“Candidatus Trichorickettsia mobilis”, a Rickettsiales bacterium, can be transiently transferred from the unicellular eukaryote Paramecium to the planarian Dugesia japonica

Ciliate host isolation, culturing, and identification

Paramecium multimicronucleatum monoclonal strain US_Bl 16I1 was established from a cell isolated from a freshwater sample collected from the Yellowwood Lake (39°11′29,0″N, 86°20′31,4″W), IN, USA and cultivated in San Benedetto mineral water (San Benedetto S. p. A., Italy).

The strain was maintained in the laboratory in an incubator at a temperature of 19 ± 1 °C and on a 12:12 h light/dark cycle (light source: NATURAL L36W/76 and FLORA L36W/77 neon tubes, OSRAM, Berlin, Germany). Instead of using bacteria as common food source for paramecia (Krenek, Berendonk & Petzoldt, 2011), cells were fed monoclonal cultures of flagellated green algae, i.e., Chlorogonium sp. (freshwater) or, alternatively, Dunaliella tertiolecta (brackish water, 1‰ of salinity) to minimize bacterial load in cell cultures. They were fed two to three times per week to obtain a mass culture (1.5 L) (Castelli et al., 2015) suitable for transfection experiments. Living ciliates were observed with a Leitz (Weitzlar, Germany) microscope equipped with Differential Interference Contrast (DIC) at a magnification of 100–1250× for general morphology and swimming behavior according to Modeo et al. (2013a) and Nitla et al., (2019). DIC observation of ciliates was also aimed at checking for macronuclear endosymbiont motility. Feulgen-stained ciliates were observed to obtain information on the nuclear apparatus. Microscope images were captured with a digital camera (Canon PowerShot S45). Cell measurements were performed using ImageJ (https://imagej.nih.gov/ij/). Morphological identification was conducted according to literature data (Fokin, 2010/2011).

Characterization of ciliate and endosymbiont

Characterization of the ciliate host and its bacterial endosymbiont was performed through the “full cycle rRNA approach”  (Amann, Ludwig & Schleifer, 1995), i.e., through the nuclear 18S rRNA gene (generally considered the preferred marker to study molecular taxonomy and phylogeny of eukaryotic organisms—just as an example see Illa et al., 2019), ITS region and partial 28S gene (ciliate), and 16S rRNA gene (endosymbiont) sequencing, combined with Fluorescence In Situ Hybridization (FISH) experiments. Additionally, mitochondrial cytochrome c oxidase subunit I (COI) gene was sequenced as well, to display haplogroup variation in P. multimicronucleatum (Barth et al., 2006).

Planarian culturing

Planarians used in this work belonged to the species Dugesia japonica, asexual strain GI (Rossi et al., 2018). Animals were kept in a solution of: CaCl₂ 2.5 mM, MgSO₄ 0.4 mM, NaHCO₃ 0.8 mM, and KCl 77µM, hereafter referred to as culturing water, at 18 °C in dim light conditions and fed with chicken liver (purchased from local food stores) prepared according to King & Newmark (2013), once a week. Non-regenerating specimens, within 5–8 mm of length, were used for all experimental procedures, after being starved for about two weeks.

Transmission electron microscopy (TEM)

TEM preparations were obtained both for P. multimicronucleatum cells and planarians to study the ultrastructure of the ciliate-associated endosymbiont “Ca. Trichorickettsia mobilis” and to verify the success of transfection experiments (i.e., to check the animals for the presence of transferred ciliate endosymbionts in their tissues with particular attention to bacteria showing possible signs of division) respectively.

TEM preparations of P. multimicronucleatum cells were obtained according to Modeo et al. (2013b). Briefly, cells were fixed in a 1:1 mixture of 2.5% (v/v) glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4), and 2% (w/v) OsO₄ in distilled water, ethanol and acetone dehydrated, and embedded in Epon-Araldite mixture. Ultrathin sections were stained with 4% (w/v) uranyl acetate followed by 0.2% (w/v) lead citrate.

TEM preparations of planarian specimens were obtained as previously described (Rossi et al., 2014; Cassella et al., 2017), with minor modifications. Planarians were fixed with 3% glutaraldehyde in 0.1 M cacodylate buffer, and post-fixed with 2% osmium tetroxide for 2 h. After ethanol dehydration, samples were embedded in Epon-Araldite mixture. Ultrathin sections were stained with uranyl acetate and lead citrate and observed with a Jeol 100 SX Transmission Electron Microscope.

