Abstract

The application of molecular techniques to accurately identify protozoan species can correct previous misidentifications based on traditional morphological identification. Colpodea ciliates have many toxicological and cytological applications, but their subtle morphological differences and small body size hinder species delineation. Herein, we used Cox I and β-tubulin genes, alongside fluorescence in situ hybridization (FISH), to evaluate each method in delineating Colpodea species. For this analysis, Colpoda harbinensis n. sp., C. reniformis, two populations of C. inflata, Colpoda compare grandis, and five populations of Paracolpoda steinii, from the soil in northeastern China, were used. We determined that (1) the Cox I gene was more suitable than the β-tubulin gene as a molecular marker for defining intra- and interspecific level relationships of Colpoda. (2) FISH probes designed for Colpoda sp., C. inflata, Colpoda compare grandis, and Paracolpoda steinii, provided rapid interspecific differentiation of Colpodea species. (3) Colpoda harbinensis n. sp. was established and mainly characterized by its size in vivo (approximately ), a reniform body in outline, one macronucleus, its spherical shape, a sometimes nonexistent micronucleus, 11–15 somatic kineties, and five or six postoral kineties. In conclusion, combining oligonucleotide probes, DNA barcoding, and morphology for the first time, we have greatly improved the delineation of Colpodea and confirmed that Cox I gene was a promising DNA barcoding marker for species of Colpodea, and FISH could provide useful morphological information as complementing traditional techniques such as silver carbonate.

1. Introduction

The increasing diversity of ciliates requires multiple methodological tools for their correct identification [18]. However, given the constraints of professional or industrial practices, achieving accurate and rapid identification can be challenging via a single method. In the past, Ciliophora identification relied mainly on either morphological and ultrastructural features or small subunit (SSU) ribosomal (r) DNA sequence analysis [916]. Although morphological analysis is a valuable technique for identifying ciliates, it can be time-consuming and laborious [17], while SSU rDNA sequence analysis has limitations in distinguishing between closely related species.

Several surveys of DNA barcoding in Ciliophora have shown a high prevalence [1822]. The Cox I gene is a suitable marker for resolving the interspecific and intraspecific relationships of Paramecium spp. [22]. The internal transcribed spacer 2 region (ITS2) is also a strong barcoding candidate for identifying the closely related Tintinnids [23]. Molecular phylogenies and genetic measurements based on variable regions of nuclear genes demonstrated that the ITS2 and LSU-D1/D2 regions are more suitable for delineating Euplotes [24]. Fluorescent probes targeting small subunit ribosomal RNA (SSU-r RNA) have been designed and optimized for fluorescence in situ hybridization (FISH), resulting in the accurate and rapid identification of pathogenic ciliates (e.g., Pseudocohnilembus persalinus, Boveria labialis, and B. subcylindrica) [2528]. FISH allows for molecular identification of targeted organisms in mixed populations, overcoming the negatives of morphological methods and producing timely detection results. However, there are currently no available fluorochrome-labeled oligonucleotide probes for the genus Colpoda.

The class Colpodea (Small and Lynn [29]) comprises approximately 60 genera and 200 species, with most living in terrestrial and semiterrestrial habitats, such as mosses, leaf litter, soil, and tree holes [3034]. However, this is likely only a subset of the total diversity, with a high number of species likely undiscovered [35]. Colpodea is typically characterized by high technical requirements for staining, environmental sensitivity, susceptibility to dormant cysts, and few multigene sequences, resulting in long-standing problems with species identification and taxon attribution [13, 30, 3641]. To date, the identification of ciliates of Colpodea has relied solely upon morphological features and SSU rDNA sequence analysis. However, with the conservative evolution of SSU rDNA alongside various issues such as asynchronous evolution with morphology, delineation remains problematic. Therefore, other methods, including DNA barcoding and oligonucleotide probes, should be developed to accurately and rapidly identify Colpodea. The uses of DNA barcoding and FISH are universally applicable tools that can identify ciliates and confirm taxonomic relationships previously based on ultrastructural and other morphological features [22, 2628, 42].

Nonetheless, there is still no universal gene marker for species discrimination of ciliates. In the present investigation, we assessed the suitability of DNA barcoding and oligonucleotide probe techniques to delineate ten newly isolated Chinese populations of five Colpodea species. Specifically, we investigated the barcoding utility of β-tubulin and the mitochondrial cox1 genes, both at the congeneric and conspecific levels, in order to analyze the reliabilities of molecular identification methods for ciliates of Colpodea.

