Paradileptus elephantinus was collected from the Pontoon Dock of Lake Weishan Wetland Park (Figure 1B; N34°46′12″, E117°09′36″), Jining, China, on 24th April 2020. The physicochemical parameters of the sampling site were as follows: water temperature 17.8°C, atmospheric pressure 763.7 mm Hg, dissolved oxygen concentration 11.65 mg/L, salinity 0.63 ppt, and pH 9.39. Paradileptus conicus was isolated from an aquaculture pond of Weishan Special Aquaculture Base (Figure 1C; N34°46′18″, E117°09′54″), Jining, China, on 4th May 2020. The physicochemical parameters of the sampling site were as follows: water temperature 22.6°C, atmospheric pressure 754.8 mm Hg, dissolved oxygen concentration 7.11 mg/L, salinity 0.28 ppt, and pH 8.46. The two samples were collected from the water surface using a 20 μm mesh-sized plankton net and transferred into several Petri dishes for processing in the laboratory as soon as possible after collection (Bai et al., 2020).

Living cells were isolated with micropipettes and observed at 100–1000× magnifications using bright field and differential interference contrast microscopy (Olympus BX53) (Wu et al., 2020; Zhang et al., 2020). The infraciliature was revealed using the protargol staining method according to Wilbert (1975). Measurements and counts of stained specimens were conducted at magnifications of 100× and 1000×. Drawings of live cells were based on photomicrographs and free-hand sketches, while those of stained cells were accomplished with the help of a camera lucida at a magnification of 1000× (Lu et al., 2019). Terminology and systematics are mainly according to Vd’ačný et al. (2011a) and Vd’ačný and Foissner (2012).

The global distribution patterns of the Paradileptus species were mainly derived from previous reports that include morphological descriptions and recognizable illustrations. In addition, we selected several ecological reports in geographic regions not covered by morphological studies to show the range of Paradileptus distribution (Figure 1A and Table 1). In order to display the distribution of Paradileptus more accurately, we supply a list of the nominal species names as originally reported and the names by which they are now known based on the findings of the present study (Table 1). The literature used for this analysis was mainly derived from Vd’ačný and Foissner (2012).

For each species, a single cell was isolated from the original sample and washed five times with 0.22 μm filtered in situ water to remove potential contaminants. Genomic DNA was extracted from the cleaned cells using the DNeasy Blood and Tissue Kit (QIAGEN, Hilden, Germany) following the manufacturer’s instructions but modified by using 1/4 of the suggested volume for each solution. Q5^® Hot Start High-Fidelity 2× Master Mix DNA polymerase (New England BioLabs) was used to amplify the 18S and ITS-5.8S rDNA using universal eukaryotic primers 82F (5′-GAAACTGCGAATGGCTC-3′) and ITS-R (5′-TACTGATATGCTTAAGTTCAGCGG-3′) (Sogin, 1989; Chi et al., 2021). PCR amplifications were performed according to the following procedure: initial denaturation at 98°C for 30 s, followed by 18 cycles of amplification (98°C, 10 s; 69–51°C touchdown, 30 s; 72°C, 1 min), and another 18 cycles (98°C, 10 s; 51°C, 30 s; 72°C, 1 min), with a final extension of 72°C for 5 min (Lian et al., 2020). PCR products were sequenced bidirectionally in Tsingke Biological Technology Company (Qingdao, China) and assembled by SeqMan (DNAStar).

All available 18S, 5.8S, and ITS rDNA sequences of free-living litostomateans from known morphospecies were downloaded from the GenBank database and were compiled into four datasets each of which was used for separate phylogenetic analyses. The first dataset included 83 18S rDNA sequences of P. elephantinus and P. conicus, their related rhynchostomatians, other litostomateans, and armophoreans (outgroup taxa) and was used to construct the 18S tree of the Litostomatea. The second dataset contained 40 18S rDNA sequences of rhynchostomatians and spathidiids and was used to generate the phylogenetic tree focusing on the subclass Rhynchostomatia. The third dataset comprising 26 5.8S and ITS rDNA sequences of the two Paradileptus species, all available rhynchostomatians, and five spathidiids (outgroup taxa) was used to construct the ITS-5.8S tree focusing on the subclass Rhynchostomatia. The 18S, 5.8S, and ITS rDNA sequences of the rhynchostomatians and spathidiids were concatenated by SeaView v4 (Gouy et al., 2009) to form the fourth dataset that was used to generate the concatenated tree. See Supplementary Table 1 for sequence sources of these datasets.

