 204 (30.3) 17.3 (1.62) PM-P3 0.342* (0.003) 2.99* (0.02) 337* (38.2) 8.92* (0.95) 24.5 (1.87) 1.04* (0.09) 13.7* (0.51) 327* (64.5) 23.8 (3.82) PM-P4 0.179* (0.014) 1.33* (0.08) 332* (31.8) 4.00* (0.13) 22.8 (2.45) 0.89* (0.26) 14.7* (2.97) 384* (75.0) 27.2 (10.6) PM-N2 0.412* (0.007) 0.55* (0.01) 249 (4.5) 1.47* (0.04) 20.2* (1.95) – 12.4* (0.98) –– PM-N3 0.346* (0.013) 0.89* (0.06) 276 (24.6) 3.21* (0.05) 15.7* (1.67) – 17.7* (0.31) –– PM-N4 0.084* (0.002) 0.26* (0.01) 271 (18.5) 0.96* (0.03) 10.4* (0.71) – 26.2* (0.01) –– a b c TA-R not included in ANOVA tests; all values are means (n = 2) with the standard deviation in parentheses; weeks of starvation; shaded cells:

*p < 0.05, indicates signiﬁcant difference from PM-R parameter, ANOVA, Tukey’s HSD, performed for each column: Fv/Fm: F(6,7) = 713, p < 0.05;

Chl per cell: F(6,7) = 416, p < 0.05; C per cell: F(6,7) = 6.3, p < 0.05; Chl:C: F(6,7) = 55.8, p < 0.05; N per cell: F(6,7) = 13.8, < 0.05; P per cell:

F(3,4) = 20.3, p < 0.05; C:N: F(6,7) = 44.9, p < 0.05; C:P: F(3,4) = 16.6, p < 0.05; N:P: F(3,4) = 154, N.S. to the control, and in -P cells varied between 1.33 and oxeia, at least 30% of P. minimum had ingested prey (Fig. 2), 3.51 pg chl a per cell. In contrast, chlorophyll concentra- with an average of 3.9 food vacuoles per feeding cell (not tions of -N cells were all below 1 pg chl a per cell follow- shown). In a few cases, cells were packed with so many ing ~2 wk of stress; however Fv/Fm remained above 0.2 food vacuoles (> 10), that it was not possible to accurately until ~3–4 wk of N starvation (Table 2). C per cell determine their number (not shown). Two cryptophyte spe- increased in the -P treatment after 3 wk of stress, while cies, G. theta and P. sulcata, were not ingested by P. mini- chl:C was significantly lower in all treatments compared to mum. Results of a one-way ANOVA (F(7,8) = 310.4; p < 0.001) P. minimum under nutrient replete conditions (PM-R). As revealed that feeding by P. minimum on the two strains of T. expected, N per cell was lower in all -N treatments, while amphioxeia was significantly higher than on all other species P per cell was lower in -P treatments relative to PM-R; (Tukey HSD, p < 0.001), however feeding was also higher however, P per cell was not measured in -N cells on the larger (Table 1) T. amphioxeia strain GCEP01 relative (Table 2). Among the -P experiments, C:P varied between to GCEP02 (Tukey HSD, p < 0.001). No relationship was 204 and 384, while N:P varied from 17.34 to 27.18 found between cryptophyte prey size (Table 1) and fre- (Table 2). In contrast, PM-R cells had a C:P of 48 and a N: quency of feeding by P. minimum when all species were P of 7.22. N-starved P. minimum cells had C:N ratios of considered (not shown). Based on these results, all further 12.4–26.2, while PM-R cells were 6.7 (Table 2). experiments, focused on the grazing kinetics of P. minimum, used T. amphioxeia (GCEP01). Cryptophyte prey preference Grazing and growth response of P. minimum on T. Prorocentrum minimum did not ingest all cryptophyte spe- amphioxeia cies it was offered, and frequency of feeding was dramatically higher (5–89) when in the presence of T. amphioxeia Several experiments were conducted for -P and -N treat- (Fig. 1, 2). After 24 h exposure to two strains of T. amphi- ments in order to determine the role of nutrient starvation on the induction of phagotrophy and the impact of feeding on the growth of P. minimum. P starvation resulted in significantly higher feeding by P. minimum than all other treatments. In both -N and stationary phase (PM-S) cells (not shown), feeding was difficult to measure, and nutrient replete (PM-R) cells were never observed to feed. Mea- sureable grazing for N-starved cells was only found in one experiment (N3), where T. amphioxeia culture media was removed and replaced with supernatant from -N treated P. minimum, prior to introducing to P. minimum cells. The rationale for this treatment was to remove any dissolved inorganic N that may carry over when T. amphioxeia prey was added. PM-R cultures had a growth rate (l) of 0.538 0.08, which was the highest observed for any treatment. Nutri- Figure 2 Frequency of phycoerythrin-containing food vacuoles in ent-starved P. minimum had higher l when fed T. amphi- phosphorous (P) and nitrogen (N) starved Prorocentrum minimum oxeia compared to unfed cultures, however the effects of cells fed a variety of cryptophyte species and strains. feeding on the growth rate of P. minimum were

© 2014 The Author(s) Journal of Eukaryotic Microbiology © 2014 International Society of Protistologists 434 Journal of Eukaryotic Microbiology 2015, 62, 431–443 Johnson P. minimum Mixotrophy dependent upon the T. amphioxeia to P. minimum ratio (T/ In all cases when T/P exceeded 5, g dropped below 1 d1 P) and the duration of nutrient stress prior to feeding. and approached zero when T/P exceeded 10 (Table 3). Maximum impact of feeding on growth was seen after Clearance rates (F) of -P treated cells also declined with 2–3 wk of P-starvation (experiments P2, P3), and resulted increasing prey density, and reached a maximum rate of in a dramatic increase of l after 48 h (Fig. 3). In experi- 162 nl per P. minimum per day (Fig. 4). Ingestion by -N P. ment P2, l increased from < 0.1 d1 without prey to over minimum on washed T. amphioxeia cells was low, relative 0.5 d1 above a T/P of 1. In contrast, P3 l was negative to P-starved cells for the same duration, reaching a maxi- without prey, and increased to ~0.3 d1 above a T/P of mum of about 1 prey per P. minimum per day, and clear- 2.5 (Fig. 3). In experiment P4, which was the most P- ance rates of nearly 100 nl per P. minimum per day. All stressed (Table 2), average growth due to mixotrophy ingestion rates (IR) for P. minimum on T. amphioxeia fol- never exceeded 0.2 d1 and declined when T/P exceeded lowed a hyperbolic function, and were highest when P 5 (Fig. 3). Prorocentrum minimum l in the -N experiment starved, reaching a maximum between 3.5 and 5 prey per was also strongly negative without prey and remained so P. minimum per day (Fig. 5). The differences between until T/P reached 5 (Fig. 3). nutrient treatments based on the combined dependent Grazing coefficients (g) declined exponentially with variables for growth and grazing were highly significant, F increasing T. amphioxeia concentration in all treatments, (3,12) = 87.6, p < 0.0001; Wilks’ Λ = 0.023. This MANOVA and maximum g varied between 1.3 and 2.5 d1 (Table 3). was conducted by pooling all saturated parameter values, which included prey: predator ratios between 2 and 10. Post hoc univariate ANOVAs for each parameter revealed significant differences between all -P treatments and N3 for F and IR, while only one of the l comparisons were different (Table 4). Relatively few differences were found for these parameters among the P treatments, while no differences were found for treatment comparisons for g (Table 4). Both the maximum IR (IRmax) and the initial slope of the IR (aIR) increased with increasing duration of P-starvation. Prorocentrum minimum treated for 4 wk in - P conditions (P4) had the highest overall IRmax (5.18 prey per P. minimum per day), which reached saturation the fastest, with an aIR of 1.34 0.24. In contrast, N-starved P. minimum (N3) had an IRmax of 1.30 and a aIR of 0.32 0.11. A one-way ANOVA (F(3,4) = 12.76, p = 0.016) and post hoc Tukey comparisons between aIR parameters revealed that P4 was significantly different from P3, and N3 (p < 0.05), but not P2.

