A Case Study of the Morphological and Molecular Variation within a Ciliate Genus: Taxonomic Descriptions of Three Dysteria Species (Ciliophora, Cyrtophoria), with the Establishment of a New Species

Three Dysteria species, D. crassipes Claparède & Lachmann, 1859; D. brasiliensis Faria et al., 1922; and D. paracrassipes n. sp., were collected from subtropical coastal waters of the East China Sea, near Ningbo, China. The three species were studied based on their living morphology, infraciliature, and molecular data. The new species D. paracrassipes n. sp. is very similar to D. crassipes in most morphological features except the preoral kinety, which is double-rowed in the new species (vs. single-rowed in D. crassipes). The difference in the small ribosomal subunit sequences (SSU rDNA) between these two species is 56 bases, supporting the establishment of the new species. The Ningbo population of D. crassipes is highly similar in morphology to other known populations. Nevertheless, the SSU rDNA sequences of these populations are very different, indicating high genetic diversity and potentially cryptic species. Dysteria brasiliensis is cosmopolitan with many described populations worldwide and four deposited SSU rDNA sequences. The present work supplies morphological and molecular information from five subtropical populations of D. brasiliensis that bear identical molecular sequences but show significant morphological differences. The findings of this study provide an opportunity to improve understanding of the morphological and genetic diversity of ciliates.


Introduction
Ciliates are single-celled eukaryotes that are highly developed, ubiquitous in freshwater, marine and terrestrial as biotopes, speciose, and morphologically diverse [1][2][3][4][5]. Since Antonie van Leeuwenhoek made the first descriptions of ciliates, approximately 8000 freeliving and epibiotic species have been recognized [6]. Most of them are free-living species, and others are symbionts, either commensals or parasites, mostly on the external surfaces of their hosts but sometimes internally [7]. In the first 200 years, the taxonomic identifications or descriptions of ciliates were mainly based on living morphology. However, their morphological features can sometimes reflect certain aspects of prevailing environmental conditions, such as food abundance and the presence of predators [8][9][10][11]. Consequently, populations of the same species may present high morphological plasticity, resulting in significant problems in taxonomic identification [5,9,10,[12][13][14][15]. Over the last century, introduction of silver staining methods provided recognition of ciliary patterns and nuclear morphology; these detailed structures allowed a better taxonomy of ciliates [6,16]. In recent years, molecular technologies have been widely used in ciliate research and have greatly facilitated the study of ciliate taxonomy and systematics. However, there are numerous examples in ciliate research showing that morphological and molecular evolutions are not always concordant [10,17]. In many cases, taxa with similar or even identical morphology show highly divergent molecular information, while diverse morphotypes share the same molecular information. These anomalies have brought new controversies to ciliate research [12,17].
Cyrtophorians are highly divergent ciliates with dorsoventrally or laterally flattened bodies. Since the beginning of the 21st century, numerous cyrtophorians have been reported from various habitats, and many of these have been used as model organisms in morphological and genetic diversity research [1,5,10,12,[18][19][20][21][22]. Dysteria Huxley, 1857, is a specialized cyrtophorian genus with a highly bilaterally compressed body and somatic cilia densely arranged in kineties that are restricted to a narrow ventral groove between two lateral plates. Other diagnostic features include the right body plate, which is arched and is slightly larger than the flattened left plate and the podite, which is located at the rear end of the sulcus and is an attachment organelle that enables an organism to attach to submerged substrata, such as rocks and plants [2,5,10,19,[23][24][25][26]. Relatively few morphological features of Dysteria (e.g., body shape, cell size, dorsal spine) can be used for species identification, which has led to many misidentifications, synonyms, and the recognition of cryptic species [5,9,10,12,[24][25][26][27][28][29]. In the last two decades, the application of silver staining methods and molecular technology has enabled several historical confusions and errors to be resolved and for new species to be described [10,12,30]. To date, about 45 nominal Dysteria species (including Dysteria paracrassipes n. sp.) have been recorded, 26 of which have been investigated using staining methods including the type species D. armata [31]. The species and genetic diversity of Dysteria, therefore, remain poorly understood.
In this study, three Dysteria species (i.e., D. paracrassipes n. sp., D. crassipes, and D. brasiliensis) were isolated from subtropical brackish wetlands in China. Their morphological features were investigated based on observations in vivo and following protargol staining. The small subunit ribosomal DNA (SSU rDNA) of each was sequenced, and the molecular phylogeny of the genus Dysteria was analyzed. The main aim of the study is to improve knowledge and understanding of the morphological and molecular diversity of Dysteria.

