Photosynthesis and other factors affecting the establishment and maintenance of cnidarian–dinoflagellate symbiosis

Abstract

Coral growth depends on the partnership between the animal hosts and their intracellular, photosynthetic dinoflagellate symbionts. In this study, we used the sea anemone Aiptasia, a laboratory model for coral biology, to investigate the poorly understood mechanisms that mediate symbiosis establishment and maintenance. We found that initial colonization of both adult polyps and larvae by a compatible algal strain was more effective when the algae were able to photosynthesize and that the long-term maintenance of the symbiosis also depended on photosynthesis. In the dark, algal cells were taken up into host gastrodermal cells and not rapidly expelled, but they seemed unable to reproduce and thus were gradually lost. When we used confocal microscopy to examine the interaction of larvae with two algal strains that cannot establish stable symbioses with Aiptasia, it appeared that both pre- and post-phagocytosis mechanisms were involved. With one strain, algae entered the gastric cavity but appeared to be completely excluded from the gastrodermal cells. With the other strain, small numbers of algae entered the gastrodermal cells but appeared unable to proliferate there and were slowly lost upon further incubation. We also asked if the exclusion of either incompatible strain could result simply from their cells' being too large for the host cells to accommodate. However, the size distributions of the compatible and incompatible strains overlapped extensively. Moreover, examination of macerates confirmed earlier reports that individual gastrodermal cells could expand to accommodate multiple algal cells.

This article is part of the theme issue ‘Sculpting the microbiome: how host factors determine and respond to microbial colonization’.

1. Introduction

Symbioses between animals and microbes are evolutionarily ancient and typically are mutualisms that benefit both partners. Among the studies dealing with such mutualisms, most have focused on the interactions of animal hosts with either bacteria [1] or eukaryotic algae such as dinoflagellates [2]. A major goal in attempting to understand these mutualisms is to identify the underlying cellular and molecular mechanisms by which the hosts specifically acquire beneficial microbes and these microbes then colonize and remain stable in the hosts. These mechanisms are particularly important to study, because mutualistic animal–microbe interactions may degrade rapidly under stress.

The reef-building corals provide prominent examples of such stress-sensitive mutualisms. Corals are foundation species of their vitally important ecosystems but are threatened worldwide by climate change and other anthropogenic stresses. The health and growth of corals and many other cnidarians depend on a mutualistic relationship with dinoflagellate algae of the family Symbiodiniaceae [3], which live within vesicles (symbiosomes) in the host's gastrodermal cells and provide it with energy and organic building blocks through their photosynthesis [4–11]. Under heat and other stresses, this relationship can break down (bleaching), with harmful or even fatal consequences to the host [12–17]. A particular cnidarian host strain can typically establish effective symbioses with more than one, but not all, strains of Symbiodiniaceae and shows preferences among possible partners [18–33], and there is good evidence that the stress resistance of the holobiont depends on the properties of both the host and the algae [14,31,34–40]. Thus, there is great interest in the molecular mechanisms that determine the specificity of algal colonization and control symbiosis stability under both nonstressful and stressful conditions. However, despite much effort, these mechanisms remain very poorly understood.

One important question is whether symbiosis establishment and maintenance depend on algal photosynthesis, as might be expected given the centrality of this process to the metabolic economy of the holobiont. However, the available data on this point are conflicting. It has been reported many times that reduction or elimination of photosynthesis by incubation in darkness or by exposure to the photosystem-II inhibitor DCMU results in a loss of algae from the host tissue [7,17–19,26,41–49]. In some cases, this loss has appeared to proceed eventually to completion [7,17–19,26,45,47,49], but in other cases, small numbers of algae have been reported to remain in the hosts even after long periods without photosynthesis [19,44,45,48]. Moreover, it has been reported that the algae may sometimes function as heterotrophs, receiving sustenance from the host that allows their continued maintenance, and even proliferation, in the absence of photosynthesis [44,50]. Most recently, Jinkerson et al. [51] reported that neither establishment nor maintenance of the symbiosis depends on photosynthesis, albeit with somewhat different outcomes in the three cnidarian species examined. Specifically, they reported that colonization of aposymbiotic hosts could occur in the absence of photosynthesis in all three species, but that algal numbers then declined slowly in the jellyfish Cassiopea, were maintained for long periods but without any increase in numbers in the anemone Aiptasia, and increased through algal reproduction in the coral Acropora.

Another important question concerns the stage at which discrimination between compatible and incompatible algal strains is made. Cnidarian gastrodermal cells must be able to take up compatible algae without digesting them during colonization, but they must also be able to take up and digest food particles during feeding. In addition, there is abundant evidence that they will take up, at least transiently, a variety of incompatible (or less favoured) algal types [18,19,21–24,29–31,52–61] as well as nonliving material of appropriate sizes, such as carmine particles, dead algal cells, and plastic microbeads [29,51–53,56,61–64]. This apparent lack of selectivity during initial uptake has suggested that discrimination might be exercised primarily after the phagocytosis of potential symbionts, resulting in a ‘winnowing’ that leads to the final and stable symbiosis [8,18,21,22,31,54,55,57–61]. However, there is also considerable evidence that in some cases, discrimination against incompatible algae occurs during phagocytosis or even during uptake into the gastric cavity [19,23,27–30,52,53,56]. Clarification of this issue appears essential for progress in illuminating the molecular mechanism(s) of symbiotic-partner selection.

In studies to date of these molecular mechanisms, most attention has been focused on the hypotheses that modulation of the host's innate-immunity system (e.g. complement, TGFβ, Toll-like-receptor, or NF-κB signalling) is needed for stable symbiosis [32,61,65–70] and that partner recognition is mediated by interactions between cell-surface molecules of the host and algae [8,30,32,58,60,71–73]. However, it has also been suggested that something as simple as algal-cell size may play a major role in symbiont selection by hosts, with larger cell types simply being too big for successful phagocytosis and/or retention [29,63,74]. This hypothesis derives some credibility from the effects of target size on phagocytosis by mammalian macrophages [75,76], so it merits further examination even though it seems unlikely to explain more than a subset of the incompatibilities that have been observed.

