Cultures and samples

Samples containing Raphidocystis tubifera (HF-68Z strain) were collected from bottom sediments of the puddle that formed after snow melt in the floodplain of the Ild River, Yaroslavl Oblast, Russia (57°57′26.1″N 38°03′50.1″E), on 07 April 2020. Samples containing R. ambigua (HF-58Z strain) were collected from bottom sediments among shoreline aquatic macrophytes of Shchuchye Lake, Aqmola Oblysy, Kazakhstan (52°58′19.60″N 70°14′35.20″E), in August 2018. Samples containing R. symmetrica (HF-80Z strain) were collected from bottom sediments among shoreline stands of Thalia dealbata Fraser ex Roscoe of Shahu Lake, Wuhan, China (30°34′21.1″N 114°20′37.8″E), on 26 May 2024. Samples containing R. marginata (HF-64Z strain) were collected from bottom sediments of Davsha thermal spring, Republic of Buryatia, Russia (54°21′23.0”N 109°29’58.7”E), on 30 August 2019. No permit was required for collection of water samples at these locations.

The samples were transported to the laboratory at 4°C, after which each sample was inoculated into a 60 mm Petri dish. One rice grain was added to stimulate bacterial growth. The samples were incubated at 20–22°C in the dark to prevent the growth of autotrophic organisms and periodically examined with upright and inverted light microscopes. To obtain clonal cultures, single cells of centrohelids were picked out using a pulled glass micropipette with a fine tip and transferred into Petri dishes with autoclaved mineral water (Aqua Minerale, PepsiCo, Inc., Moscow Oblast, Russia). The cell culture of the flagellate Parabodo caudatus (Dujardin, 1841) Moreira et al., 2004 (BAS-1 isolate, IBIW RAS), which was grown in the same autoclaved mineral water supplemented with Aeromonas sobria bacteria (strain ICISC19, Institute for Cellular and Intracellular Symbiosis Collection, Russian Academy of Science, Russia), was used as prey. Strain HF-58Z was isolated in February 2019 and perished in February 2020. Strain HF-64Z has persisted since March 2020, strain HF-68Z since April 2020, and strain HF-80Z since May 2024.To assess the ability of raphidocystids to feed on cyanobacteria, toxic and nontoxic strains of Microcystis aeruginosa (strains 905 and 928, respectively) and Aphanizomenon sp. (strain 1399) were used. The strains of cyanobacteria were obtained from the Freshwater Algae Culture Collection, Institute of Hydrobiology, Chinese Academy of Sciences, and were grown on Waris-H medium [40].

Microscopy

An AxioScope A1 upright light microscope (Carl Zeiss, Jena, Germany) with DIC and phase contrast and water immersion objectives (63×) was used for the observation of living cells, and an inverted microscope Axio Observer 5 (Carl Zeiss, Jena, Germany) with phase contrast (objectives 20×) was used for preparing the cell cultures. Light microscopy images were taken with an MC-20 camera (Lomo-Microsystems, Saint Petersburg, Russia) and an MC-1009/S video camera (AVT Horn; Aalen, Germany).

Preparations for studying skeletal elements were air dried and carried out according to previously described methods [11] and observed in transmission JEM-1011 (Jeol, Tokyo, Japan) and scanning JSM-6510 LV (Jeol, Tokyo, Japan) electron microscopes. The acceleration voltage was 80 kV for TEM and 15–30 kV for SEM. Skeletal elements were measured using ImageJ 1.52a software [41].

Molecular phylogeny

The centrohelid heliozoans were grown in a clonal culture and collected by centrifugation (1000 × g, room temperature) onto the 0.8 μm membrane of a Vivaclear mini column (Sartorius Stedim Biotech Gmng, VK01P042) (HF-58Z strain) or collected on a polycarbonate membrane filter (pore size of 1 μm) using a 60 ml plastic Luer-Lok™ syringe (BD, New Jersey, Cat. No. REF 309653) and a 25 mm filter holder (Swinnex, Millipore) (HF-68Z strain). Genomic DNA was isolated using the MasterPure™ Complete DNA and RNA Purification Kit (Epicentre, Madison, WI, USA, Cat. No. MC85200). The 18S rRNA gene was amplified using EconoTaq Plus Green 2 Master Mix (Lucigen, Middleton, WI, USA, Cat. No. 30033–1). The eukaryote-specific forward primer Thx25F 5′-CAT ATG CTT GTC TCA AAG ATT AAG CCA-3′ [42] and the centrohelid-specific reverse primer Helio1979R 5′-CAC ACT TAC WAG GAY TTC CTC GTT SAA GAC G-3′ [7] were used for the HF-58Z and HF-68Z strains. To amplify the 18S rRNA gene of the HF-80Z strain, a nested PCR approach was used. Twenty centrohelid cells were picked out using a glass micropipette and added to a PCR mixture containing the forward primer Thx25F and the reverse primer 1801R 5′-TGA TCC TTC TGC AGG TTC ACC T-3′ [43]. The amplification product was used for a second (nested) PCR using the eukaryote-specific forward primer GGF 5′-CTT CGG TCA TAG ATT AAG CCA TGC-3′ and reverse primer GGR 5′-CCT TGT TAC GAC TTC TCC TTC CTC-3′ [44]. The PCR program for amplification was as follows: initial denaturation at 95°C for 3 min; 35 cycles of 95°C for 30 sec, 52°C for 30 sec, and 72°C for 1.5 min; and a final extension at 72°C for 5 min.

