Characterization of the Parasitic Dinoflagellate Amoebophrya sp. Infecting Akashiwo sanguinea in Coastal Waters of China.

The endoparasitic dinoflagellate Amoebophrya infects a number of free-living marine dinoflagellates, including harmful algal bloom species. The parasitoid eventually kills its host and has been proposed to be a significant loss factor for dinoflagellate blooms in restricted coastal waters. For several decades, the difficulties of culturing host-parasitoid systems have been a great obstacle for further research on the biology of Amoebophrya. Here, we established an Akashiwo sanguinea-Amoebophrya sp. coculture from Chinese coastal waters and studied the parasitoid's generation time, dinospore survival and infectivity, as well as its host specificity. The lifespan of Amoebophrya sp. ex. A. sanguinea was approximately 58 h. The infective dinospores can survive up to 78 h in ambient waters but gradually lose their infectivity. The parasitoid was unable to infect other dinoflagellate species, its infection rate reached as high as 91% when the ratio of dinospores to host cells was 20:1. The high infectivity of dinospores suggests that the Amoebophrya strain was capable of removing a considerable fraction of host biomass within a short period, but that it is probably unable to maintain high infection levels under nonbloom conditions of its host, due to limited survival and time constraints in encountering host cells.

Transcriptomic profiling of Alexandrium fundyense during physical interaction with or exposure to chemical signals from the parasite Amoebophrya.

Toxic microalgae have their own pathogens, and understanding the way in which these microalgae respond to antagonistic attacks may provide information about their capacity to persist during harmful algal bloom events. Here, we compared the effects of the physical presence of the parasite Amoebophrya sp. and exposure to waterborne cues from cultures infected with this parasite, on gene expression by the toxic dinoflagellates, Alexandrium fundyense. Compared with control samples, a total of 14,882 Alexandrium genes were differentially expressed over the whole-parasite infection cycle at three different time points (0, 6 and 96 h). RNA sequencing analyses indicated that exposure to the parasite and parasitic waterborne cues produced significant changes in the expression levels of Alexandrium genes associated with specific metabolic pathways. The observed upregulation of genes associated with glycolysis, the tricarboxylic acid cycle, fatty acid ß-oxidation, oxidative phosphorylation and photosynthesis suggests that parasite infection increases the energy demand of the host. The observed upregulation of genes correlated with signal transduction indicates that Alexandrium could be sensitized by parasite attacks. This response might prime the defence of the host, as indicated by the increased expression of several genes associated with defence and stress. Our findings provide a molecular overview of the response of a dinoflagellate to parasite infection.

Genomic insights into processes driving the infection of Alexandrium tamarense by the Parasitoid Amoebophrya sp.

The regulatory circuits during infection of dinoflagellates by their parasites are largely unknown on the molecular level. Here we provide molecular insights into these infection dynamics. Alexandrium tamarense is one of the most prominent harmful algal bloom dinoflagellates. Its pathogen, the dinoflagellate parasitoid Amoebophrya sp., has been observed to infect and control the blooms of this species. We generated a data set of transcripts from three time points (0, 6, and 96 h) during the infection of this parasite-host system. Assembly of all transcript data from the parasitoid (>900,000 reads/313 Mbp with 454/Roche next-generation sequencing [NGS]) yielded 14,455 contigs, to which we mapped the raw transcript reads of each time point of the infection cycle. We show that particular surface lectins are expressed at the beginning of the infection cycle which likely mediate the attachment to the host cell. In a later phase, signal transduction-related genes together with transmembrane transport and cytoskeleton proteins point to a high integration of processes involved in host recognition, adhesion, and invasion. At the final maturation stage, cell division- and proliferation-related genes were highly expressed, reflecting the fast cell growth and nuclear division of the parasitoid. Our molecular insights into dinoflagellate parasitoid interactions point to general mechanisms also known from other eukaryotic parasites, especially from the Alveolata. These similarities indicate the presence of fundamental processes of parasitoid infection that have remained stable throughout evolution within different phyla.

Amoebophrya spp. from the bloom-forming dinoflagellate Cochlodinium polykrikoides: parasites not nested in the "Amoebophrya ceratii complex".

