Systematics and Multi-Gene Phylogeny of the Subfamily Nothoholostichinae (Ciliophora, Hypotrichia), With Integrative Description of a New Marine Species Nothoholosticha luporinii n. sp.

Morphogenesis of ciliated protists attracts a lot of attention, because their huge morphological diversity is related to formation of ciliary structures during cell division. In the present work, the morphology and morphogenesis as well as the phylogenetic position of a new, marine hypotrich ciliate, Nothoholosticha luporinii n. sp., were investigated. The new species is characterized by having a combination of the following features: a bicorona whose anterior row contains four frontal cirri and posterior row includes only two cirri, a single buccal cirrus, midventral complex composed of about 30 cirral pairs, one pretransverse cirrus, 3–6 transverse cirri, one left and one right marginal cirral row; three bipolar dorsal kineties; contractile vacuole located in about 2/3 of the body length, two types of cortical granules, and many macronuclear nodules scattered throughout the cytoplasm. The morphogenesis of N. luporinii follows the ontogenetic mode of Pseudokeronopsis, a well-known and closely related genus except that the macronucleus fuses into a single mass in the middle fission stage. Phylogenetic analyses based on the rDNA operon classify Nothoholosticha in the family Pseudokeronopsidae and support the distinctness of the new taxon as well as the monophyletic origin of the subfamily Nothoholostichinae.

So far, only two species have been assigned to the genus Nothoholosticha, namely, N. fasciola (Kahl, 1932) Li et al., 2009 (type species) and N. flava Li et al., 2016. In this study, a new species, Nothoholosticha luporinii n. sp., has been discovered in the intertidal sediment of Chizhou Island near the city of Shenzhen in the southern China. Its morphology, ontogenesis, and complete ribosomal operon (SSU rDNA, the ITS1-5.8S-ITS2 region, and LSU rDNA) have been studied to further extended our knowledge about the diversity and phylogeny of pseudokeronopsids.

Sampling and Cultivation
Samples including sea water and sediment were collected from the intertidal zone in the Daya Bay, Chizhou Island, near the city of Shenzhen, southern China (22 • 38 11 N,114 • 38 32 E) on 1st April 2018, when the water temperature was 26 • C and salinity was 32 . The original sample was divided into aliquots that were used to establish raw cultures in Petri dishes. Single specimens of Nothoholosticha luporinii n. sp. were isolated from the raw cultures and used to set up clonal cultures in filtered in situ sea water at room temperature (25 • C). Some rice grains were added to stimulate the growth of bacteria, which served as prey organisms for ciliates.

Taxonomic Methods and Terminology
Nothoholosticha luporinii n. sp. was investigated using a combination of detailed in vivo observation and protargol impregnation. Living cells were observed under a microscope Olympus BX 53 (Olympus Corporation, Tokyo, Japan) using bright field illumination and differential interference contrast optics at a magnification of 100-1,000×. Protargol impregnation followed the Wilbert's method and served to reveal the nuclear apparatus and ciliary pattern of the new species (Wilbert, 1975). Also, the morphogenetic processes were rebuilt from the protargol-impregnated preparations. Stained cells were investigated mostly at high magnification (1,000×).
In vivo measurements were made from microphotographs of freely swimming specimens, while measurements on protargolimpregnated specimens were conducted using an ocular micrometer. Illustrations of living cells were based on freehand sketches and photographs, while those of impregnated specimens were made at 1,000× magnification with the help of a drawing device. All illustrations were finally processed in Adobe PhotoShop CS5. To distinguish parental and daughter structures during the morphogenetic processes, new (daughter) structures are painted solid, while old ciliary structures are depicted by contour. General terminology and systematics mostly follow Berger (2006) and Lynn (2008).

