**3. Results**

#### *3.1. Diversity of Paramecium and Its Bacterial Symbionts Revealed by Extensive Sampling in Several Regions of Mexico*

From all sampled Mexican localities, paramecia were found in 19 waterbodies. Strains from 30 populations of six morphospecies, namely representatives of the *P. aurelia* species complex, *P. jenningsi*, *P. caudatum*, *P. multimicronucleatum*, *P. bursaria*, and *P. putrinum* were introduced in the laboratory cultures. The most frequently recorded species was *P. multimicronucleatum* (15 populations from 9 waterbodies). Six populations from different waterbodies contained representatives of the *P. aurelia* species complex, while other morphospecies were relatively rare (Table 1). It was not uncommon to find several *Paramecium* species in the same community. The greatest diversity was detected in the lakes, ponds, and streams of the Cantera Oriente reserve (Mexico City), where we isolated a number of strains of *P. caudatum, P. multimicronucleatum, P. bursaria*, and *P. putrinum.* In several populations, some *Paramecium* specimens were inhabited by bacterial endosymbionts (Table 1, Figure 1). Most of these symbiotic bacteria still have to be identified and are currently being studied. Among the most interesting findings were presumably *Trichorickettsia* sp. abundantly present in host cytoplasm in several *P. putrinum* strains from Cantera Oriente (Figure 1A), as well as unknown cytoplasmic (Figure 1C,D) and intranuclear (Figure 1E,F) symbionts in different strains of *P. multimicronucleatum*. Tiny bacteria able to produce R-bodies resembling *Caedibacter* sp. or *Caedimonas* sp. [30] were found in cytoplasm of *P. tetraurelia* cells from Xochimilco Lake (Mexico City), while other cytoplasmic symbionts were detected in the *P. octaurelia* strain from Cenote Azul (Quintana Roo).



1


**Table 1.** *Cont.*

ND—not detected; 2 NA—non-applicable; 3 Similar in ChP5-3 and ChP3-4 strains; 4 Similar in K6, K8 and K11-3 strains.

**Figure 1.** Diversity of *Paramecium* and its bacterial symbionts discovered in Mexico. (**A**) *Paramecium putrinum* cell (strain K8) with abundant bacteria in cytoplasm; (**B**) *Paramecium jenningsi* cell (strain DK) with two species-characteristic micronuclei; (**C**) Cytoplasmic bacteria in squashed cell of *Paramecium multimicronucleatum* (strain SMM80-11); (**D**) Cytoplasmic bacteria in squashed cell of *Paramecium multimicronucleatum* (strain T42); (**E**) Congregations of bacteria in the macronucleus of *Paramecium multimicronucleatum* cell (strain ChP5-3); (**F**) Bacteria in the macronucleus of *Paramecium multimicronucleatum* cell (strain R51). Symbiotic bacteria are marked with the grey arrows. Mac = macronucleus, Mic = micronucleus. Scale bars: 6 μm (**A**), 10 μm (**B–F**).

#### *3.2. Phylogenetic Analysis of the Collected Strains*

The strain attribution to certain morphospecies by DIC microscopy of the living cells was confirmed by COI gene sequencing and further positioning of a strain within the *Paramecium* phylogenetic tree. Molecular characterization by COI and COII genes sequencing is the fastest and most reliable way to discriminate between di fferent species of the *P. aurelia* complex [9,10]. All sibling species form separate branches on trees inferred from these molecular markers [10,31], and COI and COII gene barcodes make it possible to identify each species of the *P. aurelia* complex, eliminating the need for laborious round-robin mating tests with representatives of all species of the complex. COI gene sequencing can also reveal di fferent haplotypes that cluster into intraspecific groups within *P. multimicronucleatum*[16,32] and into reproductively isolated syngens in *P. bursaria* [5]. However, the haplotypes revealed within *P. caudatum* do not form pronounced and well-supported branches [3,14]. Thus, analysis of COI gene sequences of Mexican *Paramecium* strains (Figure 2) showed that *P. bursaria* isolated from the lakes in Cantera Oriente (Mexico City) and Amealco (Querétaro) belonged to syngen R3, which is known to be widespread in Far East Russia, China, Japan, and South America [5]. *Paramecium multimicronucleatum* strains sorted into three branches within this morphospecies cluster. *Paramecium jenningsi* from Mexico City grouped together with strains of the species found in Asia and Africa (Figure 3). Finally, we succeeded in recovering four known species of the *P. aurelia* complex: *P. primaurelia*, *P. triaurelia*, *P. tetraurelia*, and *P. octaurelia*. Strains from two populations (Dinamos, Mexico City and Amealco, Querétaro) clustered in a separate branch as a sister species for *P. biaurelia* (Figure 3). This suggested to us that we may have discovered a novel member of the complex, so, these strains were thoroughly studied in order to figure out if they actually represented a new species of the *P. aurelia* complex.