Etymology
The species-group name "paracrassipes" is a composite of the Greek adjective "para-" (beside) and the species-group name "crassipes", indicating its similarity to Dysteria crassipes.

Type Materials
Seven slides with protargol-stained specimens have been deposited in the Laboratory of Protozoology, Ocean University of China (OUC), including one slide (registration number: YTT-20200528-01) with the holotype specimen circled in black ink and six slides (registration numbers: YTT-20200528-02, 03 . . . 07) with paratype specimens.

Morphological Description
Body 45-60 × 30-35 µm in vivo, bilaterally compressed, oval in outline when viewed from lateral aspect, both ends broadly rounded, ventral side straight and dorsal side slightly convex ( Figures 1A and 2A). When observed from ventral aspect, right body plate arched and slightly larger than flattened left plate ( Figures 1D and 2B). Fine equatorial transverse stripe located at midbody on left plate ( Figure 2C). Cytoplasm colorless to grayish, filled with different-sized food vacuoles, lipid droplets, and ingested diatoms ( Figures 1A and 2A Ciliature as shown in Figure 1B,E and Figure 2D-G. Four right kineties including two frontoventral kineties and two inner kineties, progressively shorter in length from right to left; longest frontoventral kineties extend anteriorly to dorsal margin, each composed of 84-110 basal bodies; two inner kineties commence near cytostome, innermost row terminating at level of podite and composed of 46-55 basal bodies ( Figures 1E and 2D, Table 1). Cilia of right kineties about 12 µm long in vivo. Four to eight left kineties, densely arranged near equatorial area ( Figures 1E and 2G). Equatorial fragment composed of 4-18 basal bodies ( Figures 1E and 2D,E, Table 1), cilia about 9 µm long in vivo. Terminal fragment positioned near front end of frontoventral kineties and composed of 4-7 basal bodies ( Figures 1B and 2F, Table 1). Two single-rowed circumoral kineties nearly equal in length and transversely arranged in parallel, located near cytostome ( Figures 1B and 2E). Double-rowed preoral kinety, obliquely arranged in front of cytostome ( Figures 1B and 2E). Three single-rowed left frontal kineties, transversely oriented, between preoral kinety and circumoral kineties ( Figure 1B,E and Figure 2E). Glandule visible in protargol-stained specimens, about 3-6 µm in diameter, located near rear end of innermost right kinety ( Figure 1E, Table 1).   All data based on protargol-impregnated specimens. Abbreviations: CV, coefficient of variation in %; M, median; Max, maximum; Mean, arithmetic mean; Min, minimum; n, number of specimens investigated; SD, standard deviation.

Dysteria Crassipes Claparède & Lachmann, 1859
Dysteria crassipes was briefly described in an original report by Claparède and Lachmann [32]. Kahl [21] provided some supplementary morphological information in his short redescription. Gong et al. [9] supplied the first detailed description, including both the living morphology and ciliature, and an improved diagnosis. In recent years, several temperate and tropical populations with similar morphological characteristics to those described previously have been reported [5,12]. Here, we supply a detailed description of the morphology of a subtropical population.

Morphological Description of Ningbo Population
Body 50-65 × 35-45 µm in vivo, bilaterally compressed, approximately rectangular in outline with both ends rounded, right plate slightly larger than left plate ( Figures 1F and 2H). Ventral margin straight, dorsal margin slightly convex near equator. Thin, transversely oriented stripe near equator on left plate ( Figure 2J). Cytoplasm colorless to grayish, filled with food vacuoles and lipid droplets ( Figures 1F and 2H). Cytostome about 5 µm in diameter, located about 25% down length of body. Two pharyngeal rods, conspicuous in vivo, each about 20 µm in length ( Figure 2I). Two contractile vacuoles each about 6-8 µm in diameter, usually one in anterior third and the other in posterior third of body ( Figures 1F,H and 2H).
Ciliature, as shown in Figure 1G,K and Figure 2L-N. Four right kineties, including two frontoventral kineties and two inner kineties that are progressively shorter from right to left; innermost right kinety commences about 25% down length of body and terminates near base of podite ( Figures 1K and 2L). Two frontoventral kineties almost equal in length and extend anteriorly to dorsal margin, each composed of 93-131 basal bodies (Figures 1K and 2L). Five to nine left kineties, densely arranged in equatorial area (Figures 1K and 2N). Equatorial fragment composed of 5-20 basal bodies ( Figures 1K and 2L,N). Terminal fragment short and curved, containing 3-7 basal bodies ( Figures 1G and 2L). Two circumoral kineties nearly equal in length and transversely oriented in parallel. Single-rowed preoral kinety, obliquely arranged in front of cytostome. Three left frontal kineties transversely oriented between preoral kinety and circumoral kineties (Figures 1G and 2M). Glandule conspicuous after protargol staining, about 3-6 µm in diameter, located near rear end of innermost right kinety ( Figures 1K and 2L).