2. Material and Methods

2.1. Ciliate Isolation, Observation, and Identification

Five species were collected from soil in northeastern China and treated with nonflooded Petri dish cultures as described in Foissner et al. [43]. After isolation, specimens were maintained in Petri dishes in the laboratory for three days. Clonal cultures were then established and maintained at room temperature in boiled water amended with a grain of wheat to enrich natural bacteria as food for the ciliates. Isolated cells were observed and photographed in vivo using differential interference contrast microscopy. The silver carbonate [44] was used to reveal the infraciliature in different morphogenetic stages. Stained specimens were counted and measured at magnifications of ×100–1250, and mapping was performed with the help of a drawing device. Classification and terminology are mainly according to Foissner [30] and Lynn [45].

2.2. DNA Extraction, PCR Amplification, and Sequencing

Five cells from each monoclone were isolated under the stereomicroscope using micropipettes and washed with double distilled water to remove contaminants. Cells were then transferred to an Eppendorf tube with a small amount of water. Total genomic DNA of the cells was extracted with the DNeasy & Tissue Kit (Shanghai, QIAGEN, Germany) according to the manufacturer’s instructions.

The β-tubulin and the Cox I genes were amplified using the polymerase chain reaction (PCR). PCR primers are listed in Table 1, and conditions of the respective PCR reactions are summarized in Table 2. Sequencing was performed Shanghai Sangon Biological Engineering and Technical Service Company (Shanghai, China). 36 new molecular sequences of β-tubulin and Cox I genes were generated from five species of Colpodea. All the sequences were aligned using Clustal W implemented in BioEdit 7.0.1 [46].

2.3. Cell Fluorescence In Situ Hybridization (FISH)

Probes (Table 3) were designed using the probe design tool as implemented in the ARB software package for the SSU-rDNA sequences of the present Colpoda harbinensis n. sp, C. inflata, Colpoda compare grandis, and Paracolpoda steinii. Generated probes were checked against the GenBank sequence collection by a standard nucleotide-nucleotide BLAST search [47]. FISH was used to visualize Colpodea spp. above both in field samples and a mixture of species as well as Coleps hirtus that frequently occurred in the same habitats as the negative control. Cells were fixed with 50% Bouin’s solution and filtered onto a 2 μm-pore-size cellulose nitrate membrane (25 mm in diameter) using low under pressure. The membrane was then washed five times with 2 ml of filtered sterile water. The basic hybridization follows the protocol of Stoeck et al. [48] and Zhan et al. [26].

2.4. Phylogenetic Analyses

Phylogenetic trees were inferred using maximum likelihood (ML) and Bayesian inference (BI) methods. ML analyses were constructed by RAxML-HPC2 v8.2.12 [49], and BI analyses were constructed by MrBayes v3.2.7a [50], both on the CIPRES Science Gateway (URL: http://www.phylo.org/sub_sections/portal). The ML and BI trees based on 18S rRNA gene were constructed according to the GTR + I + G model chosen by the MrModeltest v.2.0 program [51]. ML analysis was done using rapid bootstrap with 1,000 nonparametric bootstrap replicates. Bayesian posterior probabilities were calculated by running four chains for 10,000,000 generations, with the cold chain sampling every 10,000 generations. The first 25% of sampled trees were discarded as burn-in. (ML/BI) was considered as low, 75%/0.75–90%/0.90 (ML/BI) as moderate, and >90%/0.90 (ML/BI) as high. MEGA 7.0 [52] was utilized to visualize tree topologies.

2.5. Haplotype Networks

A β-tubulin haplotype network was constructed for Paracolpoda steinii and Colpoda inflata, using the TCS method [53] as implemented in PopART ver. 1.7 [54]. Mutations in β-tubulin sequences were displayed as line segments on the haplotype network.

