Endosymbiotic Green Algae in Paramecium bursaria : A New Isolation Method and a Simple Diagnostic PCR Approach for the Identification

: Paramecium bursaria is a single ‐ celled model organism for studying endosymbiosis among ciliates and green algae. Most strains of P. bursaria bear either Chlorella variabilis or Micractinium conductrix as endosymbionts. Both algal genera are unicellular green algae characterized by cup ‐ shaped chloroplasts containing a single pyrenoid and reproduction by autospores. Due to their size and only few morphological characteristics, these green algae are very difficult to discriminate by microscopy only. Their cultivation is laborious and often unsuccessful, but we developed a three ‐ step isolation method, which provided axenic cultures of endosymbionts. In addition to the time ‐ consuming isolation, we developed a simple diagnostic PCR identification method using specific primers for C. variabilis and M. conductrix that provided reliable results. One advantage of this approach was that the algae do not have to be isolated from their host. For a comparative study, we investigated 19 strains of P. bursaria from all over the world (new isolates and available laboratory strains) belonging to the five known syngens (R1–R5). Six European ciliate strains belonging to syngens R1 and R2 bore M. conductrix as endosymbiont whereas C. variabilis was discovered in syngens R1–R5 having worldwide origins. Our results reveal the first evidence of C. variabilis as endosymbiont in P. bursaria in Europe. combinations (CvarF/ITS055R and MconF/ITS055R) were tested on 18 investigated Paramecium bursaria strains. All PCR amplifications were done on a thermocycler with the following program: 5 min initial denaturation at 95 °C, followed by 30 cycles (1 min at 95 °C, 2 min at 55 °C and 3 min at 68 °C), and final synthesis for 10 min at 68 °C.


Introduction
Symbiosis of green algae with protists and invertebrates has been studied for more than 100 years ( [1] and references therein). Living in a mutualistic relationship has advantages for both ciliate and algae. The ciliate can survive starvation under nutrient limitation and prevent against damages induced by solar irradiation. The algae are protected against infection by chloroviruses, which lyse the endosymbionts outside of their hosts [2][3][4]. Such mixotrophic ciliates are widely distributed in many freshwater habitats and belong to different phylogenetic lineages [5][6][7]. The endosymbiotic green algae are commonly assigned to Chlorellalike organisms or simply named as zoochlorellae [1,8,9]. So far, Paramecium bursaria is the most studied model ciliate in endosymbiosis research because of its easiness of cultivating and cloning and it can be identified rather easily from its morphology [10,11]. P. bursaria contains up to 500 green algal endosymbionts, which are located in perialgal vacuoles [12]. In contrast, identification of the endosymbionts of P. bursaria based solely on morphology is almost impossible (Figure 1), and the isolation and cultivation of these zoochlorellae is quite difficult and time-consuming, which was already recognized by Pringsheim [13] and Loefer [14]. The few algal strains available in public culture collections were characterized by Pröschold et al. [1] using an integrative approach based on their morphology and phylogenetic analyses. Pröschold et al. [1] identified four different green algal species isolated from different Paramecium bursaria strains: Chlorella variabilis, C. vulgaris, Micractinium conductrix, and an unidentified species of Scenedesmus. However, most strains of Paramecium bursaria bore either Chlorella variabilis or Micractinium conductrix as endosymbionts and they were assigned to an American (or Southern) and a European (or Northern) group by Gaponova et al. [15], Hoshina et al. [16][17][18], and Hoshina and Imamura [19,20], respectively. A clear identification of the endosymbionts is of special interests since the discovery of highly specific chloroviruses, which infected these green algae when they were released from their hosts ( [21] and references therein). The aim of this study was to develop (i) a protocol for the isolation of green algal endosymbionts and (ii) a quick and precise identification method without previously isolating them from their hosts. The isolation method focuses on the separation of green algal endosymbionts from other free-living organisms and the special nutrient requirement for the growth of these zoochlorellae. The easy diagnostic PCR approach uses species-specific primers, which focused on the internal transcribed spacer region 2 (ITS-2) of the nuclear ribosomal operon, often used for species delimitation among the Chlorellaceae before ( [1,22] and references therein). Moreover, the ITS-2 region was selected on the basis of the exact delineation at the species level using the compensatory base change (CBC) concept introduced by Coleman [23]. This concept uses the CBCs in the conserved region of the ITS-2 secondary structure. Coleman [23][24][25] found that if two specimens differed in at least one CBC in the conserved region of ITS-2 (helices II and III), both were not able to mate and therefore represented two different biological species. Pröschold et al. [1] demonstrated that the endosymbiotic species had several CBCs in their ITS-2 secondary structures and can be clearly distinguished. In addition, Pröschold et al. [1] raised the question if the occurrence of a green algal taxon was correlated with the geographical origin of its host or/and with the affiliation to a certain ciliate syngen (= biological species). Bomford [26] investigated the mating behavior among several isolates of P. bursaria and discovered six syngens by conjugation experiments. Greczek-Stachura et al. [9] confirmed the syngen pattern by sequencing of the nuclear ITS rDNA, the mitochondrial cytochrome c oxidase subunit I (COI), and the histone H4 gene. However, they focused only on the ciliate phylogeny and therefore, which green algal endosymbionts were present in the different syngens remains unknown.

