Biological control potential of red-tide marine dinoflagellate blooms by an Amoebophrya parasitic killer

Abstract

Despite the widespread prevalence of the genus Karenia, a notorious contributor to harmful algal blooms (HABs) worldwide, there are no reported instances of Amoebophrya spp. infecting Karenia. In 2020, we isolated infected Karenia mikimotoi cells, performed repeated inoculations of these infected cells to uninfected host cultures, and isolated a parasite strain of Amoebophrya from Osaka Bay, Japan. This strain exhibited infectivity toward K. mikimotoi and Karenia papilionacea but not Karenia selliformis or 14 other dinoflagellate species, highlighting its remarkable host specificity. Laboratory coculture experiments revealed pronounced host growth suppression upon introducing the parasite. The prevalence of infection surged rapidly, reaching a peak of 97% at 60 h post addition of the parasite, underscoring its potent infectivity. Our findings suggest that Amoebophrya sp. OSK2020 strain can inhibit K. mikimotoi, making them a promising natural biological control agent for HABs with a minimal impact on marine microbial ecosystems.

A strain of Amoebophrya isolated from Osaka Bay, Japan infects Karenia mikimotoi and K. papilionacea while showing no infectivity towards K. selliformis or 14 other dinoflagellate species. In laboratory experiments, infection with Amoebophyra caused a substantial suppression of host growth.

Introduction

Parasitism, long recognized as one of the most prevalent lifestyles on Earth, plays a pivotal role in evolutionary and ecological processes at various scales, including food webs and ecosystems^1,2. In aquatic ecosystems, the interactions between parasites and hosts assume a more substantial ecological role than previously acknowledged, as parasites have been observed to be widespread across the trophic levels of the food web, infecting a range of organisms, from top predators to phytoplankton producers^3,4. A striking example of the impact of the parasite is the documented collapse of the eastern oyster (Crassostrea virginica) fishery in Chesapeake Bay, USA. The populations of eastern oysters in the Chesapeake Bay have decreased sharply due to a combination of factors, including overfishing and infectious diseases, with the most prevalent diseases attributed to the protozoan parasites Haplosporidium costale, Haplosporidium nelsoni, and Perkinsus marinus^5,6.

While our understanding of parasitism in aquatic ecosystems, particularly involving single-celled microorganisms, remains limited, recent studies utilizing DNA metabarcoding have unveiled a vast diversity of planktonic parasites in the ocean, including Syndiniales, Perkinsozoa, and Chytridiomycota^7–10. These parasites are widespread in coastal waters and the open ocean, and their high infection rates can substantially impact algal population dynamics, leading to decreased bloom-forming harmful or toxic microalgae in coastal regions. Such changes affect the structure of microbial food webs, which have economic implications^11,12. Biological control methods of algal blooms are environmentally friendly and may provide a cost-effective alternative to physical and chemical strategies. As potential biological control agents of harmful algal blooms (HABs), viruses, bacteria, fungi, and protists, including dinoflagellates, can impair or kill microalgae as reported in reviews^13–15. Among such candidates, two genera of parasites, Amoebophrya (Syndiniales) and Parvilucifera (Perkinsozoa), have especially garnered attention^16–23, and the life cycles of these parasitoids are broadly similar. Infections commence when a free-living zoospore infiltrates a host cell, developing into a trophocyte that expands while entirely consuming the contents of the host cell. Subsequently, the trophocyte nucleus undergoes division and transforms into a sporocyte, facilitating the production of zoospores. These newly generated zoospores are released into the marine environment and seek new hosts^24,25. These zoospores exert a more substantial influence on the microbial food web than previously anticipated, as they are preyed upon by predators such as ciliates^26,27.

The endoparasitic dinoflagellate genus Amoebophrya is a potential biological control agent among various parasites and pathogens. This is because of its widespread presence in coastal environments and its ability to parasitize over 75 dinoflagellate species, many of which are HAB species^23,28. While the genus Parvilucifera has been reported in numerous coastal areas worldwide, most reports originate from Europe⁹, with a few instances documented in Asian countries, such as Korea²³. Causal relationships between HAB species and their parasites have been observed in coastal regions worldwide. A well-known example involves the toxic dinoflagellate, Alexandrium minutum, which proliferated along the French coast from the 1990s to 2000 before its sudden decline. Several studies have identified the parasite Amoebophrya as a likely factor regulating the proliferation of this toxic dinoflagellate^21,24. In our previous study, we also observed a connection between the population dynamics of the toxic dinoflagellate Alexandrium catenella and Amoebophrya sp. in Osaka Bay, Japan, where the maximum infection rate reached 73% during the host’s bloom²⁹.