DNA Extraction, PCR Amplification, and Sequencing
Single cells were picked, carefully washed five times in filtered in situ marine water, and lysed in 45 µl of Cell Lysis Buffer (DNeasy Blood and Tissue Kit, Qiagen, Hilden, Germany). To confirm the sequencing results, altogether three different types of samples were prepared: the first contained just one cell, the second comprised two cells, and finally the third included three cells. Genomic DNA was extracted with the DNeasy Blood and Tissue Kit, but only 1/4 of the suggested volume for all reagents was used as suggested by Lu et al. (2020). Amplification of SSU rDNA was achieved with the primers 82-F (5 -GAA ACT GCG AAT GGC TC-3 ) (Jerome et al., 1996) and 18S-R (5 -TGA TCC TTC TGC AGG TTC ACC TAC-3 ) (Medlin et al., 1988). Fragments containing ITS-5.8S rDNA and LSU rDNA were amplified with the primers ITS-F (5 -GTA GGT GAA CCT GCG GAA GGA TCA TTA-3 ) (Miao et al., 2008) and 28S-R2 (5 -AAC CTT GGA GAC CTG AT-3 ) (Moreira et al., 2007), using the same thermo cycler program as described by Huang et al. (2014). PCR products were purified using the EasyPure R Quick Gel Extraction Kit (TransGen Biotech Co., Ltd., Beijing, China) and subsequently cloned using the pEASY R -Blunt Cloning Kit (TransGen Biotech Co., Ltd., Beijing, China). Recombinant plasmids were sequenced in both directions on an ABI-PRISM 3730 automatic sequencer (Applied Biosystems, Tsingke Biological Technology Company, Qingdao, China) with the PCR primers. To obtain high quality sequences, two internal sequencing primers were used for SSU rDNA: 900F (5 -CGA TCA GAT ACC GTC CTA GT-3 ) and 900R (5 -ACT AGG ACG GTA TCT GAT CG-3 ), and also two internal primers for LSU rDNA: F2 (5 -GGA GTG TGT AAC AAC TCA CCT GC-3 ) and R3 (5 -CAT TCG GCA GGT GAG TTG TTA CAC-3 ) (Zhao et al., 2014). Subsequently, the newly obtained sequences were carefully inspected, trimmed, and assembled into contigs using SeqMan Pro ver. 7.1.0 (Anson and Myers, 1997). Sequences obtained from all samples were identical and therefore only those derived from the single-cell sample were included into the subsequent phylogenetic analyses.

Molecular Phylogeny
The newly obtained sequences were blasted against the nucleotide NCBI database 1 . The BLASTn algorithm revealed that the new species belongs to the core urostylids (subclass Hypotrichia). Sequences of all related urostylids, except for those without associated publication information, were included into the phylogenetic analyses. The taxon sampling in the single-gene dataset (SSU rDNA) and in the concatenated, multi-gene dataset (SSU rDNA + ITS1-5.8S-ITS2 + LSU rDNA) mostly followed Huang et al. (2014) and Zhao et al. (2014). SeaView ver. 4 was used to prepare the concatenated dataset (Galtier et al., 1996;Gouy et al., 2010). Oxytricha granulifera (accession no. AF508762), Stylonychia lemnae (accession no. AF508773), Stylonychia mytilus (accession no. AF508774), Sterkiella nova (accession no. AF508771), and Sterkiella histriomuscorum (accession no. FJ545743) were used as outgroup taxa. GenBank accession numbers are provided in Supplementary Tables S1, S2. Sequences were aligned online using the MAFFT algorithm on the GUIDANCE2 server 2 with the following parameters: the 6mer pairwise method, the maximum number of 100 iterations, and 100 bootstrap repeats (Landan and Graur, 2008;Sela et al., 2015). The 5 and 3 ends of the resulting alignments were trimmed manually in the program BioEdit ver. 7.0 (Hall, 1999). The number of unmatched nucleotides and the pairwise SSU rDNA sequence identities within the subfamily Nothoholostichinae were calculated in the program BioEdit, using the "sequence difference count matrix" and "sequence identity matrix" options, respectively. The single-gene alignment contained 1,530 nucleotide positions, while the multigene alignment comprised 3,078 positions.
Maximum likelihood (ML) analyses were performed in RAxML-HPC2 ver. 8.2.10 on XSEDE (Stamatakis, 2014) on the CIPRES Science Gateway, 3 using the GTR + gamma evolutionary model and 1,000 bootstrap replicates. Bayesian inference (BI) was carried out in MrBayes ver. 3.2.6 on XSEDE (Ronquist et al., 2012) with the GTR + I + G evolutionary model selected by MrModeltest ver. 2.2 via the Akaike information criterion (Nylander, 2004). Markov chain Monte Carlo simulations were run for six million generations with a sampling frequency of 100 and a burn-in of 6,000 trees (10%). Remaining trees were used to calculate the 50%-majority rule consensus trees and their posterior probabilities. ML and BI trees were computed as unrooted and were rooted using the outgroup taxa in