Transfection experiments

Two transfection experiments with the same protocol were sequentially carried out over a period of four months. Each experiment was conducted by treating a fixed number of planarians with ciliate-enriched food, i.e., liver paste mixed with P. multimicronucleatum homogenate; from now on these animals will be referred to as “treated planarians”. As for experimental control, planarians fed plain liver paste were used; these animals will be referred to as “control planarians”. Each transfection experiment was performed according to the following protocol:

1. 96 planarians were selected by eye from culture (see above), washed in fresh culturing water and collected in a Petri dish with 50 µg/ml gentamicin sulfate (Sigma, Saint Luis, MO, USA) dissolved in their culturing water (see above). This preliminary antibiotic treatment was performed to minimize potential endogenous bacteria contaminants and was not harmful to the animals (Roberts-Galbraith, Brubacher & Newmark, 2016). Planarians were left in antibiotic treatment for 24 h under regular culturing conditions concerning temperature and light (see above) and kept under visual control during that period to verify their viability during the antibiotic treatment and by the time of transfer. Then, planarians were washed six times in their fresh culturing water, and split into two equal groups of 48 individuals each and accommodated in two different Petri dishes for the transfection procedure by means of feeding.

2. In order to get a bacterial load consistent with previous experiments (Abnave et al., 2014), 1.5 L of P. multimicronucleatum mass culture (cell concentration: ∼4 × 10⁵ cell/L) were selected. Indeed, considering that FISH experiments indicated roughly 100 endosymbionts/cell (see Results section), in total ciliates were estimated to contain ∼6 × 10⁷ symbiont cells. P. multimicronucleatum cells were filtered with a nylon filter (pore size: 100 µm), washed twice in sterile San Benedetto mineral water to minimize potential bacterial contaminants, concentrated and harvested by means of centrifugation (400× g per 10 min), so to reduce the medium volume to 2 ml. Then, cells were mechanically squashed and homogenized by syringing (syringe needle: 22GA, 0.70 mm in diameter) for 20 min at room temperature (adapted from Vannini et al., 2017). Cell homogenate was checked under the stereomicroscope so to exclude the presence of intact Paramecium cells, centrifuged (10,000× g per 10 min), and the resulting pellet was resuspended in 50 µl of planarian food (homogenized liver paste) by direct resuspension.

3. A group of 48 planarians was fed P. multimicronucleatum-enriched liver paste (treated planarians), while the other group was in parallel fed plain liver paste (control planarians). For feeding, food was spread over the bottom of the Petri dish. Animals were allowed to feed freely for a period of 2 h under regular culturing conditions (see above). Attention was paid to planarian survival and feeding behavior during this period. Two washing steps were then carried out removing the medium and adding fresh planarian culturing water. Finally, the two groups of animals were left in their fresh culturing water and in regular cultivation conditions in the two Petri dishes until the collection of specimens for the next TEM and PCR analyses at the three timepoints (see below) of the experiments. Animals were kept under visual control throughout the experiment to regularly check their viability.

4. At day 1, 3, and 7 after feeding (experimental timepoints), from each of the two Petri dishes containing treated and control planarians a group consisting of 16 animals were sampled: 4 animals were fixed and processed for TEM, and 12 were immediately frozen and stored at −80 °C, and dedicated to DNA extraction (4 animals per each sample).

