Parasitism of Photosynthetic Dinoflagellates by Three Strains of Amoebophrya (Dinophyta): Parasite Survival, Infectivity, Generation Time, and Host Specificity1

J. Phycol. 38, 520–528 (2002)

PARASITISM OF PHOTOSYNTHETIC DINOFLAGELLATES BY THREE STRAINS OF AMOEBOPHRYA (DINOPHYTA): PARASITE SURVIVAL, INFECTIVITY, GENERATION TIME, AND HOST SPECIFICITY1

D. Wayne Coats2 and Myung Gil Park3 Smithsonian Environmental Research Center, P.O. Box 28, Edgewater, Maryland 21037, USA

Amoebophrya ceratii (Koeppen) Cachon is an obli- the cytoplasm and/or nucleus of dinoflagellates are gate parasite of dinoflagellates and may represent a generally believed to have a commensal or mutualistic species complex. However, little is known about the relationship with their host (Silva and Franca 1985, biology and host range of different strains of Amoe- Doucette et al. 1998); however, Kirchner et al. (1999) bophrya Cachon. Here, we determined parasite gen- recently linked bacterial infections with decreased eration time and dinospore infectivity, survival, and growth of Noctiluca scintillans (Macartney) Kofoid and ability to infect nonprimary hosts for strains of Amoe- Swezy, suggesting that some bacteria–dinoflagellate bophrya from Akashiwo sanguinea (Hirasaka) G. Hansen associations may be parasitic. Most other intracellular et Moestrup, Gymnodinium instriatum (Freudenthal et symbionts of dinoflagellates are clearly parasitic and Lee) Coats comb. nov., and Karlodinium micrum (Lead- have marked effects on host survival and reproduc- beater et Dodge) J. Larsen. Akashiwo sanguinea was tion (Elbrächter and Schnepf 1998). readily infected, with parasite prevalence reaching Parasite-induced mortality and associated decreases 100% in dinospore:host inoculations above a 10:1 ra- in reproduction can have strong influences on dino- tio. Parasitism also approached 100% in G. instria- flagellate population dynamics. For example, fungal tum, but only when inoculations exceeded a 40:1 ra- parasites cause major losses of cysts and vegetative tio. Karlodinium micrum appeared partially resistant cells of the freshwater dinoflagellate Ceratium hirund- to infection, as parasite prevalence saturated at 92%. inella (O. F. Müller) and thereby contribute to short- Parasite generation time differed markedly among term and the long-term reductions in host density Amoebophrya strains. Survival and infectivity of dino- (Canter and Heaney 1984, Sommer et al. 1984, Heaney spores decreased over time, with strains from G. in- et al. 1988). In marine environments, parasitic dino- striatum and A. sanguinea unable to initiate infections flagellates of the genus Amoebophrya Cachon and the after 2 and 5 days, respectively. By contrast, dino- perkinsozoa Parvilucifera infectans Norén et Moestrup spores from Amoebophrya parasitizing K. micrum re- infect numerous taxa of photosynthetic and hetero- mained infective for up to 11 days. Akashiwo sanguinea trophic dinoflagellates, including several toxic species and G. instriatum were not infected when exposed to (Coats 1999, Norén et al. 1999). Outbreaks of these dinospores from nonprimary Amoebophrya strains. parasites have been linked to declines of dinoflagel- Karlodinium micrum, however, was attacked by dino- late blooms in several marine systems (Cachon 1964, spores of Amoebophrya from the other two host spe- Taylor 1968, Nishitani and Chew 1984, Coats et al. cies, but infections failed to reach maturity. Observed 1996). differences in host–parasite biology support the hy- Amoebophrya species appear well adapted to exploit pothesis that Amoebophrya ceratii represents a com- dinoflagellate blooms in enriched coastal environ- plex of host-specific species. Results also suggest that ments (Yih and Coats 2000) where parasite preva- Amoebophrya strains have evolved somewhat diver- lence ( percent hosts infected) can exceed 80% gent survival strategies that may encompass sexual- (Coats 1999). Particularly high infection levels in the ity, heterotrophy during the “free-living” dinospore toxic dinoflagellate Alexandrium catenella (Whedon stage, and dormancy. and Kofoid) Balech led Taylor (1968) to suggest that Key index words: Akashiwo sanguinea; Gymnodinium in- parasites, specifically Amoebophrya ceratii (Koeppen) striatum; Karlodinium micrum; plankton; protist; red tide Cachon, might be intentionally used to control red tides; however, his idea was later rejected by Nishitani et al. (1985), because A. ceratii appeared to lack the Dinoflagellates serve as hosts for a variety of intra- host specificity needed to target any particular harm- cellular symbionts, including viruses, bacteria, fungi, ful bloom-forming species. Subsequently, Coats et al. and protists (Elbrächter and Schnepf 1998, Norén et (1996) demonstrated that cultures of Amoebophrya cer- al. 1999, Tarutani et al. 2001). Bacteria that inhabit atii ex Akashiwo sanguinea (Hirasaka) G. Hansen et Moestrup (formally Gymnodinium sanguineum Hirasaka) failed to infect several other dinoflagellate species and suggested that A. ceratii represented a species 1 Received 26 October 2001. Accepted 5 March 2002. 2 Author for correspondence: e-mail [email protected]. complex composed of several host-specific taxa. Re- 3 Present address: Red Tide Research Center, Kunsan National Uni- cent molecular studies support that argument by show- versity, Kunsan 573-701, Republic of Korea. ing marked genetic diversity among strains of Amoe-

520

