**1. Introduction**

The genus *Paramecium* Müller [1] (Peniculida, Oligohymenophorea, Ciliophora) includes two species that are known to maintain intracellular algae. *Paramecium chlorelligerum* Kahl [2,3] lives in symbiotic association with green algae belonging to the genus *Meyerella* Fawley & K. Fawley [4] (Chlorellaceae, Trebouxiophyceae, Chlorophyta) and is considered as an endemic, extremely rare species [3,5]. The other *Paramecium* species with algal symbionts is *Paramecium bursaria* Focke [6], the 'green' *Paramecium*, frequently found in freshwater habitats around the world [7]. Evolutionary, this species has been assumed to be the first that diverged within the genus [8,9]. This green ciliate harbours hundreds of symbiotic green algae within its cytoplasm [7,10–14] either belonging to *Chlorella variabilis* Shihira & R. W. Krauss [15] or to *Micractinium conductrix* (K. Brandt) Pröschold & Darienko [16,17] (Chlorellaceae, Trebouxiophyceae, Chlorophyta).

The algae escape lysosomal fusion by incorporation in perialgal vacuoles [18–20]. In this facultative, mutualistic symbiosis, the algae provide photosynthesis products for their host [21–23] and also form a protective layer against UV-radiation damage [24]. Moreover, they profit from protection against the lytic *Paramecium bursaria* chlorovirus [25,26]. In addition, the host increases its symbiont's motility by providing transport to brightly illuminated areas optimal for photosynthesis [27]. The facultative nature of this symbiosis allows the separate cultivation of both organisms, making the association an easy-to-access model system to study symbiotic interactions [28–30]. Albeit under laboratory conditions the symbiosis is not obligate, it is the natural condition in the environment. Both algae have not been retrieved as free-living organisms from nature so far and, therefore, are considered as highly adapted to their endosymbiotic lifestyle [31]. Similarly, aposymbiotic *P. bursaria* cells have been isolated from natural freshwater sources only on rare occasions [32]. As *Chl. variabilis* and *M. conductrix* apparently are mainly restricted to their endosymbiotic lifestyle within *P. bursaria*, these associations have been considered as highly specific. However, recent studies report these microalgae as intracellular symbionts from different host organisms such as *Tetrahymena* [33], *Hydra* [34], *Frontonia vernalis* [35], and sponges [10]. Furthermore, *P. bursaria* can also harbour different organisms either co-occurring with *Chlorella*-like algae, i.e., bacteria [36,37], or in their place, i.e., yeasts [38,39]. Therefore, the symbiotic relationship between *P. bursaria* and its intracellular green algae might not be as exclusive as previously assumed.

An interesting case of double algal infection in *P. bursaria* by *Chlorella* and another green alga was reported by two Japanese groups [31,40]. The intracellularly uncommon, additional alga was characterized as *Choricystis parasitica* (K. Brandt) Pröschold & Darienko [16,17] (Trebouxiophyceae, Chlorophyta). Cells of this alga were observed between the trichocysts near the host's cortex [31,40]. This localization corresponds to that of the natural algal endosymbionts'. The cortex area is devoid of lysosomes, thus an intracellular symbiont is likely protected here against lysosomal fusion [41] and hence against host defence mechanisms, which was interpreted as an advanced symbiotic status [31,40]. Due to its small cell size (1.5–3.0 μm in length, 1.0–1.5 μm in width; [10]), *Chor. parasitica* will be referred to as picoalga.

Characterization of symbiotic systems often requires interdisciplinary expertise to fulfill the standards of the respective scientific disciplines. Molecular characterization of a single or more phylogenetic markers is typically part of current descriptions of new taxa. Sequence analysis allows evolutionary interpretation in absence of fossil records and furthermore facilitates the assignment of new isolates to described taxa. The choice of the molecular markers depends on multiple factors and varies between organism groups and the desired level of phylogenetic resolution. In case of *P. bursaria*, the SSU (small subunit or 18S ribosomal RNA) gene sequence is commonly applied for phylogenetic analyses of members of the genus *Paramecium* [5,42–44]. Some recognized *Paramecium* species are actually sibling or cryptic species complexes, i.e., their members are morphologically indistinguishable but reproductively isolated and thus are called syngens. The most prominent example are the syngens of the *Paramecium aurelia* complex, which was first studied in detail by Sonneborn [45]. Those syngens are nowadays recognized as separate species. Similarly, the existence of multiple syngens was described for *P. bursaria* [46]. Bomford's description of *P. bursaria* has been the most extensive one, unfortunately his strain collection was lost. Only few strains remain and are scattered across different laboratories. A representative collection of *P. bursaria* strains assigned to five syngens, R1 to R5 [47], is maintained at the RC CCM collection (World Data Centre for Microorganisms, RN 1171), Saint Petersburg State University, Saint Petersburg, Russia. Molecular phylogenetic analyses confirmed *P. bursaria* to be a complex of at least five cryptic species [47–50] using different molecular markers (e.g., ITS1-5.8S-ITS2-5'LSU fragments, mitochondrial cytochrome c oxidase subunit I gene, histone H4 gene; [5,48,51]).

