Simultaneous Evaluation of Life Cycle Dynamics between a Host Paramecium and the Endosymbionts of Paramecium bursaria Using Capillary Flow Cytometry

Endosymbioses are driving forces underlying cell evolution. The endosymbiosis exhibited by Paramecium bursaria is an excellent model with which to study symbiosis. A single-cell microscopic analysis of P. bursaria reveals that endosymbiont numbers double when the host is in the division phase. Consequently, endosymbionts must arrange their cell cycle schedule if the culture-condition-dependent change delays the generation time of P. bursaria. However, it remains poorly understood whether endosymbionts keep pace with the culture-condition-dependent behaviors of P. bursaria, or not. Using microscopy and flow cytometry, this study investigated the life cycle behaviors occurring between endosymbionts and the host. To establish a connection between the host cell cycle and endosymbionts comprehensively, multivariate analysis was applied. The multivariate analysis revealed important information related to regulation between the host and endosymbionts. Results show that dividing endosymbionts underwent transition smoothly from the division phase to interphase, when the host was in the logarithmic phase. In contrast, endosymbiont division stagnated when the host was in the stationary phase. This paper explains that endosymbionts fine-tune their cell cycle pace with their host and that a synchronous life cycle between the endosymbionts and the host is guaranteed in the symbiosis of P. bursaria.

Over the last several decades, flow cytometry (FCM), a powerful and valuable tool for studies of cell biology and microbiology, has become widely used. Based on hydrodynamic focusing with a sheath fluid, this technique can assess cell optical information. Moreover, FCM has functions that include several procedures such as cell counting, detection of biomarkers, and cell sorting. In fact, FCM can save more time and provide higher sensitivity to optical properties than microscopic observation can. These features of FCM have provided an important boost to the apparatus as a driving force for important breakthroughs in science, protein engineering, and healthcare. Commercial flow cytometers with standard flow cell tips having diameters of 50-72 μ m 39 are generally optimized for standard suspended cells with 1-30 μ m diameter 40 . Using flow cytometry is challenging for analysis and sorting of larger cells, even plant cells, which are large (ca. 20-100 μ m), thereby limiting the use of a typical nozzle in commercial flow cytometry 41 . For a smooth and clog-free run, the flow nozzle should be at least four times the objective particle size 41 . Therefore, the FCM technique has been used only rarely to investigate complex systems between eukaryote cells and small eukaryotic symbionts such as P. bursaria, which is a large cell for examination using conventional FCM technique. Instead, exosymbiotic algae isolated from P. bursaria, but not host cells, have been analyzed mainly using FCM 42 because it has been assumed that host ciliates are damaged and broken by the passage pressure of hydrodynamic focusing during FCM measurements using standard flow cell tips. Actually, P. bursaria, having cell width and length of ca. 50 μ m and 120-150 μ m, is larger than the suspended cells applicable to standard FCM. Some reports of the literature describe the flow cytometric analysis of P. bursaria using a special size instrument nozzle of 200 μ m diameter 43 . However, many data obtained only from their algae are insufficient to elucidate the mechanisms of endosymbiosis because algae have already been disengaged from control by the host. A recent study has demonstrated that capillary FCM, not hydrodynamic focusing, can detect intact P. bursaria 44 . That fact implies that this system might be useful as a powerful tool for studies of symbioses such as that shown by P. bursaria.
This report presents a method of simultaneous detection of life cycle dynamics between hosts and endosymbionts of P. bursaria using capillary FCM with no special size nozzle. This study evaluated the relation between the host and endosymbiont life cycle using multivariate principal component analysis (PCA). Results show that the host cell cycle accommodated the cell cycle pace of endosymbiotic algae.

Results
Detection of population dynamics of P. bursaria host cells using capillary FCM. Before evaluating the coordination between the host and endosymbiont division, the growth of P. bursaria host cells was tracked (black circle in Fig. 2A). In addition to the total number of host cells, the generation time (white squares in Fig. 2A) was calculated. Here, the generation time of P. bursaria signifies the time-to-next division after the prior division. When the number of paramecia increased logarithmically and reached the stationary phase, the generation time was extended gradually from ca. 16 hr to more than 24 hr. This delayed generation time was fundamentally equal to the reduction of the division time of P. bursaria (white squares in Fig. 2B).