Dysteria Brasiliensis Faria et al., 1922
Dysteria brasiliensis has been redescribed many times since it was originally reported by Faria et al. [33]. However, the morphological characteristics of the various populations differ. Here, we supply details of the morphology of five populations collected from subtropical coastal waters in China.

Morphological Description of Ningbo Population-I
This population is mainly illustrated in Figure 3F,G and Figure 4D,G,I,M-O. Body about 105-125 × 40-55 µm in vivo, bilaterally compressed, when viewed from lateral aspect triangular to elongate rectangular, front end blunt and round, posterior end slightly narrowed, ratio of width to length about 1:3, ventral margin straight, dorsal margin slightly convex at the equator, right plate slightly larger than left plate ( Figures 3F and 4D). Dorsal spine conspicuous, hook-shaped, about 10-15 µm long, located at posterior end of cell ( Figure 4I). Cytoplasm colorless, filled with numerous food vacuoles and a few lipid droplets. Cytostome conspicuous in vivo, located about 20% down length of body; two strong and conspicuous pharyngeal rods extending obliquely to posterior end of body, each about 35 µm long ( Figure 3F). Two contractile vacuoles, each about 5 µm in diameter, one located in anterior quarter and the other in posterior quarter of cell. Podite slender, 8-10 µm long in vivo, located at rear end of ventral sulcus ( Figure 4H). Macronucleus ellipsoidal and heteromeric, micronucleus not detected (Figures 3F and 4D). Both plates densely covered by rod-shaped bacteria ( Figure 4G).
Ciliature as shown in Figures 3G and 4I,M-O. Five right kineties, including two frontoventral kineties ( Figure 3G). Two frontoventral kineties, equal in length, extending almost entire body length, each containing about 182-293 basal bodies, cilia about 10 µm in vivo. The other three right kineties commence near cytostome but terminate at different levels, becoming progressively shorter from right to left, innermost kinety being shortest ( Figure 3G). Five to nine left kineties closely arranged in midregion of body. Equatorial fragment composed of 7-33 kinetosomes ( Figure 4M). Terminal fragment arc-shaped, consisting of 9-17 basal bodies. Two circumoral kineties almost equal in length and arranged in parallel, obliquely oriented above cytostome. Single preoral kinety, obliquely oriented. Three horizontal left frontal kineties orthogonal to and optically intersect the two circumoral kineties ( Figure 3G).

SSU rRNA Gene Sequences and Phylogenetic Analyses
The SSU rDNA sequences derived from the three Dysteria species were submitted to the NCBI database with lengths, GC contents, and GenBank accession numbers as follows: Dysteria paracrassipes n. Comparing the SSU rDNA sequences of these populations, the Ningbo population of D. crassipes differed by 18-25 bases from its congener populations (corresponding to 98.93%-98.48% similarity). In contrast, the SSU rDNA sequences among the nine populations of D. brasiliensis (five of which were derived from the present study) were more conserved with a maximum difference of only 6 bases ( Figure 5).
The maximum likelihood (ML) and Bayesian inference (BI) trees had nearly identical topologies; therefore, only the ML tree is shown here with nodal support values from both algorithms (Figure 6). In the SSU rDNA tree, the new sequence of D. paracrassipes n. sp. clusters with D. cristata (KC753488) and Dysteria sp. (KY922819) with full support (ML/BI, 100%/1.00). The new sequence of D. crassipes (Ningbo population) is sister to a fully supported clade of D. crassipes populations (ML/BI, 97%/1.00). The five newly sequenced D. brasiliensis populations, three known D. brasiliensis populations, and another four unidentified Dysteria species form a strongly supported clade (ML/BI, 90%/0.96) that is sister to the D. compressa + D. monostyla clade.