The amplified DNA fragments were purified with a QIAquick PCR Purification Kit (Qiagen, Hilden, Germany, Cat. No. 433160764).

The PCR products of the HF-58Z and HF-80Z strains were sequenced via Sanger dideoxy sequencing at Evrogen, Moscow, Russia. The PCR products of the HF-68Z strain were sequenced via Sanger dideoxy sequencing using Applied Biosystems 3500 (Applied Biosystems, USA; Hitachi, Japan) at the Papanin Institute for Biology of Inland Waters, RAS. The PCR products for all the strains were sequenced in one replicate, with the exception of strain HF-80Z, which was sequenced in two replicates with 100% sequence identity, using the primers mentioned above and two universal internal primers, 18SintF (forward): 5′-GGT AAT TCC AGC TCC AAT AGC GTA-3′ and 18SintR (reverse): 5′-GTT TCA GCC TTG CGA CCA TAC T-3′. The resulting sequences were assembled from four overlapping reads using the Geneous R6 ver. 6.0.6 program. The 18S rRNA gene sequences generated in this study were deposited in the NCBI GenBank database under the following accession numbers: PQ530865– PQ530867.

Multiple sequence alignment was constructed using the L-INS-i algorithm in MAFFT ver. 7.475 [45] and trimmed using an automated trimming heuristic followed by a gap threshold filter of 0.7 in TrimAl ver. 1.2 [46]. The aligned sequences were visualized and checked with AliView ver. 1.28 [47]. The resulting alignment consisted of 1586 sites and was used to construct the phylogenetic trees. Phylogenetic trees were reconstructed using Bayesian and maximum likelihood (ML) methods.

The maximum likelihood phylogeny was inferred using IQ-TREE ver. 1.6.12 [48], with 1000 nonparametric bootstrap pseudoreplicates under the best fit model (TN + F + R3 model) determined by the in-built ModelFinder [49].

To infer the 50%-majority rule Bayesian phylogenetic tree, the parallel MPI version of MrBayes ver. 3.2.7a [50] was used with four categories of gamma-distributed among-site rate variation under the GTR + I + GAMMA substitution model. To calculate the posterior probabilities of individual nodes, four independent Metropolis-coupled Markov chains were run for 20 million generations and summarized after the 50% burn-in. The convergence of log-likelihood values and model parameters for chains was verified using a plot and convergence diagnostics provided by the MrBayes sump utility. The average standard deviation of the bipartition frequencies was 0.0046 at the end of the Markov chain Monte Carlo simulations.

Cyanobacteria feeding experiment

An experiment was conducted to test the ability of raphidocystid heliozoans to consume cyanobacterial cells. Initially, the centrohelids Raphidocystis tubifera HF-68Z, R. marginata HF-64Z, and R. symmetrica HF-80Z were grown on Parabodo caudatus. After consuming the bodonid prey and increasing the number of cells, the raphidocystid cells (50–70 cells) were transferred to Petri dishes containing 3 ml of the cyanobacterium Microcystis aeruginosa (toxic strain 905, ~ 784123 cells ml^-1), Microcystis aeruginosa (nontoxic strain 928, ~ 596204 cells ml^-1), and Aphanizomenon sp. strain 1399 (2703 colonies/filaments ml^-1), respectively. The experiment was not carried out for R. ambigua HF-58Z due to the loss of the culture.