Cell sizes of P. bursaria change slightly, as shown in Fig. 2C-F, through the Paramecium cell cycle: P. bursaria is depicted at the interphase (Fig. 2C), paramecia are presented in Fig. 2D,E at the nuclear division phase; and Paramecium is portrayed at cytokinesis (Fig. 2F). The P. bursaria presented in Fig. 2F is divided into two daughter cells a few minutes later. This study was conducted to detect cell cycle-dependent sizes using capillary FCM. To extract signals derived from host paramecia using FCM, the gating method described in a previous study 44 was adopted. Results revealed that three distinguishable populations were detected according to cell size based on forward scatter signals (FSS) (Fig. 2G): The population of small cells of P. bursaria immediately after cell division (dotted line, designated as A D ), that of cells at cytokinesis as in Fig. 2F (dotted line, B D ), and that of the intermediate cell sizes at the interphase or nuclear division phase (dotted line, Int). FCM data from paramecia at the early stationary phase converged on the intermediate sizes, whereas those at the log phase (1-day to 4-day of incubation) diverged from the population for small cells after cell division to that at cytokinesis. Consequently, FCM was able to detect the cell-cycle-dependent size change of P. bursaria.

Detection of population dynamics of endosymbionts in P. bursaria using capillary FCM.
Gradual reduction of the proliferation activity of P. bursaria ( Fig. 2A,B) is generally related both to depletion of nutritional resources and to deterioration of culture conditions because of cell waste. Microscopic comparison of the number of endosymbionts in individual P. bursaria at the interphase with that at the division phase in earlier studies 15,45 has shown that P. bursaria at the division phase had almost double the number of endosymbionts as those present at the interphase. However, it remains unclear whether the endosymbionts have a growth rate that matches the culture-condition-dependent reduction of the host proliferation activity.
To confirm the optical properties of endosymbionts throughout the population dynamics of P. bursaria, FCM was used for this study. To extract signals for endosymbionts of P. bursaria from FCM data including host signals, gating strategies for endosymbionts described in the previous study 44 were used. Whereas the size and chlorophyll contents of host cells were evaluated from the R1 population for intact P. bursaria host cells, those for endosymbiotic algae were evaluated simultaneously from the R2 population for endosymbiotic algae (see Methods). First, this study evaluated correlation between the host and algal life cycle using PCA of multivariate analysis. The dataset from FCM data was subjected to PCA method, which reduces multiple-dimensional information to arbitrary one-dimensional information. The primary (PC1), secondary (PC2), and tertiary (PC3) components respectively presented 51.7%, 36.5%, and 11.8% of information in data from the contribution rate (Fig. 3A). Figure 3B presents the principal component loading of PC1 and PC2. Each loading showed that all parameters, including the algal size (FSS), red fluorescence intensity, and culture time of P. bursaria (Day), were positively correlated with PC1 (Fig. 3B). Particularly, correlation factors for both the algal size and the red fluorescence intensity were more strongly positive with PC1 than the culture duration of P. bursaria was. By contrast, the algal size and the culture duration respectively showed inverse and positive correlation with PC2. Figure 3C presents the culture Scientific RepoRts | 6:31638 | DOI: 10.1038/srep31638 condition-dependent variation of endosymbionts as expressed by the score plot of PC1 versus PC2. The characteristics of both algal size and red fluorescence intensity were mainly reflected as the variation of endosymbionts on the positive PC1-axis (Fig. 3B), whereas only the culture duration of P. bursaria mainly affected the variation of endosymbionts on the positive PC2-axis (Fig. 3B). This result reflects that the variation of endosymbionts changed the culture duration of P. bursaria dependently (Fig. 3C). Results show that both the cell size and chlorophyll fluorescence intensity of endosymbionts can be an indicator to assess the variation.
