Introduction

The genus Tetrahymena Furgason, 1940 comprises not only free-living representatives, which occur in a variety of freshwater and soil habitats (e.g., Corliss 1973; Doerder 2014, 2019; Foissner 1987, 1998, 2000; Pitsch et al. 2017; Quintela-Alonso et al. 2013; Zahid et al. 2014), but also endoparasites or commensals, which inhabit a variety of animals including planarians, mollusks, insects, fish, and even mammals (e.g., Corliss 1960; Lynn et al. 2000, 2018; Rataj and Vďačný 2020; Van As and Basson 2004). The common morphological feature of this genus is the oral apparatus that is composed of four ciliary structures, i.e., a paroral membrane and three oral polykinetids (Furgason 1940). Corliss (1969, 1970) provided the first insights into the intrageneric taxonomy of Tetrahymena and recognized three groups: (1) the pyriformis complex, containing bacteria-feeding microphagous ciliates with tendencies to become parasites; (2) the rostrata complex, gathering facultative histophagous and/or parasitic species often dividing inside of a cyst; and (3) the patula complex of non-parasitic, weakly histophagous but occasionally cannibalistic macrostome forms. However, Corliss’ three complexes were useful constructs at the time, and they have long since been abandoned as they have no phylogenetic relevance. Moreover, distinguishing species within the three complexes based solely on morphological characteristics was very difficult because most Tetrahymena species share a rather similar body shape and size as well as a uniform somatic ciliature and overlapping numbers of ciliary rows. Therefore, DNA barcoding based especially on the mitochondrial gene encoding for cytochrome c oxidase subunit 1 (COI) was employed to identify morphologically highly similar species that also might possess identical 18S rRNA gene sequences (e.g., Doerder 2019; Lynn and Strüder-Kypke 2006; Rataj and Vďačný 2020; Simon et al. 2008). Molecular taxonomic methods became a standard and a quite unambiguous method for separation of Tetrahymena species, as the intraspecific variation is typically up to 2% while the interspecific variation ranges from about 4 to 11% in the COI gene (Chantangsi et al. 2007; Doerder 2014, 2019; Kher et al. 2011; Lynn and Doerder 2012; Rataj and Vďačný 2020). Despite the intensive research in the past decades, it is still hard to determine the overall species richness within the genus Tetrahymena, as many habitats and potential host organisms remain to be studied.

Although Tetrahymena infections have been comparatively rarely recorded, there are distinctly more reports from aquatic than from terrestrial animals. More specifically, tetrahymenid endoparasites were already found in freshwater planarians (Rataj and Vďačný 2020; Reynoldson and Bellamy 1973; Wright 1969, 1981), in black flies, alderflies, mosquitoes, and midges (Batson 1983, 1985; Corliss 1960; Jerome et al. 1996; Lynn et al. 1981), in larvae of freshwater bivalves (Lynn et al. 2018; Prosser et al. 2018), in adult crayfish (Edgerton et al. 1996), and in a broad variety of fish ranging from guppies (e.g., Hatai et al. 2001; Hoffman et al. 1975; Imai et al. 2000) to salmons (Ferguson et al. 1987). In contrast, there is just a single report of Tetrahymena parasitizing terrestrial vertebrates, namely T. farleyi Lynn et al., 2000, which was isolated from the urine of a Dalmatian dog (Lynn et al. 2000). As concerns terrestrial invertebrates, Tetrahymena infections have been mostly reported from various slugs and snails (for a review, see Van As and Basson 2004). Nevertheless, mollusks still represent an insufficiently explored host-group of Tetrahymena species (Antipa and Small 1971; Corliss 1960; Prosser et al. 2018; Van As and Basson 2004). In the present contribution, we examined 150 specimens belonging to 19 bivalve and gastropod species. Tetrahymena species were, however, detected only twice. Analyses of their 16S and 18S rRNA genes as well as of their COI gene demonstrated significant genetic differences, justifying the establishment of two new species, T. foissneri sp. n. and T. unionis sp. n.

Material and methods

Material collection and processing

Altogether, 105 terrestrial mollusks belonging to 11 species were collected at 10 localities, and 45 aquatic mollusks belonging to eight species were sampled at five localities in Slovakia, Central Europe (Supplementary Table S1). Tetrahymenid ciliates were detected only in two species, viz., in the invasive Lusitanian slug, Arion vulgaris (Moquin-Tandon, 1855), collected from a sidewalk at the Stratená ulica street, Bratislava (48° 12′ 11″ N, 17° 09′ 04″ E) on November 11, 2019, and in the swollen river mussel, Unio tumidus (Philipsson, 1788), collected from the River Hornád near the village of Milhosť, south of the city of Košice (48° 31′ 45″ N, 21° 18′ 20″ E) on May 10, 2020. Sampled mollusks were transferred together with in situ substrates to the laboratory at the Department of Zoology, Comenius University in Bratislava. Slugs were directly dissected; their digestive (esophagus, stomach, digestive gland, and gut), reproductive (albumen gland, ovotestis), and excretory (kidney) organs were extracted and investigated for the presence of ciliates. The mantle cavity of mussels was rinsed with water using a micropipette before dissection. Ciliates were found in great abundance in water expelled from the mantle cavity and the excurrent siphon of mussels when they were placed in Petri dishes containing tap water. The extruded water contained much organic debris and peeled off epithelial tissues as well. After rinsing the mantle cavity, mussels were dissected, and their internal organs were examined for the presence of ciliates. Tetrahymenid ciliates were manually picked with the aid of Pasteur micropipettes adjusted as described by Foissner (2014). Cultures of both isolated Tetrahymena species were established in Petri dishes, using tap water as a medium and pieces of host tissues as a food source. Cultures were maintained for several months in a refrigerator at a temperature of 6 °C. This temperature was chosen, as both species were sampled during cooler parts of the year. No attempts to establish axenic cultures were made, as both Tetrahymena species thrived on host tissues.

Taxonomic and molecular methods

Living ciliates were investigated under an optical microscope Zeiss Axio Imager 2 (Carl Zeiss, Oberkochen, Germany) equipped with differential interference contrast optics at low and high magnifications. The ciliary and silverline pattern and the nuclear apparatus were revealed with the dry silver nitrate (Klein 1958) and protargol (Wilbert 1975) impregnation methods. Only fresh material was used for living observations and all staining procedures, as some morphological features might change during cultivation. Microphotographs were captured by a Canon EOS 80D camera (Canon Inc., Ota City, Tokyo, Japan). In vivo measurements were based on microphotographs and were carried out with the help of the calibrated software ImageJ ver. 1.49 (Schneider et al. 2012). Measurements on mounted specimens were conducted with an ocular micrometer. Illustrations were prepared in Inkscape ver. 0.92.4 and Adobe Photoshop CC. Terminology mostly follows Lynn (2008).