DISCUSSION Prorocentrum minimum was found to ingest a variety of cryptophyte algae, but had a strong preference for T. amphioxeia that was not based on size alone (Table 1). Feed- ing on the cryptophyte T. amphioxeia was strongly induced by P-starvation, and moderately induced under N- starvation (Fig. 4, 5). Inducible mixotrophy by P. minimum appeared to be particularly sensitive to dissolved inorganic nitrogen (DIN) availability, as washed prey were readily ingested by –N P. minimum while unwashed prey were not. However, the role of DIN in regulating feeding responses needs to be studied further. Feeding was never observed in P. minimum when in nutrient replete conditions, and was observed at low rates for cells in stationary phase (not shown). These characteristics suggest that P. minimum is an inducible mixotroph, feeding in response Figure 3 Growth rate (l) of Prorocentrum minimum as a function of to limiting inorganic P or N, and fitting with Stoecker’s Teleaulax amphioxeia prey levels. Prorocentrum minimum was grown (1998) type II mixotroph designation. These findings sup- under identical conditions (n = 2), except either starved of phospho- port previous studies on field assemblages of P. minimum rous (P) for 2, 3, or 4 wk, or nitrogen (N) for 3 wk. Prorocentrum mini- in Chesapeake Bay, where the frequency of mixotrophy mum l is plotted vs. T. amphioxeia cell and carbon (C) concentrations was observed to increase when sample bottles were (A) or prey/predator ratios (B). Dashed line shows growth rate of zero. enriched with nitrate or phosphate, but not both (Stoecker

Table 3. Prorocentrum minimum grazing on Teleaulax amphioxeia. Prey (C) and predator (P) concentrations, ratio, growth rate of predator (l) and prey (k), grazing constant (g), and growth efﬁciency (GE) after 48 h based on ingested nitrogen (N) or phosphorous (P), for each experiment. Data are mean values with standard deviation in parentheses (n = 2)

1 1 1 NorP [C0][P0] (cells/ml) C0:P0 l (d ) k (d ) g (d )GE(%)

P2 0 16,056 – 0.084 (0.004) –– – 7,584 16,272 0.5 0.406 (0.045) 0.640 (0.036) 1.35 (0.06) 70.0 (22.8) 14,103 13,244 1.1 0.501 (0.036) 0.686 (0.034) 1.11 (0.02) 47.0 (0.96) 23,710 10,957 2.2 0.501 (0.034) 0.670 (0.030) 1.04 (0.07) 24.8 (0.72) 34,447 6,794 5.1 0.506 (0.083) 0.559 (0.012) 0.48 (0.08) 24.2 (10.1) 43,244 4,447 9.7 0.450 (0.099) 0.574 (0.018) 0.27 (0.01) 18.3 (2.77) P3 0 17,885 – 0.169 (0.059) –– – 8,008 14,788 0.5 0.044 (0.038) 0.230 (0.034) 2.51 (0.59) 21.7 (7.49) 15,346 14,615 1.1 0.150 (0.013) 0.177 (0.025) 2.52 (0.47) 16.6 (5.01) 29,731 12,346 2.4 0.284 (0.056) 0.153 (0.024) 1.69 (0.16) 14.9 (2.15) 57,464 12,644 4.5 0.269 (0.038) 0.088 (0.064) 0.58 (0.14) 22.5 (4.52) 84,308 7,117 11.8 0.489 (0.028) 0.153 (0.014) 0.42 (0.04) 17.6 (1.18) 110,534 5,305 20.8 0.489 (0.045) 0.055 (0.091) 0.13 (0.09) 40.2 (20.3) P4 0 12,567 – 0.049 (0.001) –– – 3,378 14,294 0.2 0.007 (0.028) 0.276 (0.058) 1.64 (0.30) 0 5,943 11,868 0.5 0.202 (0.001) 0.265 (0.032) 1.52 (0.15) 18.1 (3.10) 12,478 12,686 1.0 0.122 (0.018) 0.418 (0.017) 1.34 (0.14) 5.69 (2.25) 25,286 9,237 2.7 0.203 (0.027) 0.480 (0.008) 1.05 (0.081) 5.99 (1.69) 47,899 9,104 5.3 0.134 (0.019) 0.489 (0.051) 0.58 (0.08) 4.20 (1.12) 69,527 7,296 9.5 0.035 (0.009) 0.543 (0.027) 0.33 (0.01) 0 N3 0 13,957 – 0.172 (0.058) –– – 8,064 15,053 0.5 0.266 (0.106) 0.218 (0.137) 1.24 (0.04) 0 15,055 13,914 1.1 0.231 (0.162) 0.260 (0.105) 0.95 (0.21) 0 26,756 13,092 2.0 0.016 (0.108) 0.163 (0.023) 0.58 (0.03) 17.0 (10.8) 50,086 9,237 5.4 0.071 (0.027) 0.126 (0.016) 0.31 (0.05) 19.5 (1.75) 77,244 7,659 10.1 0.177 (0.092) 0.016 (0.036) 0.14 (0.04) 35.7 (14.7) 108,546 5,587 19.4 0.168 (0.116) 0.007 (0.004) 0.09 (0.03) 29.2 (15.5)