Based on molecular phylogenetic analyses of SSU and ITS (internal transcribed spacer) regions of nuclear-encoded rRNA genes, the genus and species concepts within the Chlorellaceae remain provisional due to the lack of bootstrap support in molecular analyses [10,52–55]. The family is divided into two major clades: (i) the well-supported *Parachlorella* clade and (ii) the moderately supported *Chlorella* clade [56,57]. Furthermore, the ITS2 region has been used for species delineation of members

of the Chlorellaceae [10,50,55,58]. The two common endosymbionts of *P. bursaria, Chl. variabilis* and *M. conductrix*, both vary in the ITS2 secondary structure and can be differentiated using their ITS2 barcodes [50].

This study addresses the diversity of intracellular algae in *P. bursaria*. Therefore, ciliate hosts and algal endosymbionts were identified using different molecular markers. In order to analyse the interaction between the three different algal species detected and *P. bursaria*, experimental re- and cross-infections were performed and the fate of the intracellular algae was monitored. Differences in the establishment of stable associations will allow to draw conclusions regarding the compatibility of the different algae as endosymbionts of *P. bursaria* and the specificity of such interactions.

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

#### *2.1. Strains and Cultivation*

All *P. bursaria* strains (Table 1) were cultivated in Dryl's solution (0.1 M Na2HPO4, 0.1 M NaH2PO4, 0.1 M CaCl2, 0.1 M sodium citrate, pH 6.8) at 20 ◦C in a light chamber (EHRET Type KLT/S 4, Ehret, Emmendingen, Germany) with a 12 h/12 h light-dark cycle. The cultures were fed weekly with the bacterium *Raoultella planticola* DMSZ 3069 as food organism (bacterized CM). Therefore, the bacteria were inoculated in 0.25% Cerophyll medium (CM; [59] prepared from wheat grass powder (GSE-Vertrieb), Saarbrücken, Germany) and a stigmasterol (Sigma-Aldrich, Munich, Germany) concentration of 500 μg/L. After two days of incubation at 20 ◦C, the bacterized CM was ready for use.


**Table 1.** *Paramecium bursaria* and algal strains used in this study. Obtained sequences and their accession numbers are indicated.

1: Pond at Rurikojy temple, Yamaguchi, Japan; 2: Culture collection of Dr. K. Eisler, University of Tübingen, Germany; 3: Friedrichsee, Saxony-Anhalt, Germany; 4: Pond 'Pferdetränke', Oldenburg, Germany; 5: Nymphensee, Rangsdorf, Germany; 6: Loch Katrine, Scotland, United Kingdom; 7: Saint Petersburg, pond in the park, Russia; 8: Forest pond, Vyborg, Saint Petersburg region, Russia; 9: Pond, Ardmoore, Oklahoma, USA; \* Zagata et al., 2016 [60]; part.: partial.

#### *2.2. Observation and Size Determination of Intracellular Algae*

The intracellular algae were investigated by detecting the autofluorescence of the chlorophyll by fluorescence microscopy (Axio Imager.M2, Zeiss, Jena, Germany). Therefore, *Paramecium* cells were fixed with Bouin's fixative solution (Sigma-Aldrich) and transferred to a microscope slide. Cells were examined using the Cy5 channel (EX BP 640/30, BS FT 660, EM BP 690/50) and an exposure time

ranging from 3000 to 7000 ms in order to discriminate between the individual intracellular algae. To determine the size of the intracellular micro- and picoalgae, the diameter of 15 cells per strain was measured.

#### *2.3. DNA Extraction and Amplification of Molecular Markers*

All polymerase chain reactions (PCR) were performed with a PeqStar 2× thermocycler (Peqlab, VWR International, Darmstadt, Germany) using TaKaRa reagents and TaKaRa ExTaq polymerase (TaKaRa Bio, Otsu, Japan). Primer sequences (Table 2) and amplification protocols (Supplementary Tables S1 and S2) are provided. The SSU-ITS1-5.8S-ITS2-5'LSU rRNA gene region of *P. bursaria* was amplified using the primer combinations Penic\_F82 and Penic\_R1280 and Penic\_F661 and 28S\_R457. Primers Penic\_F82, Penic\_F661 and 28S\_R457 were used for sequencing.