Individual cells of both tetrahymenid species were put in 180 μl of cell lysis buffer (Promega, Fitchburg, WI, USA) and stored at 6 °C pending DNA extraction. Five single-cell samples were prepared for T. foissneri sp. n. and four single-cell samples for T. unionis sp. n. The extraction of genomic DNA, PCR amplification, and sequencing followed our previous study on tetrahymenids parasitizing planarians (Rataj and Vďačný 2020). PCR primers are summarized in Table 1 and PCR conditions are listed in Table 2. Electropherograms of the newly obtained sequences were carefully inspected in Chromas ver. 2.6.6 (Technelysium Pty Ltd., South Brisbane, Australia), trimmed and assembled into contigs using BioEdit ver. 7.2.5 (Hall 1999).

Table 1 Primers used for amplification of three molecular markers analyzed in two new Tetrahymena species parasitizing slugs and mussels
Full size table
Table 2 Conditions of PCR reactions used for amplification of three molecular markers analyzed in two new Tetrahymena species parasitizing slugs and mussels
Full size table

Type and localization of unique nucleotides suitable for taxa diagnoses were searched for in the barcoding region of the COI gene, using a custom Python script. The query group contained all specimens of the species in question, while the reference group comprised their nearest relatives, as revealed by the BLASTn search and the present phylogenetic analyses. The reference alignments are provided in Supplementary material (Alignments 1 and 2). Only nucleotide states that were shared by all members of the query group and were different from the states in the reference group were used as diagnostic characters.

The ZooBank registration number of the present work is urn:lsid:zoobank.org:pub:41559C77-2CB2-4AA3-BD58-05C9974CD5B3.

Phylogenetic methods

To determine the phylogenetic positions of the two new Tetrahymena species, four datasets were assembled. The first dataset contained 108 concatenated sequences of the nuclear 18S rRNA gene and the mitochondrial COI gene (for taxon sampling and GenBank accession numbers, see Supplementary Table S2). The second alignment contained 114 COI sequences, and the third alignment consisted of 95 sequences coding for the 16S rRNA molecule (GenBank entries are shown in respective figures). The fourth dataset consisted of 30 concatenated sequences of two mitochondrial markers, the 16S rRNA and the COI gene (Supplementary Table S3). Taxon sampling in all datasets followed Doerder (2019) and Rataj and Vďačný (2020).

16S and 18S rRNA gene sequences were aligned using the MAFFT algorithm on the GUIDANCE2 server (http://guidance.tau.ac.il/ver2/) (Sela et al. 2015), with parameters as specified in Zhang and Vďačný (2020). The protein-coding COI gene sequences were aligned, using amino acid sequences as predicted with the invertebrate mitochondrial genetic code in MEGA X ver. 10.2 (Kumar et al. 2018). The COI alignment was divided into three partitions, each representing one codon position, using a custom Python script. Three different algorithms were used to build phylogenetic trees: neighbor-joining (NJ) as implemented in MEGA X (Kumar et al. 2018), maximum likelihood (ML) as implemented in IQ-Tree ver. 1.6.10 (Nguyen et al. 2015), and Bayesian inference as implemented in MrBayes ver. 3.2.7 (Ronquist et al. 2012). NJ analyses included the maximum composite likelihood method, gamma-distributed rates among sites, a heterogeneous pattern among lineages, a pairwise deletion option to exclude alignment gaps, and 5000 bootstrap replicates. ML analyses had the following settings: a starting BioNJ tree; each partition was allowed to have a different speed; the best fitting evolutionary substitution models were selected for each partition under the BIC criterion using the in-built ModelFinder program; 1000 ultrafast bootstrap replicates; and the bnni algorithm to reduce overestimating nodal support (Hoang et al. 2018). Bayesian inference included Markov chain Monte Carlo (MCMC) simulations, which were run for three to ten million generations with a sampling frequency of 100. Two different programs, jModelTest ver. 2.1.10 (Darriba et al. 2012) and IQ-Tree (Nguyen et al. 2015), were used to determine the best evolutionary model and its parameters for each molecular marker/codon partition. Parameters of the evolutionary models used are listed in Supplementary Table S4 and S5. Convergence of MCMC analyses was inspected using the in-built diagnostics of MrBayes. Results of convergence analyses are summarized in Supplementary Table S6. Given the results of these analyses, we present here Bayesian trees inferred from the first, third, and fourth datasets under the IQ-Tree evolutionary models and a Bayesian tree inferred from the second dataset under the jModelTest evolutionary models. The first 25% of sampled trees were considered as the burn-in fraction, and they were excluded before the construction of 50%-majority rule consensus trees. ML and Bayesian analyses were conducted on the CIPRES portal ver. 3.1 (http://www.phylo.org/) (Miller et al. 2010). All trees were computed as unrooted and were rooted using the outgroup taxa in FigTree ver. 1.2.3 (http://tree.bio.ed.ac.uk/software/ figtree/).

Bayesian delimitation of four closely related Tetrahymena species within the ‘paravorax’ clade (T. glochidiophila, T. pennsylvaniensis, T. nigricans, and T. unionis) was performed with the program BP&P ver. 2.2 (Yang 2015). The coalescent analyses were based on two markers, the nuclear 18S rRNA gene and the mitochondrial COI gene. Species delimitation together with estimation of a species tree was conducted using the following settings: a gamma prior for the θ parameter at G(1, 660) and a gamma prior for the τ parameter at G(14, 520); divergence times and rates among individual markers as estimated from the Dirichlet distribution (Yang and Rannala 2010); heredity scalars of individual markers as estimated from the data using a gamma prior G(4, 4); and a large fine-tuning parameter (ε = 15) assigned to the reversible jump (rj) algorithm to guarantee a good mixing. The θ and τ parameters were estimated by running an A00 analysis. The posterior distribution and mean of both parameters at the root were calculated from the log file in Tracer ver. 1.7 (Rambaut et al. 2018). Consequently, the posterior distributions were fitted with a gamma distribution. During rjMCMC simulations, every second iteration was taken, the first 100,000 samples were discarded as burn-in, and one million samples were saved. The maximum clade credibility species tree was calculated in TreeAnnotator ver. 2.6.0 (Bouckaert et al. 2014).

Results

Description of Tetrahymena foissneri sp. n.

ZooBank registration number: urn:lsid:zoobank.org:act:8EE817FB-292F-440B-B79D-B6EB31E4783B.