PARASITISM OF DINOFLAGELLATES 521 bophrya from different host species (Gunderson et al. ferring infected dinoflagellates from Chesapeake Bay plankton 2000). Thus, the possibility of using parasites to con- to cultures of the complementary host species. Parasites were propagated by sequentially transferring aliquots of infected trol harmful algal blooms is now receiving renewed at- host culture into uninfected host stock at roughly 2-day inter- tention (Anderson 1997, Norén et al. 1999, Erard-Le vals for A. sanguinea and G. instriatum and twice weekly for K. Denn et al. 2000). micrum. Stock cultures were maintained at 20 C under a 14:10 Species of Amoebophrya that infect dinoflagellates light:dark cycle of cool-white fluorescent light at 175 mol pho- –2 –1 –2 –1 have a relatively simple life cycle, consisting of an in- tons m s for A. sanguinea and 95 mol photons m s for G. instriatum and K. micrum. fective dispersal stage, the dinospore, an intracellular All experiments were conducted using host cultures in ex- growth stage, the trophont, and an extracellular re- ponential growth and recently formed (6 h) dinospores of productive stage, the vermiform (Cachon 1964, Ca- Amoebophrya spp. To obtain recently formed dinospores, stock chon and Cachon 1987). Dinospores are small (8–10 parasite cultures were gravity filtered through Nucleopore filters (12-m pore size for A. sanguinea, 8 m for G. instriatum, 5 m long) biflagellate cells that attach to the surface of m for K. micrum) (Costar-Nucleopore, Cambridge, MA, USA) host organisms, penetrate through the host pellicle, to yield filtrates containing dinospores of unknown age. Al- and develop into trophonts either in the cytoplasm or though dinospores of our three Amoebophrya strains were of sim- in the nucleus of the host, depending on the species ilar size, host size differed greatly across species. Thus, filters of of dinoflagellate being infected (Cachon 1964, Park different pore sizes were used in harvesting dinospores to en- sure efficient removal of remaining host cells. Dinospores were et al. 2002). The trophont stage grows inside the host inoculated into uninfected host culture in sufficient quantities for 2–3 days (20–23 C; Coats and Bockstahler 1994, (dinospore:host ratio of 20:1 to 30:1) to produce infection lev- Yih and Coats 2000), during which time it undergoes els 80%. Dinospores produced from the subsequent parasite nuclear and flagellar replication to produce a large generation were harvested within 6 h ( recently formed dinospores) using filtration procedures described above. multinucleate and multiflagellate cell known to as a Parasite prevalence versus inoculum size. Parasite prevalence was “beehive.” At maturity, the beehive metamorphoses determined as a function of inoculum size in four separate ex- into a vermiform that ruptures through the host cell periments, one for A. sanguinea, one for G. instriatum, and two membrane. Vermiforms persist for only a short time for K. micrum. For each experiment, sets of triplicate scintilla- 3 1 (20 min at 23 C), with completion of cytokinesis tion vials containing 10 or 20 mL of host cells at 1 10 mL were inoculated with dinospores and incubated for 24–40 h un- giving rise to hundreds to thousands of new dino- der growth conditions described above. Incubation time dif- spores (Coats and Bockstahler 1994). Survival time of fered for the three host species (36, 40, and 35 or 36 h for A. dinospores outside the host is unknown; however, sanguinea, G. instriatum, and K. micrum, respectively) but was al- grazing by ciliates may influence their ability to trans- ways shorter than the time needed for the parasite to reach maturity. Inoculum size for each set of triplicate vials was adjusted mit infections to new host cells (Maranda 2001, Jo- to give dinospore:host ratios of 1:2, 1:1, 5:1, 10:1, 20:1, and 40:1 hansson and Coats, unpublished data). for A. sanguinea, 1:1, 2:1, 5:1, 10:1, 20:1, 30:1 40:1 and 100:1 for Here we compare aspects of host–parasite biology K. micrum, and 1:1, 2:1, 5:1, 10:1, 20:1, 40:1, 80:1, and 120:1 for for strains of Amoebophrya isolated from bloom-form- G. instriatum. After incubation, samples were preserved with ing dinoflagellates of Chesapeake Bay. The host–para- CaCO3 buffered formaldehyde (2% final concentration) and examined by inverted epifluorescence microscopy (Leitz Dia- site systems we examined were Amoebophrya sp. ex A. vert, 450- to 490-nm excitation, 520 nm barrier filter, 400) to sanguinea, Amoebophrya sp. ex Gymnodinium instriatum visualize the distinctive green autofluorescence of the parasite (Freudenthal et Lee) Coats comb. nov. (formerly Gyro- (Coats and Bockstahler 1994). Parasite prevalence ( percent dinium instriatum Freudenthal et Lee), and Amoebophrya cells infected) was determined by scoring at least 100 hosts cells per sample as infected or uninfected. sp. ex Karlodinium micrum (Leadbeater and Dodge) J. Data for each host species were fitted to a single two-param- Larsen (formerly Gyrodinium galatheanum [Braarud] Tay- eter exponential rise to maximum. The equation for the curve lor). Our goal was to better define the biological di- fit was y a*(1 eb*x), where a is the maximum infection versity of Amoebophrya strains that infect different di- level (Imax ) and b is /Imax. Alpha ( ) represents the slope of noflagellate species. For each host–parasite system, we the initial linear portion of the fitted curve and reflects the po- tential of dinospores to infect host cells. Estimates for were examined the influence of inoculum size (i.e. dino- derived as Imax*b, with SE for determined by propagation of spore:host ratio) on parasite prevalence, determined error terms. duration of the infection cycle, estimated total para- Dinospore survival and infectivity. Replicate 300- to 400-mL site generation time, and assessed survival and infec- volumes of recently formed dinospores harvested as above from each host species were held under standard growth conditions tivity of dinospores with age. We also evaluated the in- (20 C with a 14:10 light:dark cycle of cool-white fluorescent fluence of inoculum size on parasite development light at 100–200 mol photonsm2s1) and subsampled over and assessed the ability of each parasite strain to in- time to estimate dinospore survival and infectivity. Persistence fect the other host species. of green autofluoresence was taken as an indication of dinospore survival and was used to track cell abundance over time. materials and methods Samples were preserved with formalin and examined for dinospore abundance using a hemocytometer and a Zeiss Axio- Laboratory cultures. Chesapeake Bay isolates of A. sanguinea, scope (200) (Zeiss, Jena, Germany) equipped for epifluores- G. instriatum, and K. micrum were maintained as stock cultures cence microscopy (Filter Cube 09, 450- to 490-nm excitation, in f/2-Si medium (Guillard and Ryther 1962). The medium was 510-nm beam splitter, 520-nm barrier filter). Counts were made formulated using 15 ppt Chesapeake Bay water supplemented in duplicate, with each based on examination of 10 hemocy- with 5% (v/v) GR soilwater (Starr and Zeikus 1993). tometer grids equivalent to 1 mm3 of sample. Data for each par- Three strains of parasitic dinoflagellates belonging to the asite strain were fitted to a single two-parameter exponential genus Amoebophrya, one each in the host species A. sanguinea, G. decay curve. For our purposes, this model corresponds to the kt instriatum, and K. micrum, were established in culture by trans- exponential growth function Nt N0e , where N0 is the dino- 522 D. WAYNE COATS AND MYUNG GIL PARK spore abundance (cellsmL1) at the beginning of the experi- with dinospores from K. micrum to give dinospore:host ratios of ment (T0), Nt is the dinospore abundance at subsequent times, 10:1, 50:1, and 50:1, respectively. Flasks were incubated as above, t is the time in days, and k is the the negative growth constant except that light was supplied at 80 mol photonsm2s1. Af- (i.e. mortality rate). ter inoculation, 2-mL subsamples were taken from each flask at To assess the ability of dinospores of different ages to infect 12-h intervals over 4 days and fixed with CaCO3 buffered form- naive host cells, subsamples from each of the replicate 300- to aldehyde (2% final concentration) for epifluorescence deter- 400-mL volumes of recently formed dinospores were added to mination of parasite prevalence as above. triplicate scintillation vials at each sampling period and mixed Data analysis. Data are reported in the text as mean SE. with equivalent volumes of the corresponding host culture to Statistical analyses and curve fits were performed using Jandel produce 10- or 20-mL volumes of host cells at 1 103 mL1. Scientific Software (SigmaPlot 6.0 and SigmaStat 2.0) (SPSS This approach generated a dinospore:host ratio of 5:1 for A. Science, Chicago, IL, USA). sanguinea, 15:1 for K. micrum, and 22:1 for G. instriatum for the initial (T0) sampling period. Ratios for subsequent sampling results times varied with survival of dinospores. Scintillation vials were incubated for 36 h as above, fixed using concentrated Bouin’s Nomenclature. Specimens from our cultures of G. in- fluid (Coats and Heinbokel 1982), and stained for determina- striatum were morphologically indistinguishable from tion of parasite prevalence and parasite load (i.e. number of Gyrodinium instriatum Freudenthal et Lee, except that parasites inside each infected host). Preserved samples for A. close examination revealed the presence of a horse- sanguinea and G. instriatum were processed using quantitative shoe-shaped apical groove running in an anticlock- protargol staining (Montagnes and Lynn 1993), whereas those for K. micrum were stained with alum hematoxylin (Galigher wise direction. Because an apical groove of this con- and Kozloff 1971). Stained samples were examined using Zeiss figuration is characteristic of the genus Gymnodinium optics (1000–1250), with parasite prevalence and parasite Kofoid et Swezy emend. G. Hansen et Moestrup (Daug- load determined following procedures of Coats et al. (1996). bjerg et al. 2000), we refer to this host species as Gym- Dinospore success (i.e. the percent of dinospores that invaded host cells) at each sampling period was expressed relative to di- nodinium instriatum (Freudenthal et Lee) Coats comb. nospore abundance at T0 and at the time samples were taken. nov.; basionym: Gyrodinium instriatum Freudenthal et Generation time. For each host species, replicate 200- or 300- Lee (Freudenthal and Lee 1963). mL cultures at initial densities of approximately 1 103 mL 1 Parasite prevalence versus inoculum size. Parasite prev- were inoculated with recently formed dinospores, incubated alence showed an exponential increase to a maximum under growth conditions described above, and sampled over time to determine host abundance, parasite prevalence, stage relative to inoculum size in all three host species (Fig. of infection, and dinospore abundance. To examine possible 1). Estimates for maximum infection levels (Imax) and effects of inoculum size on parasite generation time, treat- initial slope of the fitted curves ( ) were Imax 98.2 ments for each host species included inoculations at low (ap- 0.82; 66.0 2.68 (r2 0.9921; P 0.0001) for A. proximately 1:1) and high dinospore:host ratios (20:1, 40:1, 2 sanguinea, Imax 97.9 1.03; 9.3 0.36 (r and 115:1 for A. sanguinea, K. micrum, and G. instriatum, respectively). 0.9928; P 0.0001) for G. instriatum, and Imax 91.8 Estimates for host abundance were obtained from Bouin- 2.37; 19.6 1.91 (r2 0.9042; P 0.0001) for K. fixed samples by enumerating cells present in microscope micrum (experiments 1 and 2 combined). transects ( 100) of triplicate Sedgwick-Rafter chambers. For Dinospores of Amoebophrya sp. from A. sanguinea each chamber, successive transects were examined until 100 cells had been counted or five transects ( half the chamber were most aggressive of the three parasite strains, with area) had been scanned. Parasite prevalence was determined parasite prevalence rising sharply to near maximum from protargol-stained samples as described previously, with levels by a dinospore:host ratio of 10:1. Inoculations parasitized hosts partitioned by stage of infection using criteria above a 10:1 ratio typically showed 100% infection of patterned after those of Coats and Bockstahler (1994). For this analysis, early infections were classified as host cells containing parasites that had a single nucleus or multiple nuclei that were arranged in an irregular pattern. Late infections were host cells that contained parasites having multiple nuclei arranged in a circular or spiraled fashion. Dinospore abundance was obtained from formalin-fixed samples as above. Temporal differences in the occurrence of early and late stage infections were used to estimate parasite intracellular development time. The areas defined by plots for