**Table 2.** Oligonucleotides used for molecular characterization of both *P. bursaria* and symbiotic algae.


For the intracellular algae, the nuclear encoded SSU gene and the ITS region were amplified using a semi-nested PCR approach to obtain sufficient DNA amounts for direct sequencing. For the initial amplification (primer combination: AF and ITS055R) as well as for the two semi-nested amplification reactions (primer combinations: AF and Chlo\_R841 and Chlo\_G800F and ITS055R) the protocol described by Pröschold and colleagues [10] was used. AF, Chlo\_F238, Chlo\_G800F, Chlo\_R841 and ITS055R were used as sequencing primers. For strains Scot and Old-Pf amplification was carried out with primer combination Chlo\_G800F and ITS055R (Supplementary Tables S1 and S2) and both primers for sequencing. Additionally, a diagnostic PCR allowing to discriminate between *Chl. variabilis* and *M. conductrix* was carried out as described elsewhere [50].

To amplify the SSU gene of *Choricystis* from strain Frieds the primer combination Chori\_F238 and Chori\_R841 was used. Those primers were designed to specifically match *Choricystis*-like algae based on preliminary sequence data. Chori\_F238 was used for sequencing.

Purified PCR products (QIAquick PCR Purification Kit, Qiagen, Hilden, Germany) were sequenced at GATC Biotech (Konstanz, Germany). Accession numbers of obtained sequences are provided in Table 1.

#### *2.4. Molecular Characterization of Paramecium bursaria*

In order to confirm our morphological identification of the ciliates as members of *P. bursaria*, we sequenced the SSU rRNA gene sequence and the ITS region of our newly obtained strains. Both markers are commonly applied for characterization at generic and species level [5,42–44,47,50]. The obtained sequences were imported into two datasets: (i) The SSU sequences were included into a dataset of representatives of most *Paramecium* species received from the SILVA SSU Ref NR 99 release

138 database [64]. In addition, nine other members of Peniculidae were included as outgroup resulting in a dataset of 59 sequences with a length of 1493 bp. (ii) The ITS sequences were included into an alignment of 32 *P. bursaria* strains (506 bp) as described in previous studies [47,50]. For syngen assignment, we added sequences of additional strains (Ard10, Bob2, Ek), which were tested for their mating behaviour by RC CCM to increase the number of sequenced strains per syngen. The ITS sequences were aligned according to their secondary structures.

The best fitting evolutionary models for these datasets were determined using the automated model selection tool implemented in PAUP (version 4.0a, built 167) [65]. The best model (SSU: GTR+I+G; ITS: K81uf+I) was chosen following the Akaike Information Criterion (AIC). For the Bayesian analyses, Bayesian Information Criteria (BIC) were chosen (SSU: GTR+I+G; ITS: GTR+I) for calculation in MrBayes (version 3.2.6 x64; [66]). Non-paramectric Maximum Likelihood (ML) analysis was estimated on 1000 pseudoreplicates (PHYML 2.4.5 from the ARB software package; [67]). Bayesian Interference (BI) analysis was carried out running three runs with one cold and three heated Markov Chain Monte Carlo chains for 1,000,000 generations and sampling the first 25% of the generations as burn-in.

#### *2.5. Molecular Characterization of Green Algal Endosymbionts in Paramecium bursaria*

For the molecular characterization of the algal endosymbionts we used sequences covering the SSU-ITS1-5.8S-ITS2 region for discrimination at species level in accordance with previous studies [10,55,56,62,68,69]. Especially the ITS2 region of this nuclear encoded ribosomal operon is often used for species discrimination between members of the Chlorellaceae [53,55,56,70]. The here performed amplification of the ITS2 region covered helices I-III but only partial helix IV. The analysed dataset comprised 50 sequences of representative members of the Chlorellaceae, which were aligned according to the secondary structures predicted by the software mfold [71], following the approach of Pröschold and colleagues [10,55]. This software used the thermodynamic model (minimal energy) for RNA folding. The best fitting evolutionary model for this dataset was predicted and ML and BI analyses were conducted as described above except that BI analysis was run for 5,000,000 generations. Additionally, the Neighbour Joining method (implemented in PAUP) was used. The analysis was run with 1000 bootstrap replicates and a 50% majority rule consensus tree was calculated.