In this study, we reexamine some of these questions using Aiptasia, a subtropical sea anemone that has been widely adopted as a model system for examining many aspects of coral biology [8,49,77–81]. Of particular value have been precise counts of algal cell numbers in adult anemones by flow cytometry and detailed three-dimensional imaging of Aiptasia larvae by confocal microscopy, which allowed us to determine with confidence whether particular algal cells were actually within the gastrodermal cells and not just on their apical surfaces or in the gastric cavity.

2. Results

(a) . Apparent dependence of symbiosis maintenance on algal photosynthesis

Because it remains uncertain whether the maintenance of cnidarian–dinoflagellate symbiosis depends on algal photosynthesis (see Introduction), we reexamined this question using two related strains of Aiptasia that differ only in their algal symbionts (see Material and methods). Upon incubation in complete darkness but in the absence of other stresses, both strains lost algae steadily over a 30-day period (figure 1a,b). Although the weekly feedings (see Material and methods) allowed even the animals held in the dark to maintain their size or even grow slightly during the experiments (electronic supplementary material, figure S1A, B), the loss of algae was equally evident when scoring was on a per-animal (figure 1a,b) or per-unit-protein (electronic supplementary material, figure S1C, D) basis. In a separate experiment using only strain CC7-SSB01, the loss of algae continued steadily until at least 60 days (figure 1c) and indeed resembled a first-order decay over that time (electronic supplementary material, figure S1E, F). It seemed possible that the apparent loss of algae could have resulted, at least in part, from a loss of chlorophyll fluorescence by individual algal cells, such that they were no longer scored as algae during analysis of the flow-cytometry data. However, inspection of the data showed that although the median fluorescence of the algal cells did indeed decline during the experiment, they were still detected and scored as algae, so that the diminished fluorescence made little or no contribution to the apparent reduction in algal numbers (electronic supplementary material, figure S2).

Figure 1.

Apparent dependence of symbiosis maintenance and establishment in adult anemones on algal photosynthesis. (a,b) Symbiotic anemones of strains CC7 (a) and CC7-SSB01 (b) were held in continuous darkness or under a 12 h light : 12 h dark cycle (control) with a feeding and subsequent water change once per week (see Material and methods). Samples were taken at intervals and scored for the number of algal cells per anemone by Guava flow cytometry. Means and standard deviations are shown for (control) 5 to 7 anemones per time point and (dark samples) 16–20 anemones (from three tanks – see Material and methods) per time point. (c) Anemones of strain CC7-SSB01 were incubated and analysed as in a and b for 60 d. In addition, some CC7 and CC7-SSB01 animals were held in continuous darkness for 302 d (three CC7, six CC7-SSB01) or 470 d (26 CC7, 36 CC7-SSB01), then returned to a 12 h : 12 h light-dark cycle, held for greater than 9 months, and examined visually and by fluorescence stereomicroscopy for evidence of repopulation by algae. 0*, no repopulation was seen. (d) Aposymbiotic anemones of strain CC7 were exposed for 24 h to algal strain SSB01 at approximately 10⁴ cells ml⁻¹ in either artificial seawater (ASW) or ASW containing 10 µM or 25 µM DCMU; lighting was 12 h : 12 h light-dark except where noted. After a change of ASW to remove DCMU and any remaining free algal cells, the animals were incubated in ASW without feeding for an additional 14 d in the dark (continuous dark) or under a 12 h : 12 h light-dark cycle (the others); the ASW was changed every 2 to 3 d to avoid fouling of the tanks. Animals were imaged at intervals using both transmitted-light (BF) and fluorescence (F) stereo microscopy. Images shown are representative of eight anemones observed under each condition at each time point in each of three independent trials. Red, chlorophyll fluorescence of algae; green, endogenous fluorescence of the host.

At times greater than 60 days in such experiments, the algal numbers were too low to count reliably by flow cytometry. However, when CC7 and CC7-SSB01 animals were held in darkness for 302 or 470 days, then returned to a 12 h : 12 h light-dark cycle, none of them showed any sign of recovery of their algal populations over greater than nine months of observation by visual inspection and fluorescence stereomicroscopy (figure 1c). We conclude that the progressive loss of algae from symbiotic Aiptasia held in the dark continues until no viable algae remain, and thus that symbiosis maintenance appears to depend on algal photosynthesis (see Discussion).

(b) . Apparent dependence of symbiosis establishment in adult anemones on algal photosynthesis

We next asked if we could see any evidence of colonization when we exposed aposymbiotic adult Aiptasia for 24 h to algae from a clonal, axenic culture under conditions in which photosynthesis was eliminated (constant darkness) or drastically reduced (presence of the photosystem-II-specific inhibitor DCMU) during exposure of the animals to the algae. Under the normal 12 h : 12 h light-dark cycle, colonization was readily detectable by fluorescence stereomicroscopy by 10 days and became progressively more conspicuous at later times (figure 1d, row 1; electronic supplementary material, figure S3, row 1). By contrast, algae were not detected within the host even after 15 days in the absence (or near absence) of photosynthesis (figure 1d, rows 2–4; electronic supplementary material, figure S3, rows 2–4). Control experiments showed that the concentrations of DCMU used indeed reduced photosynthetic function drastically (electronic supplementary material, figure S4A), but with only a small loss of algal viability (electronic supplementary material, figure S4B, C). Taken together, these results suggest that algal uptake by the host, algal proliferation within the host, or both are impaired when photosynthesis is compromised. However, in experiments with adult anemones, we could not readily distinguish among these possibilities or evaluate the severity of the effects.

