In this study, we demonstrated that maintenance of intracellular algal symbiosis decreased crystal retention in the host cytoplasm. Although the mechanism underlying the reduction in the abundance of intracellular crystals is unclear, two potential causes associated with the infection process of Chlorella sp. in alga-free P. bursaria cells have been proposed: first, the host cells may excrete crystals into the culture solution; second, the presence of guanosine metabolites (i.e., crystal precursors), which is a prerequisite for the establishment of endosymbiosis. We observed crystals frequently excreted through the host cytoproct during algal endosymbiosis in alga-free P. bursaria (Kodama, unpublished data), which supports the first possibility; however, the exocytosis of intact crystals has never been reported in Paramecium [9]. Since the crystals in the ciliates are “stores’ of waste products of guanosine metabolism abundant under rich food supply and scarce under starvation, the low number of crystals in algae-bearing P. bursaria may be explained by the lack of excess nutrients due to their symbiotic Chlorella sp. Long crystals formed in the case of cultures kept in the dark may be explained by the appearance of excess nutrients in the ciliate because of autophagy of the algae in the dark. In the course of autophagy the amount of metabolites must increase. The absence of crystal formation in cultures kept in the dark without feeding was in good agreement with this explanation. Thus, the presence of crystals or guanosine metabolites (i.e., crystal precursors), rather than the presence of crystals, may be a prerequisite for establishing a symbiotic relationship. We performed an infection experiment with the expectation that addition of crystals isolated from alga-free P. bursaria to Chlorella sp. would increase the rate of algal endosymbiosis. However, we found no difference in endosymbiosis rates between alga-free P. bursaria ingesting a mixture of Chlorella sp. and crystals and control P. bursaria cells ingesting only Chlorella sp. (Kitatani and Kodama, unpublished data). A previous study demonstrated that the number of host mitochondria and trichocysts was significantly reduced with increasing numbers of endosymbiotic algae [24, 25, 34]. Similarly, algal reinfection may decrease the number of host crystals. The entry of algae into host P. bursaria cells may trigger the ejection of crystals occupying a major space in the cytoplasm to secure the space required for endosymbiotic association. Recently, [30] revealed a significantly decreased abundance of crystals, which almost disappeared when the original symbiotic Chlorella sp. was used to inoculate alga-free P. bursaria cells, whereas free-living Chlorella sp. induced a smaller decrease in the number of crystals, thus supporting the second possibility.

Pilátová et al. [8] suggested “that purine crystals, possibly present in the last eukaryotic common ancestor, were the first type of biocrystals in eukaryotes contingent on the emergence of cell compartmentalization in early eukaryotes. Owing to the low-solubility and high-capacity, purine inclusions possibly have emerged through an adaptation to nitrogen detoxification, protection against exposure to high levels of ammonia or nitrates, and utilization of vacuoles as a versatile sequestration space.”Although the role of crystals in the process of establishing or maintaining endosymbiosis between P. bursaria and zoochlorellae remains unclear, this is the first study to report the effect of endosymbiotic algae in P. bursaria on the prevalence of host cytoplasmic crystals. Daniels [35] demonstrated that starving for 1–2 days induces complete disappearance of crystals from the Amoeba and crystals reappear when the Amoeba are fed with prey. Therefore, it has been speculated that crystals act as a source of nutrients for Amoeba [35]. In this study, algae-bearing P. bursaria Yad1g1N contained few crystals (Fig. 1a, middle); however, the luminosity of Yad1g1N was also 7.8 (Fig. 1c), which may be attributed to the crystals in Chlorella, as demonstrated by [8]. A photosynthetic product, mainly maltose, is provided to the host P. bursaria [15, 36]; hence, starvation is not expected in host cells under constant light conditions. In fact, algae-bearing P. bursaria grow faster than alga-free cells under starvation conditions [31]. The absence of crystals in the cytoplasm of algae-bearing cells, even when nutrients are available, suggests that crystalline components may be used to maintain endosymbiosis; however, more research is needed in the future.

