Oxytricha atypica, Fan & Yao & Luo & Dong & Xu & Chen & Bourland & Zhao & Huang, 2021

Fan, Xinpeng, Yao, Shunli, Luo, Xiaotian, Dong, Tianyao, Xu, Yuan, Chen, Lingyun, Bourland, William, Zhao, Yan & Huang, Jie, 2021, Some morphologically distinguishable hypotrich ciliates share identical 18 S rRNA gene sequences - taxonomic insights from a case study on Oxytricha species (Protista, Ciliophora), Zoological Journal of the Linnean Society 193 : -

OXYTRICHA ATYPICA SP. NOV.

ZooBank registration: urn:lsid:zoobank.org:act:4D6567E6-DBD7-41DD-92C5-D30B087220E0 .

Type locality: Farmland near Zulfi City, Riyadh, Saudi Arabia (26°22’01” N, 44°46’03” E) .

Etymology: The species-group name atypica (Latin adjective, feminine gender, untypical) refers to the atypical number of post-oral ventral cirri compared with other Oxytricha species.

Deposition of type slide: The slide containing the holotype specimen ( Fig. 1B, C, R) has been deposited in the Biological History Museum, East China Normal University, China (registry no. TWJ2013051601-01). One paratype slide has been deposited in the Laboratory of Protozoology , East China Normal University, China (registry no. TWJ2013051601-02) .

Diagnosis: Size in vivo 80–130 × 30–50 µm, shape elongated ellipsoid. Extrusomes (cortical granules) colourless, densely arranged on dorsal and ventral side in vivo, extrude as ellipsoidal structures with concavity on the exterior end. Usually two macronuclear nodules and two micronuclei. Adoral zone occupies 36% of body length in vivo, composed of about 29 membranelles on average. Paroral and endoral in Stylonychia -pattern. Right and left marginal row containing about 28 and 25 cirri, respectively. Seven post-oral ventral cirri. Five dorsal kineties. Usually four caudal cirri, two of which on dorsal kinety 4.

A COMPARISON OF O. GRANULIFERA -RELATED POPULATIONS

O. granulifera , the type species of Oxytricha , established by Foissner & Adam (1983), was separated into two subspecies, namely, O. granulifera granulifera and Oxytricha granulifera quadricirrata , by Blatterer & Foissner (1988). However, in the monograph of Berger (1999), the latter was elevated to species rank. Therefore, O. granulifera comprises two valid subspecies, O. granulifera chiapasensis established by Méndez-Sánchez et al. (2018) and O. granulifera granulifera . Méndez-Sánchez et al. (2018) revealed that populations of Oxytricha granulifera do not form a monophyletic group but are intermingled with Paraurostyla viridis (Stein, 1859) ( AF508766 , possibly a misidentification) and Architricha indica ( KJ000536 ). Our 18S rRNA gene tree showed a similar topology on the subdivisions of O. granulifera (i.e. three clades were recognized) and further confirmed the polyphyly of O. granulifera . Moreover, we discovered that O. atypica , the new species with a non-18 FVT cirral pattern, fell within O. granulifera in phylogenetic trees based on both the 18S rRNA gene and the18S-ITS1-5.8S-ITS2-28S rRNA region. Considering that A. indica is an 18-FVT oxytrichid characterized by its multiple marginal rows which develop through five independent marginal primordia arising “within-row” ( Gupta et al., 2006; Xu et al., 2015), we revealed that the so-called O. granulifera populations, rather than being monophyletic, were interrupted by taxa with entirely different ciliature patterns. The following morphologic similarities might explain, to some extent, the cluster of the three species: (1) the FVT cirri arise from six primordia, which utilize six parental cirri in their origin as is typical of Oxytricha species ; (2) they all bear elliptical colourless cortical granules (extrusomes) with a hollow

*Except “Body size in vivo ”, all the other characters were shown by the range and mean which were separated with “/”.§Caudal cirri number included in the cirri in left marginal row in the original report. “−” data not available.

*AM, adoral membranelles; LMC, left marginal cirri; PVC, post-oral ventral cirri; RMC, right marginal cirri;“**”, P <0.01; “*”, 0.01 ≤ P <0.05; “ns”, P ≥ 0.05. §Only the characters separate a certain population out of all the others among the five so-called O. granulifera populations (i.e. S3, Tibet, Type, Xi’an and Mexico) are shown, for more details see Supporting Information ( Table S2) .

cavity at their external ends as both the present and previous studies showed ( Zhang et al., 2014). However, the reason why A. indica and O. atypica are closer to some populations of O. granulifera , but not to others, remain unclear. By comparing all populations of morphologically characterized so-called Oxytricha granulifera (S3, Xi’an, Tibet, Mexico and Type), we observed overlap of taxonomically important morphometric and morphological features, excepting the number of dorsal kineties ( Table 2). However, principle component analysis of the selected characters shows Type is clearly separated from others ( Fig. 8). Pairwise tests show that: (1) the body length and width and the number of right marginal row cirri separate Type (P <0.01); (2) the body length, the adoral membranelle number, and the numbers of cirri in both left and right marginal rows separate Tibet (P <0.01 or 0.01 ≤ P <0.05); and (3) the body width and the number of both right and left marginal row cirri separate Mexico (P <0.01) ( Table 3). These findings lead us to reconsider the implications of morphological dissimilarities for the identification of O. granulifera populations. Since dorsal kinety number might vary within a given population, as seen in O. granulifera chiapasensis ( Méndez-Sánchez et al., 2018: table 1), and might also vary among individuals from a clonal culture (from authors’ unpublished data), we cannot adopt five dorsal kineties as a specific feature of O. granulifera granulifera .