Diagnosis: Size about 35–58 × 20–37 μm in vivo. Body narrowly ovoidal to ovoidal, anterior end acute, posterior end broadly rounded. Macronucleus globular to ellipsoidal, a single globular micronucleus. Contractile vacuole dorsal and subterminal, two or three excretory pores. About 25 to 34 ciliary rows including two or three postoral kineties. Oral apparatus about 9–11 μm long, forming a Tetrahymena-like pattern. Silverline system composed of primary meridians connecting kinetids and secondary meridians running in parallel and regularly alternating with primaries. No caudal cilia.

Diagnostic molecular characters in COI gene (with respect to T. limacis EF070288): 3 C, 9 C, 12 C, 16 C, 32 G, 54 T, 55 C, 78 G, 114 T, 141 C, 168 A, 195 T, 201 C, 237 G, 252 T, 297 C, 300 T, 309 C, 321 G, 324 T, 333 A, 345 C, 348 T, 351 T, 354 C, 384 A, 414 A, 426 G, 468 C, 471 C, 484 T, 510 T, 515 C, 516 T, 517 G, 557 T, 561 T, 582 G, 606 C, 612 C, 613 T, 615 T, 672 A, 703 A, 708 G, 726 T, 765 A, 783 A, 786 G, 800 C, 804 A, 819 T, 831 T, 843 G, 847 G, 885 T, 909 T.

Type locality: A sidewalk at Stratená ulica street, Bratislava, Slovakia (48° 12′ 11″ N, 17° 09′ 04″ E).

Type host: Arion vulgaris (Moquin-Tandon, 1855).

Type material: A DNA sample of holotype specimen has been deposited in Natural History Museum, Vajánskeho nábrežie 2, 810 06 Bratislava, Slovakia (ID Collection Code 01427572).

Additional material: One paratype slide containing protargol-impregnated specimens (reg. no. 2021/1-ZTY) and one paratype slide containing silver nitrate-impregnated specimens (reg. no. 2021/2-ZTY) have been deposited at Department of Zoology, Comenius University in Bratislava, Ilkovičova 6, 842 15 Bratislava, Slovakia.

Gene sequences: The nuclear 18S rRNA gene, the mitochondrial 16S rRNA gene, and the mitochondrial cytochrome c oxidase subunit I sequences of the holotype specimen have been deposited in GenBank (https://www.ncbi.nlm.nih.gov/genbank/) under the following accession nos. MW827176, MW827185, and MW828683, respectively.

Dedication: We named this species in memory of an outstanding protistologist, Prof. Dr. Wilhelm Foissner (1948–2020), Paris Lodron University of Salzburg, Salzburg, Austria, as a small token of appreciation for guiding many students into ciliatology, for altruistic sharing of his great knowledge, and for his constructive criticism, which made people think and work scientifically.

Description: Body size in vivo approximately 35–58 × 20–37 μm, with a length:width ratio of about 1.3–1.8:1 (n = 13). Shape narrowly ovate to ovate, anterior end acute, posterior end broadly rounded; flexible but not contractile (Figs. 1a, d–g; 2a–c; and 3i, j). Macronucleus typically located in body center, globular to ellipsoidal in vivo and after methyl green staining, sometimes slightly deformed and with irregular surface after protargol impregnation, on average 12.3 μm in largest diameter after protargol impregnation; nucleoli numerous, globular, and evenly distributed over macronuclear surface. Micronucleus nearby or attached to macronucleus, globular and about 2.2–3.5 μm in largest diameter after protargol impregnation (Figs. 1a, c–h; 2d–f; 3a, b, d–f; and Table 3). A single conjugation pair observed after methyl green staining, each partner contained a degenerating macronuclear nodule and two or three maturation derivatives (Fig. 2g). One contractile vacuole located dorsally and distinctly subterminally, about 4.5–6.5 μm across during diastole; two or three excretory pores well recognizable after protargol impregnation (Figs. 1a, d–h; 2a, c, e, f; and 3e, h). Cytoplasm colorless, studded with innumerable lipid droplets about 0.3–1.5 μm across and some food vacuoles about 3.5–5.5 μm across (Figs. 1a and 2c, d). Cortex flexible, no cortical granules recognizable. Locomotion by slowly swimming and rotating about longitudinal body axis or rarely by gliding on the bottom of Petri dish. Feeds on host tissues and possibly also on bacteria in cultures.

Fig. 1
figure 1

Tetrahymena foissneri sp. n. from life (a, d–g), after protargol impregnation (b, c, h), and after dry silver nitrate impregnation (i–l). a Right view, showing the ovoid body. Black arrow indicates the acute anterior body end. b, c Ciliary pattern and nuclear apparatus of ventral and dorsal sides. Black arrowheads mark the anterior end of the ciliary rows which very likely begin with dikinetids (blue area) and white arrow shows the preoral suture. d–g Lateral overviews, showing the variability of body shape and size as well as of the nuclear apparatus, which consists of one macronuclear nodule (shaded grey) and a single micronucleus (shaded blue). Black arrows indicate the acute anterior body end. Drawn to scale. h Ciliary pattern of the right side. Black arrowhead marks the anterior dikinetids (blue area). i Ciliary and silverline pattern of the ventral side. j–l Details of the silverline system in the anterior (j) and posterior (l) body portions. The silverline system (k) consists of primary and secondary silverline meridians. Explanations = basal bodies (BB), contractile vacuole (CV), excretory pores (EP), macronucleus (MA), micronucleus (MI), oral apparatus (OA), paroral membrane (Pa), postoral kineties (PK), primary silverline meridian (PM), pellicular pores (PP), secondary silverline meridian (SM), somatic kineties (SK), adoral membranelles 1–3 (M1–M3). Scale bars = 2 μm (k), 8 μm (j, l), and 20 μm (a–i)

Full size image
Fig. 2
figure 2

Tetrahymena foissneri sp. n. from life (a–d) and after methyl green staining (e–g). a Overviews, showing some cells isolated from the gut of the slug Arion vulgaris. Arrows indicate the acute anterior body end. b Detail of the anterior body portion, showing the acute anterior body end (arrow). c Ventrolateral overview. Arrow indicates the narrowed anterior cell end. d Detail of the anterior body, showing the course of somatic kineties, the nuclear apparatus, and numerous lipid droplets. Arrow indicates the tapered anterior body end. e–g Overviews of the vegetative cells (e, f) and of a conjugation pair (g), showing the nuclear apparatus and contractile vacuoles. Explanations = adoral membranelles (AM), contractile vacuole (CV), degenerating macronuclear nodules (DMA), lipid droplets (LD), macronucleus (MA), micronucleus (MI), maturation derivatives (MT), paroral membrane (Pa), somatic kineties (SK). Scale bars = 12 μm (d), 20 μm (c, e–g), and 40 μm (a)