Initial alignments were performed using the MAFFT algorithm with default parameters on the web server GUIDANCE2¹ (Zhang et al., 2019). Columns with scores less than 95% and residues with scores less than 90% were filtered, thus rendering the alignment suitably refined for the phylogenetic analysis. Maximum likelihood (ML) analysis was performed with RAxML-HPC2 on XSEDE v.8.2.12 (Stamatakis, 2014) on the CIPRES Science Gateway (Miller et al., 2010). Bayesian inference (BI) analysis was performed with MrBayes on XSEDE v.3.2.7a (Ronquist et al., 2012; Wang et al., 2019) on the CIPRES Science Gateway using the GTR + I + G evolutionary model as the best-fit model selected by MrModeltest v.2.2 (Nylander, 2004) according to the Akaike Information Criterion (AIC). Markov chain Monte Carlo (MCMC) simulations were then run with two sets of four chains using the default settings for 10,000,000 generations, with a sample frequency of 100 generations. The first 10% of trees were discarded as burn-in. All remaining trees were used to calculate the posterior probabilities. TreeView v.1.6.6 (Page, 1996) and MEGA v.5.2 (Tamura et al., 2011) were used to visualize tree topologies. BioEdit (Hall, 1999) was used to analyze the genetic difference between the two Paradileptus sequences.

Flexible but non-contractile planktonic dileptids; body trunk usually ovoidal or conical; oral field broad, dish-like, and prolonged anteriorly into a spiral proboscis; contractile vacuoles small and numerous, distributed throughout body; two types of extrusomes attached to proboscis oral bulge; somatic kineties difficult to recognize, somatic kinetosomes of body trunk loosely arranged; dorsal brush diffuse and staggered; right side of circumoral kinety accompanied by a perioral kinety, left side by numerous oblique preoral kineties; freshwater habitat.

Amphileptus flagellatus Rousselet, 1890.

Paradileptus is seemingly cosmopolitan having been recorded from 21 countries representing five continents (Africa, Asia, Europe, North America, and South America). It has been reported most frequently in Europe but has not been found in Antarctica or Oceania. It mainly occurs in freshwater habitats such as lakes, reservoirs, ponds, and rivers.

The abundances of the two species reported here differed significantly, i.e., there were about 20 cells of Paradileptus conicus per 10 ml but only about two cells of P. elephantinus per 10 ml. Nevertheless, both species show a wide distribution and have been reported from four continents (Asia, Europe, North America, and South America). Paradileptus moniliger, the earliest reported species in the genus, has only been recorded in Africa and Europe, and P. flagellatus only in Europe and North America (Figure 1A and Table 1).

The 18S rDNA sequences of the two Chinese populations were deposited in GenBank with lengths, G + C contents, and accession numbers as follows: Paradileptus elephantinus 1518 bp, 42.23%, MZ147012; P. conicus 1518 bp, 41.83%, MZ147013. The topologies of the BI and ML trees based on 18S rDNA data were highly concordant, therefore only the BI tree is presented (Figure 6A). All dileptids grouped together in the subclass Rhynchostomatia with strong support (BI 1.00, ML 99%) and were divided into two well-supported monophyletic orders, namely, Dileptida (BI 1.00, ML 97%) and Tracheliida (BI 1.00, ML 99%). The two sequences of Paradileptus clustered together with low support (BI 0.77, ML 23%) as a sub-clade that was sister group to Dileptus margaritifer (BI 0.96, ML 28%) within Dileptida. The other 18S rDNA tree focusing on the subclass Rhynchostomatia had a very similar topology (Supplementary Figure 1).

The ITS-5.8S rDNA region sequences of the two Chinese populations were deposited in GenBank with lengths, G + C contents, and accession numbers as follows: Paradileptus elephantinus 392 bp, 33.42%, MZ574467; P. conicus 391 bp, 31.71%, MZ574468. The topologies of the BI and ML trees based on ITS-5.8S rDNA data were generally concordant, therefore only the BI tree is presented (Figure 7). The phylogenetic tree inferred from ITS-5.8S rDNA data had a similar overall topology to that inferred from the 18S rDNA data, i.e., all rhynchostomatians were divided into two orders. The two Paradileptus species grouped together with maximal support (BI 1.00, ML 100%).

Phylogenetic reconstructions based on concatenated sequences of 18S, 5.8S, and ITS region by ML and BI methods had similar topologies; therefore, only the BI tree is presented (Figure 8). Relationships within the subclass Rhynchostomatia were generally consistent with those inferred from the single-gene analyses, the main difference being that the order Dileptida was divided into two clades instead of three, and the clustering of Paradileptus elephantinus and P. conicus was strongly supported (BI 1.00, ML 99%).

