Gastronauta paraloisi sp. n. and a population of Trithigmostoma cucullulus were isolated from a water sample taken from the surface of a freshwater river in Shenzhen, China (22°33′59.57″N, 114°6′52.80″E). The river lays in the inner city and influxes into a sea gulf approximately 10 kilometers away. This river receives household sewage water, and is thus nutrient-rich. The second population of T. cucullulus was isolated from an artificial freshwater lake in Qingdao, China (36°3′49.66″N, 120°20′42.86″E). The third population of T. cucullulus was found in a puddle near the coast in Zhanjiang, China (21°16′3.80″N, 110°25′31.08″E). The water salinity was about 3‰, and the sample was taken by scratching the substrate that contained diatoms. Immediately after the sampling, the ciliates were maintained by adding rice grains to the originally collected indigenous water in bottles. The attempt to obtain the pure culture with distilled water or Volvic water and rice grains failed. Instead, raw cultures were established by adding rice grains to Petri dishes filled with indigenous water (filtered through 0.65 μm-membranes to maintain indigenous bacteria). The rice grains supported the growth of bacteria as a food source for ciliates.

Living cells collected from raw cultures were observed using bright field and differential interference contrast microscopy (Olympus BX53). Wilbert’s (1975) protargol-staining method was applied to reveal the ciliature pattern and nuclear apparatus. Morphometric measurement was mainly conducted on specimens after protargol preparation. Illustrations of live specimens were based on photomicrographs and notes, while those of protargol-stained cells were made with the aid of a drawing device and photomicrographs. Terminology mainly follows Foissner (2000), and classification system refers to Lynn (2008).

For each of the ciliates, one single cell was picked from the raw culture under a stereomicroscope (SZH-ILLD-200, Olympus, Japan) and washed three times with distilled water in order to remove eukaryotic contaminants. The cell was then immediately transferred into a 1.5 ml-tube with 60 μl ATL buffer (Qiagen, Hilden, Germany). Genomic DNA was then extracted using the DNeasy Tissue Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions for animal tissues. Fifty ng of DNA (measured with a NanoDrop2000, peQLab, Biotechnology GmbH, Germany) from the extract were used for the subsequent amplification of the 18S rDNA. Q5 Hot Start High-Fidelity DNA Polymerase (New England Biolabs Inc., United States) and the primer pair EukA and EukB (Medlin et al., 1988) were used for PCR. Obtained amplicons were purified with the MinElute kit (Qiagen, Germany) and then sent directly for sequencing. Bi-directional Sanger sequencing was done to obtain a near-complete 18S rDNA sequence (Pan et al., 2019). The NCBI accession numbers of the assembled sequences of Gastronauta paraloisi sp. n., as well as the Qingdao and Zhanjiang populations of Trithigmostoma cucullulus are MW072507, MW116158, and MW116159, respectively. Sequence similarities were calculated using BioEdit 7.0.9.0.

With the newly obtained 18S rDNA sequences included in the present work, a total of 80 SSU rDNA sequences were used for the phylogenetic analyses (accession numbers shown in Table 1). Three Suctoria species, Acineta compressa, Ephelota gemmipara and Prodiscophrya sp., were selected as outgroups, as they are phylogenetically close to Cyrtophoria. Sequences were aligned using the MUSCLE package on the European Bioinformatics Institute website¹. The resulting alignment was then manually edited by trimming both ends. This resulted in a consensus matrix of 1,721 nucleotide positions. Two phylogenetic tools were used. A maximum-likelihood (ML) analysis was constructed with 1,000 bootstrap replicates using RAxML-HPC2 v. 8.2.12 (Stamatakis, 2014) online (CIPRES Science Gateway; Miller et al., 2010), with the evolutionary model GTR + I + Γ. A Bayesian inference (BI) analysis was run using the MrBayes on XSEDE 3.2.7a package (Ronquist and Huelsenbeck, 2003) on the CIPRES Science Gateway; optional model was GTR + I + Γ, selected using MrModeltest 2.2 with the criterion of Akaike Information Criterion (Nylander, 2004). Markov chain Monte Carlo simulations were run for ten million generations with a sample frequency of every 100th generations; the first 10% of the simulations were discarded as burn-in. The topology of the best ML tree was visualized using MEGA 5.0 (Tamura et al., 2011).