abundance of early and late stage infections versus time were determined by integration, with the mid-point of each stage derived as half of the integrated area for the corresponding plot. The time inter- val between mid-points for early and late stage infections was multiplied by two to estimate the duration of the parasite’s intracellular phase. Estimates for total generation times (i.e. time required for infection of host cells, intracellular development, vermiform emergence, and extracellular maturation) were derived from the temporal occurrence of new dinospores after infection of host cells. Plot areas from the appearance of new di- Fig. 1. Parasite prevalence as a function of inoculum size nospores to peak concentrations were calculated by integration, for strains of Amoebophrya infecting Akashiwo sanguinea, Karlo- with elapsed time corresponding to half of the integrated area dinium micrum, and Gymnodinium instriatum. Host density was taken as total generation time. maintained at 103 cellsmL1, with dinospore density varied to Cross infection. Six flasks containing 100 mL of culture at 1 yield dinospore:host ratios of 1:1 to 120:1. Data represent repli- 103 cellsmL1 were established for each host species. Two flasks cate measurements for each experimental treatment. Karlodin- per host species were inoculated with dinospores from infected ium micrum 1 and K. micrum 2 indicate results from two separate A. sanguinea, two with dinospores from G. instriatum, and two experiments. PARASITISM OF DINOFLAGELLATES 523 host cells. Dinospores from G. instriatum were least ag- that successfully established infections (Fig. 3A) de- gressive, with parasite prevalence showing a more clined with dinospore age, more or less paralleling gradual increase to saturation; however, 100% infec- patterns in dinospore survival; however, the ability of tion levels were detected at a dinospore:host ratio of surviving dinospores (i.e. those still detectable by epif- 120:1. Parasitism in K. micrum saturated at an inocu- luorescence microscopy) to infect host cells exhibited lum size intermediate to that of A. sanguinea and G. a markedly different pattern (Fig. 3B). Success of sur- instriatum (dinospore: host ratio of 20:1) but failed to viving dinospores declined rapidly with age for Amoe- reach 100% infection levels, even at a dinospore:host bophrya strains infecting G. instriatum and A. sanguinea, ratio of 100:1. Parasite prevalence in K. micrum only with values dropping below detection levels (0.1%) averaged 91.3 1.37% (n 12) at dinospore:host ra- by days 2 and 5, respectively. By contrast, the success tios 20:1. of surviving dinospores from K. micrum increased by a Dinospore survival and infectivity. Size fractionation factor of two during the first day, stabilized at 18% of infected host cultures provided stocks of recently over the following 3 days, and then gradually declined formed dinospores (6 h old) with initial densities of with some infections still being established when di- 9.5 2.0 103 mL1 from A. sanguinea, 30 3.5 nospores were 11 days old. 103 mL1 from K. micrum, and 43 5.3 103 mL1 Generation time. Inoculation of A. sanguinea and G. from G. instriatum. From the onset of the experiment, instriatum with dinospores at a 1:1 ratio produced in- dinospore abundance for Amoebophrya strains from G. creasing numbers of infected hosts over the first 20– instriatum and K. micrum exhibited an exponential de- 24 h, with values stabilizing at 470 9 cellsmL1 (n 1 1 cay, falling below detection levels ( 250 cells mL ) 6; T24-T48) and 100 6 cells mL (n 7; T20-T44), re- by days 3 and 13, respectively (Fig. 2). By contrast, di- spectively, before death of host cells and formation of nospores harvested from A. sanguinea increased in new dinospores (Fig. 4, A and C). Comparable inocu- number by 25% during the first 16 h, suggesting lation of K. micrum resulted in a more gradual in- some cell division following size fractionation of the crease in parasitized hosts over 36 h, with subsequent 1 original culture. After 16 h, dinospores from A. san- values averaging 170 7 cells mL (T36-T44; n 3) guinea declined exponentially, falling below detection before the appearance of new dinospores (Fig. 4E). levels by day 5. Exponential curves fitted to data for By contrast, inoculation of cultures at high dino- dinospore abundance gave growth constants (k) of spore:host ratios resulted in maximum numbers of 1.020 0.149 (r2 0.9567; P 0.0001), 0.476 parasitized cells within 16 h for all three host species 0.070 (r2 0.9442; P 0.0001), and 0.257 0.0240 (r2 0.9572; P 0.0001) for Amoebophrya strains infecting G. instriatum, A. sanguinea, and K. micrum, respectively. The ability of dinospores to initiate infections differed considerably among the three Amoebophrya strains and varied dramatically with dinospore age (Fig. 3). Only 4.7 0.34% of recently formed dinospores from G. instriatum successfully invaded host cells, compared with 9.0 0.81% and 18.0 1.76% for dinospores from K. micrum and A. sanguinea, respectively. The percentage of harvested dinospores

Fig. 3. Ability of dinospores of different age to infect Fig. 2. Survival of Amoebophrya spp. dinospores from in- Akashiwo sanguinea, Karlodinium micrum, and Gymnodinium instri- fected Akashiwo sanguinea, Karlodinium micrum, and Gymnodinium atum. (A) Percent of harvested dinospores that successfully in- instriatum. Dinospores were harvested 6 h after formation, fected host cells. (B) Percent of surviving dinospores (i.e. those with abundance determined over time using epifluorescence detectable by epifluorescence microscopy) that established in- microscopy. Error bars indicate SE. fections. Error bars indicate SE. 524 D. WAYNE COATS AND MYUNG GIL PARK

Fig. 4. Time-course studies of Amoebophrya spp. infecting Akashiwo sanguinea, Gymnodinium instriatum, and Karlodinium micrum inoculated at high and low dinospore:host ratios. Bars show cumulative host abundance with white, gray, and black regions represent- ing uninfected hosts, early infections, and late infections, respectively. Open circles are dinospore abundance. Errors bars indicate SE.