The PCA analysis results (Fig. 3) prompted us to produce plots of FSS for endosymbionts versus the red fluorescence intensity for endosymbionts (Fig. 4). This result also demonstrates that the variation of endosymbionts changed the culture duration of P. bursaria dependently, as depicted in Fig. 3. As presented in Fig. 5A, the cell size and chlorophyll contents of exosymbiotic algae are correlated strongly with the algal cell cycle 46 . For this study, the distribution of endosymbionts in Fig. 4 was categorized provisionally into three populations: a cell population of both small size and high chlorophyll contents (designated as Region I), a population of both large size and high chlorophyll contents (Region II) and a population of low chlorophyll contents (Region III) based on both algal size and chlorophyll contents (Fig. 5B). Here, Regions I, II, and III were assumed respectively as populations of a growing alga, populations including autospores, and populations of a unicellular cell immediately after division. Furthermore, the ratio of algae in each region was estimated and compared to each culture term (Fig. 5C). The ratio of algae at regions I and II increased gradually with incubation time, although those at region III decreased conversely.

Comparison of population dynamics of endosymbionts in P. bursaria with their host paramecia.
In general, the algal ratio of each cell cycle phase at some instant is related closely to the actual duration of the cell cycle phase. As described above, the generation time of P. bursaria was delayed culture-condition-dependently ( Fig. 2A). This study compared the time-dependent population changes of endosymbionts in P. bursaria (Fig. 5C) with the Paramecium generation time ( Fig. 2A). When P. bursaria host cells were at the logarithmic phase, FCM was able to detect little algae at regions I and II before division (Fig. 6). In contrast, more algae at regions I and II were estimated (Fig. 6) when host cells were at the early stationary phase. That result leads ultimately to the delayed generation time of endosymbionts. Consequently, results confirmed the synchronization of culture-time-dependent behaviors between endosymbiotic algae and their host paramecia.

Discussion
Recently, FCM has offered powerful and effective procedures to elucidate numerous biological, medical, and bioengineering challenges. Most flow cytometers use hydrodynamic focusing with sheath flow to analyze particles such as cells one-by-one. However, FCM has not been used to analyze symbiosis models directly because representatively complex symbioses models such as P. bursaria, H. viridissima, and coral are sufficiently larger than suitable cells for FCM measurements. A recent study detected intact P. bursaria cells using capillary FCM, not hydrodynamic focusing with sheath flow fluid 44 . In addition to the detection of P. bursaria host cells, this study was intended to examine life cycle behavior between the host and the endosymbionts in P. bursaria using capillary FCM. In fact, the capillary FCM in this study was able to detect subtle changes at the microscopic level and to discriminate P. bursaria cells at the interphase from those at the division phases (Fig. 2G). The FCM system was also able to discriminate endosymbionts from the host P. bursaria using FSS and red fluorescence intensity, as described in an earlier study 44 . Technical advantages of microcapillary FCM over other conventional techniques are the implementation of simultaneous detection and evaluation of both endosymbionts and their host P. bursaria. Moreover, no need exists for any special size nozzle. Many time-saving and sensitive procedures are therefore available to conduct symbiosis studies.
A single-cell analysis of P. bursaria using microscopy revealed that the number of endosymbionts almost doubles when the host cell of P. bursaria is at the division phase 15,45 . This result demonstrates that the host cell of P. bursaria strictly controls the number of endosymbionts in one way or another. Endosymbionts must arrange their cell cycle schedule accordingly if the culture-condition-dependent change delays the generation time of P. bursaria ( Fig. 2A). However, it remains poorly understood whether the endosymbionts act in concert with the culture-condition-dependent behaviors of P. bursaria. From data of both Paramecium numbers using microscopy and those of endosymbionts using FCM, this study investigated the life cycle behaviors of endosymbionts and the host Paramecium. To establish a connection between the host cell cycle (Fig. 2) and multivariate data of endosymbionts obtained using FCM comprehensively, the PCA method was applied (Fig. 3). That method derived two important factors related to the connection between the host and the endosymbionts cell cycles from data: the endosymbiont size and chlorophyll fluorescence intensity. Graphs of size versus chlorophyll fluorescence of endosymbionts detected that the ratio of algae after division decreased gradually over time, instead of those before division increasing (Fig. 4). Results show that endosymbiotic algae in regions I and II divided immediately when the host Paramecium was at the logarithmic phase (Figs 4, 5 and 6). Algae in regions I and II divided instantly. A smooth transition of endosymbionts occurred from algae at the growing and division phases (regions I and II) to algae at region III. In contrast, endosymbiotic algae in regions I and II stagnated when the host Paramecium was at the stationary phase (Figs 4, 5 and 6). It also showed that algae at regions I and II divided at a crawl and that slow