3. Results

3.1. Morphological Description of Chinese Populations of Four Known Colpodea Species
3.1.1. Colpoda reniformis (Figures 1(a) and 1(b))

Specifications are as follows: size in vivo, body monk’s cap nephroid in shape, with left margin slightly curved and the right margin “C”-shaped (Figures 1(a) and 1(b)). Diagonal groove was present (Figure 1(a)). One macronucleus, nearly spherical, is located in the middle of the body, and no micronucleus was observed (Figure 1(b)). Contractile vacuole situated in the posterior 1/3 of the body, approximately 4 μm in diameter during diastole. Extrusomes were conspicuous and numerous, approximately 2 μm (Figure 1(a)): 27–39 somatic kineties, oral located 1/2 of the body, amd 13–15 postoral kineties (Figures 1(a) and 1(b)).

3.1.2. Colpoda Compare grandis (Figures 1(c)1(e))

Specifications are as follows: cell in size, round reniform in outline, laterally flattened, and no postoral sack (Figures 1(c) and 1(d)), brownish cytoplasm usually contained food vacuoles, one macronucleus, roughly spherical, positioned in the middle and anterior part of the cell, no micronucleus (Figure 1(d)), contractile vacuole located posteriorly, and approximately 4 μm in diameter during diastole. No extrusomes were observed. There is a forward swimming in a spiral pattern in the water. Somatic cilia were closely arranged, approximately 10 μm long. Diagonal groove was not observed: left oral polykinetid on vestibular bottom, elongated square (Figure 1(e)): 28–30 somatic kineties and 12–14 postoral kineties.

3.1.3. Two Populations of Colpoda inflata (Figures 1(f)1(m))

Population 1 had a body size of in vivo, while population 2 was slightly larger, with a body size of about . Other characteristics of the two populations were similar: elongated reniform in outline, with soft, rough cortex, and slightly dark endoplasm (Figures 1(f)1(k)). Oral is located 1/2 of the body. One macronucleus is roughly spherical or oval, anterior, or posterior to the middle of the body; single micronucleus, either oval or crescent-shaped, is closely adjacent to the macronucleus (Figures 1(h), 1(j), and 1(k)). One contractile positioned at the end of the body (Figures 1(g) and 1(i)). Depending on the refraction, granules of different sizes appeared brownish yellow or black under bright-field light microscopy. Diagonal grooves were absent: 23–25 somatic kineties. Left oral polykinetid on elongate elliptic (Figures 1(l) and 1(m)): seven or eight postoral kineties.

3.1.4. Five Populations of Paracolpoda steinii (Figure 2)

Five populations were present in this collection, and all interpopulation variation was within the variable range. Population 1 had a greater range of individual size variation than the other four populations (). Population 3 had a slightly longer body length than population 2, but a similar body width ( vs. ) (Figures 2(c)2(h)). Populations 4 and 5 were very similar in body size in vivo ( vs. ) (Figures 2(i)2(m)). Other characteristics were almost identical: lateral appearance reniform, preoral portion remarkably short (1/4–1/3 of body length), usually slightly ventrally inclined, flattened slightly to 2: 1, in ventral and dorsal aspect pyriform to moderately broadly wedge-shaped, and distinct diagonal grooves (Figures 2(a), 2(c), 2(f), 2(i), and 2(c)). Macronucleus slightly to distinctly ellipsoid is usually near the center of the cell. Micronucleus calotte-shaped was attached to macronucleus (Figures 2(b), 2(d), 2(e), 2(g), 2(h), 2(j), 2(l), and 2(m)). Contractile vacuole was located at the posterior end, approximately 3 μm long, during diastole with small collecting vesicles and a single excretory pore in the center of the posterior pole. Oral apparatus in anterior third. Oral polykinetids were protrude. Left polykinetids were vertically distributed. Left polykinetids are elliptical, occasionally slightly wedge-shaped or rectangular (Figures 2(n)2(o)), and moves rapidly, mostly rotating toward the back of the body or marching directly forward. Somatic cilia were approximately 8 μm long: 9–11 somatic kineties.

3.2. Phylogenetic Analyses Based on 18S rRNA Gene Sequence Data

Phylogenetic trees were constructed using ML and BI and produced similar topologies; therefore, only the ML trees and their support values from both methods are shown. According to the 18S-rRNA gene tree, all four orders within Colpodea were monophyletic (Figure 3). Colpodida and Cyrtolophosidida clustered together to form a clade, with Bursariomorphida as a sister clade, while the order Platyophryida occupied the basal position within Colpodea.