Phyllopharyngea de Puytorac et al., 1974

Cyrtophoria Fauré-Fremiet in Corliss, 1956

Chlamydodontida Deroux, 1976

Gastronautidae Deroux, 1994

Gastronauta Engelmann in Bütschli, 1889

Gastronauta paraloisi sp. n.

Synonym: Gastronauta membranaceus sensu da Silva and da Silva-Neto, 2001

Several stages of morphogenesis were observed in protargol-stained specimens. In an early divider (Figures 1F, 2D), the middle portions of the innermost four left kineties become thickened and curved inward, forming the opisthe’s oral primordia; meanwhile, an equatorial fragment occurs to the right of right kineties. In a middle stage (Figures 1G, 2E), equatorial fragment is elongated to form the anlage of opisthe’s terminal fragments (dorsal brush). The four oral primordia then dispatch from their original left kineties: anlage IV drifts in an anti-clockwise direction followed by anlagen III, II and I in succession. The fifth left kinety is interrupted in the middle. The middle part of right kineties concave to the left. In a late divider (Figures 1H, 2F), the four parts of the oral primordia merge to form the circumoral kinety of the opisthe. The outermost left kinety and the innermost two right kineties are interrupted in the middle. The anlage of terminal fragments develops into terminal fragments of the opisthe.

Chilodonellidae Deroux, 1970

Trithigmostoma Jankowski, 1967

Trithigmostoma cucullulus (Müller, 1786) Jankowski, 1967

Slides of the Qingdao, Zhanjiang and Shenzhen population of Trithigmostoma cucullulus were deposited in the Laboratory of Protozoology, Ocean University of China, with accession numbers QZS2015050601, QZS2015040301, and QZS2015120401, respectively.

Cell size 68–80 × 32–40 μm in vivo, and approximately 80 × 35 μm after protargol-staining. Body ellipsoid-shaped, both ends broadly rounded, inconspicuous protrusion located at left anterior end of cell (Figures 3A,B, 4E); dorsoventrally compressed with a width: thickness ratio of about 2:1 (Figure 3C). Cell acontractile but flexible. Anterior body end transparent. Cytostome in anterior 1/4 to 1/3, surrounded by 11 to 13 (on average 12) nematodesmal rods (Figures 3A,B,E). Approximately eight to 12 contractile vacuoles (diameter about 3 μm) counted from live cells, irregularly distributed on ventral side (Figure 3A), pulsating at an interval of 5–10 s; three to eight excretory pores detected in protargol-stained specimens (Figures 3E,F, 4C). Macronucleus in mid-body or slightly below, about 22 × 10 μm in vivo and 23 × 10 μm after staining; not (or inconspicuously) centrally heteromeric (Figures 3A,B,E,F, 4C). Cytoplasm usually containing many diatoms (Figures 3A,D, 4E,E). Usually crawling slowly on substrate, occasionally swimming in water; strongly thigmotactic.

Sixteen to 18 somatic kineties composed of three parts: right, postoral, and left kineties (Figures 3E,F, 4C). Constantly eight right kineties, all but the innermost one extending to left anterior cell end; posteriorly extending to rear end of cell. Two postoral kineties, ending at posterior 1/6 of cell, right one usually shortened (Figures 3E,F, 4C). Six to eight left kineties, all but the innermost one commences from preoral kinety, posteriorly shortened from right to left; the innermost left kinety usually starts at the level of cytostome; the outermost left kinety very short, only comprising 20 to 25 basal bodies.

Two circumoral kineties parallel to each other, slightly above cytostome, with the outer one about twice the length of the inner one; preoral kinety obliquely arranged ahead of cytostome and left kinety (Figures 3E,F, 4C). Single and oblique terminal fragment subapically located on dorsal side, comprising 15 to 24 basal bodies (Figures 3D, 4D). Equatorial fragment seldom detected, containing six to 20 basal bodies.