Members of Amoebophrya ceratii complex are known to infect a number of free-living dinoflagellates including harmful algal bloom species. In August and October 2012, Amoebophrya infections during two bloom events of the dinoflagellate Cochlodinium polykrikoides were observed along southern coastal waters of Korea. Microscopic observations and molecular data revealed that two different Amoebophrya parasites infected the same host species. In addition, while one developed in the host's nucleus, the other in the host's cytoplasm. Phylogenetic analyses showed that the two parasites were not nested in the previously recognized "Amoebophrya ceratii complex clade", which contained sequences of parasites infecting numerous dinoflagellate species. Instead, they branched as sister taxa to the isolate (possibly Amoebophrya) from radiolarians Hexacontium gigantheum. Our result indicates that the two Amoebophrya parasites infecting C. polykrikoides may be different species from those inside the "complex."

Comparative biological traits of perkinsozoan parasitoids infecting marine dinoflagellates.

The number of perkinsozoan parasitoid species known to infect dinoflagellates has increased to 11 over the last two decades. However, most of the current knowledge about the autecology of perkinsozoan parasitoids of dinoflagellates has derived from studies of one or two species, thereby making it difficult to directly compare their biological traits at the same time and even their potentials as biological control agents if they are to be exploited to mitigate harmful dinoflagellate blooms in the field. This study investigated total generation time, the number of zoospores produced per sporangium, zoospore size, swimming speed, parasite prevalence, zoospore survival and success rate, and host range and susceptibility for five perkinsozoan parasitoids. Four of the species (Dinovorax pyriformis, Tuberlatum coatsi, Parvilucifera infectans, and P. multicavata) were from the family Parviluciferaceae and one (Pararosarium dinoexitiosum) was from the family Pararosariidae, with dinoflagellate Alexandrium pacificum employed as a common host. Distinct differences in the biological traits of the five perkinsozoan parasitoid species were found, suggesting that the fitness of these parasitoids for the common host species differs. These results thus offer useful background information for the understanding of the impacts of parasitoids on the natural host population and for the design of numerical modeling including the host-parasitoid systems and biocontrol experiments in the field.

Confirmation of the wide host range of Parvilucifera corolla (Alveolata, Perkinsozoa).

Marine parasites of the genus Parvilucifera have been described as endoparasitoids of dinoflagellates. Recently, the species Parvilucifera corolla was described, but its host range was not examined. Here, the host selectivity of P. corolla was screened, including 110 strains of dinoflagellates (24 genera) and other microalgal groups as potential hosts. Infections and the full life cycle of the parasitoid were observed in 73 strains (16 genera) of dinoflagellates. Parvilucifera corolla did not infect most chlorophytes, cryptophytes, chrysophytes, diatoms, haptophytes and raphidophytes but one strain of Pyramimonas (chlorophyte) was infected, although without viable sporangia. In Symbiodinium natans, a transition to the coccoid stage was induced above a certain parasite:host ratio. These results confirm P. corolla as a generalist parasitoid of dinoflagellates, with important differences in host range regarding other species of the genus.

Taxonomy and Molecular Phylogenetics of Ensiculiferaceae, fam. nov. (Peridiniales, Dinophyceae), with Consideration of their Life-history.

In the current circumscription, the Thoracosphaeraceae comprise all dinophytes exhibiting calcified coccoid cells produced during their life-history. Species hitherto assigned to Ensiculifera and Pentapharsodinium are mostly based on the monadoid stage of life-history, while the link to the coccoid stage (occasionally treated taxonomically distinct) is not always resolved. We investigated the different life-history stages and DNA sequence data of Ensiculifera mexicana and other species occurring in samples collected from all over the world. Based on concatenated ribosomal RNA gene sequences Ensiculiferaceae represented a distinct peridinalean branch, which showed a distant relationship to other calcareous dinophytes. Both molecular and morphological data (particularly of the coccoid stage) revealed the presence of three distinct clades within Ensiculiferaceae, which may include other dinophytes exhibiting a parasitic life-history stage. At a higher taxonomic level, Ensiculiferaceae showed relationships to parasites and endosymbionts (i.e., Blastodinium and Zooxanthella) as well as to dinophytes harbouring diatoms instead of chloroplasts. These unexpected phylogenetic relationships are corroborated by the presence of five cingular plates in all such taxa, which differs from the six cingular plates of most other Thoracosphaeraceae. We herein describe Ensiculiferaceae, emend the descriptions of Ensiculifera and Pentapharsodinium, erect Matsuokaea and provide several new combinations at the species level.

Parvilucifera multicavata sp. nov. (Alveolata, Perkinsozoa), a New Parasitoid Infecting Marine Dinoflagellates Having Abundant Apertures on the Sporangium.