(Fig. 4, B, D, and F), with the abundance of infected to complete the infection cycle before the end of the hosts averaging 1120 24 (n 8; T12-T48), 700 47 experiment, with high numbers of late stage infec- 1 (n 9; T12-T44), and 890 27 (n 8; T16-T44) tions (220 12 cells mL ) remaining in the final cellsmL1 for A. sanguinea (20:1), G. instriatum (115:1), sample. Infections occurred in the nucleus of A. san- and K. micrum (40:1), respectively, until host densities guinea, in the cytoplasm of G. instriatum, and in either began to decline. the nucleus or cytoplasm of K. micrum (Fig. 5). Host Early and late stage infections showed discrete cells exposed to high densities of dinospores (i.e. peaks during the first parasite generation in all treat- 20:1, 40:1, and 115:1 dinospore:host treatments) were ments except for the 40:1 inoculation of K. micrum infected by multiple dinospores. In those treatments, (Fig. 4, A–F). In that treatment, many parasites failed more than one late-stage parasite developed inside PARASITISM OF DINOFLAGELLATES 525

Fig. 5. Protargol-stained specimens showing location of Amoebophrya spp. inside host cells. Scale bars, 10 m. (a) Akashiwo sanguinea with intranuclear parasite (arrow). (b) Gymnodinium instriatum with intracytoplasmic parasite (arrow); host nucleus (N). (c) An uninfected Karlodinium micrum showing the nucleus (N), and an infected K. micrum with an intranuclear parasite (arrow). (d) Karlo- dinium micrum with intracytoplasmic parasite (arrow); host nucleus (N). (e) Karlodinium micrum with two parasites, one in the nucleus (white arrow) and one in the cytoplasm (black arrow). most cells of G. instriatum and K. micrum, but with very in all low dinospore:host inoculations and in the high rare exception, only a single mature parasite was gen- inoculum treatment for G. instriatum, thus permitting erated in A. sanguinea. calculation of total parasite generation time (i.e. time Parasite intracellular development time averaged required for infection of hosts, intracellular develop- 57 3.3 h, 46 1.1 h, and 34 3.8 h, respectively, ment, vermiform emergence, and extracellular matu- for G. instriatum, A. sanguinea and K. micrum cultures ration). The persistence of late stage infections and/ inoculated at a dinospore:host ratio of 1:1. One-way or lack of a clear peak in dinospore density prevented analysis of variance (ANOVA) revealed significant dif- accurate estimates of parasite generation time for ference among means (P 0.03), with pair-wise com- high dinospore:host inoculations of A. sanguinea and parisons indicating longer development time for in- K. micrum. Estimates for parasite generation time aver- fections of G. instriatum relative to those of K. micrum aged 71 0.6 h, 67 0.7 h, and 59 0.02 h in 1:1 (P 0.05; Tukey test). Estimates of parasite develop- treatments for G. instriatum, A. sanguinea, and K. mi- ment time for high dinospore inoculations of A. san- crum, respectively, with all means being significantly guinea (58 0.4 h; 20:1 dinospore:host ratio) were different (one-way ANOVA; Tukey test; P 0.05). consistently longer than for low inoculations, whereas Data for the 115:1 dinospore:host inoculation of G. in- those of G. instriatum (46 0.01 h; 115 dinospore: striatum indicated a significantly shorter parasite gen- host ratio) were consistently shorter than for corre- eration (55 2.5 h; P 0.025, t-test) than that for the sponding low inoculations. Comparison of data using corresponding 1:1 treatment. two-way ANOVA showed no significant difference for Cross infection. Dinospores of Amoebophrya sp. from host species or inoculum level (low versus high dino- K. micrum failed to infect A. sanguinea and G. instria- spore:host ratio) but did show a significant interactive tum, even though inoculum size (dinospore:host ratio effect (P 0.003), with pair-wise analysis indicating a of 50:1) was sufficient to produce high parasite preva- significant increase in development time for high ver- lence (93.1 1.22% from 12 h to 96 h after inocula- sus low inoculations of A. sanguinea (P 0.004) and a tion; n 8) in the primary host species (Fig. 6A). significant decrease in development time for high ver- Inoculation of G. instriatum with complementary dino- sus low inoculations of G. instriatum (P 0.004; Tukey spores (dinospore:host ratio of 50:1) generated infec- test). High dinospore inoculations also appeared to tion levels of 83%–98% during the first parasite gen- increase the infection cycle of Amoebophrya sp. in K. eration and 100% thereafter (Fig. 6B). The same micrum (cf. temporal patterns for late-stage infections; inoculum produced no infections in A. sanguinea but Fig. 4, E and F); however, the persistence of large did result in 26.5 4.50% infection of K. micrum at numbers of infected cells in that treatment prevented T12. Interestingly, parasite prevalence in K. micrum de- accurate determination of parasite intracellular devel- clined to zero over the following 36 h, and there was opment time. no indication that a second generation of infections Dinospore densities exhibited distinct peaks or developed in subsequent samples. Similarly, inocula- reached stationary levels by the end of the experiment tion of A. sanguinea with complementary dinospores 526 D. WAYNE COATS AND MYUNG GIL PARK