Full size image
Fig. 3
figure 3

Tetrahymena foissneri sp. n. after protargol (a–h) and after dry silver nitrate (i–m) impregnation. a, b Ciliary pattern and the nuclear apparatus of ventral (a) and dorsal (b) sides of a representative specimen. White arrow marks the preoral suture. c Detail of the oral apparatus, showing the paroral membrane and three adoral membranelles. White arrow marks the preoral suture. d Top view, showing the nuclear apparatus. e, f Right (e) and left (f) side overviews, showing the ciliary pattern and nuclear apparatus. g, h Detail of the anterior (g) and posterior (h) body portions. Black arrowheads mark the dikinetids at the anterior end of the ciliary rows. i, j Overviews, showing the silverline system on the ventral (i) and dorsal (j) sides. Black arrow indicates the acute anterior body end. k–m Details of the silverline pattern in the middle (k), anterior (l), and posterior (m) body portions. Red arrows indicate the short intermeridional connectives. Explanations = excretory pores (EP), macronucleus (MA), micronucleus (MI), paroral membrane (Pa), primary silverline meridian (PM), secondary silverline meridian (SM), somatic kineties (SK), adoral membranelles 1–3 (M1–M3). Scale bars = 7 μm (c, g, h, k–m) and 20 μm (a, b, d, e, f, i, j)

Full size image
Table 3 Morphometric data on Tetrahymena foissneri sp. n. (upper line) and Tetrahymena unionis sp. n. (lower line)
Full size table

Somatic cilia about 5–8 μm long in vivo, slightly more densely spaced anteriorly than posteriorly, arranged in 25–34 meridional rows (Table 3). Two or three postoral ciliary rows, begin just behind oral apparatus and extend to rear body end. Ventral ciliary rows start from almost anterior body end to mid-portion of oral apparatus, run right and left of buccal cavity to reach posterior body pole. Lateral and dorsal ciliary rows commence subapically leaving an unciliated apical area, follow body curvature to terminate in posterior body region. All ciliary rows consist of monokinetids, except for lateral and dorsal kineties which begin with a dikinetid each. Anterior dikinetids form a paratene (circle) around unciliated apical area (Figs. 1a–c, h–j, l; 2d–f; and 3a, b, e, g, l–j). A more or less distinct suture made by ventral ciliary rows, extends from apical body end to beginning of oral apparatus (Figs. 1b and 3a, c). Caudal cilium not observed in vivo, no basal bodies recognizable in posterior pole area after protargol and dry silver nitrate impregnation.

Silverline system composed of primary meridians connecting kinetids within somatic kineties and secondary meridians running in parallel and regularly alternating with primaries (Figs. 1i–l and 3i–m). Inconspicuous outgrowths (peaks) emerge on left side of primary meridians between each two kinetids. Secondary meridians unevenly dotted by deeply impregnated pellicular pores resembling basal bodies. Cross-silverlines loosely and irregularly branch off from primary meridians to connect with secondary meridians, forming rectangular and comparatively large meshes which resemble a platyophryid pattern (Fig. 1k).

Oral apparatus situated in a shallow cavity of ventral body side, i.e., in anterior fifth to third of body length, about 6.7–7.7 μm long after dry silver nitrate impregnation. Oral ciliature composed of a paroral membrane and three membranelles (oral polykinetids) as usual in tetrahymenids. Paroral membrane J-shaped and composed of very narrowly spaced dikinetids. Membranelle M1 longest (3.7 μm long after protargol impregnation), composed of four rows of basal bodies, anterior row about half of length of the three remaining rows. Membranelle M2 similar in fine structure to M1 but slightly shorter (3.3 μm long after protargol impregnation). Membranelle M3 inconspicuous because only ca. 2.0 μm long after protargol impregnation, composed of three or four rows gradually shortened from left to right (Figs. 1b, i, j; 2c; 3a, c, f, l; and Table 3).

Prevalence and intensity: T. foissneri was detected only in a single specimen of the Lusitanian slug A. vulgaris, which was sampled at the type locality. Although no signs of pathological changes were observed, a mass Tetrahymena infection (several hundred ciliates) was detected in the slug’s digestive tract. During the years 2019–2020, altogether, 24 A. vulgaris specimens, originated from three further localities in the surroundings of the type locality of T. foissneri, were examined for the presence of ciliates (Supplementary Table S1). However, none was colonized by tetrahymenids. Likewise, any other investigated slug (Deroceras sp., Limax maximus, and Tandonia kusceri) or snail (Alinda biplicata, Caucasotachea vindobonensis, Cepaea hortensis, Clausilia pumila, Cochlodina laminata, Helicodonta obvoluta, Helix pomatia, and Oxychilus sp.) was infected by Tetrahymena, indicating that T. foissneri might be a rare species.

Description of Tetrahymena unionis sp. n.

ZooBank registration number: urn:lsid:zoobank.org:act:6399B30E-AC7D-4125-892F-2B587C329D57.

Diagnosis: Size about 35–65 × 18–40 μm in vivo. Body narrowly ellipsoidal to broadly ovoidal, anterior end slightly tapered, posterior end broadly to narrowly rounded. Macronucleus globular to ellipsoidal, a single globular micronucleus. Contractile vacuole dorsal and subterminal, two excretory pores. About 24 to 33 ciliary rows including two postoral kineties. Oral apparatus about 10–14 μm long, forming a Tetrahymena-like pattern. Silverline system composed of only primary meridians with relatively long outgrowths (peaks). One long caudal cilium.

Diagnostic molecular characters in COI gene (with respect to T. glochidiophila MF693881, T. nigricans MN991323, and T. pennsylvaniensis KJ028745): 60 G, 111 A, 123 G, 165 G, 184 C, 189 T, 195 C, 249 C, 270 C, 294 C, 303 C, 339 G, 345 A, 348 T, 354 A, 358 C, 363 C, 369 A, 426 G, 513 A, 531 T, 552 T, 669 C, 672 A, 681 C, 711 T, 747 A, 810 C, 822 G, 825 A.

Type locality: River Hornád near the village of Milhosť, Slovakia (N 48° 31′ 45″, E 21° 18′ 20″).

Type host: Unio tumidus Philipsson, 1788.

Type material: A DNA sample of holotype specimen has been deposited in Natural History Museum, Vajánskeho nábrežie 2, 810 06 Bratislava, Slovakia (ID Collection Code 01427608).