Cultivation and Molecular Characterization of Paramecium bursaria
The investigated strains were collected from around the world and cultivated in modified Bold Basal Medium (3N-BBM+V; medium 26a in [27]) with the addition of 30 mL of soil extract per liter final medium (called S/BBM). The soil extract was prepared as described in Schlösser [28]. Origin and details about the investigated Paramecium bursaria strains are listed in Table 1. All cultures were maintained at 15-21 °C under a light:dark cycle of 12:12 h (photon flux rate up 50 μmol m −2 s −1 ). Genomic DNA of the green algae was extracted using the DNeasy Plant Mini Kit (Qiagen GmbH, Hilden, Germany). The SSU and ITS rDNA were amplified using the Taq PCR Mastermix Kit (Qiagen GmbH, Hilden, Germany) with the primers EAF3 and ITS055R [29]. The SSU and ITS rDNA sequences of the Paramecium bursaria strains were aligned according to the secondary structures, resulting in a dataset of 19 sequences (2197 bp). GenBank accession numbers of all newly deposited sequences were given in the phylogenetic tree and Table 1. For the phylogenetic analyses, we calculated the log-likelihood values of 56 models using the automated selection tool implemented in PAUP version 4.0b167 [30] to test which evolutionary model fitted best for the dataset. The best model according to the Akaike criterion by PAUP was chosen for the analyses. The settings of the best model are given in the figure legend. The following methods were used for the phylogenetic analyses: Distance, maximum parsimony, and maximum likelihood, all included in PAUP version 4.0b167 [30].
The secondary structures were folded using the software mfold [31], which uses the thermodynamic model (minimal energy) for RNA folding. The visualization of the structures was done using the program PseudoViewer 3 [32].

Isolation of the Green Algal Endosymbionts
The endosymbiotic green algae are often not easy to distinguish from free-living green algae in the surrounding media. Therefore, a special isolation method needed to be developed to avoid that free-living algae, or contaminations grow during the isolation process. With the following procedure as demonstrated in the work scheme (Figure 2), we obtained several axenic clonal strains in culture from different mixotrophic ciliates, such as Paramecium bursaria. For the isolation of their green algal endosymbionts, single ciliate cells were washed several times and transferred into fresh S/BBM medium (step 1). After starvation and digestion of any food, after approximately 24 h, cells were washed again (step 2) and the ciliates were transferred onto agar plates containing basal medium with beef extract (ESFl; medium 1a according to Schlösser, [28]). Before placement of the ciliates onto agar plates, 50 μL of an antibiotic mix (mixture of 1% penicillin G, 0.25% streptomycin, and 0.25% chloramphenicol) were added to prevent bacterial growth. The agar plates were kept under the same conditions as described. After growth (~40 days), the algal colonies were transferred onto agar slopes (1.5%) containing ESFl medium and were kept under the described culture conditions (step 3).

Diagnostic PCR Amplification
The isolation of the green algal endosymbionts is time-consuming and not always successful, especially if Chlorella variabilis is the endosymbiont. To investigate which green algal endosymbiont was present in a set of P. bursaria strains without isolating them, we developed a diagnostic PCR method using species-specific primers. Pröschold et al. [1] demonstrated that the SSU and ITS rDNA of Chlorella variabilis and Micractinium conductrix contain three introns and one intron, respectively, which makes an easy PCR amplification not possible. Therefore, we amplified the ITS-2 sequences for both C. variabilis and M. conductrix, a sequence which is diagnostic. For designing species primers, the internal transcribed spacer 1 (ITS-1) rDNA sequences of all representatives (58 species) belonging to the Chlorellaceae were compared to find characteristic positions. The following diagnostic primers for both species were designed based on the SSU and ITS rDNA sequences (C.  These primers were tested in combination with the reverse primer ITS055R [29], both on the isolated axenic algal cultures and their hosts. For identification of the endosymbionts, both primer combinations (CvarF/ITS055R and MconF/ITS055R) were tested on 18 investigated Paramecium bursaria strains. All PCR amplifications were done on a thermocycler with the following program: 5 min initial denaturation at 95 °C, followed by 30 cycles (1 min at 95 °C, 2 min at 55 °C and 3 min at 68 °C), and final synthesis for 10 min at 68 °C.