those infections was not determined (Coats et al. 1996, Yih and Coats 2000). Multiple infection of G. instriatum and K. micrum in our high dinospore:host treatments resulted in the formation of more than one “beehive” stage, with parasites restricted to the cytoplasm of G. instriatum but occurring simultaneously in the cytoplasm and nucleus of K. micrum. Interest- ingly, infection of A. sanguinea by multiple dinospores produced a single mature parasite, suggesting either that one of the trophonts became dominant and prevented growth of the others or that several trophonts fused to form a syncytium before maturation of the parasite. Although sexuality has not been reported for Amoebophrya spp., development of a syncytium might provide a mechanism for genetic recombination via fusion of nuclei from different dinospores. Infectivity and survival of dinospores exhibited distinct differences among our three strains of Amoe- bophrya. Recently formed dinospores from A. sanguinea were roughly twice as likely to infect host cells as were dinospores from K. micrum and four times more likely than dinospores from G. instriatum. Fur- thermore, dinospores from G. instriatum and A. sanguinea were short lived and unable to initiate infections after 2–5 days, respectively, whereas dinospores from K. micrum maintained high infectivity for 4–5 days, with some being able to parasitize hosts after 11 days. The prolonged survival of dinospores from K. micrum is surprising, given that Amoebophrya spp. are thought to be strictly parasitic. Our observations sug- Fig. 6. Infection of Akashiwo sanguinea, Gymnodinium sanguineum, and Karlodinium micrum when exposed to dinospores gest that dinospores of Amoebophrya sp. from K. micrum from three strains of Amoebophrya. Dinospore inoculations were may be able to use bacteria and/or dissolved organic adjusted to give 80%–90% infection levels in the primary host substances as a source of nutrition or may be able to species. Error bars indicate SE. form resting stages that become active after a short period of dormancy. Susceptibility of host cells to infection also differed (inoculum size 10:1) produced high parasite preva- among the three host–parasite systems. Akashiwo san- lence (91.8 1.41% until T72) that increased to 100% guinea and G. instriatum appeared completely suscepti- in the second parasite generation (Fig. 6C). Equiva- ble to parasitism, as high dinospore:host treatments lent inoculations did not stimulate parasitism of G. in- showed 100% infection levels. By contrast, approxi- striatum but generated low infection levels (5%) in mately 10% of the K. micrum cells remained uninfected K. micrum. Parasites that formed in K. micrum from di- at intermediate to high dinospore:host treatments. Al- nospores derived from A. sanguinea also appear un- though we cannot rule out the possibility that higher in- able to produce a second generation of infections. oculations would have produced 100% parasite prevalence, available data suggest that some specimens of K. discussion micrum are resistant to infection. The source of this ap- Cachon (1964) noted that Amoebophrya ceratii formed parent resistance is unknown but may be related to cell cytoplasmic infections in athecate dinoflagellates and cycle events that render K. micrum unattractive to dino- nuclear infections in all thecate species, except Proro- spores or capable of fending off invading parasites. centrum micans Ehrenberg. Subsequent reports dem- Observed differences in endurance and infectivity onstrated that nuclear infections can occur in some of dinospores among Amoebophrya strains, along with athecate dinoflagellates but continued to supported discrepancies in the susceptibility of host species to in- the idea that infections were site specific within each fection, suggest that the three host–parasite systems host species (Coats and Bockstahler 1994, Coats et al. have evolved somewhat divergent survival strategies. 1996). Amoebophrya sp. from K. micrum is a clear excep- Although A. sanguinea and G. instriatum appear wholly tion to the general pattern of parasites being specific susceptible to infection, dinospores of their parasites to particular regions of host cells, as infections oc- are short lived and quickly loose the ability to infect curred in both the nucleus and cytoplasm. Prior stud- host cells. Thus, these strains of Amoebophrya seem ca- ies have also shown that individual host cells can be pable of rapidly exploiting dense host populations but infected by multiple dinospores; however, the fate of less likely to maintain high infection levels under non- PARASITISM OF DINOFLAGELLATES 527 bloom conditions, due to limited survival and time tively, K. micrum may be able to successfully mount a constraints in encountering host cells. By contrast, di- defense against dinospores from the other two host nospores of Amoebophrya sp. from K. micrum are very species. As mentioned