**4. Discussion**

#### *4.1. Overview of Paramecium Diversity Revealed in Mexico*

Screening of samples collected by us in Mexico showed unsurprisingly that *Paramecium* is widely represented in fresh waterbodies in this part of the world. The territory of Mexico, and in general, of Central America, remains unexplored by ciliatologists compared to Europe or some parts of the USA. Several *Paramecium* species, which we found in the current studies, were reported for the first time from these vast territories. To compensate the lack of knowledge about the *Paramecium* biogeography of Central America, we summarized the retrospective records of *Paramecium* according to [34–47]; all *Paramecium* records available from Mexico as well as Central and South America are shown in Table 4.

Besides the most important discovery of a new member of the *P. aurelia* species complex, it is worth noting several new facts on *Paramecium* biogeography. First is the occurrence of *P. putrinum* in the waterbodies of Mexico City. This species is not very frequent but can be considered common for temperate climate zones [1,11]. To our best knowledge, it has never been documented from tropical climates, but our observation is the southern-most collection of *P. putrinum* and may represent the southern extent of the species range. The same concerns *P. triaurelia* collected in Chapultepec Lake, Mexico City, as this species was not previously isolated as far south or in a tropical environment [4]. Record of *P. jenningsi* is notable, as this species is rare, and, unlike *P. putrinum*, usually inhabits waterbodies of tropical or subtropical zones [48]. This is only the third finding of *P. jenningsi* in the region, after Panama and Florida, USA [49]. *Paramecium jenningsi* strain DK from Mexico belonged to the same genotypic group as *P. jenningsi* strains collected in Japan and Madagascar [50]. Interestingly, *P. bursaria* strains found in two localities in Mexico belonged to syngen R3, common in Japan, China and also known from South America [5]. As for *P. caudatum*, all Mexican strains appeared to be related to each other and grouped in molecular phylogenetic trees with strains from Brazil, China and the European part of Russia (Figure 2), thus, confirming that the genotypes within these species do not show any special geographical pattern [3,14,32]. *Paramecium multimicronucleatum* strains found in Mexico belonged to three branches within this morphospecies. All branches include strains from all over the world.



CdM = Ciudad de México, Chi = Chiapas, EdM = Estado de México, Gro = Guerrero, Hgo = Hidalgo, Mor = Morelos, Oax = Oaxaca, Pue = Puebla, Qro = Querétaro, Q.R. = Quintana Roo, Sin = Sinaloa, Ver = Veracruz, Yuc = Yucatán, CA = Central America, SA = South America; records of this study are highlighted in red. \* according to [1,4,51].

#### *4.2. Paramecium Quindecaurelia n. sp. and a Species Concept in Paramecium*

Since Sonneborn [4] gave species rank to fourteen syngens of the *P. aurelia* complex, only one more member of this complex has been discovered: *P. sonneborni* [52]. Actually, it would be amazing if there were no more currently unknown sibling species of this complex in nature. Here, we report the sixteenth species of the *P. aurelia* complex, suggesting the name *Paramecium quindecaurelia* n. sp., as numerical tradition was interrupted with the description of *P. sonneborni*.

Morphologically, all species of the *P. aurelia* complex are indistinguishable [4], except *P. sonneborni*, which has a unique micronuclei morphology [52]. Morphometric characteristics do not show any significant variation among the sibling species of the complex [33]. The number of kineties is not considered important to differentiate between the *P. aurelia* species, while features of oral cortex are very conservative in *Paramecium* and cannot be used as a species characteristic within this genus [53]. Thus, morphological analysis of *P. quindecaurelia* n. sp., expectedly, revealed that it has no discriminating features and in particular, is extremely similar to its closest relative, *P. biaurelia*.

We reconstructed the complete phylogenetic tree for the *P. aurelia* species complex inferred from the COI gene sequence (Figure 3). *Paramecium quindecaurelia* n. sp. branches in the same cluster with *P. biaurelia*, but the COI gene identity between these two species does not exceed 95.0%, while within each of these species, it is not less than 99.3% (Table 2). The comparison of COII gene sequences of *P. quindecaurelia* n. sp. and *P. biaurelia* strains gives very similar values: 94.8–95.1% between two species and 99.6–99.9% within each species (Supplementary Table S2). The phylogenetic distance between *P. quindecaurelia* n. sp. and *P. biaurelia* inferred from COI gene sequence analysis is of the same range as between the most closely related sister species *P. primaurelia* and *P. pentaurelia* (Figure 3, Table 3) where identity of the COI gene sequences is 94.0–94.4%. According to Sonneborn's data [19], *P. primaurelia* and *P. pentaurelia* were genetically isolated from each other but had identical isozyme patterns while all other species of the *P. aurelia* complex were characterized by unique zymograms. The only additional differences between these two species were unstable mating type O in *P. pentaurelia*, while in the clonal life of *P. primaurelia* the mating types never changed [19], and "a very weak and unreliable mating reaction, not leading to conjugation, occurred between type E of *P. pentaurelia* and type O of *P. septaurelia*", but the same was not true for *P. primaurelia* [4]. Thus, *P. primaurelia* and *P. pentaurelia* are also very similar physiologically to each other.