The phylum Perkinsozoa is known as an exclusively parasitic group including the parasites of shellfish, fish, dinoflagellates, cryptophytes, and tadpoles and at present comprises seven genera across three families (Parviluciferaceae, Perkinsidae, and Xcellidae), with the genus Parvilucifera having the most abundant species in the family Parviluciferaceae. During intensive sampling along the Korean coast in August and September 2017, a new species of the genus Parvilucifera was discovered and successfully established in cultures. Morphological and ultrastructural observations revealed that the new parasitoid shares almost all known diagnostic characters with other species of Parvilucifera, except that its sporangium has a higher number of apertures although with smaller diameters than those in P. infectans. Molecular phylogenetic trees based on both nuclear small subunit (SSU) and concatenated SSU and large subunit (LSU) ribosomal DNA (rDNA) sequences revealed that the new parasitoid was nested within the family Parviluciferaceae and had a sister relationship with P. infectans. Based on morphological, ultrastructural, and molecular data, we propose to erect a new species, P. multicavata sp. nov., for the new parasitoid found in this study.

Expressed sequence tags from Amoebophrya sp. infecting Karlodinium veneficum: comparing host and parasite sequences.

Parasitic dinoflagellates of the genus Amoebophrya play important roles in the ecology of estuaries and open ocean environments. Little is known of the cell and molecular biology of Amoebophrya, but the genus is intermediate on phylogenetic trees between apicomplexans and typical dinophycean dinoflagellates. Here, we constructed four cDNA libraries, from different stages after infecting the host, Karlodinium veneficum, with Amoebophrya sp. These libraries were used to generate 898 expressed sequence tags (ESTs), with sequences attributed to either the host or parasite, based on AT bias, codon usage, and occurrence during infection. Overall, 209 sequences were attributable to the parasite and 685 to the host. The 50 putative parasite sequences with good protein matches in GenBank were used to find the same protein from host ESTs. For 26 genes, both host and parasite sequences were identified, of which 20 encoded ribosomal proteins. PCR for seven predicted parasite and two host genes were used to confirm attributions. The most common host and parasite ESTs were compared to see if multiple gene copies were present. The host plastocyanin gene had multiple sequence variants, but parasite rps27a contained only one polymorphism, likely due to an amplification error. Amplification, cloning, and sequencing of five parasite protein-coding genes suggested that the parasite has a single sequence for each gene, but three host genes were found to have multiple variants. The genome of Amoebophrya sp. infecting K. veneficum appears to have an organization more similar to other eukaryotes than to the tandem gene arrangements found in dinoflagellates.

Tuberlatum coatsi gen. n., sp. n. (Alveolata, Perkinsozoa), a New Parasitoid with Short Germ Tubes Infecting Marine Dinoflagellates.

Perkinsozoa is an exclusively parasitic group within the alveolates and infections have been reported from various organisms, including marine shellfish, marine dinoflagellates, freshwater cryptophytes, and tadpoles. Despite its high abundance and great genetic diversity revealed by recent environmental rDNA sequencing studies, Perkinsozoa biodiversity remains poorly understood. During the intensive samplings in Korean coastal waters during June 2017, a new parasitoid of dinoflagellates was detected and was successfully established in culture. The new parasitoid was most characterized by the presence of two to four dome-shaped, short germ tubes in the sporangium. The opened germ tubes were biconvex lens-shaped in the top view and were characterized by numerous wrinkles around their openings. Phylogenetic analyses based on the concatenated SSU and LSU rDNA sequences revealed that the new parasitoid was included in the family Parviluciferaceae, in which all members were comprised of two separate clades, one containing Parvilucifera species (P. infectans, P. corolla, and P. rostrata), and the other containing Dinovorax pyriformis, Snorkelia spp., and the new parasitoid from this study. Based on morphological, ultrastructural, and molecular data, we propose to erect a new genus and species, Tuberlatum coatsi gen. n., sp. n., from the new parasitoid found in this study. Further, we examined and discussed the validity of some diagnostic characteristics reported for parasitoids in the family Parviluciferaceae at both the genus and species levels.

A precedented nuclear genetic code with all three termination codons reassigned as sense codons in the syndinean Amoebophrya sp. ex Karlodinium veneficum.

