**1. Introduction**

*Paramecium* is one of the most studied genera of ciliates. Currently, at least fourteen morphological species of *Paramecium* are recognized as valid, and several more require reinvestigation [1–3]. Most of the morphological species include a number of genetically isolated groups, referred to as syngens. In some cases, syngens have been elevated to full species (the *P. aurelia* species complex, [4]) or could be described at least as genetic species due to the well-proved absolute reproductive barrier separating them, such as syngens in *P. bursaria* [5]. Additionally, several cryptic species of *Paramecium* have been reported [3], ye<sup>t</sup> they are disputed as true species due to the failure to establish their laboratory culturing and to collect more specimens in nature. Finally, some contested *Paramecium* species were documented only once from specific localities and were not subjected to thorough morphological, physiological or molecular description [1,6,7].

Most of the *Paramecium* morphospecies have recognizable morphological traits, allowing easy and fast identification of new representatives isolated from nature [1]. Numerous molecular phylogenetic analyses of inter- and intraspecific diversity of *Paramecium* have been performed, and molecular barcoding of different genes has allowed researchers to discriminate between morphospecies when morphological features were blurred [8,9], or even between sibling species or syngens within morphological species [5,10–12]. While the 18S rRNA gene has proven to be too conservative to disclose the intraspecies relationships in *Paramecium*, the cytochrome C oxidase subunit I (COI) gene is routinely used as a barcoding sequence in *Paramecium* molecular phylogenetic analysis [3,9,13,14]. This gene is sufficiently divergent and permits the inference of reliable genus level tree configuration and existence of intraspecific groups within each *Paramecium* morphospecies [14–16].

Extensive sampling, even in well-studied territories, allows researchers to find new *Paramecium* species or to confirm and validate species previously described [2,3,17,18]. Moreover, as many geographic regions remain poorly studied for *Paramecium* species occurrence, it is very possible that knowledge of the diversity of this genus is incomplete. Central America has not been surveyed systematically for *Paramecium* diversity. In this study, we provide new data on *Paramecium* occurrence and distribution in Mexico, including description of a new species of the *P. aurelia* complex found in two remote localities. Additionally, we report the presence of bacterial symbionts in some collected *Paramecium* strains since their occurrence has been considered as a possible criterion for *Paramecium* species [19].

#### **2. Materials and Methods**

#### *2.1. Sampling and Maintenance of Paramecium Strains*

About 130 freshwater samples were taken from more than 40 different waterbodies (natural and artificial lakes and ponds, canals, streams, drains, wetlands, and even a water reservoir in a roof-top garden) in several localities of seven states of Mexico: Ciudad de México, Estado de México, Hidalgo, Querétaro, Veracruz, Quintana Roo, and Yucatán in January–March 2019. The volume of each water sample collected was 10–30 mL. Samples from the same waterbody were taken at a distance of at least 20 m from each other and, thus, were considered as representing separate populations of ciliates. The samples were quickly transported to the laboratory and screened for paramecia within 24 h after sampling. Ciliates were detected under a Nikon SMZ 800 (Nikon Corporation, Tokyo, Japan) stereomicroscope. Then, all samples were kept on rice grains for 10–14 days and monitored every three days for previously unnoticed paramecia to show up. Several cells from each "positive" sample were isolated separately into depression slides; when possible, we isolated up to 10 cells from each sample, aiming to represent paramecia of different sizes and cell shapes. The ciliates introduced in culture were maintained on lettuce medium bacterized the day before use with *Enterobacter cloacae*, and supplemented with 0.8 mg/<sup>L</sup> of β-sitosterol (Merck, Darmstadt, Germany), as described earlier [19]. All currently alive strains used in the study are available upon request from RC CCM collection (World Data Centre for Microorganisms, RN 1171), Saint Petersburg State University, Saint Petersburg, Russia.

#### *2.2. DIC Microscopy and Stainings*

Live cells observations were made with differential interference contrast (DIC) microscopy with a Nikon Labophot-2 microscope equipped with a Nikon Digital Sight DS2Mv (Nikon Corporation) camera. We observed the cytological features important for quick species identification in *Paramecium*, namely cell size and shape, size, number and structure of micronuclei, structure of contractile vacuoles, and presence of algal symbionts [1]. Several staining techniques were employed, including the Feulgen procedure in De Lamater protocol, Harris hematoxylin, silver nitrate impregnation after Champy's fixation, silver carbonate and protargol [20]. Morphometric measurements were taken from stained cells.