PCR verification of transfer success

Treated and control planarians were processed for genomic DNA extraction, purification, and PCR amplification with endosymbiont-specific primers. For each experimental condition (treated and control planarians) and each experimental timepoint (1, 3, and 7 days after feeding), genomic DNA was extracted from frozen samples by using the Wizard Genomics DNA purification kit (Promega, Madison, WI, USA). One microliter of purified DNA was analysed by gel electrophoresis to check for integrity. DNA was quantified using a Nanodrop spectrophotometer. For each experimental timepoint of both experimental conditions (i.e., treated and control planarians), similar amounts of genomic DNA were used for amplification using the ampli-Taq-gold DNA polymerase (Applied Biosystems, Foster City, CA, USA) and for a nested PCR assay. The first primer pair used, specific for “Ca. Trichorickettsia mobilis”, was RickFla_ F69 5′-GTTAACTTAGGGCTTGCTC-3′and Rick_R1455 5′-CCGTGGTTGGCTGCCT-3′ (Vannini et al., 2014); PCR conditions were: 3 min at 94 °C; 5 cycles consisting of 30 s at 94 °C, 30 s at 58 °C, 2 min at 72 °C each; 10 cycles consisting of 30 s at 94 °C, 30 s at 54 °C, 2 min at 72 °C each; 25 cycles consisting of 30 s at 94 °C, 30 s at 50 °C, 2 min at 72 °C each; ending with 7 min at 72 °C. The second, nested primer pair used was RickFla_F87 5′-CTCTAGGTTAATCAGTAGCAA-3′and Rick_R1270 5′-TTTTAGGGATTTGCTCCACG-3′ (Vannini et al., 2014). For nested PCR assay, one microliter of PCR product of the first PCR assay was used as a template; PCR conditions were as above.

For each experimental condition and each timepoint of the two transfection experiments, the DNA amplification was performed in duplicate. Samples were considered positive if a single band of the expected size was recorded after nested amplification. Sequencing of amplicons was carried out to confirm the presence of “Ca. Trichorickettsia mobilis” using the primer RickFla_F87 (see above) and Sanger sequencing (BMR Genomics, Padova, Italy).

Samples obtained from planarians fed with plain liver paste were used as control. As positive control, genomic DNA purified from P. multimicronucleatum monoclonal strain US_Bl 16I1 was processed.

Host morphological and molecular identification

Ciliate strain US_Bl 16I1 (Figs. 1A, 1B; Figs. S1 and S2 in Article S1) was confirmed in morphological inspections as Paramecium multimicronucleatum Powers & Mitchell, 1910 considering features of key-characters such as cell size, number and features of micronuclei (mi), and number and features of contractile vacuoles (Figs. S1 and S2 in Article S1), as described in previous literature (Wichterman, 1986; Allen, 1988; Fokin, 1997; Fokin & Chivilev, 2000; Fokin, 2010/2011). The molecular analysis of the combined (partial) 18S rRNA-ITS1-5S rRNA gene-ITS2-(partial) 28S rRNA gene sequence (2,792 bp, GenBank accession number: MK595741) confirmed the species assignation by morphological identification with a 100% sequence identity with the sequences of other P. multimicronucleatum strains already present in GenBank presenting either 18S rRNA gene portion only (AB252006 and AF255361), or the ITS portion (AY833383, KF287719, JF741240 and JF741241). COI gene sequence identity of strain US_Bl 16I1 (760 bp, accession number: MK806287) is highest with another P. multimicronucleatum haplotype (FJ905144.1; 96.3%).

Endosymbiont identification and features

The 16S rRNA gene amplification was performed to identify endosymbionts associated with P. multimicronuclatum US_Bl 16I1 strain (sequence length 1,563 bp, accession number: MK598854), which allowed to assign it to “Ca. Trichorickettsia mobilis”. Specifically, among the three subspecies previously described, it was more similar to “Ca. Trichorickettsia mobilis subsp. hyperinfectiva”, an endosymbiont described in the cytoplasm of Paramecium calkinsi (99.8% identity of the novel sequence with type strain MF039744.1) (Sabaneyeva et al., 2018). FISH experiments (Figs. 1C–1E) performed by using bacterial universal probe EUB338 (Fig. 1C) and the species-specific probe Trichorick_142 (Fig. 1D) confirmed this result and indicated that, contrarily to the P. calkinsi symbiont, the bacterium was localized in the ciliate macronucleus (roughly with a presence of about 100 endosymbionts per cell). Additionally, the full overlapping of bacterial universal probe EUB338 and “Ca. Trichorickettsia”-specific probe signals indicated that this symbiont constituted the total set of intracellular bacteria in host P. multimicronucleatum US_Bl 16I1 cells (Fig. 1E, merge).