In the genera Akashiwo, Alexandrium, Dinophysis, and Margalefidinium (Cochlodinium), which are particularly noteworthy among HAB species, numerous instances of infection by Amoebophrya have been documented in field studies; in addition, the infection process and host range have also been elucidated in laboratory experiments^{20,27,30–33}. Despite Amoebophrya’s widespread global distribution, only several dinoflagellate host–parasite systems have been established in cultures thus far²³. This may be attributed to the relative difficulty of isolating Amoebophrya spp. and the substantial effort required for long-term maintenance. Developing more streamlined maintenance and management protocols for these parasites will be crucial for advancing future research.

Karenia mikimotoi is a HAB-forming dinoflagellate species that has negatively impacted the aquaculture industry for over 80 years, resulting in mass mortalities of marine products such as fish and shellfish. Blooms of the species have been documented since the 1930s in temperate coastal waters across Asia, Europe, Oceania, South Africa, and South America³⁴. In Asia, K. mikimotoi is recognized as one of the most damaging algal species due to the substantial economic losses it has caused. For example, in Japanese coastal areas, K. mikimotoi was first reported to form blooms in Gokasho Bay in 1934³⁵. Since the 1960s, K. mikimotoi has bloomed frequently, causing damage to fisheries; the most substantial economic loss, 43 million USD, was recorded in 1984³⁶. Additionally, a massive bloom of K. mikimotoi occurred along the coast of China, resulting in the mass mortality of farmed abalones and economic losses exceeding 330 million USD in 2021³⁷. Blooms of K. mikimotoi have become a common environmental disaster in China, occurring almost annually, with 152 blooms reported between 1998 and 2012³⁷.

Algicidal bacteria or virus-like particles are also recognized to suppress the growth of K. mikimotoi^38–41. However, their host range has not been adequately studied, and their impact on marine ecosystems is unpredictable. Therefore, there is an urgent need to isolate the parasite Amoebophrya, which specifically infects the harmful dinoflagellate K. mikimotoi. Despite large-scale surveys conducted along the Chinese coasts^42,43, Amoebophrya spp. Infection with the genus Karenia has never been found or isolated in these surveys or globally. In one laboratory culture study⁴⁴, it was reported that the perkinsid Parvilucifera sinerae could infect Karenia brevis, which has caused substantial fishery damages in the United States and Mexico⁴⁵. However, the experiment showed that K. brevis infection occurred relatively slowly, and the host culture remained alive for up to 2 weeks after infection.

During a bloom of K. mikimotoi in Osaka Bay in 2020 (Fig. 1), we observed that many cells of K. mikimotoi appeared to be infected by Amoebophrya. We collected the infected cells of K. mikimotoi from seawater and subsequently inoculated them into uninfected cultured cells of K. mikimotoi. Through this process, we successfully isolated and cultivated a strain, identified as Amoebophrya sp. OSK2020. This study presents the nuclear rDNA sequences, infection process, host range, and infecting ability of the Amoebophrya strain.

Fig. 1. Maps showing the sampling point.

Permission was granted to use the map image.

Results

Genetic information of Amoebophrya sp.

To elucidate the phylogenetic position of the OSK2020 strain within the genus Amoebophrya, the nuclear 18S rDNA sequence (1742 bp) was determined. The analysis of the maximum likelihood (ML) tree (Fig. 2) revealed that the sequence of OSK2020 did not match any registered sequences. The highest similarity (98.6%) was to Amoebophrya sp. 2 ex Margalefidinium polykrikoides (accession number: KF791348). The Basic Local Alignment Search Tool (BLAST) search compared uncultured/environmental sequences, dividing the 18S rDNA into several parts. However, no matches to any uncultured/environmental sequences were found.

Fig. 2. Unrooted ML phylogenetic tree of the nuclear 18S rDNA in the genus Amoebophrya clustered by their dinoflagellate hosts (shown in pictures) and outgroups.

A culture strain, Amoebophrya sp. OSK2020 ex Karenia mikimotoi, isolated in this study, is indicated in red letters. Bootstrap values of >50% on the tree produced using MEGA ver. X software is the percentage of 100 bootstrap replicates at the respective nodes. The final dataset contains 1579 informative sites. The scale bar represents the number of substitutions per site.