Type Material
The protargol slide (no. ZTY2018040101_1) with the holotype specimen (Figures 1K,L, 2I) marked with an ink circle, and eight paratype slides (no. ZTY2018040101_2-9), are deposited in the Laboratory of Protozoology, Ocean University of China, Qingdao, China.

Gene Sequences
The nuclear SSU rDNA, ITS1-5.8S-ITS2 and LSU rDNA sequences have been deposited in GenBank under the following accession nos. MW035040, MW035039, and MW035042.

Dedication
We dedicate this species to Prof. Dr. Pierangelo Luporini (University of Camerino, Italy) in recognition of his great contributions to ciliatology.
Adoral zone of membranelles occupies 20-30% of body length; bipartite, i.e., divided into crown and lapel region separated by a conspicuous gap (Figures 1A,K, 2I-K and Table 1). Crown region composed of 7-13 membranelles arranged in an arch-shaped pattern along anterior cell pole, length of membranellar cilia 18-22 µm. Lapel region composed of 23-39 membranelles forming a Gonostomumlike pattern, i.e., extends along left body margin to one fifth or third of body length, where it bends rather abruptly rightwards to run almost in parallel with undulating membranes, length of membranellar cilia up to 17 µm. Undulating membranes arranged in a Pseudokeronopsis-like pattern, i.e., paroral and endoral almost straight, extend in parallel and only partially overlap. Endoral membrane commences posterior to buccal cirrus and runs to buccal vertex, 18-35 µm long after protargol impregnation. Paroral membrane begins anterior to endoral, remarkably shorter than endoral, i.e., only 6-15 µm long in protargol preparations (Figures 1K, 2I-K).

Oral Primordium and Cirral Streaks
Morphogenesis commences with de novo formation of small groups of basal bodies adjacent to left cirri of the midventral complex about in the mid-body ( Figure 6A). Groups of proliferating basal bodies join to form a longitudinal field, i.e., the oral primordium of the opisthe. Simultaneously, the proter's oral primordium develops as a single anarchic field of closely spaced basal bodies in the region of the buccal vertex (Figures 3A,  6B,C). New adoral membranelles differentiate within the oral primordium of both the proter and the opisthe in a posteriad direction (Figures 3B,C, 6D-F). The undulating membranes (UM) anlage (streak I) forms to the right of and possibly from the oral primordium both in the proter and the opisthe (Figures 3C, 4A, 6F, arrowheads). The anterior portion of the UM anlage splits a single cirrus that migrates anteriorly to become the leftmost frontal cirrus in the anterior row of the bicorona (Figures 4C, 6J,K, arrowheads). Then, the UM anlage divides longitudinally to give rise to the paroral membrane and the endoral membrane (Figures 4C, 6J-N, arrowheads). Meanwhile, multiple frontal-midventral-transverse (FVT) cirral anlagen develop as series of oblique streaks to the right of the oral primordia (Figures 3C, 4A,C, 6F,J,K, arrows). Streak II (FVT  , frontal-midventral-transverse cirral anlagen (red arrows), the anlagen for marginal cirral rows and dorsal kineties, and the dividing macronuclear mass. (C,D) Ventral (C) and dorsal (D) views of a late divider, showing the buccal cirri (red double-arrowheads), frontal cirri (red arrowheads), undulating membranes (purple arrowheads), frontal-midventral-transverse cirral anlagen differentiating into cirri (red arrows), and dividing macronucleus. DKA, dorsal kinety anlagen; LMA, left marginal cirral row anlagen; Ma, macronucleus; Mi, micronuclei; n, last frontal-midventral-transverse cirral anlage; n-1, n-2, n-3, n-4, second, third, fourth, and fifth rearmost frontal-midventral-transverse cirral anlagen; RMA, right marginal cirral row anlagen; S1-3, first, second, and third streaks. Scale bars = 140 µm.
anlage I) generates the second frontal cirrus of the anterior row of the bicorona as well as the buccal cirrus, which migrates toward the newly formed paroral membrane in mid-dividers; streak III (FVT anlage II) splits the third frontal cirrus of the anterior row of the bicorona as well as the left frontal cirrus of the posterior row of the bicorona; streak IV (FVT anlage III) produces the rightmost frontal cirrus of the anterior row of the bicorona as well as the right frontal cirrus of the posterior row of the bicorona; the rearmost (rightmost) streak develops the rightmost transverse cirrus and the 3-6 frontoterminal cirri, which migrate anteriorly to their species-specific position during the late division stages; streak n − 1 (n represents the last FVT cirral anlage) provides the rearmost midventral cirral pair, a single pretransverse cirrus and one transverse cirrus. Streaks n − 2 to n − 5 (deduced from morphometric data) contribute one midventral cirral pair and a single transverse cirrus each (Figures 4C, 5A,B, 6L,N-Q). The remaining streaks (FVT anlagen) provide one midventral cirral pair each (Figures 5A,B). When the formation of the new oral apparatus is almost completed in each daughter cell, new cirri migrate to their final positions.