Mycosporine-like amino acids (MAAs) produced by symbionts may be related to host protection through the accumulation of sunscreen compounds in tissues [37]. The existence of MAAs in symbiotic ciliates has been reported in marine and freshwater species; however, the presence of MAA has not been confirmed in algae-bearing P. bursaria [37]. Summerer et al. [38] reported that exposure to artificial UV radiation (UVR) + photosynthetically active radiation (PAR) and ‘‘high’’ PAR (160 mmol m−1 s−1) showed an immediate aggregation of algae-bearing P. bursaria into several dense ‘‘spots’’ of approximately 1–3 mm in diameter in a Petri dish. Furthermore, Summerer et al. [38] reported that P. bursaria can protect against UV damage by accumulation as well as by symbiont dislocation. One of the functions of protist crystals is to protect them from UV radiation [39]. As an alternative to UV protection by symbiotic zoochlorellae, alga-free P. bursaria may increase its retained crystals.

Although it has also been reported that alga-free P. bursaria is found in nature [41], some algae were present in P. bursaria cells collected from the field (data not shown). Therefore, it can be said that the presence of endosymbiotic zoochlorellae is typical for P. bursaria. When the number of algae is artificially reduced by culturing them under constant dark conditions with food bacteria, crystals may be generated from substances obtained during the digestion processes of both algae and bacteria, and they may be stored in the cytoplasm. Thus, storing the crystals in the cytoplasm of alga-free P. bursaria cells may provide opportunities for endosymbiosis.

Why do crystals in the cytoplasm became larger as the number of symbiotic algae decreases (Fig. 2)? Analysis of the crystal length revealed that the minimum length was 0.2 μm, while the maximum length was 25 μm in P. multimicronucleatum, the larger species in the genus Paramecium [1]. As shown in Fig. 2b, in the crystals isolated from the algae-reduced cells, the minimum length was 3.4 μm and the maximum length was 33.5 μm; the maximum length was larger than that of P. multimicronucleatum (0.2–25 μm; [1]). The crystal structure is surrounded by a membrane [9]; hence, the crystals are considered to potentially grow inside the vesicle and increase due to the binding of crystals wrapped in another vesicle membrane. It is possible that the components obtained from the digestion of symbiotic algae by P. bursaria are involved in this growth method. Figure 2a (left and middle) shows crystals grown after algal digestion under constant dark conditions. Crystals were absent in the absence of bacteria (Fig. 2a, right). This interesting change accompanied by algal reduction possibly indicates that the crystals are not made from algal-digested components alone, but involve bacterial-digested components. While analyzing the size of the crystals, Hausmann et al. [9] reported that Paramecia fed on bacteria contained small crystal particles; after being fed on protein or meat extracts, they contained numerous large crystals. Foraminifers contained crystals after feeding on copepods or ciliates; however, no crystals were found after a diet restricted to diatoms. However, the direct relationship between food digestion and crystal formation has yet to be determined.

Regarding morphology of the crystals, the crystals isolated from the alga-free P. bursaria had a round shape (Fig. 4a), whereas those isolated from the algae-reduced P. bursaria had angular shapes, such as rods and plates (Fig. 4b). The crystals of protists are one of the criteria used for species identification [40]; however, our results showed that the size and shape of the crystals in P. bursaria changed significantly with changes in the number of symbiotic algae (Fig. 4). Interestingly, alga-free P. bursaria cells fed compatible Chlorella sp., such as the original symbiotic algae, lost their intracellular crystals during the algal infection process, but not when fed less compatible Chlorella sp., such as free-living Chlorella sp. [30]. Since one of the roles of crystals is as a storage reservoir for purines and organic nitrogen [3], the amount and composition of photosynthetic products of symbiotic algae may affect the host crystals.

Crystals of seven protist species (Mayorella sp., Cochliopodium bilimbosum, Trichamoeba villosa, Chaos diffluens, Chilomonas paramecium, Halteria grandinella, and Paramecium multimicronulceatum) have been shown to dissolve immediately after treatment with strong acids or bases [1]. Furthermore, their crystals were dissolved in water within 15 min. As shown in Table 1, P. bursaria crystals also showed a similar trend of high solubility in strong acids and bases, but they were not dissolved in water. A comparative study of the solubilities and melting points of paramecium and other protist crystals in various solutions may help predict the constituents or functions of the crystals.

As shown in Fig. 3a, we successfully isolated a large number of high-purity crystals from the Paramecium cells, revealing high stability at − 20 °C (Table 2). The crystals contain guanine, suggesting that successful large-scale cultures of P. bursaria could lead to an environmentally safe fertilizer.