Growth efficiency (GE) for N or P was calculated for P. minimum (Pm) grazing on T. amphioxeia (Ta), assuming the dissolved fraction is minimal, ½ð ÞðlÞð Þ and using the equation: GE ¼ Pm cells Pm N or P content 100. Here the GE is simply the ratio of the growth in the P. minimum popula- ½ðTa cellsÞðgÞðTa N or P contentÞ tion to their intake of T. amphioxeia, normalized to N or P content, and represented as a percent. et al. 1997). This response can be caused when the addi- increased at C:P ratios > 130 and N:P ratios > 19 (Smalley tion of an N or P source on its own, causes a dramatic et al. 2003), which were similar to values observed here shift in the N/P ratio, inducing starvation of the other for P. minimum starved 2 wk (P2; Table 2). Feeding by C. nutrient (if scarce) over relatively short time scales. furca under N-starvation occurred mostly at C:N ratios above 8, and N:P ratios < 10 (Smalley et al. 2003). Surpris- ingly, no measurable feeding was observed in this study Inducible mixotrophy in dinoflagellates for N-starved cultures that were offered unwashed prey. The induction or increase of phagotrophy by photosyn- In the mixotrophic dinoflagellate, Karlodinium veneficum thetic protists in response to limiting dissolved inorganic (=K. micrum =Gyrodinium galatheanum), feeding occurs nutrients (e.g. N, P, Fe) is one of the more common and under nutrient-replete conditions, but increases under star- well-studied forms of mixotrophy in protists (Stoecker vation of N, P, or both nutrients (Li et al. 2000). As seen 1998). Several other photosynthetic dinoflagellates have with P. minimum (Tables 2 and 4) and C. furca (Smalley been shown to ingest prey primarily under nutrient starva- et al. 2003), P-starvation in K. veneficum results in both tion. In a study of Ceratium furca, feeding was only found the highest grazing response as well as highest cellular C to occur when cultures were N or P starved for at least content (Li et al. 2000). In all three studies, N-starvation 11–16 d, and increased with further nutrient starvation had the strongest effect on cellular chl content (Table 2), (Smalley et al. 2003). Similar to the results presented which is likely a consequence of impaired protein synthe- here, grazing by C. furca was most strongly induced sis (Dong et al. 2013). Inducible mixotrophy has also been through P-starvation. While C:P and C:N ratios for feeding demonstrated in another dinoflagellate, Heterocapsa triqu- P. minimum are similar to those observed for C. furca, the etra, which fed only when grown in media deplete of N current study did not sample in sufficient detail to deter- and P in both light and darkness (Legrand et al. 1998). mine threshold ratios for precisely when feeding is Heterocapsa triquetra ingested a fluorescently labeled induced. In C. furca, feeding in P-starved cultures unidentified flagellate and the diatom Thalassiosira

Figure 5 Ingestion rate (IR) of Prorocentrum minimum as a function Figure 4 Clearance rate (F) of Prorocentrum minimum as a function of Teleaulax amphioxeia prey levels. Prorocentrum minimum was of Teleaulax amphioxeia prey levels. Prorocentrum minimum was grown under identical conditions (n = 2), except either starved of grown under identical conditions (n = 2), except either starved of phosphorous (P) for 2, 3, or 4 wk, or nitrogen (N) for 3 wk. Prorocen- phosphorous (P) for 2, 3, or 4 wk, or nitrogen (N) for 3 wk. Prorocen- trum minimum IR is plotted vs. T. amphioxeia cell and carbon (C) con- trum minimum F is plotted vs. T. amphioxeia cell and carbon (C) concentrations (A) or prey/predator ratios (B). centrations (A) or prey/predator ratios (B). pseudonana, but at much lower rates, up to 0.4 prey per between various treatments for g, however, may seem dino per day, than reported here (Legrand et al. 1998). counterintuitive, but implies that population-level grazing pressure on T. amphioxeia was similar in both treatments, Phosphorous starvation while cell-specific IRs varied. Much of the difference in When experiencing P-stress, P. minimum will express intra- these rates were driven by lower l of prey-only control cellular and cell-surface alkaline phosphatases (AP), which treatments in N3 (Table 3), which were only slightly lower hydrolyze phosphate groups off a variety of molecules (Dy- than P3, but ~59 and 39 lower than P2 and P4 respec- hrman and Palenik 1999, 2001; Sakshaug et al. 1984). tively. Furthermore, feeding by moderately P-stressed cells Expression of AP is induced during late-log or early station- resulted in higher growth rates of P. minimum than -N cells ary phase cultures grown in P-deficient medium (Dyhrman stressed for the same duration. Mixotrophic growth under and Palenik 1999; Sakshaug et al. 1984). In this study, mild -P conditions is similar to the maximum phototrophic feeding by P. minimum was documented at C:P levels of growth rate observed for this strain of P. minimum in nutri- ~200 and above (Table 2), therefore induction of mixotro- ent-replete media. These results suggest that P. minimum phy may follow that of AP activity. Feeding by P. minimum is very efficient at utilizing ingested prey as a source of P. appears to be a major source of P under limiting conditions, as ingestion rates were up to 59 greater than under N-star- Nitrogen starvation vation, and both IR and F were significantly higher com- In this study, feeding by N-starved P. minimum was only pared to N3 (Table 3, Fig. 5). The lack of any difference measurable in one experiment, when T. amphioxeia

Table 4. Results of multivariate analysis of variance post hoc univariate ANOVA’s with Tukey’s HSD to identify differences in mean values of log-transformed growth rate (l)ofProrocentrum minimum and its grazing coefﬁcients (g), clearance rates (F), and ingestion rates (I)onTeleaulax amphioxeia across a series of prey/predator ratios and nutrient stress treatments

l gF I

a Nutrient treatment 14.68(3)*** 3.552(3) 10.28(3)*** 44.74(3)*** Comparison Difference in log-transformed mean P2-N3 0.175*** 0.070 0.439** 0.307*** P3-N3 0.107 0.137 0.499*** 0.441*** P4-N3 0.031 0.168 0.362** 0.531*** P3-P2 0.068 0.066 0.060 0.134 P4-P2 0.207*** 0.098 0.077 0.224*** P4-P3 0.138** 0.032 0.137 0.090