Marginal and Dorsal Anlagen
The proter's left and right marginal cirral row anlagen form within the parental marginal cirral rows, very likely by dissociation of some cirri, about at level of the growing proter's oral primordium. Similarly, the opisthe's left and right marginal cirral row anlagen develop within the parental marginal cirral rows about at level of the opisthe's oral primordium (Figures 3C, 6I). The marginal row anlagen extend posteriorly, gradually producing new cirri already in early mid-dividers and stretch toward both ends of the dividing cell to form new ones for each daughter cell (Figures 4A,C, 5A,B). The morphogenesis of the dorsal side ciliature begins in mid-dividers ( Figure 3D). Specifically, within-row primordia appear in the parental dorsal kineties at two sites, viz., anterior and posterior to the prospective fission area (Figure 4B). The new dorsal kineties elongate and obtain their characteristic positions in late dividers (Figures 4D, 5C, 6M).

Nuclear Division
The macronuclear nodules fuse in mid-dividers to a branched mass (Figures 3D, 6G), becoming oblong before and during the cell fission ( Figure 4B). The elongate macronucleus divides into two pieces in late mid-dividers ( Figure 4D). Each piece develops into a tree-dimensional macronuclear reticulum that gradually fragments into individual ellipsoidal nodules in very late dividers ( Figure 5C). Micronuclei divide only once during the middle stages of binary fission (Figures 3D, 4B,D, 6G). More specifically, the micronuclei become spindle-shaped when the macronuclear nodules are fused into a branched mass (Figures 3D, 6G). Then, the micronuclei assume a dumbbell-shaped morphology as the macronuclear mass elongates into an oblong structure constricted in the middle ( Figure 4B). Finally, the daughter micronuclei are connected by an internal fiber bundle that conspicuously elongates in late dividers. During the postdivisional patterning, the micronuclei move to the scattered macronuclear nodules (Figure 5C).