Paradileptus elephantinus was originally reported by Šveç (1897) under the name Dileptus elephantinus, but it was omitted from Paradileptus when Wenrich (1929) established this genus. Wenrich (1929) also described a new species, P. robustus, but Kahl (1935) considered this to be a junior synonym of P. elephantinus. The Chinese population of P. elephantinus closely resembles the original population with respect to its body shape, number and distribution of contractile vacuoles, moniliform macronucleus, and habitat (Šveç, 1897). The main difference is the cell size (265–420 μm in Chinese population vs. 200–250 μm). Considering the size range of the body of P. elephantinus (length 180–600 μm), we consider these two forms to be conspecific.

Paradileptus flagellatus and P. moniliger remain insufficiently described since there is no detailed living or infraciliature information for either. But they can still be separated from the Chinese population of P. elephantinus by the body shape (posterior end rounded or slightly pointed in P. elephantinus vs. posteriorly end sharply tapered with a short tail in P. moniliger) and the macronucleus (moniliform in P. elephantinus vs. two macronuclear nodules in P. flagellatus) (Ehrenberg, 1838; Wenrich, 1929).

According to the present and previous studies, Paradileptus elephantinus and P. conicus differ significantly in their morphology in vivo and infraciliature. The trunk of the body is oval in outline in P. elephantinus (vs. inverted-conical in P. conicus) and the extrusomes differ in length (type I, 11.6–13.4 μm and type II, 3.1–3.7 μm in P. elephantinus vs. type I, 4.0–5.7 μm and type II, 1.7–2.3 μm in P. conicus). In terms of its infraciliature, P. elephantinus can be distinguished from P. conicus by the number of preoral kineties (83–112, on average about 95 in P. elephantinus vs. 60–85, on average about 67 in P. conicus).

Paradileptus conicus was first described by Wenrich (1929) who characterized it as follows: “total length usually 100–200 μm, body conical in shape, tapering posteriorly to a spike-like projection; broad anterior end occupied by a cytostome and a peristomial field, surrounded by a flange or rim from which the spirally wound proboscis arises as an extension; contractile vacuoles numerous, distributed over the body and along the posterior part of the proboscis; macronucleus beaded, composed of from four to eight segments”. The Chinese population closely resembles the original population. Paradileptus conicus can be clearly distinguished from P. flagellatus and P. moniliger by its body shape (inverted-conical trunk in P. conicus vs. ovoidal trunk with rounded posterior end in P. flagellatus vs. ovoidal trunk with posterior sharply tapered to form a short tail in P. moniliger) (Ehrenberg, 1838; Rousselet, 1890). It also can be distinguished from P. flagellatus by its moniliform macronucleus (vs. two macronuclear nodules) (Rousselet, 1890).

For illustrations of selected key characters, see Figure 9.

Rhynchostomatia was established by Jankowski (1980) as one of three subclasses of the class Litostomatea. This subclass comprises two orders, Tracheliida and Dileptida (Vd’ačný et al., 2017). Phylogenetic analyses based on 18S, 5.8S, and ITS rDNA sequence data demonstrated that each of these two orders is monophyletic (Figures 6A, 7, 8), which is consistent with previous studies (Vd’ačný et al., 2011b, 2014; Jang et al., 2014; Vd’ačný and Rajter, 2015; Huang et al., 2018).

The topology of the 18S rDNA tree (Figure 6A) shows that the two Paradileptus species sequenced here nest within the Dileptida where they form a clade that clusters with other dileptids in the following order: Dileptus margaritifer (DQ487195) followed by Pelagodileptus trachelioides (AB558117) followed by Pseudomonilicaryon fraterculum (HM581677). Paradileptus can be clearly separated from D. margaritifer by morphological features such as the body shape (wide body with a spiral proboscis in Paradileptus vs. narrow body with a non-spiral proboscis) and the macronucleus (moniliform or as two nodules in Paradileptus vs. many scattered macronuclear nodules) (Vd’ačný and Foissner, 2012). In addition, it can be distinguished from Pseudomonilicaryon fraterculum by the mode of locomotion (free-swimming vs. gliding in P. fraterculum) and the body shape (wide vs. narrow to cylindrical in P. fraterculum) (Vd’ačný and Foissner, 2012). Pelagodileptus trachelioides is a planktonic dileptid with a moniliform macronucleus that superficially resembles Paradileptus. However, Paradileptus can be distinguished by the shape of its oral opening (roundish vs. narrowly elliptical in P. trachelioides) and the presence (vs. absence in P. trachelioides) of a strongly broadened proboscis base (Foissner et al., 1999; Vd’ačný and Foissner, 2012).