Analysis conducted using data pooled from prey:predator ratios of 2, 5, and 10. Top: results of a univariate ANOVA with the F value and degrees of freedom in parentheses; bottom: results for a Tukey HSD test, reporting the difference in the means of the log-transformed values for each parameter compared, and using a Bonferroni adjusted a level of 0.025. aNutrient treatments: phosphorous (-P) or nitrogen (-N) limited for 2 (P2), 3 (N3, P3), and 4 (P4) weeks. **p < 0.01, ***p < 0.001. culture medium (F/2) was replaced with spent media Mixotrophy in Prorocentrum spp. from the -N P. minimum culture, and occurred at C:N of 17.7 (Table 2). Thus, induction of phagotrophy in P. mini- This study focused on cryptophyte algae due to their pre- mum can occur under moderate to severe N stress, viously established trophic link with P. minimum (Li et al. which also suggests that mixotrophy may be among the 1996; Stoecker et al. 1997). Prorocentrum minimum, last options for N-starved cells. The need to remove car- however, is known to consume other prey, including small ryover DIN prior to feeding, indicates that N-induced Strobilidium sp. (Li et al. 1996), Isochrysis galbana, Amph- mixotrophy in P. minimum is tightly regulated by the idinium carterae, Heterosigma akashiwo (Jeong et al. availability of nutrients. Perhaps this is due to dinoflagel- 2005), Skeletonema costatum (Yoo et al. 2009), Synecho- lates having a greater diversity of mechanisms of acquir- coccus sp. (Jeong et al. 2005), and heterotrophic bacteria ing, storing, and scavenging N relative to P. (Seong et al. 2006; Wikfors and Fernandez 2013). One Prorocentrum minimum can utilize nitrate (NO3), ammo- study of a P. minimum strain isolated from the Chesa- nium (NH4), urea, and dissolved-free amino acids, but peake Bay region found that bacterivory occurred when has a relatively low affinity for their uptake, consistent cells were in stationary phase, but was not observed with being adapted to eutrophic conditions (Glibert et al. when P-starved (Wikfors and Fernandez 2013). In another 2012). It is also known to express ectocellular leucine study, the frequency of feeding on cultured Strobilidium aminopeptidase, which may have a role in scavenging sp. added to natural assemblages containing P. minimum leucine residues of peptides and proteins (Salerno and was found to be relatively high, with 20% of P. minimum Stoecker 2009). In addition to the apparent stronger role having ingested fluorescently labeled ciliates after 4 h (Li of DIN in regulating feeding by P. minimum, N-starvation et al. 1996). Other studies that have utilized protist prey, resulted in an overall lower grazer response relative to -P however, have reported low grazing rates or frequencies. treatments. Furthermore, washing T. amphioxeia with This study is the first to provide detailed grazing kinetics spent media from -P P. minimum cultures (Experiment of P. minimum on protist prey, as well as the first to use P2, Table 2) did not appear to alter grazing rates (Fig. 4, controlled laboratory experiments to understand the rela- 5). This difference, however, could be attributed in part tionship between nutrient starvation and induction of feed- to the molar ratio of N:P in F/2 media, and the propor- ing. Perhaps surprisingly, this study also determined that tionally greater removal of P from media during exponen- at least one strain of P. minimum strongly prefers T. am- tial growth of the cryptophyte prey. Previous studies phioxeia to a broad spectrum of other cryptophyte spe- have shown that the presence of free inorganic P above cies. Previous laboratory studies measuring grazing on 9 lM dramatically reduces feeding in C. furca after 24 h cryptophytes by P. minimum have either reported low (Smalley et al. 2003). Further experiments are needed in feeding frequencies, up to 5% on Pyrenomonas sp. order to determine the role of dissolved nutrients in reg- (=Rhodomonas sp.) (Salerno 2005), or simply verified that ulating phagotrophy in P. minimum. The major differ- R. salina and an unidentified cryptophyte species were ences observed here for grazing rates between N and P ingested (Jeong et al. 2005). While ingestion rates were starvation were surprising, and perhaps can be explained not measured for R. salina prey here, frequency of P. mini- by differences in signaling pathways related to N or P mum with ingested R. salina was 5.34 1.48% after recycling and scavenging. Alternatively, N-starvation 24 h (Fig. 2). would lead to greater overall decline in protein synthesis Culture studies of Prorocentrum donghaiense, Prorocen- (e.g. Dong et al. 2013), effecting photosynthetic as well trum micans, and P. minimum have described a wide vari- as phagotrophic metabolism. ety of potential prey species, their feeding mechanisms,

© 2014 The Author(s) Journal of Eukaryotic Microbiology © 2014 International Society of Protistologists 438 Journal of Eukaryotic Microbiology 2015, 62, 431–443 Johnson P. minimum Mixotrophy and reported growth rates and grazing kinetics on a num- feeding (Table 5). However, growth efficiencies calculated ber of those species (Jeong et al. 2005; Yoo et al. 2009). from P uptake (GEP) due to prey ingestion revealed a Cryptophyte prey are engulfed by Prorocentrum spp. decline in efficiency with increasing time of P starvation through sutures on the side of their cell, most frequently (Table 3). This decline in GEP for the most P-starved treat- occurring in the anterior and posterior ends, as well as the ment (P4) appeared to be counterbalanced by the highest upper right or left sides of the cell (Jeong et al. 2005). In IR and aIR parameters, indicating that these cells were contrast with results reported here for P. minimum, mixo- ingesting prey not only at the greatest rate, but that IR trophy in P. donghaiense and P. micans apparently does saturated faster than other treatments. The dramatic not require nutrient starvation, as experiments described decline in GEP may have been due to accumulated cellular by Jeong et al. (2005) were conducted in F/2 media. Fur- stress and loss of chl a (Table 2). Digestion of chloroplasts thermore, growth of both P. donghaiense and P. micans is and other organelles or cytoplasm occurs during auto- maximum under mixotrophic conditions (Jeong et al. phagy, and is mediated by lysosomal activity. Autophagy 2005), while maximum growth of P. minimum occurred in is associated with senescence in plants (Wada et al. this study under nutrient replete conditions or in mildly P- 2009), and has been demonstrated in the marine chloro- stressed cultures given prey (Table 2). Ingestion rates phyte Nannochloropsis oceanica to occur during N starva- reported here for -P P. minimum were over 29 greater tion (Dong et al. 2013). In Prorocentrum lima and than for P. donghaiense and P. micans grazing on an Prorocentrum maculosum, lysosomes have been reported unidentified cryptophyte (Jeong et al. 2005). However, to increase in size with cell culture age (Zhou and Fritz direct comparisons between these studies are compli- 1994). Decreases in incorporation efficiency of ingested C, cated, due to various differences in experimental design N and P have also been observed in the mixotrophic (e.g. light levels, prey species), as well as uncertainty chrysophyte Poterioochromonas malhamensis in nutrient regarding the growth state of P. donghaiense and P. mi- stressed (i.e. stationary phase) cultures (Caron et al. cans used for grazing experiments. 1990). In all cases, GEP for P. minimum was higher after 48 h, relative to 24 h (not shown), perhaps due to the need to repair accumulated cellular stress caused by auto- Mixotrophy as a nutrient source phagy prior to division. However, in the longest P-stressed Estimates of C, N, and P budgets suggest that at saturat- treatment, GEP was negative or near 0 after 24 h (not ing prey concentrations, ingestion of T. amphioxeia can shown) and remained low after 48 h (Table 3). These dif- provide ~50–200%, 50–190%, and 85–330% of cellular C, ferences suggest that the degree of cellular stress N, and P content per day, respectively (Table 5). Esti- imposed by nutrient starvation is a critical determinant of mated C-fixation rates (Pyield), based on observed chl a per GE, and that for P-starved cells the benefits of feeding cell and published maximum photosynthetic rates normal- increase over time. Synergistic effects of mixotrophy, Pchl ized to cellular chl a ( max), reveal that phagotrophy con- such as repair of damaged cellular components (e.g. chlo- tributes substantially to the C budget of nutrient-starved roplasts) from nutrients acquired through feeding, likely Pchl cells. However, in reality max would likely be substantially occur relatively quickly, however we did not measure lowered depending on the severity and type of nutrient- changes in chl per cell in P. minimum after feeding. In starvation (Adolf et al. 2006a), thus these maximum val- contrast with P-starvation GEN was slightly higher after ues likely under estimate the contribution of mixotrophy to 24 h compared to 48 h, which may indicate that these the cellular C budget. Nevertheless, ingestion exceeded moderately N-stressed cells were able to access and photosynthesis as a source of C at saturated ingestion respond quickly to N acquired from ingested T. amphi- rates for all experiments where P. minimum was actively oxeia prey.