(c) . Apparent dependence of algal proliferation, but not of initial algal uptake, on photosynthesis in anemone larvae

To examine more closely the effects of a loss of photosynthesis on symbiosis establishment, we turned to experiments with Aiptasia larvae, in which confocal microscopy allowed us to determine the localization (outside the larva, in the gastric cavity, or within the gastrodermal tissue) of each algal cell (electronic supplementary material, figure S5 and video S1). Five-day-old larvae were exposed to algae of strain SSB01 for 24 h either in constant darkness or under 12 h light followed by 12 h dark; in some cases, DCMU was also present to depress photosynthesis. After 24 h, the larvae were rinsed free of unassociated algae and DCMU and incubated further under a 12 h : 12 h light-dark cycle (most cases) or in continuous darkness (where indicated). Samples were taken for examination by confocal microscopy at the end of the initial 24-h incubation and 3 days and 9 days later (thus, 1-d, 4-d, and 10-d samples).

Under control conditions (12 h light : 12 h dark, no DCMU), algae were already numerous within the gastrodermal cells of most larvae by the end of the initial 1-day exposure period, and the fraction of larvae containing algae, the number of algal cells per larva, and the fraction of those algae that were within the gastroderm all continued to increase during the further incubation (figure 2a, row 1; table 1, row 1; figure 2b). By contrast, under conditions in which photosynthesis was eliminated or reduced, the fractions of larvae containing algae, the numbers of algal cells per larva, and the fractions of those algae that were within the gastroderm were all reduced relative to the control at the end of the 1-day exposure period, and the disparities increased further during the subsequent incubations (figure 2a, rows 2–4; table 1, rows 2–4; figure 2b). Importantly, however, those numbers were not zero. Thus, it appears that in the absence (or near absence) of photosynthesis, algae can still be taken up into the gastroderm of larvae, albeit with somewhat reduced efficiency. Once there, however, they proliferate little or not at all, as seen most clearly by focusing on the larvae that contained at least one alga (figure 2b, ‘continuous dark²’). Indeed, these algae are lost gradually over time if photosynthesis continues to be suppressed (as expected from the dark-bleaching experiments with adults: see above).

Figure 2.

Apparent dependence on photosynthesis of algal proliferation, but not of initial algal uptake by gastrodermal cells, in anemone larvae. Aiptasia larvae at approximately 5 d post-fertilization were exposed for 24 h to algal strain SSB01 at approximately 10⁵ cells ml⁻¹ in ASW under either continuous dark or 12 h light followed by 12 h dark. In some cases (as indicated), the ASW also contained 10 µM or 25 µM DCMU during the 24-h exposure of larvae to the algae. At 24 h, larvae from each culture were sampled for imaging by confocal fluorescence microscopy (1-d data), rinsed with ASW to remove free algae and (where relevant) DCMU, incubated further under continuous dark (where indicated) or a 12 h : 12 h light-dark regimen (the other cultures), and sampled again for imaging at 4 d and 10 d. (a) Single optical sections representative of observations made on the larvae containing algae in each sample. Red, chlorophyll fluorescence of algae; green, staining of host-cell boundaries by Alexa Fluor 488-phalloidin; n.d., not determined. (b) Histograms showing mean numbers (± SEMs) of algal cells scored manually as being within the gastroderm or free in the gastric cavity based on three-dimensional reconstructions of larvae from images like those in panel (a). For each case, ≥100 larvae were scored in each of three or six independent trials. continuous dark¹, scoring all larvae; continuous dark², scoring only larvae containing one or more algae. Different letters represent significantly different values in the numbers of algae in the gastroderm, as determined by a Tukey's test for pairwise comparisons; ANOVA, p < 0.001. See table 1 for additional details.

Table 1.

Extents of infection of Aiptasia larvae exposed to various algal strains under various infection conditions. All experiments used algal strain SSB01 except for those of Rows 5 and 6. In all experiments, algae from a stock culture in IMK + Cas medium (see Materials and methods) were grown for 14 days in IMK medium under continuous light prior to the start of infection. In all experiments, larvae at approximately 5 days post-fertilization were exposed to algae at approximately 10⁵ cells ml⁻¹ for 24 h (12 h light : 12 h dark except for Row 2); exposures were in ASW without additives except for Rows 3 and 4. In all experiments, larvae were sampled at the end of the 24-h infection period (1-d data), washed free of most algae (see Materials and methods), transferred to fresh ASW without additives or added algae, and incubated further; the further incubation was under a 12 h : 12 h light-dark regimen except in the case of Row 2 (note b). The data presented are the sums of three (most cases) or six (where noted) independent trials. n.d., not determined. In parentheses are shown the percentages of larvae containing one or more algae (in either gastric cavity or gastroderm) and the percentages of the algae observed that were in the gastric cavity or the gastroderm, respectively.

row	experiment	1 d			4 d			10 d
		larvae with: without algae	algae in gastric cavity	algae in gastroderm	larvae with: without algae	algae in gastric cavity	algae in gastroderm	larvae with: without algae	algae in gastric cavity	algae in gastroderm
1	12 h L:12 h D	358:264a (58%)	251a (16%)	1323a (84%)	360:244a (60%)	196a (8%)	2352a (92%)	230:101 (69%)	45 (2%)	2811 (98%)
2	continuous darkb	218:399a (35%)	426a (49%)	442a (51%)	243:373a (39%)	198a (34%)	385a (66%)	106:191 (36%)	100 (42%)	138 (58%)
3	DCMU (10 µM)c	125:198 (39%)	219 (48%)	238 (52%)	82:181 (31%)	79 (37%)	134 (63%)	n.d.	n.d.	n.d.
4	DCMU (25 µM)c	89:245 (27%)	155 (49%)	164 (51%)	43:195 (18%)	39 (28%)	101 (72%)	n.d.	n.d.	n.d.
5	strain SSA03d	48:214 (18%)	165 (100%)	0 (0%)	5:215 (2%)	13 (100%)	0 (0%)	n.d.	n.d.	n.d.
6	strain SSE01d	193:429a (31%)	465a (87%)	69a (13%)	82:572a (13%)	103a (66%)	53a (34%)	41:274 (13%)	68 (79%)	18 (21%)

^aData are the sums of six independent trials.