To clarify the genetic diversity of the OSK2020 strain, nuclear 28S rDNA sequences (D1/D2 region; 755 bp) from the culture strain and cells of five naturally infected K. mikimotoi isolated from Osaka Bay (2020Oct5-1, 2020Oct5-2, 2020Oct5-3, 2020Oct5-4, and 2020Oct5-5) were analyzed. The results (Fig. 3) show that only cell ID 2020Oct5-1 differed by one base from the other sequences. In line with the 18S rDNA findings described above, a BLAST search showed the 28S rDNA sequence of Amoebophrya sp. OSK2020 did not match any uncultured/environmental sequences.

Fig. 3. Unrooted ML phylogenetic tree of the nuclear 28S rDNA in the genus Amoebophrya and an outgroup.

Sequences from the culture strain, Amoebophrya sp. OSK 2020 ex Karenia mikimotoi, and natural cells of Amoebophrya obtained from five infected cells of K. mikimotoi are highlighted in red letters. Bootstrap values exceeding 50% on the tree, generated using MEGA version X software, are provided as percentages of 100 bootstrap replicates at the respective nodes. The final dataset comprises 734 informative sites, with the scale bar denoting the number of substitutions per site.

Host range of Amoebophrya sp. OSK2020

Cross-infection experiments were conducted using Amoebophrya sp. OSK2020 and several dinoflagellate hosts to assess the host range (Table 1). The results confirmed that Amoebophrya sp. OSK2020 infected 6 strains of K. mikimotoi and one strain of Karenia papilionacea, but not Karenia selliformis. Although the host cultures were maintained until mortality, no parasitism was observed in other dinoflagellate hosts.

Table 1.

Host specificity of Amoebophrya sp. OSK2020 ex Karenia mikimotoi demonstrated by laboratory culture experiments

Infected host species (strain)	Uninfected host species
Karenia mikimotoi (KmTG09)	Akashiwo sanguinea
Karenia mikimotoi (KmTG01)	Alexandrium catenella
Karenia mikimotoi (KmUW3)	Alexandrium pacificum
Karenia mikimotoi (KmUW5)	Gonyaulax polygramma
Karenia mikimotoi (KmUW9)	Gonyaulax spinifera
Karenia mikimotoi (KmHM3)	Gymnodinium impudicum
Karenia papilionacea	Heterocapsa circularisquama
	Karenia selliformis
	Levanderina fissa
	Margalefidinium polykrikoides
	Prorocentrum koreanum
	Prorocentrum mexicanum
	Prorocentrum minimum
	Scrippsiella trochoidea
	Takayama xiamenensis

Parasitism was confirmed by the green autofluorescence of an Amoebophrya trophont within the host cells. Locations include TG: Ago Bay, UW: Uwajima, and HM: Lake Hamana.

Infection process and time-course experiment in host–parasite culture system

Figure 4 illustrates the infection process of Amoebophrya sp. OSK2020 in a stock culture. The OSK2020 parasite strain displayed the typical life cycle of Amoebophrya, consistent with previous findings²⁴. Dinospores exhibited bright blue–green autofluorescence, which did not originate from the chloroplasts. This characteristic of the parasite makes it easy to distinguish red autofluorescence emitted by the host chloroplasts by using microscopy (Fig. 4). Only one growing trophont was typically observed during infection. However, occasional instances of multiple trophonts with a host cell were also noted, a phenomenon observed in another dinoflagellate host–Amoebophrya systems⁴⁶. The size of dinospores was initially relatively large (5–7 µm) outside of the host cells of K. mikimotoi, gradually decreasing in size (3–4 µm) with age outside of the hosts, which aligned with observations of other Amoebophrya spp.^29,32. The duration from infection to host cell colonization until dinospore release was approximately 3–4 days. Between 7 and 10 days after parasite inoculation, a high prevalence of the parasite resulted in the eventual demise of most host cells (Figs. 4, 5, Supplementary Fig. 1).

Fig. 4. Light and epifluorescence micrographs of the life cycle stages of Amoebophrya sp. OSK2020 ex Karenia mikimotoi in culture stocks.

a, b Red autofluorescence indicates a healthy (noninfected) host cell with chloroplasts. c, d Early infection stage, featuring a small trophont within the host cell, confirmed by green autofluorescence of Amoebophrya. e, f Mid-to-late infection stage. g, h Late infection stage with multiple trophonts present. i, j The emerging vermiform (aggregation of multiple dinospores) of Amoebophrya from the host cell. k Dinospores (three green dots) dispersed from a vermiform attached to a new host cell (red), which is already infected. All scale bars = 20 μm.