Phylogenetic Analyses
Both maximum likelihood (ML) and Bayesian (BI) trees were constructed to determine the phylogenetic position of N. luporinii n. sp. (Figures 7, 8). Although the taxon sampling slightly differed between the single-and multi-gene datasets, N. luporinii was consistently assigned to the order Urostylida with very strong or full statistical support. All members of the family Pseudokeronopsidae, including N. luporinii, always clustered together with Anteholosticha pulchra with full statistical support. Monophylies of the Pseudokeronopsinae and Nothoholostichinae were strongly to fully statistically supported in phylogenetic analyses of both single-and multigene datasets. The Pseudokeronopsinae clustered with the Nothoholostichinae in the multi-gene trees, but with very weak support (50% ML, 0.79 BI). On the other hand, their sister-group relationship was not recognized in the single-gene trees, as the Pseudokeronopsinae grouped with A. pulchra, but with very weak support (63% ML, 0.62 BI).
According to the single-gene and multi-gene trees, N. luporinii clustered with full statistical support in the monophyletic subfamily Nothoholostichinae, which encompasses N. fasciola, N. flava, Heterokeronopsis pulchra Pan et al., 2013, andApoholosticha sinica Fan et al., 2014a. Within this subfamily, Numbers near branches denote bootstrap values for maximum likelihood (ML) and posterior probabilities for Bayesian inference (BI). Asterisk indicates the disagreement between ML and BI trees. GenBank accession numbers are provided after species names. For taxon sampling and GenBank accession numbers in the collapsed clades, see Supplementary Table S1. Scale bar corresponds to three substitutions per 100 nucleotide positions. the genus Nothoholosticha was depicted as monophyletic with very weak support in the single-gene analyses (73% ML, 0.91 BI), but with full statistical support in the multi-gene analyses. Nothoholosticha luporinii was shown to be most closely related to N. fasciola, but this relationship was very poorly statistically supported in all analyses. According to the multigene analyses, Nothoholosticha was sister to the monotypic genus Heterokeronopsis. And, finally, Apoholosticha was placed in a sister position to the Nothoholosticha-Heterokeronopsis cluster in the multi-gene trees.
The number of unmatched nucleotide positions and the pairwise SSU rDNA sequence similarities among members of the subfamily Nothoholostichinae are summarized in Table 2. Nothoholosticha luporinii differs from N. flava by 13 nucleotide positions (99.1% sequence identity), from N. fasciola by 14 nucleotides (99.1% identity), from H. pulchra by 21 nucleotides (98.6% identity), from T. silvanetoi by 26 nucleotides (98.3% identity, the 121 Ns positions in the T. silvanetoi sequence were coded according to other related species as this region was fully conserved), from A. sinica population 1 by 26 nucleotides (98.3% identity), and from A. sinica population 2 by 25 nucleotides (98.4% identity).

DISCUSSION
The New Species Nothoholosticha luporinii Nothoholosticha luporinii n. sp. can be easily separated from members of the genera Anteholosticha, Antiokeronopsis, Pseudokeronopsis, and Uroleptopsis by having an atypical bicorona composed of four anterior and two posterior frontal cirri (vs. only three frontal cirri in Anteholosticha and more than four anterior coronal cirri in the three latter genera); from members of the genus Apoholosticha by possessing (vs. lacking) buccal cirri; and from Heterokeronopsis by exhibiting (vs. lacking) frontoterminal and transverse cirri. On the other hand, the new species highly resembles taxa assigned to the genera Nothoholosticha and Tetrakeronopsis (Figure 9). Hitherto, Nothoholosticha comprises two species, N. fasciola and N. flava, while Tetrakeronopsis is monotypic and includes only T. silvanetoi.
Our newly discovered species cannot be confused with N. fasciola, the type species of Nothoholosticha, because it lacks frontoterminal cirri (Li et al., 2009). On the other hand, N. luporinii highly resembles N. flava in the body shape and size as well as in the number and arrangement of the macronuclear nodules and cirri (Li et al., 2016). Morphologically, N. luporinii can be distinguished from N. flava only by the lower number of crown (7-13 vs. 13-18) and lapel (23-39 vs. 36-47) adoral membranelles. However, the proper identification requires molecular data, because both differentiating characters partially overlap. The SSU rDNA sequences of N. luporinii and N. flava differ in 13 nucleotide positions ( Table 2), which undoubtedly supports the distinctness of both taxa.
Because of the high morphological similarity, it might be speculated whether N. luporinii should not be classified as a subspecies of N. flava. However, the 13 different nucleotide positions in the SSU rDNA stand strongly against this suggestion. This is, indeed, a very pronounced genetic difference as the SSU rRNA gene is highly conservative with a rate of only 1.24-3.96 × 10 −4 substitutions per site per one million years (Wright and Lynn, 1997;Vd'ačný, 2015;Vd'ačný et al., 2019). Kumar et al. (2017) showed that one base pair difference in SSU rDNA is sufficient to separate two closely related hypotrich taxa. Although Bistichella variabilis (HQ699895), Uroleptoides magnigranulosus (AM412774), and Orthoamphisiella breviseries (AY498654) are classified in different genera on the basis of morphological data, they differ only by 5-9 nucleotide positions (99.5-99.7% identity) in their SSU rDNA sequences (He and Xu, 2011). In this light, we find the species status of N. luporinii proposed in the present study to be justified.
Finally, N. luporinii can be distinguished from T. silvanetoi by the higher number of frontoterminal cirri (3-6 vs. invariably 2), the lower number of pretransverse cirri (1 vs. 2), the lower number of crown (7-13 vs. 12-17) and lapel (23-39 vs. 40-52) adoral membranelles as well as by the arrangement of the smaller type of bright brown-reddish cortical granules (clustered in a flower-like pattern around dorsal bristles vs. loosely scattered throughout the cortex). The morphological differences are supported also by the genetic differences in SSU rDNA sequences. Thus, there are 21 unmatched nucleotides between N. luporinii and T. silvanetoi ( Table 2).