^bBoth the 24-h infection period and further incubation were in continuous darkness.

^cThe 24-h infection period used ASW containing 10 µM (Row 3) or 25 µM (Row 4) DCMU, but further incubation was in ASW without DCMU.

^dExperimental conditions were identical to those of Row 1 except for the use of a different algal strain.

(d) . Evidence for both pre- and post-phagocytosis mechanisms for discrimination against incompatible algal strains

Previous studies have shown that in contrast to strain SSB01 (Breviolum minutum; ITS Clade B1), Symbiodiniaceae strains SSA03 (Symbiodinium pilosum; ITS2 Clade A2) and SSE01 (Effrenium voratum; ITS2 Clade E) are not able to colonize an Aiptasia CC7 adult host, despite their photosynthetic activity, and that a similar discrimination is manifested also by Aiptasia larvae [26,27]. The confocal-microscopy method used here allowed us to shed more light on the mechanisms of this discrimination against incompatible strains.

With algal strain SSA03, although a substantial number of algae were observed within larval gastric cavities after the initial 1-day exposure period, and a few such algae remained after continued incubation to 4 days, we did not observe a single algal cell within the larval gastroderm at either time (figure 3a, row 1; figure 3c; table 1, row 5). Thus, we infer that there is either an efficient mechanism for preventing initial phagocytosis of SSA03 cells or an efficient mechanism for quickly recognizing such cells as incompatible after phagocytosis and digesting them or releasing them by exocytosis.

Figure 3.

Evidence for both pre- and post-phagocytosis mechanisms for discrimination by Aiptasia larvae against incompatible algal strains. Experimental conditions and analyses were like those in figure 2 (12 h : 12 h light-dark experiment) except for the use of incompatible algal strains SSA03 and SSE01. (a) Single optical sections representative of observations made on the larvae in each sample. (b) Higher-magnification images of portions of two individual larvae exposed to strain SSE01. Asterisk, the algal cell followed through the full z-stack of optical sections in electronic supplementary material video S2. In a and b: red, chlorophyll fluorescence of algae; green, staining of host-cell boundaries by Alexa Fluor 488-phalloidin. (c) Histograms like those in figure 2b. The 12 h : 12 h light-dark data from figure 2b are repeated here (SSB01) to facilitate comparison. A Student's t-test indicated a significant difference (p < 0.0001) between SSB01 and SSE01 at 10 d. At 1 d and 4 d, different letters represent significantly different values, as determined by a Tukey's test for pairwise comparisons; ANOVA, p < 0.001. See table 1 for additional details.

By contrast, although uptake of strain SSE01 was also much less than that of the compatible strain SSB01, some algal cells were unequivocally within the host gastroderm after the initial 1-day exposure period (figure 3a, row 2; figure 3b; figure 3c; electronic supplementary material, video S2), and these algae were lost only slowly (along with algae in the gastric cavity) during the subsequent days of incubation (figure 3c; table 1, row 6). Thus, we infer that although there appears to be some level of discrimination against SSE01 cells during initial entry to the gastroderm, there is an additional mechanism(s) that prevent(s) SSE01 cells that do enter the gastroderm from establishing a stable symbiosis and proliferating there, even though they are not rapidly digested or expelled.

(e) . Evidence against the hypothesis that excessive algal cell size is a major determinant of host–algal incompatibility

It has been suggested that some algal strains may be excluded from particular hosts simply because the algal cells are too large to be accommodated by the host's gastrodermal cells (see Introduction). Examination by confocal microscopy of Aiptasia larvae exposed to algae of strain SSB01, SSA03 or SSE01 allowed us to examine this possibility for these particular combinations of host and algae. We were able to get a precise measurement of each algal cell's diameter by measuring it in the particular optical section in which its periphery was in optimal focus (electronic supplementary material, figure S5; video S1). Although we found no SSB01 or SSE01 cells within the gastroderm with diameters greater than 7 µm, we also found a wide range of diameters for the cells of all three strains, with many SSA03 and SSE01 cells in the gastric cavity or outside the larvae that were smaller than many of the SSB01 or SSE01 cells that were within the gastroderm (figure 4a). Moreover, examination of dissociated cells of Aiptasia strains CC7-SSA01 and CC7-SSB01 showed that individual gastrodermal cells could expand to accommodate two, three or more algal cells (figure 4b; electronic supplementary material, figure S6; videos S3 and S4). Thus, it seems unlikely that either SSA03 or SSE01 is incompatible with the Aiptasia host simply on the basis of excessive cell size (see Discussion).

Figure 4.

Evidence that algal cell size is not a major determinant of host–algal compatibility for Aiptasia larvae and the algal strains examined here. (a) In the experiments of figures 2 and 3, the diameters of individual algal cells outside the larvae, in their gastric cavities, or within their gastrodermal cells in the 1-d samples were measured using ImageJ software in the three-dimensional reconstructions of fluorescence images; measurements were made in the optical section in which the periphery of the cell was in best focus (see Materials and methods, electronic supplementary material, figure S5, and video S1). In each case, ≥120 randomly selected algal cells were measured; means and SEMs are indicated. (b) Epifluorescence (images 1, 2, and 4) or confocal-microscopy (image 3) images of dissociated CC7-SSA01 gastrodermal cells (see Materials and methods) showing that individual host cells can contain multiple algal cells. Red, chlorophyll fluorescence of algae; blue, staining of host cell nuclei by Hoechst 33342; grey, DIC imaging.

3. Discussion

In this study, we have used the experimentally tractable Aiptasia model system to explore several important aspects of the establishment and maintenance of cnidarian–dinoflagellate symbiosis.