Fig. 5. Epifluorescence micrographs in culture stocks.

a Early stage of infection, showing numerous actively swimming host cells (red autofluorescence), some of which are infected (green autofluorescence). b Late stage of infection, where the host cells have largely disappeared, and numerous dinospores are observed. All scale bars = 200 μm.

To assess the infection ability of Amoebophrya sp. OSK2020, host abundance, prevalence, and dinospore abundance were measured for 168 h (Fig. 6). The control used was a noninfected culture of K. mikimotoi, for which the abundance gradually increased, reaching 37,333 ± 5002 cells/mL at 168 h (Fig. 6a). Conversely, in the experimental plot where the parasite was added, the host density increased from the beginning of incubation until 36 h, then decreased rapidly, and disappeared completely at 168 h (Fig. 6a, b). Figure 6b illustrates the host density by infection stage in the experimental plot depicted in Fig. 6a (red bars) and its prevalence. The first infection was confirmed at 36 h after the start of the culture, after which the infection rate sharply increased. The maximum prevalence of 97% ± 3% during the experiment was recorded after 60 h, showing early and mid-to-late-stage infections. The proportion of mid-to-late-stage infection to the infected host cells increased toward the latter half of the experiment. The abundance of dinospores gradually decreased from the initial density of 16,000 cells mL⁻¹ to 473 ± 42 cells mL⁻¹ at 48 h, possibly owing to penetration into host cells or natural mortality (Fig. 6c). Many new dinospores released from infected host cells appeared at 60 h, reaching a maximum density of 38,400 ± 1200 cells mL⁻¹ and then gradually decrease. Although the experiment ended at 168 h (7 days), we continued to observe the dinospores and confirmed their motility for up to 240 h (10 days) without host cells.

Fig. 6. Time-course study of the host dinoflagellate Karenia mikimotoi infected with Amoebophrya sp. OSK2020.

a Cell densities of K. mikimotoi in infected and uninfected (control) cultures. b Breakdown of infection stages (uninfected, early, and mid-late) of K. mikimotoi cells as shown in the treatment (red bars) in Fig. 7a above, with the line graph depicting the prevalence (on the right axis) of Amoebophrya. c Cell density of dinospores (Amoebophrya outside the host cells). Error bars indicate the standard deviation of cell counting (n = 3).

Discussion

Many ecological aspects regarding the infection of Amoebophrya spp. in K. mikimotoi (Amoebophrya sp. ex K. mikimotoi) remain unclear, including its distribution in coastal waters, native or invasive status, impact on host field dynamics, and survival strategies in the absence of host cells in the water column. The distribution of Amoebophrya sp. ex K. mikimotoi in coastal regions is especially interesting. In a previous study, we isolated Amoebophrya sp. OSK2019 from Osaka Bay infecting the genus Alexandrium and observed Amoebophrya spp. infecting various dinoflagellate hosts such as Prorocentrum micans, Prorocentrum minimum, Scrippsiella sp., and Tripos furca²⁹. Osaka Bay has been a hotspot for dinoflagellate blooms⁴⁷, potentially contributing to the diverse range of Amoebophrya parasites. Recent advancements in metabarcoding techniques enable the precise exploration of phytoplankton composition, diversity, and distribution, shedding light on the importance of parasites such as Amoebophrya^48–51. Despite global metabarcoding analyses, the sequence is identical to Amoebophrya sp. OSK2020, obtained in this study, has not been registered in GenBank, indicating its presence is exclusive to Osaka Bay. The blooms of K. mikimotoi have recurred in Osaka Bay since the 1990s, with a peak density of 72,050 cells/mL recorded in 2020 (Fig. 7). Records of the maximum cell density of K. mikimotoi worldwide have been reviewed previously³⁴: the maximum cell density in Japanese coastal areas ranged from 800 to 47,000 cells/mL, and was 34,000 cells/mL in China, 70,000 cells/mL in Norway, 20,000 cells/mL in Scotland, 70,000 cells/mL in Spain, 24,000 cells/mL in South Africa, and 9,000 cells/mL in Chile. Thus, it is evident that Osaka Bay is one of the areas most densely populated by K. mikimotoi. Therefore, it is reasonable to assume that Amoebophrya sp., which infects K. mikimotoi as a host, exists in Osaka Bay. However, it is not known whether Amoebophrya sp. OSK2020 originally inhabited Osaka Bay, which became apparent with an increase in its host population or its introduction to Osaka Bay from another sea area. Future research, such as DNA analysis of the water column or sediment using specific primers for Amoebophrya spp., is expected to provide further insight.