Table 5. Estimated C contributions from photosynthesis and C, N, and P contributions from phagotrophy for mixotrophic Prorocentrum minimum (PM). All ingestion budgets were calculated from saturated ingestion rates

Ingested T. amphioxeiab Photosynthesisa Treatment % C per Pm per day % C per Pm per day % N per Pm per day % P per Pm per day

PM-R 159 ––– P2 86.9 120 (28.9) 117 (28.1) 199 (48.0) P3 68.1 75.5 (20.7) 73.5 (20.2) 138 (37.9) P4 29.9 198 (21.7) 193 (21.2) 330 (36.1 N3 27.3 50.9 (10.9) 49.5 (10.6) 84.5 (18.2)

chl a ½ðchlÞðP Þð14hÞ Estimates for C contributions from photosynthesis for P. minimum were calculated using the equationP ¼ C , where chl is the chl 1 1 observed chl a per cell, P is 5.46 pg C (per pg chl a) h , which is the mean maximum photosynthetic rate (Pmax) for P. minimum in the surface mixed layer of Chesapeake Bay between February and June (Harding and Coats 1988), used to calculate the photosynthetic yield (Pyield)over a 14 h time period at 50 lmol photons/m2/s, and C is the observed cellular C content for P. minimum. Ta bIngestion budgets for C, N, and P were calculated using the equation I ¼ IRN , where NTa and NPm are the nutrient (C, N, or P) content of T. NPm amphioxeia and P. minimum, respectively.

Previous studies of dinoflagellates have shown that growth for P. minimum in this study, moderate to severe mixotrophic GEC, accounting for both photosynthesis and N or P starvation resulted in much lower growth rates, phagotrophy, is around 35 (Adolf et al. 2006a) to 45% and required relatively high levels of prey to approach lmax (Hansen et al. 2000), and is similar to when growing pho- (Fig. 3). Together, these results indicate that for P. mini- totrophically. In the dinoflagellate Fragilidium subglobosum mum, phagotrophy is only advantageous under nutrient C loss due to respiration has been shown to be lower stress and that cells can maintain optimal growth through during mixotrophic than phototrophic growth (Hansen phototrophy. However, while mixotrophy in P. minimum et al. 2000). This increase in efficiency, however, is coun- may act as a “lifeline”, it is an efficient mechanism for terbalanced by a greater value of excreted or egested C, assimilating nutrients when cells are moderately to as well as down-regulated photosynthesis (Adolf et al. severely stressed. 2006a; Hansen et al. 2000). Since photosynthesis was not The phototrophic proficiency of P. minimum, as well as measured here, direct comparisons of GEC from previous its diverse and flexible mechanisms for acquiring nutrients, studies are not possible. However, using the same may explain why it is capable of forming annual spring assumptions described in Table 4 for estimating photosyn- blooms and persistent red tides in coastal ecosystems. In thesis, mixotrophic GEC were estimated to be 31 (P2), 32 Chesapeake Bay, P. minimum has been described as under- (P3), 10 (P4), and 19% (N2) at saturating prey concentra- going an annual cycle, where cells enter the bay in winter, tions (not shown). While these values are lower than pre- become entrained below the pycnocline and transported vious measurements, some of this difference could again north where they surface and bloom during spring near the be explained by over-estimation of contributions from pho- Patapsco River (Tyler and Seliger 1978). During this trans- Pchl tosynthesis due to assuming a high max. Measurements port sub-pycnocline cells have a similar photoacclimation of GEC in F. subglobosum and K. veneficum, which utilize state to cultures growing in low-light (~60 lmol photons/ phagotrophy in a facultative manner to maximize growth m2/s), despite being exposed to light levels below their potential, are also likely to have GEC closer to phototrophic compensation irradiance (Harding and Coats 1988). These values simply because they feed prior to accumulating populations lack vacuoles or other evidence of cellular major cellular stress. In contrast, P. minimum requires stress seen in cultures grown in extremely low light (Coats substantial nutrient limitation to induce feeding on crypto- and Harding 1988). While certain phagotrophic phototrophs phyte algae, and it appears that increased duration of P may use mixotrophy when light-limited (Stoecker 1998), evi- starvation, which results in degradation of the photosys- dence is lacking for increased mixotrophy in pycnocline pop- tem and very high C:P and N:P ratios, continues to lower ulations of P. minimum during spring (Stoecker et al. 1997). GEC. Furthermore, when exposed to temporary darkness, in situ populations of P. minimum significantly decrease their ingestion of cryptophyte algae relative to illuminated incuba- A perspective tions (Stoecker et al. 1997), a phenomenon seen in other Phagotrophy is an ancient eukaryotic trait that has been mixotrophic (Caron et al. 1993; Li et al. 1999) and heterotro- retained in most photosynthetic lineages (Raven 1997). phic protists (Strom 2001). Rather feeding by P. minimum in Compared to photoautotrophy, the evolutionary “cost” of Chesapeake Bay is positively correlated with the abundance maintaining the genetic and cellular capacity for phagotro- of cryptophyte algae and NH4, the latter of which is consis- phy is believed to be low (Raven 1997). In mixotrophic tent with P-starvation (Stoecker et al. 1997). In coastal sys- algae, phagotrophy competes with photosynthesis for cel- tems, eutrophication causes dramatic deviations in N:P lular resources, and often results in lowered C-fixation ratios, frequently leading to induced limitation of N or P rates, loss of chlorophyll, and decreased chl:C (Adolf et al. (Kemp et al. 2005). This “artificial” limitation has been cor- 2006a; Flynn and Mitra 2009; Skovgaard et al. 2000). Fur- related with increased harmful algal bloom activity as well thermore, dinoflagellates tend to have higher respiration as the occurrence of mixotrophy (Burkholder et al. 2008). In rates and low chl:C, relative to their phototrophic counter- Chesapeake Bay, historical analysis of P. minimum abun- parts (Adolf et al. 2006a; Falkowski et al. 1985). As a dance suggests that significantly elevated DIN:PO4 molar result of these and other less well understood variables ratios (~170) are associated with bloom events (Li 2011). (e.g. growth factors), some mixotrophic dinoflagellates Phosphorous limitation in Chesapeake Bay is widespread have been shown to attain maximum growth when mixo- during spring, and DIN:PO4 ratios of incoming river water trophic (Adolf et al. 2006a; Skovgaard et al. 2000). Many frequently exceed 60 (Kemp et al. 2005). These observa- such dinoflagellate species tend to be persistent mixo- tions, coupled with cellular C, N, and P data presented here, trophs, feeding even when sufficient inorganic nutrients suggest that mixotrophy is induced under moderate to are available. In contrast, P. minimum has a chl:C ratio severe nutrient stress, particularly P starvation. Considering typical for a phytoplankter under optimal phototrophic that such conditions are likely to occur during dense growth conditions (Table 2). Furthermore, P. minimum has blooms, mixotrophy may help P. minimum reach maximum a high capacity for photosynthetic growth, with maximum densities and/or persist at bloom levels when DIP becomes rates of several strains exceeding 1 d1 (Glibert et al. limiting. It remains unclear, however, if mixotrophic grazing 2012), and a capacity to photoacclimate to low light (Har- on cryptophyte algae plays an important role in the initiation ding and Coats 1988). While mixotrophic growth under of P. minimum blooms. Studies of annual winter blooms of slight P starvation was similar to maximum phototrophic P. minimum in a sub-tropical tidally mixed estuary reveal