It is not surprising that *P. quindecaurelia* n. sp. is able to mate with *P. biaurelia*, and that *P. primaurelia* can conjugate with *P. pentaurelia* [19]. However, existence of the pronounced reproductive barrier between *P. quindecaurelia* n. sp. and *P. biaurelia*, despite the isolation is not absolute, further confirms that we found a novel member of the *P. aurelia* complex. The survival of F2 progeny in crosses between *P. biaurelia* and *P. quindecaurelia* n. sp. is slightly greater than zero. It is assumed that the reproductive species criterion is reliable but not absolute, as in many zoological species interspecific hybrids are known, and sometimes they also are considered as separate species (for example, water frogs from the *Pelophylax esculentus* complex—[54]). The F1 interspecies hybrids of *P. biaurelia* and *P. quindecaurelia* n. sp. cross are perfectly viable, but lethality in the F2 generation is very high, so the hybrids can be considered effectively sterile. Unfortunately, there are no data in the literature on F1 and F2 survival in crosses between *P. primaurelia* and *P. pentaurelia*. In the third pair of closely related species, *P. tetraurelia* and *P. octaurelia* (Figure 3), F1 hybrids had problems with survival and growth, and F2 post-autogamous progeny never survived [55]. Still, their COI gene sequence similarity is just 83.9%, and the phylogenetic distance between the latter two species is bigger than between *P. primaurelia* and *P. pentaurelia* or between *P. biaurelia* and *P. quindecaurelia* n. sp. (see Figure 3 and Tables 2 and 3). Different molecular markers have different resolution in *Paramecium*. For example, comparison of 18S rRNA gene provides good overview of the whole genus phylogeny [8,15], while it becomes useless if applied to the *P. aurelia* complex [8]. Mitochondrial COI and COII genes are rather conserved within *P. aurelia* sibling species (polymorphism in these genes among strains of the same species does not exceed 1–2%), while even the closest sibling species differ at least in 5% of these gene sequences ([9,13] and this work). At the same time, in related to *Paramecium* genus *Tetrahymena*, the divergence for more than 1% in the *COI* gene sequence is considered reliable interspecific di fference [56]. Obviously, only molecular phylogenomic analysis would allow us to understand the diagnostic level of molecular divergence between closely related species of ciliates.

In our opinion, at present, our data are su fficient to claim that *P. quindecaurelia* n. sp. is a separate species and not just a divergent group in *P. biaurelia*. Interestingly, *P. biaurelia*, the most common species of the *P. aurelia* complex in cold and moderate climate zones, at least in Europe [57], is not known from tropical environments [4,58], which may shelter its twin species, *P. quindecaurelia* n. sp.

#### *4.3. Paramecium Quindecaurelia n. sp. Taxonomic Summary*

*Diagnosis.* Classical species of the *Paramecium aurelia* complex, sister species of *P. biaurelia. Type locality.* Los Dinamos National Park, Mexico City (19◦1602" N/99◦1731" W).

*Type slides.* Several holotype and paratype slides have been deposited in the collection of microscopical slides of the Department of Invertebrate Zoology, Saint Petersburg State University, Russia.

*Type culture.* The type strain D88-8 and other strains of the species are maintained in the RC CCM culture collection (World Data Centre for Microorganisms, RN 1171) of Saint Petersburg State University, Saint Petersburg, Russia.

*Sequence availability*. The nucleotide sequence of the mitochondrial cytochrome c oxidase I gene of the type strain was deposited in the NCBI GenBank database under accession number MT078138, the sequence of the mitochondrial cytochrome c oxidase II gene under accession number MT318928.

*Zoobank Registration LSID*: http://zoobank.org/urn:lsid:zoobank.org:act:B5A24294-3165-40DA-A425-3AD2D47EB8E7.

*Further remarks.* The strains of this species are able to conjugate with *P. biaurelia* strains. No endosymbionts have been detected in the species so far.