(a) . Role of algal photosynthesis

Our results indicate that for the host and algal strains studied here, long-term maintenance of the symbiosis depends on algal photosynthesis. Precise flow-cytometer counts of algal numbers in adult polyps incubated in continuous darkness showed a steady but slow loss of algae that eventually went to completion (as judged by the absence of repopulation after return to a light-dark cycle), even though regular feeding allowed the animals to maintain their biomass. Although it is conceivable that this loss of algae results from the absence of some other light-dependent signal, the observations by multiple investigators that treatment with the photosystem-II inhibitor DCMU also leads to a complete loss of algae [7,17,19,23,26,43,47,49] strongly support the conclusion that the absence of photosynthesis is the relevant factor.

Our observations are consistent with those in most previous studies [18,19,41,42,44–46,48], although Steen [44] did also report that some animals that were seemingly aposymbiotic after prolonged (greater than 4 years) incubation in darkness became repopulated without intentional reinoculation when placed back on a light-dark cycle, possibly from a residual internal population of algae. However, the precise conditions of the long incubation in darkness were not reported, leaving open the possibility that accidental reintroduction(s) of algae had occurred. Also seemingly inconsistent with our results is the report by Jinkerson et al. [51] that Symbiodiniaceae cells are maintained at nearly constant numbers in an Aiptasia host over long periods in the dark despite an inability to proliferate under these conditions. As these investigators used strains nominally identical to those used in our own experiments, we cannot explain the discrepancy, although we do note both that the methods used by Jinkerson et al. [51] to determine algal numbers were less precise than those used here and that the microscopy methods that they used did not have sufficient resolution to determine reliably whether the algae observed were inside or outside the gastrodermal cells.

That the decline in algal numbers in the dark is so slow suggests that algae are lost stochastically in the absence of reproduction, rather than being actively released or digested as can occur upon heat stress [12,16,46,82–86], cold shock [17,87,88], or exposure to certain noxious chemicals such as menthol [89–91]. A lack of active clearance of algae in the dark appears to make teleonomic sense, as it would presumably be more costly for the host in terms of energy and/or limiting nutrients such as nitrogen [92–94] to re-grow its algae each day, or after any transient period of accidental shading, than simply to maintain a population of nonproliferating (and thus presumably metabolically quiescent) algae. A dependence on re-acquisition of compatible algae from the environment would also seem to be a risky strategy. Although we saw no indication of algal reproduction in the dark either in adult anemones or in larvae (see below), it is clear that at least some strains of Symbiodiniaceae can grow heterotrophically in the dark in culture [26]. Thus, it seems quite possible that under some conditions, and/or in some hosts whose lifestyle makes this advantageous (e.g., animals exposed to a pronounced seasonal cycle), heterotrophic algal growth and reproduction might occur in the dark using energy and nutrients provided by the host [44,50,51].

When we explored the initiation of symbiosis in the absence of photosynthesis by exposing aposymbiotic adult polyps to a compatible strain of algae in the dark or in the presence of DCMU, we saw no evidence of colonization. However, the stereomicroscopy used could easily have missed modest numbers of algae in the animals (such as reported by Jinkerson et al. [51]). Thus, we turned to experiments with larvae, where analysis of complete z-stacks of optical sections from confocal microscopy allowed precise counts of the numbers of algae and unequivocal determination for each algal cell of whether it had been taken up into the host gastroderm or was only in the gastric cavity. We found that when photosynthesis was eliminated or reduced by darkness or DCMU, compatible algae were still ingested by the larvae and taken up into their gastrodermal cells, but in smaller numbers than when photosynthesis was allowed. The differences were already apparent after the initial 24 h of incubation with the algae and affected primarily gastrodermal uptake. (Indeed, the effect may have been entirely at this level, as our data did not provide an independent measure of ingestion into the gastric cavity.) Upon further incubation after washing away most uningested algae, the discrepancies were magnified: whereas algal numbers in the gastroderm increased rapidly over 10 days under a 12 h : 12 h light-dark cycle, there was a steady, although slow, decline when incubation was in the dark, similar to that seen with adult anemones. Although we cannot rule out the possibility that some algal reproduction occurred in the dark but was outweighed by a greater loss of algae to the gastric cavity, and then the surrounding seawater, the more parsimonious explanation is that there was simply no algal reproduction, even as algal cells were gradually released from (or digested by) the gastrodermal cells they had entered (or as these gastrodermal cells themselves were sloughed off the epithelial layer).

Taking our results with larvae and adults together, it appears that for the strains used here, compatible algae can be taken up into host gastrodermal cells in the absence of photosynthesis, and are tolerated there by the host, but cannot reproduce and thus are gradually lost over time.

(b) . Multiple modes of discrimination against incompatible algae

Analysis by confocal microscopy also revealed striking differences in the interactions of Aiptasia larvae with the two incompatible algal strains examined here. Cells of strain SSA03 (S. pilosum) were ingested into the gastric cavity in numbers that seemed similar to those of the compatible strain SSB01 (B. minutum), but we did not find a single SSA03 cell within the gastroderm. Our data do not allow us to distinguish between (i) a total absence of phagocytic entry and (ii) entry by phagocytosis followed by very rapid exocytosis (or digestion), although the former possibility seems more parsimonious. It appears possible that the total rejection of SSA03 cells relates to their pilose (hairy) surfaces [95,96], but more focused experiments will be needed to test this possibility.

Cells of strain SSE01 (E. voratum) were also ingested abundantly into the gastric cavity. However, in contrast to SSA03 cells, they did enter the gastroderm from there, although in much smaller numbers than did cells of the compatible strain SSB01. (The initial exposure of larvae to algae was much too short for the large difference in algal cells in the gastroderm at 24 h to be explained solely by the reproduction of the SSB01 cells after internalization.) Wolfowicz et al. [29] have also reported observing small numbers of SSE01 cells within the gastroderm of Aiptasia larvae at 4 days after initial exposure. Thus, it appears that there must be some discrimination, albeit incomplete, against SSE01 cells at the time of phagocytosis. However, in addition, the SSE01 cells that did enter the gastroderm did not appear to proliferate there under a 12 h : 12 h light-dark cycle, and their numbers declined slowly over time, much as when larvae containing SSB01 cells were held in the dark. As strain SSE01 can grow photoautotrophically in culture [26], the cells would presumably also be able to photosynthesize in a host during the light phase of the daily cycle. However, if E. voratum is really a free-living, rather than symbiotic, species in nature, as generally believed [3], the cells may have no mechanism to release photosynthate to a host, and the host may, in turn, fail to provide the algal cells with other nutrients needed for proliferation. The question then would be why the host cells do not either treat the SSE01 cells as particles of food and digest them or egest them rapidly as if they were inert beads [51,63].