© 2014 The Author(s) Journal of Eukaryotic Microbiology © 2014 International Society of Protistologists 440 Journal of Eukaryotic Microbiology 2015, 62, 431–443 Johnson P. minimum Mixotrophy that peaks in their abundance follow the demise of crypto- ACKNOWLEDGMENTS phyte algal assemblages, however mixotrophic grazing has not been observed (Litaker et al. 2002). Not considered in I would like to thank the National Science Foundation OCE this study or previous field studies in Chesapeake Bay, is awards 1031718 and 1436169, and The Andrew W. the potential contribution of bacteria, cyanobacteria, and Mellon Foundation Endowed Fund for Innovative pico-eukaryotes to their nutrient acquisition. Previous Research, which provided support to complete this research by Jeong et al. (2005) and Wikfors and Fernandez research. I would like to thank David Beaudoin for his (2013) has demonstrated that P. minimum can have a sub- assistance in processing chlorophyll samples and in pre- stantial grazing impact on cyanobacteria and heterotrophic paring slides for microscopy, Drs Donald Anderson and bacteria respectively. Mengmeng Tong for generously providing the Teleaulax The high degree of specificity found here by P. minimum amphioxeia, Dr. Allen Place for kindly providing the Storea- for ingesting T. amphioxeia prey was surprising given the tula major culture, and Ryan Stephansky and Louie Kerr diversity of cryptophyte species they may encounter and for their assistance with confocal microscopy. I would also that feeding is primarily a mechanism to satisfy nutrient like to thank Dr. Diane K. Stoecker, Dr. Hae Jin Jeong, starvation. However, while counterintuitive, similar exam- and an anonymous reviewer for helpful comments on this ples of prey specificity can be found among mixotrophic di- manuscript. noflagellates; F. subglobosum feeds exclusively on Ceratium spp., while C. furca, Gyrodinium uncatenum, and LITERATURE CITED G. instriatum feed only on ciliates (Jeong et al. 2010). In all of these examples, including herein (Table 1), prey size is Adolf, J. E., Stoecker, D. K. & Harding, L. W. 2006a. The balance of not the definitive factor in determining its suitability for autotrophy and heterotrophy during mixotrophic growth of ingestion. One possible explanation for such preference Karlodinium micrum (Dinophyceae). J. Plankton Res.,28:737–751. may be adaptation to select prey based on unique chemical Adolf, J., Yeager, C. L., Miller, W. D., Mallonee, M. & Harding Jr, L. 2006b. Environmental forcing of phytoplankton floral compo- or behavior cues, perhaps due to sharing a common niche. sition, biomass, and primary productivity in Chesapeake Bay, In the Baltic Sea, vertical distribution of Teleaulax sp. during USA. Estuar. Coast. Shelf Sci., 67:108–122. cyanobacterial blooms has been shown to be unique com- Bird, D. F. & Kalff, J. 1986. Bacterial grazing by planktonic lake pared to co-occurring cryptophytes (i.e. Plagioselmis sp. algae. Science, 231:493–495. and Hemiselmis sp.), where during the day Teleaulax sp. Burkholder, J. M., Glibert, P. M. & Skelton, H. M. 2008. Mixotro- were found in a deep layer (20–30 m) and at higher concen- phy, a major mode of nutrition for harmful algal species in trations closer to the surface at night (Hajdu et al. 2007). eutrophic waters. Harmful Algae, 8:77–93. While it is unclear if this behavior is typical of Teleaulax Caron, D. A., Porter, K. G. & Sanders, R. W. 1990. Carbon, nitro- spp., it could make it stand out more to certain predators. In gen, and phosphorus budgets for the mixotrophic phytoflagel- Chesapeake Bay, high concentrations of cryptophytes can late Poterioochromonas malhamensis (Chrysophyceae) during bacterial ingestion. Limnol. Oceanogr., 35:433–443. be found both in the main stem of the Bay (Adolf et al. Caron, D. A., Sanders, R. W., Lim, E. L., Marrase, C., Amaral, L. 2006b), as well as its numerous tributaries (Johnson et al. A., Whitney, S., Aoki, R. B. & Porters, K. G. 1993. Light-depen- 2013). The composition of these cryptophyte communities dent phagotrophy in the freshwater mixotrophic chrysophyte Di- is frequently dominated by Teleaulax spp. (Johnson, M. D., nobryon cylindricum. Microb. Ecol., 25:93–111. et al. in prep). The persistent abundance of T. amphioxeia in Coats, D. W. & Harding, L. W. 1988. Effect of light history on the coastal ecosystems, combined with its unique behavior, ultrastructure and physiology of Prorocentrum mariae-lebouriae may partly explain why P. minimum has adapted to prefer (Dinophyceae). J. Phycol., 24:67–77. it. These results underscore the need for basic research on R Development Core Team 2012. R: A Language and Environ- mixotrophs in order to understand their role in marine micro- ment for Statistical Computing. R Foundation for Statistical bial food webs and adaptations for selecting certain prey. Computing, Austria, Vienna. Dong, H.-P., Williams, E., Wang, D.-Z., Xie, Z.-X., Hsia, R.-C., Jenck, Mixotrophy is the result of coexisting but disparate evo- A., Halden, R., Li, J., Chen, F. & Place, A. R. 2013. Responses of lutionary trajectories and has given rise to a myriad of Nannochloropsis oceanica IMET1 to long-term nitrogen starvation physiological phenotypes and ecological niches. The role and recovery. Plant Physiol., 162:1110–1126. and frequency of feeding by photosynthetic protists varies Dyhrman, S. T. & Palenik, B. 1999. Phosphate stress in cultures greatly, from inducible mixotrophs, like P. minimum, which and field populations of the dinoflagellate Prorocentrum mini- feed to acquire a limiting nutrient, to persistent mixo- mum detected by a single-cell alkaline phosphatase assay. trophs that ingest prey as their major nutritional source Appl. Environ. Microbiol., 65:3205–3212. (e.g. Skovgaard et al. 2000). While mixotrophy is an impor- Dyhrman, S. T. & Palenik, B. 2001. A single-cell immunoassay for tant mechanism for acquiring nutrients for numerous phy- phosphate stress in the dinoflagellate Prorocentrum Minimum – toplankton species, it remains understudied and the (Dinophyceae). J. Phycol., 37:400 410. Falkowski, P. G., Dubinsky, Z. & Wyman, K. 1985. Growth-irradiance impact of mixotrophy on microbial food webs is largely relationships in phytoplankton. Limnol. Oceanogr.,30:311–321. underappreciated (Flynn et al. 2013). Understanding the Flynn, K. J. & Mitra, A. 2009. Building the “perfect beast”: model- ecological role of mixotrophy in coastal (Burkholder et al. ling mixotrophic plankton. J. Plankton Res., 31:965–992. 2008) and open ocean (Hartmann et al. 2012) phytoplank- Flynn, K. J., Stoecker, D. K., Mitra, A., Raven, J. A., Glibert, P. ton is critical for predicting their productivity and biogeo- M., Hansen, P. J., Graneli, E. & Burkholder, J. M. 2013. Misuse chemical role in marine foodwebs. of the phytoplankton-zooplankton dichotomy: the need to