Tibet could be separated from Type and three other populations of O. granulifera . When re-examining the original data and slides, it was noticed that Tibet might also have distinctly more dorsal bristles in each of the dorsal kineties 1–5 based on the limited number of specimens in the late stage of morphogenesis (See Supporting Information). For instance, dorsal kinety 1 contained 32 dikinetids whereas all the other isolates of O. granulifera contained less than 20 ( Foissner & Adam, 1983; Kwon & Shin, 2013; Shao et al., 2014; Méndez-Sánchez et al., 2018). This discrepancy and the pairwise tests analysis implied that Tibet might represent an independent morphospecies rather than a population of O. granulifera and can be characterized by the greater body length and the greater number of adoral membranelles, marginal cirri and dorsal bristles. However, we choose not to establish a new species due to the lack of interphase voucher specimens showing this possible diagnostic feature (i.e. number of dorsal bristles). Thus, we accept the Tibet population as Oxytricha cf. granulifera and await its further characterization. It is noteworthy that, as is the case for O. atypica and Type, there is only one nucleotide difference in the 18S rRNA gene sequence between Tibet and Type.

Among all the populations we discussed here, O. granulifera chiapasensis, Mexico, has the lowest average numbers of adoral membranelles and right and left marginal cirri ( Table 2; Méndez-Sánchez et al., 2018). In contrast to the case of Tibet, this might indicate that the relatively fewer number of membranelles and cirri can be a potential diagnostic feature for the subspecies in addition to the features provided by Méndez-Sánchez et al. (2018). The validity of O. granulifera chiapasensis as a subspecies was supported by the 18S rRNA gene tree where an independent clade comprising this subspecies and an Indian population was formed, separated from the Type of O. granulifera granulifera .

The Saudi Arabia (S3) and Xi’an populations (Xi’an) have a longer adoral zone, relative to body length, and fewer left and right marginal cirri than Type; S3 also has fewer adoral membranelles ( Table 2; Foissner & Adam, 1983; Shao et al., 2014). However, we hesitate to use these features in delimitation of the two populations since they do not always show confident separation in the other pairwise tests i.e. S3 vs. Xi’an, S3 vs. Mexico, Xi’an vs. Mexico (Supporting Information, Table S2). Thus, we currently accept these populations as O. granulifera . Further study is needed as COI gene sequences become available for more populations, especially the sequence for a population from the type locality.

Considering the genetic difference and the phylogenetic relationships ( Figs 4B, 6), it is highly possible that sequences without morphological data ( AM412770 , JX899421 , KU715983 , AF164122 , AF508762 , AM412771 ), especially those in Clade I and II, represent cryptic species/new subspecies within O. granulifera or even new species of Oxytricha .

INSIGHTS FROM THE CASE STUDY

Should the taxonomy of hypotrich ciliates be based on clonal cultures?

In the taxonomic assessment of ciliates, studies of mixed (non-clonal) and clonal cultures of what is considered a single ciliate species serve different but complementary purposes. Characterization of a natural population (and its variability, great or small) of a species should be based on mixed cultures ( Zhang et al., 2018; Jung & Berger, 2019; Zhu et al., 2019), while characterization of nuclear and mitochondrial genomic architecture, metabolomics and so on are facilitated by the use of clonal cultures ( Zhao et al., 2013; Tarcz et al., 2014). However, notable morphological variations (e.g. number of marginal rows, number of caudal cirri) often exist among the individuals of a given mixed monoprotist culture of hypotrichids ( Lu et al., 2015) and euplotids ( Jiang et al., 2010, 2013). Although some of the variations are confirmed to be discrepancies of two daughter cells formed during morphogenesis ( Jiang et al., 2013; authors’ unpublished data), the possibility that raw-culture and “monoprotist” cultures contain a mixture of different species (sibling species) cannot be excluded. Therefore, in practice, for the “population” with obscure morphological variations between individuals, it is recommended to sequence the respective marker genes for multiple individuals to test their conspecificity. If sequences of multiple-individuals are non-identical, several clonal cultures should then be established and both morphologic and molecular studies should be performed for each culture to test whether the “populations” represent a single species or a mixture of different species. Additionally, multiple intra-individual sequences are suggested to reveal the intra-genomic sequence variation if any. It has been shown that extremely high rDNA copy number ( Gong et al., 2013; Wang et al., 2017, 2019) and heterogeneity occurring within mitochondrial DNA (mtDNA) occur at an exceptionally high level in some ciliates ( Zhao et al., 2013), and these properties could cause both intraspecific and intra-individual variation of rDNA and mtDNA. In any case, whether the intraspecific and intra-individual sequence differences come from the heterozygosity of two different alleles, should be determined by double checking the original chromatograms of the nucleotide that differs. It is also necessary to use high-fidelity polymerase for PCR to avoid experimental errors as suggested previously ( Wang et al., 2017).