Additional material: One paratype slide containing protargol-impregnated specimens (reg. no. 2021/3-ZTY) and one paratype slide containing silver nitrate-impregnated specimens (reg. no. 2021/4-ZTY) have been deposited at Department of Zoology, Comenius University in Bratislava, Ilkovičova 6, 842 15 Bratislava, Slovakia.

Gene sequences: The nuclear 18S rRNA gene, the mitochondrial 16S rRNA gene, and the mitochondrial cytochrome c oxidase subunit I sequences of the holotype specimen have been deposited in GenBank (https://www.ncbi.nlm.nih.gov/genbank/) under the following accession nos. MW827181, MW827190, and MW828688, respectively.

Etymology: The specific epithet is a singular genitive case of the Latin noun unio, union·is [f], meaning a Tetrahymena from Unio. The species-group name is to be treated as an adjective used as a substantive in the genitive case, because of its derivation from the host’s generic name (Art. 11.9.1.4. of the International Commission on Zoological Nomenclature 1999).

Description: Body size in vivo approximately 35–65 × 18–40 μm, with a length:width ratio of about 1.3–2.7:1 (n = 8). Shape narrowly elliptical to broadly ovate, anterior end slightly tapered, posterior end narrowly to broadly rounded; flexible but not contractile (Figs. 4a, b, e–g and 5a–c). Macronucleus usually located in body center, globular to ellipsoidal, on average 17 μm in largest diameter after protargol impregnation; innumerable circular nucleoli evenly distributed over macronucleus. A single micronucleus typically attached to posterior side of macronucleus, globular and about 2.3–4.4 μm across after protargol impregnation (Figs. 4a, b, d–g; 5h; 6a–f, j; and Table 3). One contractile vacuole situated dorsally and distinctly subterminally, about 4.5–7.0 μm in diameter during diastole; two interkinetal excretory pores each situated slightly right of ciliary rows (Figs. 4i, l and 6h). Cytoplasm colorless, contains numerous lipid droplets about 0.5 μm across and some to many food vacuoles up to 5.5 μm in diameter and usually containing pieces of host tissues (Figs. 4a, b and 5e, f). Cortex flexible, no cortical granules recognizable. Locomotion by slowly swimming while rotating about longitudinal body axis or rarely by gliding on the bottom of Petri dish. Feeds on host tissues and possibly also on bacteria in cultures.

Fig. 4
figure 4

Tetrahymena unionis sp. n. from life (a, b, e–g), after protargol impregnation (c, d, j), and after dry silver nitrate impregnation (h, i, k, l). a, b Ventral overviews, showing two different body shapes. c, d Ciliary pattern and nuclear apparatus of the ventral (c) and dorsal (d) sides. Black arrowhead marks dikinetids at the anterior end of the ciliary rows (blue area). e–g Overviews, showing the variability of body shape and size, the contractile vacuoles as well as the nuclear apparatus, which consists of one macronuclear nodule (shaded grey) and a single micronucleus (shaded blue). Drawn to scale. h, i Ciliary and silverline pattern of the ventral (h) and lateral (i) sides. White arrow in (h) shows the preoral suture. j–l Details of the anterior (j, k) and posterior (l) body portions. Black arrowhead marks dikinetids at the beginning of the ciliary rows (blue area), white arrow shows the preoral suture. Explanations = caudal cilium (CC), contractile vacuole (CV), excretory pores (EP), intermeridional silverline connectives (IC1, IC2), macronucleus (MA), micronucleus (MI), oral ribs (OR), paroral membrane (Pa), postoral kineties (PK), primary silverline meridian (PM), somatic kineties (SK), adoral membranelles 1–3 (M1–M3). Scale bars = 12 μm (j–l) and 20 μm (a–i)

Full size image
Fig. 5
figure 5

Tetrahymena unionis sp. n. from life. a Overview, showing many cells expelled from the mantle cavity by Unio tumidus after exposure to tap water. b, c Ventral views of representative specimens, showing the two contrasting body shapes, the contractile vacuoles, and the oral apparatus. d–f Dorsal (d) and ventrolateral (e, f) overviews, showing the caudal cilium, contractile vacuole, oral apparatus, somatic kineties, and cytoplasmic inclusions. g, h Details of the oral (g) and nuclear (h) apparatus. Explanations = caudal cilium (CC), contractile vacuole (CV), food vacuoles (FV), lipid droplets (LD), macronucleus (MA), micronucleus (MI), oral apparatus (OA), paroral membrane (Pa), somatic kineties (SK), adoral membranelles 1–3 (M1–M3). Scale bars = 8 μm (g, h), 20 μm (b–f), and 100 μm (a)

Full size image
Fig. 6
figure 6

Tetrahymena unionis sp. n. after protargol (a–j) and dry silver nitrate (k–o) impregnation. a–c Ciliary pattern and nuclear apparatus of the ventral (a, c) and dorsal (b) sides of representative specimens. White arrows mark the preoral suture. d–f Ciliary pattern and nuclear apparatus of mid-dividers. Black arrows show the broken somatic kineties, white arrows indicate the preoral suture. g, i Details of the anterior body end (g) and oral apparatus (i). Black arrowheads in (g) mark dikinetids at the anterior end of the ciliary rows, white arrowheads denote the monokinetids, white arrow indicates the preoral suture. h Detail of the posterior body portion, showing the excretory pores of the contractile vacuole and the basal body of the caudal cilium. j Detail of the nuclear apparatus. k, l Overviews of the silverline system on the ventral (k) and dorsal (l) sides. White arrow marks the preoral suture, red arrows show the intermeridional connectives. m–o Details of the ventral (m) and dorsal (o) sides of the anterior body portion, as well as of the posterior body end (n). White arrow marks the preoral suture, red arrows show the intermeridional connectives. Explanations = caudal cilium (CC), excretory pores (EP), macronucleus (MA), micronucleus (MI), oral ribs (OR), paroral membrane (Pa), somatic kineties (SK), adoral membranelles 1– 3 (M1–M3). Scale bars = 6 μm (g–j), 8 μm (m–o), and 20 μm (a–f, k, l)

Full size image

Somatic cilia about 5.5–7.5 μm long in vivo, distinctly more densely spaced anteriorly than posteriorly, arranged in 24–33 meridional rows (Table 3). Ciliary pattern similar to that of previous species: (i) typically two postoral ciliary rows, (ii) ventral ciliary rows begin from almost anterior body end to mid-portion of oral apparatus and run right and left of buccal cavity, (iii) lateral and dorsal ciliary rows commence subapically leaving a small, horseshoe-shaped unciliated apical area, (iv) an indistinct and short suture anterior to oral apparatus, (v) all ciliary rows composed entirely from monokinetids, except for lateral and dorsal kineties which begin with a dikinetid each (Figs. 4a–d, h–l; 5d, e; and 6a–g, k–o). Caudal cilium stiff and conspicuously longer than ordinary somatic cilia, i.e., about 11 μm long in vivo, seen in all slender specimens investigated in detail (Figs. 4a, f, g and 5c, d) but not in well fed cells (Fig. 4b, e), very likely lost during manipulation as the basal body of caudal cilium was observed in slender and broad specimens after protargol impregnation (Figs. 4d and 6b, c, f, h) and dry silver nitrate impregnation (Figs. 4i, l and 6l, n).