For Shenzhen and Zhanjiang populations, only illustrations, microphotographs and morphometric data are provided (Figures 3C, 4A,B,F–H and Table 3).

Three SSU rDNA sequences were obtained. The sequence of Gastronauta paraloisi sp. n. is highly divergent from available reference sequences in the GenBank database. By pairwise comparison, its closest relative, with a sequence similarity of 92.8%, is Phascolodon vorticella KX258192. The two Trithigmostoma cucullulus populations in the present work (Qingdao population, T. cucullulus QD; Zhanjiang population, T. cucullulus ZJ) have a similarity of 99.31%. The result of pairwise comparison of related sequences is shown in Figure 6F.

In the phylogenetic trees inferred from SSU rDNA sequences (Figure 5), four Trithigmostoma sequences form a monophyletic clade with full support. Gastronauta paraloisis sp. n. clusters with Trithigmostoma species with moderate support values (ML/BI, 80/0.87). This branch is then sister to the clade formed by other Chilodonellidae taxa represented by genera Chilodonella, Odontochlamys, Phascolodon, and Pseudochilodonopsis, with full support. The Gatronautidae/Chilodonellidae clade is located within the order Chlamydodontida. As in previous studies (Gao et al., 2012; Chen et al., 2016; Qu et al., 2017; Wang et al., 2017), both Chlamydodontidae and Lynchellidae are monophyletic (arrows in Figure 5 indicate their monophylies).

The subclass Cyrtophoria contains two orders, Dysteriida and Chlamydodontida, which are mainly characterized by the patterns of flattened body (dorsoventrally or laterally flattened), and the presence or absence of adhesive organelle (podite or secretory organelle) (Lynn, 2008). The dorsoventrally flattened body and the absence of adhesive organelle confirm that Gastronauta belongs to the order Chlamydodontida. By this criterion, Kahl (1931) put Gastronauta under the family Chlamydodontidae. The range of “Chlamydodontidae” by Kahl equates to order Chlamydodontida in today’s classification. While, the special structures, for example, a closed circular circumoral kinety and short vertical fragments in the left anterior body, make Gastronauta unique among all families in Chlamydodontida. Corliss (1979) tended to assign it to the family Lynchellidae because they both contain short arc kineties before cytostome (a large portion of kineties interrupted by cytostome). But their differences are also clear: (1) Gastronauta has a complete circular oral kinety, whereas Lynchellidae has parallel oral kinety fragments; and (2) Gastronauta has a centrally heteromeric macronucleus, while Lynchellidae species possess a parallelly heteromeric macronucleus. Thus, Deroux (1994) established the family Gastronautidae for the genus. Foissner (2000) also briefly discussed the familial assignment of Gastronauta, and he agreed with Deroux (1994)’s classification. On top of that, Foissner split the genus into two, based on the presence/absence of the postoral barren gap between right and left kineties: Gastronauta has the gap, while Paragastronauta does not. Lynn (2008) also adopted Gastronautidae, under subclass Chlamydodontida. Gastronauta is currently comprised of six species (including the new species), namely: G. membranaceus Engelmann in Bütschli, 1889 (type); G. fontzoui Nie & Ho, 1943; G. derouxi Blatterer & Foissner, 1992; G. aloisi Oberschmidleitner & Aescht, 1996; G. multistriata Xu et al., 2016, and G. paraloisi sp. n. Paragastronauta contains only one species, P. (Gastronauta) clatratus (Deroux, 1976) Foissner, 2000. The key morphometric characteristics of gastronautid species are shown in Table 4. A key to the identification of gastronautids based on morphological characteristics is provided below.