Generic Classification of Nothoholosticha luporinii
Generic classification of hypotrichs is traditionally based on the cirral and oral patterns as well as on the morphogenesis of the ventral and the dorsal ciliature (for reviews, see Berger, 1999Berger, , 2006Berger, , 2008Berger, , 2011. The generic classification of N. luporinii is a difficult matter, because its cirral pattern is a mixture of features found in the type species of Nothoholosticha (an "extra" cirrus between the midventral complex and transverse cirri) and the monotypic genus Tetrakeronopsis (multiple frontoterminal cirri). More specifically, N. fasciola, the type species of Notholosticha, lacks frontoterminal cirri (Li et al., 2009), which are present not only in two further Nothoholosticha species, N. flava and N. luporinii, but also in T. silvanetoi (Paiva et al., 2014;Li et al., 2016;present study). The extra cirrus, situated between the midventral complex and transverse cirri, is interpreted as the last cirrus of the midventral complex in N. fasciola and N. flava (Li et al., 2009(Li et al., , 2016, while as a pretransverse cirrus in N. luporinii. Regardless of the terminology, it is very likely the same cirrus, which is derived from the second rearmost frontal-midventral-transverse cirral streak n − 1 (Figures 4C,  5A,B). Tetrakeronopsis exhibits even two cirri between the midventral complex and transverse cirri. Because these two extra cirri are situated ahead of the rightmost transverse cirri, they also might be designated as pretransverse cirri. The right pretransverse cirrus is derived from the rearmost streak n and the left pretransverse cirrus from the second rearmost streak n − 1 (Paiva et al., 2014). In this light, all Nothoholosticha species have retained only the left pretransverse cirrus and N. fasciola further lost the frontoterminal cirri, which are derived from the rearmost streak n in N. luporinii and Tetrakeronopsis (Figure 9). Because N. luporinii is more closely related to N. fasciola than to T. silvanetoi in SSU DNA phylogenies (Figure 7), we classify our new species in the genus Nothoholosticha. However, the results of the present phylogenetic analyses question the generic diagnostic value of the loss of frontoterminal and pretransverse cirri in N. fasciola, which are derived from the two rearmost frontal-midventral-transverse cirral anlagen. Nevertheless, we prefer to await discovery of further species to improve the diagnosis of the genus Nothoholosticha.