transition of endosymbionts occurred from algae at the division phase to algae at region III. The relation between the host Paramecium and the endosymbiotic algae, which behave like chloroplasts, is analogous to that between a photosynthetic eukaryotic cell and their chloroplast(s). The synchronous division between chloroplasts and algae in most algal species is achieved by expression control of the nucleus-encoded chloroplast division genes and proteins 9,10 . In practice, most chloroplast genes derived from ancestral cyanobacterium have been lost or relocated to the host nucleus through EGT events [6][7][8][9][10] . Therefore, organelles such as chloroplasts no longer survive outside of the eukaryotic cell. In contrast to organelles, exosymbiotic algae isolated from P. bursaria can proliferate independently from the host ciliate 32 . That capability demonstrates that algae isolated from P. bursaria, unlike organelles such as chloroplasts, can self-manage their cell cycle pace. However, these results support the inference that endosymbionts find it necessary to hasten or slow their cell cycle pace with their host cell cycle and that a synchronous life cycle between the endosymbionts and the host Paramecium is guaranteed in the P. bursaria symbiosis system.
This study compared the synchronous division cycle strategies of endosymbionts and the host in P. bursaria with several strategies in other symbiosis systems (Fig. 7). To realize synchronous division between organelles and their host cells, sophisticated strategies including EGT and LGT have been performed. Particularly, LGT is an important factor in establishing a stable endosymbiosis with other eukaryotes because the new host cell can take over the function from EGT-containing organelle division genes by LGT. Actually, LGT and synchronization of division cycles in several ongoing endosymbioses (upper left and lower right in Fig. 7) are apparently less successful than well-established and integrated symbioses in secondary and tertiary endosymbioses (upper right in Fig. 7). Several photosynthetic animals including sacoglossan sea slugs have evolved to acquire only plastids (kleptoplasts) from algal prey 4 . However, they seem to neglect or fail to obtain endosymbiotic related genes as LGT from the algal prey. In fact, the plastids are not transmitted vertically and do not undergo division in the  31 concluded from RNA-seq analysis that there is no evidence that the evolution of kleptoplastry in sea slugs involves LGT events. The plastid from endosymbionts seems also not to undergo division in the marine ciliate M. rubrum [25][26][27][28] , which sustains not only plastids but also a transcriptionally active nucleus from algal prey. Therefore, it is considered that M. rubrum may need to feed periodically to replace ageing plastids or plastids diluted out by host cell division 25 . New individuals of these photosynthetic animals and ciliates must necessarily acquire new plastids from algal prey to obtain photosynthetic benefits. The endosymbiosis in these photosynthetic organisms is apparently an unauthorized strategy without LGT events or with incomplete  LGT. The ongoing plastid-acquisition types of endosymbiosis are defective in their ability to maintain permanent endosymbiosis compared to secondary and tertiary endosymbioses (upper left in Fig. 7). In contrast to the transient endosymbioses explained above, the endosymbiosis in P. bursaria seems to apply a coordinated strategy by which endosymbionts fine-tune their cell cycle pace with their host Paramecium (Fig. 6 and lower right in Fig. 7). In practice, some endosymbiotic algae can at least undergo cell division and are transmitted to daughter cells. Furthermore, P. bursaria has the capability of re-establishing the endosymbiosis between the host and their endosymbionts for emergency, even if the host cell loses the endosymbiotic algae 32,37,38 (* , lower middle in Fig. 7). All intact alga-embracement types, however, are not always coordinated and are not superior to the plastid-acquisition types of endosymbiosis. A katablepharid flagellate, Hatena arenicola, has an endosymbiotic alga, which is slightly atrophied rather than intact 8 . When the flagellate host cell is at the division phase, the endosymbiont does not divide synchronously with the host. After host cell division, the flagellate harboring the endosymbiont and the other with no endosymbiont are produced 8 (upper middle in Fig. 7). It is particularly interesting that the daughter cell with no alga can uptake the partner alga and can re-establish endosymbiosis between the host and the endosymbiont 8 just as algae-free P. bursaria can, although the appearance of algae-free P. bursaria is practically from artificial conditions [34][35][36][37] . The basal metazoan H. viridis also has endosymbiotic algae 12,17,18 . The symbiotic partners are necessary for the hydra sexual reproduction, particularly for the production of female gonads 18 , although the presence or absence of endosymbiotic algae is unaffected in P. bursaria 47 . Moreover, P. bursaria host cells are ready to cooperate with photosymbiotic partners in tolerance to reactive oxygen species, which might be generated from energetic metabolism containing photosynthesis, even if algae-free Paramecium is 43,48 . Considering several endosymbioses, the endosymbiosis in P. bursaria is coordinated fortuitously with endosymbionts, not so much with control of endosymbiotic algae by the host, as with that of organelles such as chloroplasts, for photosynthetic function and transmission of endosymbionts after cell division.