The newly sequenced species Paracolpoda steinii was sister to the clade clustered by P. steinii (KJ607914) and Bromeliothrix metopoides (100% ML, 0.9 BI). All nine newly sequenced species were clustered within the core of the Colpodea clade. The two newly sequenced species, Colpoda compare grandis and C. reniformis, formed a sister group, which then grouped with C. henneguyi and Bresslauides discoideus. The newly sequenced Colpoda harbinensis n. sp., C. inflata pop1, and C. inflata pop2 clustered together. The seven Paracolpoda steinii sequences, including the five newly sequenced populations, clustered together as a sister group to Bromeliothrix metopoides with full support (100% ML, 1.00 BI).

3.3. DNA Barcoding of the Colpoda
3.3.1. The Utility of Cox I Gene Tested for Accurate Identification

The Cox I amplification primers MOU08–121 and MOU08–122 (Table 1) yielded a single DNA band of the predicted length (~945 bp) from Colpoda compare grandis, C. inflata pop. 2, Paracolpoda steinii pop. 2, and Paracolpoda steinii pop. 3 isolates. Therefore, each PCR product was cloned, and the partial Cox I sequences were deposited in GenBank under the respective accession numbers OM752200, OM752201, OM752202, and OM752203. Their GC contents were 28.56%, 26.42%, 27.87%, and 27.75%, respectively, with sequence differences shown in Figure 4. Base variations between populations of Colpoda compare grandis, C. inflata, and Paracolpoda steinii were large, ranging from 12.01% to 14.88%, while the base variation between individuals within the Paracolpoda steinii population was small, at 0.35%.

3.3.2. The Utility of β-Tubulin Gene Tested for Accurate Identification

The β-tubulin amplification primers 349A and 349B (Table 1) generated a total of 33 DNA sequences of predicted length (~980 bp) from C. inflata (populations 1–2), Paracolpoda steinii (populations 1–5), and C. harbinensis n. sp. isolates. The interspecific genetic distances of β-tubulin of Colpoda ranged from 0.59% to 8.80%, and intraspecific genetic distances ranged from 0.89% to 5.81%. The TCS network of β-tubulin genes revealed the C. inflata, the largest difference between pop. 1 and pop. 2 was 60 genetic steps (5a and 2a), while the smallest difference was nine genetic steps (1b and 4). There were large genetic step differences among individuals within the same pop, e.g., 60 genetic step differences between 5a and 4 in pop. 2 (Figure 5(a)). Within Paracolpoda steinii, pop. 3 (7) and pop. 5 (11b) differed by 69 genetic steps, while pop. 4 (8a) differed from pop. 2 (3a) and pop. 3 (6a) by only one genetic step. In addition, there were large genetic step differences among the offspring individuals produced from the same individual by monoclonal cultures, such as 10 and 9 genetic step differences between 3a and 3b and 4a and 4b in pop. 2, respectively (Figure 5(b)).

3.4. Detection and Identification Using FISH

Our five probes were evaluated with the probe match tool in the ARB software package, revealing that they were specific to Colpoda (Table 3). There are one to six mismatches between the probes of different Colpoda species. After conducting fluorescence in situ hybridization with each of the five probes, Colpoda compare grandis, C. harbinensis n. sp., Paracolpoda steinii pop. 4, P. steinii pop. 5, C. inflata pop. 1, and C. inflata pop. 2 all exhibited red fluorescent signals (Figures 6(a)6(h) and 6(m)6(p)), clearly distinguishable from the faint autofluorescence signals achieved with negative-control hybridizations using the TSBs probe to hybridize the untargeted ciliates C. reniformis (Figures 6(i) and 6(j)) and Coleps hirtus (Figures 6(k) and 6(l)). FISH also provided some morphological information such as body shape, macronucleus shape, and macronucleus number (Figures 6(b), 6(d), 6(f), 6(h), 6(n), and 6(p)). The signal intensity became weaker when the formamide (FA) concentration increased in the hybridization buffers, and the fluorescence signals with more than 10% FA were weaker than those of the positive control. Therefore, 10% of formamide in the hybridization was the optimal concentration for the stringency of our probes.