(c) . Possible role of excessive cell size in rendering some algal strains incompatible with some hosts

It has been suggested that algal cell size may play a role in symbiont selection by hosts, with larger cell types simply being too big for successful phagocytosis and/or retention [29,63,74]. However, we could find no evidence to support this possibility in our comparison of strains SSB01, SSA03, and SSE01, despite the seemingly very different mechanisms by which the latter two strains are discriminated against by Aiptasia larvae. Moreover, although particle size and shape are known to affect the kinetics of phagocytosis by many cell types, such as mammalian macrophages [75,76], it is also known that at least some phagocytic cells can ingest particles considerably larger than themselves [97]. In this regard, our confocal microscopy also has confirmed and extended earlier reports [88,98–100] that cnidarian gastrodermal cells can expand to accommodate multiple (five or more) algal cells. Taken together, these data suggest that it is unlikely that algal cell size, by itself, plays a major role in the selection of potential Symbiodiniaceae partners by cnidarian hosts.

(d) . Conclusions and future directions

There remain many important questions about the molecular mechanisms of host–algal recognition and discrimination in cnidarian–dinoflagellate symbioses (see Introduction and Rosset et al. [32]). Although expulsion of intact algal cells appears to be the primary mechanism by which symbionts are eliminated from the host in response to heat and/or light stress [17], it remains unclear whether a similarly active mechanism occurs in response to a lack of algal photosynthesis. Given that algal loss under these conditions is a slow process that spans months (see Results), further careful research will be needed to determine whether it represents an active expulsion process or simply host-cell renewal in the absence of algal proliferation. In addition, answering questions about the determinants of initial algal uptake will depend both on a clear understanding of the precise stage(s) at which the discrimination among particular strains actually takes place (which will require additional experimentation at the level of resolution used here) and on the application of genetic methods [101,102] to test rigorously whether particular molecules and interactions really have the roles proposed for them. Among the high-priority targets for gene knockdowns or knockouts are the glucose transporter GLC8 [7,47,49,86] and the sterol transporter NPC2 [11,47]. Determining the roles of these nutrient transporters in initiating or maintaining symbiosis may clarify how host exposure to algae in the light versus in the dark leads to such different outcomes. Another crucial area for future exploration will be the nutritional (and perhaps other) impacts of dinoflagellate–bacterial interactions [103–107], as the full complexity of the microbiome may be critical in mediating cnidarian–dinoflagellate symbioses.

4. Material and methods

(a) . Organisms and culture conditions

We used clonal strains of both anemones and dinoflagellate algae (family Symbiodiniaceae) in this study. Aiptasia strain CC7 (male) is of presumed Florida origin; it harbours a largely or entirely homogeneous population of Symbiodinium linucheae (ITS2 Clade A4) algae [3,17,108]. Aiptasia strain PLF3 (female) is also of presumed southeastern-US origin; it harbours predominantly Breviolum minutum (ITS2 Clade B1) algae [3], as described previously by Grawunder et al. [109], who refer to this Aiptasia strain as F003. Strains CC7 and PLF3 are closely related to each other despite their difference in endogenous algal populations [109]. Aposymbiotic CC7 anemones were generated as described previously [19,26,110,111]. Strains CC7-SSA01 and CC7-SSB01 were generated several years ago by exposing aposymbiotic CC7 animals to SSA01 or SSB01 algae from culture (see below) and subsequently maintaining the symbiotic animals under routine culture conditions with regular testing of algal genotypes. Anemone stocks were maintained routinely at 27°C in artificial seawater (ASW) prepared using Coral Pro Salt (Red Sea, Houston) at 33.5 ppt in deionized water (dH2O) and fed twice per week with freshly hatched Artemia nauplii followed by a water change later the same day. Lighting was on a 12 h : 12 h light-dark cycle using white (4100 K) fluorescent bulbs (Philips ALTO II 25 W) at an irradiance of 25 µmol photons m⁻² s⁻¹ of photosynthetically active radiation as measured using a GMSW-SS quantum metre (Apogee) [26].

Aiptasia larvae were obtained using a spawning protocol developed in our laboratory by S. Perez. Each of several small polycarbonate tanks contained one large (greater than 15 mm oral-disc diameter) CC7 and one large PLF3 anemone in approximately 300 ml of ASW. The tanks were held at 27°C, and the anemones were induced to spawn by exposing them to a light regimen of 25 days of 12 h light : 12 h dark (as described above) followed by 5 days of 16 h light : 8 h ‘dark’ (i.e., white lights off but low-intensity blue actinic LED illumination (453 nm) provided throughout the 8-h period). When spawning occurred, fertilization in the original tank led to the formation of swimming larvae within 24 h; these larvae were rinsed with ASW on a 40-µm mesh filter and transferred to glass bowls, then allowed to develop for 5 days without exposure to symbiotic algae and with daily rinses with ASW, by which point they had well-developed mouths.

The clonal, axenic Symbiodiniaceae strains SSA01 (Symbiodinium linucheae; ITS2 Clade A4), SSB01 (Breviolum minutum; ITS2 Clade B1), SSA03 (Symbiodinium pilosum; ITS2 Clade A2), and SSE01 (Effrenium voratum; ITS2 Clade E) have been described previously [3,17,26,79]. It should be noted that although strain SSB01 was derived from Aiptasia strain H2 [26], which is phylogenetically rather distant from Aiptasia strains CC7 and PLF3 [109], it is also closely related to the algae that are predominant in PLF3 anemones [109].