Discussion
Dysteria is one of the most commonly reported cyrtophorian genera, and due to continuous studies of its morphology and phylogeny, especially during the last two decades, a high species diversity of this genus has been revealed [9,[23][24][25]27,34]. In addition to the establishment of new species, many poorly known species have been redescribed using modern methods, including live observation, silver staining, and molecular analysis [5,10]. This in turn brings challenges to species circumscription and identification when there is a discrepancy between the morphological and molecular data. Research on different populations of the same nominal morphospecies has revealed examples both of intra-specific variations and cryptospecies among morphologically similar but molecularly different populations [10]. The present work describes one new species and several populations of two known species. By making detailed comparisons among species and populations using both morphological and molecular data, we hope to improve understanding of the species diversity in Dysteria.

Comments on Dysteria paracrassipes n. sp.
With reference to its living morphological characteristics and ciliature, Dysteria paracrassipes n. sp. closely resembles D. crassipes Claparède & Lachmann, 1859. However, there is a significant difference between these two species; that is, the preoral kinety is doublerowed in D. paracrassipes n. sp. (vs. single-rowed in D. crassipes). In previous studies, the structure and position of the preoral kinety were largely ignored. We carefully examined and verified this structure of many well illustrated D. crassipes populations, all of which are single-rowed [9]. Therefore, this is a stable and reliable character for separating these two species. In addition, the SSU rDNA sequence of D. paracrassipes n. sp. differs significantly and shares a poor genetic similarity with the SSU rDNA sequences of the D. crassipes populations (Ningbo population, 56 bases, 96.38%; Zhuhai population, 56 bases, 96.20%; Haikou population, 55 bases, 96.47%; Shenzhen population, 55 bases, 96.33%; Daya Bay population, 57 bases, 96.46%), providing further support for the validity of the new species.

Comments on Dysteria crassipes Claparède & Lachmann, 1859
Dysteria crassipes was originally reported by Claparède and Lachmann [32] and briefly redescribed by Kahl [21]. Its morphology, however, remained poorly known with information confined to features such as body size and shape. Gong et al. [9] provided the first detailed morphological description using modern methods based on a population collected from Chinese coastal waters of the Yellow Sea. Dysteria crassipes has since been reported repeatedly in coastal waters of China, for example, by Chen et al. [17], who documented the SSU rDNA sequences of three populations from the South China Sea but without morphological data, and Wang et al. [5], who provided morphological and molecular information of a South China Sea population.
The Ningbo population collected from subtropical waters of East China Sea is morphologically consistent both with the Qingdao populations (temperate coastal waters of the Yellow Sea) and the Haikou populations (tropical waters of the South China Sea) [5,9,10]. However, the SSU rDNA sequences of these Chinese populations are quite different; that is, the number of base-pair differences ranges from 5 to 25, and sequence identities among the populations range from 98.39% to 99.71% (Table S1, Figure 5). Because of the absence of distinct species-level morphological differences, we consider the current isolate to be a subtropical population of D. crassipes. However, it seems likely that there are several cryptic species within the D. crassipes complex. Faria et al., 1922 Dysteria brasiliensis is a well-known species that has been investigated many times since being originally described by Faria et al. [33]. Song and Packroff [34] provided the first detailed morphological description using modern methods. In their investigation of different Yellow Sea populations, Gong et al. [9] showed that the morphology of this species is highly variable, for example, in terms of the body shape and the presence/absence of the dorsal spine. Qu et al. [10] added some morphological information based on two populations, one collected from the Bohai Sea and the other from the South China Sea [10]. Wang et al. [12] provided morphological and molecular information based on another tropical population, further enriching knowledge of the morphological and molecular diversity of this species.

Comments on Dysteria Brasiliensis
The present work includes a comprehensive study on the morphological and molecular data of five subtropical populations of D. brasiliensis. A comparison of the four known and five new SSU rDNA sequences revealed that the molecular information is highly consistent, with a maximum difference of only six bases and sequence identities of 99.62% to 100% ( Figure 5). However, there were significant morphological variations among these populations, mainly in body shape and dorsal spine structure (Table S2). These could be roughly divided into three types (i.e., slim and curved ( Figure 3A,K,L), slim with a prominent dorsal spine ( Figure 3D,M,N), and triangular or ellipsoidal with a curved dorsal spine ( Figure 3F,H,J). Similarly, the characteristics of the dorsal spine also vary greatly. Among the five populations collected from Ningbo, some individuals had no dorsal spine, some had a short hook-shaped dorsal spine, and some had a long spear-shaped dorsal spine that was up to one-third of the body length ( Figure 3D,E,N and Figure 4E,F). Some researchers have regarded the dorsal spine structure as a species character [21,32], whereas Gong et al. [9] did not accept this assertion, noting that the dorsal spine was not a key feature for Dysteria identification. Based on the morphological and molecular information of multiple populations of D. brasiliensis, we suggest that variations in the body shape and features of the dorsal spine of Dysteria are probably responses to the environment, including food abundance and predator pressure.