Amoebophrya is part of an enigmatic, diverse, and ubiquitous marine alveolate lineage known almost entirely from anonymous environmental sequencing. Two cultured Amoebophrya strains grown on core dinoflagellate hosts were used for transcriptome sequencing. BLASTx using different genetic codes suggests that Amoebophyra sp. ex Karlodinium veneficum uses the three typical stop codons (UAA, UAG, and UGA) to encode amino acids. When UAA and UAG are translated as glutamine about half of the alignments have better BLASTx scores, and when UGA is translated as tryptophan one fifth have better scores. However, the sole stop codon appears to be UGA based on conserved genes, suggesting contingent translation of UGA. Neither host sequences, nor sequences from the second strain, Amoebophrya sp. ex Akashiwo sanguinea had similar results in BLASTx searches. A genome survey of Amoebophyra sp. ex K. veneficum showed no evidence for transcript editing aside from mitochondrial transcripts. The dynein heavy chain (DHC) gene family was surveyed and of 14 transcripts only two did not use UAA, UAG, or UGA in a coding context. Overall the transcriptome displayed strong bias for A or U in third codon positions, while the tRNA genome survey showed bias against codons ending in U, particularly for amino acids with two codons ending in either C or U. Together these clues suggest contingent translation mechanisms in Amoebophyra sp. ex K. veneficum and a phylogenetically distinct instance of genetic code modification.

Description, host-specificity, and strain selectivity of the dinoflagellate parasite Parvilucifera sinerae sp. nov. (Perkinsozoa).

A new species of parasite, Parvilucifera sinerae sp. nov., isolated from a bloom of the toxic dinoflagellate Alexandrium minutum in the harbor of Arenys de Mar (Mediterranean Sea, Spain), is described. This species is morphologically, behaviourally, and genetically (18S rDNA sequence) different from Parvilucifera infectans, until now the only species of the genus Parvilucifera to be genetically analyzed. Sequence analysis of the 18S ribosomal DNA supported P. sinerae as a new species placed within the Perkinsozoa and close to P. infectans. Data on the seasonal occurrence of P. sinerae, its infective rates in natural and laboratory cultures, and intra-species strain-specific resistance are presented. Life-cycle studies in field samples showed that the dinoflagellate resting zygote (resting cyst) was resistant to infection, but the mobile zygote (planozygote) or pellicle stage (temporary cyst) became infected. The effects of light and salinity levels on the growth of P. sinerae were examined, and the results showed that low salinity levels promote both sporangial germination and higher rates of infection. Our findings on this newly described parasite point to a complex host-parasite interaction and provide valuable information that leads to a reconsideration of the biological strategy to control dinoflagellate blooms by means of intentional parasitic infections.

Ultrastructure of a novel tube-forming, intracellular parasite of dinoflagellates: Parvilucifera prorocentri sp. nov. (Alveolata, Myzozoa).

We have characterized the intracellular development and ultrastructure of a novel parasite that infected the marine benthic dinoflagellate Prorocentrum fukuyoi. The parasite possessed a combination of features described for perkinsids and syndineans, and also possessed novel characters associated with its parasitic life cycle. Reniform zoospores, about 4 microm long, possessed a transverse flagellum, alveoli, a refractile body, a mitochondrion with tubular cristae, a syndinean-like nucleus with condensed chromatin, micronemes, bipartite trichocysts with square profiles (absent in perkinsids) and oblong microbodies. Like Parvilucifera, the zoospores also possessed a shorter posterior flagellum, a heteromorphic pair of central microtubules in the anterior axoneme and a reduced pseudoconoid positioned directly above an orthogonal pair of basal bodies. Early developmental stages consisted of a sporangium about 5-15 microm in diam that contained spherical bodies and amorphous spaces. The undifferentiated sporangium increased to about 20-25 microm in diam before being enveloped by a wall with a convoluted mid-layer. The sporangium differentiated into an unordered mass of zoospores that escaped from the cyst through a pronounced germ tube about 4-5 microm in diam and 10-15 microm long. Weakly developed germ tubes have been described in Perkinsus but are absent altogether in Parvilucifera and syndineans. Comparison of these data with other myzozoans led us to classify the parasite as Parvilucifera prorocentri sp. nov., Myzozoa. Although we were hesitant to erect a new genus name in the absence of molecular sequence data, our ultrastructural data strongly indicated that this parasite is most closely related to perkinsids and syndineans, and represents an intriguing candidate for the cellular identity of a major subclade of Group I alveolates.

Different life cycle strategies of the dinoflagellates Fragilidium duplocampanaeforme and its prey Dinophysis acuminata may explain their different susceptibilities to the infection by the parasite Parvilucifera infectans.