above, some cells of K. micrum long lived, providing increased opportunity to find appear resistant to dinospores from their correspond- hosts cells even when present at low densities. None- ing Amoebophrya sp. Perhaps all K. micrum have the theless, persistence of K. micrum in the face of high ability to defend against invading parasites to some parasite prevalence seems assured, as some host cells degree, with the response simply being more success- appear resistant to infection. ful when mounted against dinospores from a “for- The life cycle of Amoebophrya sp. consists of an intra- eign” Amoebophrya sp. cellular phase during which parasites grow to maturity Amoebophrya strains cultured in association with A. inside host cells and an extracellular phase in which sanguinea, G. instriatum, and K. micrum exhibited marked the mature parasite emerges from the host cell, di- differences in host–parasite biology. These differences vides to produce dinospores, and then infects new along with results of our cross-infections experiments hosts. Earlier studies have estimated the duration of and reported genetic divergence among the parasite the intracellular phase of Amoebophrya sp. ex A. san- strains (Gunderson et al. 2000) support the conten- guinea and provided information on the timing of tion that Amoebophrya ceratii is a species complex con- dinospore formation but have not determined the en- sisting of several host-specific taxa (Coats et al. 1996). tire generation time of the parasite (Coats and Bock- Host specificity may make Amoebophrya spp. more at- stahler 1994, Yih and Coats 2000). Data for our three tractive than nonspecific parasites like Parvilucifera in- host–parasite systems indicate that the intracellular fectans (Erard-Le Denn et al. 2000) in the biological phase only represents 60%–80% of total parasite gen- control of bloom-forming dinoflagellate. eration time. In addition, the duration of the intracellular phase appeared to be influenced by parasite We thank Mr. Sean Cooney for technical assistance, Dr. Aishao load. Inoculation of cultures at a 1:1 dinospore:host Li for the original culture of Karlodinium micrum, Dr. Wonho ratio typically generated one parasite per infected host Yih for critical comments, and Dr. Øjvind Moestrup for assis- cell, with intracellular development time averaging 34 tance in identification of host species and guidance in di- 3.8 h, 46 1.1 h, 57 3.3 h in K. micrum, A. sanguinea, noflagellate nomenclature. Supported by NSF award OCE- and G. instriatum, respectively. Inoculation of cultures 9730695 to D. W. Coats. at high dinospore:host ratios stimulated infection of host cells by several dinospores, with parasite genera- Anderson, D. M. 1997. Turning back the harmful red tide. Nature 388:513–4. tion time in A. sanguinea increasing by 12 h (to 58 Cachon, J. 1964. Contribution à l’étude des péridiniens parasites. 0.4 h) but that of G. instriatum decreasing by 11 h (to Cytologie, cycles évolutifs. Ann. Sci. Nat. Zool. 6:1–158. 46 0.01 h). High dinospore inoculations also result Cachon, J. & Cachon, M. 1987. Parasitic dinoflagellates. In Taylor, in a significantly shorter total parasite generation time F. J. R. [Ed.] The Biology of Dinoflagellates. Blackwell Science, Ox- for G. instriatum infections relative to low dinospore: ford, pp. 571–610. Canter, H. M. & Heaney, S. I. 1984. Observations on zoosporic host treatments. That multiple infection by dino- fungi of Ceratium spp. in lakes of the English Lake District; im- spores had opposite effects on intracellular develop- portance for phytoplankton population dynamics. New Phytol. ment times of A. sanguinea and G. instriatum may be 97:601–12 related to corresponding differences in parasite inter- Coats, D. W. 1999. Parasitic life styles of marine dinoflagellates. J. Euk. Microbiol. 46:402–9. action. Competition between invading dinospores or Coats, D. W. & Bockstahler, K. R. 1994. Occurrence of the parasitic fusion of trophonts into a syncytium may retard para- dinoflagellate Amoebophrya ceratii in Chesapeake Bay popula- site maturation in A. sanguinea, whereass independent tions of Gymnodinium sanguineum. J. Euk. Microbiol. 41:586–93. growth of multiple parasites in G. instriatum may re- Coats, D. W., Adam, E. J., Gallegos, C. L. & Hedrick, S. 1996. Para- sult in more rapid utilization of host resources. sitism of photosynthetic dinoflagellates in a shallow subestuary of Chesapeake Bay, USA. Aquat. Microb. Ecol. 11:1–9. Amoebophrya strains from A. sanguinea, G. instriatum, Coats, D. W. & Heinbokel, J. F. 1982. A study of reproduction and and K. micrum either failed to establish infections or other life cycle phenomena in planktonic protists using an were unable to successfully complete their life cycle in acridine orange fluorescence technique. Mar. Biol. 67:71–9. the presence of nonprimary host species. Results are Daugbjerg, D. Hansen, G., Larsen, J. & Moesrup, Ø. 2000. Phylog- eny of some of the major genera of dinoflagellates based on ul- consistent with the parasites being host specific and trastructure and partial LSU rDNA sequence data, including