Fig. 7. Field dynamics of Karenia mikimotoi (host) in Osaka Bay, Japan.

a The annual maximum cell densities (cells mL⁻¹) of K.mikimotoi in Osaka Bay from 1990 to 2023. Data presented are from the Research Institute of Environment, Agriculture, and Fisheries, Osaka Prefecture. Seawater samples were collected within the upper 10 m of the water column using a 10-meter hose. Surveys were conducted at more than 10 fixed sampling points in Osaka Bay; the maximum values among those points are shown in the graph. b Fluctuation of the cell density of K. mikimotoi in Osaka Bay, specifically for the year 2020. Similarly, the maximum values among the fixed sampling points are shown in the graph.

The OSK2020 strain of the parasite was capable of infecting K. mikimotoi and Karenia papilionacea but not Karenia selliformis, even though they are of the same genus, as demonstrated by laboratory experiments. The result indicates that the high host specificity of OSK2020 within Amoebophrya, considering its safety during field applications. In a previous study, for instance, Amoebophrya sp. ex Alexandrium could infect all tested species of the genus Alexandrium, Prorocentrum micans, Prorocentrum minimum, and Tripos furca²⁹. In other cases of Amoebophrya spp., varying degrees of host specificity were apparent, ranging from unspecific^20,52,53 to extremely species-specific^22,32,46,54. However, Parvilucifera (Perkinsozoa) has a more comprehensive host range than Amoebophrya spp. For example, Parvilucifera corolla and Parvilucifera sinerae infected 16 out of 24 and 15 out of 23 tested genera of dinoflagellate hosts, respectively, in laboratory culture experiments^55,56. The study of mechanisms by which these parasites infect their hosts is of value, and there are several insights into the host attraction and recognition mechanisms of parasites. Amoebophrya uses a microfilament polymerization-based motility system to enter the host cell⁵⁷. Parvilucifera sinerae is activated by filtered exudates containing the host material, identified as the algal metabolite dimethylsulphide (DMS)^58,59. Some chytrids also recognize chemical cues, possibly related to photosynthetic exudates of their hosts⁶⁰. Currently, there is limited information on how Amoebophrya spp. can recognize specific hosts, but the narrow host range of Amoebophrya appears to be disadvantageous for stable survival. However, when used as a biological control agent for HABs, the narrow host range may confer advantages because of the high species specificity of Amoebophrya sp. OSK2020 indicates safety for the microbial ecosystem, as they suppress only the blooms of harmful Karenia species and does not affect other plankton.

One ecological question remains regarding the life history of Amoebophrya sp. infection of Karenia species. As illustrated in Fig. 7b, in Osaka Bay, the bloom of K. mikimotoi has been detected in seawater for only approximately 2 months during the year; almost no K. mikimotoi cells have been detected in other periods. Other Karenia species are also rarely observed throughout the year in Osaka Bay. Therefore, the question arises of where and in what form does Amoebophrya survives during the absence of the host? Several dinoflagellates produce overwintering resting cysts as part of their sexual life cycle, but this has not been reported in the parasitic dinoflagellate Amoebophrya spp. Our previous fluorescent microscopy study indicated that some cysts of A. catenella obtained from a natural seawater sample contained a parasitic Amoebophrya sp. that was infected before the encystment process²⁹. This finding shows that Amoebophrya may coexist and survive in the cysts of infected hosts in sediment. Similar to our observations, Amoebophrya infection has also been confirmed in cysts of the host dinoflagellate Scrippsiella trochoidea; after 10 months, these cysts were able to germinate, and new infective cycles of the parasite were rapidly observed⁵³. Although Amoebophrya spp. may not infect hosts once they have formed cysts, it has been reported that the chytrid parasite Dinomyces arenysensis can infect not only the vegetative cells of the host dinoflagellate S. trochoidea but also cells after cyst formation⁶¹. Discovering and isolating parasites capable of infecting HAB cysts could decrease the cyst population in sediment, thereby mitigating the occurrence of HAB in water. Cysts of K. mikimotoi have been documented in culture experiments and surveys along the coast of China^62,63. Therefore, assuming that host switching does not occur in the water, it is conceivable that Amoebophrya sp. ex K. mikimotoi could coexist within the host cysts in sediment and therefore survive the unfavorable conditions. In contrast, Amoebophrya dinospores that can infect K. mikimotoi may occur at very low densities in the water, even in the absence of the host. In the future, attempts to detect the DNA of Amoebophrya spp. from seawater or sediment samples may provide insights into their life history. Thus, the 28S rDNA sequence of Amoebophrya sp. OSK2020, provided in this study, is expected to facilitate the design of a specific primer/probe for DNA quantification.