3.5. Establish of New Species

Here are the following classifications of new species:

Class: Colpodea Small and Lynn, 1981

Order: Colpodida Puytorac et al., 1974

Family: Colpodidae Bory De St. Vincent, 1826

Genus: Colpoda Müller, 1773

Species: Colpoda harbinensis sp. nov

Diagnosis is as follows: size in vivo approximately , reniform in outline; narrower toward anterior and wider towards posterior; one spherical macronucleus, micronucleus sometimes nonexistent; 11–15 somatic kineties; five or six postoral kineties; left oral polykinetid elongate elliptic, composed of an average of 13 kineties; a few pronounced diagonal grooves present; and soil habitat. Type locality is as follows: soil from Hulan Beet Research Institute of Heilongjiang University (45°5947N, 126°3818E), Harbin, Heilongjiang province, northeastern China. Type specimens were as follows: the slide containing the holotype specimen (Figures 7(d) and 7(e)) and a paratype slide (registration number SYM–2020301011–02) are deposited in the Laboratory of Protozoology, Harbin Normal University. ZooBank registration was as follows: present work: urn:lsid:http://zoobank.org/:pub:2E33F1C0–CF47–4126–B317–C3505BC41C46. New species: urn:lsid:zoobank. Org:act: 485F1 A9C–4078–4F62–8C9C–06A6527EE730. Etymology was as follows: the species group name “harbinensis” indicates that this species was isolated from a sampling site in Harbin, Heilongjiang province, northeastern China.

3.5.1. Morphological Description (Figure 7 and Table 4)

Cell has a size approximately in vivo, usually about , length to width ratio close to 1.5 : 1 in life (Table 4): reniform in outline (Figures 7(a), 7(d)7(i), 7(k), and 7(l)) and straight keel and distinctly projecting ventrally, with four or five notches (Figures 7(a) and 7(g)7(j)). Buccal field occupies approximately one fifth of body length, funnel opening about 7 μm wide in vivo. Cytoplasm colorless contains several minute (<0.5 μm) crystals, mainly concentrated in the lower right corners, glistening under interference contrast illumination; only a few pronounced diagonal grooves were observed (Figures 7(g)7(i)). Macronucleus globular to slightly ellipsoid, on average, was generally above mid-body right of median (Figures 7(a), 7(b), 7(e), 7(n), and 7(p); Table 4). Micronucleus ellipsoid-shaped was attached to macronucleus, about in vivo (Figures 7(b) and 7(n); Table 4), sometimes nonexistent. Contractile vacuole was slightly ahead of posterior end, approximately 4 μm in diameter during diastole (Figures 7(a) and 7(g)7(i)), without tubular drainage pore. Cortex inconspicuous, flexible, extrusomes was recognizable in vivo (Figures 7(c) and 7(i)). Cytoplasm contains numerous granules, variably sized bacteria-filled food vacuoles, and crystals (Figures 7(a) and 7(g)7(j)) and moderately fast spiral movement on a substrate and rapid spiral swimming in water.

Typical Colpoda ciliature pattern was as follows: somatic cilia (approximately 8 μm long) was closely arranged (Figures 7(a) and 7(g)7(i); Table 4) and was densely arranged in the anterior part of the oral cavity, distinctly spiral, and roughly “S”-shaped, ranging in number from 11 to 15, each composed of monokinetids (Figures 7(a), 7(d), 7(e), 7(k), and 7(l); Table 4). Left oral polykinetid situated on elongate elliptic and consisting of an average of 13 minute kineties: five postoral kineties (Table 4; Figures 7(a), 7(d), and 7(k)).

3.5.2. Gene Sequence Data

The SSU rDNA sequence of Colpoda harbinensis sp. nov. has been deposited in the GenBank database with the accession number, length, and G + C content as follows: MZ557804, 1716 bp, and 44.23%.

4. Discussion

4.1. Comparison of Known Species with Original Descriptions
4.1.1. Colpoda reniformis Kahl, 1931

Our population of C. reniformis is similar to previous populations, as they share a distinctly nephrogenic body shape in vivo and an ellipsoid macronucleus between their vestibulum and dorsal side but is distinct in their large body size ( in the present study vs. 90–100 μm) and absence of micronucleus (vs. presence in the previous populations [30, 55].