assign organisms as mixotrophs within plankton functional Hagy, J. D., Harding, L. W., Houde, E. D., Kimmel, D. G., Miller, types. J. Plankton Res., 35:3–11. W. D., Newwll, R. I. E., Roman, M. R., Smith, E. M. & Steven- Frost, B. W. 1972. Effects of size and concentration of food parti- son, J. C. 2005. Eutrophication of Chesapeake Bay: historical cles on the feeding behavior of the marine planktonic copepod trends and ecological interactions. Mar. Ecol. Prog. Ser., 303:1– calanus pacificus. Limnol. Oceanogr., 17:805–815. 29. Glibert, P. M., Burkholder, J. M. & Kana, T. M. 2012. Recent Kim, S., Kang, Y. G., Kim, H. S., Yih, W., Coats, D. W. & Park, M. insights about relationships between nutrient availability, forms, G. 2008. Growth and grazing responses of the mixotrophic and stoichiometry, and the distribution, ecophysiology, and food dinoflagellate Dinophysis acuminata as functions of light inten- web effects of pelagic and benthic Prorocentrum species. sity and prey concentration. Aqua. Microb. Ecol., 51:301–310. Harmful Algae, 14:231–259. Kolber, Z. S., Prasil, O. & Falkowski, P. G. 1998. Measurements Guillard, R. R. & Ryther, J. H. 1962. Studies of marine planktonic of variable chlorophyll fluorescence using fast repetition rate diatoms: 1. Cyclotella nana Hustedt, and Detonula confervacia techniques: defining methodology and experimental protocols. (Cleve) Gran. Can. J. Microbiol., 8:229–239. Biochim. Biophys. Acta Bioenerg., 1367:88–106. Hajdu, S., Hoglander,€ H. & Larsson, U. 2007. Phytoplankton verti- Legrand, C., Graneli, E. & Carlsson, P. 1998. Induced phagotrophy cal distributions and composition in Baltic Sea cyanobacterial in the photosynthetic dinoflagellate Heterocapsa triquetra. blooms. Harmful Algae, 6:189–205. Aquat. Microb. Ecol., 15:65–75. Hansen, P. J., Skovgaard, A., Glud, R. N. & Stoecker, D. K. 2000. Li, J. 2011. The Effect of Ambient N: P Ratio and Light on the Physiology of the mixotrophic dinoflagellate Fragilidium subgl- Nitrogen Uptake and Growth of Select Estuarine and Oceanic obosum. II. Effects of time scale and prey concentration on pho- Dinoflagellates. University of Maryland, College Park, MD, 272 tosynthetic performance. Mar. Ecol. Prog. Ser., 201:137–146. p. Harding, L. W. & Coats, D. W. 1988. Photosynthetic physiology Li, A., Stoecker, D. K. & Adolf, J. E. 1999. Feeding, pigmentation, of Prorocentrum mariae-lebouriae (Dinophyceae) during its sub- photosynthesis and growth of the mixotrophic dinoflagellate Gy- pycnocline transport in Chesapeake Bay. J. Phycol., 24:77–89. rodinium galatheanum. Aquat. Microb. Ecol., 19:163–176. Hartmann, M., Grob, C., Tarran, G. A., Martin, A. P., Burkill, P. H., Li, A., Stoecker, D. K. & Coats, D. W. 2000. Mixotrophy in Gyrodi- Scanlan, D. J. & Zubkov, M. V. 2012. Mixotrophic basis of nium galatheanum (Dinophyceae): grazing responses to light Atlantic oligotrophic ecosystems. Proc. Natl Acad. Sci., intensity and inorganic nutrients. J. Phycol., 36:33–45. 109:5756–5760. Li, A., Stoecker, D., Coats, D. W. & Adam, E. J. 1996. Ingestion Heil, C. A., Glibert, P. M. & Fan, C. 2005. Prorocentrum minimum of fluorescently labeled and phycoerythrin-containing prey by (Pavillard) Schiller: a review of a harmful algal bloom species of mixotrophic dinoflagellates. Aquat. Microb. Ecol., 10:139–147. growing worldwide importance. Harmful Algae, 4:449–470. Litaker, R. W., Tester, P. A., Duke, C. S., Kenney, B. E., Pinckney, Heinbokel, J. F. 1978. Studies on the functional role of tintinnids J. L. & Ramus, J. 2002. Seasonl niche stratedy of the bloom- in the Southern California Bight. I. Grazing and growth rates in forming dinoflagellate Heterocapsa triquestra. Mar. Ecol. Prog. laboratory cultures. Mar. Biol., 47:177–189. Ser., 232:45–62. Heisler, J., Glibert, P. M., Burkholder, J. M., Anderson, D. M., Montagnes, D. J., Berges, J. A., Harrison, P. J. & Taylor, F. 1994. Cochlan, W., Dennison, W. C., Dortch, Q., Gobler, C. J., Heil, Estimating carbon, nitrogen, protein, and chlorophyll a from vol- C. A., Humphries, E., Lewitus, A., Magnien, R., Marshall, H. G., ume in marine phytoplankton. Limnol. Oceanogr., 39:1044– Sellner, K., Stockwell, D. A., Stoecker, D. K. & Suddleson, M. 1060. 2008. Eutrophication and harmful algal blooms: a scientific con- Nygaard, K. & Tobiesen, A. 1993. Bacterivory in algae: a survival sensus. Harmful Algae, 8:3–13. strategy during nutrient limitation. Limnol. Oceanogr., 38:273– Hodgkiss, I. & Ho, K. 1997. Are changes in N: P ratios in coastal 279. waters the key to increased red tide blooms? Hydrobiologia, Paerl, H. W. 1997. Coastal eutrophication and harmful algal 352:141–147. blooms: importance of atmospheric deposition and groundwater Jeong, H. J. & Latz, M. I. 1994. Growth and grazing rates of the as “New” nitrogen and other nutrient sources. Limnol. Ocea- heterotrophic dinoflagellates Protoperidinium spp. on red tide nogr., 42:1154–1165. dinoflagellates. Mar. Ecol. Prog. Ser., 106:173–185. Raven, J. A. 1997. Phagotrophy in phototrophs. Limnol. Ocea- Jeong, H. J., Park, J. Y., Nho, J. H., Park, M. O., Ha, J. H., nogr., 42:198–205. Seong, K. A., Jeng, C., Seong, C. N., Lee, K. Y. & Yih, W. H. Sakshaug, E., Graneli, E., Elbrachter,€ M. & Kayser, H. 1984. 2005. Feeding by red-tide dinoflagellates on the cyanobacterium Chemical composition and alkaline phosphatase activity of nutri- Synechococcus. Aquat. Microb. Ecol., 41:131–143. ent-saturated and P-deficient cells of four marine dinoflagel- Jeong, H., Yoo, Y., Kim, J., Seong, K., Kang, N. & Kim, T. 2010. lates. J. Exp. Mar. Biol. Ecol., 77:241–254. Growth, feeding and ecological roles of the mixotrophic and Salerno, M. 2005. Ecophysiology of mixotrophic dinoflagellates: A heterotrophic dinoflagellates in marine planktonic food webs. comparison of the northern Adriatic Sea and the Chesapeake Ocean Sci. J., 45:65–91. Bay. Universita Degli Studi Di Trieste. Johnson, M. D., Stoecker, D. K. & Marshall, H. G. 2013. Seasonal Salerno, M. & Stoecker, D. K. 2009. Ectocellular glucosidase and dynamics of Mesodinium rubrum in Chesapeake Bay. J. Plank- peptidase activity of the mixotrophic dinoflagellate Prorocen- ton Res., 35:877–893. trum minimum (Dinophyceae). J. Phycol., 45:34–45. Jones, R. I. 1994. Mixotrophy in planktonic protists as a spectrum Sanders, R. W. 1991. Mixotrophic protists in marine and freshwa- of nutritional strategies. Mar. Microb. Food Webs, 8:87–96. ter ecosystems. J. Protozool., 38:76–81. Keller, M. D., Shapiro, L. P., Haugen, E. M., Cucci, T. L., Sherr, Seong, K. A., Jeong, H. J., Kim, S., Kim, G. H. & Kang, J. H. E. B. & Sherr, B. F. 1994. Phagotrophy of fluorescently 2006. Bacterivory by co-occurring red-tide algae, heterotrophic labeled bacteria by an oceanic phytoplankter. Microb. Ecol., nanoflagellates, and ciliates. Mar. Ecol. Prog. Ser., 322:85–97. 28:39–52. Skovgaard, A. 2000. A phagotrophically derivable growth factor in Kemp, W. M., Boynton, W. R., Adolf, J., Boesch, D. F., Boicourt, the plastidic dinoflagellate Gyrodinium resplendens (Dinophy- W. C., Brush, G., Cornwell, J. C., Fisher, T. R., Glibert, P. M., ceae). J. Phycol., 36:1069–1078.