In the time-course culture experiment of the host and Amoebophrya sp. OSK2020, a prevalence of up to 97% was observed, accompanied by remarkably high algicidal effects in the host (see Fig. 6). Similarly, the maximum prevalence of other Amoebophrya spp. in laboratory settings has been consistently high: 98.2% in Amoebophrya sp. ex Akashiwo sanguinea⁴⁶, 97.9% in Amoebophrya sp. ex Gymnodinium instriatum⁴⁶, 91.8% in Amoebophrya sp. ex Karlodinium micrum⁴⁶, and 98.7% in Amoebophrya sp. ex Alexandrium catenella (Group I)³¹. Conversely, laboratory experiments have reported a slightly lower prevalence for the genus Parvilucifera (Perkinsozoa): 80% in P. sinerae for Alexandrium minutum²⁵ and 79% in Parvilucifera rostrata for A. minutum⁶⁴. This difference indicates that Amoebophrya spp. may substantially affect host dynamics in controlled laboratory environments and field waters. However, field prevalence is influenced by various factors, including host abundance^16,54, temperature⁶⁵, salinity⁴², nutrient availability⁶⁶, and host-to-parasite ratio^32,46. Thus, Amoebophrya spp. are highly infectious and proliferative but also vulnerable. As demonstrated by our experiments and reported elsewhere⁴⁶, most Amoebophrya dinospores perish within 10 days without the host. Less than 20% of the dinospores released into the water can successfully infect new hosts; moreover, if new hosts are not encountered, their infectivity may diminish completely within 2–5 days⁴⁶. This short survival time poses a substantial disadvantage when we introduced this parasite into the field water. We added only fresh and active dinospores into the ocean to avoid introducing the cultured host that was isolated from other areas. Therefore, using a plankton net may be necessary to remove the host cells from the culture medium before introduction into the sea. However, the short survival time of the parasite could also act as an advantage, as the dinospores do not persist in seawater for prolonged periods, preventing substantial impacts on the marine microbial ecosystem.

In future research, several topics need further clarification. Laboratory culture experiments should be conducted using Amoebophrya sp. OSK2020 under various parameters, and field investigations of host–parasite dynamics should also be performed. Additionally, it is crucial to develop preservation methods to maintain Amoebophrya sustainably in the long term, as the dual cultivation of uninfected and infected host cultures and frequent transfers are highly time-consuming and labor-intensive. Furthermore, if Amoebophrya is to be utilized as a biological control agent for HAB species, verifying its safety for marine ecosystems, fishery products, and human health is essential. Several steps may be necessary before introducing a parasite into the field water. These include transporting seawater containing host blooms from the field to the laboratory to confirm the pathogenicity of the parasite, and then setting up an isolated mesocosm in the field for infection experiments. In addition to research, working closely with the local government and the aquaculture producers is also crucial.

As primary producers, dinoflagellates are crucial for energy transfer and resource cycling within marine ecosystems. In recent years, the global proliferation of harmful or toxic dinoflagellate blooms has been increasingly likely to stem from human activities associated with population growth and coastal development^67,68. However, it is essential to acknowledge that dinoflagellates are, rather than inherently harmful, integral constituents of marine ecosystems. The continuous coevolution of hosts and parasites presents a challenging scenario as they are potentially affected by human intervention. Rapid environmental transformations may lead to the depletion of parasite species, making future efforts into parasite conservation a global task⁶⁹. Advancing our understanding of parasites is imperative to safeguard marine environments and preserve parasite species diversity.