To elucidate symbiosis in P. bursaria, the effects of some chemicals such as protein synthesis inhibitors [34][35][36]49 and cytoskeletal depolymerization inhibitors 15 on P. bursaria were examined microscopically. Studies based on microscopic observation have generally involved great amounts of time for obtaining data. Moreover, they have not been done with high-throughput experiments. FCM can save time better than microscopic observations can. When Gerashchenko et al. 42 first introduced FCM to study the endosymbiosis of algae in P. bursaria, almost all commercially available flow cytometers applied hydrodynamic focusing with sheath fluid in the biomedical markets of that time. Therefore, instead of analyzing P. bursaria host cells directly, exosymbiotic algae isolated from their host paramecia were the main target of FCM measurements 42,46 . However, their algae are far from airtight experiment materials to elucidate the mechanisms of endosymbiosis because algae have not been controlled stringently by the host. Newly and commercially supplied capillary FCM was able to detect intact P. bursaria 44 . Although the hydrodynamic focusing system has been superior to the microcapillary system in terms of resolution ability to detect small cells, this system with microcapillary technology holds enormous potential for application to studies of symbioses such as those of P. bursaria. In addition to detecting endogenous fluorescence of their algae, the combination of fluorescence labeling antibodies for cell cycle dependent biomolecules and flow cytometry might provide valuable insights to support symbiosis studies. Taken together, this system with microcapillary technology can offer a powerful tool for studies of symbiosis and evolution of eukaryotic cells.

Methods
Strain and culture condition of P. bursaria. After collecting P. bursaria from a river Asida-kawa (Hiroshima prefecture, Japan), paramecia were cultured in a lettuce infusion containing Klebsiella pneumonia bacteria one-by-one into a culture dish. A cloned P. bursaria syngen I (AS-2, mating type IV) of paramecia was used for this study. The cloned paramecia were cultured in lettuce infusion containing K. pneumonia as food under an LD cycle (12 h light/12 h dark) at ca. 1100 lux natural white fluorescence light and 23 ± 2 °C.

Culture condition of Chlorella-like exosymbiotic algae isolated from P. bursaria. Chlorella-like
exosymbiotic algae (SA-1 strain) isolated from P. bursaria 32 were used. The cloned strain of exosymbiotic algae on an agar plate containing CA medium 32 were cultured under a constant light condition at 3500 lux of natural white fluorescent light and 25 ± 1 °C.
Growth curve of P. bursaria. Paramecia at the logarithmic to stationary phase were collected and counted using a stereomicroscope. Paramecia at an initial density of 20 cells/ml were cultured in fresh lettuce medium containing K. pneumonia under an LD cycle (12 h light/12 h dark) at ca. 1100 lux of natural white fluorescence light and 24 °C. The paramecia (average ± standard deviation) were counted over time using stereomicroscopy. The generation time (g) (hr) was calculated from data obtained as given in equations (1)-(3). Here, N 0 and N respectively denote the initial density of paramecia (cells/ml) before incubation and the numbers of paramecia (cells/ml) after incubation. Therein, t denotes the incubation time (hr). Therefore, t g represents the frequency of cell division.