Phylogenetic position

The three molecularly characterized centrohelids in this study belong to the family Raphidocystidae: Panacanthocystida (Fig 7). Raphidocystis tubifera HF-68Z is grouped with full support with the unidentified marine centrohelid Helio10 (AY749619) from Whale Bay, Raglan, New Zealand, with which they are 96.5% similar. Two environmental sequences from the rainforest of Panama are grouped together in a sister clade.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Bayesian phylogenetic tree generated from 18S rRNA gene sequences of 42 centrohelids (seven of which represent the outgroup).

Bayesian posterior probabilities (BPP) and Maximum Likelihood (ML) bootstrap values are indicated at branches (BPP > 0.50 and ML bootstrap > 50% are shown; dt – different topology). Filled circles of different colours indicate values of BPP = 1.00 and ML bootstrap = 100%. The sequences generated in this study are highlighted in bold, colored in red and marked with black stars symbols. The percentage of similarity of the sequences is indicated to the right of the names of the sequences. Abbreviations: B, brackish environment: F, freshwater environment; M, marine environment; S, soil environment.

https://doi.org/10.1371/journal.pone.0322585.g007

In the sister group to the clade of the four above-mentioned, R. symmetrica HF-80Z clustered with full support with the freshwater sequences H18.6 (AY749501) and “R. brunii” (AY749618). A discussion of the cospecificity of these strains is given below. BLASTn analysis revealed 98.4% similarity between HF-80Z and H18.6 and 93.3% similarity with “R. brunii”.

The HF-58Z strain with complete support was grouped within a clade with other representatives of Raphidocystis ambigua from the United Kingdom (AY305008 and AF534708). The similarity between these sequences is 98.9%. The environmental sequences from the soils of the USA and Panama were placed more basally.

Feeding of raphidocystids on the toxic and nontoxic freshwater cyanobacteria

All tested strains of raphidocystids showed significant uptake of cells of both the toxic and nontoxic strains of Microcystis aeruginosa (Fig 8). At the same time, the density of the cyanobacterial cell suspension did not noticeably decrease. The death of heliozoans after their transfer to cyanobacteria was not recorded. Cell divisions of heliozoans were not observed in cyanobacterial cultures. On the 10th day of observation, in all the cases, active cells of raphidocystids with elongated axopodia were recorded. When the feeding intensities of raphidocystids on bodonid and cyanobacterial cultures are compared, more active feeding is observed on the flagellates Parabodo caudatus. In this case, raphidocystid cells, which simultaneously consume up to ten bodonid cells, were observed. During feeding on cyanobacteria, despite their high density, a maximum of 3–4 cyanobacterial cells were simultaneously consumed by heliozoan cells. At the same time, interactions between the kinetocysts of heliozoan axopodia and the sheaths of cyanobacteria have never been observed. It is likely that the absorption of cyanobacteria can occur through close contact of their cells with the surface of the centrohelid cell body without anchoring the cyanobacterial cells by the kinetocysts of axopodia, probably due to the presence of a complex cell wall and extracellular covering in Microcystis, which includes a peptidoglycan layer and polysaccharide fibrils [99]. In an experiment using Aphanizomenon sp., feeding on cyanobacteria was not recorded for any species of raphidocystid heliozoans. At the same time, the trichomes of Aphanizomenon sp. were used by raphidocystids as a substrate on which they anchored.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Feeding of raphidocystid heliozoans by cyanobacteria.

(A1, B1, C1, A3, B3, C3 in DIC; A2, B2, C2, A4, B4, C4 in bright field). A1–A4 – R. tubifera HF-68Z with ingested cells of M. aeruginosa; B1–B4 – R. marginata HF-64Z with ingested cells of M. aeruginosa; C1– C4 – R. symmetrica HF-80Z with ingested cells of M. aeruginosa. Red arrows show ingested cyanobacterial cells. Scale bars: 10 μm.

https://doi.org/10.1371/journal.pone.0322585.g008

Morphology and phylogeny of raphidocystidae

A comparison of the literature data revealed some differences between strain HF-68Z and previously studied R. tubifera strains. Thus, previous studies reported larger Sc.1 scales in other samples, ranging from 15–25 μm [57], 20.0–37.5 μm [65], 9.70–46.53 μm [70], and up to 39 μm in the meromictic lake Trekhtsvetnoe (White Sea) (Y. Mindolina, personal comm.). Our study also revealed larger plate scales, with lengths reaching up to 9.50 μm and widths up to 4.54 μm, than those reported in previous studies by Croome [72], Mikrjukov [62], and Rees et al. [57]. These studies reported maximum lengths of 7.5 μm, 4.8 μm, and 6.1 μm, and maximum widths of 2.3 μm, 2.0 μm, and 3.1 μm, respectively.