Results
The isolation of the green algal endosymbionts using the method described above (Figure 2) was successful for some of the Paramecium bursaria strains (SAG 27.96, CIL-16, CIL-19, and CIL-20). However, it was not always successful and was too time-consuming. Therefore, we used the diagnostic PCR approach for the identification of the endosymbionts. The primer combinations CvarF/ITS055R and MconF/ITS055R were highly species-specific as shown in Figure 4   We tested this approach using these primer combinations on 18 Paramecium bursaria strains from different geographical origin and various syngens (see Figure 5 for the origin of the strains). Additionally, we sequenced the SSU and ITS rDNA sequences of the ciliate strains (the accession numbers are given in Figure 5). Despite their low genetic variability among the isolates (only 1.7%), the phylogenetic analyses ( Figure 5) of the SSU and ITS rDNA sequences confirmed their subdivision of P. bursaria strains into five lineages representing the known syngens R1-R5 [9], which was well or moderately supported in all bootstrap analyses.
Among the different syngens, both endosymbionts were discovered using the diagnostic PCR approach. All ciliate syngens but not all strains harbored C. variabilis, whereas M. conductrix was only present in syngens R1 and R2, which originated from Europe ( Figure 5). The ITS-2 sequences of these Chlorella variabilis (12 obtained with the primers CvarF/ITS055R) and Micractinium conductrix (6 obtained with the primers MconF/ITS055R) were identical to those of the reference strains NC64A = CCAP 211/84 (FN298923; [1]) and SAG 241.80 (FM205851; [1]), respectively. The phylogenetic tree shown was inferred using the maximum likelihood method based on the datasets (2197 aligned positions of 19 taxa) using computer program PAUP 4.0a167. For the analyses, the best model was calculated by PAUP 4.0a167. The setting of the best model was given as follows: GTR + I (base frequencies: A 0.2994, C 0.1832, G 0.2278, T 0.2896; rate matrix A-C 2.4955, A-G 4.2183, A-U 4.9756, C-G 0.7513, C-U 9.6155, G-U 1.0000) with the proportion of invariable sites (I = 0.9303). The branches in bold are highly supported in all bootstrap analyses (bootstrap values >50% calculated with PAUP using the maximum likelihood, neighborjoining, and maximum parsimony). The clades are named after the syngens (color-coded) proposed by Greczek-Stachura et al. [9]. The accession numbers are given after the strain numbers. The endosymbiotic green algae identified using the diagnostic PCR are highlighted (Mcon-Micractinium conductrix and Cvar-Chlorella variabilis) after the origin of the Paramecium bursaria strains. The reference strain of each syngen is marked with an asterisk. The green algal endosymbionts isolated and available for further investigations are highlighted in blue.