**Figure 2.** Phylogenetic position of *Paramecium caudatum* and *Paramecium multimicronucleatum* strains collected in this study on the mitochondrial COI gene tree. The sequences of *P. chlorelligerum* and *P. bursaria* are included as outgroups. Groups I, II and III within *P. multimicronucleatum* (see Table 1) are indicated. The tree was computed by the bootstrapping procedure (500 bootstraps) and approximate likelihood ratio test method PhyML 3.1/3.0 aLRT. Numbers at nodes represent posterior probabilities higher than 0.4. The scale bar represents the branch length, corresponding to 0.3 substitution per site. The COI gene sequence accession numbers of the strains collected in this study are shown in blue.

**Figure 3.** The phylogenetic tree of the *Paramecium aurelia* species complex inferred from mitochondrial COI gene sequences. *Paramecium multimicronucleatum* was used as an outgroup. The red arrow indicates position of *P. quindecaurelia* n. sp. The red brace shows the pair of sister species *P. biaurelia* and *P. quindecaurelia* n. sp.; the green brace shows the pair of sister species *P. primaurelia* and *P. pentaurelia*; the blue brace shows the pair of sister species *P. tetraurelia* and *P. octaurelia*. The tree was computed by the bootstrapping procedure (500 bootstraps) and approximate likelihood ratio test method PhyML 3.1/3.0 aLRT. Numbers at nodes represent posterior probabilities higher than 0.3. The scale bar represents the branch length, corresponding to 0.5 substitution per site. The COI gene sequence accession numbers of the strains collected in this study are shown in blue.

#### *3.3. New Species of the Paramecium Aurelia Complex, P. Quindecaurelia n. sp.*

The ciliates were collected from a small lake near the highway in Amealco, Querétaro (population A65) and from the lake drain in Los Dinamos, the National Park in the mountains on Mexico City territory (population D88). Several cells from both populations were isolated, and the obtained strains were introduced into laboratory cultures. They can be maintained in a variety of temperatures but demonstrated the best growth rate, three divisions per day, at 21–23 ◦C. The ciliates look like typical representatives of the *P. aurelia* complex (Figure 4). The cells are rather large, and maximum body length of 80 fixed specimens ranged from 104 to 168 μm with a mean length of 125 μm (16.9 μm standard error). Maximum body width of 80 specimens ranged from 19 to 39 μm with a mean width of 28 μm (8.0 μm standard error). The average kineties number was 55. The length of the infundibulum ranged from 15 to 36 μm with an average of 26 μm. All cells had one roundish macronucleus that was longer than wide (Figure 4I), as well as two micronuclei of the vesicular type (Figure 4G,H) typical for *P. aurelia* [1,4]. The average length of the macronucleus was 32 μm, and average width was 12 μm. Two contractile vacuoles were present per cell, and each vacuole had one pore and six or seven canals in different cells (Figure 4E,F). The macronucleus is fragmented into 35–40 pieces during autogamy, and two macronuclear anlagen are formed in the cell (Figure 4J). All morphometric data confirm that these strains are very similar to *P. biaurelia* ([33], Supplementary Table S1).