Based on the overall living morphology and the general ciliary pattern, especially the number of kineties and the two leftward curved innermost right kineties, one could be inclined to assign the new ciliate to Gastronauta aloisi. However, the new species can be clearly distinguished from the latter by the combination of the following characteristics: (1) body size. Our form is generally larger than G. aloisi (60–120 μm in vivo and about 68 × 45 μm after staining vs. 50–70 μm in vivo and 45 × 35 μm in G. aloisi); (2) number of arc kineties above oral slit (three vs. two); (3) number of vertical kinety fragments (three or four vs. two); and (4) the position of the anterior end of the innermost right kinety (almost joining the innermost left kinety vs. slightly left to the middle of the barren gap between right and left kineties) (Oberschmidleitner & Aescht, 1996). Further, the new species can be clearly distinguished from other Gastronauta species mainly by the two curved innermost right kineties (morphometric comparison see the identification key above and Table 4).

Foissner (2016) synonymized a Brazilian population of Gastronauta membranaceus with G. aloisi mainly because of the feature of the two curved postoral right kineties. However, this form has three arc kineties and the anterior end of the innermost right kinety connects the innermost left kinety (da Silva and da Silva-Neto, 2001), indicating it should be conspecific with the current new form. However, the Brazilian population is slightly different from our form by the position of the macronucleus (in anterior body vs. in mid-body or slightly below) and the number of terminal fragments (5–7 vs. 3–4), which could be considered as intraspecific variations.

da Silva and da Silva-Neto (2001) isolated a population of Gastronauta paraloisi sp. n. (see section “Discussion”) from an experimental reactor of an activated sludge sewage treatment plant. Although the environmental parameters were not measured in our study, our population was isolated from a similar habitat. The sampling river is part of the sewer system of the inner city, and it seems saprobic, especially in rush hours in the morning and evening. Right after the sampling, the food vacuoles were checked, and many small algae (diameter about 1–2 μm) were found. When the raw culture was established with rice grains added to the indigenous water, G. paraloisi sp. n. thrived and cells tended to gather around the grains (bacteria-rich), indicating it is omnivorous (feeding on bacteria and small algae). As demonstrated by Šimek et al. (2019), unselective grazers that feed on a broader size spectrum from bacteria to small algae may have a considerable competitive advantage in hypertrophic environments, which explains the occurrence of a high abundance of G. paraloisi sp. n. in a saprobic environment. The fact that the attempt to establish the pure culture solely with rice grains and instilled water failed shows that the cultivation of this species may not be monoxenic.

Jankowski (1967) established the genus Trithigmostoma from Chilodonella, mainly based on the lack of a barren gap between right and left kineties. The type is Trithigmostoma cucullulus. Foissner (1987, 1988) then summarized and made new combinations for some Chilodonella species. So far, the genus is comprised of five species, namely: T. bavariensis (Kahl, 1931) Foissner, 1987; T. chattoni (Mac Dougall, 1936) Foissner, 1988; T. cucullulus (Müller, 1786) Jankowski, 1967; T. srameki (Sramek-Husek, 1952) Foissner, 1988; and T. steini (Blochmann, 1895) Foissner, 1988. Details about the species identification refer to Foissner (1987, 1988) and Blatterer & Foissner (1992). Three other genera, Odontochlamys Certes (1891), Thigmogaster Deroux, 1976 and Pseudochilodonopsis Foissner, 1979 were also separated from Chilodonella (Qu et al., 2015), and together with Trithigmostoma, are now assigned to the family Chilodonellidae. However, characteristics such as a lack of barren gap between right and left kineties, almost all right kineties reaching posterior cell end, and many irregularly distributed contractile vacuoles, make Trithigmostoma very special in Chilodonellidae.

Trithigmostoma cucullulus is one of the earliest identified ciliates. It was originally reported by Müller (1786), using the name Kolpoda cucullulus (Aescht, 2001). Because of its wide distribution and relatively high abundance, Trithigmostoma cucullulus has been recorded worldwide and its morphology is well studied. Since the original report, the generic assignment of this species (species name) has been changed several times, for example to Chilodon cucullulus (Müller, 1786) Klein, 1927, Chilodonella cucullulus (Müller, 1786) Kahl, 1931 and Trithigmostoma cucullus (Müller, 1786) Foissner (1988)Jankowski, 1967 (spelling mistake was corrected by Foissner, 1988). Foissner et al., 1991 summarized the morphology of the species in detail. Our populations correspond well with the previous populations as well as with a recent estuarine population from China (salinity 4.3‰; Chen et al., 2018). Although reported many times in history, T. cucullulus demonstrates very stable morphological features. In addition, considering its worldwide distribution, it might be a cosmopolitan species.