Silverline system composed of only primary meridians extending within somatic ciliary rows. Conspicuous outgrowths (peaks) emerge on left side of primary meridians between each two kinetids. Two intermeridional silverlines transversely connect right and left ciliary rows of dorsal side in anterior body portion, possibly transverse intermeridional silverlines also on ventral side but too faintly impregnated and thus difficult to recognize. All posterior basal bodies radiate silverlines to make a circle around basal body of caudal cilium (Figs. 4h, i, k, l and 6k–o).

Oral apparatus situated in a relatively deep cavity of ventral body side, i.e., in about anterior fifth of body length, ca. 8–15 μm long after dry silver nitrate impregnation (Table 3). Oral ciliature as usual in tetrahymenids, i.e., composed of a paroral membrane and three membranelles. Paroral membrane J-shaped and composed of very narrowly spaced dikinetids. Membranelle M1 approximately 7.2 μm long, membranelle M2 about 5.8 μm long, and membranelle M3 only ca. 2.1 μm long after protargol impregnation; each membranelle composed of three almost equally long rows of basal bodies (Table 3). Patches of irregularly arranged basal bodies right of membranelles in some specimens, very likely reorganizing postdividers. Cytopharyngeal fibers originate from basal bodies of paroral membrane, extend towards midline of oral cavity, well recognizable in some silver nitrate-impregnated specimens, only weakly impregnated with the protargol method used (Figs. 4a–c, h, j, k; 5b, c, e–g; and 6a, c, i, k, m).

Notes on division: Although only few mid-dividers were found in the protargol slides, they document that the ontogenesis of T. unionis follows the tetrahymenid mode. More specifically, binary fission is homothetogenic (i.e., the main body axes of the proter and the opiste have the same orientation) and occurs in freely motile condition. Stomatogenesis is monoparakinetal, that is, the rightmost postoral kinety produces the oral anlage of the opisthe. The parental oral structures are not involved in the formation of the daughter oral ciliature and are not reorganized. In mid-dividers, the somatic ciliary rows split in the middle leaving a barren area at the posterior region of the proter and at the anterior region of the opisthe (Fig. 6f, h).

Prevalence and intensity: T. unionis was detected in all five Unio tumidus specimens collected from the type locality. Although all mussels were heavily colonized with ciliates (more than five hundred individuals per mussel), no signs of pathological changes were detected. Despite our research on multiple native (Anodonta anatina) and invasive mussel species (Corbicula fluminea, Dreissena polymorpha, Sinanodonta woodiana), tetrahymenids were not detected in them. Likewise, no tetrahymenids were found in native (Theodoxus danubialis, Viviparus contectus) and invasive (Potamopyrgus antipodarum) aquatic pulmonate mollusks. Tetrahymena unionis is thus very likely a rare species.

Phylogenetic analyses

Altogether, 27 new sequences of three molecular markers (16S, 18S rRNA, and COI genes) were obtained from T. foissneri sp. n. and T. unionis sp. n. Their GenBank accession numbers, length, and guanine-cytosine content are summarized in Table 4. Three different algorithms (maximum likelihood, Bayesian inference, and neighbor-joining) were used to determine the phylogenetic positions of the two new Tetrahymena species, using the 16S and 18S rRNA as well as the COI gene sequences (Figs. 7, 8, and 9 and Suppl. Fig. S1). Although the overall tree topologies slightly differed depending on the molecular marker(s) and taxon sampling, T. foissneri and T. unionis were consistently classified within the same clades across all phylogenetic analyses. Specifically, T. foissneri was depicted within the ‘borealis’ clade in a sister position to T. limacis with full statistical support in the COI trees (Fig. 8) as well as in the 18S rRNA gene + COI trees (Fig. 7). As no 16S rRNA gene sequences are available for T. limacis, T. foissneri was left as an orphan lineage within the ‘borealis’ clade in the 16S rRNA gene trees (Fig. 9) as well as in the 16S rRNA gene + COI trees (Suppl. Fig. S1).

Table 4 Characterization of new sequences of Tetrahymena foissneri sp. n. and Tetrahymena unionis sp. n.
Full size table
Fig. 7
figure 7

Phylogenetic tree based on the nuclear 18S rRNA gene and the mitochondrial cytochrome c oxidase subunit I (COI) sequences, showing the systematic position of tetrahymenids newly isolated from mollusks. The genus Ichthyophthirius was used as an outgroup. Bootstrap values for maximum likelihood conducted in IQ-Trees, posterior probabilities for Bayesian inference conducted in MrBayes, and NJ bootstrap values were mapped onto the best scoring IQ-tree. Asterisks indicate a mismatch between maximum likelihood and Bayesian and/or NJ tree topologies. Sequences marked by red color were obtained during this study. Fully statistically supported nodes are marked with red solid circles. The strain MP80 was mislabeled as T. elliotti by Chantangsi et al. (2007). According to Doerder (2019), T. elliotti is related to T. gruchyi. GenBank accession numbers can be found in Supplementary Table S2. The scale bar represents the expected number of nucleotide substitutions per site

Full size image
Fig. 8
figure 8

Phylogenetic tree based on the mitochondrial COI sequences, showing the systematic position of tetrahymenids newly isolated from mollusks. The genus Ichthyophthirius was used as an outgroup. Bootstrap values for maximum likelihood conducted in IQ-Trees, posterior probabilities for Bayesian inference conducted in MrBayes, and NJ bootstrap values were mapped onto the best scoring IQ-tree. Asterisks indicate a mismatch between maximum likelihood and Bayesian and/or NJ tree topologies. Sequences marked by red color were obtained during this study. Fully statistically supported nodes are marked with red solid circles. The strain MP80 was mislabeled as T. elliotti by Chantangsi et al. (2007). The scale bar represents the expected number of nucleotide substitutions per site