Furthermore, we compared the algal endosymbionts' secondary structures of ITS2 helices I-III to that of the type strains of *Chl. variabilis* and *M. conductrix*. In case of the symbionts of strains Bob2, Ek, and RanNy these algae were identified by means of diagnostic PCR (Supplementary Figure S1).

To identify the intracellular picoalgae of *P. bursaria* observed in strain Frieds, we used the obtained partial SSU gene sequence and compared it with 41 representative sequences for the *Trebouxia* lineage (Trebouxiophyceae; following Darienko and colleagues [62]) with the *Prasiola* clade (four species) and the genus *Neocystis* (two species) as outgroup. The alignment was obtained as described for *Paramecium* SSU and comprised 948 positions. Model selection (TIM+I+G) for this dataset was performed as detailed above. Since this model is not implemented in MrBayes, it was substituted with GTR +I+G. ML and BI analyses were performed as described above for *P. bursaria*. The settings of all used evolutionary models are provided in Supplementary Table S3.

#### *2.6. Establishment of Aposymbiotic Paramecium bursaria*

Algal-free *P. bursaria* were generated by treatment with cycloheximid (Roth, Karlsruhe, Germany). Therefore, symbiotic *P. bursaria* cultures not fed for at least seven days were incubated in 10 μg/mL cycloheximid solution and kept under constant light conditions for approximately three days, then the cells were washed with Dryl's solution by filtering over a 10 μm membrane and fed with bacterized CM. This filtering step was repeated three to five times over the course of two weeks to assure complete elimination of the algae. If necessary, a second consecutive cycloheximid treatment was performed seven to ten days after the initial exposure. Microscopic inspection via fluorescence microscopy confirmed the elimination of intracellular algae. Aposymbiotic status was confirmed if after 10 seconds of exposure time no autofluorescence signal of the algae's chloroplasts was detected.

#### *2.7. Pulse-Chase Infection Experiments*

To investigate the symbiosis specificity between *P. bursaria* and three different symbiotic algal species, we exposed aposymbiotic cells (strains JPN, RanNy, Scot; affiliated to different syngens, see below) to isolated *Chl. variabilis*, *M. conductrix*, or *Chor. parasitica* (Figure 1). Aposymbiotic *P. bursaria* generated by treatment with cycloheximid were exposed (recipient) to isolated algae obtained from symbiotic *P. bursaria* cultures (donor; Supplementary Table S4). In re-infection assays, the supplied algae belonged to the same algal species as the recipient strain originally harboured, while in cross-infection experiments a different species was provided. We followed the fate of the algae after their uptake and differentiated between digestion, expulsion, and endosymbiotic maintenance by fluorescence microscopy using the above listed parameters. Each experiment comprised three replicates. In total, 20 aposymbiotic *Paramecium* cells were exposed to a suspension of the respective alga for 5 min with a ratio of 3 × 10<sup>3</sup> algae per ciliate cell. Then the paramecia were washed and transferred to 200 μL Dryl's solution inoculated with 20 μL of bacterized CM. 20 μL bacterized CM were added every second day for seven to ten days. Afterwards, the amount of supplied food was adjusted with respect to cell densities. A regular weekly feeding schedule was implemented after two to three weeks post infection (p.i.). The establishment of the symbiosis was monitored using the fluorescence signal of the algae's chlorophyll as described above. Instead of Bouin's solution, the *Paramecium* cells were fixed with 2% paraformaldehyde (PFA), which resulted in lower background signal and thus an improved signal to noise ratio. To monitor the process of re-establishment of symbiosis, *P. bursaria* cells were microscopically screened after two, four, seven, ten, and 14 days. Successful infection was confirmed when individually enclosed algae were localized beneath the host's cell cortex.

**Figure 1.** Pulse-chase infection experiments to assess the symbiosis specificity between *Paramecium bursaria* and its symbiotic algae. Symbiotic cultures of *P. bursaria* (donor) were mechanically lysed to obtain algal suspensions of *Chlorella variabilis* and *Micractinium conductrix*, which were used in re- and cross-infection experiments, respectively. Detailed information about strains used in the experiments (algae as well as host) is provided in Supplementary Table S4. Algal culture of *Choricystis parasitica* was obtained via mechanical lysis of *P. bursaria* strain Frieds for subsequent extracellular cultivation. This algal culture was then used in infection experiments with aposymbiotic *P. bursaria* cells (recipient). After 5 min of exposure (pulse), the paramecia were washed to remove extracellular algae and placed into fresh medium inoculated with food bacteria (chase). Status of symbiosis was monitored regularly via autofluorescence microscopy.