also suggest that host specificity in Amoebophrya strains the erection of three new genera of unarmoured dinoflagel- involves more than parasite recognition of host cells, lates. Phycologia 39:302–17. as Amoebophrya spp. from A. sanguinea and G. instria- Doucette, G. J., Kodama, M., Franca, S. & Gallacher, S. 1998. Bacte- tum were able to invade K. micrum but did not develop rial interactions with harmful algal bloom species: bloom ecology, toxigenesis, and cytology. In Anderson, D. M., Cembella, to maturity. Karlodinium micrum is considered to be a A. D. & Hallegraeff, G. M. [Eds.] Physiological Ecology of Harmful toxic dinoflagellate (Li et al. 2000) and thus may rep- Algal Blooms. NATO ASI Series G: Ecological Sciences, Vol. 41, resent a very different environment for Amoebophrya Springer-Verlag, Heidelberg, pp. 619–47. spp. than the other two nontoxic host species. Dino- Elbrächter, M. & Schnepf, E. 1998. Parasites of harmful algae. In Anderson, D. M., Cembella, A. D. & Hallegraeff, G. M. [Eds.] spores of Amoebophrya strains from A. sanguinea and G. Physiological Ecology of Harmful Algal Blooms. NATO ASI Series G: instriatum that did invade K. micrum may have suc- Ecological Sciences, Vol. 41, Springer-Verlag, Heidelberg, pp. cumbed to inappropriate growth conditions. Alterna- 351–63. 528 D. WAYNE COATS AND MYUNG GIL PARK

Erard-Le Denn, E., Chrétiennot-Dinet, J.-M. & Probert, I. 2000. Nishitani, L. & Chew, K. K. 1984. Recent developments in paralytic First report of parasitism on the toxic dinoflagellates Alexan- shellfish poisoning research. Aquaculture 39:317–29. drium minutum Halim. Estuar. Coast. Shelf Sci. 50:109–13. Nishitani, L., Erickson, G. & Chew, K. K. 1985. Role of the parasitic Freudenthal, H. D. & Lee, J. J. 1963. Glenodinium halli n. sp. and Gy- dinoflagellate Amoebophrya ceratii in control of Gonyaulax catenella rodinium instriatum n. sp., dinoflagellates from New York wa- populations. In Anderson, D. M., White, A. W. & Baden, D. G. ters. J. Protozool. 10:182–9. [Eds.] Toxic Dinoflagellates Elsevier, New York, pp. 225–30. Galigher, A. E. & Kozloff, E. N. 1971. Essentials of Practical Microtech- Norén, F., Moestrup, Ø. & Rehnstam-Holm, A.-S. 1999. Parvilucifera nique, 2nd ed. Lea & Febiger, Philadelphia, 531 pp. infectans Norén et Moestrup gen. et sp. nov. (Perkinsozoa phy- Guillard, R. R. L. & Ryther, J. H. 1962. Studies on marine plank- lum nov.): a parasitic flagellate capable of killing toxic microal- tonic diatoms. I. Cyclotella nana Hustedt and Detonula conferva- gae. Eur. J. Protistol. 35:233–54. cea (Cleve) Gran. Can. J. Microbiol. 8:229–39. Park, M. G., Cooney, S. K., Yih, W. H. & Coats, D. W. 2002. Effects Gunderson, J. H., Goss, S. H. & Coats, D. W. 2000. rRNA sequence of two strains of the parasitic dinoflagellate Amoebophrya on differences among Amoebophrya strains infecting dinoflagellates growth, photosynthesis, light absorption, and quantum yield of in Chesapeake Bay. J. Euk. Microbiol. 47(2):4A. bloom-forming dinoflagellates. Mar. Ecol. Prog. Ser. 227:281–92 Heaney, S. I., Lund, J. W. G., Canter, H. M. & Gray, K. 1988. Popu- Silva, E. S. & Franca, S. 1985. The association dinoflagellate-bacte- lation dynamics of Ceratium spp. in three English lakes, 1945– ria: their ultrastructural relationship in two species of di- 1985. Hydrobiol. 161:133–48. noflagellates. Protistologica 21:429–46. Kirchner, M., Sahling, G., Schütt, C., Döpke, H & Uhlig, G. 1999. Sommer, U., Wedemeyer, C. & Lowsky, B. 1984. Comparison of po- Intracellular bacteria in the red tide-forming heterotrophic di- tential growth rates of Ceratium hirundinella with observed pop- noflagellate Noctiluca scintillans. Arch. Hydrobiol. Spec. Issues Ad- ulation density changes. Hydrobiologia 109:159–64. vanc. Limnol. 54:297–310. Starr, R. C. & Zeikus, J. A. 1993. UTEX—The culture collection of al- Li, A., Stoecker, D. K. & Coats, D. W. 2000. Spatial and temporal as- gae at the University of Texas at Austin. J. Phycol. 29(suppl.):94. pects of Gyrodinium galatheanum in Chesapeake Bay: distribu- Tarutani, K., Nagasaki, K., Itakura, S. & Yamaguchi, M. 2001. Isola- tion and mixotrophy. J. Plank. Res. 22:2105–24. tion of a virus infecting the novel shellfish-killing dinoflagel- Maranda, L. 2001. Infection of Prorocentrum minimum (Dino- late Heterocapsa circularisquama. Aquat. Microb. Ecol. 23:103–11. phyceae) by the parasite Amoebophrya sp. (Dinoflagellea). J. Taylor, F. J. R. 1968. Parasitism of the toxin-producing dinoflagel- Phycol. 37:245–8. late Gonyaulax catenella by the endoparasitic dinoflagellate Montagnes, D. J. S. & Lynn, D. H. 1993. A quantitative protargol Amoebophrya ceratii. J. Fish. Res. Bd. Can. 25:2241–5. stain (QPS) for ciliates and other protists. In Kemp, P. F., Yih, W. & Coats, D. W. 2000. Infection of Gymnodinium sanguineum Sherr, B. F., Sherr, E. B. & Cole, J. J. [Eds.] Handbook of Methods by the dinoflagellate Amoebophrya sp.: effect of nutrient envi- in Aquatic Microbial Ecology. Lewis Publishers, Boca Raton, pp. ronment on parasite generation time, reproduction, and infec- 229–40. tivity. J. Euk. Microbiol. 47:504–10.