Some marine dinoflagellates form ecdysal cyst (=temporary cysts) as part of their life cycle or under unfavorable growth conditions. Whether the dinoflagellates form ecdysal cysts or not may influence susceptibility to parasitism. In this study, parasite prevalence relative to inoculum size of the parasitoid Parvilucifera infectans zoospores for two dinoflagellate hosts (i.e., Fragilidium duplocampanaeforme and Dinophysis acuminata), which have different life cycle strategies, was examined. Further, susceptibility of cysts to parasitism, encystment signal, duration of encystments, and effects of induced encystment on diel periodicity, using ecdysal cyst-forming F. duplocampanaeforme were explored. The percent hosts infected by P. infectans plotted as a function of inoculum size showed a sharp increase to a maximum in D. acuminata, but a gradual linear rise in F. duplocampanaeforme: while the parasite prevalence in D. acuminata increased to a maximum of 78.8 (±2.4%) by a zoospore:host ratio of 20:1, it in F. duplocampanaeforme only reached 8.9 (±0.3%), even at a zoospore:host ratio of 120:1. In F. duplocampanaeforme, infections were observed only in the vegetative cells and not observed in ecdysal cysts. When exposed to live, frozen, and sonicated zoospores and zoospore filtrate, F. duplocampanaeforme formed ecdysal cysts only when exposed to live zoospores, suggesting that temporary cyst formation in the dinoflagellate resulted from direct contact with zoospores. When the Parvilucifera zoospores attacked and struggled to penetrate F. duplocampanaeforme through its flagellar pore, the Fragilidium cell shed all thecal plates, forming a 'thecal cloud layer', in which the zoospores were caught and immobilized and thus could not penetrate anymore. The duration (35±1.8h) of ecdysal cysts induced with addition of zoospores was significantly longer than that (15±0.8h) of normally formed cysts (i.e., without addition of zoospores), thereby resulting in delayed growth as well as influencing the pattern of diel periodicity. The results from this study suggest that in addition to the classical predator-prey interaction and allelopathic interaction, parasitism and its accompanying defense can make the food web dynamics much more complicated than previously thought.

Life-cycle, ultrastructure, and phylogeny of Parvilucifera corolla sp. nov. (Alveolata, Perkinsozoa), a parasitoid of dinoflagellates.

Recent studies of marine protists have revealed parasites to be key components of marine communities. Here we describe a new species of the parasitoid genus Parvilucifera that was observed infecting the dinoflagellate Durinskia baltica in salt marshes of the Catalan coast (NW Mediterranean). In parallel, the same species was detected after the incubation of seawater from the Canary Islands (Lanzarote, NE Atlantic). The successful isolation of strains from both localities allowed description of the life cycle, ultrastructure, and phylogeny of the species. Its infection mechanism consists of a free-living zoospore that penetrates a dinoflagellate cell. The resulting trophont gradually degrades the dinoflagellate cytoplasm while growing in size. Once the host is consumed, schizogony of the parasitoid yields a sporocyte. After cytokinesis is complete, the newly formed zoospores are released into the environment and are ready to infect new host cells. A distinguishing feature of the species is the radial arrangement of its zoospores around the central area of the sporocyte during their formation. The species shows a close morphological similarity with other species of the genus, including P. infectans, P. sinerae, and P. rostrata.

The 'other' coral symbiont: Ostreobium diversity and distribution.

Ostreobium is an endolithic algal genus thought to be an early-diverging lineage of the Bryopsidales (Ulvophyceae, Chlorophyta). Ostreobium can live in low-light conditions on calcium carbonate substrata in tropical conditions. It is best known as a symbiont of corals, where it lives deep within the animal skeleton and exchanges nitrogen and carbon, as well as providing nutrients and photoassimilates. In contrast to the relatively well-studied role of the photosynthetic zooxanthellae symbionts in coral (Symbiodinium), Ostreobium phylogeny, diversity and distribution are all poorly understood. Here, we describe the phylogenetic position and diversity of Ostreobium based on plastid 16S ribosomal DNA (rDNA), 18S rDNA and rbcL genes from a nuclear genome survey and complete plastid genome, and determined its environmental diversity and distribution by screening the publicly available environmental data for those genes. The results shed light on the phylogeny and the ecology of the 'other' coral symbiont.

Significance of plankton community structure and nutrient availability for the control of dinoflagellate blooms by parasites: a modeling approach.