The scale of the transitional shape compared with Sc.2 and plate-scales (Fig 2D) is marked in Fig 5 in [51]. Notably, in samples studied directly from the environment [57,70,72 etc.], the amount of Sc.1 is usually greater than that in samples cultivated in the laboratory for 4 years (HF-68Z strain).

Sequencing of the studied clones provided the first insights into the phylogenetic placement of R. tubifera within the centrohelids. The assumption of Zlatogursky et al. [8] about the phylogenetic position of R. tubifera close to R. ambigua and ‘R. brunii’, which possess plate scales with a reticular texture, was confirmed.

It is worth noting that the sequences closest to R. tubifera were isolated from significantly different biotopes. Strain HF-68Z was isolated from a freshwater biotope, sequence Helio10 from seawater, and the sequences LN582189 and LN583237 from the tropical forest of Panama. The vast majority of findings of R. tubifera were made in freshwater biotopes. However, Tong [75] noted two scales of R. tubifera from Southampton water (salinity ~ 28‰) north of the Solent. Apparently, they may belong to the freshwater heliozoans brought by the River Test current. Nonetheless, the presence of the close sequence of ‘Helio10’ from the sea waters of Whale Bay, New Zealand, allows us to assume the existence of a closely related species to R. tubifera, adapted to living in sea water.

The type species of the genus Raphidocystis, R. lemani (Penard, 1891) Penard, 1904, is quite morphologically similar to R. tubifera and is not often noted by researchers. The morphology of the covering scales, as well as their variability, is poorly studied in R. lemani. This species must be isolated in clonal culture to obtain molecular phylogenetic and detailed morphometric data. On the basis of the available morphological data, this species is expected to form a lineage close to R. tubifera in future phylogenetic reconstructions.

The revealed morphological characteristics of clone HF-58Z generally coincide with those described previously for R. ambigua [1,2 etc]. This species is reported for the first time for the water bodies of Kazakhstan. The similarity of the sequence of our strain with the available R. ambigua sequences is also very high (98.9%). The sequences AY305008 and AF534708 belong to the same strain from River Cam, Cambridgeshire, UK. It was deposited in the CCAP culture collection (1568/1) and subsequently sequenced independently twice [3,100]. The obtained sequences of the strains were 99% similar to each other, with 11 nucleotide substitutions out of 2209 nucleotides. NCBI also currently includes the sequence LC682598 from Wake city, Okayama, Japan, which is annotated as Raphidocystis ambigua. The length of the LC682598 sequence is significantly shorter than that of sequences from the UK and Kazakhstan (1240 bp). When the tree is reconstructed, this sequence forms a long branch, grouping within a clade with other Raphidocystis ambigua sequences.

A number of problems have been accumulated regarding Raphidocystis marginata and morphologically similar centrohelids. This species was first described in 1981 by F. Siemensma on the bases of light microscopy (as Raphidiophrys marginata). Soon after, Nicholls and Dürrschmidt [51] reported specimens with slightly larger scales, but well consistent with the shape and structure of the scales of R. marginata and studied them using electron microscopy. Siemensma and Roijackers [28] subsequently confirmed the morphology refined by Nicholls and Dürrschmidt by obtaining electron microscopic data for this species from the Netherlands. Soon, information appeared about brackish-water (Baltic Sea, 6‰, [91]) and marine (35‰, [92]) forms, which were also identified as R. marginata.

In 1995, Kinoshita and colleagues described a new species, Raphidiophrys contractilis, similar to R. marginata, from a brackish water pond (3.5‰), justifying the novelty by the smaller scale and cell sizes and by the absence of broadly rounded poles of plate-scales, although in the provided Figs 4–5, poles are broadly rounded [101]. Mikrjukov in his monograph [2], carried out a revision, transferring R. marginata and R. contractilis to the previously described genus Polyplacocystis as synonyms of Polyplacocystis coerulea (Penard, 1904) Mikrjukov et Croome, 1998. The latter was described from the fresh waters of Switzerland as a small (13–14 µm) heliozoan [27], but there were no data on the structure of its scales. It was noted that the cell covers consist of small ‘sequins’ or very thin elongated rods that are straight or slightly curved [27]. Due to the fact that neither the shape of the scales nor their sizes nor any of their elements (e.g., margins, ribs, surface texture, etc.) were described and the type was not preserved, the specimen found by Penard is currently impossible to adequately compare with any known species. After the phylogenetic position of the genus Raphidocystis was clarified [8], both Raphidiophrys marginata and Raphidiophrys contractilis were transferred to Raphidocystis.