Skovgaard, A., Hansen, P. J. & Stoecker, D. K. 2000. Physiology effects due to Prorocentrum minimum blooms in Chesapeake of the mixotrophic dinoflagellate Fragilidium subglobosum.I. Bay. Harmful Algae, 4:525–531. Effects of phagotrophy and irradiance on photosynthesis and Tyler, M. A. & Seliger, H. H. 1978. Annual subsurface transport of carbon content. Mar. Ecol. Prog. Ser., 201:129–136. a red tide dinoflagellate to its bloom area: water circulation Smalley, G. W., Coats, D. W. & Stoecker, D. K. 2003. Feeding in patterns and organism distributions in the Chesapeake Bay. the mixotrophic dinoflagellate Ceratium furca is influenced by Limnol. Oceanogr., 23:227–246. intracellular nutrient concentrations. Mar. Ecol. Prog. Ser., Wada, S., Ishida, H., Izumi, M., Yoshimoto, K., Ohsumi, Y., Mae, 262:137–151. T. & Makino, A. 2009. Autophagy plays a role in chloroplast Stoecker, D. K. 1998. Conceptual models of mixotrophy in plank- degradation during senescence in individually darkened leaves. tonic protists and some ecological and evolutionary implications. Plant Physiol., 149:885–893. Eur. J. Protistol., 34:281–290. Wikfors, G. H. & Fernandez, E. 2013. Induced bacteriovory in a Stoecker, D. K. & Gustafson, D. E. 2003. Cell-surface proteolytic declining culture of the mixotrophic dinoflagellate Prorocentrum activity of photosynthetic dinoflagellates. Aquat. Microb. Ecol., minimum (Pavillard) Schiller. Int. J. Ecol., 2013:4. 30:175–183. Yoo, Y. D., Jeong, H. J., Kim, M. S., Kang, N. S., Song, J. Stoecker, D. K., Johnson, M. D., de Vargas, C. & Not, F. 2009. Acquired Y., Shin, W., Kim, K. Y. & Lee, K. 2009. Feeding by photo- phototrophy in aquatic protists. Aquat. Microb. Ecol., 57:279–310. trophic red-tide dinoflagellates on the ubiquitous marine dia- Stoecker, D., Li, A., Coats, D., Gustafson, D. & Nannenl, M. tom Skeletonema costatum. J. Eukaryot. Microbiol., 56:413– 1997. Mixotrophy in the dinoflagellate Prorocentrum minimum. 420. Mar. Ecol. Prog. Ser., 152:1–12. Zhou, J. & Fritz, L. 1994. The PAS/accumulation bodies in Proro- Strom, S. L. 2001. Light-aided digestion, grazing and growth in centrum lima and Prorocentrum maculosum (Dinophyceae) are herbivorous protists. Aquat. Microb. Ecol., 23:253–261. dinoflagellate lysosomes. J. Phycol., 30:39–44. Tango, P. J., Magnien, R., Butler, W., Luckett, C., Luckenbach, M., Lacouture, R. & Poukish, C. 2005. Impacts and potential