Dinoflagellate blooms are frequently observed under temporary eutrophication of coastal waters after heavy rains. Growth of these opportunistic microalgae is believed to be promoted by sudden input of nutrients and the absence or inefficiency of their natural enemies, such as grazers and parasites. Here, numerical simulations indicate that increasing nutrient availability not only promotes the formation of dinoflagellate blooms but can also stimulate their control by protozoan parasites. Moreover, high abundance of phytoplankton other than dinoflagellate hosts might have a significant dilution effect on the control of dinoflagellate blooms by parasites, either by resource competition with dinoflagellates (thus limiting the number of hosts available for infection) or by affecting numerical-functional responses of grazers that consume free-living parasite stages. These outcomes indicate that although both dinoflagellates and their protozoan parasites are directly affected by nutrient availability, the efficacy of the parasitic control of dinoflagellate blooms under temporary eutrophication depends strongly on the structure of the plankton community as a whole.

New Insights into the Parasitoid Parvilucifera sinerae Life Cycle: The Development and Kinetics of Infection of a Bloom-forming Dinoflagellate Host.

Parvilucifera sinerae is a parasitoid of dinoflagellates, the major phytoplankton group responsible for harmful algal bloom events. Here we provide a detailed description of both the life cycle of P. sinerae, based on optical, confocal, and transmission electron microscopy observations, and its infection kinetics and dynamics. P. sinerae completes its life cycle in 3-4 days. The zoospore encounters and penetrates the host cell within 24h after its addition to the host culture. Inside the host, the parasitoid develops a trophocyte, which constitutes the longest stage of its life cycle. The trophocyte replicates and divides by schizogony to form hundreds of new zoospores contained within a sporangium. Under laboratory conditions, P. sinerae has a short generation time, a high rate of asexual reproduction, and is highly prevalent (up to 80%) in the Alexandrium minutum population. Prevalence was shown to depend on both the parasitoid inoculum size and host density, which increase the encounter probability rate. The parasitoid infection parameters described in this study are the first reported for the genus Parvilucifera. They show that P. sinerae is well-adapted to its dinoflagellate hosts and may be an important factor in the termination of A. minutum blooms in the natural environment.

Marine dinoflagellates show induced life-history shifts to escape parasite infection in response to water-borne signals.

Many dinoflagellate species form dormant resting cysts as a part of their life cycle, and in some freshwater species, hatching of these cysts can be delayed by the presence of water-borne signals from grazing zooplankton. Some marine dinoflagellates can form temporary cysts, which may function to resist unfavourable short-term environmental conditions. We investigated whether the marine dinoflagellate Alexandrium ostenfeldii is able to induce an increased resistance to the parasitic flagellate Parvilucifera infectans by forming temporary cysts. We performed several laboratory experiments where dinoflagellates were exposed either to direct contact with parasites or to filtered water from cultures of parasite-infected conspecifics (parasite-derived signals). Infection by P. infectans is lethal to motile A. ostenfeldii cells, but temporary cysts were more resistant to parasite infection. Furthermore, A. ostenfeldii induced a shift in life-history stage (from motile cells to temporary cysts) when exposed to parasite-derived water-borne signals. The response was relaxed within a couple of hours, indicating that A. ostenfeldii may use this behaviour as a short-term escape mechanism to avoid parasite infection. The results suggest that intraspecies chemical communication evoked by biotic interactions can be an important mechanism controlling life-history shifts in marine dinoflagellates, which may have implications for the development of toxic algal blooms.

Intracellular bacteria: the origin of dinoflagellate toxicity.

Dinoflagellate blooms of the same species have been registered either as toxic or nontoxic and, in the latter case, toxicity may be of different types. A hypothesis has been formulated according to which the bacteria having in some way taken part in the toxin formation are either inside the dinoflagellate cell or in the nutritive liquid. The presence of intracellular bacteria in those microorganisms has been studied mainly in material from cultures, a few from the sea, and several strains were isolated from different species. Experiments with crossed inoculations have shown that the bacterial strain from Gonyaulax tamarensis caused the cells of some other species to become toxic. From nontoxic clonal cultures of Prorocentrum balticum, Glenodinium foliaceum, and Gyrodinium instriatum, after inoculation of that bacterial strain, cultures were obtained whose cell extracts showed the same kind of toxicity as G. tamarensis. No toxic action could be found in the extracts of the bacterial cells form the assayed strains. The interference of intracellular bacteria in the metabolism of dinoflagellates must be the main cause of their toxicity.