In TEM-processed ciliate cells, the endosymbionts were confirmed to be hosted in the macronucleus; they showed a two-membrane cell wall typical of Gram-negative bacteria and appeared rod-shaped, with rounded to narrower ends (Fig. 2). They were surrounded by a clear halo and not encircled by any symbiosomal vacuole, i.e., they appeared in direct contact with host nuclear material. Their size was ∼1.2–2.1 × 0.5–0.6 µm, and they were scattered throughout the macronucleus of the ciliate (Figs. 2A–2D). Sometimes in their cytoplasm several electron-lucid “holes” (diameter: ∼0.2 µm), were observed (Fig. 2A). Bacterial cytoplasm was electrondense; no other structures were visible except for abundant ribosomes (Fig. 2B). Endosymbionts displayed thin (diameter: ∼9 nm) and short flagella distributed all around the cell (Figs. 2A, 2C–2D) which sometimes formed a putative tail emerging from a cell end (Fig. 2D). The presence of longer flagella was evident after negative staining procedure (Fig. 3): besides some short flagella occurring all around the cell, at least a few longer and thicker flagella (∼2000 × 12 nm) arose from one of cell ends.

Some active bacterial motility, likely obtained by means of their flagella, was recorded inside the macronucleus of intact P. multimicronucleatum cells during observation under DIC microscope at 1,000×. Endosymbionts were seen to move through the chromatin bodies covering short distances; additionally, after ciliate cell squashing, the released bacteria were still capable of movement (L Modeo, pers. obs., 2017).

Transfection experiments

To verify RLO transfer to planarians, we fed animals either with ciliate-enriched liver paste (treated planarians) or plain liver paste (control planarians). 100% survival of the planarians (control and treated animals) was confirmed throughout the transfection experiments; in particular, the survival of treated planarians was not affected by the treatment and the animals did not show any morphological or behavioral alteration. The success of each of the two independent transfection experiments performed was investigated by means of PCR experiments and TEM observation on treated and control planarians at the different experimental timepoints, according to the described experimental procedure.

The DNA of “Ca. Trichorickettsia mobilis” was detected at all experimental timepoints (day 1, 3, and 7 after feeding) in genomic DNA preparations obtained from all investigated planarians fed with P. multimicronucleatum lysate-enriched liver paste. The size of the nested amplicon (1,360 bp) obtained from treated planarian samples matched the size of the amplicon obtained in DNA purified from ciliate monoclonal strain US_Bl 16I1 (positive control) (Fig. 4). The sequencing of the obtained amplicons confirmed the specificity for “Ca. Trichorickettsia mobilis” (100% sequence identity with the sequenced obtained from the symbiont). No amplification product was obtained in genomic DNA samples from control planarians fed with plain liver paste (negative control) (Fig. 4).

Ultrastructural observation was conducted on all TEM-processed specimens per each timepoints of the experiments, for both experimental groups, i.e., treated and control planarians.

In tissues collected from all the investigated planarians fed with Paramecium US_Bl 16I1 lysate-enriched liver paste, the presence of bacteria with morphology and sizes (see below for details) fitting with those of ciliate endosymbionts within intestinal cells was detected at all timepoints of the two experiments (Figs. 5 and 6).

At day 1 after feeding, several bacteria were recognizable in planarian enterocytes. Some of them appeared to be in division (Figs. 5A–5D). In several cases, the membrane of bacteria-including phagosomes was damaged and interrupted, and bacteria showing a two-membrane cell wall and a surrounding clear halo were detected free in the cytoplasm of planarian intestinal cells (i.e., in direct contact with their cytoplasm) (Figs. 5E, 5F). Flagella (diameter: ∼9 nm) could also be detected on bacterial surface (Figs. 5B, 5F). Additionally, extruded trichocysts (extrusive organelles typical of Paramecium), presumably included in Paramecium homogenate and ingested by treated planarians, were easily recognizable inside intestinal cell phagosomes (Fig. 5D).

A similar scenario was observed at day 3 after feeding (Figs. 6A–6D), with several bacteria in apparent division as well (Figs. 6C, 6D). At day 7 after feeding, a few bacteria (i.e., roughly less bacteria than the several observed after 1 and 3 days) were still well visible in planarian enterocytes (Fig. 6E), also free in the cytoplasm (Fig. 6F).

No bacteria were ever observed in tissues other than intestine in treated planarians; similarly, no bacteria were observed in tissues of TEM-processed control animals (Fig. S3).