Preparation of specimens for light microscopy. Paramecia were collected and fixed with 5% (v/v) formalin. Microscopic images were made of collected paramecia fixed with 5% formalin. Evaluation for P. bursaria size using microcapillary FCM. To estimate the Paramecium size, a microcapillary flow cytometer (Muse TM Cell Analyzer; Merck Millipore), but not one based on hydrodynamic focusing with sheath flow, was used as described in an earlier report 44 . P. bursaria fixed with 5% (v/v) formalin was analyzed at flow rates of 0.59 μ l/s (sample load time < 2 min/single measurement). Cells that went through the rectangular capillary with 100-μ m round bore were analyzed using this instrument equipped with a green laser operating at 532 nm for several optical properties. FSS were collected to ascertain cell size. The fluorescence of chlorophyll was detected in the red fluorescence channel through a 680/30 nm band pass filter simultaneously. To obtain sufficient data including intact P. bursaria, FCM measurements were repeated several times (5-20 times). The paramecium amounts in the logarithmic phase did not yet reach sufficient numbers. Therefore, the shorter the culture duration was, the more repeated measurements for intact P. bursaria cells were needed.
Individual signals in each FCS3.0 file were extracted using software (FCSExtract utility ver. 1.02; Stowers Institute for Medical Research). They were subsequently re-analyzed using standard spreadsheet software (Access and Excel; Microsoft Corp.). When P. bursaria cells with endosymbiotic algae were analyzed using FCM, two distinct populations were detected (designated as R1 for intact P. bursaria cells and R2 for their algae released from several broken P. bursaria cells, respectively) as described in the previous report 44 . The size variation in P. bursaria was expressed as a distribution histogram of relative size from FSS.
Evaluation of size and chlorophyll contents of endosymbiotic algae in P. bursaria using microcapillary FCM. To estimate the optical properties of endosymbiotic algae, the microcapillary FCM was used as described above. The earlier report described that the passage rate of intact P. bursaria through the capillary was 79 ± 9.9% 44 . A P. bursaria cell at interphase has ca. 300-500 endosymbiotic algae in cytoplasm 15,45 . Consequently, the capillary FCM was able to analyze a statistically sufficient number of endosymbiotic algae released from several broken P. bursaria during FCM measurements. Whereas the size and chlorophyll contents of P. bursaria host cells were evaluated from the R1 population as described in the previous report 44 , those for endosymbiotic algae were evaluated simultaneously from the R2 population.
Correlative evaluation using PCA of multivariate analysis. This study was undertaken to evaluate the correlativity between host and algal life cycles over time. To evaluate the correlativity between multiple properties including incubation time of P. bursaria, and the size and chlorophyll contents of endosymbiotic algae, PCA of multivariate analysis was used for this study. As a dimensional reduction technique, PCA reduces multiple-dimensional information to arbitrary one-dimensional information, which is a dataset from a new axis produced by PCA. For this study, the dataset produced from FCM data was subjected to PCA analyses using software for multivariate analysis (Institute of Statistical Analyses, Inc.). Based on results of the correlation matrix analysis for the data obtained, the author calculated the contribution rates of respective components, the factor loading of each parameter, and the score plot of each component were calculated. Here, each factor loading indicates correlation factors between each parameter and each component (PC1-3).
Interim evaluation for cell cycle of endosymbiotic algae in P. bursaria. Previous studies determined the algal cell cycle from the size and chlorophyll contents of exosymbiotic algae isolated from P. bursaria 46 . After detection of signals for endosymbiotic algae released from P. bursaria, an interim evaluation for the cell cycle of endosymbiotic algae was done using signals of the size and chlorophyll contents. Endosymbionts were categorized into three populations: cell population of both small size and high chlorophyll contents (<75 on the forward scatter signal and ≥15 on the red fluorescence intensity) (region I); the population having both large size and high chlorophyll contents (≥75 on the forward scatter signal and ≥15 on the red fluorescence intensity) (region II); and the population with low chlorophyll contents (<15 on the red fluorescence intensity)(region III). Then, the respective time-dependent change ratios of algal numbers for all regions were estimated at different incubation times of P. bursaria.