Full size image
Fig. 9
figure 9

Phylogenetic tree based on the mitochondrial 16S rRNA gene sequences, showing the systematic position of tetrahymenids newly isolated from mollusks. Bootstrap values for maximum likelihood conducted in IQ-Trees, posterior probabilities for Bayesian inference conducted in MrBayes, and NJ bootstrap values were mapped onto the best scoring IQ-tree. Asterisks indicate a mismatch between maximum likelihood and Bayesian and/or NJ tree topologies. Sequences marked by red color were obtained during this study. Fully statistically supported nodes are marked with red solid circles. The scale bar represents the expected number of nucleotide substitutions per site

Full size image

Tetrahymena unionis was placed within the ‘paravorax’ clade with full statistical support in all analyses (Figs. 7, 8, and 9 and Suppl. Fig. S1). Tetrahymena pennsylvaniensis isolated from freshwater, T. glochidiophila isolated from glochidia of freshwater mussels, and T. nigricans isolated from freshwater planarians were revealed as the nearest relatives of T. unionis with maximum statistical support. Phylogenetic relationships among these four species were, however, not well resolved either in the single-gene or in the two-gene trees. To further address this problem, 18S rRNA and COI gene sequences of these four species were analyzed using the Bayesian coalescent approach, which accounts for ancestral polymorphism and incomplete lineage sorting. Bayesian coalescent analyses delimited all four Tetrahymena species with a posterior probability of 1.00. Moreover, T. glochidiophila was recognized to be sister to T. nigricans with full statistical support. However, the phylogenetic position of T. unionis remained unresolved also in the coalescent species trees, as it was depicted sister to the T. pennsylvaniensis, T. glochidiophila, and T. nigricans cluster but with a negligible posterior probability of 0.52 (not shown).

Discussion

Erection of two new Tetrahymena species

To establish new ciliate species, Warren et al. (2017) recommended preparing detailed morphological descriptions and taxonomic analyses of relevant distinguishing diagnostic characters. However, most Tetrahymena species have a rather similar body shape and size as well as overlapping numbers of ciliary rows. The uniformity of the genus Tetrahymena makes the erection of new species based solely on morphological characteristics to be very difficult in practice (e.g., Lynn et al. 2018; Quintela-Alonso et al. 2013). Thus, following Doerder (2019) as well as Rataj and Vďačný (2020), three molecular markers (16S, 18S rRNA, and COI genes) were used to define and separate the two new tetrahymenid ciliates isolated from mollusks. According to Chantangsi et al. (2007), the COI gene is taxonomically most important, as tetrahymenid ciliates exhibit an average of about 10% interspecific genetic divergence while less than 1% intraspecific variability in the barcoding region of their COI sequences. Doerder (2014, 2019) applied a 4% interspecific divergence as a threshold for species delimitation using the barcoding region of the COI gene. We also considered the genetic comparison and especially the COI divergence as the most important part in the argumentation for the establishment of new Tetrahymena species.

Tetrahymena foissneri sp. n., which was isolated from the invasive Lusitanian slug Arion vulgaris, differs from its nearest relative T. limacis by 6.34% in the COI gene. On the other hand, their 18S rRNA gene sequences are identical and their 16S rRNA gene sequences share a 99.88% identity. The high similarity of the nuclear and mitochondrial rRNA genes indicates a close kinship of these two species parasitizing slugs. However, their taxonomic distinctness and species status are supported not only by COI gene sequences but also by three morphological traits. The parasitic-phase forms of T. limacis have an anterior body end differentiated into a conspicuous thorn (Kozloff 1946; Warren 1932), while the anterior end of freshly isolated, parasitic-phase forms of T. foissneri is acute and never apiculate (Figs. 1a, d–g; 2a–c; and 3i, j). Nonetheless, it is important to mention that the anterior end of T. limacis becomes acute after cultivation (Corliss 1973), causing that the body shape of the parasitic-phase forms of T. foissneri and the cultured forms of T. limacis to become highly similar. Therefore, the morphology of freshly isolated ciliates needs to be compared. Both species can be furthermore separated by the diameter of the micronucleus, which never exceeds 2.5 μm in T. limacis (Corliss 1973) but ranges from about 2.3 to 4.4 μm in T. foissneri (Table 3). Finally, T. limacis has three transverse intermeridional silverline connectives in the anterior body region (Kozloff 1946), which were never observed in T. foissneri (Figs. 1i, j and 3i, j, l).

Tetrahymena unionis n. sp., which originated from the swollen river mussel Unio tumidus, is closely related to T. pennsylvaniensis, T. glochidiophila, and T. nigricans, all belonging to the ‘paravorax’ clade. The COI sequences of the new species differ from those of T. pennsylvaniensis by 8.29%, from T. glochidiophila by 7.51–7.56%, and from T. nigricans by 8.52%. Tetrahymena unionis shares a 99.82–99.88% identity in the 18S rRNA gene and a 98.42–99.19% identity in the 16S rRNA gene with its three nearest relatives. These genetic differences along with the present Bayesian coalescent delimitation analyses undoubtedly corroborate the species status of the new Slovak isolate. As there are no detailed morphological data on T. pennsylvaniensis (Doerder 2019), no comparison is possible. Although the morphological description of T. nigricans is very brief (Rataj and Vďačný 2020), the new species is much smaller (35–65 × 18–40 μm vs. 75–90 × 55–70 μm) and has a different body shape (narrowly elliptical to broadly ovate, anterior end slightly tapered, posterior end broadly to narrowly rounded vs. elliptical with both body ends broadly rounded). Moreover, T. nigricans originated from the invasive North American planarian Girardia tigrina, while T. unionis was isolated from a native bivalve Unio tumidus. Albeit the morphological description of Lynn et al. (2018) is very detailed, no relevant morphological distinguishing characters between T. glochidiophila and T. unionis were found. Thus, besides molecular data, the host organism might be so far considered as a further relevant difference. Specifically, the former species parasitizes glochidia of the genus Lampsilis (family Unionidae), while the latter species was isolated from adults of Unio tumidus. Additionally, the type locality of T. glochidiophila is in North America, while that of T. unionis in Central Europe. Nevertheless, much more research needs to be conducted whether host preference and the geographic place of origin might be counted among the relevant taxonomic differences.