Phylogenetic Analyses Based on SSU rDNA Sequences
As shown in Figure 6, Dysteria paracrassipes n. sp. is sister to the D. cristata KC753488 + Dysteria sp. KY922819 clade. Furthermore, most of the morphological features of the new species are similar with Dysteria cristata (e.g., ovoidal body shape, approximately 45-60 µm × 30-35 µm in vivo, with four right kineties and four to eight left kineties), thus supporting the phylogenetic affiliation of these two taxa. However, D. paracrassipes differs from D. cristata in having four right kineties (vs. three in D. cristata) [10,27]. Although the five D. brasiliensis populations investigated here differ slightly in the base differences of their SSU rDNA (maximum difference of six bases) and cluster together in the phylogenetic tree, they vary greatly in their morphology (i.e., significant differences in body size and shape, dorsal spine, and ratio of dorsal spine to body length (Table S2)). Accordingly, we speculate that the presence and shape of the dorsal spine of D. brasiliensis do not correspond to systematic signals in the SSU rDNA sequences and are not taxonomically informative characters.

Sample Collection, Observation, and Identification
Three Dysteria species were collected on 28 May 2020 from subtropical brackish coastal waters of the East China Sea at Ningbo, China (Figure 7). Dysteria paracrassipes n. sp. was collected from a coastal wetland on Meishan Island (29 • 46 23" N, 121 • 57 17" E), where the water temperature was about 26 • C and salinity was 4.5. Dysteria crassipes was collected on 26 February 2019 from a brackish lake (29 • 45 51" N, 121 • 54 2" E) connected to the East China Sea by channels, where the water temperature was about 12 • C and salinity was 20.0. Five populations of D. brasiliensis were collected from 2019 to 2020: population-I was collected from a brackish fish-culturing pond (29 • 33 50" N, 121 • 42 49" E), where the water temperature was about 28 • C and salinity was 26.0; populations II-V were collected from the same brackish lake as above (29 • 45 51" N, 121 • 54 2" E) in different seasons, so the water temperature varied between 28 and 34 • C and the salinity ranged from 17.0 to 19.0. Samples were taken from the surface of the sediment using a sterile syringe and the dilute with untreated water from the collection site. Clonal cultures were maintained for a few days in Petri dishes at room temperature using filtered habitat water. Rice grains or wheat grass juice was added to promote the growth of bacteria as food source for the ciliates. All cultured ciliated died within about 1 week; therefore, we were not able to maintain either taxon for a long time. However, it was possible to isolate enough individuals of three species to provide a detailed morphological description. Cells were observed in vivo using bright field and differential interference contrast (DIC) microscopy (Leica DM2500) at 400-1000× magnification. Protargol staining was used to reveal the ciliature and nuclear apparatus following the method of Wilbert [35]. Counts, measurements, and drawings of stained specimens were performed at 1000× magnification.

DNA Extraction, PCR Amplification, and Gene Sequencing
For each population, a single cell was isolated using a micropipette. Specimens were washed three to five times using filtered habitat water and twice using ultrapure water. Three to five parallel molecular samples of every population would be prepared. Extraction of genomic DNA was performed using a DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany), following the manufacturer's instructions. Two samples were randomly chosen and sent for sequencing. No intraspecific variations were identified. The SSU rDNA was amplified using the primers 18SF (5 -AAC CTG GTT GAT CCT GCCi AGT-3 ) and 18SR (5 -TGA TCC TTC TGC AGG TTC ACC TAC-3 ) [36]. To minimize the possibility of amplification errors, a Q5 Hot Start High-Fidelity DNA Polymerase (New England Biolabs Co., Ltd., M0493, Beijing) was used.
The polymerase chain reaction (PCR) conditions used in the amplification were as follows: initial denaturation for 30 s at 98 • C, 35 cycles of 10 s, denaturation for 20 s at 98 • C, primer annealing for 100 s at 56 • C, primer elongation at 72 • C, final primer elongation for 5 min at 72 • C. Sequencing was performed bidirectionally at Tsingke Biological Technology Company (Beijing, China).