**Figure 4.** Morphological features of *Paramecium quindecaurelia* n. sp. (**A**) DIC live micrograph of a specimen; OA—oral aperture, CV—contractile vacuole with 6 or 7 canals; MAC—macronucleus. Silver nitrate impregnated cells: (**B**) Ventro-lateral and dorso-lateral cell projections. CP—cytoproct; (**C**) Buccal overture with buccal ciliature; (**D**) Cytoproct region; (**E**) One pore (CVP) per one contractile vacuole is characteristic. (**F**) DIC live micrograph showing the contractile vacuole (CV), the macronucleus (MAC) and two micronuclei (marked by the arrows). Micronuclei have typical vesicular shape shown by DIC (**G**) and on Feulgen stained cell (**H**,**I**). (**J**) Feulgen stained specimen in autogamy: two new macronuclear anlagen (MA) surrounded by numerous fragments of old macronucleus. Scale bars: 8 μm (**A**,**B**,**J**), 4 μm (**C**,**D**,**E**,**F**), 3 μm (**G**,**H**,**I**).

The COI gene sequencing was performed to assign A65 and D88 strains to a certain species of the *P. aurelia* complex. The strains from the same populations had identical sequences (Table 2), while between A65 and D88 strains the sequence was 99.3% similar (5 single nucleotide polymorphisms (SNPs) per 760 bp-long sequence). The best match in GenBank for both strains was KX008305.1, belonging to *P. biaurelia*. This sequence was 94.9% similar to A65 (39 SNPs per 760 bp-long sequence) and 94.5% similar to D88 (42 SNPs per 760 bp-long sequence). In the *P. aurelia* phylogenetic tree inferred from the COI gene sequences, strains A65 and D88 formed a separate branch sister to *P. biaurelia* (Figure 3, red brace), mirroring another pair of closely related species, *P. primaurelia* and *P. pentaurelia* (Figure 3, green brace). The COI gene sequence similarity between and within the latter two species is of the same range (Table 3).

The first cells were entering autogamy only after 20 divisions in daily re-isolations cycle, while synchronous autogamy of the culture was observed after 25–27 vegetative divisions. Two macronuclear anlagen were formed after the sexual process, also typical for the *P. aurelia* species. We never observed selfing (intrastrain conjugation) in A65 and D88 strains, thus, the mating type determination is not stochastic as in some species of the complex [31]. Furthermore, we never obtained conjugation by mixing mature cells of di fferent strains from both populations, so, presumably, all isolated strains were of the same mating type. In closely related *P. biaurelia*, the mating types are inherited maternally [19,29], and it is very likely that cytoplasmic inheritance is characteristic for these strains also.

Since we were unable to obtain conjugation between A65 and D88 strains, we also tested the possibility of conjugation between these strains and tester strains of *P. biaurelia*. Several conjugating couples were observed in crosses between IST and D88 and between Rie ff and D88 strains after incubation of slightly starved reactive cells at + 18 ◦C, while only single couples were observed in mating tests of A65 with Rie ff and IST strains. No conjugation was observed between Ts strain and either A65 or D88. In the control mating tests between *P. biaurelia* strains, conjugation was well pronounced at the same conditions. None of the few exconjugants in crosses utilizing A65 survived.

Exconjugant cells were isolated and F1 clones were established for the crosses IST × D88 and Rie ff × D88. All F1 clones fit well and grew like healthy paramecia. F2 progeny were obtained by autogamy from all F1 clones. Survival rate of F2 clones was 3% (2 of 60) in the IST × D88 cross, and 8% (5 of 60) in the Rie ff × D88 cross. At the same time, survival rate of F2 clones obtained in the IST intrastrain cross was 95%, and in the cross of *P. biaurelia* strains Rie ff × Ts, it was 77%. Moreover, during the following month of maintenance, no F2 clones from the IST × D88 cross survived, and only two F2 clones remained viable from the Rie ff × D88 cross. This rate of survival is very low and could be due to a technical error of the experiments. For example, some cells could have remained vegetative and were occasionally selected from autogamous culture of F1 heterozygous clones or some F1 cells might regenerated old macronuclei in autogamy.


*Diversity* **2020**, *12*,197


**Table 3.** COI gene sequence identity values for sister species *P. pentaurelia* and *P. primaurelia*.