Tetrahymena infections of mollusks

Only little attention has been paid to the diversity of tetrahymenids infecting mollusks in the past two decades. To date, only three Tetrahymena species have been reported from terrestrial gastropods, namely, T. limacis and T. pyriformis mostly from their digestive system, and T. rostrata mostly from their renal organ (for a review, see Van As and Basson 2004). With regard to aquatic mollusks, Dobrzańska (1958) recorded two Tetrahymena species in three bivalve species: Pisidium casertanum, P. obtusale, and Sphaerium lacustre. She identified them as T. limacis and T. pyriformis, which was, however, questioned by Corliss (1960) due to the lack of any morphological information. Since no DNA samples from these two tetrahymenid isolates are available, their identity cannot be tested with modern molecular methods. Very likely, the T. pyriformis of early studies of parasitism may not be the species T. pyriformis recognized today. To our best knowledge, the last reports of Tetrahymena infecting aquatic mollusks come from the bivalve Lampsilis siliquoidea, L. fasciola, and L. cardium (family Unionidae) whose glochidia are parasitized by T. glochidiophila (Lynn et al. 2018; Prosser et al. 2018). Despite our comparatively broad sampling (150 specimens belonging to six bivalve and 13 gastropod species), we detected tetrahymenids only twice. They could not be assigned to any of the aforementioned species parasitizing mollusks. Molecular analyses indicated that the first species, T. foissneri, is related to T. limacis from the ‘borealis’ clade, and the other species, T. unionis, is related to T. glochidiophila from the ‘paravorax’ clade. Although only 10.52% of the examined mollusk species were infected, we believe that a discovery of further new Tetrahymena species is to be expected with increased host sampling.

Van As and Basson (2004) very carefully reviewed records of terrestrial gastropods infected by Tetrahymena species. According to their list, wild populations of terrestrial gastropods rather commonly host T. limacis and T. rostrata, while T. pyriformis is considered a facultative parasite of gastropods, as it was able to infect the slug Deroceras reticulatum in experimental studies (Kozloff 1956a). Interestingly, the majority of reports of Tetrahymena infections come from the genus Deroceras (family Agriolimacidae) especially from D. reticulatum, which seems to be the most common host of T. limacis (Arias and Crowell 1963; Borden 1948; Brooks 1968; Kozloff 1946; Michelson 1971; Warren 1932; Windsor 1959). Further hosts of T. limacis are slugs of the families Arionidae, Milacidae, and Limacidae as well as snails of the families Bradybaenidae, Daudebardiidae, Helicidae, Hygromiidae, Succineidae, Vitrinidae, and Zonitidae (Brooks 1968; Kazubski 1958, 1959, 1960; Kazubski and Szablewski 1978; Kozloff 1956b).

Tetrahymena rostrata has a similarly long list of hosts and has been reported from the slug families Agriolimacidae, Arionidae, and Milacidae as well as from the snail families Clausiliidae, Cochlicopidae, Discidae, Oleacinidae, Valloniidae, Vitrinidae, and Zonitidae (Brooks 1968; Kazubski 1958, 1959, 1960; Kazubski and Szablewski 1978). This Tetrahymena species is also rather frequently encountered as a facultative parasite or a histophagous scavenger of other invertebrates. It is also readily found free-living in various edaphic habitats ranging from moss through leaf-litter to soil (e.g., Corliss 1973; Foissner 1998, 2000). On the other hand, T. limacis is edaphic only in the sense that it is occasionally found free-living in its host’s environment (Corliss 1973). Unfortunately, we did not examine soil samples from the site where the host of T. foissneri was found and, therefore, we cannot exclude that it might also have a free-living phase.

It seems that T. limacis is much more specific for mollusks than T. rostrata. However, the broad host spectrum of both Tetrahymena species needs to be tested with molecular methods in the future, as morphology-based identification of tetrahymenids needs to be confirmed by genetic data. As T. foissneri has been recorded only once, its host specificity is open for further analyses of wild populations of gastropods. Nevertheless, T. limacis and T. foissneri appear to be ecologically more similar, as they have a strong affinity to the alimentary tract including the digestive gland, intestine, and rectum. On the other hand, T. rostrata apparently prefers the renal organ (Van As and Basson 2004; Segade et al. 2009).

Three possible routes of transmission of tetrahymenids to new mollusk hosts were speculated. Michelson (1971) suggested that transmission can occur through feces in the case of T. limacis, as it is able to produce resting cysts. Spreading via ingestion of fecal material contaminated by resting cysts would, in turn, enable a wide host range and low host specificity. Although we did not observe cyst formation in T. foissneri, its mass occurrence in the alimentary tract is also indicative that its transmission might be mediated by the fecal-oral route. Corliss (1973) speculated that the dorsal integumentary pouch might also serve as a portal of entry into an adult slug. In this case, transfer to new hosts might be facilitated by the gregarious behavior of slugs during feeding or breeding. Finally, trans-ovum transmission has been recorded for T. rostrata in D. reticulatum (Brooks 1968). Ciliates invade the albumen gland, which leads to their incorporation in eggs. Embryos became infected through ingestion during aspiration of the albumen. By contrast to the fecal-oral route, the trans-ovum transmission might result in high host specificity.

Although we did not observe any pathological effects of tetrahymenids associated with mollusks, high infections of T. limacis have fatal results on their gastropod hosts, especially when ciliates enter various internal organs including the kidney (Brooks 1968; Michelson 1971). Tetrahymena rostrata might be responsible for the density-dependent regulation in agriolimacids infesting lowland pastures (Barker 1993, 2002). According to experimental studies, it reduces the growth rate, feeding, and fecundity of its hosts (Van As and Basson 2004). Epizootic mortality in the natural populations of agriolimacids is very likely also mediated by the onset of hot weather in late spring to early summer (Barker 1993). Segade et al. (2009) also observed pathological effects of T. rostrata on snails. Specifically, their examinations of farmed helicid snails (Helix aspersa aspersa and Helix aspersa maxima) revealed a severe destruction of the renal epithelium in heavily infected hosts. There is also some evidence that T. rostrata may have further deleterious effects on Helix aspersa, i.e., swelling of the mantle collar and a marked inability to retract the body into the shell. Mantle swelling was observed also in Deroceras reticulatum infected experimentally with T. rostrata (Brooks 1968).

With regard to bivalves, the viability of glochidia is comparatively strongly affected by tetrahymenids. Their viability declined by more than 60% after exposition to T. glochidiophila for 72 h (Prosser et al. 2018). Despite this fact, T. glochiophila very likely does not have a significant effect on the reproduction of Lampsilis species in the wild, as the 24-h window of glochidia infectivity upon release from the marsupia may be too short for ciliates to develop high abundances. However, as T. glochidiophila has been found also free-living, there might be a pool of ciliates that can infect glochidia and hence further negatively affect the bivalve reproduction. Unfortunately, we did not examine water samples from the site where the host of T. unionis was found and, therefore, we cannot exclude that it might also have a free-living phase.