4.1.2. Colpoda Compare grandis Smith, 1899

Colpoda compare grandis has many features that are similar to those of C. grandis: body reniform in vivo (about 2 : 1) with a distinct indentation at its vestibular entrance sometimes absent, laterally flattened, no postoral sack, contractile vacuole, cytopyge near its posterior end, extrusomes conspicuous and numerous, left oral polykinetid on the vestibular bottom, and elongate square [30, 56]. However, Colpoda compare grandis differs from C. grandis by the shape of the macronuclei (round vs. distinctly oval in C. grandis; Smith [56]). However, the morphology of macronucleus alone is not sufficient to distinguish Colpoda species. Considering the slightly variable shape of the macronucleus in Colpoda, the insufficient number of specimens investigated in this study, and the close phylogenic relationship with C. grandis based on the SSU-rRNA gene sequences, we temporarily identify our isolate as Colpoda compare grandis.

4.1.3. Colpoda inflata Stokes, 1884

Both the two Chinese populations of C. inflata have typical “L”-shaped body with a marked preoral narrowing and a hemispherical postoral portion, similar numbers of somatic kineties, and postoral kineties with those of previous studies [5759]. The body size of pop. 1 did not differ much from previous studies; although, the body size of pop. 2 was much larger ( in the previous populations compared to in the present study) [59].

4.1.4. Paracolpoda steinii Maupas, 1883

Compared with the previous studies, the four Chinese populations of P. steinii are similar in the following characteristics: dikinetid, two longer caudal cilia, a distinctly ellipsoidal macronucleus placed in the posterior half of the body, and a comma-shaped micronucleus [30, 59, 60]. The main difference of the Chinese populations is their larger body size ( in our populations vs. ), which may have resulted from the increased nutrition of our cultures.

4.2. Phylogeny of Genus Colpoda

Among the polygenes with small subunit ribosomal RNA genes (SSU-rRNA), the genus Colpoda was nonmonophyletic, consistent with previous studies [13, 36]. Typical Colpoda species are unlikely to unite into a single clade because they are spread throughout the order Colpodida, and some species (e.g., Colpoda maupasi and C. ecaudata) often form unexpected clades with two or more genera that have little in common morphologically [13]. This is also observed in previously constructed phylogenies (e.g., [10, 35, 36, 56]) by Foissner et al. [61]. Dunthorn et al., [19] even proposed that there exists a strongly radiating Colpoda, in which several species subsequently evolved independently to form new genera and families. We augmented the taxon sampling within the genus Colpoda with seven newly sequenced taxa, and our results support these earlier analyses, indicating a nonmonophyletic topology of Colpoda. In the 18S-rRNA gene phylogenetical analysis, five Colpoda species (C. reniformis, Colpoda compare grandis, C. inflata, Paracolpoda steinii, and C. harbinensis n. sp.) appeared in the core of Colpodidae with medium to high support. Paracolpoda steinii pops. 1–4 were sister to the clade clustered by P. steinii and Bromeliothrix metopoides. This discrepancy may be due to the fact that the SSU rRNA gene is too conservative in Colpoda to differentiate species.

4.3. DNA Barcoding of the Colpodea Species
4.3.1. The Utility of the Cox I Gene Inaccurate Identification

Extensive barcode analyses of the animal kingdom indicate that sequence divergences in mitochondrial genes encoding Cox I can distinguish closely related animal species [6264]. In the model protist genus Tetrahymena, intraspecific Cox I divergence is typically >4% [6567]. Interestingly, Colpoda compare grandis, C. inflata, and Paracolpoda steinii differed by 12.01%–14.88% in the Cox I gene, strongly suggesting that the three species were distinct. In contrast, the intraspecific genetic variation of Paracolpoda steinii was only 0.35%, indicating that the Cox I gene could represent an applicable DNA barcoding region for accurate and rapid identification of Colpoda. However, based on our experience, we conclude that it is difficult to design primers to amplify the Cox I gene in Colpoda.

4.3.2. The Utility of the β-Tubulin Gene Inaccurate Identification

The β-tubulin gene is another strong candidate gene for the delineation of Colpoda, given that it displays a diverse array of microtubules composed of tubulin with highly similar sequences [68, 69]. Specific regions of the β-tubulin gene are highly conserved, making it possible to design universal primers, while regions containing hypervariable sequences can be used to generate species-specific primers for accurate identification. In this study, there were no clear boundaries between intra- and interspecific genetic distances for each of the Colpoda. The intraspecies variation in the β-tubulin gene in the Colpoda was considerable, as indicated by the haplotype network, with a difference of 60 genetic steps between pop. 1 and pop. 2 in C. inflata genetic steps (5a and 2a) (Figure 5(a)) and 69 genetic steps between pop. 3 (7) and pop. 5 (11b) in Paracolpoda steinii (Figure 5(b)). Therefore, the β-tubulin gene may be less suitable for Colpoda DNA barcoding than Cox I.