In the present phylogenetic analyses, three sequences of Trithigmostoma cucullulus did not form a monophyletic group, but formed a clade within which T. steini X71134 (Leipe et al., 1994) nested. In general living morphology, T. cucullulus is very similar to T. steini. The main differences are the number of somatic kineties (18–22 in T. cucullulus vs. 25–33 in T. steini) and the number of contractile vacuole pores (2–11 vs. 10–40), which are mainly detectable after staining (Foissner et al., 1991; Chen et al., 2018; and data from the present work). Therefore, it is likely that T. steini X71134 (morphological data not available) was misidentified, or that the morphological discrepancy of these two species could not be shown by the molecular phylogeny inferred from SSU rDNA sequences.

The special positions of Gastronautidae (Gastronauta) and Trithigmostoma are also revealed by phylogeny derived from the SSU rDNA sequences. They cluster but also branch away from the core Chilodonellidae genera of the family Chilodonellidae. As mentioned above, Trithigmostoma has a special position within Chilodonellidae. The familial assignment to Chilodonellidae is beyond doubt, but its special characteristics also make it close to Chlamydodontidae and Lynchellidae. Chen et al. (2016) briefly discussed the status of this genus using phylogeny from SSU rDNA sequences, and demonstrated that it is a peripheral genus within Chilodonellidae. In another analysis inferred from mitochondrial small subunit ribosome DNA, the genus is also revealed to have an uncertain position in Chilodonellidae (Wang et al., 2017). Chen et al. (2018) concluded that Trithigmostoma is a basal genus in Chilodonelliae. Thus, based on previous studies and our own we agree that Trithigmostoma represents a basal group in Chilodonellidae, and it may be closely related to the common ancestor (still undiscovered or extinct) of Chilodonellidae, Chlamydodontidae and Lynchellidae.

Gastronautidae has more specialized and intermediate characteristics among the order Chlamydodontida (Figure 6). As mentioned above, both Gastronautidae and Lynchellidae have kineties interrupted by cytostome, thus there is no clear suture formed between right and left kineties in the left anterior body (violet dashes in Figure 6). Gastronautidae, Lynchellidae and Chilodonellidae usually have two (or sometimes several in Chilodonellidae) regularly distributed contractile vacuoles (vs. many irregularly distributed in Chlamydodontidae). Gastronautidae and Chilodonellidae have a centrally heteromeric macronucleus (sometimes inconspicuous in Trithgmostoma) (marked green in Figure 6) and usually well-separated right and left kineties (except for Trithigmostoma and Paragastronauta) (red double arrows in Figure 6), while the other two families have a parallelly heteromeric macronucleus and unseparated right and left kineties (Atopochilodon in Lynchellidae has only unobvious separation). In addition, Gastronautidae’s multiple terminal fragments make it close to Chlamydodontidae and Lynchellidae but not to Chilodonellidae (blue fragments in Figure 6). More importantly, Gastronautidae can be clearly separated from others by its unique, closed circular oral kinety (marked orange in Figure 6), which confirms the rationality of the establishment of Gastronautidae. Thus, we assume Gastronautidae represents an evolutionary link among Chlamydodontida ciliates. However, the poor-matching of morphology of Gastronauta and Trithigmostoma, along with the grouping of these two genera in SSU rDNA phylogeny, indicates that data concerning the key evolutionary link(s) of cyrtophorian ciliates are still missing. But up to date, there is only one sequence of Gastronauta available and the molecular representative of the type species, G. membranaceus, is still lacking, which largely constrains the phylogenetic resolution. Further study with high phylogenetic resolution needs to be conducted with more morphological and molecular data from Gastronautidae and other related groups.