Discussion
The green algal endosymbionts are often designated as Chlorella-like organisms or as simple zoochlorellae because of the difficulties in identifying them based solely on the morphology (see Figure 1). Especially, identification at the species level is almost impossible using light microscopic observations. The isolation of clonal cultures from the ciliate hosts is quite difficult and time-consuming, but the described method above resulted in some axenic strains of green algal endosymbionts. The main problem for isolating endosymbionts from their hosts is the slow growth and the requirement of additional nutrients, which is included in supplements, such as soil and beef extract. Especially, Chlorella variabilis requires organic compounds and vitamins [33,34]. Several attempts have been undertaken to isolate green algal endosymbionts from their hosts. Loefer [14] was the first to obtain the endosymbiont in culture. He isolated this alga by taking green algae from the sediment of an axenic P. bursaria culture and spreading it on agar plates containing tap water with unknown organic compounds. Since then, several methods for the isolation of endosymbionts have been described [35][36][37][38]. Similar to our approach, washing of the ciliate and the usage of antibiotics and transfer onto agar plates were used in different variants. However, the crucial points of our approach are the starvation of the ciliate for 24 h before rupture on agar plates and the microscopical check during all steps, providing some security that the isolated algae are the endosymbionts of Paramecium bursaria. Especially, the last point is of great importance. For example, Hoshina and Imamura [20] described that the strain CCAP 1660/13 of P. bursaria had an additional endosymbiont (Coccomyxa sp.); however, Pröschold et al. [1] revealed that this alga was not an endosymbiont and represented only a free-living alga co-occurring in the culture of this P. bursaria strain.
Another critical point is the choice of culture media for the endosymbionts. As highlighted, most of them need organic compounds for growth. Therefore, it is mostly likely that the three protocols provided in Achilles-Day and Day [39] resulted in the cultivation of free-living green algae, which are co-cultivated with the hosts. As described in Achilles-Day and Day [39], all green algal endosymbionts grew on media without organic nutrients. In contrast, the three steps of our method (Figure 2) rely on the microscopical control at each step as well as the elimination of contaminants and free-living algae growing outside in the medium. This method can result in axenic clonal cultures of green algal endosymbionts when they are needed for further investigations.
If it is only required to know which endosymbiont species is in a strain of Paramecium bursaria, the presented diagnostic PCR approach revealed an easy and fast method for species identification. Diagnostic PCR approaches have been successfully established in several approaches, such as for the identification of harmful algae ( [40] and references therein). In P. bursaria, Tanaka et al. [41] used a PCR-based approach to demonstrate the success of the elimination of green algal endosymbionts from their hosts. This was based on the small subunit of the rubisco gene (rbcS) gene and did not focus on the identification of the endosymbionts. Here, we used the ITS-2 for our diagnostic PCR approach because this gene has been used for species delimitation as described above, and for all described species belonging to the Chlorellaceae, ITS rDNA sequences are available in GenBank, which is the only reliable dataset for species identification within this group until now ( [22] and references therein). With species-specific primers ( Figure 3) in PCR amplifications, Chlorella variabilis and Micractinium conductrix could be exactly identified, which were confirmed by ITS-2 sequencing (Figures 4 and 5). Both endosymbionts were differently distributed among the five syngens of P. bursaria. Whereas the syngens R3-R5 exclusively had C. variabilis as an endosymbiont, both endosymbionts could be discovered in strains belonging to syngens R1 and R2, which originated exclusively from Europe. Gaponova et al. [15] also found M. conductrix in P. bursaria isolates collected in north Karelia (Russia). It appears that this green algal endosymbiont occurred only in Europe, whereas C. variabilis was distributed worldwide. Consequently, the subdivision into the ʺAmericanʺ and ʺEuropeanʺ endosymbionts groups, i.e., C. variabilis and M. conductrix, respectively, as proposed by Hoshina and Imamura [20] and the references therein needs to be revised as C. variabilis was also found in P. bursaria isolates originating from Europe: Ciliate strains PB-19 (Poland, syngen R1), CIL-46 (Germany, syngen R2), CIL-19, and CIL-20 (both from Austria, syngen R3). In contrast to our findings here, Summerer et al. [42,43] described that ciliate isolates from the two locations about 50 km apart (called PbPIB and PbW) bore Chlorella sp. as endosymbionts based on ITS-1 rDNA sequences. Unfortunately, neither the cultures nor the DNA of these endosymbionts are available anymore for comparative investigations. However, in our investigations, both Austrian P. bursaria strains and both algal strains revealed C. variabilis, which are the first European isolates. Considering the findings of Jeanniard et al. [4], who demonstrated that specific chloroviruses infected these algal species were widely distributed, indicating that hosts containing C. variabilis and M. conductrix also occurred in these freshwater habitats. As demonstrated above, our diagnostic PCR approach provided a quick and precise identification of the endosymbiotic green algae occurring in P. bursaria. This promising method will discover new endosymbionts not only in the model ciliate P. bursaria but also in other ciliate and invertebrate hosts and finally elucidate the biogeographic patterns of endosymbiotic green algal species.
The characteristics of both endosymbionts found in Paramecium bursaria are summarized in Pröschold et al. [1]. Both species differed in the ITS-2 secondary structures and their ITS-2 barcode ( Figure 6). Despite the variations in the helices I-IV, both species were differentiated by one compensatory base change (CBC) and one hemi-CBC (one-sided base change) in the conserved region (ITS-2 barcode). The genomes of both species were sequenced [44,45]. The genome of the strain CCAP 211/84 (NC64A) Chlorella variabilis (ITS-2 barcode: CVAR in Figure 6) has a size of 46.2 Mb and contains 12 chromosomes [44]. Micractinium conductrix, strain SAG 241.80 (ITS-2 barcode: MCON in Figure 6), has a larger genome (60.8 Mb with more than 13 chromosomes; [45]). The chloroplast and mitochondrial genomes of both species are similar in size (125 vs. 129 Kb and 78 vs. 75 Kb; respectively; [45]). Fan et al. [46] questioned the separation of both species into two different genera based on a comparison of the chloroplast and mitochondrial genomes. However, only a few taxa of Chlorella and Micractinium were included in this study. Before generic revision can be taken into account, more species of both genera need to be investigated using an integrative approach.

Conclusions
The endosymbiotic green algae of Paramecium bursaria can be clearly identified at the species level using the diagnostic PCR approach. This approach is also applicable for other mixotrophic ciliates if species-specific primers were designed. For the design of these primers, it is necessary to know which green algal endosymbionts occur in the green algal-ciliate association. This can be provided by isolation of the endosymbionts using the three-step method described above. Axenic cultures of isolated endosymbionts allow further genomic studies, such as the sequencing of whole genomes and plastomes.

Conflicts of Interest:
The authors declare no conflict of interest.