Methods

Cultures of host and parasite

The host strain K. mikimotoi KmTG09 was isolated from Ago Bay, Japan 2013. A stock culture of K. mikimotoi was sustained in Daigo’s IMK culture medium (Nihon Pharmaceutical, Osaka, Japan). The host culture was cultivated at 24 °C and a salinity of 30 under cool-white fluorescent lamps, providing 100 µmol photons m⁻² s⁻¹ under a 14-hour light/10-hour dark cycle. The parasite strain Amoebophrya sp. OSK2020 ex K. mikimotoi was isolated from an infected cell of K. mikimotoi in a seawater sample (water temperature: 24 °C, salinity: 30) collected at Tanigawa Port (34.322 N, 135.120 E), Osaka Bay, Japan, on October 5, 2020 (Fig. 1). In Tanigawa Port, the cell density of K. mikimotoi and the prevalence were not measured throughout the bloom; however, the prevalence on the single measurement day (October 5) was 32%. An infected cell of K. mikimotoi was inoculated into an uninfected culture of K. mikimotoi. Subsequently, the infection of Amoebophrya spread, facilitating the successful isolation of the Amoebophrya strain. The host and parasites were confirmed to be culturally axenic. Amoebophrya sp. was maintained in a stock culture with the host cells of K. mikimotoi by transferring aliquots (0.5 ml) of the infected host culture into the uninfected host culture (20 ml) at 4-day intervals. As the timing of subcultures may vary, even under the same culture conditions, frequent microscopic observations are necessary to maintain Amoebophrya strains. Our laboratory has stably maintained this host–parasite culture system for over 4 years.

Isolation of natural cells of Amoebophrya sp. ex K. mikimotoi

To ascertain the genetic diversity of Amoebophrya sp. ex K. mikimotoi, five naturally infected host cells (cell ID = 2020Oct5-1, 2020Oct5-2, 2020Oct5-3, 2020Oct5-4, and 2020Oct5-5, respectively) were isolated from the same seawater sample from which Amoebophrya sp. OSK2020 was isolated. For single-cell analysis, the five infected cells of K. mikimotoi were isolated via micropipetting under an inverted fluorescent microscope (IX71, Olympus, Tokyo, Japan) equipped with the U-MNBV2 cube (420–440 nm excitation, 475 nm emission). Cells of K. mikimotoi were washed at least three times with filter-sterilized seawater (0.1 μm pore size). Individual cells were placed into a 0.2 mL PCR tube with 15 μL of 5% Chelex® suspension (Bio-Rad, Hercules, CA), and DNA was extracted following an established method⁷⁰.

PCR, sequencing, and phylogenetic analysis

PCR targeting nuclear 18S rDNA and 28S rDNA of our culture strain Amoebophrya sp. OSK2020 and 28S rDNA from field samples were performed, and sequencing was conducted using our previous methods^29,71. The 18S rDNA sequences were amplified using three primer pairs, S1F: 5’-AAC CTG GTT GAT CCT GCC AG-3’/S1R: 5’-CTA CGA GCT TTT TAA CTG CAA CAA-3’; S2F: 5’-CTG AGA AAC GGC TAC CAC ATC-3’/S2R: 5’-TGG TAA GTT TTC CCG TGT TGA GTC-3’; and S3F: 5’-AGC TTG CGG CTT AAT TTG ACT C-3’/S3R: 5’-CTA CGG AAA CCT TGT TAC GAC-3’⁷¹. The 28S rDNA primer set (n28S-Amoeb-1F: 5′-ACC CRC TGA ATT TAA GCA TAT AAG-3′, n28S-Amoeb-824R: 5′-ACT TTC ATT CCG CRT ACA GGT-3′) is designed explicitly for Amoebophrya spp. and does not amplify the host DNA²⁹. Therefore, the sequences of Amoebophrya spp. can be easily determined via direct sequencing without gene cloning. PCR was performed on a Veriti thermal cycler (Thermo Fisher Scientific, MA, USA) in a 20 μL reaction mixture containing 1.0 μL of template DNA, 0.2 mM of each dNTP, 1× PCR buffer, 1.5 mM Mg²⁺, 1.0 U of KOD -Plus- ver. 2 (TOYOBO Co., Ltd., Osaka, Japan) with intensive 3’ → 5’ exonuclease activity, and 0.2 μM of each primer. The PCR program included an initial denaturation at 94 °C for 2 min, followed by 40 cycles at 98 °C for 10 s, 58 °C for 30 s, and 68 °C for 60 s. The amplicons were resolved on 1.5% agarose gels. They were purified using the ExoSAP-IT PCR Product Cleanup Reagent (Thermo Fisher Scientific, USA). The Amoebophrya DNA was sequenced directly, without gene cloning, using a DYEnamic ET Terminator Cycle Sequencing Kit (GE Healthcare, Little Chalfont, UK) and analyzed on a 3730xl DNA Analyzer (Thermo Fisher Scientific, USA). Forward and reverse sequences were aligned utilizing the GENETYX software (Genetyx Corporation, Tokyo, Japan). The Amoebophrya sequences obtained in this study were checked against GenBank using the Nucleotide Basic Local Alignment Search Tool (BLASTN). Unrooted phylogenetic trees of 18S rDNA and 28S rDNA were generated by MEGA version X software⁷² using the ML method with default settings. The topology of the phylogenetic tree was evaluated using the bootstrap method with 100 replicates.