Algal cultures were grown at 27°C in IMK (ASW containing 0.25 g l⁻¹ Daigo's IMK powder; FujiFilm Wako Chemicals) or IMK + Cas (IMK supplemented with 4 g l⁻¹ casamino hydrolysate; Affymetrix/USB) medium [26,112,113]. Lighting conditions were as described for the anemone stock cultures but at 10 µM photons m⁻² s⁻¹. Stock cultures were maintained in IMK + Cas and passaged once a month by 1 : 10 or 1 : 20 dilution. The concentrations of algal cells in cultures were monitored using a Guava easyCyte HT2 laser flow cytometer (EMD Millipore) as described previously [114].

(b) . Assessment of the effects of DCMU on algal photosynthesis and survival

To assess the effect of DCMU treatment on algal photosynthesis, we measured the maximum quantum yield of photosystem II (PSII) by chlorophyll-fluorescence emission after a 10-min dark adaptation according to the equation F_v/F_m = (F_m – F₀)/F_m [115]. Cells were collected by centrifugation at 200 g for 5 min and resuspended at 10⁷ cells ml⁻¹ in ASW containing 20% Ficoll to avoid cell settling. Measurements of Fv/Fm were then made using a JTS-10 spectrophotometer (Bio-Logic; Grenoble, France). To assess the effect of DCMU treatment on cell viability, samples of the treated cultures were spread at appropriate concentrations on agar medium containing Marine Broth (Millipore Sigma #76448) plus glucose at 37.4 g l⁻¹. The plates were maintained under a 12 h : 12 h light-dark cycle for 2 weeks at 27°C, and the numbers of colonies were counted and compared to those obtained from control cultures without DCMU. Cells from the DCMU-treated cultures were also imaged using a Nikon Eclipse E600FN microscope with a 100x/1.40 NA oil-immersion objective, a Hamamatsu C4742-98-24NR ORCA II CCD camera, and MetaMorph 7.5.2 control software (Molecular Devices). Colour images were created by using red (ET632/60m), green (ET535/50m), and blue (ET455/50m) emission filters (all from Chroma) in the transmitted-light path to record single-colour images, which were then merged and colour-balanced using the ImageJ 1.52e image-processing software.

(c) . Bleaching of adult anemones in the dark

To assess the effect of a loss of photosynthesis (by incubation in the dark) on maintenance of the symbiosis, separate experiments followed (i) CC7 anemones for 30 days, (ii) CC7-SSB01 anemones for 30 days, and (iii) CC7-SSB01 anemones for 60 days. In each such experiment, three light-excluding black polycarbonate tanks contained test anemones, whereas one ordinary (transparent) polycarbonate tank contained control anemones incubated under the light conditions described above. For each tank, feeding with Artemia larvae and water changes (either later the same day or on the day following feeding) were done weekly. For each tank, five or six anemones were removed on Day 0 and analysed individually for numbers of algae (by flow cytometry) and total protein (by BCA assay) as described previously [114]. At intervals thereafter, five to seven anemones were removed from each tank and analysed similarly. In a separate set of long-term experiments, one black polycarbonate tank containing CC7 anemones and two containing CC7-SSB01 anemones were incubated for 470 days, with feeding and water changes as above, and examined intermittently for the presence of algae using a Leica MZ16 FA fluorescence stereomicroscope. At 302 and 470 days, some animals were returned to a 12 h : 12 h light-dark cycle and monitored for possible repopulation using the fluorescence stereomicroscope.

(d) . Algal colonization of adult and larval anemones

To prepare algal cells for colonization of adult or larval Aiptasia, approximately 2 × 10⁷ cells from a growing stock culture (at approx. 1.5 × 10⁶ cells ml⁻¹) were collected by centrifugation at 100 g for 5 min at room temperature, washed once with autoclave-sterilized ASW, transferred to 50 ml of IMK medium in a 250-ml flask, and grown without agitation or a change of medium for 2 weeks in continuous white light (as above).

For each test of colonization of adult anemones, eight small (2–5 mm in oral-disc diameter) aposymbiotic CC7 anemones that had not been fed for 3–4 days were transferred from a stock tank to a new polycarbonate tank containing 300 ml of ASW, allowed to settle and attach over a period of 24 h (12 h light, then 12 h dark) at 27°C, and then exposed for 24 h at 27°C to approximately 10⁴ cells ml⁻¹ of strain SSB01 that had been prepared as described above. These 24-h exposures took place (i) in ASW under 12 h light followed by 12 h dark, (ii) in ASW in the dark (black polycarbonate tanks), (iii) in ASW containing 10 µM 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU, Sigma-Aldrich #D2425; added from a 25 mM stock in ethanol), and (iv) in ASW containing 25 µM DCMU; the DCMU experiments both also used a 12 h light : 12 h dark regimen. The anemones were then collected from each tank, washed twice with ASW, and transferred to a new tank with fresh ASW (no additives) for the duration of the experiment, during which lighting was on a 12 h : 12 h light-dark cycle, except for the ‘continuous dark’ experiment, which used the black tanks throughout. The anemones were not fed during the experiment, and the ASW was changed every 2–3 days to avoid fouling of the tanks.