The DNA-based taxonomy of hypotrich ciliates

Ever since 1990, sequences of the nuclear rRNA gene have been employed with exponentially increasing frequency to assist taxonomic assignment and systematic revision in the subclass Hypotrichia ( Schlegel et al., 1991; Hewitt et al., 2003; Moon et al., 2020). In the alpha-taxonomy of hypotrich ciliates, the 18S rRNA gene is currently the most frequently used marker (e.g. Shao et al., 2014; Kaur et al., 2018). However, there is no consensus as to the range of intra- and interspecific sequence divergence in the 18S rRNA gene. Schmidt et al. (2006) sequenced 20 clones of Stylonychia mytilus Ehrenberg, 1838 and Stylonychia lemnae Ammermann & Schlegel, 1983 and reported that a difference at only a single nucleotide position was detected between these two sibling species. Park et al. (2019) revealed the genetic divergence can be as low as 0.1% between two species of Pseudokeronopsis Borror & Wicklow, 1983 . Kim and Min (2019) showed a pairwise distance between Oxytricha seokmoensis Kim & Min, 2019 and its relatives ranging from 1.5 to 3.0%. The interspecific sequence identity of two genera, Notohymena Blatterer & Foissner, 1988 ( Aponotohymena sensu Foissner, 2016 ) and Paraurostyla Borror, 1972 , were measured by Xu et al. (2020), showing a range of genetic divergence of 0.8–3.6% and 0.5–3.4%, respectively. In our case, we found that three different morphospecies represented by O. atypica (S7), O. cf. granulifera (Tibet) and O. granulifera (Xi’an) share identical 18S rRNA gene sequences ( Fig. 4B). This underscores the importance of caution in using 18S rRNA gene sequences as evidence of conspecificity in lineages having a high degree of conservation of the 18S rRNA gene such as hypotrichs, especially in cases that display subtle morphological differences between the population under study and the original population. Sequence comparison of O. atypica (S7) and O. granulifera (S3) reveals that these two morphospecies differ by only one nucleotide in the 18S rRNA gene. This agrees with the data of Schmidt et al. (2006) showing that a speciesspecific nucleotide pattern comprising one or few, but stable, nucleotide difference(s) of the 18S rRNA gene might exist between closely related species. Therefore, small differences of rRNA gene sequences between different morphospecies can also be used as evidence of species delimitation. Among the rRNA gene regions, previous studies revealed that ITS1-5.8S-ITS2 and 28S rRNA performed better than the 18S rRNA gene for species identification and phylotype differentiation as their evolutionary rate is fast enough to provide higher diversity resolution than the latter ( Behnke et al., 2004; Vollmer & Palumbi, 2004; Beszteri et al., 2005; Windsor et al., 2006). More nucleotide differences were detected in the ITS1-5.8S-ITS2 and D1–2 of 28S rRNA genes of O. atypica (S7), O. cf. granulifera (Tibet) and O. granulifera (S3) and phylogenetic analyses including the two regions provided better discrimination of the three morphospecies ( Figs 4A, 7B, C). These results confirm the barcoding value of the ITS and D1–2 of the 28S rRNA gene in hypotrich ciliate taxonomy as previous studies concluded ( Santoferrara et al., 2013; Stoeck et al., 2014; Zhan et al., 2019).

Although the COI gene has been successfully used in the barcoding of Tetrahymena Furgason, 1940 , Paramecium O.F. Muller, 1773 and other groups of ciliates ( Barth et al., 2006; Lynn & Strüder-Kypke, 2006; Chantangsi & Lynn, 2008; Gentekaki & Lynn, 2009; Jung et al., 2011; Kher et al., 2011; Tarcz et al., 2012, 2013, 2014; Zhao et al., 2013, 2016; Liu et al., 2016), this marker has only recently been applied in the hypotrich ciliates ( Park et al., 2019). With primers developed by Park et al. (2019), we detected a COI gene sequence divergence of 11.8% between the two morphospecies (S3 and S7) of Oxytricha , similar to the interspecific divergence (13.6%) in two species of Pseudokeronopsis ( Park et al., 2019) . However, whether this threshold (i.e. approximately 10–15% divergence of the COI gene) is suitable for species level identification of all hypotrichids remains uncertain, since previous studies revealed that the levels of both intra- and interspecific variability in different ciliates might be genus specific and should be determined for each given lineage (Strüder- Kypke & Lynn, 2010; Zhao et al., 2013). However, the unique feature of wobble modifications in mitochondrial COI genes (resulting from higher rates of synonymous than non-synonymous substitutions) between the two morphospecies results in the COI gene carrying more genetic variation than that of the nuclear rRNA genes and consequently showing promise as a marker in the taxonomy of morphologically similar hypotrich ciliates as suggested in Park et al. (2019).