#### *2.3. Molecular Identification of Paramecium Strains and Bacterial Symbionts*

*Paramecium* strains isolated from all samples and attributed to different morphospecies were subjected to sequencing of the mitochondrial COI gene. Additionally, in order to reconstruct a complete molecular phylogenetic tree of the *P. aurelia* species complex, COI gene sequences were obtained for the *P. primaurelia* strains Ir 4-2 (Russia) and FT11 (Pakistan), *P. pentaurelia* strains NR-2 (USA) and Nr1-9 (Russia), *P. septaurelia* strains 227 and 38 (USA). The total cell DNA was extracted from 100 to 200 cells of each strain using the GenElute Mammalian Genomic DNA Purification Kit (Sigma, Germany), according to the protocol "Genomic DNA from tissue". The PCR was performed using Encyclo Taq polymerase (Evrogen, Russia). The 767 bp-long partial COI gene sequences were amplified using the primers F388 and R1184, which are suitable for the majority of *Paramecium* species as described by Strüder-Kypke et al. [9]. For *P. primaurelia*, *P. triaurelia*, *P. pentaurelia* and *P. septaurelia*, the primers COI-long F (GATAAGGCTTGAGATGGCATACCCAGGAAG), and COI-longR (CAAAACCCATGTAAGCCATAACGTAGACAG) were designed, and 35 cycles of PCR were performed with an annealing temperature of 60 ◦C. Additionally, the partial sequences were obtained for the mitochondrial cytochrome C oxidase subunit II (COII) gene of some strains of *P. biaurelia* and presumably new species of the *P. aurelia* complex (see below; GenBank accession numbers MT318927–MT318930). In all cases, the same primers were used for PCR and sequencing. The partial 16S rRNA gene sequence for bacterial symbionts inhabiting *P. putrinum* strain K8 was amplified by PCR using the primers 16S alfa F19 and 16S alfa R1517, and sequenced with the primer 16S F343 [21]. All oligonucleotides were synthesized by Eurofins DNA (Germany). The PCR products were directly purified and sequenced at the Core Facility Center "Molecular and Cell Technologies" (St Petersburg State University, Saint Petersburg, Russia).

#### *2.4. Molecular Phylogenetic Analysis*

The COI gene sequences obtained in this study (GenBank accession numbers MT078136–MT078152, MT318931–MT318935) were aligned with COI gene sequences manually selected from GenBank or retrieved from ParameciumDB [22] for the strains with sequenced mitochondrial genomes [23]. Manual selection of entries was performed, since many COI gene sequences in GenBank are either too short, incorrectly assigned to a species or simply missing for some species in GenBank. The longer sequences were trimmed manually to obtain the 767 bp-long COI gene fragment, and the incomplete COI gene sequences from Genbank were chosen so that all sequences in the final set were at least 600 bp long. MUSCLE algorithm was used for alignment (online multiple alignment program [24]). Phylogenetic trees were constructed to infer the relationships of all isolated *Paramecium* strains using Phylogeny.fr [25,26]. The trees were computed by the bootstrapping procedure (500 bootstraps) and approximate likelihood ratio test method PhyML 3.1/3.0 aLRT [27]. The Maximum Likelihood analysis was performed with the HKY model.

## *2.5. Mating Tests*

For the strains of the presumably new species of the *P. aurelia* complex, preparation of cell lines for mating tests and studies of autogamy were carried out by method of daily re-isolations [28]. The sexually reactive cultures were mixed with each other and also with *P. biaurelia* tester strains IST, Rieff, and Ts from the RC CCM collection. Strains Rieff and IST belong to the same mating type compatible with the Ts strain mating type. The conjugation was observed only between +18 ◦C and +21 ◦C. The conjugating couples were picked with the Pasteur pipette, then exconjugant cells were isolated into separate microaquariums, and F1 clones were established. The fragment of the nuclear *mtB* gene involved in mating type control [29] was amplified for F1 clones in 32 cycles of PCR using the primers bi-mtBF (GCACACCCTCTTAAAATAAGT) and bi-mtBR (AAATCTCGCAAACAACTACTG) with an annealing temperature of 55 ◦C. This fragment was sequenced using the same primers to confirm the heterozygosity of F1 clones (i.e., to confirm the exchange of pronuclei in conjugation), as allelic single nucleotide polymorphisms were visible on chromatograms as double peaks (data not shown). F2 progeny were obtained after autogamy of F1 clones, and survival rates of the F2 generation were counted. The F2 clones were considered as viable if they completed more than 6 divisions after autogamy, since old macronuclei remain functional during the first 5–6 vegetative divisions in *Paramecium* [19], and confer otherwise inviable cells to survive during that period.