Compared to the 18S tree, nodal support for the ITS-5.8S and concatenated trees was generally higher, in particular for the resolution of the Paradileptus clade (BI 1.00, ML 100%; BI 1.00, ML 99%). The concatenated alignment might amplify the phylogenetic signal of single markers resulting in highly resolved and robust trees for rhynchostomatians (Figures 7, 8).

The findings of the present study support previous reports that suggest species of Paradileptus, and in particular P. elephantinus and P. conicus, are cosmopolitan (Figure 1A and Table 1; Foissner et al., 1999; Vd’ačný and Foissner, 2012). The Chinese population of P. conicus was isolated from an aquaculture pond where food resources are rich, whereas P. elephantinus was isolated from Lake Weishan where food resources are poorer. Considering that both the morphological and the 18S rDNA sequence data (Figure 6B) of these two taxa provide reliable and robust resolution of their separation at species level, we hypothesize that they might have undergone a putative speciation process via food preference and niche differentiation. The presence of greater numbers of preoral kineties, the larger body, the longer extrusomes and the larger buccal cavity of P. elephantinus suggest that its prey probably differs significantly from that of P. conicus, which may also be an adaptation to life in resource-poor habitats.

Rhynchostomatians are raptorial feeders whose predatory lifestyle has led to the development of special structures for prey recognition and capture. The primary feature is the apical proboscis which carries a dorsal brush that may be used for sensory feedback, extrusomes (toxicysts) to paralyze and/or kill prey organisms, and differentiated kineties such as the circumoral kinety (∼paroral membrane) and preoral kineties (∼adoral zone of membranelles) to aid feeding (Visscher, 1923; Dragesco, 1962; Vd’ačný and Foissner, 2012; Vd’ačný et al., 2017; Chi et al., 2021). Furthermore, differences in hunting strategies and preferred prey among dileptids (Vd’ačný and Foissner, 2012) seem to be related to these differentiated structures. Exploring this phenotypic divergence might improve understanding of the distribution patterns of ryhnchostomatians and their adaptations to different environments.

According to Vd’ačný et al. (2017), the most common type of proboscis bears two kinds of extrusomes and a multi-rowed dorsal brush, which will influence speciation, extinction, and net diversification of rhynchostomatians. Therefore, we speculate that: (1) there are competitive advantages in having two types of extrusomes compared to a single type of extrusome; (2) a multi-rowed dorsal brush provides distinct advantages over a two-rowed dorsal brush by increasing sensory function during locomotion; and (3) the increased number of differentiated kineties on the proboscis improves the efficiency of predation and the dietary niche differentiation of Paradileptus elephantinus and P. conicus and could contribute to their separation as different species.

In addition to the dorsal brush, extrusomes, circumoral kinety, and preoral kineties, the rhynchostomatian proboscis also bears one or two perioral kineties, i.e., the first one or two kineties on the right side of the circumoral kinety. Most rhynchostomatians possess one perioral kinety whereas the planktonic genera Paradileptus and Pelagodileptus have two perioral kineties, which is thought to increase the efficiency of food acquisition (Foissner et al., 1999; Vd’ačný and Foissner, 2012). In the present study, we revealed that there are five to eight recognizable kineties on the right side of the circumoral kinety in P. elephantinus, and two to three in P. conicus. We also found that the first kinety to the right of the circumoral kinety has densely spaced kinetosomes, whereas those of the remaining right kineties are loosely arranged, and that the first right kinety and circumoral kinety are closely adjacent, whereas there is significantly larger gap to the second right kinety. This arrangement is found in almost all other rhynchostomatians in which the first right kinety on the right side of the circumoral kinety is the perioral kinety and the remaining right kineties are somatic kineties (for details, see Vd’ačný and Foissner, 2012). Therefore, we conclude that the first kinety to the right of the circumoral kinety in P. elephantinus and P. conicus is the perioral kinety. It is noteworthy that the arrangement of kineties on the right side of the circumoral kinety in schematic drawings of Pelagodileptus trachelioides is also similar to almost all other rhynchostomatians (Vd’ačný and Foissner, 2012). In particular, Packroff and Wilbert (1991) labeled to the 2nd kinety to the right of the circumoral kinety as “somatic kinety 1,” whereas the first right kinety was marked the right circumoral kinety. We suggest that this latter structure is the perioral kinety and that all rhynchostomatians are characterized by the possession of a single perioral kinety.

Finally, we speculate that the densely arranged right kineties on the proboscis (vs. loosely arranged somatic kineties in the trunk) may be a taxonomically informative character for species separation and identification in Paradileptus. Unfortunately, the lack of ultrastructural and ontogenetic information of the proboscis in Paradileptus makes the origin of these densely ciliated kineties uncertain, which should be investigated in further studies.