Classification Framework of the Subfamily Nothoholostichinae
The structure of the bicorona represents an important subfamily-level character that was used to divide the family Pseudokeronopsidae into two subfamilies by Paiva et al. (2014): the Nothoholostichinae, with an atypical bicorona whose anterior portion is formed by four frontal cirri, and the Pseudokeronopsinae, with a typical bicorona whose anterior portion is formed by more than four frontal cirri. The present phylogenetical analyses corroborate this subdivision as well as the monophyletic origin of the Nothoholostichinae (Figures 7, 8).
Interestingly, there are as many as three monotypic genera within the subfamily Nothoholostichinae. However, they can be easily distinguished from each other either by the loss or by the retention of some cirri, with respect to the type genus Nothoholosticha. Thus, Apoholosticha lost the buccal cirrus, Heterokeronopsis lost the frontoterminal and transverse cirri, while Tetrakeronopsis maintained both the left and the right pretransverse cirrus (Pan et al., 2013;Fan et al., 2014a;Paiva et al., 2014;Hu et al., 2015 ; Figure 9), which seems to be the plesiomorphic condition in the family Pseudokeronopsidae. Remarkably, frontoterminal cirri were lost in Heterokeronopsis and N. fasciola, but not in any other Nothoholosticha species. Homoplastic nature of this character thus questions its taxonomic significance at genus level. On the other hand, the loss of the buccal cirrus in Heterokeronopsis and of the right pretransverse cirrus in Nothoholosticha might serve as good generic characters at the present state of knowledge. Likewise, the long midventral row found in Heterokeronopsis, seems to be another good generic diagnostic feature for separation of genera within the subfamily Nothoholostichinae.

Multi-Gene Phylogenetic Analyses of the Family Pseudokeronopsidae
The SSU rRNA gene is very conservative and it is generally known that it bears mainly information for deeper nodes of phylogenetic trees. Therefore, manifold studies pursued to improve the knowledge about phylogenetic interrelationships within the subclass Hypotrichia using mostly sequences of that gene (e.g., Luo et al., 2017Luo et al., , 2018Song and Shao, 2017;Lyu et al., 2018;Kim and Min, 2019;Chen et al., 2020;Dong et al., 2020;Paiva, 2020;Park et al., 2020;Wang J. et al., 2020;Xu et al., 2020). On the other hand, the ITS1-5.8S-ITS2 region as well as the LSU rRNA gene are much faster evolving parts of the rDNA operon and hence bear phylogenetic signal also for more recent divergences (Abraham et al., 2019). Therefore, the concatenation of SSU rDNA, ITS1-5.8S-ITS2 region and LSU rDNA might lead to better resolved phylogenetic trees, as also evidenced in the present study (Figures 7, 8). For instance, Anteholosticha pulchra, a typical urostylid, clustered in a sister position to the subfamily Pseudokeronopsinae in the SSU rDNA tree, causing non-monophyly of the family Pseudokeronopsidae (Figure 7), as defined by Paiva et al. (2014). On the other hand, A. pulchra was placed outside the Pseudokeronopsidae in the multi-gene trees although the statistical support remained poor (Figure 8). This position is, however, much more consistent with morphological classifications, because A. pulchra possesses three enlarged frontal cirri while pseudokeronopsids have a bicorona. The internal branching pattern within the subfamilies Pseudokeronopsinae and Nothoholostichinae as well as the placement of the new species within the genus Nothoholosticha were also much better statistically supported in multi-gene than in the single-gene analyses (Figures 7, 8). Likewise, the sistergroup relationship of Uroleptopsis citrina and Pseudokeronopsis flava within the Pseudokeronopsinae obtained much better support in the multi-gene trees. Thus, the whole rDNA operon strongly suggests that the genus Pseudokeronopsis is nonmonophyletic and might be split into multiple genera in future. Pseudokeronopsis species are, however, highly similar in terms of their nuclear apparatus and cirral patterns (for details, see Li et al., 2017). This was the main reason why multiple species were synonymized, misidentified, or their species status was questioned. Nevertheless, the distinctness of most Pseudokeronopsis species was corroborated by analyses of SSU rDNA sequences along with ITS-5.8S rDNA sequences (Li et al., 2017). To summarize, the combination of the traditionally used SSU rDNA sequences with ITS region and LSU rDNA sequences improves phylogenetic inferences and classifications of pseudokeronopsids in specific and of ciliates in general.

DATA AVAILABILITY STATEMENT
GenBank accession numbers of sequences used in phylogenetic analyses can be found in the Supplementary Material.

AUTHOR CONTRIBUTIONS
WS and CS conceptualized the project. TZ and TC performed the laboratory work. TZ, YW, and JM prepared the data sets and conducted analyses. TZ and PV wrote the first draft of the article. YW, CS, and WS revised the manuscript. All authors approved the final version of the manuscript.