Cross-infection experiment using multiple host cultures

To assess the host range of Amoebophrya sp. OSK2020, a cross-infection experiment was conducted using six strains of K. mikimotoi and 16 host dinoflagellate species: Akashiwo sanguinea, Alexandrium catenella (Group I), Alexandrium pacificum, Gonyaulax polygramma, Gonyaulax spinifera, Gymnodinium impudicum (NIES-3597), Heterocapsa circularisquama, Karenia papilionacea, Karenia selliformis, Levanderina fissa, Margalefidinium polykrikoides, Prorocentrum koreanum, Prorocentrum mexicanum, Prorocentrum minimum (NIES-237), Scrippsiella trochoidea, and Takayama xiamenensis. The experimental design was based on previous studies^20,29,32,53. Specifically, all experiments were conducted using host cultures in exponential growth and recently formed (< 12 h) dinospores of Amoebophrya sp. OSK2020. Dinospores of Amoebophrya sp. OSK2020 were obtained from host cells through an 8-µm pore size filter (Nuclepore Track-Etch Membrane, Whatman, UK). When the host cultures reached a sufficient density, 2 mL aliquots were transferred to a 24-well microplate. To each well, dinospores were inoculated in a 20:1 ratio of dinospore: host, and the microplate was then incubated for 1 week at 20 °C under a 14-hour light:10-hour dark cycle of cool-white fluorescent light providing 100 µmol photons m⁻² s⁻¹. Parasitism was observed using a fluorescence microscope to detect the green autofluorescence of the parasite within the host⁷³, by analyzing a minimum of 100 cells.

The time-course experiment in the host–parasite culture system

To evaluate the impact of parasitism on the cell density of K. mikimotoi, we conducted a laboratory infection experiment in which dinospores of Amoebophrya sp. OSK2020 were obtained by filtering a stock parasite culture through a filter with 8 µm pores. For the infected treatments, triplicate host cultures of exponentially growing K. mikimotoi (strain KmTG09) with an initial density of 1600 cells mL⁻¹ were inoculated with recently formed (< 12 h old) dinospores of Amoebophrya sp. OSK2020 at an initial concentration of 16,000 cells mL⁻¹ (the dinospore: host ratio was 10:1) in a total volume of 20 mL. All flasks were incubated at 24 °C and salinity 30 under a 14 hour light:10-hour dark cycle of cool-white fluorescent light providing 100 µmol photons m⁻² s⁻¹. The infection process was monitored over time (every 12 h) to quantify the abundance of the hosts (uninfected and infected) and parasitoid dinospores. Triplicate subsamples were fixed with buffered formaldehyde (final concentration 1%) and immediately counted under the epifluorescence microscope. The triplicate parasite prevalence (%) was calculated by scoring at least 100 host cells per sample as uninfected cells, early-stage infection, or mid to late-stage infection, if the sample contained enough host cells.

Statistics and reproducibility

All error values are reported as the standard deviation of three replicates. The time-course experiments in the host–parasite culture system were repeated independently using separate flasks to ensure reproducibility. Microphotographs were repeatedly taken, and representative images are shown.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

G.N., M.Y., and K.Y. designed the study. K.Y., M.Nakajima, N.N., and M.Natsuike collected water samples. G.N. and S.M. performed DNA analysis and time-course culture experiments. M.Y. performed the host range experiment. W.S.-O. provided feedback on experimental planning and procedures. All authors contributed to the interpretation of the results. G.N. and M.Y. drafted the first version of the manuscript, and K.Y., M.Nakajima, N.N., M.Natsuike, and W.S.-O. provided critical feedback. All authors read and approved the final version of this manuscript.

Peer review

Peer review information

Communications Biology thanks Kevin Lafferty and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Linn Hoffmann, George Inglis, and David Favero. A peer review file is available.

Data availability

The 18S ribosomal RNA gene sequence of Amoebophrya sp. OSK2020, a culture strain, has been deposited in the National Center for Biotechnology Information database (accession number: LC801622). Additionally, the 28S ribosomal RNA gene sequences of Amoebophrya sp. (culture strain and natural cell) have been deposited in the National Center for Biotechnology Information database (accession numbers: LC801623 and LC801624). The raw data supporting the figures are provided in Supplementary Data 1–4.

Code availability

There is no custom code associated with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s42003-025-09141-1.