Occurrence of the Parasitic Dinoflagellate Amoebophrya ceratii in Chesapeake Bay Populations of Gymnodinium sanguineum

Chesapeake Bay populations of the red-tide dinoflagellate Gymnodinium sanguineum were regularly infected by the parasitic dinoflagellate Amoebophrya ceratii during the summers of 1988-1991. Infections developed inside the nucleus of G. sanguineum and were always lethal to the host. Parasite generation time was ~ 40 h at 23° C, with the intracellular, trophont phase lasting 39.5 ± 0.3 h, and the extracellular, vermiform stage persisting for ~ 20 min. Near surface accumulations of G. sanguineum sometimes exceeded 1,000 cells/ml; however, host abundance was relatively low when integrated over the surface mixed layer of each station (mean = 12.2 cells/ml ± 2.96 SE; n = 60). Parasitized hosts were encountered in 75% of the samples where host abundance was > 1 per ml, and epidemic outbreaks (20-40% hosts infected) were observed on several occasions. Epidemic infections were generally located several meters below surface accumulations of G. sanguineum and were always restricted to a narrow region near the pycnocline. Consequently, integrated station values for parasite prevalence were low, with an average 2.7% (± 0.31 SE; n = 60). Parasite induced mortality removed up to 8% of G. sanguineum populations per day, but averaged < 2% of host biomass throughout the Bay. Thus, parasitism by A. ceratii does not appear to be a major factor regulating G. sanguineum bloom in the main stem of Chesapeake Bay. Supplementary key words. Infection level, parasitism, population dynamics, red-tide. THE heterotrophic dinoflagellate Amoebophrya ceratii is an obligate, intracellular parasite of other dinoflagellates that is known to occur in coastal waters of the North Atlantic [1,8, 9, 12], the North Pacific [14, 15], and the Mediterranean Sea [2]. A. ceratii is not a host-specific parasite, as infections have been reported in species of more than 15 dinoflagellate genera including a few toxic forms [2, 4, 11, 12, 14]. Dinoflagellates infected by A. ceratii become reproductively incompetent [8, 14] and are eventually killed by the parasite. The lelhal nature of A. ceratii and its high infection levels in some host species have led to the suggestion that this parasite might serve as an agent for the biological control for toxic dinoflagellate blooms [15, 16]. Amoebophrya ceratii infections are initiated when dinospores, the biflagellated dispersal stage of the parasite, attach to the outer membrane and penetrate into the cytoplasm of the host. As many as 12 dinospores can attack a single host, but rarely does more than one reach maturity [2]. Once inside the host, To whom correspondence should be addressed. the parasite differentiates into the trophont (vegetative) stage that may remain in the cytoplasm or may invade the host's nucleus. In either case, the parasite soon increases in size and begins a series of nuclear divisions that are accompanied by flagellar replication, but proceed without cytoplasmic fission. With continued growth, the anterior portion of the parasite (the episome) is partially enclosed by the posterior hyposome to form a flagellar cavity, the mastigocoel [3]. Thus, trophonts of late infections are large, polynucleate and multiflagellate organisms that occupy most of the host cell and have a characteristic "beehive" [9] appearance. At maturity, the trophont ruptures through the host's pellicle and transforms into a strongly motile vermiform stage that divides to produce numerous infective dinospores. Cachon [2] reported that Amoebophrya ceratii occurred sporadically in a number of Mediterranean host species, but highest infection levels usually occurred near the end of dinoflagellate blooms. Heavy infections (30-40% in Gonyaulax cantenella; 80% in Ceratium fusus) were also observed along the western coast of North America, where parasitism by A. ceratii was linked to rapid declines in host populations and implicated as COATS & BOCKSTAHLER-PARASITIC DINOFLAGELLATES 587 an important factor in preventing bloom formation [14, 15]. By contrast, Fritz and Nass [9] found that < 2% of Dinophysis norvegica collected from coastal waters of Nova Scotia were infected by A. ceratii and argued that parasitism had little effect on host populations. The same study reported infection levels of 50% in Scrippsiella trochoidea., To reliably assess the effect of Amoebophrya ceratii on host populations requires recognition of all parasite life history stages and knowledge of parasite development time. Heavy infections would be of little consequence if the duration of infection greatly exceeded host generation time. Were the reverse true, then even low infection levels could have a significant influence on host populations. Thus far, only Nishitani et al. [14] have estimated parasite prevalence using techniques that reveal all intracellular stages of the parasite, and data on development time of A. ceratii have been unavailable. Here we present data on the occurrence of Amoebophrya ceratii in Chesapeake Bay populations of Gymnodinium sanguineum over a four-year period. We also provide information on parasite morphology and development time, and estimate the impact of A. ceratii on this host species. MATERIALS AND METHODS Field samples for documentation of temporal and spatial occurrence of Amoebophrya ceratii in Chesapeake Bay populations of Gymnodinium sanguineum were collected during cruises conducted at approximately monthly intervals between June and October of 1988-1991, On each of 15 cruises, vertical CTD (conductivity-temperature-depth) -Niskin bottle casts were taken at 10 stations located along the longitudinal axis of the Bay (Fig. 1). CTD profiles provided data on conductivity and temperature, which defined the depth of the pycnpcline and facilitated sampling within the surface mixed layer. The number and vertical position of Niskin samples depended on the depth and stratification of the water column, but six to eight were taken at most stations with the upper four to five positioned at 2 to 3-meter intervals between the surface and the pycnocline. A 500-mI subsample was taken from each Niskin bottle and preserved in modified Bouin's solution [15]. Estimates of Gymnodinium sanguineum abundance were obtained by enumerating cells in 1-ml aliquots of Bouin's preserved whole-water sample using Sedgwick-Rafter chambers and a Zeiss microscope (x 125). Samples that contained > 1 dinoflagellate/ml were processed by the quantitative protargol staining (QPS) technique [13] and examined at 500 x using Zeiss oilimmersion optics. Parasite prevalence (i.e. percent G. sanguineum infected by Amoebophrya ceratii) was determined by examining > 50 (usually 100) protargol-stained hosts per sample, and the number of infected hosts per ml was calculated as host abundance multiplied by parasite prevalence. Integrated values for the number of Gymnodinium sanguineum per ml and number infected hosts/ml were generated for the surface mixed layer of each station as depth-weighted averages of data for samples taken between the surface and the greater of two depths* represented by either the pycnocline or the deepest sample containing > 1 G. sanguineum per ml. Integrated data were used to calculate station values for parasite prevalence, and parasite induced mortality of G. sanguineum was estimated as: Proportion of host's population killed per day _ integrated parasite prevalence