4.4. Species Identification by FISH

In this study, five probes were developed to accurately identify Colpoda (Table 3). Using Coleps hirtus instead of Colpoda species as a negative control is more effective to test the probe’s specificity. Following Fried and Foissner [25], we evaluated our probes with the ARB software package and the GenBank BLAST tool to analyze the probe’s specificity. Previous studies have already demonstrated the power of this method for specific delineation. Nevertheless, the probes still require consolidation with the support of isolation and/or sequencing of Colpoda. Our study reveals that FISH can be used for rapid and interspecific identification of Colpoda and can also provide some morphological information such as body shape, macronucleus shape, and macronucleus number, which will help verify morphotypes in mixed taxa samples. However, while Colpoda species are geographically dispersed (e.g., Korea, U.S.A, and China), limited molecular data from disparate isolates are available [12, 13, 34, 36, 70]. The FISH probes designed here can potentially be used to investigate the geographic distribution of Colpoda and potentially even their dispersal.

4.5. Morphological Comparison of Colpoda harbinensis n. sp. with Other Congeners

The most important criteria for species identification in Colpoda are their body size and shape, oral characteristics, and the number of somatic kineties [30]. Considering the body shape, size, and number of somatic kineties, three specific species should be compared with the new species: Colpoda inflata, C. maupasi, and C. cucullus. Compared with Colpoda harbinensis n. sp. (Figure 8), Colpoda inflata has a different body shape (mainly “L” shaped) and more somatic kineties (20–25 vs. 11–15 in C. harbinensis n. sp.) [30, 57, 58]. This distinctive L-shape is produced by a marked preoral narrowing, and a hemispherical postoral portion which juts out at almost right angle vs. reniform in outline with their posterior ends broadly rounded in C. harbinensis n. sp. Colpoda maupasi is more elongated in shape ( vs. in C. harbinensis n. sp.), with more somatic kineties (15–18 vs. 11–15 in C. harbinensis n. sp.) [30, 59, 71]. Colpoda cucullus can be easily separated from C. harbinensis by having more somatic kineties (26–38 vs. 11–15 in C. harbinensis n. sp.) and postoral kineties (8–12 vs. 3–5 in C. harbinensis n. sp.) [30, 58, 72].

5. Conclusion

In conclusion, our analysis is consistent with previous study showing that no single marker can delineate microbial species [73]. Combining morphological and molecular biology techniques can greatly improve the delineation of Colpodea. We suggest that Cox I is a promising DNA barcoding marker for species of Colpodea, as shown in this and previous studies [22, 65, 74, 75]. However, difficulties with amplification may challenge its utility in identifying this group. The FISH can provide some morphological information, thus complementing traditional techniques such as silver carbonate. Furthermore, the establishment of a character-based database may be a useful tool for resolving conflicts between morphological or molecular approaches to the differentiation of not only Colpodea but also ciliate species in general.

In conclusion, we investigated and compared the morphological features of Colpoda reniformis, Colpoda compare grandis, Colpoda inflata, and Paracolpoda steinii, revealed the phylogeny of Colpoda, explored the feasibilities of Cox I and β-tubulin as DNA barcoding, and supplied the identification of Colpoda species using oligonucleotide probes. In addition, we have established a new species of Colpoda. The novelty of this study mainly displays in following several aspects: (1) molecular techniques are used for the identification of Colpoda for the first time; (2) oligonucleotide probes and haplotype network analysis are firstly conducted for the identification of Colpoda species; and (3) the comparative exploration is made for the feasibility of Cox I and β-tubulin genes as DNA barcoding.

Data Availability

The data presented in the study are deposited in the NCBI database repository, accession numbers: OM752200, OM752201, OM752202, and OM752203.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This work was supported by the Natural Science Foundation of China (project number: 31970498) and Shenzhen Science and Technology Program (Grant No. KQTD20190929172630447). Many thanks are given to Xinyuan Zhang and Rong Gao, for their help on sampling.