Analysis of the currently available morphological data does not allow us to clearly draw a boundary between the abovementioned raphidocystid species, since the quality of the data varies greatly among different authors. A number of works present only single scales without describing the morphology of the covering elements [53,55,56,60,66,67,71,88,89,93]. Other works contain only images of living cells [33,90], whereas others present the morphometry of single specimens [70].

The morphometric characteristics of the studied R. marginata strain are similar to those previously reported [28,51]. Sharp differences in scale morphology were noted for a strain from a marine aquarium studied by Mikrjukov [92]. This strain is characterized by a more elongated and narrow shape, greater length (8.0–9.0 µm) and smaller width (1.7–2.0) and, accordingly, a greater length-to-width ratio (3.4–5.4 [measurements according to Fig 4 in [92], n = 11] versus 1.9–2.9 in our study). In our opinion, it is another species of Raphidocystis closely related to R. marginata. Another strain isolated from the Black Sea and considered R. marginata by Mikrjukov (Fig 11 in [29]), in our opinion, also belongs to other species owing to an even greater length-to-width ratio (8.7, measurements according to Fig 11 in [29]; n = 1]). In addition to differences in the morphology of the covering elements, the abovementioned species were isolated from marine waters (35‰ and 18‰, respectively), whereas the HF-64Z clone is freshwater.

Although R. marginata strain HF-64Z and R. contractilis strain NIES-2498 [101,102] have similar plate scales, the size limits for both scale length and width and cell diameter are lower in R. contractilis (Fig 9). The length-to-width ratio in the original description of R. contractilis is almost the same as that of the HF-64Z strain (2.0–3.0), but in the subsequent work [102], the microphotographs demonstrate a significantly greater length-to-width ratio (up to 3.8). Therefore, considering the differences in the size and shape of the scales and the differences in the mineralization of the biotopes from which the strains were isolated, we are not inclined to equate these species.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Comparison of the morphometric parameters of Raphidocystis contractilis strain Nies-2498 (Wan et al. 2023) and R. marginata strain HF-64Z.

https://doi.org/10.1371/journal.pone.0322585.g009

The morphology of ‘Polyplacocystis’ coerulea revealed by Plotnikov and Ermolenko [65] is closer to that of strain HF-64Z, but this specimen was found in a lake with higher salinity (8‰) than our strain and R. contractilis sensu Kinoshita et al. (1995). Therefore, to gain a deeper understanding of the taxonomy and phylogeny of raphidocystids, along with molecular genetic data, salinity experiments are needed to test whether freshwater strains are able to successfully survive and reproduce in brackish-water environments and, vice versa, brackish-water strains in freshwater conditions.

We observed thin ribs on the scales of our R. marginata strain, a feature not previously described in the literature. Although not previously reported, these structures are clearly discernible in the provided images (Fig 4D, [89]; Fig 28, [28]).

The morphometric parameters of the studied R. symmetrica strain also generally correspond to those previously described for other strains. However, the cells we studied possess slightly larger lengths, up to 10.45 μm versus 9.6 μm in [86], and 8.4–8.7 μm in other studies [12,70,83,96,97], which can probably be explained by the larger sampling size of the current study.

The average size of the well-fed R. symmetrica HF-80Z cells was 20.7 µm, matching the average size in the original description (20–25 µm), which was based on light microscopy data only [27]. The general appearance of the cells we studied and the arrangement of their scales at the bases of the axopodia also correspond to those shown by other authors (Fig 10). Penard [27] described two types of scales: long scales located at the base of the axopodia and short scales covering the cell body. In the strain we studied, the distributions of scales of different lengths were the same, but no division of scales into two clear size classes was revealed.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 10. General view of living cells of Raphidocystis symmetrica.