For each test of colonization of anemone larvae, approximately 300 5-day-old larvae were placed in 500 µl of ASW or (where noted) ASW containing 10 or 25 µM DCMU (added as described above) in a well of a 24-well polypropylene plate. Algae of strain SSB01, SSA03 or SSE01, prepared as described above, were then added to a concentration of approximately 10⁵ cells ml⁻¹ and gently mixed with the larvae. The mixture was then incubated for 24 h at 27°C in continuous dark (where indicated; achieved by placing the 24-well plate in a sealed black polycarbonate tank) or 12 h light followed by 12 h dark (all other cases); lighting was white light (as above) at 25 µmol photons m⁻² s⁻¹. After 24 h, the larvae were rinsed free of unassociated algae by sieving with a 40-µM mesh and then washing them into a new well with 500 µl of fresh ASW. This step also removed the DCMU where it had been present. Incubation was then continued at 27°C in continuous darkness (where indicated) or under 12 h light : 12 h dark (all other cases; lighting as described above); the ASW was replaced daily by rinsing with the 40-µm mesh. At intervals, approximately 100 larvae from each well were collected with a pipette and fixed by incubation in 0.02% glutaraldehyde plus 4% paraformaldehyde in ASW at 4°C for 5 min followed by incubation in 4% paraformaldehyde in ASW at 4°C for 1 h. Larvae were then washed three times (for 5 min each) with phosphate-buffered saline (PBS) and five times (for 30 min each) with PBS containing 0.2% Triton X-100 (PBT), then stained with Alexa Fluor 488-phalloidin (Thermo Fisher Scientific, cat. no. A12379; 1:40 in PBT) for 20 min at room temperature and washed three more times with PBS for 5 min each.

(e) . Microscopic evaluation of colonization success

To assess colonization of adult anemones, the animals were imaged at intervals in both transmitted-light and fluorescence modes using a Leica MZ16 FA fluorescence stereomicroscope and a Leica DFC 500 digital camera.

To assess colonization of larvae, the fixed larvae (see above) were mounted on glass slides in 3 : 1 glycerol:PBS and imaged using a Leica SP8 laser-scanning confocal microscope equipped with a Leica 40x/1.25 NA oil-immersion objective, Hamamatsu sCMOS camera, and Leica LAS X software. z-stacks were collected in the channels for Alexa Fluor 488-phalloidin (495 nm excitation, 518 nm emission) and chlorophyll autofluorescence (561 nm excitation, 637 nm emission). Moving through a z-stack of high-resolution, 0.3-µm optical sections, the numbers of algae within individual larvae were quantified manually, with each algal cell identified by its chlorophyll autofluorescence. Localization of each algal cell within the gastric cavity or gastroderm of the larva was scored using the phalloidin staining, which labels the boundaries of gastrodermal cells. The number of larvae having no algae in either location was also recorded.

To determine algal cell sizes, we measured cell diameters using ImageJ 1.51 g software. We first set a scale under the ‘Analyze’ menu, based on the scale bar generated from confocal images of larvae containing algae, and then used this calibrated scale to determine individual cell diameters. Going through the z-stack of images, each cell was measured in the optical section in which its periphery was in focus (see electronic supplementary material, figure S5 and video S1). To avoid possible bias, for each larva in which any algal diameters were to be measured, we measured all the algal cells within, on the surface of, or near to the larva.

(f) . Preparation and imaging of dissociated anemone cells

Symbiotic CC7-SSA01 and CC7-SSB01 anemones with oral-disc diameters ≤1 mm were dissociated by rotating them in 1.5-ml microcentrifuge tubes containing 200 µl of dissociation-and-staining solution for 20–60 min at room temperature in the dark. The solution used was based on that described by Hu et al. [116]; it contained 50 µg ml⁻¹ Liberase (Sigma-Aldrich, #05401119001), 40 mg ml⁻¹ L-cysteine, and 10 µg ml⁻¹ Hoechst 33342 HCl (Cayman Chemical, #15547) in ASW. Every 10–20 min, the dissociating anemones were pipetted in and out of a 200-µl pipette tip to break up tissue clumps. The dissociated cells were fixed by adding 40 µl fetal bovine serum (FBS; Sigma-Aldrich #F2442) and 500 µl 3.7% paraformaldehyde in ASW (diluted from 16% paraformaldehyde solution; Electron Microscopy Sciences #15710) and incubating in the dark at 4°C for 30 min with mixing by inversion every 10 min. The cells were then collected by centrifugation at 4000 g for 5 min at 4°C and resuspended in 200 µl ASW + 20 µl FBS before loading either onto slides and sealing coverslips with lanolin or into the wells of a 4-well Nunc Lab-Tek Chambered Coverglass (ThermoFisher, #155383). Cells were then imaged immediately either with a Nikon Eclipse microscope (model E600FN) equipped with a mercury-lamp light source and a Hamamatsu camera (model C4742-98-24NR) using MetaMorph software (MDS Analytical Technologies, version 7.6.0.0), or with a Nikon Eclipse Ti microscope (model TI-DH) equipped with a Yokogawa CSU-W1 spinning disc confocal scanning unit (model CSUW1CA-D), 405-nm and 488-nm laser light sources, and an Andor Ixon 3 camera (model DU-897D-C00-#BV) using NIS-Elements software (Nikon, AR version 4.30.02). All images included DIC or phase contrast, Hoechst 33342 fluorescence (405/450) and chlorophyll autofluorescence (488/600 LP) channels. Magnifications of 200–600× were used. In some cases, z-stacks with 1-µm steps were acquired. Images were processed using Fiji/ImageJ v2.1.0/1.53c.

Ethics

This work did not require ethical approval from a human subject or animal welfare committee.

Data accessibility

All data are included in the manuscript and electronic supplemental material [117].

Declaration of AI use

We have not used AI-assisted technologies in creating this article.

Authors' contributions

C.T.: conceptualization, formal analysis, investigation, methodology, writing—original draft, writing—review and editing; G.R.: conceptualization, formal analysis, investigation, methodology, writing—original draft, writing—review and editing; P.C.: conceptualization, investigation, methodology, writing—review and editing; C.K.: conceptualization, investigation, methodology, writing—review and editing; M.P.: investigation, methodology; S.C.: investigation, methodology; A.G.: conceptualization, investigation, methodology, resources, writing—review and editing; J.P.: conceptualization, funding acquisition, investigation, methodology, resources, supervision, writing—original draft, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

Funding for this study was provided by the Gordon and Betty Moore Foundation (grant nos 2629.01), NSF-IOS EDGE Award no. 1645164, and Simons Foundation (grant no. LIFE 336932). A Faculty Research Grant from the University of San Diego was awarded to C.T. to support completion of this manuscript.