A – R. symmetrica HF-80Z strain (PhC), the inset shows plate-scales (DIC); B – drawing of R. symmetrica (182 p., as Raphidiophrys symmetrica) by Penard [27]; C – drawing of R. symmetrica (77 p., Plaat 19, as Raphidiophrys symmetrica) by Siemensma [61]. Scale bar: A –10 μm, inset – 5 μm.

https://doi.org/10.1371/journal.pone.0322585.g010

The strain of R. symmetrica we studied forms a trichotomy with the environmental sequence AY749501 and the strain identified as ‘R. brunii’ on the phylogenetic tree. The latter species was identified by Cavalier-Smith and Chao [103] on the basis of cell size and the number of axopodia. The 18S rRNA sequence of this strain was obtained previously [7]. The available microphotographs (Fig 8 in [103]; Fig 4 in [7]) show that the cell covers consist of one type of scales with a reticular surface texture, characteristic of R. ambigua, R. tubifera and R. symmetrica. Unfortunately, the original description of R. brunii [104] does not contain enough information about the scale structure, so it is difficult to accept the conclusion of Cavalier-Smith and Chao [103] about the cospecificity of the clone they studied and R. brunii. Cavalier-Smith and Chao [103] do not provide scale sizes, preventing comparison with the strain we studied. R. brunii is characterized by a cell size of 12–13 µm according to the original description [104]. In the existing figure (Fig 8 in [103]), the cell diameter is approximately 12 μm, which also corresponds to the lower limit of the cell diameter of the strain we studied. The similarity of the 18S rRNA gene sequences of the clone HF-80Z and the so-called “R. brunii” is only 93.3%, suggesting that they are two different species since the difference threshold for many described centrohelid species is typically on the order of 3%. It should be noted that Mikrjukov [105] and Prokina et al. [106] considered marine centrohelids with very narrow scales as R. brunii (R. brunii sensu Mikrjukov 1999 and R. brunii sensu Prokina et al. 2021).

Importantly, species with a reticular scale surface structure (R. tubifera, R. ambigua, R. symmetrica and R. brunii sensu Cavalier-Smith and Chao 2012) form a distinct clade on the Raphidocystidae tree. In contrast, species with smooth surfaces of scales (R. pallida, R. contractilis, and R. glabra) are placed closer to the base of the raphidocystid tree. Interestingly, when R. ambigua forms cysts, it can also form plate scales without a reticular texture [107]. Therefore, scales with smooth surfaces are apparently characteristic of the common ancestor of all Raphidocystidae, which confirms the view of Drachko et al. [108], and scales with a reticular structure were acquired evolutionarily later.

Cysts within Raphidocystidae species were apparently a later evolutionary acquisition too. Only R. tubifera, R. ambigua and R. symmetrica possess the ability to encyst. While in R. pallida, R. contractilis and R. glabra, the presence of cysts has not been revealed. Cyst formation was also not observed in R. marginata in this study. The ability to form cysts in raphidocystids was likely acquired in connection with the transition to life in conditions with variable moisture levels. For example, it is known that R. tubifera, R. ambigua and R. symmetrica can survive in the conditions of temporary waterbodies (HF-68Z in this study), dry mosses (R. ambigua [58]), and soils (R. symmetrica, unpublished data).

Despite the fact that representatives of Raphidocystidae (e.g., Raphidocystis glabra and R. contractilis) are characterized by the replacement of siliceous skeletal elements with organic spicules during their life cycle [8,102], this phenomenon was not recorded in our study.

Feeding on cyanobacteria

The results of the feeding experiment revealed the ability of centrohelids of the genus Raphidocystis to use cyanobacteria as food. That is, the abundance of cyanobacteria may potentially be subject to ‘top-down control’ by raphidocystids in aquatic communities. Similarly, feeding on cyanobacteria has been shown for the naked amoebae Mayorella [109], Hartmannella, Vexillifera, Vannella, Korotnovella, Copromyxa [110], Acanthamoeba castellanii [111], Naegleria [112], heterolobose Vrihiamoeba [113], vampirelids [114] and some others. It has been found that mixotrophic representatives of the chrysophytes Ochromonas and Poterioochromonas malhamensis (Pringsheim 1952) Andersen, Graf, Malakhov et Yoon, 2017 can quickly consume toxic Microcystis cells and effectively destroy both intracellular and extracellular microcystins [38,115–120]. Interestingly, heterotrophic protists can be highly resistant to cyanotoxins and exhibit feeding strategies that differ significantly from those of metazoans [39]. Thus, heterotrophic protists have great potential for controlling harmful cyanobacterial blooms and restoring the phytoplankton composition in eutrophic waters. Regarding the results of this study, it should be emphasized that raphidocystids do not consume the filamentous colonies of Aphanizomenon sp., and although they are able to consume Microcystis cells, their number is not significantly reduced. Further research is needed to examine the conditions that increase cyanobacterial feeding in Raphidocystis. This could be useful for the development of biological methods and technologies for the control of toxic cyanobacterial blooms, which are potentially less harmful to the environment and may represent a cost-effective alternative to physical and chemical methods.